12
Theoretical Analysis Drawing pictures during learning from scientific text: testing the generative drawing effect and the prognostic drawing effect Annett Schmeck a, *, Richard E. Mayer b , Maria Opfermann a , Vanessa Pfeiffer c , Detlev Leutner a a Department of Instructional Psychology, University of Duisburg-Essen, Berliner Platz 6-8, D-45127 Essen, Germany b Psychology, University of California at Santa Barbara, Santa Barbara, CA 93106, USA c Didactics of Biology, University of Duisburg-Essen, Universitätstraße 2, D-45117 Essen, Germany ARTICLE INFO Article history: Available online 23 July 2014 Keywords: Text comprehension Drawing Generative learning activities Generative drawing effect Prognostic drawing effect A B ST R AC T Does using a learner-generated drawing strategy (i.e., drawing pictures during reading) foster students’ engagement in generative learning during reading? In two experiments, 8th-grade students (Exp. 1: N = 48; Exp. 2: N = 164) read a scientific text explaining the biological process of influenza and then took two learning outcome tests. In Experiment 1, students who were asked to draw pictures during reading (learner- generated drawing group), scored higher than students who only read (control group) on a multiple- choice comprehension test (d = 0.85) and on a drawing test (d = 1.15). In Experiment 2, students in the learner-generated drawing group scored significantly higher than the control group on both a multiple- choice comprehension test (d = 0.52) and on a drawing test (d = 1.89), but students who received author- generated pictures in addition to drawing or author-generated pictures only did not. Additionally, the drawing-accuracy scores during reading correlated with comprehension test scores (r = .623, r = .470) and drawing scores (r = .620, r = .615) in each experiment, respectively. These results provide further evi- dence for the generative drawing effect and the prognostic drawing effect, thereby confirming the benefits of the learner-generated drawing strategy. © 2014 Elsevier Inc. All rights reserved. 1. Introduction Suppose you want to enable students to study a scientific text by themselves for deep level understanding. In this case, you will have to ensure that students engage in generative learning pro- cesses during reading, such as organizing material into coherent mental representations, and integrating the representations with each other and with relevant knowledge activated from long-term memory (de Jong, 2005; Mayer, 2004, 2009; Wittrock, 1990). A pos- sible way to accomplish this goal is to encourage students to use a learner-generated drawing strategy (Alesandrini, 1984; Schwamborn, Mayer, Thillmann, Leopold, & Leutner, 2010; van Meter & Garner, 2005), in which they receive a text to read and are instructed to draw pictures that reflect the main elements and relations described in the text. The goal of the present study is to examine a generative drawing effect (i.e., engaging in appropriate drawing activities during learning from text improves performance on tests of learning outcomes) and a prognostic drawing effect (i.e., the quality of drawing during learning from text predicts performance on subsequent tests of learning outcomes). 1.1. Theoretical framework for the learner-generated drawing strategy A straightforward way to encourage students to use a learner- generated drawing strategy when learning from verbal instruction is to ask them to generate an external visual representation of a to- be learned content. The drawing that is generated has a representational quality, similar to the characteristics of a repre- sentational illustration (cf., Alesandrini, 1984; van Meter & Garner, 2005). By representational, we mean that learners make drawings which are intended to show what depicted objects look like (Carney & Levin, 2002). This requirement excludes nonrepresentational graphic constructions such as diagrams and concept maps. Thus, our definition of drawing is that the learner creates a visual rep- resentation intended to depict what is described in text. Drawing can be seen as a learning strategy intended to influ- ence how learners process information during learning (Pashler et al., 2007; Weinstein & Mayer, 1986). By drawing, learners are no longer passive consumers of information and knowledge; they are * Corresponding author. Address: Annett Schmeck, University of Duisburg-Essen, Berliner Platz 6-8, D-45127 Essen, Germany. Fax: +49 2011834350. E-mail address: [email protected] (A. Schmeck). http://dx.doi.org/10.1016/j.cedpsych.2014.07.003 0361-476X/© 2014 Elsevier Inc. All rights reserved. Contemporary Educational Psychology 39 (2014) 275–286 Contents lists available at ScienceDirect Contemporary Educational Psychology journal homepage: www.elsevier.com/locate/cedpsych

Drawing pictures during learning from scientific text: testing the generative drawing effect and the prognostic drawing effect

  • Upload
    detlev

  • View
    214

  • Download
    1

Embed Size (px)

Citation preview

Theoretical Analysis

Drawing pictures during learning from scientific text testing thegenerative drawing effect and the prognostic drawing effectAnnett Schmeck a Richard E Mayer b Maria Opfermann a Vanessa Pfeiffer cDetlev Leutner a

a Department of Instructional Psychology University of Duisburg-Essen Berliner Platz 6-8 D-45127 Essen Germanyb Psychology University of California at Santa Barbara Santa Barbara CA 93106 USAc Didactics of Biology University of Duisburg-Essen Universitaumltstraszlige 2 D-45117 Essen Germany

A R T I C L E I N F O

Article historyAvailable online 23 July 2014

KeywordsText comprehensionDrawingGenerative learning activitiesGenerative drawing effectPrognostic drawing effect

A B S T R A C T

Does using a learner-generated drawing strategy (ie drawing pictures during reading) foster studentsrsquoengagement in generative learning during reading In two experiments 8th-grade students (Exp 1 N = 48Exp 2 N = 164) read a scientific text explaining the biological process of influenza and then took twolearning outcome tests In Experiment 1 students who were asked to draw pictures during reading (learner-generated drawing group) scored higher than students who only read (control group) on a multiple-choice comprehension test (d = 085) and on a drawing test (d = 115) In Experiment 2 students in thelearner-generated drawing group scored significantly higher than the control group on both a multiple-choice comprehension test (d = 052) and on a drawing test (d = 189) but students who received author-generated pictures in addition to drawing or author-generated pictures only did not Additionally thedrawing-accuracy scores during reading correlated with comprehension test scores (r = 623 r = 470) anddrawing scores (r = 620 r = 615) in each experiment respectively These results provide further evi-dence for the generative drawing effect and the prognostic drawing effect thereby confirming the benefitsof the learner-generated drawing strategy

copy 2014 Elsevier Inc All rights reserved

1 Introduction

Suppose you want to enable students to study a scientific textby themselves for deep level understanding In this case you willhave to ensure that students engage in generative learning pro-cesses during reading such as organizing material into coherentmental representations and integrating the representations witheach other and with relevant knowledge activated from long-termmemory (de Jong 2005 Mayer 2004 2009 Wittrock 1990) A pos-sible way to accomplish this goal is to encourage students to use alearner-generated drawing strategy (Alesandrini 1984 SchwambornMayer Thillmann Leopold amp Leutner 2010 van Meter amp Garner2005) in which they receive a text to read and are instructed to drawpictures that reflect the main elements and relations described inthe text The goal of the present study is to examine a generativedrawing effect (ie engaging in appropriate drawing activities duringlearning from text improves performance on tests of learning

outcomes) and a prognostic drawing effect (ie the quality of drawingduring learning from text predicts performance on subsequent testsof learning outcomes)

11 Theoretical framework for the learner-generateddrawing strategy

A straightforward way to encourage students to use a learner-generated drawing strategy when learning from verbal instructionis to ask them to generate an external visual representation of a to-be learned content The drawing that is generated has arepresentational quality similar to the characteristics of a repre-sentational illustration (cf Alesandrini 1984 van Meter amp Garner2005) By representational we mean that learners make drawingswhich are intended to show what depicted objects look like (Carneyamp Levin 2002) This requirement excludes nonrepresentationalgraphic constructions such as diagrams and concept maps Thusour definition of drawing is that the learner creates a visual rep-resentation intended to depict what is described in text

Drawing can be seen as a learning strategy intended to influ-ence how learners process information during learning (Pashleret al 2007 Weinstein amp Mayer 1986) By drawing learners are nolonger passive consumers of information and knowledge they are

Corresponding author Address Annett Schmeck University of Duisburg-EssenBerliner Platz 6-8 D-45127 Essen Germany Fax +49 2011834350

E-mail address annettschmeckuni-duede (A Schmeck)

httpdxdoiorg101016jcedpsych2014070030361-476Xcopy 2014 Elsevier Inc All rights reserved

Contemporary Educational Psychology 39 (2014) 275ndash286

Contents lists available at ScienceDirect

Contemporary Educational Psychology

journal homepage wwwelseviercom locate cedpsych

actively involved in the cognitive processes of selecting organiz-ing and integrating the information to be learned Thus learner-generated drawing is a cognitive learning strategy that is aimed tofoster learning from text and if used adequately drawing can in-crease learning outcomes (Ainsworth Prain amp Tytler 2011Alesandrini 1984 van Meter amp Garner 2005)

The processes underlying drawing are described in van Meterand Garnerrsquos (2005) generative theory of drawing construction(GTDC) which is based on Mayerrsquos (2005) model of multimedialearning It is assumed that learners benefit from using the drawingstrategy as drawing requires them to engage in generative learn-ing processes during reading First learners select the relevant keyinformation from the text Second the selected key information isorganized to build up an internal verbal representation of the textinformation Third learners construct an internal nonverbal (visual)representation of the text information and connect it with the verbalrepresentation and with relevant prior knowledge To construct thevisual representation which is the basis for the external drawingthe learner has to rely mainly on the verbal representation and thuslearner-generated drawing demands an integration of the verbal andnonverbal representation

Additionally van Meter and Garner (2005) describe metacognitiveprocesses fostered by the drawing activity ldquoAttempts at construct-ing the nonverbal representation can send learners back to eitherthe verbal representation or the text as difficulties building the in-ternal image are encounteredrdquo (van Meter amp Garner 2005 p 317)That is as the drawing process itself is not linear metacognitive pro-cesses of monitoring and regulation are stimulated by drawing (cfvan Meter 2001 van Meter Aleksic Schwartz amp Garner 2006)

12 Empirical framework for the learner-generated drawing strategy

Following the GTDC (cf van Meter amp Garner 2005) the drawingstrategy is beneficial as it fosters deep cognitive processing includ-ing organizing and integrating material (which can be calledgenerative processing Mayer 2009) as well as metacognitive self-monitoring and regulation processes Research on drawing howeverhas produced somewhat mixed results (see Alesandrini 1984 vanMeter amp Garner 2005 for overviews) in which some studies re-ported positive effects of drawing on text comprehension (egAlesandrini 1981 Hall Bailey amp Tillman 1997 Leopold amp Leutner2012 Lesgold DeGood amp Levin 1977 Lesgold Levin Shimron ampGuttman 1975 Schwamborn et al 2010 van Meter 2001 vanMeter et al 2006) whereas others did not (eg Leutner Leopoldamp Sumfleth 2009 Rasco Tennyson amp Boutwell 1975 Tirre Manelisamp Leicht 1979) Benefits of drawing appear to be related to thequality of studentsrsquo drawings during learning Students who producehigh-quality drawings during reading tend to score better onposttests of learning outcome than do students who produce low-quality drawings during reading (eg Greene 1989 Hall et al 1997Leopold 2009 Lesgold et al 1975 1977 Schwamborn et al 2010van Meter 2001 van Meter et al 2006)

121 Effectiveness of learner-generated drawingsFollowing van Meter and Garner (2005) reasons for the mixed

empirical results concerning drawing can be seen attributed to thetype of test used for assessing learning outcomes as well as in theform of support that assists learners in the drawing process Firstbenefits of drawing are more likely to be revealed on tests that assesshigher-order knowledge of to-be learned content for example testson comprehension and transfer (eg Alesandrini 1981 Leopold ampLeutner 2012) or problem solving (van Meter 2001 van Meter et al2006) Leutner et al (2009) for example found no positive effectof drawing compared with a control group on a multiple choice teston factual knowledge Leopold and Leutner (2012) however showedsuperior effects of the drawing strategy on transfer test perfor-

mance van Meter et al (2006) accordingly found no effects ofdrawing activity on a multiple choice recognition test however stu-dents in the drawing group scored significantly higher on a problem-solving test With regard to the GTDC (van Meter amp Garner 2005)it seems that benefits of drawing can be found if the learningoutcome test complies with characteristics of the verbal and non-verbal representations which are generated by drawing

Second positive effects of drawing often appear under the con-dition that instructional support is provided to constrain andstructure the drawing activity (eg Lesgold et al 1975 1977Schwamborn et al 2010 van Meter 2001 van Meter et al 2006)That is drawing is more effective when the learnersrsquo generation ofthe drawing is assisted by some kind of additional information vanMeter (2001) and van Meter et al (2006) for example showed thatthe provision of author-generated pictures after drawing en-hanced the benefits of the drawing strategy By comparing their owndrawing with a provided one learners get to know what theirdrawing should look like and this might lead them back to revisetheir own drawing and thus their mental model Following the GTDC(van Meter amp Garner 2005) this should improve comprehensionLesgold et al (1975) in turn supported first grade students withcutout figures and instructed them to organize these into an accu-rate pictorial representation while listening to a prose story Thislearner-generated illustration activity facilitated prose learning asindicated by higher recall of story propositions only when stu-dents were given the correct pieces for the illustration or had theillustration done for them When students had to select the piecesfor each illustration out of a pool of cutouts the learner-generatedillustration activity had either a negative or no effect (cf Lesgoldet al 1977) Following these results Schwamborn et al (2010) pro-posed that a pure unsupported drawing instruction might bear therisk that managing the mechanics of drawing itself is difficult forthe learners resulting in insufficient remaining capacity for makingsense of the text through generative processes of organization andintegration which might diminish the benefits of drawing definedby van Meter and Garnerrsquos GTDC To counter this risk in the studyof Schwamborn et al (2010) students in the drawing groups re-ceived baseline instructional support while learning a lesson onwashing which provided them with a drawing prompt that in-cluded a legend showing all the relevant elements for drawing anda partly pre-drawn background for their paper-pencil based draw-ings That is students could use the presented elements as prototypesfor their own drawings and integrate them by pencil in the givenpre-drawn backgrounds Results showed that students who wereinstructed to generate drawings during learning scored signifi-cantly higher on the subsequent comprehension tests than studentswho only read the text

Using cutout-figures (cf Lesgold et al 1975 1977) or a drawingprompt (cf Schwamborn et al 2010) during drawing seems toprovide sufficient constraints and leave enough cognitive capaci-ties for learners to benefit from the drawing strategy Thus cognitiveprocessing including selecting organizing and integrating materi-al should be encouraged resulting in an improved mental modelwhich in turn should improve comprehension (cf GTDC van Meteramp Garner 2005)

In line with the GTDC and the reported results derived from re-search on drawing Schwamborn et al (2010) proposed a generativedrawing effect that is students gain a better understanding of a sci-entific text when they are asked to draw illustrations representingthe content of each paragraph they read This work highlights theimportance of drawing support such as the provision of drawingsof all key elements and a background for the drawing

122 Quality of learner-generated drawingsPrevious studies that measured the quality of studentsrsquo draw-

ings during learning all showed positive correlations between

276 A Schmeck et alContemporary Educational Psychology 39 (2014) 275ndash286

the quality of studentsrsquo drawings during learning and their learn-ing outcomes (eg Greene 1989 Hall et al 1997 Lesgold et al1975 1977 Schwamborn et al 2010 van Meter 2001 van Meteret al 2006) The quality of learner-generated drawing is also re-ferred to as the drawing accuracy (eg van Meter 2001 van Meteret al 2006) and is defined as ldquothe degree to which completeddrawings resemble the represented object(s)rdquo (van Meter amp Garner2005 p 299) In a study by Hall et al (1997) for example collegestudents learning a mechanics lesson with the instruction to drawproduced better learning outcomes on a transfer test than a textonly control group but only if they produced higher quality draw-ings In the study by Schwamborn et al (2010) students who wereable to generate high accuracy drawings scored significantly higheron learning outcome tests than did those who generated loweraccuracy drawings In addition to this Schwamborn and col-leagues also found that the quality of the generated drawings duringlearning correlated positively with the comprehension test scoresBased on these results Schwamborn et al (2010) proposed the prog-nostic drawing effect Students who produce high-quality drawingsduring reading a scientific text tend to score better on posttestsof learning outcome than do students who produce low-qualitydrawings during reading

13 Overview of the experiments

Research on drawing ndash ie the generative and the prognosticdrawing effect ndash is promising however at the present time thereis a need for a more solid evidence base and for a closer examina-tion of theoretical issues First the generalizability is limited atthis point as replication studies using learning outcome tests thatare sensitive to the underlying process of drawing as well as newlearning materials other than the washing lesson (eg Schwambornet al 2010) or the birds wing (van Meter 2001) are yet missingIn their report for the US National Research Council entitledScientific Research in Education Shavelson and Towne (2002 p 4)for example highlighted the need to ldquoreplicate and generalizeacross studiesrdquo as one of the six essential scientific principles ofeducational research It has to be mentioned at this point that whengeneralizing results to new domains or lessons one should care-fully consider whether these are comparable at all In ourexperiments we aimed at generalizing results by Schwamborn et al(2010) who worked with a science text explaining the causalsteps regarding the chemistry of washing to a new lesson that ishowever comparable in that the text we used also describedcausal steps of a process in this case regarding the infection withinfluenza and regarding the immune response That is althoughthere were differences between the two domains (chemistry versusbiology) the lessons showed structural similarities and thusallow for comparing results and drawing conclusions regardinggeneralizability

Second research on drawing indicates that some form of supportis needed to assist learners during drawing Schwamborn et al(2010) for example introduced a drawing prompt as helpful supportfor learners to benefit from drawing They proposed that the re-sulting positive drawing effect is based on studentsrsquo engagementin generative learning activities during reading due to drawing (con-sistent with the GTDC of van Meter amp Garner 2005 see also de Jong2005 Mayer 2004 2009 Wittrock 1990) However the results re-ported by Schwamborn et al (2010) might rather reflect amultimedia effect (Mayer 2005 2009) than the proposed drawingeffect as the learning lesson used ndash a scientific text and a drawingprompt consisting of pictorial elements and backgrounds ndash createda multimedia lesson In other words the results of Schwamborn et al(2010) might not be due to the drawing activity but rather due tothe multimedia effect that students ldquolearn better from words andpictures than from words alonerdquo (Mayer 2009 p 223) In this case

the words are presented in the lessons and pictures are generatedby the students so a control group that receives author-generatedpictures is warranted

Third research on drawing mostly used only one way to supportthe drawing strategy at a specific time That is instructionalsupport was added during learning (ie by using cut-out figures ora drawing prompt (cf Lesgold et al 1975 1977 Schwamborn et al2010) or after learning (ie by providing pictures van Meter 2001van Meter et al 2006) Less is known about whetheradding instructional support not only during learning but also afterlearning can additionally enhance the benefits of the drawingstrategy

Fourth research on drawing should include motivational andcognitive aspects that may have an impact on the effectiveness ofthe learner-generated drawing strategy Studentsrsquo current motiva-tion for example is a one condition for successful learning A studentfor example who has low motivation to learn may invest lesseffort in learning than students who are highly motivated to learn(cf Vollmeyer amp Rheinberg 2000) Studentsrsquo spatial ability maybe a further condition for successful learning when workingwith visualizations (cf Houmlffler 2010 Houmlffler Schmeck amp Opfermann2013) A high-spatial-ability student for example may haveadvantages in learning with visualizations compared with a low-spatial-ability student That is preexisting motivational and cogni-tive differences between students before learning might havean influence on the learning outcome and thus should becontrolled

In addition recent research has shown that not only experi-mental conditions (such as the kind of picture) and the abovementioned ldquoclassicalrdquo covariates can have an impact on how suc-cessful learning takes place but that these effects can be mediatedby the amount of mental effort someone invests while learning orworking on a lesson and by how difficult someone perceives adomain or lesson to be (cf Leutner et al 2009 SchwambornThillmann Opfermann amp Leutner 2011) These aspects of cogni-tive load (invested mental effort and perceived task difficulty) werethus included as additional variables in our studies as well Thuswe conducted the following two experiments using a science textexplaining the biological process of influenza In Experiment 1 weimplemented an experimental drawing condition and a reading onlycontrol condition in order to determine how both the generativeand the prognostic drawing effect would extend to a new contextAnalogous to the study of Schwamborn et al (2010) students inthe drawing condition received a baseline instructional support bymeans of a drawing prompt that included a legend showing all therelevant elements for drawing and a partly pre-drawn back-ground for their drawing (as shown in Fig 1)

In Experiment 2 we again implemented an experimental drawingcondition and a reading only control condition and we addition-ally implemented author-generated pictures in order to test whetherthe generative drawing effect was caused by the simple presenceof illustrations rather than the generation of illustrations That iswe implemented a text plus picture condition (which we called theauthor-generated picture condition) to test whether the reportedgenerative drawing effect of Schwamborn et al (2010) is based onstudentsrsquo engagement in generative learning activities during readingrather than on the pictorial representations given by the drawingprompt In addition we implemented a drawing plus picture con-dition (in which students both draw and are given a picture) to testwhether the reported generative drawing effect can be enhancedby instructing students to compare their own drawing with anauthor-generated picture In short we tested whether combiningdifferent forms of support to the drawing strategy additionally en-hances the benefits of the drawing strategy In both experimentslearning outcome tests that are sensitive to the underlying processof drawing were used

277A Schmeck et alContemporary Educational Psychology 39 (2014) 275ndash286

2 Experiment 1

21 Participants and design

Forty-eight German eighth graders in higher track secondaryschools participated in this study The mean age was 137 years(SD = 06) and there were 22 girls and 26 boys The study was basedon a between-subjects design with two levels of text learning(learner-generated drawings and control) as the single factor Twenty-

four students served in the control group and 24 served in thedrawing group

22 Materials

All materials were paper-pencil based The materials consistedof five adjunct questionnaires two learning booklets two cogni-tive load rating scales and two posttests The five adjunctquestionnaires were intended to determine whether the groups were

Fig 1 Screenshot of the drawing prompt for the first paragraph in the drawing versions of the learning booklet Note Translated from the German original

278 A Schmeck et alContemporary Educational Psychology 39 (2014) 275ndash286

equivalent on basic characteristics They included a participant ques-tionnaire a comprehension pretest a spatial ability test and amotivation questionnaire The participant questionnaire solicited in-formation concerning the studentsrsquo age and sex The comprehensionpretest consisted of 25 multiple-choice items and was intended toassess studentsrsquo prior-knowledge of information covered in the textStudentsrsquo spatial ability was measured with a 10 multiple-choicepaper-folding items taken from a battery of cognitive tests devel-oped by Ekstrom French and Harman (1976) The motivationquestionnaire assessed studentsrsquo current motivation for doing thelearning task after reading the instructions before the lesson It con-sisted of nine items from the challenge and interest subscales of theQuestionnaire on Current Motivation (QCM) developed by RheinbergVollmeyer and Burns (2001) Cognitive load by means of investedmental effort was measured using the 7-point subjective rating scaledeveloped by Paas (1992) which ranges from (1) very low mentaleffort to (7) very high mental effort Cognitive load by means of per-ceived task difficulty was measured using the 7-point subjective ratingscale developed by Kalyuga Chandler and Sweller (1999) whichranges from (1) very easy to (7) very difficult These subjective mea-sures have been criticized for assessing cognitive load with only singleitems (eg Bruumlnken Plass amp Leutner 2003) However several studiesshowed the effectiveness of the rating scale by showing that the vari-ation in learnersrsquo cognitive load ratings depended on variations intask complexity or instructional design (for overviews see PaasTuovinen Tabbers amp Van Gerven 2003 Van Gog amp Paas 2008) Inthis regard Sweller Ayres and Kalyuga (2011) conclude that ldquothesimple subjective rating scale regardless of the wording used (mentaleffort or difficulty) has perhaps surprisingly been shown to be themost sensitive measure available to differentiate the cognitive loadimposed by different instructional proceduresrdquo (p 74) For that reasonand due to the economic applicability we decided to use this kindof cognitive load measurement while acknowledging the limita-tions of a short self-report instrument

The two learning booklets each included a science text on thebiology of the influenza The text explained the causal steps re-garding an infection with influenza and regarding the immuneresponse which is an unfamiliar subject for eighth graders in highertrack secondary schools due to the German curriculum The text con-sisted of approximately 850 words (in German) and was divided intoseven paragraphs (as shown in Table 1)

The drawing version of the booklet contained seven pairs of facingpages with a text paragraph on the left page and a two-part drawingprompt on the right page The first part of the drawing prompt in-cluded a legend showing all the relevant elements (in total eightelements) for drawing a picture for that text paragraph (as shownin the top of Fig 1) The second part of the drawing prompt in-cluded a partly pre-drawn background for studentsrsquo drawing (asshown in the bottom of Fig 1) Overall students had to make sevendrawings ie one drawing to each paragraph

The control version of the learning booklet contained four pairsof facing pages with one of the seven text paragraph on each page

Students in both groups learned with exactly the same text mate-rial To make sure that students in the control group learnedwith the same amount of information as students in the drawinggroup all elements of the drawing prompt as well as the spatialrelations between these elements were also described in the sciencetext

The two posttests intended to assess the learning outcomes werea comprehension posttest and a drawing posttest The comprehen-sion posttest (Cronbachrsquos alpha = 083) consisted of 25 multiple-choice items (the same items as in the comprehension pretest) andwas intended to assess studentsrsquo comprehension of the factual andconceptual information covered in the text as well as their abilityto transfer what was presented to new situations An item exampleis ldquoT-helper cells do not only recognize viruses but also agents thatare extraneous to the body Which medication would you admin-ister to a patient who has received a new kidney (a) a medicinethat suppresses the immune response of the body (b) a medicinethat activates the immune response of the body (c) a medicine thatcontains antigens or (d) a medicine that contains blood of the kidneydonorrdquo [(a) is the correct answer] The drawing test (Cronbachrsquosalpha = 081) was intended to assess studentsrsquo comprehension of theconceptual information presented in the science text by means ofdrawing That is students had to reproduce the main ideas givenin the text by drawing It consisted of three drawing items in whichstudents were asked to draw sketches depicting key concepts of thetext and their spatial relations An item example for the drawingtest is ldquoHow does an influenza virus invade a cell and how is it re-producedrdquo The science text the drawing prompt and the learningoutcome tests were constructed by the first author in cooperationwith a biology teacher The materials were adapted fromSchwamborn et al (2010) however using another science domainand including measures of individual learning times and cognitiveload

23 Procedure

Participants were tested in the schoolsrsquo classrooms Within theirclasses they were randomly assigned to one of the two groupsGroups were tested in separate classrooms in order to insure thatstudents in the drawing group did not feel rushed when studentsin the control group completed the task early Each student wasseated at an individual desk First students were given the partic-ipant questionnaire and the comprehension pretest to complete attheir own rate Second students filled in the paper-folding test witha 3 min time limit Third students were given instructional book-lets corresponding to their assigned group After they had read theinstructions for reading the booklets studentsrsquo current motivationfor doing the learning task was assessed Next students started learn-ing with the text material corresponding to their treatment groupStudents were instructed to carefully read the text on the biologyof the influenza in order to comprehend the material Students inthe drawing condition were instructed to read the text and addi-tionally to draw pictures for each text paragraph using the drawingprompt representing the main ideas of each text paragraph Thatis students had to use the pictorial elements given in the legendsuch as the virus as templates for their own paper-pencil baseddrawing across the pre-dawn background Students in the controlgroup were instructed to read the text for comprehension but werenot instructed to engage in drawing Students in both groups learnedat their own pace whereby individual learning time was mea-sured by the instructors in the classrooms Fourth in order to ensurecomparable testing procedures after finishing learning with thewhole learning material students in both groups directly rated theamount of mental effort he or she had invested during learning andthe amount of difficulty he or she had perceived during learningFifth students received the comprehension posttest consisting of

Table 1Text from the second paragraph of the influenza lesson

How the influenza virus replicates

Once inside the influenza virus uses your somatic cell to produce new particlesof the influenza-virus The glycoproteins move toward the membrane of thesomatic cell and stick out into the outside of the cell The capsules of thevirus however are assembled inside the somatic cell Next these newassembled capsules of the virus leave your somatic cell By moving throughthe somatic cell membrane the capsules are enveloped with the membraneand its glycoproteins which then plays the role of the virus membraneThus several new influenza viruses are located outside your somatic cell

Note Translated from the German original

279A Schmeck et alContemporary Educational Psychology 39 (2014) 275ndash286

the comprehension posttest and the drawing posttest Students had20 minutes time for completion and did not have access to thescience text or their drawings Finally students were thanked anddebriefed As students learned at their own rate the whole proce-dure took about 70ndash90 minutes depending on the individual testingtimes

24 Results and discussion

241 ScoringThe dependent variables were studentsrsquo scores on the compre-

hension and drawing posttests studentsrsquo rating on the mental effortand the difficulty scales and the drawing accuracy score indicat-ing the quality of learner-generated drawings produced by studentsin the drawing group during the learning phase

The comprehension test score (pre- and posttest) for each studentwas computed by awarding 1 point for each correct answer and byadding up the points to obtain the total comprehension score (outof a total possible of 25 points) Actual scores ranged from 3 to 24points with a mean of 13 points (SD = 53) Following Schwambornet al (2010) scoring of the drawing test was carried out by count-ing the total number of correct main ideas in each learnerrsquos answeracross the three drawing items The main ideas were drawn out fromboth expert visualizations and a checklist specifying important re-lational features Students could earn a maximum of 19 points onthe drawing test Two student assistants (teacher trainees in biology)scored the quality for each of the three drawings for each studentwith an acceptable inter-rater agreement of GoodmanndashKruskalgamma of 090 Actual scores ranged from 0 to 185 points with amean of 77 points (SD = 47) Total scores of both the comprehen-sion and the drawing test were transferred into z-standardized scoresto make them comparable across studies

The drawing accuracy score (concerning drawing during learn-ing in the drawing group) was computed by using a coding schemeadapted from Schwamborn et al (2010) which was based on expertdrawings and a checklist specifying important relational featuresof the drawings Students could earn a maximum drawing-accuracyscore of 22 points Again the two student assistants scored each ofthe seven learner-generated drawings for each student with an ac-ceptable interrater agreement of GoodmanndashKruskal gamma of 92Both coding schemes were constructed by the first author and abiology teacher Actual scores ranged from 4 to 21 points with amean of 133 points (SD = 50) The total drawing accuracy score wasagain transferred into a z-standardized score

In addition the spatial ability test was scored by tallying thenumber correct out of 10 and the motivation questionnaire wasscored by tallying the nine ratings on both subscales to a total scoreof motivation Finally for comparing performance across the dif-ferent tests the proportion correct on each test was computed bydividing the studentrsquos obtained score by the total possible score

242 Are the groups equivalent on basic characteristicsBefore looking at treatment effects on the dependent variables

we analyzed whether the two groups differed on several control vari-ables A chi-square analysis indicated that there were no significantdifferences regarding gender (p = 562) Separate univariate analy-ses of variance (ANOVAs) revealed that the groups did not differsignificantly on age F lt 1 on spatial ability F lt 1 or on motiva-tion F(1 46) = 360 p = 064 However groups differed significantlyon prior knowledge F(1 46) = 3890 p lt 001 partial eta2 = 46 inthat students in the drawing group scored significantly lower onthe comprehension pretest (M = 10 SD = 15) than students in thecontrol group (M = 34 SD = 12) Thus we included studentsrsquo priorknowledge in the following analyses

243 Is there support for the generative drawing effectMean proportion correct and SDs on the comprehension and

drawing posttests for both groups are presented in Table 2 Repeat-ed measures univariate analyses of variance (ANOVA) with thecomprehension pre- and post-test scores as the within-subject factorsand group (drawing versus control) as the between-subject factorshowed a main effect over time indicating that overall partici-pants reached significant knowledge gains between thecomprehension pretest and the comprehension posttest F(146) = 9897 p lt 001 partial eta2 = 68 An interaction additionallyshowed that these knowledge gains were significantly higher forthe drawing group than for the control group F(1 46) = 4617p lt 001 partial eta2 = 50

For the drawing test a repeated measures ANOVA was not pos-sible since these items were only used in the posttest In this casea univariate analysis of covariance (ANCOVA) predicting the drawingtest score with group (drawing versus control) as the factorial in-dependent variable and prior knowledge as a covariate showed thatthe drawing group scored significantly better than the control groupon the drawing posttest F(1 45) = 1349 p = 001 partial eta2 = 231

Cohenrsquos d favoring the drawing group over the control group was085 for the comprehension posttest and 115 for the drawingposttest all of which are considered large effects Thus there is strongsupport for the generative drawing effect as predicted

Additionally results revealed that the drawing group needed sig-nificantly more learning time (M = 2108 min SD = 424) than thecontrol group (M = 1738 min SD = 333) F(1 46) = 1134 p = 002partial eta2 = 20 Thus to test whether learning time mediates thepositive effect of drawing on text comprehension additional me-diation analyses (Baron amp Kenny 1986) were calculated by includinglearning time as an additional predictor in the aforementioned linearmodel A mediation effect would be detected if in this case effectsof drawing on text comprehension would significantly decreaseResults of the mediation analyses showed that the effect of drawingon both comprehension test scores and drawing test scores was notfully mediated by learning time That is including learning time stillrevealed the interaction between group (drawing versus control)and time (pre- versus post) in that the drawing group had signifi-cantly higher knowledge gains than the control group on thecomprehension test items (p lt 001) Furthermore the drawing groupalso still outperformed the control group on the drawing posttestafter controlling for learning time (p = 009)

Furthermore results revealed that students in the drawing grouprated their invested mental effort during learning significantly higher(M = 504 SD = 112) than students in the control group (M = 396SD = 165) F(1 46) = 705 p = 011 partial eta2 = 13 There was nodifference between the two groups on the perceived difficulty item(drawing group M = 408 SD = 150 control group M = 425SD = 122 F lt 1) Thus consistent with predictions concerning thegenerative drawing effect there is partial support for the idea that

1 The exclusion of studentsrsquo prior knowledge in the reported analyses does notchange the reported pattern of results

Table 2Mean proportion correct on the comprehension test and drawing test for two groupsndash Experiment 1

Group Type of test

n Comprehension test Drawing test

M SD M SD

Drawing 24 61 20 52 27Control 24 44 20 28 11

Note Asterisk () indicates significant difference from control group at p lt 05

280 A Schmeck et alContemporary Educational Psychology 39 (2014) 275ndash286

drawing causes students to engage in more generative processingduring learning

Taken together the results suggest that the drawing strategy en-courages students to engage in generative processing during learningas is indicated by their higher learning outcomes Thus the dataprovide further evidence for the generative drawing effect consis-tent with the results of Schwamborn et al (2010) Additionallyresults indicate that students in the drawing condition seem to investmore mental effort than students in the control group without per-ceiving higher levels of difficulty

244 Is there support for the prognostic drawing effectMean proportion correct on drawing accuracy during learning

was 59 (SD = 23) A correlation analysis revealed that the drawing-accuracy score of learner-generated drawings correlated significantlywith the comprehension posttest score r = 620 p lt 001 and withthe drawing posttest score r = 623 p lt 001 Additional correla-tion analyses revealed that the drawing-accuracy score of learner-generated drawings correlated significantly negatively with theperceived difficulty score r = minus489 p = 015 There were no signif-icant correlations between the drawing accuracy score and eitherthe invested mental effort score r = minus134 p = 533 the prior knowl-edge test score r = minus004 p = 984 the spatial ability test score r = 072p = 739 or the motivation test score r = 086 p = 690 Thus as pre-dicted the data provide further evidence for the prognostic drawingeffect consistent with the results of Schwamborn et al (2010)

In sum the results of Experiment 1 are consistent with the pre-diction that students learn better from a science text when they areasked to draw illustrations representing the main ideas of the textand that the quality of the generated drawings during learning cor-relates positively with studentsrsquo text comprehension (egSchwamborn et al 2010 van Meter 2001 van Meter et al 2006)

However it might be argued that the reported results are dueto the way we supported the strategy use In other words the re-ported positive effect of the learner-generated drawing strategy mightnot be caused by studentsrsquo engagement in generative learning ac-tivities during reading (de Jong 2005 Mayer 2004 2009 Wittrock1990) but rather by the additional pictorial information given in thedrawing prompt Additionally looking at studentsrsquo learning out-comes our results indeed show positive effects of drawing howevermean scores of learning outcomes for the drawing group aremedium-sized Thus it might be argued that the way we sup-ported the strategy use was not fully sufficient In other words thereported positive effect of the learner-generated drawing strategyie the generative drawing effect might be increased by giving stu-dents instructional support in addition to the drawing prompt (vanMeter 2001 van Meter et al 2006 van Meter amp Garner 2005) Toaddress these issues we added two experimental conditions byimplementing author-generated pictures in the design in Experi-ment 2

3 Experiment 2

One possible issue with Experiment 1 is the type of control groupused In Experiment 1 following Schwamborn et al (2010) we useda reading only control group in which the control group learnedwith verbal information only In the drawing group however stu-dents not only learned with verbal information but also with pictorialinformation given by the drawing prompt Based on theories of mul-timedia learning the use of different forms of representations suchas texts and pictures can promote learning in that ldquopeople learnbetter from words and pictures than from words alonerdquo (ie mul-timedia principle Mayer 2009 p 223) because in this caseboth a (verbal) propositional representation as well as a (pictori-al) mental model are built up and are optimally integrated into oneschema that can be stored in long-term memory (Schnotz 2005)

This assumption is also in line with the dual-coding approach statedby Paivio (1986) In this regard it might be argued that the re-ported drawing effect is actually a multimedia effect that is basedon the presentation of text and picture rather than a generativedrawing effect that is based on studentsrsquo active engagement indrawing activities during reading In other words instead of askingpeople to draw pictures representing the main ideas of the textgiving them text and author-generated pictures representing themain ideas of the text might be as good or even better Thus weincluded a condition in Experiment 2 in which we added author-generated pictures to the text

An additional issue with Experiment 1 is whether the reportedgenerative drawing effect can be enhanced by using various formsof supporting the strategy First there is evidence that using adrawing prompt during learning seems to be effective in support-ing the learner-generated drawing strategy by minimizing thecreation of extraneous processing (cf Schwamborn et al 2010 seealso Exp 1) Second research has shown that instructing studentsto compare their own drawing with an author-generated picturemight be also effective in supporting the learner-generated drawingstrategy as self-monitoring processes are enhanced (cf van Meter2001) Up to now however there is no empirical evidence whetherthe combination of both ways to support the drawing strategy hasan additive effect on learning outcomes Thus we included a furthercondition in Experiment 2 in which we combined both forms ofstrategy support

The main purpose of Experiment 2 was to test the generativedrawing and prognostic drawing effects of learner-generated drawingas in Experiment 1 but this time also compared with another controlgroup (ie author-generated pictures) Additionally we were in-terested in testing whether the benefits of the learner-generateddrawing strategy can be increased when we instructionally supportstudents not only with a drawing prompt but also with an author-generated picture after the drawing process In this new treatmentwe instructed students to draw a picture of the text content andthen to compare their own drawing with an expert picture

31 Participants and design

The participants were 168 German eighth graders from highertrack secondary schools The mean age was 138 years (SD = 06)and there were 112 girls and 56 boys The study was based on a2 times 2-between-subjects design with learner-generated drawing (yesno) and author-generated picture (yesno) as factors Forty studentsserved in the drawing group 44 students served in the author-generated picture group 41 students served in the drawing + author-generated picture group and 43 students served in the control group

32 Materials

The materials were identical to those used in Experiment 1 exceptthat we used a shortened version of the comprehension pretest thatconsisted of 19 rather than 25 items (Cronbachrsquos alpha = 70) andslightly extended versions of both the comprehension posttest (28items Cronbachrsquos alpha = 84) and the drawing test (four items witha maximum score of 21 points Cronbachrsquos alpha = 78) The pretestwas shortened because the first experiment showed that the re-spective items were either much too easy or much too difficult andthus unsuitable to differentiate between successful and unsuccess-ful learners thus we deleted these items in the second experimentFurthermore we decided to add some items to the comprehen-sion posttest in the second experiment because during data analysisof the first experiment and after receiving some feedback fromexperts in the domain of biology we recognized that a few itemsassessing transfer ability could be added These transfer itemshowever would have been unsuitable to be included in the pretest

281A Schmeck et alContemporary Educational Psychology 39 (2014) 275ndash286

because they are too difficult to answer without prior training inthe topic Additionally author-generated pictures were used in thenew conditions The author-generated pictures were static func-tional pictures representing the main ideas of each paragraph andconsisted of pictorial elements identical to those provided in thedrawing prompt (as shown in Fig 2) These pictures were con-structed by the first author in cooperation with a biology teacher

The drawing version of the booklet was identical to that usedin Experiment 1 (as shown in Fig 1) The control version of the learn-ing booklet was identical to that used in Experiment 1 The author-generated picture version of the booklet consisted of seven pairsof facing pages with a text paragraph on the left page and a corre-sponding author-generated picture (such as in Fig 2) on the rightpage The drawing + author-generated picture version of the bookletcontained the material from the drawing version consisting of sevenpairs of facing pages with a text paragraph on the left page and atwo-part drawing prompt on the right page In addition attachedto each page there was an additional page that students could foldout after having generated their drawing When unfolding this ad-ditional page a picture of that text paragraph right aside the drawingprompt was provided and there was an additional instruction to

compare the learner-generated drawing with the author-generatedpicture Author-generated pictures were the same as in the author-generated picture version of the booklet

33 Procedure

The procedure was identical to that used in Experiment 1 exceptthat there were two additional groups learning with author-generated pictures Students in the author-generated picturecondition were instructed to read the text and additionally to lookat pictures representing the main ideas of each text paragraph Stu-dents in the drawing + author-generated picture version of thebooklet were instructed to read the text to draw pictures for eachtext paragraph using the drawing prompt representing the mainideas of each text paragraph and finally to compare their picturewith an author-generated picture representing main ideas of eachparagraph correctly Students in all groups learned at their own pacewhereby individual learning time was measured by the instruc-tors in the classrooms Again to ensure that studentsrsquo in both drawinggroups did not feel rushed when students in the non-drawing

Fig 2 Author-generated pictures for the seven paragraphs in the author-generated picture versions of the learning booklet Note Pictures are scaled-down from the orig-inal format

282 A Schmeck et alContemporary Educational Psychology 39 (2014) 275ndash286

group completed the task early groups were tested in separateclassrooms

34 Results

341 ScoringAll tests instruments were scored with the same procedures used

in Experiment 1 Again two student assistants (teacher trainees inbiology) scored each of the drawing test items and each of the sevenlearner-generated drawings for each student with acceptableinterrater agreements (drawing test GoodmanndashKruskal gamma of90 drawing-accuracy GoodmanndashKruskal gamma of 94) Actualscores ranged from 1 to 28 points (M = 153 points SD = 58) for thecomprehension test from zero to 21 points (M = 109 points SD = 53)for the drawing test and from 275 to 215 points (M = 141 pointsSD = 48) for drawing accuracy Again total scores of comprehen-sion drawing and accuracy were transferred into z-standardizedscores

342 Are the groups equivalent on basic characteristicsBefore looking at treatment effects on the dependent variables

we analyzed whether the four experimental groups differed onseveral control variables A chi-square analysis indicated that therewere no significant differences regarding gender (p = 097) Sepa-rate univariate analyses of variance (ANOVAs) revealed that thegroups also did not differ significantly on age F lt 1 on spatial abilityF(3 164) = 120 p = 312 or on motivation F(3 164) = 122 p = 305However groups differed significantly on prior knowledge F(3164) = 104 p = 010 partial eta2 = 07 in that students in the controlgroup scored significantly higher on the comprehension pretest(M = 25 SD = 17) than students in both (p lt 05) the author-generated picture group (M = 17 SD = 16) and the drawing + author-generated picture group (M = 15 SD = 12) the drawing group(M = 20 SD = 14) did not differ significantly from the other groupsThus we included studentsrsquo prior knowledge as a covariate in thefollowing analyses

343 Is there support for the generative drawing effectA major goal in this experiment was to determine whether asking

students to generate drawings to represent science text is a moreeffective learning strategy than asking students to learn with textalone or with text and author-generated pictures In other wordswe wanted to determine whether we could replicate and extendthe learner-generated drawing effect Additionally we were inter-ested in whether giving students an author-generated picture afterdrawing can increase the benefits of the learning strategy Mean pro-portion correct and standard deviations on the comprehension anddrawing tests for the four groups are presented in Table 3

The left portion of Table 3 summarizes the mean proportioncorrect on the comprehension test A two-factorial analysis ofcovariance (ANCOVA) predicting learning outcomes (comprehen-sion posttest score) with learner-generated drawing (yesno) and

author-generated picture (yesno) as the factorial independent vari-ables and prior knowledge as a covariate showed a significantpositive main effect of learner-generated drawing F(1 163) = 398p = 048 partial eta2 = 02 a significant interaction effect F(1163) = 626 p = 013 partial eta2 = 04 but no main effect of author-generated pictures F lt 1 In addition multiple pairwise comparisons(with p lt 05) showed that the drawing group performed signifi-cantly better than each of the three other groups which did not differsignificantly from each other Cohenrsquos d favoring the drawing groupover the author-generated picture group was 49 over the learner-generated + author-generated picture group was 57 and over thecontrol group was 52

The right portion of Table 3 summarizes the mean proportioncorrect on the drawing posttest Again a two-factorial analysis ofcovariance (ANCOVA) predicting learning outcome (drawing testscore) with learner-generated drawing (yesno) and author-generatedpicture (yesno) as the factorial independent variables and priorknowledge as a covariate showed a significant positive main effectof learner-generated drawing F(1 163) = 6260 p lt 001 partialeta2 = 28 a significant positive main effect of author-generated pic-tures F(1 163) = 1104 p = 001 partial eta2 = 06 and a significantinteraction effect F(1 163) = 1658 p lt 001 partial eta2 = 09 In ad-dition multiple pairwise comparisons (with p lt 05) showed thatboth the drawing group and the drawing + author-generated picturegroup performed significantly better than the author-generatedpicture group (d = 68 d = 59) and the control group (d = 187d = 188) In turn the author-generated picture group performed sig-nificantly better than the control group (d = 95) The drawing groupand the drawing + author-generated picture group did not differ sig-nificantly from each other (d = 15)2 Overall these results areconsistent with Experiment 1 and provide additional support forthe generative drawing effect

In accordance with Experiment 1 we were interested in whetherdifferences in learning time among the experimental groups mediatethe positive effect of drawing on text comprehension First an ANOVApredicting learning time with learner-generated drawing (yesno)and author-generated picture (yesno) as the factorial indepen-dent variables showed a significant main effect of learner-generateddrawing F(1 164) = 39226 p lt 001 partial eta2 = 71 a significantmain effect of author-generated picture F(1 164) = 1685 p lt 001partial eta2 = 09 and a significant interaction effect F(1 164) = 490p = 028 partial eta2 = 03 Linear contrasts (with p lt 05) revealedthat the drawing group (M = 1938 min SD = 380) and thedrawing + author-generated picture group (M = 2340 min SD = 551)needed significantly more learning time than the author-generatedpicture group (M = 960 min SD = 397) and the control group(M = 834 min SD = 251) Thus to test whether learning time me-diates the positive effect of drawing on text comprehensionadditional mediation analyses (Baron amp Kenny 1986) were calcu-lated by including learning time as an additional predictor in theaforementioned linear model Results of the mediation analysesshowed that the effects of drawing on the comprehension posttestand the drawing posttest scores (see multiple pairwise compari-sons) are mediated by learning time to some extent That is includinglearning time in the linear model for predicting comprehension testscores still revealed a positive effect of the drawing group com-pared with the drawing + author-generated group on thecomprehension test (p = 012) However including learning time inthe linear model for predicting comprehension posttest scoresreduced the positive effect of the drawing group compared with theauthor-generated picture group (from p = 034 to p = 281) as wellas compared with the control group (from p = 002 to p = 087) being

2 The exclusion of studentsrsquo prior knowledge in the reported analyses does notchange the overall pattern of results

Table 3Mean proportion correct on the comprehension test and drawing test for the fourgroups ndash Experiment 2

Group Type of test

n Comprehension test Drawing test

M SD M SD

Learner-generated drawing 40 63 22 66 22Author-generated picture 44 53 19 50 25Learner-generated drawing +

author-generated picture41 51 20 63 19

Control 43 52 20 30 16

Note Asterisk () indicates significant difference from control group at p lt 05

283A Schmeck et alContemporary Educational Psychology 39 (2014) 275ndash286

no longer statistically significant Regarding the drawing posttestscore including learning time does not change the reported patternof results except that the positive effect of the drawing + author-generated picture group compared with the author-generated picturegroup is no longer statistically significant (from p = 004 to p = 223)

There were neither main effects of learner-generated drawingand author-generated pictures on the mental effort item (drawinggroup M = 455 SD = 025 author-generated picture group M = 459SD = 024 drawing + author-generated picture group M = 444SD = 026 control group M = 481 SD = 024 F lt 1) nor on theperceived difficulty item (drawing group M = 363 SD = 023 author-generated picture group M = 371 SD = 022 drawing + author-generated picture group M = 395 SD = 023 control group M = 393SD = 022 F lt 1)

Taken together the drawing strategy apparently fosters stu-dents to engage in generative activities indicated by their higherlearning outcomes Thus the data provide further evidence for thegenerative drawing effect predicted by Schwamborn et al (2010)In Experiment 2 benefits of the drawing activity however are me-diated by learning time and do not involve higher mental effortAdditionally there was no increased benefit when additional drawingsupport was available in the form of author-generated pictures

344 Is there support for the prognostic drawing effectA second major goal of this study was to determine whether the

prognostic drawing effect could be extended to a new context Meanproportion correct on drawing-accuracy during learning was 60(SD = 04) for the drawing group and 68 (SD = 03) for thedrawing + author generated picture group This difference betweenthe two drawing groups is not significant F(1 79) = 252 p = 116This lack of group differences allowed us to pool the data of bothdrawing groups for subsequent correlation analyses Correlation anal-yses based on the combined data from the two drawing groupsrevealed that the drawing-accuracy score of learner-generated draw-ings correlates significantly with the comprehension posttest scorer = 470 p lt 001 as well as with the drawing posttest score r = 615p lt 001 Additional correlation analyses revealed that the drawing-accuracy score of learner-generated drawings did not correlatesignificantly with the prior knowledge test score r = 095 p = 400the spatial ability test score r = 127 p = 257 the motivation testscore r = 033 p = 769 or the mental effort test score r = 042p = 712 The correlation between the drawing-accuracy score andthe perceived difficulty score was only slightly statistical signifi-cance r = minus218 p = 053 Thus the data provide further evidencefor the prognostic drawing effect consistent with the results ofSchwamborn et al (2010)

In sum results of Experiment 2 are partly consistent with theresults of Experiment 1 in that students learn better from a sciencetext when they are asked to draw illustrations representing the mainideas of the text and the quality of the generated drawings duringlearning correlates positively with studentsrsquo text comprehension (egSchwamborn et al 2010 van Meter 2001 van Meter et al 2006)

4 Discussion

41 Empirical contributions

The present set of experiments makes three empirical contri-butions to the field First this study shows strong and consistentevidence that students who are asked to generate drawings (withsufficient support) during reading a scientific text that describes acausal sequence perform better than students who read withoutdrawing both on a comprehension test (d = 085 in Experiment 1and d = 052 in Experiment 2) and on a drawing test (d = 115 in Ex-periment 1 and d = 187 in Experiment 2) Thus the generativedrawing effect can be extended to a new domain and therefore

corresponds to Shavelson and Townersquos (2002 p 4) recommenda-tion to ldquoreplicate and generalize across studiesrdquo as one of the sixessential scientific principles of educational research

Second this study shows strong and consistent evidence that thequality of drawings that students generate during learning with ascientific text that describes a causal sequence is positively relatedto subsequent performance on tests of learning outcome includ-ing a comprehension test (r = 623 in Experiment 1 and r = 470 inExperiment 2) and a drawing test (r = 620 in Experiment 1 andr = 615 in Experiment 2) Thus the prognostic drawing effect canbe replicated and extended to a new domain consistent with stan-dards for scientific research in education prescribed by Shavelsonand Towne (2002)

Third this study shows that asking learners to draw picturesduring reading a scientific text (ie learner-generated drawing groupin Experiment 2) is more effective than simply providing draw-ings (ie author-generated picture group in Experiment 2) both ona comprehension test (d = 049) and a drawing test (d = 068) Sim-ilarly adding author-generated drawings (ie learner-generatedpictures + author-generated pictures group in Experiment 2) doesnot improve the learning outcomes of students who also draw pic-tures during learning (ie learner-generated pictures group inExperiment 2) either on a comprehension test (d = minus057) or adrawing test (d = minus015) In short the act of drawing during learn-ing (with sufficient support) improves learning beyond the simpleprovision of drawings

42 Theoretical contributions

The results are consistent with the idea that drawing during learn-ing serves as a generative activity (Mayer amp Wittrock 2006Schwamborn et al 2010 van Meter amp Garner 2005 Wittrock 1990)That is the act of drawing encourages learners to engage in gen-erative cognitive processing during learning such as organizing therelevant information into a coherent structure and integrating itwith relevant prior knowledge from long-term memory In thepresent study positive effects of drawing were indicated with a com-prehension and a drawing learning outcome test and therefore arein line with the theoretical assumption derived from the GTDC thatbenefits of drawing can be found if learning outcome tests are usedthat are sensitive to the underlying process of drawing (cf van Meteramp Garner 2005) Additionally in our study the drawing activity wassupported in a way that was intended to help learners carry out theunderlying cognitive processes of drawing (ie selecting organiz-ing and integrating) successfully In this regard results of the presentstudy might supplement the theoretical framework of learner-generated drawing by providing further evidence that benefits ofdrawing defined by van Meter and Garnerrsquos GTDC can diminish ifno instructional support is given to constrain and structure thedrawing activity However a fuller understanding of the underly-ing cognitive processes of drawing and how these processes canbe influenced via drawing support requires more direct measuresof cognitive processing during learning Additionally following theidea that metacognitive processes of monitoring and regulation areautomatically activated by drawing (van Meter amp Garner 2005) afuller understanding of the metacognitive effects of drawings is alsorequired

43 Practical contributions

The present study encourages instructional designers and in-structors to incorporate drawing activities into venues involvinglearning from text which we call the generative drawing effect Oneimportant feature of a successful drawing strategy that is presentin this study and in a previous study by Schwamborn et al (2010)is that the drawing activity was supported by providing a

284 A Schmeck et alContemporary Educational Psychology 39 (2014) 275ndash286

background scene and a legend showing how to represent eachelement to constrain and structure the drawing activity Thus animportant practical implication is that students may need supportin their drawing activity so they do not need to draw from scratch

The present study also suggests a potentially useful diagnostictool to gauge the depth of student learning namely the quality ofthe drawings created by students during learning which we referto as the prognostic drawing effect Incorporating a measure of thequality of a learnerrsquos drawing during learning can be a useful toolin developing remedial instruction to give learners individual supportIt may be important to use materials that explain a cause-and-effect process and give learners drawings of the elements they needto represent the process pictorially Asking learners to simply drawpictures of elements is unlikely to be helpful whereas asking themto generate drawings that show the relations among the elementsin a schematic form is more likely to be helpful

44 Limitations and future directions

Some limitations and future directions of our study should beaddressed As noted in the theoretical contributions subsection wedid not have direct measures of cognitive processing during learn-ing so it is not possible to pinpoint how the drawing activity affectedspecific cognitive processes such as attending to relevant informa-tion organizing it and integrating it with prior knowledge We alsodid not assess metacognitive processing during learning thus it isnot possible to pinpoint how the drawing activity affected specificmetacognitive processes such as monitoring and regulation

Furthermore results of the cognitive load rating scales (in-vested mental effort and perceived task difficulty) are inconsistentWhereas in Experiment 1 an effect on mental effort but not on per-ceived task difficulty showed up (ie students in the drawing grouprated their invested mental effort during learning significantlyhigher) no effects on mental effort and task difficulty were foundin Experiment 2 Additionally in both experiments only a nega-tive correlation of perceived task difficulty with the quality of learner-generated pictures appears but no correlation of mental effort withthe quality Following de Jong (2010) those cognitive load ratingscales might have the disadvantages that they do not give a con-current measure of cognitive load and do not measure an essentialconcept in cognitive load theory namely cognitive overload (p 125)Future studies on learner-generated drawing might also use othercognitive load measures such as physiological measures as moredirect indicators of cognitive load

As noted in the practical contributions subsection we showedthe drawing effects by using a scientific text describing how a cause-and-effect system works that is the causal steps regarding aninfection with influenza and the immune response It might be pos-sible however that for other types of text producing drawings mightharm rather than promote text comprehension Thus to test whetherthe reported drawing effects can be extended future research hasto focus on other types of text such as descriptive texts as well ason other types of relations that can be conveyed with other typesof representations such as compare and contrast relations whichcan be shown in a matrix Additionally studentsrsquo learning out-comes were tested immediately after reading thus future work isneeded to investigate the longer-term effects of generative drawingon learning outcomes

Furthermore we only compared drawing with control groupsthat received no further learning strategy instructions However en-gaging in generative learning activities such as drawing requires aconsiderable amount of time Accordingly results showed that forExperiment 2 the positive effect of the drawing group on text com-prehension compared with the author-generated picture group andto the control group was mediated by learning time To rule out thatthe effects of drawing result only from additional time on task instead

of the generative activity future research should also compare thedrawing strategy with other time demanding generative learningstrategies such as summarization (cf Leopold amp Leutner 2012)

Another point that should be noted is that students in both ex-periments received some kind of multimedia materials in that evenwhen they had to draw and did not see presented pictures they wereat least provided with the basic (visual) elements for their draw-ings which they had to do on the given background which thusalso contained information In other words when students are pre-sented with important elements of the drawings which they canuse to draw themselves they will not have to put as much effortinto summarizing visually what they have just read compared withstudents who have to draw without any instructional help Futurestudies might also compare the drawing group with a summariza-tion group in which students receive a set of verbal key terms thatare similar to the drawing elements and are asked to make a textualsummary

Additionally future research is needed to validate the prognos-tic drawing effect So far we know that the quality of learner-generated pictures is related to studentsrsquo learning outcomes (iethe higher the learning outcome the higher the drawing accuracyand vice versa) and their perceived difficulty (ie the lower the per-ceived difficulty the higher the drawing accuracy and vice versa)and that it is not related to studentsrsquo prior knowledge motivationspatial ability or mental effort However less is known about whatthis might mean That is less is known regarding the causal direc-tion of this relation or the presence of a possible further moderatorvariable Do studentsrsquo efforts to produce accurate drawings lead tobetter comprehension and lower perceived difficulty Or do stu-dents who are more adept in drawing benefit more from the strategyand thus perceive the difficulty of the learning materials as beinglower Both arguments seem convincing

Finally more work is needed to determine the level of supportthat makes the drawing strategy most effective for various kinds oflearners As noted in the empirical contribution adding author-generated drawings (ie learner-generated pictures + author-generated pictures group in Experiment 2) does not improve thelearning outcomes of students who also draw pictures during learn-ing and were supported by a drawing prompt In other words thecombination of two ways of supporting the drawing strategy (iegiving a drawing prompt during reading plus an author-generatedpicture after reading) did not improve studentsrsquo learning out-comes compared with students in the drawing group as well ascompared with students in the control and author-generated pic-tures only groups This result is inconsistent with previous research(eg van Meter 2001 van Meter et al 2006) which found that com-paring own drawings to author-generated pictures normally helpslearning van Meter and colleagues (2001 2006) however provid-ed author-generated pictures plus prompting questions after thedrawing process That is students answered prompting questionsto guide the comparison process between their self-generateddrawing and the author-generated drawing In our study studentswere only instructed to generate a drawing to inspect an author-generated one and to check whether their own drawing incomparison with the author-generated one really represented themain ideas of the text paragraph correctly In other words we didnot guide the process of comparing self-generated drawings withauthor-generated ones As a potential consequence students per-formed the intended comparison process inadequately or even notall and thus did not benefit from it One reason for this inade-quate comparison process might be that students need guidancein doing the comparison process Another reason might be the factthat students do not seriously engage in generating drawings oncethey notice that there are author-generated drawings Thus futureresearch should also use additional guidance to test whether thecombination of different ways of supporting the drawing strategy

285A Schmeck et alContemporary Educational Psychology 39 (2014) 275ndash286

(ie giving a drawing prompt during reading plus an author-generated picture after reading) helps learning as well asobservational measures of the drawing process itself (ie think aloudprotocols) to shed more light on the cognitive processes underly-ing the drawing activities

Overall drawing during learning from text appears to be a po-tentially powerful strategy for improving studentsrsquo learning fromscientific text when certain boundaries and prerequisites are takeninto account

Acknowledgments

This article is based on a research project funded by the GermanResearch Foundation (DFG LE 6459-3 as part of FOR 511) We wouldlike to thank Angela Sandmann for her assistance in developing thelearning materials

References

Ainsworth S Prain V amp Tytler R (2011) Drawing to learn in science Science 3331096ndash1097

Alesandrini K L (1981) Pictorial-verbal and analytic-holistic learning strategies inscience learning Journal of Educational Psychology 73 358ndash368

Alesandrini K L (1984) Pictures and adult learning Instructional Science 13 63ndash77Baron R M amp Kenny D A (1986) The moderator-mediator variable distinction in

social psychological research Conceptual strategic and statistical considerationsJournal of Personality and Social Psychology 51 1173ndash1182

Bruumlnken R Plass J L amp Leutner D (2003) Direct measurement of cognitive loadin multimedia learning Educational Psychologist 38 53ndash61

Carney RN amp Levin JR (2002) Pictorial illustrations still improve studentsrsquo learningfrom text Educational Psychology Review 14 5ndash26

de Jong T (2005) The guided discovery principle in multimedia learning In R EMayer (Ed) The Cambridge handbook of multimedia learning (pp 215ndash228) NewYork Cambridge University Press

de Jong T (2010) Cognitive load theory educational research and instructionaldesign Some food for thought Instructional Science 38 105ndash134

Ekstrom R B French J W amp Harman H H (1976) Manual for kit of factor-referencedcognitive tests Princeton NJ Educational Testing Service

Greene T R (1989) Childrenrsquos understanding of class inclusion hierarchies Therelationship between external representation and task performance Journal ofExperimental Child Psychology 48 62ndash89

Hall V C Bailey J amp Tillman C (1997) Can student-generated illustrations be worthten thousand words Journal of Educational Psychology 89 677ndash681

Houmlffler T N (2010) Spatial ability Its influence on Learning with visualizations ndashA meta-analytic review Educational Psychology Review 22 245ndash269

Houmlffler T N Schmeck A amp Opfermann M (2013) Static and dynamic visualrepresentations Individual differences in processing In G Schraw M TMcCrudden amp D Robinson (Eds) Learning through visual displays (pp 133ndash163)Charlotte NC Information Age Publishing

Kalyuga S Chandler P amp Sweller J (1999) Managing split-attention and redundancyin multimedia instruction Applied Cognitive Psychology 13 351ndash371

Leopold C (2009) Lernstrategien und Textverstehen [Learning strategies and textcomprehension] Muumlnster Waxmann

Leopold C amp Leutner D (2012) Science text comprehension Drawing main ideaselection and summarizing as learning strategies Learning and Instruction 2216ndash26

Lesgold A M DeGood H amp Levin J R (1977) Pictures and young childrenrsquos proselearning A supplementary report Journal of Reading Behavior 9 353ndash360

Lesgold A M Levin J R Shimron J amp Guttman J (1975) Pictures andyoung childrenrsquos learning from oral prose Journal of Educational Psychology 67636ndash642

Leutner D Leopold C amp Sumfleth E (2009) Cognitive load and science textcomprehension Effects of drawing and mentally imagining text contentComputers in Human Behavior 25 284ndash289

Mayer R E (2004) Should there be a three-strikes rule against pure discoverylearning The case for guided methods of instruction The American Psychologist59 14ndash19

Mayer R E (2005) Cognitive theory of multimedia learning In R E Mayer (Ed)The Cambridge handbook of multimedia learning (pp 31ndash48) New York CambridgeUniversity Press

Mayer R E (2009) Multimedia learning (2nd ed) New York NY CambridgeUniversity Press

Mayer R E amp Wittrock M C (2006) Problem solving In P Alexander P Winne ampG Phye (Eds) Handbook of educational psychology (pp 287ndash303) Mahwah NJErlbaum

Paas F (1992) Training strategies for attaining transfer of problem-solving skill instatisticsmdashA cognitive-load approach Journal of Educational Psychology 84429ndash434

Paas F Tuovinen J Tabbers H K amp Van Gerven P W M (2003) Cognitive loadmeasurement as a means to advance cognitive load theory EducationalPsychologist 38 63ndash71

Paivio A (1986) Mental representation A dual coding approach New York OxfordUniversity Press

Pashler H Bain P Bottage B Graesser A Koedinger K McDaniel M et al (2007)Organizing instruction and study to improve student learning Washington DCNational Center for Educational Research

Rasco R W Tennyson R D amp Boutwell R C (1975) Imagery instructions anddrawings in learning prose Journal of Educational Psychology 67 188ndash192

Rheinberg F Vollmeyer R amp Burns B D (2001) FAM Ein fragebogen zurerfassung aktueller motivation in lern- und leistungssituationen [QCM Aquestionnaire to assess current motivation in learning situations] Diagnostica47 57ndash66

Schnotz W (2005) An integrated model of text and picture comprehension In RE Mayer (Ed) The Cambridge handbook of multimedia learning (pp 49ndash70) NewYork Cambridge University Press

Schwamborn A Mayer R E Thillmann H Leopold C amp Leutner D (2010) Drawingas a generative activity and drawing as a prognostic activity Journal of EducationalPsychology 102 872ndash879

Schwamborn A Thillmann H Opfermann M amp Leutner D (2011) Cognitive loadand instructionally supported learning with provided and learner-generatedvisualizations Computers in Human Behavior 27 89ndash93

Shavelson R J amp Towne L (Eds) (2002) Scientific research in education WashingtonDC National Academy Press

Sweller J Ayres P amp Kalyuga S (2011) Cognitive Load Theory New York SpringerTirre W C Manelis L amp Leicht K (1979) The effects of imaginal and verbal strategies

on prose comprehension by adults Journal of Reading Behavior 11 99ndash106van Meter P (2001) Drawing construction as a strategy for learning from text Journal

of Educational Psychology 69 129ndash140van Meter P Aleksic M Schwartz A amp Garner J (2006) Learner-generated drawing

as a strategy for learning from content area text Contemporary EducationalPsychology 31 142ndash166

van Meter P amp Garner J (2005) The promise and practice of learner-generateddrawings Literature review and synthesis Educational Psychology Review 12261ndash312

Van Gog T amp Paas F (2008) Instructional efficiency Revisiting the original constructin educational research Educational Psychologist 43 16ndash26

Vollmeyer R amp Rheinberg F (2000) Does motivation affect learning via persistenceLearning and Instruction 4 293ndash309

Weinstein C E amp Mayer R E (1986) The teaching of learning strategies In M CWittrock (Ed) Handbook of research on teaching (pp 315ndash327) New YorkMacmillan

Wittrock M C (1990) Generative processes of comprehension EducationalPsychologist 24 345ndash376

286 A Schmeck et alContemporary Educational Psychology 39 (2014) 275ndash286

  • Drawing pictures during learning from scientific text testing the generative drawing effect and the prognostic drawing effect
  • Introduction
  • Theoretical framework for the learner-generated drawing strategy
  • Empirical framework for the learner-generated drawing strategy
  • Effectiveness of learner-generated drawings
  • Quality of learner-generated drawings
  • Overview of the experiments
  • Experiment 1
  • Participants and design
  • Materials
  • Procedure
  • Results and discussion
  • Scoring
  • Are the groups equivalent on basic characteristics
  • Is there support for the generative drawing effect
  • Is there support for the prognostic drawing effect
  • Experiment 2
  • Participants and design
  • Materials
  • Procedure
  • Results
  • Scoring
  • Are the groups equivalent on basic characteristics
  • Is there support for the generative drawing effect
  • Is there support for the prognostic drawing effect
  • Discussion
  • Empirical contributions
  • Theoretical contributions
  • Practical contributions
  • Limitations and future directions
  • Acknowledgments
  • References

actively involved in the cognitive processes of selecting organiz-ing and integrating the information to be learned Thus learner-generated drawing is a cognitive learning strategy that is aimed tofoster learning from text and if used adequately drawing can in-crease learning outcomes (Ainsworth Prain amp Tytler 2011Alesandrini 1984 van Meter amp Garner 2005)

The processes underlying drawing are described in van Meterand Garnerrsquos (2005) generative theory of drawing construction(GTDC) which is based on Mayerrsquos (2005) model of multimedialearning It is assumed that learners benefit from using the drawingstrategy as drawing requires them to engage in generative learn-ing processes during reading First learners select the relevant keyinformation from the text Second the selected key information isorganized to build up an internal verbal representation of the textinformation Third learners construct an internal nonverbal (visual)representation of the text information and connect it with the verbalrepresentation and with relevant prior knowledge To construct thevisual representation which is the basis for the external drawingthe learner has to rely mainly on the verbal representation and thuslearner-generated drawing demands an integration of the verbal andnonverbal representation

Additionally van Meter and Garner (2005) describe metacognitiveprocesses fostered by the drawing activity ldquoAttempts at construct-ing the nonverbal representation can send learners back to eitherthe verbal representation or the text as difficulties building the in-ternal image are encounteredrdquo (van Meter amp Garner 2005 p 317)That is as the drawing process itself is not linear metacognitive pro-cesses of monitoring and regulation are stimulated by drawing (cfvan Meter 2001 van Meter Aleksic Schwartz amp Garner 2006)

12 Empirical framework for the learner-generated drawing strategy

Following the GTDC (cf van Meter amp Garner 2005) the drawingstrategy is beneficial as it fosters deep cognitive processing includ-ing organizing and integrating material (which can be calledgenerative processing Mayer 2009) as well as metacognitive self-monitoring and regulation processes Research on drawing howeverhas produced somewhat mixed results (see Alesandrini 1984 vanMeter amp Garner 2005 for overviews) in which some studies re-ported positive effects of drawing on text comprehension (egAlesandrini 1981 Hall Bailey amp Tillman 1997 Leopold amp Leutner2012 Lesgold DeGood amp Levin 1977 Lesgold Levin Shimron ampGuttman 1975 Schwamborn et al 2010 van Meter 2001 vanMeter et al 2006) whereas others did not (eg Leutner Leopoldamp Sumfleth 2009 Rasco Tennyson amp Boutwell 1975 Tirre Manelisamp Leicht 1979) Benefits of drawing appear to be related to thequality of studentsrsquo drawings during learning Students who producehigh-quality drawings during reading tend to score better onposttests of learning outcome than do students who produce low-quality drawings during reading (eg Greene 1989 Hall et al 1997Leopold 2009 Lesgold et al 1975 1977 Schwamborn et al 2010van Meter 2001 van Meter et al 2006)

121 Effectiveness of learner-generated drawingsFollowing van Meter and Garner (2005) reasons for the mixed

empirical results concerning drawing can be seen attributed to thetype of test used for assessing learning outcomes as well as in theform of support that assists learners in the drawing process Firstbenefits of drawing are more likely to be revealed on tests that assesshigher-order knowledge of to-be learned content for example testson comprehension and transfer (eg Alesandrini 1981 Leopold ampLeutner 2012) or problem solving (van Meter 2001 van Meter et al2006) Leutner et al (2009) for example found no positive effectof drawing compared with a control group on a multiple choice teston factual knowledge Leopold and Leutner (2012) however showedsuperior effects of the drawing strategy on transfer test perfor-

mance van Meter et al (2006) accordingly found no effects ofdrawing activity on a multiple choice recognition test however stu-dents in the drawing group scored significantly higher on a problem-solving test With regard to the GTDC (van Meter amp Garner 2005)it seems that benefits of drawing can be found if the learningoutcome test complies with characteristics of the verbal and non-verbal representations which are generated by drawing

Second positive effects of drawing often appear under the con-dition that instructional support is provided to constrain andstructure the drawing activity (eg Lesgold et al 1975 1977Schwamborn et al 2010 van Meter 2001 van Meter et al 2006)That is drawing is more effective when the learnersrsquo generation ofthe drawing is assisted by some kind of additional information vanMeter (2001) and van Meter et al (2006) for example showed thatthe provision of author-generated pictures after drawing en-hanced the benefits of the drawing strategy By comparing their owndrawing with a provided one learners get to know what theirdrawing should look like and this might lead them back to revisetheir own drawing and thus their mental model Following the GTDC(van Meter amp Garner 2005) this should improve comprehensionLesgold et al (1975) in turn supported first grade students withcutout figures and instructed them to organize these into an accu-rate pictorial representation while listening to a prose story Thislearner-generated illustration activity facilitated prose learning asindicated by higher recall of story propositions only when stu-dents were given the correct pieces for the illustration or had theillustration done for them When students had to select the piecesfor each illustration out of a pool of cutouts the learner-generatedillustration activity had either a negative or no effect (cf Lesgoldet al 1977) Following these results Schwamborn et al (2010) pro-posed that a pure unsupported drawing instruction might bear therisk that managing the mechanics of drawing itself is difficult forthe learners resulting in insufficient remaining capacity for makingsense of the text through generative processes of organization andintegration which might diminish the benefits of drawing definedby van Meter and Garnerrsquos GTDC To counter this risk in the studyof Schwamborn et al (2010) students in the drawing groups re-ceived baseline instructional support while learning a lesson onwashing which provided them with a drawing prompt that in-cluded a legend showing all the relevant elements for drawing anda partly pre-drawn background for their paper-pencil based draw-ings That is students could use the presented elements as prototypesfor their own drawings and integrate them by pencil in the givenpre-drawn backgrounds Results showed that students who wereinstructed to generate drawings during learning scored signifi-cantly higher on the subsequent comprehension tests than studentswho only read the text

Using cutout-figures (cf Lesgold et al 1975 1977) or a drawingprompt (cf Schwamborn et al 2010) during drawing seems toprovide sufficient constraints and leave enough cognitive capaci-ties for learners to benefit from the drawing strategy Thus cognitiveprocessing including selecting organizing and integrating materi-al should be encouraged resulting in an improved mental modelwhich in turn should improve comprehension (cf GTDC van Meteramp Garner 2005)

In line with the GTDC and the reported results derived from re-search on drawing Schwamborn et al (2010) proposed a generativedrawing effect that is students gain a better understanding of a sci-entific text when they are asked to draw illustrations representingthe content of each paragraph they read This work highlights theimportance of drawing support such as the provision of drawingsof all key elements and a background for the drawing

122 Quality of learner-generated drawingsPrevious studies that measured the quality of studentsrsquo draw-

ings during learning all showed positive correlations between

276 A Schmeck et alContemporary Educational Psychology 39 (2014) 275ndash286

the quality of studentsrsquo drawings during learning and their learn-ing outcomes (eg Greene 1989 Hall et al 1997 Lesgold et al1975 1977 Schwamborn et al 2010 van Meter 2001 van Meteret al 2006) The quality of learner-generated drawing is also re-ferred to as the drawing accuracy (eg van Meter 2001 van Meteret al 2006) and is defined as ldquothe degree to which completeddrawings resemble the represented object(s)rdquo (van Meter amp Garner2005 p 299) In a study by Hall et al (1997) for example collegestudents learning a mechanics lesson with the instruction to drawproduced better learning outcomes on a transfer test than a textonly control group but only if they produced higher quality draw-ings In the study by Schwamborn et al (2010) students who wereable to generate high accuracy drawings scored significantly higheron learning outcome tests than did those who generated loweraccuracy drawings In addition to this Schwamborn and col-leagues also found that the quality of the generated drawings duringlearning correlated positively with the comprehension test scoresBased on these results Schwamborn et al (2010) proposed the prog-nostic drawing effect Students who produce high-quality drawingsduring reading a scientific text tend to score better on posttestsof learning outcome than do students who produce low-qualitydrawings during reading

13 Overview of the experiments

Research on drawing ndash ie the generative and the prognosticdrawing effect ndash is promising however at the present time thereis a need for a more solid evidence base and for a closer examina-tion of theoretical issues First the generalizability is limited atthis point as replication studies using learning outcome tests thatare sensitive to the underlying process of drawing as well as newlearning materials other than the washing lesson (eg Schwambornet al 2010) or the birds wing (van Meter 2001) are yet missingIn their report for the US National Research Council entitledScientific Research in Education Shavelson and Towne (2002 p 4)for example highlighted the need to ldquoreplicate and generalizeacross studiesrdquo as one of the six essential scientific principles ofeducational research It has to be mentioned at this point that whengeneralizing results to new domains or lessons one should care-fully consider whether these are comparable at all In ourexperiments we aimed at generalizing results by Schwamborn et al(2010) who worked with a science text explaining the causalsteps regarding the chemistry of washing to a new lesson that ishowever comparable in that the text we used also describedcausal steps of a process in this case regarding the infection withinfluenza and regarding the immune response That is althoughthere were differences between the two domains (chemistry versusbiology) the lessons showed structural similarities and thusallow for comparing results and drawing conclusions regardinggeneralizability

Second research on drawing indicates that some form of supportis needed to assist learners during drawing Schwamborn et al(2010) for example introduced a drawing prompt as helpful supportfor learners to benefit from drawing They proposed that the re-sulting positive drawing effect is based on studentsrsquo engagementin generative learning activities during reading due to drawing (con-sistent with the GTDC of van Meter amp Garner 2005 see also de Jong2005 Mayer 2004 2009 Wittrock 1990) However the results re-ported by Schwamborn et al (2010) might rather reflect amultimedia effect (Mayer 2005 2009) than the proposed drawingeffect as the learning lesson used ndash a scientific text and a drawingprompt consisting of pictorial elements and backgrounds ndash createda multimedia lesson In other words the results of Schwamborn et al(2010) might not be due to the drawing activity but rather due tothe multimedia effect that students ldquolearn better from words andpictures than from words alonerdquo (Mayer 2009 p 223) In this case

the words are presented in the lessons and pictures are generatedby the students so a control group that receives author-generatedpictures is warranted

Third research on drawing mostly used only one way to supportthe drawing strategy at a specific time That is instructionalsupport was added during learning (ie by using cut-out figures ora drawing prompt (cf Lesgold et al 1975 1977 Schwamborn et al2010) or after learning (ie by providing pictures van Meter 2001van Meter et al 2006) Less is known about whetheradding instructional support not only during learning but also afterlearning can additionally enhance the benefits of the drawingstrategy

Fourth research on drawing should include motivational andcognitive aspects that may have an impact on the effectiveness ofthe learner-generated drawing strategy Studentsrsquo current motiva-tion for example is a one condition for successful learning A studentfor example who has low motivation to learn may invest lesseffort in learning than students who are highly motivated to learn(cf Vollmeyer amp Rheinberg 2000) Studentsrsquo spatial ability maybe a further condition for successful learning when workingwith visualizations (cf Houmlffler 2010 Houmlffler Schmeck amp Opfermann2013) A high-spatial-ability student for example may haveadvantages in learning with visualizations compared with a low-spatial-ability student That is preexisting motivational and cogni-tive differences between students before learning might havean influence on the learning outcome and thus should becontrolled

In addition recent research has shown that not only experi-mental conditions (such as the kind of picture) and the abovementioned ldquoclassicalrdquo covariates can have an impact on how suc-cessful learning takes place but that these effects can be mediatedby the amount of mental effort someone invests while learning orworking on a lesson and by how difficult someone perceives adomain or lesson to be (cf Leutner et al 2009 SchwambornThillmann Opfermann amp Leutner 2011) These aspects of cogni-tive load (invested mental effort and perceived task difficulty) werethus included as additional variables in our studies as well Thuswe conducted the following two experiments using a science textexplaining the biological process of influenza In Experiment 1 weimplemented an experimental drawing condition and a reading onlycontrol condition in order to determine how both the generativeand the prognostic drawing effect would extend to a new contextAnalogous to the study of Schwamborn et al (2010) students inthe drawing condition received a baseline instructional support bymeans of a drawing prompt that included a legend showing all therelevant elements for drawing and a partly pre-drawn back-ground for their drawing (as shown in Fig 1)

In Experiment 2 we again implemented an experimental drawingcondition and a reading only control condition and we addition-ally implemented author-generated pictures in order to test whetherthe generative drawing effect was caused by the simple presenceof illustrations rather than the generation of illustrations That iswe implemented a text plus picture condition (which we called theauthor-generated picture condition) to test whether the reportedgenerative drawing effect of Schwamborn et al (2010) is based onstudentsrsquo engagement in generative learning activities during readingrather than on the pictorial representations given by the drawingprompt In addition we implemented a drawing plus picture con-dition (in which students both draw and are given a picture) to testwhether the reported generative drawing effect can be enhancedby instructing students to compare their own drawing with anauthor-generated picture In short we tested whether combiningdifferent forms of support to the drawing strategy additionally en-hances the benefits of the drawing strategy In both experimentslearning outcome tests that are sensitive to the underlying processof drawing were used

277A Schmeck et alContemporary Educational Psychology 39 (2014) 275ndash286

2 Experiment 1

21 Participants and design

Forty-eight German eighth graders in higher track secondaryschools participated in this study The mean age was 137 years(SD = 06) and there were 22 girls and 26 boys The study was basedon a between-subjects design with two levels of text learning(learner-generated drawings and control) as the single factor Twenty-

four students served in the control group and 24 served in thedrawing group

22 Materials

All materials were paper-pencil based The materials consistedof five adjunct questionnaires two learning booklets two cogni-tive load rating scales and two posttests The five adjunctquestionnaires were intended to determine whether the groups were

Fig 1 Screenshot of the drawing prompt for the first paragraph in the drawing versions of the learning booklet Note Translated from the German original

278 A Schmeck et alContemporary Educational Psychology 39 (2014) 275ndash286

equivalent on basic characteristics They included a participant ques-tionnaire a comprehension pretest a spatial ability test and amotivation questionnaire The participant questionnaire solicited in-formation concerning the studentsrsquo age and sex The comprehensionpretest consisted of 25 multiple-choice items and was intended toassess studentsrsquo prior-knowledge of information covered in the textStudentsrsquo spatial ability was measured with a 10 multiple-choicepaper-folding items taken from a battery of cognitive tests devel-oped by Ekstrom French and Harman (1976) The motivationquestionnaire assessed studentsrsquo current motivation for doing thelearning task after reading the instructions before the lesson It con-sisted of nine items from the challenge and interest subscales of theQuestionnaire on Current Motivation (QCM) developed by RheinbergVollmeyer and Burns (2001) Cognitive load by means of investedmental effort was measured using the 7-point subjective rating scaledeveloped by Paas (1992) which ranges from (1) very low mentaleffort to (7) very high mental effort Cognitive load by means of per-ceived task difficulty was measured using the 7-point subjective ratingscale developed by Kalyuga Chandler and Sweller (1999) whichranges from (1) very easy to (7) very difficult These subjective mea-sures have been criticized for assessing cognitive load with only singleitems (eg Bruumlnken Plass amp Leutner 2003) However several studiesshowed the effectiveness of the rating scale by showing that the vari-ation in learnersrsquo cognitive load ratings depended on variations intask complexity or instructional design (for overviews see PaasTuovinen Tabbers amp Van Gerven 2003 Van Gog amp Paas 2008) Inthis regard Sweller Ayres and Kalyuga (2011) conclude that ldquothesimple subjective rating scale regardless of the wording used (mentaleffort or difficulty) has perhaps surprisingly been shown to be themost sensitive measure available to differentiate the cognitive loadimposed by different instructional proceduresrdquo (p 74) For that reasonand due to the economic applicability we decided to use this kindof cognitive load measurement while acknowledging the limita-tions of a short self-report instrument

The two learning booklets each included a science text on thebiology of the influenza The text explained the causal steps re-garding an infection with influenza and regarding the immuneresponse which is an unfamiliar subject for eighth graders in highertrack secondary schools due to the German curriculum The text con-sisted of approximately 850 words (in German) and was divided intoseven paragraphs (as shown in Table 1)

The drawing version of the booklet contained seven pairs of facingpages with a text paragraph on the left page and a two-part drawingprompt on the right page The first part of the drawing prompt in-cluded a legend showing all the relevant elements (in total eightelements) for drawing a picture for that text paragraph (as shownin the top of Fig 1) The second part of the drawing prompt in-cluded a partly pre-drawn background for studentsrsquo drawing (asshown in the bottom of Fig 1) Overall students had to make sevendrawings ie one drawing to each paragraph

The control version of the learning booklet contained four pairsof facing pages with one of the seven text paragraph on each page

Students in both groups learned with exactly the same text mate-rial To make sure that students in the control group learnedwith the same amount of information as students in the drawinggroup all elements of the drawing prompt as well as the spatialrelations between these elements were also described in the sciencetext

The two posttests intended to assess the learning outcomes werea comprehension posttest and a drawing posttest The comprehen-sion posttest (Cronbachrsquos alpha = 083) consisted of 25 multiple-choice items (the same items as in the comprehension pretest) andwas intended to assess studentsrsquo comprehension of the factual andconceptual information covered in the text as well as their abilityto transfer what was presented to new situations An item exampleis ldquoT-helper cells do not only recognize viruses but also agents thatare extraneous to the body Which medication would you admin-ister to a patient who has received a new kidney (a) a medicinethat suppresses the immune response of the body (b) a medicinethat activates the immune response of the body (c) a medicine thatcontains antigens or (d) a medicine that contains blood of the kidneydonorrdquo [(a) is the correct answer] The drawing test (Cronbachrsquosalpha = 081) was intended to assess studentsrsquo comprehension of theconceptual information presented in the science text by means ofdrawing That is students had to reproduce the main ideas givenin the text by drawing It consisted of three drawing items in whichstudents were asked to draw sketches depicting key concepts of thetext and their spatial relations An item example for the drawingtest is ldquoHow does an influenza virus invade a cell and how is it re-producedrdquo The science text the drawing prompt and the learningoutcome tests were constructed by the first author in cooperationwith a biology teacher The materials were adapted fromSchwamborn et al (2010) however using another science domainand including measures of individual learning times and cognitiveload

23 Procedure

Participants were tested in the schoolsrsquo classrooms Within theirclasses they were randomly assigned to one of the two groupsGroups were tested in separate classrooms in order to insure thatstudents in the drawing group did not feel rushed when studentsin the control group completed the task early Each student wasseated at an individual desk First students were given the partic-ipant questionnaire and the comprehension pretest to complete attheir own rate Second students filled in the paper-folding test witha 3 min time limit Third students were given instructional book-lets corresponding to their assigned group After they had read theinstructions for reading the booklets studentsrsquo current motivationfor doing the learning task was assessed Next students started learn-ing with the text material corresponding to their treatment groupStudents were instructed to carefully read the text on the biologyof the influenza in order to comprehend the material Students inthe drawing condition were instructed to read the text and addi-tionally to draw pictures for each text paragraph using the drawingprompt representing the main ideas of each text paragraph Thatis students had to use the pictorial elements given in the legendsuch as the virus as templates for their own paper-pencil baseddrawing across the pre-dawn background Students in the controlgroup were instructed to read the text for comprehension but werenot instructed to engage in drawing Students in both groups learnedat their own pace whereby individual learning time was mea-sured by the instructors in the classrooms Fourth in order to ensurecomparable testing procedures after finishing learning with thewhole learning material students in both groups directly rated theamount of mental effort he or she had invested during learning andthe amount of difficulty he or she had perceived during learningFifth students received the comprehension posttest consisting of

Table 1Text from the second paragraph of the influenza lesson

How the influenza virus replicates

Once inside the influenza virus uses your somatic cell to produce new particlesof the influenza-virus The glycoproteins move toward the membrane of thesomatic cell and stick out into the outside of the cell The capsules of thevirus however are assembled inside the somatic cell Next these newassembled capsules of the virus leave your somatic cell By moving throughthe somatic cell membrane the capsules are enveloped with the membraneand its glycoproteins which then plays the role of the virus membraneThus several new influenza viruses are located outside your somatic cell

Note Translated from the German original

279A Schmeck et alContemporary Educational Psychology 39 (2014) 275ndash286

the comprehension posttest and the drawing posttest Students had20 minutes time for completion and did not have access to thescience text or their drawings Finally students were thanked anddebriefed As students learned at their own rate the whole proce-dure took about 70ndash90 minutes depending on the individual testingtimes

24 Results and discussion

241 ScoringThe dependent variables were studentsrsquo scores on the compre-

hension and drawing posttests studentsrsquo rating on the mental effortand the difficulty scales and the drawing accuracy score indicat-ing the quality of learner-generated drawings produced by studentsin the drawing group during the learning phase

The comprehension test score (pre- and posttest) for each studentwas computed by awarding 1 point for each correct answer and byadding up the points to obtain the total comprehension score (outof a total possible of 25 points) Actual scores ranged from 3 to 24points with a mean of 13 points (SD = 53) Following Schwambornet al (2010) scoring of the drawing test was carried out by count-ing the total number of correct main ideas in each learnerrsquos answeracross the three drawing items The main ideas were drawn out fromboth expert visualizations and a checklist specifying important re-lational features Students could earn a maximum of 19 points onthe drawing test Two student assistants (teacher trainees in biology)scored the quality for each of the three drawings for each studentwith an acceptable inter-rater agreement of GoodmanndashKruskalgamma of 090 Actual scores ranged from 0 to 185 points with amean of 77 points (SD = 47) Total scores of both the comprehen-sion and the drawing test were transferred into z-standardized scoresto make them comparable across studies

The drawing accuracy score (concerning drawing during learn-ing in the drawing group) was computed by using a coding schemeadapted from Schwamborn et al (2010) which was based on expertdrawings and a checklist specifying important relational featuresof the drawings Students could earn a maximum drawing-accuracyscore of 22 points Again the two student assistants scored each ofthe seven learner-generated drawings for each student with an ac-ceptable interrater agreement of GoodmanndashKruskal gamma of 92Both coding schemes were constructed by the first author and abiology teacher Actual scores ranged from 4 to 21 points with amean of 133 points (SD = 50) The total drawing accuracy score wasagain transferred into a z-standardized score

In addition the spatial ability test was scored by tallying thenumber correct out of 10 and the motivation questionnaire wasscored by tallying the nine ratings on both subscales to a total scoreof motivation Finally for comparing performance across the dif-ferent tests the proportion correct on each test was computed bydividing the studentrsquos obtained score by the total possible score

242 Are the groups equivalent on basic characteristicsBefore looking at treatment effects on the dependent variables

we analyzed whether the two groups differed on several control vari-ables A chi-square analysis indicated that there were no significantdifferences regarding gender (p = 562) Separate univariate analy-ses of variance (ANOVAs) revealed that the groups did not differsignificantly on age F lt 1 on spatial ability F lt 1 or on motiva-tion F(1 46) = 360 p = 064 However groups differed significantlyon prior knowledge F(1 46) = 3890 p lt 001 partial eta2 = 46 inthat students in the drawing group scored significantly lower onthe comprehension pretest (M = 10 SD = 15) than students in thecontrol group (M = 34 SD = 12) Thus we included studentsrsquo priorknowledge in the following analyses

243 Is there support for the generative drawing effectMean proportion correct and SDs on the comprehension and

drawing posttests for both groups are presented in Table 2 Repeat-ed measures univariate analyses of variance (ANOVA) with thecomprehension pre- and post-test scores as the within-subject factorsand group (drawing versus control) as the between-subject factorshowed a main effect over time indicating that overall partici-pants reached significant knowledge gains between thecomprehension pretest and the comprehension posttest F(146) = 9897 p lt 001 partial eta2 = 68 An interaction additionallyshowed that these knowledge gains were significantly higher forthe drawing group than for the control group F(1 46) = 4617p lt 001 partial eta2 = 50

For the drawing test a repeated measures ANOVA was not pos-sible since these items were only used in the posttest In this casea univariate analysis of covariance (ANCOVA) predicting the drawingtest score with group (drawing versus control) as the factorial in-dependent variable and prior knowledge as a covariate showed thatthe drawing group scored significantly better than the control groupon the drawing posttest F(1 45) = 1349 p = 001 partial eta2 = 231

Cohenrsquos d favoring the drawing group over the control group was085 for the comprehension posttest and 115 for the drawingposttest all of which are considered large effects Thus there is strongsupport for the generative drawing effect as predicted

Additionally results revealed that the drawing group needed sig-nificantly more learning time (M = 2108 min SD = 424) than thecontrol group (M = 1738 min SD = 333) F(1 46) = 1134 p = 002partial eta2 = 20 Thus to test whether learning time mediates thepositive effect of drawing on text comprehension additional me-diation analyses (Baron amp Kenny 1986) were calculated by includinglearning time as an additional predictor in the aforementioned linearmodel A mediation effect would be detected if in this case effectsof drawing on text comprehension would significantly decreaseResults of the mediation analyses showed that the effect of drawingon both comprehension test scores and drawing test scores was notfully mediated by learning time That is including learning time stillrevealed the interaction between group (drawing versus control)and time (pre- versus post) in that the drawing group had signifi-cantly higher knowledge gains than the control group on thecomprehension test items (p lt 001) Furthermore the drawing groupalso still outperformed the control group on the drawing posttestafter controlling for learning time (p = 009)

Furthermore results revealed that students in the drawing grouprated their invested mental effort during learning significantly higher(M = 504 SD = 112) than students in the control group (M = 396SD = 165) F(1 46) = 705 p = 011 partial eta2 = 13 There was nodifference between the two groups on the perceived difficulty item(drawing group M = 408 SD = 150 control group M = 425SD = 122 F lt 1) Thus consistent with predictions concerning thegenerative drawing effect there is partial support for the idea that

1 The exclusion of studentsrsquo prior knowledge in the reported analyses does notchange the reported pattern of results

Table 2Mean proportion correct on the comprehension test and drawing test for two groupsndash Experiment 1

Group Type of test

n Comprehension test Drawing test

M SD M SD

Drawing 24 61 20 52 27Control 24 44 20 28 11

Note Asterisk () indicates significant difference from control group at p lt 05

280 A Schmeck et alContemporary Educational Psychology 39 (2014) 275ndash286

drawing causes students to engage in more generative processingduring learning

Taken together the results suggest that the drawing strategy en-courages students to engage in generative processing during learningas is indicated by their higher learning outcomes Thus the dataprovide further evidence for the generative drawing effect consis-tent with the results of Schwamborn et al (2010) Additionallyresults indicate that students in the drawing condition seem to investmore mental effort than students in the control group without per-ceiving higher levels of difficulty

244 Is there support for the prognostic drawing effectMean proportion correct on drawing accuracy during learning

was 59 (SD = 23) A correlation analysis revealed that the drawing-accuracy score of learner-generated drawings correlated significantlywith the comprehension posttest score r = 620 p lt 001 and withthe drawing posttest score r = 623 p lt 001 Additional correla-tion analyses revealed that the drawing-accuracy score of learner-generated drawings correlated significantly negatively with theperceived difficulty score r = minus489 p = 015 There were no signif-icant correlations between the drawing accuracy score and eitherthe invested mental effort score r = minus134 p = 533 the prior knowl-edge test score r = minus004 p = 984 the spatial ability test score r = 072p = 739 or the motivation test score r = 086 p = 690 Thus as pre-dicted the data provide further evidence for the prognostic drawingeffect consistent with the results of Schwamborn et al (2010)

In sum the results of Experiment 1 are consistent with the pre-diction that students learn better from a science text when they areasked to draw illustrations representing the main ideas of the textand that the quality of the generated drawings during learning cor-relates positively with studentsrsquo text comprehension (egSchwamborn et al 2010 van Meter 2001 van Meter et al 2006)

However it might be argued that the reported results are dueto the way we supported the strategy use In other words the re-ported positive effect of the learner-generated drawing strategy mightnot be caused by studentsrsquo engagement in generative learning ac-tivities during reading (de Jong 2005 Mayer 2004 2009 Wittrock1990) but rather by the additional pictorial information given in thedrawing prompt Additionally looking at studentsrsquo learning out-comes our results indeed show positive effects of drawing howevermean scores of learning outcomes for the drawing group aremedium-sized Thus it might be argued that the way we sup-ported the strategy use was not fully sufficient In other words thereported positive effect of the learner-generated drawing strategyie the generative drawing effect might be increased by giving stu-dents instructional support in addition to the drawing prompt (vanMeter 2001 van Meter et al 2006 van Meter amp Garner 2005) Toaddress these issues we added two experimental conditions byimplementing author-generated pictures in the design in Experi-ment 2

3 Experiment 2

One possible issue with Experiment 1 is the type of control groupused In Experiment 1 following Schwamborn et al (2010) we useda reading only control group in which the control group learnedwith verbal information only In the drawing group however stu-dents not only learned with verbal information but also with pictorialinformation given by the drawing prompt Based on theories of mul-timedia learning the use of different forms of representations suchas texts and pictures can promote learning in that ldquopeople learnbetter from words and pictures than from words alonerdquo (ie mul-timedia principle Mayer 2009 p 223) because in this caseboth a (verbal) propositional representation as well as a (pictori-al) mental model are built up and are optimally integrated into oneschema that can be stored in long-term memory (Schnotz 2005)

This assumption is also in line with the dual-coding approach statedby Paivio (1986) In this regard it might be argued that the re-ported drawing effect is actually a multimedia effect that is basedon the presentation of text and picture rather than a generativedrawing effect that is based on studentsrsquo active engagement indrawing activities during reading In other words instead of askingpeople to draw pictures representing the main ideas of the textgiving them text and author-generated pictures representing themain ideas of the text might be as good or even better Thus weincluded a condition in Experiment 2 in which we added author-generated pictures to the text

An additional issue with Experiment 1 is whether the reportedgenerative drawing effect can be enhanced by using various formsof supporting the strategy First there is evidence that using adrawing prompt during learning seems to be effective in support-ing the learner-generated drawing strategy by minimizing thecreation of extraneous processing (cf Schwamborn et al 2010 seealso Exp 1) Second research has shown that instructing studentsto compare their own drawing with an author-generated picturemight be also effective in supporting the learner-generated drawingstrategy as self-monitoring processes are enhanced (cf van Meter2001) Up to now however there is no empirical evidence whetherthe combination of both ways to support the drawing strategy hasan additive effect on learning outcomes Thus we included a furthercondition in Experiment 2 in which we combined both forms ofstrategy support

The main purpose of Experiment 2 was to test the generativedrawing and prognostic drawing effects of learner-generated drawingas in Experiment 1 but this time also compared with another controlgroup (ie author-generated pictures) Additionally we were in-terested in testing whether the benefits of the learner-generateddrawing strategy can be increased when we instructionally supportstudents not only with a drawing prompt but also with an author-generated picture after the drawing process In this new treatmentwe instructed students to draw a picture of the text content andthen to compare their own drawing with an expert picture

31 Participants and design

The participants were 168 German eighth graders from highertrack secondary schools The mean age was 138 years (SD = 06)and there were 112 girls and 56 boys The study was based on a2 times 2-between-subjects design with learner-generated drawing (yesno) and author-generated picture (yesno) as factors Forty studentsserved in the drawing group 44 students served in the author-generated picture group 41 students served in the drawing + author-generated picture group and 43 students served in the control group

32 Materials

The materials were identical to those used in Experiment 1 exceptthat we used a shortened version of the comprehension pretest thatconsisted of 19 rather than 25 items (Cronbachrsquos alpha = 70) andslightly extended versions of both the comprehension posttest (28items Cronbachrsquos alpha = 84) and the drawing test (four items witha maximum score of 21 points Cronbachrsquos alpha = 78) The pretestwas shortened because the first experiment showed that the re-spective items were either much too easy or much too difficult andthus unsuitable to differentiate between successful and unsuccess-ful learners thus we deleted these items in the second experimentFurthermore we decided to add some items to the comprehen-sion posttest in the second experiment because during data analysisof the first experiment and after receiving some feedback fromexperts in the domain of biology we recognized that a few itemsassessing transfer ability could be added These transfer itemshowever would have been unsuitable to be included in the pretest

281A Schmeck et alContemporary Educational Psychology 39 (2014) 275ndash286

because they are too difficult to answer without prior training inthe topic Additionally author-generated pictures were used in thenew conditions The author-generated pictures were static func-tional pictures representing the main ideas of each paragraph andconsisted of pictorial elements identical to those provided in thedrawing prompt (as shown in Fig 2) These pictures were con-structed by the first author in cooperation with a biology teacher

The drawing version of the booklet was identical to that usedin Experiment 1 (as shown in Fig 1) The control version of the learn-ing booklet was identical to that used in Experiment 1 The author-generated picture version of the booklet consisted of seven pairsof facing pages with a text paragraph on the left page and a corre-sponding author-generated picture (such as in Fig 2) on the rightpage The drawing + author-generated picture version of the bookletcontained the material from the drawing version consisting of sevenpairs of facing pages with a text paragraph on the left page and atwo-part drawing prompt on the right page In addition attachedto each page there was an additional page that students could foldout after having generated their drawing When unfolding this ad-ditional page a picture of that text paragraph right aside the drawingprompt was provided and there was an additional instruction to

compare the learner-generated drawing with the author-generatedpicture Author-generated pictures were the same as in the author-generated picture version of the booklet

33 Procedure

The procedure was identical to that used in Experiment 1 exceptthat there were two additional groups learning with author-generated pictures Students in the author-generated picturecondition were instructed to read the text and additionally to lookat pictures representing the main ideas of each text paragraph Stu-dents in the drawing + author-generated picture version of thebooklet were instructed to read the text to draw pictures for eachtext paragraph using the drawing prompt representing the mainideas of each text paragraph and finally to compare their picturewith an author-generated picture representing main ideas of eachparagraph correctly Students in all groups learned at their own pacewhereby individual learning time was measured by the instruc-tors in the classrooms Again to ensure that studentsrsquo in both drawinggroups did not feel rushed when students in the non-drawing

Fig 2 Author-generated pictures for the seven paragraphs in the author-generated picture versions of the learning booklet Note Pictures are scaled-down from the orig-inal format

282 A Schmeck et alContemporary Educational Psychology 39 (2014) 275ndash286

group completed the task early groups were tested in separateclassrooms

34 Results

341 ScoringAll tests instruments were scored with the same procedures used

in Experiment 1 Again two student assistants (teacher trainees inbiology) scored each of the drawing test items and each of the sevenlearner-generated drawings for each student with acceptableinterrater agreements (drawing test GoodmanndashKruskal gamma of90 drawing-accuracy GoodmanndashKruskal gamma of 94) Actualscores ranged from 1 to 28 points (M = 153 points SD = 58) for thecomprehension test from zero to 21 points (M = 109 points SD = 53)for the drawing test and from 275 to 215 points (M = 141 pointsSD = 48) for drawing accuracy Again total scores of comprehen-sion drawing and accuracy were transferred into z-standardizedscores

342 Are the groups equivalent on basic characteristicsBefore looking at treatment effects on the dependent variables

we analyzed whether the four experimental groups differed onseveral control variables A chi-square analysis indicated that therewere no significant differences regarding gender (p = 097) Sepa-rate univariate analyses of variance (ANOVAs) revealed that thegroups also did not differ significantly on age F lt 1 on spatial abilityF(3 164) = 120 p = 312 or on motivation F(3 164) = 122 p = 305However groups differed significantly on prior knowledge F(3164) = 104 p = 010 partial eta2 = 07 in that students in the controlgroup scored significantly higher on the comprehension pretest(M = 25 SD = 17) than students in both (p lt 05) the author-generated picture group (M = 17 SD = 16) and the drawing + author-generated picture group (M = 15 SD = 12) the drawing group(M = 20 SD = 14) did not differ significantly from the other groupsThus we included studentsrsquo prior knowledge as a covariate in thefollowing analyses

343 Is there support for the generative drawing effectA major goal in this experiment was to determine whether asking

students to generate drawings to represent science text is a moreeffective learning strategy than asking students to learn with textalone or with text and author-generated pictures In other wordswe wanted to determine whether we could replicate and extendthe learner-generated drawing effect Additionally we were inter-ested in whether giving students an author-generated picture afterdrawing can increase the benefits of the learning strategy Mean pro-portion correct and standard deviations on the comprehension anddrawing tests for the four groups are presented in Table 3

The left portion of Table 3 summarizes the mean proportioncorrect on the comprehension test A two-factorial analysis ofcovariance (ANCOVA) predicting learning outcomes (comprehen-sion posttest score) with learner-generated drawing (yesno) and

author-generated picture (yesno) as the factorial independent vari-ables and prior knowledge as a covariate showed a significantpositive main effect of learner-generated drawing F(1 163) = 398p = 048 partial eta2 = 02 a significant interaction effect F(1163) = 626 p = 013 partial eta2 = 04 but no main effect of author-generated pictures F lt 1 In addition multiple pairwise comparisons(with p lt 05) showed that the drawing group performed signifi-cantly better than each of the three other groups which did not differsignificantly from each other Cohenrsquos d favoring the drawing groupover the author-generated picture group was 49 over the learner-generated + author-generated picture group was 57 and over thecontrol group was 52

The right portion of Table 3 summarizes the mean proportioncorrect on the drawing posttest Again a two-factorial analysis ofcovariance (ANCOVA) predicting learning outcome (drawing testscore) with learner-generated drawing (yesno) and author-generatedpicture (yesno) as the factorial independent variables and priorknowledge as a covariate showed a significant positive main effectof learner-generated drawing F(1 163) = 6260 p lt 001 partialeta2 = 28 a significant positive main effect of author-generated pic-tures F(1 163) = 1104 p = 001 partial eta2 = 06 and a significantinteraction effect F(1 163) = 1658 p lt 001 partial eta2 = 09 In ad-dition multiple pairwise comparisons (with p lt 05) showed thatboth the drawing group and the drawing + author-generated picturegroup performed significantly better than the author-generatedpicture group (d = 68 d = 59) and the control group (d = 187d = 188) In turn the author-generated picture group performed sig-nificantly better than the control group (d = 95) The drawing groupand the drawing + author-generated picture group did not differ sig-nificantly from each other (d = 15)2 Overall these results areconsistent with Experiment 1 and provide additional support forthe generative drawing effect

In accordance with Experiment 1 we were interested in whetherdifferences in learning time among the experimental groups mediatethe positive effect of drawing on text comprehension First an ANOVApredicting learning time with learner-generated drawing (yesno)and author-generated picture (yesno) as the factorial indepen-dent variables showed a significant main effect of learner-generateddrawing F(1 164) = 39226 p lt 001 partial eta2 = 71 a significantmain effect of author-generated picture F(1 164) = 1685 p lt 001partial eta2 = 09 and a significant interaction effect F(1 164) = 490p = 028 partial eta2 = 03 Linear contrasts (with p lt 05) revealedthat the drawing group (M = 1938 min SD = 380) and thedrawing + author-generated picture group (M = 2340 min SD = 551)needed significantly more learning time than the author-generatedpicture group (M = 960 min SD = 397) and the control group(M = 834 min SD = 251) Thus to test whether learning time me-diates the positive effect of drawing on text comprehensionadditional mediation analyses (Baron amp Kenny 1986) were calcu-lated by including learning time as an additional predictor in theaforementioned linear model Results of the mediation analysesshowed that the effects of drawing on the comprehension posttestand the drawing posttest scores (see multiple pairwise compari-sons) are mediated by learning time to some extent That is includinglearning time in the linear model for predicting comprehension testscores still revealed a positive effect of the drawing group com-pared with the drawing + author-generated group on thecomprehension test (p = 012) However including learning time inthe linear model for predicting comprehension posttest scoresreduced the positive effect of the drawing group compared with theauthor-generated picture group (from p = 034 to p = 281) as wellas compared with the control group (from p = 002 to p = 087) being

2 The exclusion of studentsrsquo prior knowledge in the reported analyses does notchange the overall pattern of results

Table 3Mean proportion correct on the comprehension test and drawing test for the fourgroups ndash Experiment 2

Group Type of test

n Comprehension test Drawing test

M SD M SD

Learner-generated drawing 40 63 22 66 22Author-generated picture 44 53 19 50 25Learner-generated drawing +

author-generated picture41 51 20 63 19

Control 43 52 20 30 16

Note Asterisk () indicates significant difference from control group at p lt 05

283A Schmeck et alContemporary Educational Psychology 39 (2014) 275ndash286

no longer statistically significant Regarding the drawing posttestscore including learning time does not change the reported patternof results except that the positive effect of the drawing + author-generated picture group compared with the author-generated picturegroup is no longer statistically significant (from p = 004 to p = 223)

There were neither main effects of learner-generated drawingand author-generated pictures on the mental effort item (drawinggroup M = 455 SD = 025 author-generated picture group M = 459SD = 024 drawing + author-generated picture group M = 444SD = 026 control group M = 481 SD = 024 F lt 1) nor on theperceived difficulty item (drawing group M = 363 SD = 023 author-generated picture group M = 371 SD = 022 drawing + author-generated picture group M = 395 SD = 023 control group M = 393SD = 022 F lt 1)

Taken together the drawing strategy apparently fosters stu-dents to engage in generative activities indicated by their higherlearning outcomes Thus the data provide further evidence for thegenerative drawing effect predicted by Schwamborn et al (2010)In Experiment 2 benefits of the drawing activity however are me-diated by learning time and do not involve higher mental effortAdditionally there was no increased benefit when additional drawingsupport was available in the form of author-generated pictures

344 Is there support for the prognostic drawing effectA second major goal of this study was to determine whether the

prognostic drawing effect could be extended to a new context Meanproportion correct on drawing-accuracy during learning was 60(SD = 04) for the drawing group and 68 (SD = 03) for thedrawing + author generated picture group This difference betweenthe two drawing groups is not significant F(1 79) = 252 p = 116This lack of group differences allowed us to pool the data of bothdrawing groups for subsequent correlation analyses Correlation anal-yses based on the combined data from the two drawing groupsrevealed that the drawing-accuracy score of learner-generated draw-ings correlates significantly with the comprehension posttest scorer = 470 p lt 001 as well as with the drawing posttest score r = 615p lt 001 Additional correlation analyses revealed that the drawing-accuracy score of learner-generated drawings did not correlatesignificantly with the prior knowledge test score r = 095 p = 400the spatial ability test score r = 127 p = 257 the motivation testscore r = 033 p = 769 or the mental effort test score r = 042p = 712 The correlation between the drawing-accuracy score andthe perceived difficulty score was only slightly statistical signifi-cance r = minus218 p = 053 Thus the data provide further evidencefor the prognostic drawing effect consistent with the results ofSchwamborn et al (2010)

In sum results of Experiment 2 are partly consistent with theresults of Experiment 1 in that students learn better from a sciencetext when they are asked to draw illustrations representing the mainideas of the text and the quality of the generated drawings duringlearning correlates positively with studentsrsquo text comprehension (egSchwamborn et al 2010 van Meter 2001 van Meter et al 2006)

4 Discussion

41 Empirical contributions

The present set of experiments makes three empirical contri-butions to the field First this study shows strong and consistentevidence that students who are asked to generate drawings (withsufficient support) during reading a scientific text that describes acausal sequence perform better than students who read withoutdrawing both on a comprehension test (d = 085 in Experiment 1and d = 052 in Experiment 2) and on a drawing test (d = 115 in Ex-periment 1 and d = 187 in Experiment 2) Thus the generativedrawing effect can be extended to a new domain and therefore

corresponds to Shavelson and Townersquos (2002 p 4) recommenda-tion to ldquoreplicate and generalize across studiesrdquo as one of the sixessential scientific principles of educational research

Second this study shows strong and consistent evidence that thequality of drawings that students generate during learning with ascientific text that describes a causal sequence is positively relatedto subsequent performance on tests of learning outcome includ-ing a comprehension test (r = 623 in Experiment 1 and r = 470 inExperiment 2) and a drawing test (r = 620 in Experiment 1 andr = 615 in Experiment 2) Thus the prognostic drawing effect canbe replicated and extended to a new domain consistent with stan-dards for scientific research in education prescribed by Shavelsonand Towne (2002)

Third this study shows that asking learners to draw picturesduring reading a scientific text (ie learner-generated drawing groupin Experiment 2) is more effective than simply providing draw-ings (ie author-generated picture group in Experiment 2) both ona comprehension test (d = 049) and a drawing test (d = 068) Sim-ilarly adding author-generated drawings (ie learner-generatedpictures + author-generated pictures group in Experiment 2) doesnot improve the learning outcomes of students who also draw pic-tures during learning (ie learner-generated pictures group inExperiment 2) either on a comprehension test (d = minus057) or adrawing test (d = minus015) In short the act of drawing during learn-ing (with sufficient support) improves learning beyond the simpleprovision of drawings

42 Theoretical contributions

The results are consistent with the idea that drawing during learn-ing serves as a generative activity (Mayer amp Wittrock 2006Schwamborn et al 2010 van Meter amp Garner 2005 Wittrock 1990)That is the act of drawing encourages learners to engage in gen-erative cognitive processing during learning such as organizing therelevant information into a coherent structure and integrating itwith relevant prior knowledge from long-term memory In thepresent study positive effects of drawing were indicated with a com-prehension and a drawing learning outcome test and therefore arein line with the theoretical assumption derived from the GTDC thatbenefits of drawing can be found if learning outcome tests are usedthat are sensitive to the underlying process of drawing (cf van Meteramp Garner 2005) Additionally in our study the drawing activity wassupported in a way that was intended to help learners carry out theunderlying cognitive processes of drawing (ie selecting organiz-ing and integrating) successfully In this regard results of the presentstudy might supplement the theoretical framework of learner-generated drawing by providing further evidence that benefits ofdrawing defined by van Meter and Garnerrsquos GTDC can diminish ifno instructional support is given to constrain and structure thedrawing activity However a fuller understanding of the underly-ing cognitive processes of drawing and how these processes canbe influenced via drawing support requires more direct measuresof cognitive processing during learning Additionally following theidea that metacognitive processes of monitoring and regulation areautomatically activated by drawing (van Meter amp Garner 2005) afuller understanding of the metacognitive effects of drawings is alsorequired

43 Practical contributions

The present study encourages instructional designers and in-structors to incorporate drawing activities into venues involvinglearning from text which we call the generative drawing effect Oneimportant feature of a successful drawing strategy that is presentin this study and in a previous study by Schwamborn et al (2010)is that the drawing activity was supported by providing a

284 A Schmeck et alContemporary Educational Psychology 39 (2014) 275ndash286

background scene and a legend showing how to represent eachelement to constrain and structure the drawing activity Thus animportant practical implication is that students may need supportin their drawing activity so they do not need to draw from scratch

The present study also suggests a potentially useful diagnostictool to gauge the depth of student learning namely the quality ofthe drawings created by students during learning which we referto as the prognostic drawing effect Incorporating a measure of thequality of a learnerrsquos drawing during learning can be a useful toolin developing remedial instruction to give learners individual supportIt may be important to use materials that explain a cause-and-effect process and give learners drawings of the elements they needto represent the process pictorially Asking learners to simply drawpictures of elements is unlikely to be helpful whereas asking themto generate drawings that show the relations among the elementsin a schematic form is more likely to be helpful

44 Limitations and future directions

Some limitations and future directions of our study should beaddressed As noted in the theoretical contributions subsection wedid not have direct measures of cognitive processing during learn-ing so it is not possible to pinpoint how the drawing activity affectedspecific cognitive processes such as attending to relevant informa-tion organizing it and integrating it with prior knowledge We alsodid not assess metacognitive processing during learning thus it isnot possible to pinpoint how the drawing activity affected specificmetacognitive processes such as monitoring and regulation

Furthermore results of the cognitive load rating scales (in-vested mental effort and perceived task difficulty) are inconsistentWhereas in Experiment 1 an effect on mental effort but not on per-ceived task difficulty showed up (ie students in the drawing grouprated their invested mental effort during learning significantlyhigher) no effects on mental effort and task difficulty were foundin Experiment 2 Additionally in both experiments only a nega-tive correlation of perceived task difficulty with the quality of learner-generated pictures appears but no correlation of mental effort withthe quality Following de Jong (2010) those cognitive load ratingscales might have the disadvantages that they do not give a con-current measure of cognitive load and do not measure an essentialconcept in cognitive load theory namely cognitive overload (p 125)Future studies on learner-generated drawing might also use othercognitive load measures such as physiological measures as moredirect indicators of cognitive load

As noted in the practical contributions subsection we showedthe drawing effects by using a scientific text describing how a cause-and-effect system works that is the causal steps regarding aninfection with influenza and the immune response It might be pos-sible however that for other types of text producing drawings mightharm rather than promote text comprehension Thus to test whetherthe reported drawing effects can be extended future research hasto focus on other types of text such as descriptive texts as well ason other types of relations that can be conveyed with other typesof representations such as compare and contrast relations whichcan be shown in a matrix Additionally studentsrsquo learning out-comes were tested immediately after reading thus future work isneeded to investigate the longer-term effects of generative drawingon learning outcomes

Furthermore we only compared drawing with control groupsthat received no further learning strategy instructions However en-gaging in generative learning activities such as drawing requires aconsiderable amount of time Accordingly results showed that forExperiment 2 the positive effect of the drawing group on text com-prehension compared with the author-generated picture group andto the control group was mediated by learning time To rule out thatthe effects of drawing result only from additional time on task instead

of the generative activity future research should also compare thedrawing strategy with other time demanding generative learningstrategies such as summarization (cf Leopold amp Leutner 2012)

Another point that should be noted is that students in both ex-periments received some kind of multimedia materials in that evenwhen they had to draw and did not see presented pictures they wereat least provided with the basic (visual) elements for their draw-ings which they had to do on the given background which thusalso contained information In other words when students are pre-sented with important elements of the drawings which they canuse to draw themselves they will not have to put as much effortinto summarizing visually what they have just read compared withstudents who have to draw without any instructional help Futurestudies might also compare the drawing group with a summariza-tion group in which students receive a set of verbal key terms thatare similar to the drawing elements and are asked to make a textualsummary

Additionally future research is needed to validate the prognos-tic drawing effect So far we know that the quality of learner-generated pictures is related to studentsrsquo learning outcomes (iethe higher the learning outcome the higher the drawing accuracyand vice versa) and their perceived difficulty (ie the lower the per-ceived difficulty the higher the drawing accuracy and vice versa)and that it is not related to studentsrsquo prior knowledge motivationspatial ability or mental effort However less is known about whatthis might mean That is less is known regarding the causal direc-tion of this relation or the presence of a possible further moderatorvariable Do studentsrsquo efforts to produce accurate drawings lead tobetter comprehension and lower perceived difficulty Or do stu-dents who are more adept in drawing benefit more from the strategyand thus perceive the difficulty of the learning materials as beinglower Both arguments seem convincing

Finally more work is needed to determine the level of supportthat makes the drawing strategy most effective for various kinds oflearners As noted in the empirical contribution adding author-generated drawings (ie learner-generated pictures + author-generated pictures group in Experiment 2) does not improve thelearning outcomes of students who also draw pictures during learn-ing and were supported by a drawing prompt In other words thecombination of two ways of supporting the drawing strategy (iegiving a drawing prompt during reading plus an author-generatedpicture after reading) did not improve studentsrsquo learning out-comes compared with students in the drawing group as well ascompared with students in the control and author-generated pic-tures only groups This result is inconsistent with previous research(eg van Meter 2001 van Meter et al 2006) which found that com-paring own drawings to author-generated pictures normally helpslearning van Meter and colleagues (2001 2006) however provid-ed author-generated pictures plus prompting questions after thedrawing process That is students answered prompting questionsto guide the comparison process between their self-generateddrawing and the author-generated drawing In our study studentswere only instructed to generate a drawing to inspect an author-generated one and to check whether their own drawing incomparison with the author-generated one really represented themain ideas of the text paragraph correctly In other words we didnot guide the process of comparing self-generated drawings withauthor-generated ones As a potential consequence students per-formed the intended comparison process inadequately or even notall and thus did not benefit from it One reason for this inade-quate comparison process might be that students need guidancein doing the comparison process Another reason might be the factthat students do not seriously engage in generating drawings oncethey notice that there are author-generated drawings Thus futureresearch should also use additional guidance to test whether thecombination of different ways of supporting the drawing strategy

285A Schmeck et alContemporary Educational Psychology 39 (2014) 275ndash286

(ie giving a drawing prompt during reading plus an author-generated picture after reading) helps learning as well asobservational measures of the drawing process itself (ie think aloudprotocols) to shed more light on the cognitive processes underly-ing the drawing activities

Overall drawing during learning from text appears to be a po-tentially powerful strategy for improving studentsrsquo learning fromscientific text when certain boundaries and prerequisites are takeninto account

Acknowledgments

This article is based on a research project funded by the GermanResearch Foundation (DFG LE 6459-3 as part of FOR 511) We wouldlike to thank Angela Sandmann for her assistance in developing thelearning materials

References

Ainsworth S Prain V amp Tytler R (2011) Drawing to learn in science Science 3331096ndash1097

Alesandrini K L (1981) Pictorial-verbal and analytic-holistic learning strategies inscience learning Journal of Educational Psychology 73 358ndash368

Alesandrini K L (1984) Pictures and adult learning Instructional Science 13 63ndash77Baron R M amp Kenny D A (1986) The moderator-mediator variable distinction in

social psychological research Conceptual strategic and statistical considerationsJournal of Personality and Social Psychology 51 1173ndash1182

Bruumlnken R Plass J L amp Leutner D (2003) Direct measurement of cognitive loadin multimedia learning Educational Psychologist 38 53ndash61

Carney RN amp Levin JR (2002) Pictorial illustrations still improve studentsrsquo learningfrom text Educational Psychology Review 14 5ndash26

de Jong T (2005) The guided discovery principle in multimedia learning In R EMayer (Ed) The Cambridge handbook of multimedia learning (pp 215ndash228) NewYork Cambridge University Press

de Jong T (2010) Cognitive load theory educational research and instructionaldesign Some food for thought Instructional Science 38 105ndash134

Ekstrom R B French J W amp Harman H H (1976) Manual for kit of factor-referencedcognitive tests Princeton NJ Educational Testing Service

Greene T R (1989) Childrenrsquos understanding of class inclusion hierarchies Therelationship between external representation and task performance Journal ofExperimental Child Psychology 48 62ndash89

Hall V C Bailey J amp Tillman C (1997) Can student-generated illustrations be worthten thousand words Journal of Educational Psychology 89 677ndash681

Houmlffler T N (2010) Spatial ability Its influence on Learning with visualizations ndashA meta-analytic review Educational Psychology Review 22 245ndash269

Houmlffler T N Schmeck A amp Opfermann M (2013) Static and dynamic visualrepresentations Individual differences in processing In G Schraw M TMcCrudden amp D Robinson (Eds) Learning through visual displays (pp 133ndash163)Charlotte NC Information Age Publishing

Kalyuga S Chandler P amp Sweller J (1999) Managing split-attention and redundancyin multimedia instruction Applied Cognitive Psychology 13 351ndash371

Leopold C (2009) Lernstrategien und Textverstehen [Learning strategies and textcomprehension] Muumlnster Waxmann

Leopold C amp Leutner D (2012) Science text comprehension Drawing main ideaselection and summarizing as learning strategies Learning and Instruction 2216ndash26

Lesgold A M DeGood H amp Levin J R (1977) Pictures and young childrenrsquos proselearning A supplementary report Journal of Reading Behavior 9 353ndash360

Lesgold A M Levin J R Shimron J amp Guttman J (1975) Pictures andyoung childrenrsquos learning from oral prose Journal of Educational Psychology 67636ndash642

Leutner D Leopold C amp Sumfleth E (2009) Cognitive load and science textcomprehension Effects of drawing and mentally imagining text contentComputers in Human Behavior 25 284ndash289

Mayer R E (2004) Should there be a three-strikes rule against pure discoverylearning The case for guided methods of instruction The American Psychologist59 14ndash19

Mayer R E (2005) Cognitive theory of multimedia learning In R E Mayer (Ed)The Cambridge handbook of multimedia learning (pp 31ndash48) New York CambridgeUniversity Press

Mayer R E (2009) Multimedia learning (2nd ed) New York NY CambridgeUniversity Press

Mayer R E amp Wittrock M C (2006) Problem solving In P Alexander P Winne ampG Phye (Eds) Handbook of educational psychology (pp 287ndash303) Mahwah NJErlbaum

Paas F (1992) Training strategies for attaining transfer of problem-solving skill instatisticsmdashA cognitive-load approach Journal of Educational Psychology 84429ndash434

Paas F Tuovinen J Tabbers H K amp Van Gerven P W M (2003) Cognitive loadmeasurement as a means to advance cognitive load theory EducationalPsychologist 38 63ndash71

Paivio A (1986) Mental representation A dual coding approach New York OxfordUniversity Press

Pashler H Bain P Bottage B Graesser A Koedinger K McDaniel M et al (2007)Organizing instruction and study to improve student learning Washington DCNational Center for Educational Research

Rasco R W Tennyson R D amp Boutwell R C (1975) Imagery instructions anddrawings in learning prose Journal of Educational Psychology 67 188ndash192

Rheinberg F Vollmeyer R amp Burns B D (2001) FAM Ein fragebogen zurerfassung aktueller motivation in lern- und leistungssituationen [QCM Aquestionnaire to assess current motivation in learning situations] Diagnostica47 57ndash66

Schnotz W (2005) An integrated model of text and picture comprehension In RE Mayer (Ed) The Cambridge handbook of multimedia learning (pp 49ndash70) NewYork Cambridge University Press

Schwamborn A Mayer R E Thillmann H Leopold C amp Leutner D (2010) Drawingas a generative activity and drawing as a prognostic activity Journal of EducationalPsychology 102 872ndash879

Schwamborn A Thillmann H Opfermann M amp Leutner D (2011) Cognitive loadand instructionally supported learning with provided and learner-generatedvisualizations Computers in Human Behavior 27 89ndash93

Shavelson R J amp Towne L (Eds) (2002) Scientific research in education WashingtonDC National Academy Press

Sweller J Ayres P amp Kalyuga S (2011) Cognitive Load Theory New York SpringerTirre W C Manelis L amp Leicht K (1979) The effects of imaginal and verbal strategies

on prose comprehension by adults Journal of Reading Behavior 11 99ndash106van Meter P (2001) Drawing construction as a strategy for learning from text Journal

of Educational Psychology 69 129ndash140van Meter P Aleksic M Schwartz A amp Garner J (2006) Learner-generated drawing

as a strategy for learning from content area text Contemporary EducationalPsychology 31 142ndash166

van Meter P amp Garner J (2005) The promise and practice of learner-generateddrawings Literature review and synthesis Educational Psychology Review 12261ndash312

Van Gog T amp Paas F (2008) Instructional efficiency Revisiting the original constructin educational research Educational Psychologist 43 16ndash26

Vollmeyer R amp Rheinberg F (2000) Does motivation affect learning via persistenceLearning and Instruction 4 293ndash309

Weinstein C E amp Mayer R E (1986) The teaching of learning strategies In M CWittrock (Ed) Handbook of research on teaching (pp 315ndash327) New YorkMacmillan

Wittrock M C (1990) Generative processes of comprehension EducationalPsychologist 24 345ndash376

286 A Schmeck et alContemporary Educational Psychology 39 (2014) 275ndash286

  • Drawing pictures during learning from scientific text testing the generative drawing effect and the prognostic drawing effect
  • Introduction
  • Theoretical framework for the learner-generated drawing strategy
  • Empirical framework for the learner-generated drawing strategy
  • Effectiveness of learner-generated drawings
  • Quality of learner-generated drawings
  • Overview of the experiments
  • Experiment 1
  • Participants and design
  • Materials
  • Procedure
  • Results and discussion
  • Scoring
  • Are the groups equivalent on basic characteristics
  • Is there support for the generative drawing effect
  • Is there support for the prognostic drawing effect
  • Experiment 2
  • Participants and design
  • Materials
  • Procedure
  • Results
  • Scoring
  • Are the groups equivalent on basic characteristics
  • Is there support for the generative drawing effect
  • Is there support for the prognostic drawing effect
  • Discussion
  • Empirical contributions
  • Theoretical contributions
  • Practical contributions
  • Limitations and future directions
  • Acknowledgments
  • References

the quality of studentsrsquo drawings during learning and their learn-ing outcomes (eg Greene 1989 Hall et al 1997 Lesgold et al1975 1977 Schwamborn et al 2010 van Meter 2001 van Meteret al 2006) The quality of learner-generated drawing is also re-ferred to as the drawing accuracy (eg van Meter 2001 van Meteret al 2006) and is defined as ldquothe degree to which completeddrawings resemble the represented object(s)rdquo (van Meter amp Garner2005 p 299) In a study by Hall et al (1997) for example collegestudents learning a mechanics lesson with the instruction to drawproduced better learning outcomes on a transfer test than a textonly control group but only if they produced higher quality draw-ings In the study by Schwamborn et al (2010) students who wereable to generate high accuracy drawings scored significantly higheron learning outcome tests than did those who generated loweraccuracy drawings In addition to this Schwamborn and col-leagues also found that the quality of the generated drawings duringlearning correlated positively with the comprehension test scoresBased on these results Schwamborn et al (2010) proposed the prog-nostic drawing effect Students who produce high-quality drawingsduring reading a scientific text tend to score better on posttestsof learning outcome than do students who produce low-qualitydrawings during reading

13 Overview of the experiments

Research on drawing ndash ie the generative and the prognosticdrawing effect ndash is promising however at the present time thereis a need for a more solid evidence base and for a closer examina-tion of theoretical issues First the generalizability is limited atthis point as replication studies using learning outcome tests thatare sensitive to the underlying process of drawing as well as newlearning materials other than the washing lesson (eg Schwambornet al 2010) or the birds wing (van Meter 2001) are yet missingIn their report for the US National Research Council entitledScientific Research in Education Shavelson and Towne (2002 p 4)for example highlighted the need to ldquoreplicate and generalizeacross studiesrdquo as one of the six essential scientific principles ofeducational research It has to be mentioned at this point that whengeneralizing results to new domains or lessons one should care-fully consider whether these are comparable at all In ourexperiments we aimed at generalizing results by Schwamborn et al(2010) who worked with a science text explaining the causalsteps regarding the chemistry of washing to a new lesson that ishowever comparable in that the text we used also describedcausal steps of a process in this case regarding the infection withinfluenza and regarding the immune response That is althoughthere were differences between the two domains (chemistry versusbiology) the lessons showed structural similarities and thusallow for comparing results and drawing conclusions regardinggeneralizability

Second research on drawing indicates that some form of supportis needed to assist learners during drawing Schwamborn et al(2010) for example introduced a drawing prompt as helpful supportfor learners to benefit from drawing They proposed that the re-sulting positive drawing effect is based on studentsrsquo engagementin generative learning activities during reading due to drawing (con-sistent with the GTDC of van Meter amp Garner 2005 see also de Jong2005 Mayer 2004 2009 Wittrock 1990) However the results re-ported by Schwamborn et al (2010) might rather reflect amultimedia effect (Mayer 2005 2009) than the proposed drawingeffect as the learning lesson used ndash a scientific text and a drawingprompt consisting of pictorial elements and backgrounds ndash createda multimedia lesson In other words the results of Schwamborn et al(2010) might not be due to the drawing activity but rather due tothe multimedia effect that students ldquolearn better from words andpictures than from words alonerdquo (Mayer 2009 p 223) In this case

the words are presented in the lessons and pictures are generatedby the students so a control group that receives author-generatedpictures is warranted

Third research on drawing mostly used only one way to supportthe drawing strategy at a specific time That is instructionalsupport was added during learning (ie by using cut-out figures ora drawing prompt (cf Lesgold et al 1975 1977 Schwamborn et al2010) or after learning (ie by providing pictures van Meter 2001van Meter et al 2006) Less is known about whetheradding instructional support not only during learning but also afterlearning can additionally enhance the benefits of the drawingstrategy

Fourth research on drawing should include motivational andcognitive aspects that may have an impact on the effectiveness ofthe learner-generated drawing strategy Studentsrsquo current motiva-tion for example is a one condition for successful learning A studentfor example who has low motivation to learn may invest lesseffort in learning than students who are highly motivated to learn(cf Vollmeyer amp Rheinberg 2000) Studentsrsquo spatial ability maybe a further condition for successful learning when workingwith visualizations (cf Houmlffler 2010 Houmlffler Schmeck amp Opfermann2013) A high-spatial-ability student for example may haveadvantages in learning with visualizations compared with a low-spatial-ability student That is preexisting motivational and cogni-tive differences between students before learning might havean influence on the learning outcome and thus should becontrolled

In addition recent research has shown that not only experi-mental conditions (such as the kind of picture) and the abovementioned ldquoclassicalrdquo covariates can have an impact on how suc-cessful learning takes place but that these effects can be mediatedby the amount of mental effort someone invests while learning orworking on a lesson and by how difficult someone perceives adomain or lesson to be (cf Leutner et al 2009 SchwambornThillmann Opfermann amp Leutner 2011) These aspects of cogni-tive load (invested mental effort and perceived task difficulty) werethus included as additional variables in our studies as well Thuswe conducted the following two experiments using a science textexplaining the biological process of influenza In Experiment 1 weimplemented an experimental drawing condition and a reading onlycontrol condition in order to determine how both the generativeand the prognostic drawing effect would extend to a new contextAnalogous to the study of Schwamborn et al (2010) students inthe drawing condition received a baseline instructional support bymeans of a drawing prompt that included a legend showing all therelevant elements for drawing and a partly pre-drawn back-ground for their drawing (as shown in Fig 1)

In Experiment 2 we again implemented an experimental drawingcondition and a reading only control condition and we addition-ally implemented author-generated pictures in order to test whetherthe generative drawing effect was caused by the simple presenceof illustrations rather than the generation of illustrations That iswe implemented a text plus picture condition (which we called theauthor-generated picture condition) to test whether the reportedgenerative drawing effect of Schwamborn et al (2010) is based onstudentsrsquo engagement in generative learning activities during readingrather than on the pictorial representations given by the drawingprompt In addition we implemented a drawing plus picture con-dition (in which students both draw and are given a picture) to testwhether the reported generative drawing effect can be enhancedby instructing students to compare their own drawing with anauthor-generated picture In short we tested whether combiningdifferent forms of support to the drawing strategy additionally en-hances the benefits of the drawing strategy In both experimentslearning outcome tests that are sensitive to the underlying processof drawing were used

277A Schmeck et alContemporary Educational Psychology 39 (2014) 275ndash286

2 Experiment 1

21 Participants and design

Forty-eight German eighth graders in higher track secondaryschools participated in this study The mean age was 137 years(SD = 06) and there were 22 girls and 26 boys The study was basedon a between-subjects design with two levels of text learning(learner-generated drawings and control) as the single factor Twenty-

four students served in the control group and 24 served in thedrawing group

22 Materials

All materials were paper-pencil based The materials consistedof five adjunct questionnaires two learning booklets two cogni-tive load rating scales and two posttests The five adjunctquestionnaires were intended to determine whether the groups were

Fig 1 Screenshot of the drawing prompt for the first paragraph in the drawing versions of the learning booklet Note Translated from the German original

278 A Schmeck et alContemporary Educational Psychology 39 (2014) 275ndash286

equivalent on basic characteristics They included a participant ques-tionnaire a comprehension pretest a spatial ability test and amotivation questionnaire The participant questionnaire solicited in-formation concerning the studentsrsquo age and sex The comprehensionpretest consisted of 25 multiple-choice items and was intended toassess studentsrsquo prior-knowledge of information covered in the textStudentsrsquo spatial ability was measured with a 10 multiple-choicepaper-folding items taken from a battery of cognitive tests devel-oped by Ekstrom French and Harman (1976) The motivationquestionnaire assessed studentsrsquo current motivation for doing thelearning task after reading the instructions before the lesson It con-sisted of nine items from the challenge and interest subscales of theQuestionnaire on Current Motivation (QCM) developed by RheinbergVollmeyer and Burns (2001) Cognitive load by means of investedmental effort was measured using the 7-point subjective rating scaledeveloped by Paas (1992) which ranges from (1) very low mentaleffort to (7) very high mental effort Cognitive load by means of per-ceived task difficulty was measured using the 7-point subjective ratingscale developed by Kalyuga Chandler and Sweller (1999) whichranges from (1) very easy to (7) very difficult These subjective mea-sures have been criticized for assessing cognitive load with only singleitems (eg Bruumlnken Plass amp Leutner 2003) However several studiesshowed the effectiveness of the rating scale by showing that the vari-ation in learnersrsquo cognitive load ratings depended on variations intask complexity or instructional design (for overviews see PaasTuovinen Tabbers amp Van Gerven 2003 Van Gog amp Paas 2008) Inthis regard Sweller Ayres and Kalyuga (2011) conclude that ldquothesimple subjective rating scale regardless of the wording used (mentaleffort or difficulty) has perhaps surprisingly been shown to be themost sensitive measure available to differentiate the cognitive loadimposed by different instructional proceduresrdquo (p 74) For that reasonand due to the economic applicability we decided to use this kindof cognitive load measurement while acknowledging the limita-tions of a short self-report instrument

The two learning booklets each included a science text on thebiology of the influenza The text explained the causal steps re-garding an infection with influenza and regarding the immuneresponse which is an unfamiliar subject for eighth graders in highertrack secondary schools due to the German curriculum The text con-sisted of approximately 850 words (in German) and was divided intoseven paragraphs (as shown in Table 1)

The drawing version of the booklet contained seven pairs of facingpages with a text paragraph on the left page and a two-part drawingprompt on the right page The first part of the drawing prompt in-cluded a legend showing all the relevant elements (in total eightelements) for drawing a picture for that text paragraph (as shownin the top of Fig 1) The second part of the drawing prompt in-cluded a partly pre-drawn background for studentsrsquo drawing (asshown in the bottom of Fig 1) Overall students had to make sevendrawings ie one drawing to each paragraph

The control version of the learning booklet contained four pairsof facing pages with one of the seven text paragraph on each page

Students in both groups learned with exactly the same text mate-rial To make sure that students in the control group learnedwith the same amount of information as students in the drawinggroup all elements of the drawing prompt as well as the spatialrelations between these elements were also described in the sciencetext

The two posttests intended to assess the learning outcomes werea comprehension posttest and a drawing posttest The comprehen-sion posttest (Cronbachrsquos alpha = 083) consisted of 25 multiple-choice items (the same items as in the comprehension pretest) andwas intended to assess studentsrsquo comprehension of the factual andconceptual information covered in the text as well as their abilityto transfer what was presented to new situations An item exampleis ldquoT-helper cells do not only recognize viruses but also agents thatare extraneous to the body Which medication would you admin-ister to a patient who has received a new kidney (a) a medicinethat suppresses the immune response of the body (b) a medicinethat activates the immune response of the body (c) a medicine thatcontains antigens or (d) a medicine that contains blood of the kidneydonorrdquo [(a) is the correct answer] The drawing test (Cronbachrsquosalpha = 081) was intended to assess studentsrsquo comprehension of theconceptual information presented in the science text by means ofdrawing That is students had to reproduce the main ideas givenin the text by drawing It consisted of three drawing items in whichstudents were asked to draw sketches depicting key concepts of thetext and their spatial relations An item example for the drawingtest is ldquoHow does an influenza virus invade a cell and how is it re-producedrdquo The science text the drawing prompt and the learningoutcome tests were constructed by the first author in cooperationwith a biology teacher The materials were adapted fromSchwamborn et al (2010) however using another science domainand including measures of individual learning times and cognitiveload

23 Procedure

Participants were tested in the schoolsrsquo classrooms Within theirclasses they were randomly assigned to one of the two groupsGroups were tested in separate classrooms in order to insure thatstudents in the drawing group did not feel rushed when studentsin the control group completed the task early Each student wasseated at an individual desk First students were given the partic-ipant questionnaire and the comprehension pretest to complete attheir own rate Second students filled in the paper-folding test witha 3 min time limit Third students were given instructional book-lets corresponding to their assigned group After they had read theinstructions for reading the booklets studentsrsquo current motivationfor doing the learning task was assessed Next students started learn-ing with the text material corresponding to their treatment groupStudents were instructed to carefully read the text on the biologyof the influenza in order to comprehend the material Students inthe drawing condition were instructed to read the text and addi-tionally to draw pictures for each text paragraph using the drawingprompt representing the main ideas of each text paragraph Thatis students had to use the pictorial elements given in the legendsuch as the virus as templates for their own paper-pencil baseddrawing across the pre-dawn background Students in the controlgroup were instructed to read the text for comprehension but werenot instructed to engage in drawing Students in both groups learnedat their own pace whereby individual learning time was mea-sured by the instructors in the classrooms Fourth in order to ensurecomparable testing procedures after finishing learning with thewhole learning material students in both groups directly rated theamount of mental effort he or she had invested during learning andthe amount of difficulty he or she had perceived during learningFifth students received the comprehension posttest consisting of

Table 1Text from the second paragraph of the influenza lesson

How the influenza virus replicates

Once inside the influenza virus uses your somatic cell to produce new particlesof the influenza-virus The glycoproteins move toward the membrane of thesomatic cell and stick out into the outside of the cell The capsules of thevirus however are assembled inside the somatic cell Next these newassembled capsules of the virus leave your somatic cell By moving throughthe somatic cell membrane the capsules are enveloped with the membraneand its glycoproteins which then plays the role of the virus membraneThus several new influenza viruses are located outside your somatic cell

Note Translated from the German original

279A Schmeck et alContemporary Educational Psychology 39 (2014) 275ndash286

the comprehension posttest and the drawing posttest Students had20 minutes time for completion and did not have access to thescience text or their drawings Finally students were thanked anddebriefed As students learned at their own rate the whole proce-dure took about 70ndash90 minutes depending on the individual testingtimes

24 Results and discussion

241 ScoringThe dependent variables were studentsrsquo scores on the compre-

hension and drawing posttests studentsrsquo rating on the mental effortand the difficulty scales and the drawing accuracy score indicat-ing the quality of learner-generated drawings produced by studentsin the drawing group during the learning phase

The comprehension test score (pre- and posttest) for each studentwas computed by awarding 1 point for each correct answer and byadding up the points to obtain the total comprehension score (outof a total possible of 25 points) Actual scores ranged from 3 to 24points with a mean of 13 points (SD = 53) Following Schwambornet al (2010) scoring of the drawing test was carried out by count-ing the total number of correct main ideas in each learnerrsquos answeracross the three drawing items The main ideas were drawn out fromboth expert visualizations and a checklist specifying important re-lational features Students could earn a maximum of 19 points onthe drawing test Two student assistants (teacher trainees in biology)scored the quality for each of the three drawings for each studentwith an acceptable inter-rater agreement of GoodmanndashKruskalgamma of 090 Actual scores ranged from 0 to 185 points with amean of 77 points (SD = 47) Total scores of both the comprehen-sion and the drawing test were transferred into z-standardized scoresto make them comparable across studies

The drawing accuracy score (concerning drawing during learn-ing in the drawing group) was computed by using a coding schemeadapted from Schwamborn et al (2010) which was based on expertdrawings and a checklist specifying important relational featuresof the drawings Students could earn a maximum drawing-accuracyscore of 22 points Again the two student assistants scored each ofthe seven learner-generated drawings for each student with an ac-ceptable interrater agreement of GoodmanndashKruskal gamma of 92Both coding schemes were constructed by the first author and abiology teacher Actual scores ranged from 4 to 21 points with amean of 133 points (SD = 50) The total drawing accuracy score wasagain transferred into a z-standardized score

In addition the spatial ability test was scored by tallying thenumber correct out of 10 and the motivation questionnaire wasscored by tallying the nine ratings on both subscales to a total scoreof motivation Finally for comparing performance across the dif-ferent tests the proportion correct on each test was computed bydividing the studentrsquos obtained score by the total possible score

242 Are the groups equivalent on basic characteristicsBefore looking at treatment effects on the dependent variables

we analyzed whether the two groups differed on several control vari-ables A chi-square analysis indicated that there were no significantdifferences regarding gender (p = 562) Separate univariate analy-ses of variance (ANOVAs) revealed that the groups did not differsignificantly on age F lt 1 on spatial ability F lt 1 or on motiva-tion F(1 46) = 360 p = 064 However groups differed significantlyon prior knowledge F(1 46) = 3890 p lt 001 partial eta2 = 46 inthat students in the drawing group scored significantly lower onthe comprehension pretest (M = 10 SD = 15) than students in thecontrol group (M = 34 SD = 12) Thus we included studentsrsquo priorknowledge in the following analyses

243 Is there support for the generative drawing effectMean proportion correct and SDs on the comprehension and

drawing posttests for both groups are presented in Table 2 Repeat-ed measures univariate analyses of variance (ANOVA) with thecomprehension pre- and post-test scores as the within-subject factorsand group (drawing versus control) as the between-subject factorshowed a main effect over time indicating that overall partici-pants reached significant knowledge gains between thecomprehension pretest and the comprehension posttest F(146) = 9897 p lt 001 partial eta2 = 68 An interaction additionallyshowed that these knowledge gains were significantly higher forthe drawing group than for the control group F(1 46) = 4617p lt 001 partial eta2 = 50

For the drawing test a repeated measures ANOVA was not pos-sible since these items were only used in the posttest In this casea univariate analysis of covariance (ANCOVA) predicting the drawingtest score with group (drawing versus control) as the factorial in-dependent variable and prior knowledge as a covariate showed thatthe drawing group scored significantly better than the control groupon the drawing posttest F(1 45) = 1349 p = 001 partial eta2 = 231

Cohenrsquos d favoring the drawing group over the control group was085 for the comprehension posttest and 115 for the drawingposttest all of which are considered large effects Thus there is strongsupport for the generative drawing effect as predicted

Additionally results revealed that the drawing group needed sig-nificantly more learning time (M = 2108 min SD = 424) than thecontrol group (M = 1738 min SD = 333) F(1 46) = 1134 p = 002partial eta2 = 20 Thus to test whether learning time mediates thepositive effect of drawing on text comprehension additional me-diation analyses (Baron amp Kenny 1986) were calculated by includinglearning time as an additional predictor in the aforementioned linearmodel A mediation effect would be detected if in this case effectsof drawing on text comprehension would significantly decreaseResults of the mediation analyses showed that the effect of drawingon both comprehension test scores and drawing test scores was notfully mediated by learning time That is including learning time stillrevealed the interaction between group (drawing versus control)and time (pre- versus post) in that the drawing group had signifi-cantly higher knowledge gains than the control group on thecomprehension test items (p lt 001) Furthermore the drawing groupalso still outperformed the control group on the drawing posttestafter controlling for learning time (p = 009)

Furthermore results revealed that students in the drawing grouprated their invested mental effort during learning significantly higher(M = 504 SD = 112) than students in the control group (M = 396SD = 165) F(1 46) = 705 p = 011 partial eta2 = 13 There was nodifference between the two groups on the perceived difficulty item(drawing group M = 408 SD = 150 control group M = 425SD = 122 F lt 1) Thus consistent with predictions concerning thegenerative drawing effect there is partial support for the idea that

1 The exclusion of studentsrsquo prior knowledge in the reported analyses does notchange the reported pattern of results

Table 2Mean proportion correct on the comprehension test and drawing test for two groupsndash Experiment 1

Group Type of test

n Comprehension test Drawing test

M SD M SD

Drawing 24 61 20 52 27Control 24 44 20 28 11

Note Asterisk () indicates significant difference from control group at p lt 05

280 A Schmeck et alContemporary Educational Psychology 39 (2014) 275ndash286

drawing causes students to engage in more generative processingduring learning

Taken together the results suggest that the drawing strategy en-courages students to engage in generative processing during learningas is indicated by their higher learning outcomes Thus the dataprovide further evidence for the generative drawing effect consis-tent with the results of Schwamborn et al (2010) Additionallyresults indicate that students in the drawing condition seem to investmore mental effort than students in the control group without per-ceiving higher levels of difficulty

244 Is there support for the prognostic drawing effectMean proportion correct on drawing accuracy during learning

was 59 (SD = 23) A correlation analysis revealed that the drawing-accuracy score of learner-generated drawings correlated significantlywith the comprehension posttest score r = 620 p lt 001 and withthe drawing posttest score r = 623 p lt 001 Additional correla-tion analyses revealed that the drawing-accuracy score of learner-generated drawings correlated significantly negatively with theperceived difficulty score r = minus489 p = 015 There were no signif-icant correlations between the drawing accuracy score and eitherthe invested mental effort score r = minus134 p = 533 the prior knowl-edge test score r = minus004 p = 984 the spatial ability test score r = 072p = 739 or the motivation test score r = 086 p = 690 Thus as pre-dicted the data provide further evidence for the prognostic drawingeffect consistent with the results of Schwamborn et al (2010)

In sum the results of Experiment 1 are consistent with the pre-diction that students learn better from a science text when they areasked to draw illustrations representing the main ideas of the textand that the quality of the generated drawings during learning cor-relates positively with studentsrsquo text comprehension (egSchwamborn et al 2010 van Meter 2001 van Meter et al 2006)

However it might be argued that the reported results are dueto the way we supported the strategy use In other words the re-ported positive effect of the learner-generated drawing strategy mightnot be caused by studentsrsquo engagement in generative learning ac-tivities during reading (de Jong 2005 Mayer 2004 2009 Wittrock1990) but rather by the additional pictorial information given in thedrawing prompt Additionally looking at studentsrsquo learning out-comes our results indeed show positive effects of drawing howevermean scores of learning outcomes for the drawing group aremedium-sized Thus it might be argued that the way we sup-ported the strategy use was not fully sufficient In other words thereported positive effect of the learner-generated drawing strategyie the generative drawing effect might be increased by giving stu-dents instructional support in addition to the drawing prompt (vanMeter 2001 van Meter et al 2006 van Meter amp Garner 2005) Toaddress these issues we added two experimental conditions byimplementing author-generated pictures in the design in Experi-ment 2

3 Experiment 2

One possible issue with Experiment 1 is the type of control groupused In Experiment 1 following Schwamborn et al (2010) we useda reading only control group in which the control group learnedwith verbal information only In the drawing group however stu-dents not only learned with verbal information but also with pictorialinformation given by the drawing prompt Based on theories of mul-timedia learning the use of different forms of representations suchas texts and pictures can promote learning in that ldquopeople learnbetter from words and pictures than from words alonerdquo (ie mul-timedia principle Mayer 2009 p 223) because in this caseboth a (verbal) propositional representation as well as a (pictori-al) mental model are built up and are optimally integrated into oneschema that can be stored in long-term memory (Schnotz 2005)

This assumption is also in line with the dual-coding approach statedby Paivio (1986) In this regard it might be argued that the re-ported drawing effect is actually a multimedia effect that is basedon the presentation of text and picture rather than a generativedrawing effect that is based on studentsrsquo active engagement indrawing activities during reading In other words instead of askingpeople to draw pictures representing the main ideas of the textgiving them text and author-generated pictures representing themain ideas of the text might be as good or even better Thus weincluded a condition in Experiment 2 in which we added author-generated pictures to the text

An additional issue with Experiment 1 is whether the reportedgenerative drawing effect can be enhanced by using various formsof supporting the strategy First there is evidence that using adrawing prompt during learning seems to be effective in support-ing the learner-generated drawing strategy by minimizing thecreation of extraneous processing (cf Schwamborn et al 2010 seealso Exp 1) Second research has shown that instructing studentsto compare their own drawing with an author-generated picturemight be also effective in supporting the learner-generated drawingstrategy as self-monitoring processes are enhanced (cf van Meter2001) Up to now however there is no empirical evidence whetherthe combination of both ways to support the drawing strategy hasan additive effect on learning outcomes Thus we included a furthercondition in Experiment 2 in which we combined both forms ofstrategy support

The main purpose of Experiment 2 was to test the generativedrawing and prognostic drawing effects of learner-generated drawingas in Experiment 1 but this time also compared with another controlgroup (ie author-generated pictures) Additionally we were in-terested in testing whether the benefits of the learner-generateddrawing strategy can be increased when we instructionally supportstudents not only with a drawing prompt but also with an author-generated picture after the drawing process In this new treatmentwe instructed students to draw a picture of the text content andthen to compare their own drawing with an expert picture

31 Participants and design

The participants were 168 German eighth graders from highertrack secondary schools The mean age was 138 years (SD = 06)and there were 112 girls and 56 boys The study was based on a2 times 2-between-subjects design with learner-generated drawing (yesno) and author-generated picture (yesno) as factors Forty studentsserved in the drawing group 44 students served in the author-generated picture group 41 students served in the drawing + author-generated picture group and 43 students served in the control group

32 Materials

The materials were identical to those used in Experiment 1 exceptthat we used a shortened version of the comprehension pretest thatconsisted of 19 rather than 25 items (Cronbachrsquos alpha = 70) andslightly extended versions of both the comprehension posttest (28items Cronbachrsquos alpha = 84) and the drawing test (four items witha maximum score of 21 points Cronbachrsquos alpha = 78) The pretestwas shortened because the first experiment showed that the re-spective items were either much too easy or much too difficult andthus unsuitable to differentiate between successful and unsuccess-ful learners thus we deleted these items in the second experimentFurthermore we decided to add some items to the comprehen-sion posttest in the second experiment because during data analysisof the first experiment and after receiving some feedback fromexperts in the domain of biology we recognized that a few itemsassessing transfer ability could be added These transfer itemshowever would have been unsuitable to be included in the pretest

281A Schmeck et alContemporary Educational Psychology 39 (2014) 275ndash286

because they are too difficult to answer without prior training inthe topic Additionally author-generated pictures were used in thenew conditions The author-generated pictures were static func-tional pictures representing the main ideas of each paragraph andconsisted of pictorial elements identical to those provided in thedrawing prompt (as shown in Fig 2) These pictures were con-structed by the first author in cooperation with a biology teacher

The drawing version of the booklet was identical to that usedin Experiment 1 (as shown in Fig 1) The control version of the learn-ing booklet was identical to that used in Experiment 1 The author-generated picture version of the booklet consisted of seven pairsof facing pages with a text paragraph on the left page and a corre-sponding author-generated picture (such as in Fig 2) on the rightpage The drawing + author-generated picture version of the bookletcontained the material from the drawing version consisting of sevenpairs of facing pages with a text paragraph on the left page and atwo-part drawing prompt on the right page In addition attachedto each page there was an additional page that students could foldout after having generated their drawing When unfolding this ad-ditional page a picture of that text paragraph right aside the drawingprompt was provided and there was an additional instruction to

compare the learner-generated drawing with the author-generatedpicture Author-generated pictures were the same as in the author-generated picture version of the booklet

33 Procedure

The procedure was identical to that used in Experiment 1 exceptthat there were two additional groups learning with author-generated pictures Students in the author-generated picturecondition were instructed to read the text and additionally to lookat pictures representing the main ideas of each text paragraph Stu-dents in the drawing + author-generated picture version of thebooklet were instructed to read the text to draw pictures for eachtext paragraph using the drawing prompt representing the mainideas of each text paragraph and finally to compare their picturewith an author-generated picture representing main ideas of eachparagraph correctly Students in all groups learned at their own pacewhereby individual learning time was measured by the instruc-tors in the classrooms Again to ensure that studentsrsquo in both drawinggroups did not feel rushed when students in the non-drawing

Fig 2 Author-generated pictures for the seven paragraphs in the author-generated picture versions of the learning booklet Note Pictures are scaled-down from the orig-inal format

282 A Schmeck et alContemporary Educational Psychology 39 (2014) 275ndash286

group completed the task early groups were tested in separateclassrooms

34 Results

341 ScoringAll tests instruments were scored with the same procedures used

in Experiment 1 Again two student assistants (teacher trainees inbiology) scored each of the drawing test items and each of the sevenlearner-generated drawings for each student with acceptableinterrater agreements (drawing test GoodmanndashKruskal gamma of90 drawing-accuracy GoodmanndashKruskal gamma of 94) Actualscores ranged from 1 to 28 points (M = 153 points SD = 58) for thecomprehension test from zero to 21 points (M = 109 points SD = 53)for the drawing test and from 275 to 215 points (M = 141 pointsSD = 48) for drawing accuracy Again total scores of comprehen-sion drawing and accuracy were transferred into z-standardizedscores

342 Are the groups equivalent on basic characteristicsBefore looking at treatment effects on the dependent variables

we analyzed whether the four experimental groups differed onseveral control variables A chi-square analysis indicated that therewere no significant differences regarding gender (p = 097) Sepa-rate univariate analyses of variance (ANOVAs) revealed that thegroups also did not differ significantly on age F lt 1 on spatial abilityF(3 164) = 120 p = 312 or on motivation F(3 164) = 122 p = 305However groups differed significantly on prior knowledge F(3164) = 104 p = 010 partial eta2 = 07 in that students in the controlgroup scored significantly higher on the comprehension pretest(M = 25 SD = 17) than students in both (p lt 05) the author-generated picture group (M = 17 SD = 16) and the drawing + author-generated picture group (M = 15 SD = 12) the drawing group(M = 20 SD = 14) did not differ significantly from the other groupsThus we included studentsrsquo prior knowledge as a covariate in thefollowing analyses

343 Is there support for the generative drawing effectA major goal in this experiment was to determine whether asking

students to generate drawings to represent science text is a moreeffective learning strategy than asking students to learn with textalone or with text and author-generated pictures In other wordswe wanted to determine whether we could replicate and extendthe learner-generated drawing effect Additionally we were inter-ested in whether giving students an author-generated picture afterdrawing can increase the benefits of the learning strategy Mean pro-portion correct and standard deviations on the comprehension anddrawing tests for the four groups are presented in Table 3

The left portion of Table 3 summarizes the mean proportioncorrect on the comprehension test A two-factorial analysis ofcovariance (ANCOVA) predicting learning outcomes (comprehen-sion posttest score) with learner-generated drawing (yesno) and

author-generated picture (yesno) as the factorial independent vari-ables and prior knowledge as a covariate showed a significantpositive main effect of learner-generated drawing F(1 163) = 398p = 048 partial eta2 = 02 a significant interaction effect F(1163) = 626 p = 013 partial eta2 = 04 but no main effect of author-generated pictures F lt 1 In addition multiple pairwise comparisons(with p lt 05) showed that the drawing group performed signifi-cantly better than each of the three other groups which did not differsignificantly from each other Cohenrsquos d favoring the drawing groupover the author-generated picture group was 49 over the learner-generated + author-generated picture group was 57 and over thecontrol group was 52

The right portion of Table 3 summarizes the mean proportioncorrect on the drawing posttest Again a two-factorial analysis ofcovariance (ANCOVA) predicting learning outcome (drawing testscore) with learner-generated drawing (yesno) and author-generatedpicture (yesno) as the factorial independent variables and priorknowledge as a covariate showed a significant positive main effectof learner-generated drawing F(1 163) = 6260 p lt 001 partialeta2 = 28 a significant positive main effect of author-generated pic-tures F(1 163) = 1104 p = 001 partial eta2 = 06 and a significantinteraction effect F(1 163) = 1658 p lt 001 partial eta2 = 09 In ad-dition multiple pairwise comparisons (with p lt 05) showed thatboth the drawing group and the drawing + author-generated picturegroup performed significantly better than the author-generatedpicture group (d = 68 d = 59) and the control group (d = 187d = 188) In turn the author-generated picture group performed sig-nificantly better than the control group (d = 95) The drawing groupand the drawing + author-generated picture group did not differ sig-nificantly from each other (d = 15)2 Overall these results areconsistent with Experiment 1 and provide additional support forthe generative drawing effect

In accordance with Experiment 1 we were interested in whetherdifferences in learning time among the experimental groups mediatethe positive effect of drawing on text comprehension First an ANOVApredicting learning time with learner-generated drawing (yesno)and author-generated picture (yesno) as the factorial indepen-dent variables showed a significant main effect of learner-generateddrawing F(1 164) = 39226 p lt 001 partial eta2 = 71 a significantmain effect of author-generated picture F(1 164) = 1685 p lt 001partial eta2 = 09 and a significant interaction effect F(1 164) = 490p = 028 partial eta2 = 03 Linear contrasts (with p lt 05) revealedthat the drawing group (M = 1938 min SD = 380) and thedrawing + author-generated picture group (M = 2340 min SD = 551)needed significantly more learning time than the author-generatedpicture group (M = 960 min SD = 397) and the control group(M = 834 min SD = 251) Thus to test whether learning time me-diates the positive effect of drawing on text comprehensionadditional mediation analyses (Baron amp Kenny 1986) were calcu-lated by including learning time as an additional predictor in theaforementioned linear model Results of the mediation analysesshowed that the effects of drawing on the comprehension posttestand the drawing posttest scores (see multiple pairwise compari-sons) are mediated by learning time to some extent That is includinglearning time in the linear model for predicting comprehension testscores still revealed a positive effect of the drawing group com-pared with the drawing + author-generated group on thecomprehension test (p = 012) However including learning time inthe linear model for predicting comprehension posttest scoresreduced the positive effect of the drawing group compared with theauthor-generated picture group (from p = 034 to p = 281) as wellas compared with the control group (from p = 002 to p = 087) being

2 The exclusion of studentsrsquo prior knowledge in the reported analyses does notchange the overall pattern of results

Table 3Mean proportion correct on the comprehension test and drawing test for the fourgroups ndash Experiment 2

Group Type of test

n Comprehension test Drawing test

M SD M SD

Learner-generated drawing 40 63 22 66 22Author-generated picture 44 53 19 50 25Learner-generated drawing +

author-generated picture41 51 20 63 19

Control 43 52 20 30 16

Note Asterisk () indicates significant difference from control group at p lt 05

283A Schmeck et alContemporary Educational Psychology 39 (2014) 275ndash286

no longer statistically significant Regarding the drawing posttestscore including learning time does not change the reported patternof results except that the positive effect of the drawing + author-generated picture group compared with the author-generated picturegroup is no longer statistically significant (from p = 004 to p = 223)

There were neither main effects of learner-generated drawingand author-generated pictures on the mental effort item (drawinggroup M = 455 SD = 025 author-generated picture group M = 459SD = 024 drawing + author-generated picture group M = 444SD = 026 control group M = 481 SD = 024 F lt 1) nor on theperceived difficulty item (drawing group M = 363 SD = 023 author-generated picture group M = 371 SD = 022 drawing + author-generated picture group M = 395 SD = 023 control group M = 393SD = 022 F lt 1)

Taken together the drawing strategy apparently fosters stu-dents to engage in generative activities indicated by their higherlearning outcomes Thus the data provide further evidence for thegenerative drawing effect predicted by Schwamborn et al (2010)In Experiment 2 benefits of the drawing activity however are me-diated by learning time and do not involve higher mental effortAdditionally there was no increased benefit when additional drawingsupport was available in the form of author-generated pictures

344 Is there support for the prognostic drawing effectA second major goal of this study was to determine whether the

prognostic drawing effect could be extended to a new context Meanproportion correct on drawing-accuracy during learning was 60(SD = 04) for the drawing group and 68 (SD = 03) for thedrawing + author generated picture group This difference betweenthe two drawing groups is not significant F(1 79) = 252 p = 116This lack of group differences allowed us to pool the data of bothdrawing groups for subsequent correlation analyses Correlation anal-yses based on the combined data from the two drawing groupsrevealed that the drawing-accuracy score of learner-generated draw-ings correlates significantly with the comprehension posttest scorer = 470 p lt 001 as well as with the drawing posttest score r = 615p lt 001 Additional correlation analyses revealed that the drawing-accuracy score of learner-generated drawings did not correlatesignificantly with the prior knowledge test score r = 095 p = 400the spatial ability test score r = 127 p = 257 the motivation testscore r = 033 p = 769 or the mental effort test score r = 042p = 712 The correlation between the drawing-accuracy score andthe perceived difficulty score was only slightly statistical signifi-cance r = minus218 p = 053 Thus the data provide further evidencefor the prognostic drawing effect consistent with the results ofSchwamborn et al (2010)

In sum results of Experiment 2 are partly consistent with theresults of Experiment 1 in that students learn better from a sciencetext when they are asked to draw illustrations representing the mainideas of the text and the quality of the generated drawings duringlearning correlates positively with studentsrsquo text comprehension (egSchwamborn et al 2010 van Meter 2001 van Meter et al 2006)

4 Discussion

41 Empirical contributions

The present set of experiments makes three empirical contri-butions to the field First this study shows strong and consistentevidence that students who are asked to generate drawings (withsufficient support) during reading a scientific text that describes acausal sequence perform better than students who read withoutdrawing both on a comprehension test (d = 085 in Experiment 1and d = 052 in Experiment 2) and on a drawing test (d = 115 in Ex-periment 1 and d = 187 in Experiment 2) Thus the generativedrawing effect can be extended to a new domain and therefore

corresponds to Shavelson and Townersquos (2002 p 4) recommenda-tion to ldquoreplicate and generalize across studiesrdquo as one of the sixessential scientific principles of educational research

Second this study shows strong and consistent evidence that thequality of drawings that students generate during learning with ascientific text that describes a causal sequence is positively relatedto subsequent performance on tests of learning outcome includ-ing a comprehension test (r = 623 in Experiment 1 and r = 470 inExperiment 2) and a drawing test (r = 620 in Experiment 1 andr = 615 in Experiment 2) Thus the prognostic drawing effect canbe replicated and extended to a new domain consistent with stan-dards for scientific research in education prescribed by Shavelsonand Towne (2002)

Third this study shows that asking learners to draw picturesduring reading a scientific text (ie learner-generated drawing groupin Experiment 2) is more effective than simply providing draw-ings (ie author-generated picture group in Experiment 2) both ona comprehension test (d = 049) and a drawing test (d = 068) Sim-ilarly adding author-generated drawings (ie learner-generatedpictures + author-generated pictures group in Experiment 2) doesnot improve the learning outcomes of students who also draw pic-tures during learning (ie learner-generated pictures group inExperiment 2) either on a comprehension test (d = minus057) or adrawing test (d = minus015) In short the act of drawing during learn-ing (with sufficient support) improves learning beyond the simpleprovision of drawings

42 Theoretical contributions

The results are consistent with the idea that drawing during learn-ing serves as a generative activity (Mayer amp Wittrock 2006Schwamborn et al 2010 van Meter amp Garner 2005 Wittrock 1990)That is the act of drawing encourages learners to engage in gen-erative cognitive processing during learning such as organizing therelevant information into a coherent structure and integrating itwith relevant prior knowledge from long-term memory In thepresent study positive effects of drawing were indicated with a com-prehension and a drawing learning outcome test and therefore arein line with the theoretical assumption derived from the GTDC thatbenefits of drawing can be found if learning outcome tests are usedthat are sensitive to the underlying process of drawing (cf van Meteramp Garner 2005) Additionally in our study the drawing activity wassupported in a way that was intended to help learners carry out theunderlying cognitive processes of drawing (ie selecting organiz-ing and integrating) successfully In this regard results of the presentstudy might supplement the theoretical framework of learner-generated drawing by providing further evidence that benefits ofdrawing defined by van Meter and Garnerrsquos GTDC can diminish ifno instructional support is given to constrain and structure thedrawing activity However a fuller understanding of the underly-ing cognitive processes of drawing and how these processes canbe influenced via drawing support requires more direct measuresof cognitive processing during learning Additionally following theidea that metacognitive processes of monitoring and regulation areautomatically activated by drawing (van Meter amp Garner 2005) afuller understanding of the metacognitive effects of drawings is alsorequired

43 Practical contributions

The present study encourages instructional designers and in-structors to incorporate drawing activities into venues involvinglearning from text which we call the generative drawing effect Oneimportant feature of a successful drawing strategy that is presentin this study and in a previous study by Schwamborn et al (2010)is that the drawing activity was supported by providing a

284 A Schmeck et alContemporary Educational Psychology 39 (2014) 275ndash286

background scene and a legend showing how to represent eachelement to constrain and structure the drawing activity Thus animportant practical implication is that students may need supportin their drawing activity so they do not need to draw from scratch

The present study also suggests a potentially useful diagnostictool to gauge the depth of student learning namely the quality ofthe drawings created by students during learning which we referto as the prognostic drawing effect Incorporating a measure of thequality of a learnerrsquos drawing during learning can be a useful toolin developing remedial instruction to give learners individual supportIt may be important to use materials that explain a cause-and-effect process and give learners drawings of the elements they needto represent the process pictorially Asking learners to simply drawpictures of elements is unlikely to be helpful whereas asking themto generate drawings that show the relations among the elementsin a schematic form is more likely to be helpful

44 Limitations and future directions

Some limitations and future directions of our study should beaddressed As noted in the theoretical contributions subsection wedid not have direct measures of cognitive processing during learn-ing so it is not possible to pinpoint how the drawing activity affectedspecific cognitive processes such as attending to relevant informa-tion organizing it and integrating it with prior knowledge We alsodid not assess metacognitive processing during learning thus it isnot possible to pinpoint how the drawing activity affected specificmetacognitive processes such as monitoring and regulation

Furthermore results of the cognitive load rating scales (in-vested mental effort and perceived task difficulty) are inconsistentWhereas in Experiment 1 an effect on mental effort but not on per-ceived task difficulty showed up (ie students in the drawing grouprated their invested mental effort during learning significantlyhigher) no effects on mental effort and task difficulty were foundin Experiment 2 Additionally in both experiments only a nega-tive correlation of perceived task difficulty with the quality of learner-generated pictures appears but no correlation of mental effort withthe quality Following de Jong (2010) those cognitive load ratingscales might have the disadvantages that they do not give a con-current measure of cognitive load and do not measure an essentialconcept in cognitive load theory namely cognitive overload (p 125)Future studies on learner-generated drawing might also use othercognitive load measures such as physiological measures as moredirect indicators of cognitive load

As noted in the practical contributions subsection we showedthe drawing effects by using a scientific text describing how a cause-and-effect system works that is the causal steps regarding aninfection with influenza and the immune response It might be pos-sible however that for other types of text producing drawings mightharm rather than promote text comprehension Thus to test whetherthe reported drawing effects can be extended future research hasto focus on other types of text such as descriptive texts as well ason other types of relations that can be conveyed with other typesof representations such as compare and contrast relations whichcan be shown in a matrix Additionally studentsrsquo learning out-comes were tested immediately after reading thus future work isneeded to investigate the longer-term effects of generative drawingon learning outcomes

Furthermore we only compared drawing with control groupsthat received no further learning strategy instructions However en-gaging in generative learning activities such as drawing requires aconsiderable amount of time Accordingly results showed that forExperiment 2 the positive effect of the drawing group on text com-prehension compared with the author-generated picture group andto the control group was mediated by learning time To rule out thatthe effects of drawing result only from additional time on task instead

of the generative activity future research should also compare thedrawing strategy with other time demanding generative learningstrategies such as summarization (cf Leopold amp Leutner 2012)

Another point that should be noted is that students in both ex-periments received some kind of multimedia materials in that evenwhen they had to draw and did not see presented pictures they wereat least provided with the basic (visual) elements for their draw-ings which they had to do on the given background which thusalso contained information In other words when students are pre-sented with important elements of the drawings which they canuse to draw themselves they will not have to put as much effortinto summarizing visually what they have just read compared withstudents who have to draw without any instructional help Futurestudies might also compare the drawing group with a summariza-tion group in which students receive a set of verbal key terms thatare similar to the drawing elements and are asked to make a textualsummary

Additionally future research is needed to validate the prognos-tic drawing effect So far we know that the quality of learner-generated pictures is related to studentsrsquo learning outcomes (iethe higher the learning outcome the higher the drawing accuracyand vice versa) and their perceived difficulty (ie the lower the per-ceived difficulty the higher the drawing accuracy and vice versa)and that it is not related to studentsrsquo prior knowledge motivationspatial ability or mental effort However less is known about whatthis might mean That is less is known regarding the causal direc-tion of this relation or the presence of a possible further moderatorvariable Do studentsrsquo efforts to produce accurate drawings lead tobetter comprehension and lower perceived difficulty Or do stu-dents who are more adept in drawing benefit more from the strategyand thus perceive the difficulty of the learning materials as beinglower Both arguments seem convincing

Finally more work is needed to determine the level of supportthat makes the drawing strategy most effective for various kinds oflearners As noted in the empirical contribution adding author-generated drawings (ie learner-generated pictures + author-generated pictures group in Experiment 2) does not improve thelearning outcomes of students who also draw pictures during learn-ing and were supported by a drawing prompt In other words thecombination of two ways of supporting the drawing strategy (iegiving a drawing prompt during reading plus an author-generatedpicture after reading) did not improve studentsrsquo learning out-comes compared with students in the drawing group as well ascompared with students in the control and author-generated pic-tures only groups This result is inconsistent with previous research(eg van Meter 2001 van Meter et al 2006) which found that com-paring own drawings to author-generated pictures normally helpslearning van Meter and colleagues (2001 2006) however provid-ed author-generated pictures plus prompting questions after thedrawing process That is students answered prompting questionsto guide the comparison process between their self-generateddrawing and the author-generated drawing In our study studentswere only instructed to generate a drawing to inspect an author-generated one and to check whether their own drawing incomparison with the author-generated one really represented themain ideas of the text paragraph correctly In other words we didnot guide the process of comparing self-generated drawings withauthor-generated ones As a potential consequence students per-formed the intended comparison process inadequately or even notall and thus did not benefit from it One reason for this inade-quate comparison process might be that students need guidancein doing the comparison process Another reason might be the factthat students do not seriously engage in generating drawings oncethey notice that there are author-generated drawings Thus futureresearch should also use additional guidance to test whether thecombination of different ways of supporting the drawing strategy

285A Schmeck et alContemporary Educational Psychology 39 (2014) 275ndash286

(ie giving a drawing prompt during reading plus an author-generated picture after reading) helps learning as well asobservational measures of the drawing process itself (ie think aloudprotocols) to shed more light on the cognitive processes underly-ing the drawing activities

Overall drawing during learning from text appears to be a po-tentially powerful strategy for improving studentsrsquo learning fromscientific text when certain boundaries and prerequisites are takeninto account

Acknowledgments

This article is based on a research project funded by the GermanResearch Foundation (DFG LE 6459-3 as part of FOR 511) We wouldlike to thank Angela Sandmann for her assistance in developing thelearning materials

References

Ainsworth S Prain V amp Tytler R (2011) Drawing to learn in science Science 3331096ndash1097

Alesandrini K L (1981) Pictorial-verbal and analytic-holistic learning strategies inscience learning Journal of Educational Psychology 73 358ndash368

Alesandrini K L (1984) Pictures and adult learning Instructional Science 13 63ndash77Baron R M amp Kenny D A (1986) The moderator-mediator variable distinction in

social psychological research Conceptual strategic and statistical considerationsJournal of Personality and Social Psychology 51 1173ndash1182

Bruumlnken R Plass J L amp Leutner D (2003) Direct measurement of cognitive loadin multimedia learning Educational Psychologist 38 53ndash61

Carney RN amp Levin JR (2002) Pictorial illustrations still improve studentsrsquo learningfrom text Educational Psychology Review 14 5ndash26

de Jong T (2005) The guided discovery principle in multimedia learning In R EMayer (Ed) The Cambridge handbook of multimedia learning (pp 215ndash228) NewYork Cambridge University Press

de Jong T (2010) Cognitive load theory educational research and instructionaldesign Some food for thought Instructional Science 38 105ndash134

Ekstrom R B French J W amp Harman H H (1976) Manual for kit of factor-referencedcognitive tests Princeton NJ Educational Testing Service

Greene T R (1989) Childrenrsquos understanding of class inclusion hierarchies Therelationship between external representation and task performance Journal ofExperimental Child Psychology 48 62ndash89

Hall V C Bailey J amp Tillman C (1997) Can student-generated illustrations be worthten thousand words Journal of Educational Psychology 89 677ndash681

Houmlffler T N (2010) Spatial ability Its influence on Learning with visualizations ndashA meta-analytic review Educational Psychology Review 22 245ndash269

Houmlffler T N Schmeck A amp Opfermann M (2013) Static and dynamic visualrepresentations Individual differences in processing In G Schraw M TMcCrudden amp D Robinson (Eds) Learning through visual displays (pp 133ndash163)Charlotte NC Information Age Publishing

Kalyuga S Chandler P amp Sweller J (1999) Managing split-attention and redundancyin multimedia instruction Applied Cognitive Psychology 13 351ndash371

Leopold C (2009) Lernstrategien und Textverstehen [Learning strategies and textcomprehension] Muumlnster Waxmann

Leopold C amp Leutner D (2012) Science text comprehension Drawing main ideaselection and summarizing as learning strategies Learning and Instruction 2216ndash26

Lesgold A M DeGood H amp Levin J R (1977) Pictures and young childrenrsquos proselearning A supplementary report Journal of Reading Behavior 9 353ndash360

Lesgold A M Levin J R Shimron J amp Guttman J (1975) Pictures andyoung childrenrsquos learning from oral prose Journal of Educational Psychology 67636ndash642

Leutner D Leopold C amp Sumfleth E (2009) Cognitive load and science textcomprehension Effects of drawing and mentally imagining text contentComputers in Human Behavior 25 284ndash289

Mayer R E (2004) Should there be a three-strikes rule against pure discoverylearning The case for guided methods of instruction The American Psychologist59 14ndash19

Mayer R E (2005) Cognitive theory of multimedia learning In R E Mayer (Ed)The Cambridge handbook of multimedia learning (pp 31ndash48) New York CambridgeUniversity Press

Mayer R E (2009) Multimedia learning (2nd ed) New York NY CambridgeUniversity Press

Mayer R E amp Wittrock M C (2006) Problem solving In P Alexander P Winne ampG Phye (Eds) Handbook of educational psychology (pp 287ndash303) Mahwah NJErlbaum

Paas F (1992) Training strategies for attaining transfer of problem-solving skill instatisticsmdashA cognitive-load approach Journal of Educational Psychology 84429ndash434

Paas F Tuovinen J Tabbers H K amp Van Gerven P W M (2003) Cognitive loadmeasurement as a means to advance cognitive load theory EducationalPsychologist 38 63ndash71

Paivio A (1986) Mental representation A dual coding approach New York OxfordUniversity Press

Pashler H Bain P Bottage B Graesser A Koedinger K McDaniel M et al (2007)Organizing instruction and study to improve student learning Washington DCNational Center for Educational Research

Rasco R W Tennyson R D amp Boutwell R C (1975) Imagery instructions anddrawings in learning prose Journal of Educational Psychology 67 188ndash192

Rheinberg F Vollmeyer R amp Burns B D (2001) FAM Ein fragebogen zurerfassung aktueller motivation in lern- und leistungssituationen [QCM Aquestionnaire to assess current motivation in learning situations] Diagnostica47 57ndash66

Schnotz W (2005) An integrated model of text and picture comprehension In RE Mayer (Ed) The Cambridge handbook of multimedia learning (pp 49ndash70) NewYork Cambridge University Press

Schwamborn A Mayer R E Thillmann H Leopold C amp Leutner D (2010) Drawingas a generative activity and drawing as a prognostic activity Journal of EducationalPsychology 102 872ndash879

Schwamborn A Thillmann H Opfermann M amp Leutner D (2011) Cognitive loadand instructionally supported learning with provided and learner-generatedvisualizations Computers in Human Behavior 27 89ndash93

Shavelson R J amp Towne L (Eds) (2002) Scientific research in education WashingtonDC National Academy Press

Sweller J Ayres P amp Kalyuga S (2011) Cognitive Load Theory New York SpringerTirre W C Manelis L amp Leicht K (1979) The effects of imaginal and verbal strategies

on prose comprehension by adults Journal of Reading Behavior 11 99ndash106van Meter P (2001) Drawing construction as a strategy for learning from text Journal

of Educational Psychology 69 129ndash140van Meter P Aleksic M Schwartz A amp Garner J (2006) Learner-generated drawing

as a strategy for learning from content area text Contemporary EducationalPsychology 31 142ndash166

van Meter P amp Garner J (2005) The promise and practice of learner-generateddrawings Literature review and synthesis Educational Psychology Review 12261ndash312

Van Gog T amp Paas F (2008) Instructional efficiency Revisiting the original constructin educational research Educational Psychologist 43 16ndash26

Vollmeyer R amp Rheinberg F (2000) Does motivation affect learning via persistenceLearning and Instruction 4 293ndash309

Weinstein C E amp Mayer R E (1986) The teaching of learning strategies In M CWittrock (Ed) Handbook of research on teaching (pp 315ndash327) New YorkMacmillan

Wittrock M C (1990) Generative processes of comprehension EducationalPsychologist 24 345ndash376

286 A Schmeck et alContemporary Educational Psychology 39 (2014) 275ndash286

  • Drawing pictures during learning from scientific text testing the generative drawing effect and the prognostic drawing effect
  • Introduction
  • Theoretical framework for the learner-generated drawing strategy
  • Empirical framework for the learner-generated drawing strategy
  • Effectiveness of learner-generated drawings
  • Quality of learner-generated drawings
  • Overview of the experiments
  • Experiment 1
  • Participants and design
  • Materials
  • Procedure
  • Results and discussion
  • Scoring
  • Are the groups equivalent on basic characteristics
  • Is there support for the generative drawing effect
  • Is there support for the prognostic drawing effect
  • Experiment 2
  • Participants and design
  • Materials
  • Procedure
  • Results
  • Scoring
  • Are the groups equivalent on basic characteristics
  • Is there support for the generative drawing effect
  • Is there support for the prognostic drawing effect
  • Discussion
  • Empirical contributions
  • Theoretical contributions
  • Practical contributions
  • Limitations and future directions
  • Acknowledgments
  • References

2 Experiment 1

21 Participants and design

Forty-eight German eighth graders in higher track secondaryschools participated in this study The mean age was 137 years(SD = 06) and there were 22 girls and 26 boys The study was basedon a between-subjects design with two levels of text learning(learner-generated drawings and control) as the single factor Twenty-

four students served in the control group and 24 served in thedrawing group

22 Materials

All materials were paper-pencil based The materials consistedof five adjunct questionnaires two learning booklets two cogni-tive load rating scales and two posttests The five adjunctquestionnaires were intended to determine whether the groups were

Fig 1 Screenshot of the drawing prompt for the first paragraph in the drawing versions of the learning booklet Note Translated from the German original

278 A Schmeck et alContemporary Educational Psychology 39 (2014) 275ndash286

equivalent on basic characteristics They included a participant ques-tionnaire a comprehension pretest a spatial ability test and amotivation questionnaire The participant questionnaire solicited in-formation concerning the studentsrsquo age and sex The comprehensionpretest consisted of 25 multiple-choice items and was intended toassess studentsrsquo prior-knowledge of information covered in the textStudentsrsquo spatial ability was measured with a 10 multiple-choicepaper-folding items taken from a battery of cognitive tests devel-oped by Ekstrom French and Harman (1976) The motivationquestionnaire assessed studentsrsquo current motivation for doing thelearning task after reading the instructions before the lesson It con-sisted of nine items from the challenge and interest subscales of theQuestionnaire on Current Motivation (QCM) developed by RheinbergVollmeyer and Burns (2001) Cognitive load by means of investedmental effort was measured using the 7-point subjective rating scaledeveloped by Paas (1992) which ranges from (1) very low mentaleffort to (7) very high mental effort Cognitive load by means of per-ceived task difficulty was measured using the 7-point subjective ratingscale developed by Kalyuga Chandler and Sweller (1999) whichranges from (1) very easy to (7) very difficult These subjective mea-sures have been criticized for assessing cognitive load with only singleitems (eg Bruumlnken Plass amp Leutner 2003) However several studiesshowed the effectiveness of the rating scale by showing that the vari-ation in learnersrsquo cognitive load ratings depended on variations intask complexity or instructional design (for overviews see PaasTuovinen Tabbers amp Van Gerven 2003 Van Gog amp Paas 2008) Inthis regard Sweller Ayres and Kalyuga (2011) conclude that ldquothesimple subjective rating scale regardless of the wording used (mentaleffort or difficulty) has perhaps surprisingly been shown to be themost sensitive measure available to differentiate the cognitive loadimposed by different instructional proceduresrdquo (p 74) For that reasonand due to the economic applicability we decided to use this kindof cognitive load measurement while acknowledging the limita-tions of a short self-report instrument

The two learning booklets each included a science text on thebiology of the influenza The text explained the causal steps re-garding an infection with influenza and regarding the immuneresponse which is an unfamiliar subject for eighth graders in highertrack secondary schools due to the German curriculum The text con-sisted of approximately 850 words (in German) and was divided intoseven paragraphs (as shown in Table 1)

The drawing version of the booklet contained seven pairs of facingpages with a text paragraph on the left page and a two-part drawingprompt on the right page The first part of the drawing prompt in-cluded a legend showing all the relevant elements (in total eightelements) for drawing a picture for that text paragraph (as shownin the top of Fig 1) The second part of the drawing prompt in-cluded a partly pre-drawn background for studentsrsquo drawing (asshown in the bottom of Fig 1) Overall students had to make sevendrawings ie one drawing to each paragraph

The control version of the learning booklet contained four pairsof facing pages with one of the seven text paragraph on each page

Students in both groups learned with exactly the same text mate-rial To make sure that students in the control group learnedwith the same amount of information as students in the drawinggroup all elements of the drawing prompt as well as the spatialrelations between these elements were also described in the sciencetext

The two posttests intended to assess the learning outcomes werea comprehension posttest and a drawing posttest The comprehen-sion posttest (Cronbachrsquos alpha = 083) consisted of 25 multiple-choice items (the same items as in the comprehension pretest) andwas intended to assess studentsrsquo comprehension of the factual andconceptual information covered in the text as well as their abilityto transfer what was presented to new situations An item exampleis ldquoT-helper cells do not only recognize viruses but also agents thatare extraneous to the body Which medication would you admin-ister to a patient who has received a new kidney (a) a medicinethat suppresses the immune response of the body (b) a medicinethat activates the immune response of the body (c) a medicine thatcontains antigens or (d) a medicine that contains blood of the kidneydonorrdquo [(a) is the correct answer] The drawing test (Cronbachrsquosalpha = 081) was intended to assess studentsrsquo comprehension of theconceptual information presented in the science text by means ofdrawing That is students had to reproduce the main ideas givenin the text by drawing It consisted of three drawing items in whichstudents were asked to draw sketches depicting key concepts of thetext and their spatial relations An item example for the drawingtest is ldquoHow does an influenza virus invade a cell and how is it re-producedrdquo The science text the drawing prompt and the learningoutcome tests were constructed by the first author in cooperationwith a biology teacher The materials were adapted fromSchwamborn et al (2010) however using another science domainand including measures of individual learning times and cognitiveload

23 Procedure

Participants were tested in the schoolsrsquo classrooms Within theirclasses they were randomly assigned to one of the two groupsGroups were tested in separate classrooms in order to insure thatstudents in the drawing group did not feel rushed when studentsin the control group completed the task early Each student wasseated at an individual desk First students were given the partic-ipant questionnaire and the comprehension pretest to complete attheir own rate Second students filled in the paper-folding test witha 3 min time limit Third students were given instructional book-lets corresponding to their assigned group After they had read theinstructions for reading the booklets studentsrsquo current motivationfor doing the learning task was assessed Next students started learn-ing with the text material corresponding to their treatment groupStudents were instructed to carefully read the text on the biologyof the influenza in order to comprehend the material Students inthe drawing condition were instructed to read the text and addi-tionally to draw pictures for each text paragraph using the drawingprompt representing the main ideas of each text paragraph Thatis students had to use the pictorial elements given in the legendsuch as the virus as templates for their own paper-pencil baseddrawing across the pre-dawn background Students in the controlgroup were instructed to read the text for comprehension but werenot instructed to engage in drawing Students in both groups learnedat their own pace whereby individual learning time was mea-sured by the instructors in the classrooms Fourth in order to ensurecomparable testing procedures after finishing learning with thewhole learning material students in both groups directly rated theamount of mental effort he or she had invested during learning andthe amount of difficulty he or she had perceived during learningFifth students received the comprehension posttest consisting of

Table 1Text from the second paragraph of the influenza lesson

How the influenza virus replicates

Once inside the influenza virus uses your somatic cell to produce new particlesof the influenza-virus The glycoproteins move toward the membrane of thesomatic cell and stick out into the outside of the cell The capsules of thevirus however are assembled inside the somatic cell Next these newassembled capsules of the virus leave your somatic cell By moving throughthe somatic cell membrane the capsules are enveloped with the membraneand its glycoproteins which then plays the role of the virus membraneThus several new influenza viruses are located outside your somatic cell

Note Translated from the German original

279A Schmeck et alContemporary Educational Psychology 39 (2014) 275ndash286

the comprehension posttest and the drawing posttest Students had20 minutes time for completion and did not have access to thescience text or their drawings Finally students were thanked anddebriefed As students learned at their own rate the whole proce-dure took about 70ndash90 minutes depending on the individual testingtimes

24 Results and discussion

241 ScoringThe dependent variables were studentsrsquo scores on the compre-

hension and drawing posttests studentsrsquo rating on the mental effortand the difficulty scales and the drawing accuracy score indicat-ing the quality of learner-generated drawings produced by studentsin the drawing group during the learning phase

The comprehension test score (pre- and posttest) for each studentwas computed by awarding 1 point for each correct answer and byadding up the points to obtain the total comprehension score (outof a total possible of 25 points) Actual scores ranged from 3 to 24points with a mean of 13 points (SD = 53) Following Schwambornet al (2010) scoring of the drawing test was carried out by count-ing the total number of correct main ideas in each learnerrsquos answeracross the three drawing items The main ideas were drawn out fromboth expert visualizations and a checklist specifying important re-lational features Students could earn a maximum of 19 points onthe drawing test Two student assistants (teacher trainees in biology)scored the quality for each of the three drawings for each studentwith an acceptable inter-rater agreement of GoodmanndashKruskalgamma of 090 Actual scores ranged from 0 to 185 points with amean of 77 points (SD = 47) Total scores of both the comprehen-sion and the drawing test were transferred into z-standardized scoresto make them comparable across studies

The drawing accuracy score (concerning drawing during learn-ing in the drawing group) was computed by using a coding schemeadapted from Schwamborn et al (2010) which was based on expertdrawings and a checklist specifying important relational featuresof the drawings Students could earn a maximum drawing-accuracyscore of 22 points Again the two student assistants scored each ofthe seven learner-generated drawings for each student with an ac-ceptable interrater agreement of GoodmanndashKruskal gamma of 92Both coding schemes were constructed by the first author and abiology teacher Actual scores ranged from 4 to 21 points with amean of 133 points (SD = 50) The total drawing accuracy score wasagain transferred into a z-standardized score

In addition the spatial ability test was scored by tallying thenumber correct out of 10 and the motivation questionnaire wasscored by tallying the nine ratings on both subscales to a total scoreof motivation Finally for comparing performance across the dif-ferent tests the proportion correct on each test was computed bydividing the studentrsquos obtained score by the total possible score

242 Are the groups equivalent on basic characteristicsBefore looking at treatment effects on the dependent variables

we analyzed whether the two groups differed on several control vari-ables A chi-square analysis indicated that there were no significantdifferences regarding gender (p = 562) Separate univariate analy-ses of variance (ANOVAs) revealed that the groups did not differsignificantly on age F lt 1 on spatial ability F lt 1 or on motiva-tion F(1 46) = 360 p = 064 However groups differed significantlyon prior knowledge F(1 46) = 3890 p lt 001 partial eta2 = 46 inthat students in the drawing group scored significantly lower onthe comprehension pretest (M = 10 SD = 15) than students in thecontrol group (M = 34 SD = 12) Thus we included studentsrsquo priorknowledge in the following analyses

243 Is there support for the generative drawing effectMean proportion correct and SDs on the comprehension and

drawing posttests for both groups are presented in Table 2 Repeat-ed measures univariate analyses of variance (ANOVA) with thecomprehension pre- and post-test scores as the within-subject factorsand group (drawing versus control) as the between-subject factorshowed a main effect over time indicating that overall partici-pants reached significant knowledge gains between thecomprehension pretest and the comprehension posttest F(146) = 9897 p lt 001 partial eta2 = 68 An interaction additionallyshowed that these knowledge gains were significantly higher forthe drawing group than for the control group F(1 46) = 4617p lt 001 partial eta2 = 50

For the drawing test a repeated measures ANOVA was not pos-sible since these items were only used in the posttest In this casea univariate analysis of covariance (ANCOVA) predicting the drawingtest score with group (drawing versus control) as the factorial in-dependent variable and prior knowledge as a covariate showed thatthe drawing group scored significantly better than the control groupon the drawing posttest F(1 45) = 1349 p = 001 partial eta2 = 231

Cohenrsquos d favoring the drawing group over the control group was085 for the comprehension posttest and 115 for the drawingposttest all of which are considered large effects Thus there is strongsupport for the generative drawing effect as predicted

Additionally results revealed that the drawing group needed sig-nificantly more learning time (M = 2108 min SD = 424) than thecontrol group (M = 1738 min SD = 333) F(1 46) = 1134 p = 002partial eta2 = 20 Thus to test whether learning time mediates thepositive effect of drawing on text comprehension additional me-diation analyses (Baron amp Kenny 1986) were calculated by includinglearning time as an additional predictor in the aforementioned linearmodel A mediation effect would be detected if in this case effectsof drawing on text comprehension would significantly decreaseResults of the mediation analyses showed that the effect of drawingon both comprehension test scores and drawing test scores was notfully mediated by learning time That is including learning time stillrevealed the interaction between group (drawing versus control)and time (pre- versus post) in that the drawing group had signifi-cantly higher knowledge gains than the control group on thecomprehension test items (p lt 001) Furthermore the drawing groupalso still outperformed the control group on the drawing posttestafter controlling for learning time (p = 009)

Furthermore results revealed that students in the drawing grouprated their invested mental effort during learning significantly higher(M = 504 SD = 112) than students in the control group (M = 396SD = 165) F(1 46) = 705 p = 011 partial eta2 = 13 There was nodifference between the two groups on the perceived difficulty item(drawing group M = 408 SD = 150 control group M = 425SD = 122 F lt 1) Thus consistent with predictions concerning thegenerative drawing effect there is partial support for the idea that

1 The exclusion of studentsrsquo prior knowledge in the reported analyses does notchange the reported pattern of results

Table 2Mean proportion correct on the comprehension test and drawing test for two groupsndash Experiment 1

Group Type of test

n Comprehension test Drawing test

M SD M SD

Drawing 24 61 20 52 27Control 24 44 20 28 11

Note Asterisk () indicates significant difference from control group at p lt 05

280 A Schmeck et alContemporary Educational Psychology 39 (2014) 275ndash286

drawing causes students to engage in more generative processingduring learning

Taken together the results suggest that the drawing strategy en-courages students to engage in generative processing during learningas is indicated by their higher learning outcomes Thus the dataprovide further evidence for the generative drawing effect consis-tent with the results of Schwamborn et al (2010) Additionallyresults indicate that students in the drawing condition seem to investmore mental effort than students in the control group without per-ceiving higher levels of difficulty

244 Is there support for the prognostic drawing effectMean proportion correct on drawing accuracy during learning

was 59 (SD = 23) A correlation analysis revealed that the drawing-accuracy score of learner-generated drawings correlated significantlywith the comprehension posttest score r = 620 p lt 001 and withthe drawing posttest score r = 623 p lt 001 Additional correla-tion analyses revealed that the drawing-accuracy score of learner-generated drawings correlated significantly negatively with theperceived difficulty score r = minus489 p = 015 There were no signif-icant correlations between the drawing accuracy score and eitherthe invested mental effort score r = minus134 p = 533 the prior knowl-edge test score r = minus004 p = 984 the spatial ability test score r = 072p = 739 or the motivation test score r = 086 p = 690 Thus as pre-dicted the data provide further evidence for the prognostic drawingeffect consistent with the results of Schwamborn et al (2010)

In sum the results of Experiment 1 are consistent with the pre-diction that students learn better from a science text when they areasked to draw illustrations representing the main ideas of the textand that the quality of the generated drawings during learning cor-relates positively with studentsrsquo text comprehension (egSchwamborn et al 2010 van Meter 2001 van Meter et al 2006)

However it might be argued that the reported results are dueto the way we supported the strategy use In other words the re-ported positive effect of the learner-generated drawing strategy mightnot be caused by studentsrsquo engagement in generative learning ac-tivities during reading (de Jong 2005 Mayer 2004 2009 Wittrock1990) but rather by the additional pictorial information given in thedrawing prompt Additionally looking at studentsrsquo learning out-comes our results indeed show positive effects of drawing howevermean scores of learning outcomes for the drawing group aremedium-sized Thus it might be argued that the way we sup-ported the strategy use was not fully sufficient In other words thereported positive effect of the learner-generated drawing strategyie the generative drawing effect might be increased by giving stu-dents instructional support in addition to the drawing prompt (vanMeter 2001 van Meter et al 2006 van Meter amp Garner 2005) Toaddress these issues we added two experimental conditions byimplementing author-generated pictures in the design in Experi-ment 2

3 Experiment 2

One possible issue with Experiment 1 is the type of control groupused In Experiment 1 following Schwamborn et al (2010) we useda reading only control group in which the control group learnedwith verbal information only In the drawing group however stu-dents not only learned with verbal information but also with pictorialinformation given by the drawing prompt Based on theories of mul-timedia learning the use of different forms of representations suchas texts and pictures can promote learning in that ldquopeople learnbetter from words and pictures than from words alonerdquo (ie mul-timedia principle Mayer 2009 p 223) because in this caseboth a (verbal) propositional representation as well as a (pictori-al) mental model are built up and are optimally integrated into oneschema that can be stored in long-term memory (Schnotz 2005)

This assumption is also in line with the dual-coding approach statedby Paivio (1986) In this regard it might be argued that the re-ported drawing effect is actually a multimedia effect that is basedon the presentation of text and picture rather than a generativedrawing effect that is based on studentsrsquo active engagement indrawing activities during reading In other words instead of askingpeople to draw pictures representing the main ideas of the textgiving them text and author-generated pictures representing themain ideas of the text might be as good or even better Thus weincluded a condition in Experiment 2 in which we added author-generated pictures to the text

An additional issue with Experiment 1 is whether the reportedgenerative drawing effect can be enhanced by using various formsof supporting the strategy First there is evidence that using adrawing prompt during learning seems to be effective in support-ing the learner-generated drawing strategy by minimizing thecreation of extraneous processing (cf Schwamborn et al 2010 seealso Exp 1) Second research has shown that instructing studentsto compare their own drawing with an author-generated picturemight be also effective in supporting the learner-generated drawingstrategy as self-monitoring processes are enhanced (cf van Meter2001) Up to now however there is no empirical evidence whetherthe combination of both ways to support the drawing strategy hasan additive effect on learning outcomes Thus we included a furthercondition in Experiment 2 in which we combined both forms ofstrategy support

The main purpose of Experiment 2 was to test the generativedrawing and prognostic drawing effects of learner-generated drawingas in Experiment 1 but this time also compared with another controlgroup (ie author-generated pictures) Additionally we were in-terested in testing whether the benefits of the learner-generateddrawing strategy can be increased when we instructionally supportstudents not only with a drawing prompt but also with an author-generated picture after the drawing process In this new treatmentwe instructed students to draw a picture of the text content andthen to compare their own drawing with an expert picture

31 Participants and design

The participants were 168 German eighth graders from highertrack secondary schools The mean age was 138 years (SD = 06)and there were 112 girls and 56 boys The study was based on a2 times 2-between-subjects design with learner-generated drawing (yesno) and author-generated picture (yesno) as factors Forty studentsserved in the drawing group 44 students served in the author-generated picture group 41 students served in the drawing + author-generated picture group and 43 students served in the control group

32 Materials

The materials were identical to those used in Experiment 1 exceptthat we used a shortened version of the comprehension pretest thatconsisted of 19 rather than 25 items (Cronbachrsquos alpha = 70) andslightly extended versions of both the comprehension posttest (28items Cronbachrsquos alpha = 84) and the drawing test (four items witha maximum score of 21 points Cronbachrsquos alpha = 78) The pretestwas shortened because the first experiment showed that the re-spective items were either much too easy or much too difficult andthus unsuitable to differentiate between successful and unsuccess-ful learners thus we deleted these items in the second experimentFurthermore we decided to add some items to the comprehen-sion posttest in the second experiment because during data analysisof the first experiment and after receiving some feedback fromexperts in the domain of biology we recognized that a few itemsassessing transfer ability could be added These transfer itemshowever would have been unsuitable to be included in the pretest

281A Schmeck et alContemporary Educational Psychology 39 (2014) 275ndash286

because they are too difficult to answer without prior training inthe topic Additionally author-generated pictures were used in thenew conditions The author-generated pictures were static func-tional pictures representing the main ideas of each paragraph andconsisted of pictorial elements identical to those provided in thedrawing prompt (as shown in Fig 2) These pictures were con-structed by the first author in cooperation with a biology teacher

The drawing version of the booklet was identical to that usedin Experiment 1 (as shown in Fig 1) The control version of the learn-ing booklet was identical to that used in Experiment 1 The author-generated picture version of the booklet consisted of seven pairsof facing pages with a text paragraph on the left page and a corre-sponding author-generated picture (such as in Fig 2) on the rightpage The drawing + author-generated picture version of the bookletcontained the material from the drawing version consisting of sevenpairs of facing pages with a text paragraph on the left page and atwo-part drawing prompt on the right page In addition attachedto each page there was an additional page that students could foldout after having generated their drawing When unfolding this ad-ditional page a picture of that text paragraph right aside the drawingprompt was provided and there was an additional instruction to

compare the learner-generated drawing with the author-generatedpicture Author-generated pictures were the same as in the author-generated picture version of the booklet

33 Procedure

The procedure was identical to that used in Experiment 1 exceptthat there were two additional groups learning with author-generated pictures Students in the author-generated picturecondition were instructed to read the text and additionally to lookat pictures representing the main ideas of each text paragraph Stu-dents in the drawing + author-generated picture version of thebooklet were instructed to read the text to draw pictures for eachtext paragraph using the drawing prompt representing the mainideas of each text paragraph and finally to compare their picturewith an author-generated picture representing main ideas of eachparagraph correctly Students in all groups learned at their own pacewhereby individual learning time was measured by the instruc-tors in the classrooms Again to ensure that studentsrsquo in both drawinggroups did not feel rushed when students in the non-drawing

Fig 2 Author-generated pictures for the seven paragraphs in the author-generated picture versions of the learning booklet Note Pictures are scaled-down from the orig-inal format

282 A Schmeck et alContemporary Educational Psychology 39 (2014) 275ndash286

group completed the task early groups were tested in separateclassrooms

34 Results

341 ScoringAll tests instruments were scored with the same procedures used

in Experiment 1 Again two student assistants (teacher trainees inbiology) scored each of the drawing test items and each of the sevenlearner-generated drawings for each student with acceptableinterrater agreements (drawing test GoodmanndashKruskal gamma of90 drawing-accuracy GoodmanndashKruskal gamma of 94) Actualscores ranged from 1 to 28 points (M = 153 points SD = 58) for thecomprehension test from zero to 21 points (M = 109 points SD = 53)for the drawing test and from 275 to 215 points (M = 141 pointsSD = 48) for drawing accuracy Again total scores of comprehen-sion drawing and accuracy were transferred into z-standardizedscores

342 Are the groups equivalent on basic characteristicsBefore looking at treatment effects on the dependent variables

we analyzed whether the four experimental groups differed onseveral control variables A chi-square analysis indicated that therewere no significant differences regarding gender (p = 097) Sepa-rate univariate analyses of variance (ANOVAs) revealed that thegroups also did not differ significantly on age F lt 1 on spatial abilityF(3 164) = 120 p = 312 or on motivation F(3 164) = 122 p = 305However groups differed significantly on prior knowledge F(3164) = 104 p = 010 partial eta2 = 07 in that students in the controlgroup scored significantly higher on the comprehension pretest(M = 25 SD = 17) than students in both (p lt 05) the author-generated picture group (M = 17 SD = 16) and the drawing + author-generated picture group (M = 15 SD = 12) the drawing group(M = 20 SD = 14) did not differ significantly from the other groupsThus we included studentsrsquo prior knowledge as a covariate in thefollowing analyses

343 Is there support for the generative drawing effectA major goal in this experiment was to determine whether asking

students to generate drawings to represent science text is a moreeffective learning strategy than asking students to learn with textalone or with text and author-generated pictures In other wordswe wanted to determine whether we could replicate and extendthe learner-generated drawing effect Additionally we were inter-ested in whether giving students an author-generated picture afterdrawing can increase the benefits of the learning strategy Mean pro-portion correct and standard deviations on the comprehension anddrawing tests for the four groups are presented in Table 3

The left portion of Table 3 summarizes the mean proportioncorrect on the comprehension test A two-factorial analysis ofcovariance (ANCOVA) predicting learning outcomes (comprehen-sion posttest score) with learner-generated drawing (yesno) and

author-generated picture (yesno) as the factorial independent vari-ables and prior knowledge as a covariate showed a significantpositive main effect of learner-generated drawing F(1 163) = 398p = 048 partial eta2 = 02 a significant interaction effect F(1163) = 626 p = 013 partial eta2 = 04 but no main effect of author-generated pictures F lt 1 In addition multiple pairwise comparisons(with p lt 05) showed that the drawing group performed signifi-cantly better than each of the three other groups which did not differsignificantly from each other Cohenrsquos d favoring the drawing groupover the author-generated picture group was 49 over the learner-generated + author-generated picture group was 57 and over thecontrol group was 52

The right portion of Table 3 summarizes the mean proportioncorrect on the drawing posttest Again a two-factorial analysis ofcovariance (ANCOVA) predicting learning outcome (drawing testscore) with learner-generated drawing (yesno) and author-generatedpicture (yesno) as the factorial independent variables and priorknowledge as a covariate showed a significant positive main effectof learner-generated drawing F(1 163) = 6260 p lt 001 partialeta2 = 28 a significant positive main effect of author-generated pic-tures F(1 163) = 1104 p = 001 partial eta2 = 06 and a significantinteraction effect F(1 163) = 1658 p lt 001 partial eta2 = 09 In ad-dition multiple pairwise comparisons (with p lt 05) showed thatboth the drawing group and the drawing + author-generated picturegroup performed significantly better than the author-generatedpicture group (d = 68 d = 59) and the control group (d = 187d = 188) In turn the author-generated picture group performed sig-nificantly better than the control group (d = 95) The drawing groupand the drawing + author-generated picture group did not differ sig-nificantly from each other (d = 15)2 Overall these results areconsistent with Experiment 1 and provide additional support forthe generative drawing effect

In accordance with Experiment 1 we were interested in whetherdifferences in learning time among the experimental groups mediatethe positive effect of drawing on text comprehension First an ANOVApredicting learning time with learner-generated drawing (yesno)and author-generated picture (yesno) as the factorial indepen-dent variables showed a significant main effect of learner-generateddrawing F(1 164) = 39226 p lt 001 partial eta2 = 71 a significantmain effect of author-generated picture F(1 164) = 1685 p lt 001partial eta2 = 09 and a significant interaction effect F(1 164) = 490p = 028 partial eta2 = 03 Linear contrasts (with p lt 05) revealedthat the drawing group (M = 1938 min SD = 380) and thedrawing + author-generated picture group (M = 2340 min SD = 551)needed significantly more learning time than the author-generatedpicture group (M = 960 min SD = 397) and the control group(M = 834 min SD = 251) Thus to test whether learning time me-diates the positive effect of drawing on text comprehensionadditional mediation analyses (Baron amp Kenny 1986) were calcu-lated by including learning time as an additional predictor in theaforementioned linear model Results of the mediation analysesshowed that the effects of drawing on the comprehension posttestand the drawing posttest scores (see multiple pairwise compari-sons) are mediated by learning time to some extent That is includinglearning time in the linear model for predicting comprehension testscores still revealed a positive effect of the drawing group com-pared with the drawing + author-generated group on thecomprehension test (p = 012) However including learning time inthe linear model for predicting comprehension posttest scoresreduced the positive effect of the drawing group compared with theauthor-generated picture group (from p = 034 to p = 281) as wellas compared with the control group (from p = 002 to p = 087) being

2 The exclusion of studentsrsquo prior knowledge in the reported analyses does notchange the overall pattern of results

Table 3Mean proportion correct on the comprehension test and drawing test for the fourgroups ndash Experiment 2

Group Type of test

n Comprehension test Drawing test

M SD M SD

Learner-generated drawing 40 63 22 66 22Author-generated picture 44 53 19 50 25Learner-generated drawing +

author-generated picture41 51 20 63 19

Control 43 52 20 30 16

Note Asterisk () indicates significant difference from control group at p lt 05

283A Schmeck et alContemporary Educational Psychology 39 (2014) 275ndash286

no longer statistically significant Regarding the drawing posttestscore including learning time does not change the reported patternof results except that the positive effect of the drawing + author-generated picture group compared with the author-generated picturegroup is no longer statistically significant (from p = 004 to p = 223)

There were neither main effects of learner-generated drawingand author-generated pictures on the mental effort item (drawinggroup M = 455 SD = 025 author-generated picture group M = 459SD = 024 drawing + author-generated picture group M = 444SD = 026 control group M = 481 SD = 024 F lt 1) nor on theperceived difficulty item (drawing group M = 363 SD = 023 author-generated picture group M = 371 SD = 022 drawing + author-generated picture group M = 395 SD = 023 control group M = 393SD = 022 F lt 1)

Taken together the drawing strategy apparently fosters stu-dents to engage in generative activities indicated by their higherlearning outcomes Thus the data provide further evidence for thegenerative drawing effect predicted by Schwamborn et al (2010)In Experiment 2 benefits of the drawing activity however are me-diated by learning time and do not involve higher mental effortAdditionally there was no increased benefit when additional drawingsupport was available in the form of author-generated pictures

344 Is there support for the prognostic drawing effectA second major goal of this study was to determine whether the

prognostic drawing effect could be extended to a new context Meanproportion correct on drawing-accuracy during learning was 60(SD = 04) for the drawing group and 68 (SD = 03) for thedrawing + author generated picture group This difference betweenthe two drawing groups is not significant F(1 79) = 252 p = 116This lack of group differences allowed us to pool the data of bothdrawing groups for subsequent correlation analyses Correlation anal-yses based on the combined data from the two drawing groupsrevealed that the drawing-accuracy score of learner-generated draw-ings correlates significantly with the comprehension posttest scorer = 470 p lt 001 as well as with the drawing posttest score r = 615p lt 001 Additional correlation analyses revealed that the drawing-accuracy score of learner-generated drawings did not correlatesignificantly with the prior knowledge test score r = 095 p = 400the spatial ability test score r = 127 p = 257 the motivation testscore r = 033 p = 769 or the mental effort test score r = 042p = 712 The correlation between the drawing-accuracy score andthe perceived difficulty score was only slightly statistical signifi-cance r = minus218 p = 053 Thus the data provide further evidencefor the prognostic drawing effect consistent with the results ofSchwamborn et al (2010)

In sum results of Experiment 2 are partly consistent with theresults of Experiment 1 in that students learn better from a sciencetext when they are asked to draw illustrations representing the mainideas of the text and the quality of the generated drawings duringlearning correlates positively with studentsrsquo text comprehension (egSchwamborn et al 2010 van Meter 2001 van Meter et al 2006)

4 Discussion

41 Empirical contributions

The present set of experiments makes three empirical contri-butions to the field First this study shows strong and consistentevidence that students who are asked to generate drawings (withsufficient support) during reading a scientific text that describes acausal sequence perform better than students who read withoutdrawing both on a comprehension test (d = 085 in Experiment 1and d = 052 in Experiment 2) and on a drawing test (d = 115 in Ex-periment 1 and d = 187 in Experiment 2) Thus the generativedrawing effect can be extended to a new domain and therefore

corresponds to Shavelson and Townersquos (2002 p 4) recommenda-tion to ldquoreplicate and generalize across studiesrdquo as one of the sixessential scientific principles of educational research

Second this study shows strong and consistent evidence that thequality of drawings that students generate during learning with ascientific text that describes a causal sequence is positively relatedto subsequent performance on tests of learning outcome includ-ing a comprehension test (r = 623 in Experiment 1 and r = 470 inExperiment 2) and a drawing test (r = 620 in Experiment 1 andr = 615 in Experiment 2) Thus the prognostic drawing effect canbe replicated and extended to a new domain consistent with stan-dards for scientific research in education prescribed by Shavelsonand Towne (2002)

Third this study shows that asking learners to draw picturesduring reading a scientific text (ie learner-generated drawing groupin Experiment 2) is more effective than simply providing draw-ings (ie author-generated picture group in Experiment 2) both ona comprehension test (d = 049) and a drawing test (d = 068) Sim-ilarly adding author-generated drawings (ie learner-generatedpictures + author-generated pictures group in Experiment 2) doesnot improve the learning outcomes of students who also draw pic-tures during learning (ie learner-generated pictures group inExperiment 2) either on a comprehension test (d = minus057) or adrawing test (d = minus015) In short the act of drawing during learn-ing (with sufficient support) improves learning beyond the simpleprovision of drawings

42 Theoretical contributions

The results are consistent with the idea that drawing during learn-ing serves as a generative activity (Mayer amp Wittrock 2006Schwamborn et al 2010 van Meter amp Garner 2005 Wittrock 1990)That is the act of drawing encourages learners to engage in gen-erative cognitive processing during learning such as organizing therelevant information into a coherent structure and integrating itwith relevant prior knowledge from long-term memory In thepresent study positive effects of drawing were indicated with a com-prehension and a drawing learning outcome test and therefore arein line with the theoretical assumption derived from the GTDC thatbenefits of drawing can be found if learning outcome tests are usedthat are sensitive to the underlying process of drawing (cf van Meteramp Garner 2005) Additionally in our study the drawing activity wassupported in a way that was intended to help learners carry out theunderlying cognitive processes of drawing (ie selecting organiz-ing and integrating) successfully In this regard results of the presentstudy might supplement the theoretical framework of learner-generated drawing by providing further evidence that benefits ofdrawing defined by van Meter and Garnerrsquos GTDC can diminish ifno instructional support is given to constrain and structure thedrawing activity However a fuller understanding of the underly-ing cognitive processes of drawing and how these processes canbe influenced via drawing support requires more direct measuresof cognitive processing during learning Additionally following theidea that metacognitive processes of monitoring and regulation areautomatically activated by drawing (van Meter amp Garner 2005) afuller understanding of the metacognitive effects of drawings is alsorequired

43 Practical contributions

The present study encourages instructional designers and in-structors to incorporate drawing activities into venues involvinglearning from text which we call the generative drawing effect Oneimportant feature of a successful drawing strategy that is presentin this study and in a previous study by Schwamborn et al (2010)is that the drawing activity was supported by providing a

284 A Schmeck et alContemporary Educational Psychology 39 (2014) 275ndash286

background scene and a legend showing how to represent eachelement to constrain and structure the drawing activity Thus animportant practical implication is that students may need supportin their drawing activity so they do not need to draw from scratch

The present study also suggests a potentially useful diagnostictool to gauge the depth of student learning namely the quality ofthe drawings created by students during learning which we referto as the prognostic drawing effect Incorporating a measure of thequality of a learnerrsquos drawing during learning can be a useful toolin developing remedial instruction to give learners individual supportIt may be important to use materials that explain a cause-and-effect process and give learners drawings of the elements they needto represent the process pictorially Asking learners to simply drawpictures of elements is unlikely to be helpful whereas asking themto generate drawings that show the relations among the elementsin a schematic form is more likely to be helpful

44 Limitations and future directions

Some limitations and future directions of our study should beaddressed As noted in the theoretical contributions subsection wedid not have direct measures of cognitive processing during learn-ing so it is not possible to pinpoint how the drawing activity affectedspecific cognitive processes such as attending to relevant informa-tion organizing it and integrating it with prior knowledge We alsodid not assess metacognitive processing during learning thus it isnot possible to pinpoint how the drawing activity affected specificmetacognitive processes such as monitoring and regulation

Furthermore results of the cognitive load rating scales (in-vested mental effort and perceived task difficulty) are inconsistentWhereas in Experiment 1 an effect on mental effort but not on per-ceived task difficulty showed up (ie students in the drawing grouprated their invested mental effort during learning significantlyhigher) no effects on mental effort and task difficulty were foundin Experiment 2 Additionally in both experiments only a nega-tive correlation of perceived task difficulty with the quality of learner-generated pictures appears but no correlation of mental effort withthe quality Following de Jong (2010) those cognitive load ratingscales might have the disadvantages that they do not give a con-current measure of cognitive load and do not measure an essentialconcept in cognitive load theory namely cognitive overload (p 125)Future studies on learner-generated drawing might also use othercognitive load measures such as physiological measures as moredirect indicators of cognitive load

As noted in the practical contributions subsection we showedthe drawing effects by using a scientific text describing how a cause-and-effect system works that is the causal steps regarding aninfection with influenza and the immune response It might be pos-sible however that for other types of text producing drawings mightharm rather than promote text comprehension Thus to test whetherthe reported drawing effects can be extended future research hasto focus on other types of text such as descriptive texts as well ason other types of relations that can be conveyed with other typesof representations such as compare and contrast relations whichcan be shown in a matrix Additionally studentsrsquo learning out-comes were tested immediately after reading thus future work isneeded to investigate the longer-term effects of generative drawingon learning outcomes

Furthermore we only compared drawing with control groupsthat received no further learning strategy instructions However en-gaging in generative learning activities such as drawing requires aconsiderable amount of time Accordingly results showed that forExperiment 2 the positive effect of the drawing group on text com-prehension compared with the author-generated picture group andto the control group was mediated by learning time To rule out thatthe effects of drawing result only from additional time on task instead

of the generative activity future research should also compare thedrawing strategy with other time demanding generative learningstrategies such as summarization (cf Leopold amp Leutner 2012)

Another point that should be noted is that students in both ex-periments received some kind of multimedia materials in that evenwhen they had to draw and did not see presented pictures they wereat least provided with the basic (visual) elements for their draw-ings which they had to do on the given background which thusalso contained information In other words when students are pre-sented with important elements of the drawings which they canuse to draw themselves they will not have to put as much effortinto summarizing visually what they have just read compared withstudents who have to draw without any instructional help Futurestudies might also compare the drawing group with a summariza-tion group in which students receive a set of verbal key terms thatare similar to the drawing elements and are asked to make a textualsummary

Additionally future research is needed to validate the prognos-tic drawing effect So far we know that the quality of learner-generated pictures is related to studentsrsquo learning outcomes (iethe higher the learning outcome the higher the drawing accuracyand vice versa) and their perceived difficulty (ie the lower the per-ceived difficulty the higher the drawing accuracy and vice versa)and that it is not related to studentsrsquo prior knowledge motivationspatial ability or mental effort However less is known about whatthis might mean That is less is known regarding the causal direc-tion of this relation or the presence of a possible further moderatorvariable Do studentsrsquo efforts to produce accurate drawings lead tobetter comprehension and lower perceived difficulty Or do stu-dents who are more adept in drawing benefit more from the strategyand thus perceive the difficulty of the learning materials as beinglower Both arguments seem convincing

Finally more work is needed to determine the level of supportthat makes the drawing strategy most effective for various kinds oflearners As noted in the empirical contribution adding author-generated drawings (ie learner-generated pictures + author-generated pictures group in Experiment 2) does not improve thelearning outcomes of students who also draw pictures during learn-ing and were supported by a drawing prompt In other words thecombination of two ways of supporting the drawing strategy (iegiving a drawing prompt during reading plus an author-generatedpicture after reading) did not improve studentsrsquo learning out-comes compared with students in the drawing group as well ascompared with students in the control and author-generated pic-tures only groups This result is inconsistent with previous research(eg van Meter 2001 van Meter et al 2006) which found that com-paring own drawings to author-generated pictures normally helpslearning van Meter and colleagues (2001 2006) however provid-ed author-generated pictures plus prompting questions after thedrawing process That is students answered prompting questionsto guide the comparison process between their self-generateddrawing and the author-generated drawing In our study studentswere only instructed to generate a drawing to inspect an author-generated one and to check whether their own drawing incomparison with the author-generated one really represented themain ideas of the text paragraph correctly In other words we didnot guide the process of comparing self-generated drawings withauthor-generated ones As a potential consequence students per-formed the intended comparison process inadequately or even notall and thus did not benefit from it One reason for this inade-quate comparison process might be that students need guidancein doing the comparison process Another reason might be the factthat students do not seriously engage in generating drawings oncethey notice that there are author-generated drawings Thus futureresearch should also use additional guidance to test whether thecombination of different ways of supporting the drawing strategy

285A Schmeck et alContemporary Educational Psychology 39 (2014) 275ndash286

(ie giving a drawing prompt during reading plus an author-generated picture after reading) helps learning as well asobservational measures of the drawing process itself (ie think aloudprotocols) to shed more light on the cognitive processes underly-ing the drawing activities

Overall drawing during learning from text appears to be a po-tentially powerful strategy for improving studentsrsquo learning fromscientific text when certain boundaries and prerequisites are takeninto account

Acknowledgments

This article is based on a research project funded by the GermanResearch Foundation (DFG LE 6459-3 as part of FOR 511) We wouldlike to thank Angela Sandmann for her assistance in developing thelearning materials

References

Ainsworth S Prain V amp Tytler R (2011) Drawing to learn in science Science 3331096ndash1097

Alesandrini K L (1981) Pictorial-verbal and analytic-holistic learning strategies inscience learning Journal of Educational Psychology 73 358ndash368

Alesandrini K L (1984) Pictures and adult learning Instructional Science 13 63ndash77Baron R M amp Kenny D A (1986) The moderator-mediator variable distinction in

social psychological research Conceptual strategic and statistical considerationsJournal of Personality and Social Psychology 51 1173ndash1182

Bruumlnken R Plass J L amp Leutner D (2003) Direct measurement of cognitive loadin multimedia learning Educational Psychologist 38 53ndash61

Carney RN amp Levin JR (2002) Pictorial illustrations still improve studentsrsquo learningfrom text Educational Psychology Review 14 5ndash26

de Jong T (2005) The guided discovery principle in multimedia learning In R EMayer (Ed) The Cambridge handbook of multimedia learning (pp 215ndash228) NewYork Cambridge University Press

de Jong T (2010) Cognitive load theory educational research and instructionaldesign Some food for thought Instructional Science 38 105ndash134

Ekstrom R B French J W amp Harman H H (1976) Manual for kit of factor-referencedcognitive tests Princeton NJ Educational Testing Service

Greene T R (1989) Childrenrsquos understanding of class inclusion hierarchies Therelationship between external representation and task performance Journal ofExperimental Child Psychology 48 62ndash89

Hall V C Bailey J amp Tillman C (1997) Can student-generated illustrations be worthten thousand words Journal of Educational Psychology 89 677ndash681

Houmlffler T N (2010) Spatial ability Its influence on Learning with visualizations ndashA meta-analytic review Educational Psychology Review 22 245ndash269

Houmlffler T N Schmeck A amp Opfermann M (2013) Static and dynamic visualrepresentations Individual differences in processing In G Schraw M TMcCrudden amp D Robinson (Eds) Learning through visual displays (pp 133ndash163)Charlotte NC Information Age Publishing

Kalyuga S Chandler P amp Sweller J (1999) Managing split-attention and redundancyin multimedia instruction Applied Cognitive Psychology 13 351ndash371

Leopold C (2009) Lernstrategien und Textverstehen [Learning strategies and textcomprehension] Muumlnster Waxmann

Leopold C amp Leutner D (2012) Science text comprehension Drawing main ideaselection and summarizing as learning strategies Learning and Instruction 2216ndash26

Lesgold A M DeGood H amp Levin J R (1977) Pictures and young childrenrsquos proselearning A supplementary report Journal of Reading Behavior 9 353ndash360

Lesgold A M Levin J R Shimron J amp Guttman J (1975) Pictures andyoung childrenrsquos learning from oral prose Journal of Educational Psychology 67636ndash642

Leutner D Leopold C amp Sumfleth E (2009) Cognitive load and science textcomprehension Effects of drawing and mentally imagining text contentComputers in Human Behavior 25 284ndash289

Mayer R E (2004) Should there be a three-strikes rule against pure discoverylearning The case for guided methods of instruction The American Psychologist59 14ndash19

Mayer R E (2005) Cognitive theory of multimedia learning In R E Mayer (Ed)The Cambridge handbook of multimedia learning (pp 31ndash48) New York CambridgeUniversity Press

Mayer R E (2009) Multimedia learning (2nd ed) New York NY CambridgeUniversity Press

Mayer R E amp Wittrock M C (2006) Problem solving In P Alexander P Winne ampG Phye (Eds) Handbook of educational psychology (pp 287ndash303) Mahwah NJErlbaum

Paas F (1992) Training strategies for attaining transfer of problem-solving skill instatisticsmdashA cognitive-load approach Journal of Educational Psychology 84429ndash434

Paas F Tuovinen J Tabbers H K amp Van Gerven P W M (2003) Cognitive loadmeasurement as a means to advance cognitive load theory EducationalPsychologist 38 63ndash71

Paivio A (1986) Mental representation A dual coding approach New York OxfordUniversity Press

Pashler H Bain P Bottage B Graesser A Koedinger K McDaniel M et al (2007)Organizing instruction and study to improve student learning Washington DCNational Center for Educational Research

Rasco R W Tennyson R D amp Boutwell R C (1975) Imagery instructions anddrawings in learning prose Journal of Educational Psychology 67 188ndash192

Rheinberg F Vollmeyer R amp Burns B D (2001) FAM Ein fragebogen zurerfassung aktueller motivation in lern- und leistungssituationen [QCM Aquestionnaire to assess current motivation in learning situations] Diagnostica47 57ndash66

Schnotz W (2005) An integrated model of text and picture comprehension In RE Mayer (Ed) The Cambridge handbook of multimedia learning (pp 49ndash70) NewYork Cambridge University Press

Schwamborn A Mayer R E Thillmann H Leopold C amp Leutner D (2010) Drawingas a generative activity and drawing as a prognostic activity Journal of EducationalPsychology 102 872ndash879

Schwamborn A Thillmann H Opfermann M amp Leutner D (2011) Cognitive loadand instructionally supported learning with provided and learner-generatedvisualizations Computers in Human Behavior 27 89ndash93

Shavelson R J amp Towne L (Eds) (2002) Scientific research in education WashingtonDC National Academy Press

Sweller J Ayres P amp Kalyuga S (2011) Cognitive Load Theory New York SpringerTirre W C Manelis L amp Leicht K (1979) The effects of imaginal and verbal strategies

on prose comprehension by adults Journal of Reading Behavior 11 99ndash106van Meter P (2001) Drawing construction as a strategy for learning from text Journal

of Educational Psychology 69 129ndash140van Meter P Aleksic M Schwartz A amp Garner J (2006) Learner-generated drawing

as a strategy for learning from content area text Contemporary EducationalPsychology 31 142ndash166

van Meter P amp Garner J (2005) The promise and practice of learner-generateddrawings Literature review and synthesis Educational Psychology Review 12261ndash312

Van Gog T amp Paas F (2008) Instructional efficiency Revisiting the original constructin educational research Educational Psychologist 43 16ndash26

Vollmeyer R amp Rheinberg F (2000) Does motivation affect learning via persistenceLearning and Instruction 4 293ndash309

Weinstein C E amp Mayer R E (1986) The teaching of learning strategies In M CWittrock (Ed) Handbook of research on teaching (pp 315ndash327) New YorkMacmillan

Wittrock M C (1990) Generative processes of comprehension EducationalPsychologist 24 345ndash376

286 A Schmeck et alContemporary Educational Psychology 39 (2014) 275ndash286

  • Drawing pictures during learning from scientific text testing the generative drawing effect and the prognostic drawing effect
  • Introduction
  • Theoretical framework for the learner-generated drawing strategy
  • Empirical framework for the learner-generated drawing strategy
  • Effectiveness of learner-generated drawings
  • Quality of learner-generated drawings
  • Overview of the experiments
  • Experiment 1
  • Participants and design
  • Materials
  • Procedure
  • Results and discussion
  • Scoring
  • Are the groups equivalent on basic characteristics
  • Is there support for the generative drawing effect
  • Is there support for the prognostic drawing effect
  • Experiment 2
  • Participants and design
  • Materials
  • Procedure
  • Results
  • Scoring
  • Are the groups equivalent on basic characteristics
  • Is there support for the generative drawing effect
  • Is there support for the prognostic drawing effect
  • Discussion
  • Empirical contributions
  • Theoretical contributions
  • Practical contributions
  • Limitations and future directions
  • Acknowledgments
  • References

equivalent on basic characteristics They included a participant ques-tionnaire a comprehension pretest a spatial ability test and amotivation questionnaire The participant questionnaire solicited in-formation concerning the studentsrsquo age and sex The comprehensionpretest consisted of 25 multiple-choice items and was intended toassess studentsrsquo prior-knowledge of information covered in the textStudentsrsquo spatial ability was measured with a 10 multiple-choicepaper-folding items taken from a battery of cognitive tests devel-oped by Ekstrom French and Harman (1976) The motivationquestionnaire assessed studentsrsquo current motivation for doing thelearning task after reading the instructions before the lesson It con-sisted of nine items from the challenge and interest subscales of theQuestionnaire on Current Motivation (QCM) developed by RheinbergVollmeyer and Burns (2001) Cognitive load by means of investedmental effort was measured using the 7-point subjective rating scaledeveloped by Paas (1992) which ranges from (1) very low mentaleffort to (7) very high mental effort Cognitive load by means of per-ceived task difficulty was measured using the 7-point subjective ratingscale developed by Kalyuga Chandler and Sweller (1999) whichranges from (1) very easy to (7) very difficult These subjective mea-sures have been criticized for assessing cognitive load with only singleitems (eg Bruumlnken Plass amp Leutner 2003) However several studiesshowed the effectiveness of the rating scale by showing that the vari-ation in learnersrsquo cognitive load ratings depended on variations intask complexity or instructional design (for overviews see PaasTuovinen Tabbers amp Van Gerven 2003 Van Gog amp Paas 2008) Inthis regard Sweller Ayres and Kalyuga (2011) conclude that ldquothesimple subjective rating scale regardless of the wording used (mentaleffort or difficulty) has perhaps surprisingly been shown to be themost sensitive measure available to differentiate the cognitive loadimposed by different instructional proceduresrdquo (p 74) For that reasonand due to the economic applicability we decided to use this kindof cognitive load measurement while acknowledging the limita-tions of a short self-report instrument

The two learning booklets each included a science text on thebiology of the influenza The text explained the causal steps re-garding an infection with influenza and regarding the immuneresponse which is an unfamiliar subject for eighth graders in highertrack secondary schools due to the German curriculum The text con-sisted of approximately 850 words (in German) and was divided intoseven paragraphs (as shown in Table 1)

The drawing version of the booklet contained seven pairs of facingpages with a text paragraph on the left page and a two-part drawingprompt on the right page The first part of the drawing prompt in-cluded a legend showing all the relevant elements (in total eightelements) for drawing a picture for that text paragraph (as shownin the top of Fig 1) The second part of the drawing prompt in-cluded a partly pre-drawn background for studentsrsquo drawing (asshown in the bottom of Fig 1) Overall students had to make sevendrawings ie one drawing to each paragraph

The control version of the learning booklet contained four pairsof facing pages with one of the seven text paragraph on each page

Students in both groups learned with exactly the same text mate-rial To make sure that students in the control group learnedwith the same amount of information as students in the drawinggroup all elements of the drawing prompt as well as the spatialrelations between these elements were also described in the sciencetext

The two posttests intended to assess the learning outcomes werea comprehension posttest and a drawing posttest The comprehen-sion posttest (Cronbachrsquos alpha = 083) consisted of 25 multiple-choice items (the same items as in the comprehension pretest) andwas intended to assess studentsrsquo comprehension of the factual andconceptual information covered in the text as well as their abilityto transfer what was presented to new situations An item exampleis ldquoT-helper cells do not only recognize viruses but also agents thatare extraneous to the body Which medication would you admin-ister to a patient who has received a new kidney (a) a medicinethat suppresses the immune response of the body (b) a medicinethat activates the immune response of the body (c) a medicine thatcontains antigens or (d) a medicine that contains blood of the kidneydonorrdquo [(a) is the correct answer] The drawing test (Cronbachrsquosalpha = 081) was intended to assess studentsrsquo comprehension of theconceptual information presented in the science text by means ofdrawing That is students had to reproduce the main ideas givenin the text by drawing It consisted of three drawing items in whichstudents were asked to draw sketches depicting key concepts of thetext and their spatial relations An item example for the drawingtest is ldquoHow does an influenza virus invade a cell and how is it re-producedrdquo The science text the drawing prompt and the learningoutcome tests were constructed by the first author in cooperationwith a biology teacher The materials were adapted fromSchwamborn et al (2010) however using another science domainand including measures of individual learning times and cognitiveload

23 Procedure

Participants were tested in the schoolsrsquo classrooms Within theirclasses they were randomly assigned to one of the two groupsGroups were tested in separate classrooms in order to insure thatstudents in the drawing group did not feel rushed when studentsin the control group completed the task early Each student wasseated at an individual desk First students were given the partic-ipant questionnaire and the comprehension pretest to complete attheir own rate Second students filled in the paper-folding test witha 3 min time limit Third students were given instructional book-lets corresponding to their assigned group After they had read theinstructions for reading the booklets studentsrsquo current motivationfor doing the learning task was assessed Next students started learn-ing with the text material corresponding to their treatment groupStudents were instructed to carefully read the text on the biologyof the influenza in order to comprehend the material Students inthe drawing condition were instructed to read the text and addi-tionally to draw pictures for each text paragraph using the drawingprompt representing the main ideas of each text paragraph Thatis students had to use the pictorial elements given in the legendsuch as the virus as templates for their own paper-pencil baseddrawing across the pre-dawn background Students in the controlgroup were instructed to read the text for comprehension but werenot instructed to engage in drawing Students in both groups learnedat their own pace whereby individual learning time was mea-sured by the instructors in the classrooms Fourth in order to ensurecomparable testing procedures after finishing learning with thewhole learning material students in both groups directly rated theamount of mental effort he or she had invested during learning andthe amount of difficulty he or she had perceived during learningFifth students received the comprehension posttest consisting of

Table 1Text from the second paragraph of the influenza lesson

How the influenza virus replicates

Once inside the influenza virus uses your somatic cell to produce new particlesof the influenza-virus The glycoproteins move toward the membrane of thesomatic cell and stick out into the outside of the cell The capsules of thevirus however are assembled inside the somatic cell Next these newassembled capsules of the virus leave your somatic cell By moving throughthe somatic cell membrane the capsules are enveloped with the membraneand its glycoproteins which then plays the role of the virus membraneThus several new influenza viruses are located outside your somatic cell

Note Translated from the German original

279A Schmeck et alContemporary Educational Psychology 39 (2014) 275ndash286

the comprehension posttest and the drawing posttest Students had20 minutes time for completion and did not have access to thescience text or their drawings Finally students were thanked anddebriefed As students learned at their own rate the whole proce-dure took about 70ndash90 minutes depending on the individual testingtimes

24 Results and discussion

241 ScoringThe dependent variables were studentsrsquo scores on the compre-

hension and drawing posttests studentsrsquo rating on the mental effortand the difficulty scales and the drawing accuracy score indicat-ing the quality of learner-generated drawings produced by studentsin the drawing group during the learning phase

The comprehension test score (pre- and posttest) for each studentwas computed by awarding 1 point for each correct answer and byadding up the points to obtain the total comprehension score (outof a total possible of 25 points) Actual scores ranged from 3 to 24points with a mean of 13 points (SD = 53) Following Schwambornet al (2010) scoring of the drawing test was carried out by count-ing the total number of correct main ideas in each learnerrsquos answeracross the three drawing items The main ideas were drawn out fromboth expert visualizations and a checklist specifying important re-lational features Students could earn a maximum of 19 points onthe drawing test Two student assistants (teacher trainees in biology)scored the quality for each of the three drawings for each studentwith an acceptable inter-rater agreement of GoodmanndashKruskalgamma of 090 Actual scores ranged from 0 to 185 points with amean of 77 points (SD = 47) Total scores of both the comprehen-sion and the drawing test were transferred into z-standardized scoresto make them comparable across studies

The drawing accuracy score (concerning drawing during learn-ing in the drawing group) was computed by using a coding schemeadapted from Schwamborn et al (2010) which was based on expertdrawings and a checklist specifying important relational featuresof the drawings Students could earn a maximum drawing-accuracyscore of 22 points Again the two student assistants scored each ofthe seven learner-generated drawings for each student with an ac-ceptable interrater agreement of GoodmanndashKruskal gamma of 92Both coding schemes were constructed by the first author and abiology teacher Actual scores ranged from 4 to 21 points with amean of 133 points (SD = 50) The total drawing accuracy score wasagain transferred into a z-standardized score

In addition the spatial ability test was scored by tallying thenumber correct out of 10 and the motivation questionnaire wasscored by tallying the nine ratings on both subscales to a total scoreof motivation Finally for comparing performance across the dif-ferent tests the proportion correct on each test was computed bydividing the studentrsquos obtained score by the total possible score

242 Are the groups equivalent on basic characteristicsBefore looking at treatment effects on the dependent variables

we analyzed whether the two groups differed on several control vari-ables A chi-square analysis indicated that there were no significantdifferences regarding gender (p = 562) Separate univariate analy-ses of variance (ANOVAs) revealed that the groups did not differsignificantly on age F lt 1 on spatial ability F lt 1 or on motiva-tion F(1 46) = 360 p = 064 However groups differed significantlyon prior knowledge F(1 46) = 3890 p lt 001 partial eta2 = 46 inthat students in the drawing group scored significantly lower onthe comprehension pretest (M = 10 SD = 15) than students in thecontrol group (M = 34 SD = 12) Thus we included studentsrsquo priorknowledge in the following analyses

243 Is there support for the generative drawing effectMean proportion correct and SDs on the comprehension and

drawing posttests for both groups are presented in Table 2 Repeat-ed measures univariate analyses of variance (ANOVA) with thecomprehension pre- and post-test scores as the within-subject factorsand group (drawing versus control) as the between-subject factorshowed a main effect over time indicating that overall partici-pants reached significant knowledge gains between thecomprehension pretest and the comprehension posttest F(146) = 9897 p lt 001 partial eta2 = 68 An interaction additionallyshowed that these knowledge gains were significantly higher forthe drawing group than for the control group F(1 46) = 4617p lt 001 partial eta2 = 50

For the drawing test a repeated measures ANOVA was not pos-sible since these items were only used in the posttest In this casea univariate analysis of covariance (ANCOVA) predicting the drawingtest score with group (drawing versus control) as the factorial in-dependent variable and prior knowledge as a covariate showed thatthe drawing group scored significantly better than the control groupon the drawing posttest F(1 45) = 1349 p = 001 partial eta2 = 231

Cohenrsquos d favoring the drawing group over the control group was085 for the comprehension posttest and 115 for the drawingposttest all of which are considered large effects Thus there is strongsupport for the generative drawing effect as predicted

Additionally results revealed that the drawing group needed sig-nificantly more learning time (M = 2108 min SD = 424) than thecontrol group (M = 1738 min SD = 333) F(1 46) = 1134 p = 002partial eta2 = 20 Thus to test whether learning time mediates thepositive effect of drawing on text comprehension additional me-diation analyses (Baron amp Kenny 1986) were calculated by includinglearning time as an additional predictor in the aforementioned linearmodel A mediation effect would be detected if in this case effectsof drawing on text comprehension would significantly decreaseResults of the mediation analyses showed that the effect of drawingon both comprehension test scores and drawing test scores was notfully mediated by learning time That is including learning time stillrevealed the interaction between group (drawing versus control)and time (pre- versus post) in that the drawing group had signifi-cantly higher knowledge gains than the control group on thecomprehension test items (p lt 001) Furthermore the drawing groupalso still outperformed the control group on the drawing posttestafter controlling for learning time (p = 009)

Furthermore results revealed that students in the drawing grouprated their invested mental effort during learning significantly higher(M = 504 SD = 112) than students in the control group (M = 396SD = 165) F(1 46) = 705 p = 011 partial eta2 = 13 There was nodifference between the two groups on the perceived difficulty item(drawing group M = 408 SD = 150 control group M = 425SD = 122 F lt 1) Thus consistent with predictions concerning thegenerative drawing effect there is partial support for the idea that

1 The exclusion of studentsrsquo prior knowledge in the reported analyses does notchange the reported pattern of results

Table 2Mean proportion correct on the comprehension test and drawing test for two groupsndash Experiment 1

Group Type of test

n Comprehension test Drawing test

M SD M SD

Drawing 24 61 20 52 27Control 24 44 20 28 11

Note Asterisk () indicates significant difference from control group at p lt 05

280 A Schmeck et alContemporary Educational Psychology 39 (2014) 275ndash286

drawing causes students to engage in more generative processingduring learning

Taken together the results suggest that the drawing strategy en-courages students to engage in generative processing during learningas is indicated by their higher learning outcomes Thus the dataprovide further evidence for the generative drawing effect consis-tent with the results of Schwamborn et al (2010) Additionallyresults indicate that students in the drawing condition seem to investmore mental effort than students in the control group without per-ceiving higher levels of difficulty

244 Is there support for the prognostic drawing effectMean proportion correct on drawing accuracy during learning

was 59 (SD = 23) A correlation analysis revealed that the drawing-accuracy score of learner-generated drawings correlated significantlywith the comprehension posttest score r = 620 p lt 001 and withthe drawing posttest score r = 623 p lt 001 Additional correla-tion analyses revealed that the drawing-accuracy score of learner-generated drawings correlated significantly negatively with theperceived difficulty score r = minus489 p = 015 There were no signif-icant correlations between the drawing accuracy score and eitherthe invested mental effort score r = minus134 p = 533 the prior knowl-edge test score r = minus004 p = 984 the spatial ability test score r = 072p = 739 or the motivation test score r = 086 p = 690 Thus as pre-dicted the data provide further evidence for the prognostic drawingeffect consistent with the results of Schwamborn et al (2010)

In sum the results of Experiment 1 are consistent with the pre-diction that students learn better from a science text when they areasked to draw illustrations representing the main ideas of the textand that the quality of the generated drawings during learning cor-relates positively with studentsrsquo text comprehension (egSchwamborn et al 2010 van Meter 2001 van Meter et al 2006)

However it might be argued that the reported results are dueto the way we supported the strategy use In other words the re-ported positive effect of the learner-generated drawing strategy mightnot be caused by studentsrsquo engagement in generative learning ac-tivities during reading (de Jong 2005 Mayer 2004 2009 Wittrock1990) but rather by the additional pictorial information given in thedrawing prompt Additionally looking at studentsrsquo learning out-comes our results indeed show positive effects of drawing howevermean scores of learning outcomes for the drawing group aremedium-sized Thus it might be argued that the way we sup-ported the strategy use was not fully sufficient In other words thereported positive effect of the learner-generated drawing strategyie the generative drawing effect might be increased by giving stu-dents instructional support in addition to the drawing prompt (vanMeter 2001 van Meter et al 2006 van Meter amp Garner 2005) Toaddress these issues we added two experimental conditions byimplementing author-generated pictures in the design in Experi-ment 2

3 Experiment 2

One possible issue with Experiment 1 is the type of control groupused In Experiment 1 following Schwamborn et al (2010) we useda reading only control group in which the control group learnedwith verbal information only In the drawing group however stu-dents not only learned with verbal information but also with pictorialinformation given by the drawing prompt Based on theories of mul-timedia learning the use of different forms of representations suchas texts and pictures can promote learning in that ldquopeople learnbetter from words and pictures than from words alonerdquo (ie mul-timedia principle Mayer 2009 p 223) because in this caseboth a (verbal) propositional representation as well as a (pictori-al) mental model are built up and are optimally integrated into oneschema that can be stored in long-term memory (Schnotz 2005)

This assumption is also in line with the dual-coding approach statedby Paivio (1986) In this regard it might be argued that the re-ported drawing effect is actually a multimedia effect that is basedon the presentation of text and picture rather than a generativedrawing effect that is based on studentsrsquo active engagement indrawing activities during reading In other words instead of askingpeople to draw pictures representing the main ideas of the textgiving them text and author-generated pictures representing themain ideas of the text might be as good or even better Thus weincluded a condition in Experiment 2 in which we added author-generated pictures to the text

An additional issue with Experiment 1 is whether the reportedgenerative drawing effect can be enhanced by using various formsof supporting the strategy First there is evidence that using adrawing prompt during learning seems to be effective in support-ing the learner-generated drawing strategy by minimizing thecreation of extraneous processing (cf Schwamborn et al 2010 seealso Exp 1) Second research has shown that instructing studentsto compare their own drawing with an author-generated picturemight be also effective in supporting the learner-generated drawingstrategy as self-monitoring processes are enhanced (cf van Meter2001) Up to now however there is no empirical evidence whetherthe combination of both ways to support the drawing strategy hasan additive effect on learning outcomes Thus we included a furthercondition in Experiment 2 in which we combined both forms ofstrategy support

The main purpose of Experiment 2 was to test the generativedrawing and prognostic drawing effects of learner-generated drawingas in Experiment 1 but this time also compared with another controlgroup (ie author-generated pictures) Additionally we were in-terested in testing whether the benefits of the learner-generateddrawing strategy can be increased when we instructionally supportstudents not only with a drawing prompt but also with an author-generated picture after the drawing process In this new treatmentwe instructed students to draw a picture of the text content andthen to compare their own drawing with an expert picture

31 Participants and design

The participants were 168 German eighth graders from highertrack secondary schools The mean age was 138 years (SD = 06)and there were 112 girls and 56 boys The study was based on a2 times 2-between-subjects design with learner-generated drawing (yesno) and author-generated picture (yesno) as factors Forty studentsserved in the drawing group 44 students served in the author-generated picture group 41 students served in the drawing + author-generated picture group and 43 students served in the control group

32 Materials

The materials were identical to those used in Experiment 1 exceptthat we used a shortened version of the comprehension pretest thatconsisted of 19 rather than 25 items (Cronbachrsquos alpha = 70) andslightly extended versions of both the comprehension posttest (28items Cronbachrsquos alpha = 84) and the drawing test (four items witha maximum score of 21 points Cronbachrsquos alpha = 78) The pretestwas shortened because the first experiment showed that the re-spective items were either much too easy or much too difficult andthus unsuitable to differentiate between successful and unsuccess-ful learners thus we deleted these items in the second experimentFurthermore we decided to add some items to the comprehen-sion posttest in the second experiment because during data analysisof the first experiment and after receiving some feedback fromexperts in the domain of biology we recognized that a few itemsassessing transfer ability could be added These transfer itemshowever would have been unsuitable to be included in the pretest

281A Schmeck et alContemporary Educational Psychology 39 (2014) 275ndash286

because they are too difficult to answer without prior training inthe topic Additionally author-generated pictures were used in thenew conditions The author-generated pictures were static func-tional pictures representing the main ideas of each paragraph andconsisted of pictorial elements identical to those provided in thedrawing prompt (as shown in Fig 2) These pictures were con-structed by the first author in cooperation with a biology teacher

The drawing version of the booklet was identical to that usedin Experiment 1 (as shown in Fig 1) The control version of the learn-ing booklet was identical to that used in Experiment 1 The author-generated picture version of the booklet consisted of seven pairsof facing pages with a text paragraph on the left page and a corre-sponding author-generated picture (such as in Fig 2) on the rightpage The drawing + author-generated picture version of the bookletcontained the material from the drawing version consisting of sevenpairs of facing pages with a text paragraph on the left page and atwo-part drawing prompt on the right page In addition attachedto each page there was an additional page that students could foldout after having generated their drawing When unfolding this ad-ditional page a picture of that text paragraph right aside the drawingprompt was provided and there was an additional instruction to

compare the learner-generated drawing with the author-generatedpicture Author-generated pictures were the same as in the author-generated picture version of the booklet

33 Procedure

The procedure was identical to that used in Experiment 1 exceptthat there were two additional groups learning with author-generated pictures Students in the author-generated picturecondition were instructed to read the text and additionally to lookat pictures representing the main ideas of each text paragraph Stu-dents in the drawing + author-generated picture version of thebooklet were instructed to read the text to draw pictures for eachtext paragraph using the drawing prompt representing the mainideas of each text paragraph and finally to compare their picturewith an author-generated picture representing main ideas of eachparagraph correctly Students in all groups learned at their own pacewhereby individual learning time was measured by the instruc-tors in the classrooms Again to ensure that studentsrsquo in both drawinggroups did not feel rushed when students in the non-drawing

Fig 2 Author-generated pictures for the seven paragraphs in the author-generated picture versions of the learning booklet Note Pictures are scaled-down from the orig-inal format

282 A Schmeck et alContemporary Educational Psychology 39 (2014) 275ndash286

group completed the task early groups were tested in separateclassrooms

34 Results

341 ScoringAll tests instruments were scored with the same procedures used

in Experiment 1 Again two student assistants (teacher trainees inbiology) scored each of the drawing test items and each of the sevenlearner-generated drawings for each student with acceptableinterrater agreements (drawing test GoodmanndashKruskal gamma of90 drawing-accuracy GoodmanndashKruskal gamma of 94) Actualscores ranged from 1 to 28 points (M = 153 points SD = 58) for thecomprehension test from zero to 21 points (M = 109 points SD = 53)for the drawing test and from 275 to 215 points (M = 141 pointsSD = 48) for drawing accuracy Again total scores of comprehen-sion drawing and accuracy were transferred into z-standardizedscores

342 Are the groups equivalent on basic characteristicsBefore looking at treatment effects on the dependent variables

we analyzed whether the four experimental groups differed onseveral control variables A chi-square analysis indicated that therewere no significant differences regarding gender (p = 097) Sepa-rate univariate analyses of variance (ANOVAs) revealed that thegroups also did not differ significantly on age F lt 1 on spatial abilityF(3 164) = 120 p = 312 or on motivation F(3 164) = 122 p = 305However groups differed significantly on prior knowledge F(3164) = 104 p = 010 partial eta2 = 07 in that students in the controlgroup scored significantly higher on the comprehension pretest(M = 25 SD = 17) than students in both (p lt 05) the author-generated picture group (M = 17 SD = 16) and the drawing + author-generated picture group (M = 15 SD = 12) the drawing group(M = 20 SD = 14) did not differ significantly from the other groupsThus we included studentsrsquo prior knowledge as a covariate in thefollowing analyses

343 Is there support for the generative drawing effectA major goal in this experiment was to determine whether asking

students to generate drawings to represent science text is a moreeffective learning strategy than asking students to learn with textalone or with text and author-generated pictures In other wordswe wanted to determine whether we could replicate and extendthe learner-generated drawing effect Additionally we were inter-ested in whether giving students an author-generated picture afterdrawing can increase the benefits of the learning strategy Mean pro-portion correct and standard deviations on the comprehension anddrawing tests for the four groups are presented in Table 3

The left portion of Table 3 summarizes the mean proportioncorrect on the comprehension test A two-factorial analysis ofcovariance (ANCOVA) predicting learning outcomes (comprehen-sion posttest score) with learner-generated drawing (yesno) and

author-generated picture (yesno) as the factorial independent vari-ables and prior knowledge as a covariate showed a significantpositive main effect of learner-generated drawing F(1 163) = 398p = 048 partial eta2 = 02 a significant interaction effect F(1163) = 626 p = 013 partial eta2 = 04 but no main effect of author-generated pictures F lt 1 In addition multiple pairwise comparisons(with p lt 05) showed that the drawing group performed signifi-cantly better than each of the three other groups which did not differsignificantly from each other Cohenrsquos d favoring the drawing groupover the author-generated picture group was 49 over the learner-generated + author-generated picture group was 57 and over thecontrol group was 52

The right portion of Table 3 summarizes the mean proportioncorrect on the drawing posttest Again a two-factorial analysis ofcovariance (ANCOVA) predicting learning outcome (drawing testscore) with learner-generated drawing (yesno) and author-generatedpicture (yesno) as the factorial independent variables and priorknowledge as a covariate showed a significant positive main effectof learner-generated drawing F(1 163) = 6260 p lt 001 partialeta2 = 28 a significant positive main effect of author-generated pic-tures F(1 163) = 1104 p = 001 partial eta2 = 06 and a significantinteraction effect F(1 163) = 1658 p lt 001 partial eta2 = 09 In ad-dition multiple pairwise comparisons (with p lt 05) showed thatboth the drawing group and the drawing + author-generated picturegroup performed significantly better than the author-generatedpicture group (d = 68 d = 59) and the control group (d = 187d = 188) In turn the author-generated picture group performed sig-nificantly better than the control group (d = 95) The drawing groupand the drawing + author-generated picture group did not differ sig-nificantly from each other (d = 15)2 Overall these results areconsistent with Experiment 1 and provide additional support forthe generative drawing effect

In accordance with Experiment 1 we were interested in whetherdifferences in learning time among the experimental groups mediatethe positive effect of drawing on text comprehension First an ANOVApredicting learning time with learner-generated drawing (yesno)and author-generated picture (yesno) as the factorial indepen-dent variables showed a significant main effect of learner-generateddrawing F(1 164) = 39226 p lt 001 partial eta2 = 71 a significantmain effect of author-generated picture F(1 164) = 1685 p lt 001partial eta2 = 09 and a significant interaction effect F(1 164) = 490p = 028 partial eta2 = 03 Linear contrasts (with p lt 05) revealedthat the drawing group (M = 1938 min SD = 380) and thedrawing + author-generated picture group (M = 2340 min SD = 551)needed significantly more learning time than the author-generatedpicture group (M = 960 min SD = 397) and the control group(M = 834 min SD = 251) Thus to test whether learning time me-diates the positive effect of drawing on text comprehensionadditional mediation analyses (Baron amp Kenny 1986) were calcu-lated by including learning time as an additional predictor in theaforementioned linear model Results of the mediation analysesshowed that the effects of drawing on the comprehension posttestand the drawing posttest scores (see multiple pairwise compari-sons) are mediated by learning time to some extent That is includinglearning time in the linear model for predicting comprehension testscores still revealed a positive effect of the drawing group com-pared with the drawing + author-generated group on thecomprehension test (p = 012) However including learning time inthe linear model for predicting comprehension posttest scoresreduced the positive effect of the drawing group compared with theauthor-generated picture group (from p = 034 to p = 281) as wellas compared with the control group (from p = 002 to p = 087) being

2 The exclusion of studentsrsquo prior knowledge in the reported analyses does notchange the overall pattern of results

Table 3Mean proportion correct on the comprehension test and drawing test for the fourgroups ndash Experiment 2

Group Type of test

n Comprehension test Drawing test

M SD M SD

Learner-generated drawing 40 63 22 66 22Author-generated picture 44 53 19 50 25Learner-generated drawing +

author-generated picture41 51 20 63 19

Control 43 52 20 30 16

Note Asterisk () indicates significant difference from control group at p lt 05

283A Schmeck et alContemporary Educational Psychology 39 (2014) 275ndash286

no longer statistically significant Regarding the drawing posttestscore including learning time does not change the reported patternof results except that the positive effect of the drawing + author-generated picture group compared with the author-generated picturegroup is no longer statistically significant (from p = 004 to p = 223)

There were neither main effects of learner-generated drawingand author-generated pictures on the mental effort item (drawinggroup M = 455 SD = 025 author-generated picture group M = 459SD = 024 drawing + author-generated picture group M = 444SD = 026 control group M = 481 SD = 024 F lt 1) nor on theperceived difficulty item (drawing group M = 363 SD = 023 author-generated picture group M = 371 SD = 022 drawing + author-generated picture group M = 395 SD = 023 control group M = 393SD = 022 F lt 1)

Taken together the drawing strategy apparently fosters stu-dents to engage in generative activities indicated by their higherlearning outcomes Thus the data provide further evidence for thegenerative drawing effect predicted by Schwamborn et al (2010)In Experiment 2 benefits of the drawing activity however are me-diated by learning time and do not involve higher mental effortAdditionally there was no increased benefit when additional drawingsupport was available in the form of author-generated pictures

344 Is there support for the prognostic drawing effectA second major goal of this study was to determine whether the

prognostic drawing effect could be extended to a new context Meanproportion correct on drawing-accuracy during learning was 60(SD = 04) for the drawing group and 68 (SD = 03) for thedrawing + author generated picture group This difference betweenthe two drawing groups is not significant F(1 79) = 252 p = 116This lack of group differences allowed us to pool the data of bothdrawing groups for subsequent correlation analyses Correlation anal-yses based on the combined data from the two drawing groupsrevealed that the drawing-accuracy score of learner-generated draw-ings correlates significantly with the comprehension posttest scorer = 470 p lt 001 as well as with the drawing posttest score r = 615p lt 001 Additional correlation analyses revealed that the drawing-accuracy score of learner-generated drawings did not correlatesignificantly with the prior knowledge test score r = 095 p = 400the spatial ability test score r = 127 p = 257 the motivation testscore r = 033 p = 769 or the mental effort test score r = 042p = 712 The correlation between the drawing-accuracy score andthe perceived difficulty score was only slightly statistical signifi-cance r = minus218 p = 053 Thus the data provide further evidencefor the prognostic drawing effect consistent with the results ofSchwamborn et al (2010)

In sum results of Experiment 2 are partly consistent with theresults of Experiment 1 in that students learn better from a sciencetext when they are asked to draw illustrations representing the mainideas of the text and the quality of the generated drawings duringlearning correlates positively with studentsrsquo text comprehension (egSchwamborn et al 2010 van Meter 2001 van Meter et al 2006)

4 Discussion

41 Empirical contributions

The present set of experiments makes three empirical contri-butions to the field First this study shows strong and consistentevidence that students who are asked to generate drawings (withsufficient support) during reading a scientific text that describes acausal sequence perform better than students who read withoutdrawing both on a comprehension test (d = 085 in Experiment 1and d = 052 in Experiment 2) and on a drawing test (d = 115 in Ex-periment 1 and d = 187 in Experiment 2) Thus the generativedrawing effect can be extended to a new domain and therefore

corresponds to Shavelson and Townersquos (2002 p 4) recommenda-tion to ldquoreplicate and generalize across studiesrdquo as one of the sixessential scientific principles of educational research

Second this study shows strong and consistent evidence that thequality of drawings that students generate during learning with ascientific text that describes a causal sequence is positively relatedto subsequent performance on tests of learning outcome includ-ing a comprehension test (r = 623 in Experiment 1 and r = 470 inExperiment 2) and a drawing test (r = 620 in Experiment 1 andr = 615 in Experiment 2) Thus the prognostic drawing effect canbe replicated and extended to a new domain consistent with stan-dards for scientific research in education prescribed by Shavelsonand Towne (2002)

Third this study shows that asking learners to draw picturesduring reading a scientific text (ie learner-generated drawing groupin Experiment 2) is more effective than simply providing draw-ings (ie author-generated picture group in Experiment 2) both ona comprehension test (d = 049) and a drawing test (d = 068) Sim-ilarly adding author-generated drawings (ie learner-generatedpictures + author-generated pictures group in Experiment 2) doesnot improve the learning outcomes of students who also draw pic-tures during learning (ie learner-generated pictures group inExperiment 2) either on a comprehension test (d = minus057) or adrawing test (d = minus015) In short the act of drawing during learn-ing (with sufficient support) improves learning beyond the simpleprovision of drawings

42 Theoretical contributions

The results are consistent with the idea that drawing during learn-ing serves as a generative activity (Mayer amp Wittrock 2006Schwamborn et al 2010 van Meter amp Garner 2005 Wittrock 1990)That is the act of drawing encourages learners to engage in gen-erative cognitive processing during learning such as organizing therelevant information into a coherent structure and integrating itwith relevant prior knowledge from long-term memory In thepresent study positive effects of drawing were indicated with a com-prehension and a drawing learning outcome test and therefore arein line with the theoretical assumption derived from the GTDC thatbenefits of drawing can be found if learning outcome tests are usedthat are sensitive to the underlying process of drawing (cf van Meteramp Garner 2005) Additionally in our study the drawing activity wassupported in a way that was intended to help learners carry out theunderlying cognitive processes of drawing (ie selecting organiz-ing and integrating) successfully In this regard results of the presentstudy might supplement the theoretical framework of learner-generated drawing by providing further evidence that benefits ofdrawing defined by van Meter and Garnerrsquos GTDC can diminish ifno instructional support is given to constrain and structure thedrawing activity However a fuller understanding of the underly-ing cognitive processes of drawing and how these processes canbe influenced via drawing support requires more direct measuresof cognitive processing during learning Additionally following theidea that metacognitive processes of monitoring and regulation areautomatically activated by drawing (van Meter amp Garner 2005) afuller understanding of the metacognitive effects of drawings is alsorequired

43 Practical contributions

The present study encourages instructional designers and in-structors to incorporate drawing activities into venues involvinglearning from text which we call the generative drawing effect Oneimportant feature of a successful drawing strategy that is presentin this study and in a previous study by Schwamborn et al (2010)is that the drawing activity was supported by providing a

284 A Schmeck et alContemporary Educational Psychology 39 (2014) 275ndash286

background scene and a legend showing how to represent eachelement to constrain and structure the drawing activity Thus animportant practical implication is that students may need supportin their drawing activity so they do not need to draw from scratch

The present study also suggests a potentially useful diagnostictool to gauge the depth of student learning namely the quality ofthe drawings created by students during learning which we referto as the prognostic drawing effect Incorporating a measure of thequality of a learnerrsquos drawing during learning can be a useful toolin developing remedial instruction to give learners individual supportIt may be important to use materials that explain a cause-and-effect process and give learners drawings of the elements they needto represent the process pictorially Asking learners to simply drawpictures of elements is unlikely to be helpful whereas asking themto generate drawings that show the relations among the elementsin a schematic form is more likely to be helpful

44 Limitations and future directions

Some limitations and future directions of our study should beaddressed As noted in the theoretical contributions subsection wedid not have direct measures of cognitive processing during learn-ing so it is not possible to pinpoint how the drawing activity affectedspecific cognitive processes such as attending to relevant informa-tion organizing it and integrating it with prior knowledge We alsodid not assess metacognitive processing during learning thus it isnot possible to pinpoint how the drawing activity affected specificmetacognitive processes such as monitoring and regulation

Furthermore results of the cognitive load rating scales (in-vested mental effort and perceived task difficulty) are inconsistentWhereas in Experiment 1 an effect on mental effort but not on per-ceived task difficulty showed up (ie students in the drawing grouprated their invested mental effort during learning significantlyhigher) no effects on mental effort and task difficulty were foundin Experiment 2 Additionally in both experiments only a nega-tive correlation of perceived task difficulty with the quality of learner-generated pictures appears but no correlation of mental effort withthe quality Following de Jong (2010) those cognitive load ratingscales might have the disadvantages that they do not give a con-current measure of cognitive load and do not measure an essentialconcept in cognitive load theory namely cognitive overload (p 125)Future studies on learner-generated drawing might also use othercognitive load measures such as physiological measures as moredirect indicators of cognitive load

As noted in the practical contributions subsection we showedthe drawing effects by using a scientific text describing how a cause-and-effect system works that is the causal steps regarding aninfection with influenza and the immune response It might be pos-sible however that for other types of text producing drawings mightharm rather than promote text comprehension Thus to test whetherthe reported drawing effects can be extended future research hasto focus on other types of text such as descriptive texts as well ason other types of relations that can be conveyed with other typesof representations such as compare and contrast relations whichcan be shown in a matrix Additionally studentsrsquo learning out-comes were tested immediately after reading thus future work isneeded to investigate the longer-term effects of generative drawingon learning outcomes

Furthermore we only compared drawing with control groupsthat received no further learning strategy instructions However en-gaging in generative learning activities such as drawing requires aconsiderable amount of time Accordingly results showed that forExperiment 2 the positive effect of the drawing group on text com-prehension compared with the author-generated picture group andto the control group was mediated by learning time To rule out thatthe effects of drawing result only from additional time on task instead

of the generative activity future research should also compare thedrawing strategy with other time demanding generative learningstrategies such as summarization (cf Leopold amp Leutner 2012)

Another point that should be noted is that students in both ex-periments received some kind of multimedia materials in that evenwhen they had to draw and did not see presented pictures they wereat least provided with the basic (visual) elements for their draw-ings which they had to do on the given background which thusalso contained information In other words when students are pre-sented with important elements of the drawings which they canuse to draw themselves they will not have to put as much effortinto summarizing visually what they have just read compared withstudents who have to draw without any instructional help Futurestudies might also compare the drawing group with a summariza-tion group in which students receive a set of verbal key terms thatare similar to the drawing elements and are asked to make a textualsummary

Additionally future research is needed to validate the prognos-tic drawing effect So far we know that the quality of learner-generated pictures is related to studentsrsquo learning outcomes (iethe higher the learning outcome the higher the drawing accuracyand vice versa) and their perceived difficulty (ie the lower the per-ceived difficulty the higher the drawing accuracy and vice versa)and that it is not related to studentsrsquo prior knowledge motivationspatial ability or mental effort However less is known about whatthis might mean That is less is known regarding the causal direc-tion of this relation or the presence of a possible further moderatorvariable Do studentsrsquo efforts to produce accurate drawings lead tobetter comprehension and lower perceived difficulty Or do stu-dents who are more adept in drawing benefit more from the strategyand thus perceive the difficulty of the learning materials as beinglower Both arguments seem convincing

Finally more work is needed to determine the level of supportthat makes the drawing strategy most effective for various kinds oflearners As noted in the empirical contribution adding author-generated drawings (ie learner-generated pictures + author-generated pictures group in Experiment 2) does not improve thelearning outcomes of students who also draw pictures during learn-ing and were supported by a drawing prompt In other words thecombination of two ways of supporting the drawing strategy (iegiving a drawing prompt during reading plus an author-generatedpicture after reading) did not improve studentsrsquo learning out-comes compared with students in the drawing group as well ascompared with students in the control and author-generated pic-tures only groups This result is inconsistent with previous research(eg van Meter 2001 van Meter et al 2006) which found that com-paring own drawings to author-generated pictures normally helpslearning van Meter and colleagues (2001 2006) however provid-ed author-generated pictures plus prompting questions after thedrawing process That is students answered prompting questionsto guide the comparison process between their self-generateddrawing and the author-generated drawing In our study studentswere only instructed to generate a drawing to inspect an author-generated one and to check whether their own drawing incomparison with the author-generated one really represented themain ideas of the text paragraph correctly In other words we didnot guide the process of comparing self-generated drawings withauthor-generated ones As a potential consequence students per-formed the intended comparison process inadequately or even notall and thus did not benefit from it One reason for this inade-quate comparison process might be that students need guidancein doing the comparison process Another reason might be the factthat students do not seriously engage in generating drawings oncethey notice that there are author-generated drawings Thus futureresearch should also use additional guidance to test whether thecombination of different ways of supporting the drawing strategy

285A Schmeck et alContemporary Educational Psychology 39 (2014) 275ndash286

(ie giving a drawing prompt during reading plus an author-generated picture after reading) helps learning as well asobservational measures of the drawing process itself (ie think aloudprotocols) to shed more light on the cognitive processes underly-ing the drawing activities

Overall drawing during learning from text appears to be a po-tentially powerful strategy for improving studentsrsquo learning fromscientific text when certain boundaries and prerequisites are takeninto account

Acknowledgments

This article is based on a research project funded by the GermanResearch Foundation (DFG LE 6459-3 as part of FOR 511) We wouldlike to thank Angela Sandmann for her assistance in developing thelearning materials

References

Ainsworth S Prain V amp Tytler R (2011) Drawing to learn in science Science 3331096ndash1097

Alesandrini K L (1981) Pictorial-verbal and analytic-holistic learning strategies inscience learning Journal of Educational Psychology 73 358ndash368

Alesandrini K L (1984) Pictures and adult learning Instructional Science 13 63ndash77Baron R M amp Kenny D A (1986) The moderator-mediator variable distinction in

social psychological research Conceptual strategic and statistical considerationsJournal of Personality and Social Psychology 51 1173ndash1182

Bruumlnken R Plass J L amp Leutner D (2003) Direct measurement of cognitive loadin multimedia learning Educational Psychologist 38 53ndash61

Carney RN amp Levin JR (2002) Pictorial illustrations still improve studentsrsquo learningfrom text Educational Psychology Review 14 5ndash26

de Jong T (2005) The guided discovery principle in multimedia learning In R EMayer (Ed) The Cambridge handbook of multimedia learning (pp 215ndash228) NewYork Cambridge University Press

de Jong T (2010) Cognitive load theory educational research and instructionaldesign Some food for thought Instructional Science 38 105ndash134

Ekstrom R B French J W amp Harman H H (1976) Manual for kit of factor-referencedcognitive tests Princeton NJ Educational Testing Service

Greene T R (1989) Childrenrsquos understanding of class inclusion hierarchies Therelationship between external representation and task performance Journal ofExperimental Child Psychology 48 62ndash89

Hall V C Bailey J amp Tillman C (1997) Can student-generated illustrations be worthten thousand words Journal of Educational Psychology 89 677ndash681

Houmlffler T N (2010) Spatial ability Its influence on Learning with visualizations ndashA meta-analytic review Educational Psychology Review 22 245ndash269

Houmlffler T N Schmeck A amp Opfermann M (2013) Static and dynamic visualrepresentations Individual differences in processing In G Schraw M TMcCrudden amp D Robinson (Eds) Learning through visual displays (pp 133ndash163)Charlotte NC Information Age Publishing

Kalyuga S Chandler P amp Sweller J (1999) Managing split-attention and redundancyin multimedia instruction Applied Cognitive Psychology 13 351ndash371

Leopold C (2009) Lernstrategien und Textverstehen [Learning strategies and textcomprehension] Muumlnster Waxmann

Leopold C amp Leutner D (2012) Science text comprehension Drawing main ideaselection and summarizing as learning strategies Learning and Instruction 2216ndash26

Lesgold A M DeGood H amp Levin J R (1977) Pictures and young childrenrsquos proselearning A supplementary report Journal of Reading Behavior 9 353ndash360

Lesgold A M Levin J R Shimron J amp Guttman J (1975) Pictures andyoung childrenrsquos learning from oral prose Journal of Educational Psychology 67636ndash642

Leutner D Leopold C amp Sumfleth E (2009) Cognitive load and science textcomprehension Effects of drawing and mentally imagining text contentComputers in Human Behavior 25 284ndash289

Mayer R E (2004) Should there be a three-strikes rule against pure discoverylearning The case for guided methods of instruction The American Psychologist59 14ndash19

Mayer R E (2005) Cognitive theory of multimedia learning In R E Mayer (Ed)The Cambridge handbook of multimedia learning (pp 31ndash48) New York CambridgeUniversity Press

Mayer R E (2009) Multimedia learning (2nd ed) New York NY CambridgeUniversity Press

Mayer R E amp Wittrock M C (2006) Problem solving In P Alexander P Winne ampG Phye (Eds) Handbook of educational psychology (pp 287ndash303) Mahwah NJErlbaum

Paas F (1992) Training strategies for attaining transfer of problem-solving skill instatisticsmdashA cognitive-load approach Journal of Educational Psychology 84429ndash434

Paas F Tuovinen J Tabbers H K amp Van Gerven P W M (2003) Cognitive loadmeasurement as a means to advance cognitive load theory EducationalPsychologist 38 63ndash71

Paivio A (1986) Mental representation A dual coding approach New York OxfordUniversity Press

Pashler H Bain P Bottage B Graesser A Koedinger K McDaniel M et al (2007)Organizing instruction and study to improve student learning Washington DCNational Center for Educational Research

Rasco R W Tennyson R D amp Boutwell R C (1975) Imagery instructions anddrawings in learning prose Journal of Educational Psychology 67 188ndash192

Rheinberg F Vollmeyer R amp Burns B D (2001) FAM Ein fragebogen zurerfassung aktueller motivation in lern- und leistungssituationen [QCM Aquestionnaire to assess current motivation in learning situations] Diagnostica47 57ndash66

Schnotz W (2005) An integrated model of text and picture comprehension In RE Mayer (Ed) The Cambridge handbook of multimedia learning (pp 49ndash70) NewYork Cambridge University Press

Schwamborn A Mayer R E Thillmann H Leopold C amp Leutner D (2010) Drawingas a generative activity and drawing as a prognostic activity Journal of EducationalPsychology 102 872ndash879

Schwamborn A Thillmann H Opfermann M amp Leutner D (2011) Cognitive loadand instructionally supported learning with provided and learner-generatedvisualizations Computers in Human Behavior 27 89ndash93

Shavelson R J amp Towne L (Eds) (2002) Scientific research in education WashingtonDC National Academy Press

Sweller J Ayres P amp Kalyuga S (2011) Cognitive Load Theory New York SpringerTirre W C Manelis L amp Leicht K (1979) The effects of imaginal and verbal strategies

on prose comprehension by adults Journal of Reading Behavior 11 99ndash106van Meter P (2001) Drawing construction as a strategy for learning from text Journal

of Educational Psychology 69 129ndash140van Meter P Aleksic M Schwartz A amp Garner J (2006) Learner-generated drawing

as a strategy for learning from content area text Contemporary EducationalPsychology 31 142ndash166

van Meter P amp Garner J (2005) The promise and practice of learner-generateddrawings Literature review and synthesis Educational Psychology Review 12261ndash312

Van Gog T amp Paas F (2008) Instructional efficiency Revisiting the original constructin educational research Educational Psychologist 43 16ndash26

Vollmeyer R amp Rheinberg F (2000) Does motivation affect learning via persistenceLearning and Instruction 4 293ndash309

Weinstein C E amp Mayer R E (1986) The teaching of learning strategies In M CWittrock (Ed) Handbook of research on teaching (pp 315ndash327) New YorkMacmillan

Wittrock M C (1990) Generative processes of comprehension EducationalPsychologist 24 345ndash376

286 A Schmeck et alContemporary Educational Psychology 39 (2014) 275ndash286

  • Drawing pictures during learning from scientific text testing the generative drawing effect and the prognostic drawing effect
  • Introduction
  • Theoretical framework for the learner-generated drawing strategy
  • Empirical framework for the learner-generated drawing strategy
  • Effectiveness of learner-generated drawings
  • Quality of learner-generated drawings
  • Overview of the experiments
  • Experiment 1
  • Participants and design
  • Materials
  • Procedure
  • Results and discussion
  • Scoring
  • Are the groups equivalent on basic characteristics
  • Is there support for the generative drawing effect
  • Is there support for the prognostic drawing effect
  • Experiment 2
  • Participants and design
  • Materials
  • Procedure
  • Results
  • Scoring
  • Are the groups equivalent on basic characteristics
  • Is there support for the generative drawing effect
  • Is there support for the prognostic drawing effect
  • Discussion
  • Empirical contributions
  • Theoretical contributions
  • Practical contributions
  • Limitations and future directions
  • Acknowledgments
  • References

the comprehension posttest and the drawing posttest Students had20 minutes time for completion and did not have access to thescience text or their drawings Finally students were thanked anddebriefed As students learned at their own rate the whole proce-dure took about 70ndash90 minutes depending on the individual testingtimes

24 Results and discussion

241 ScoringThe dependent variables were studentsrsquo scores on the compre-

hension and drawing posttests studentsrsquo rating on the mental effortand the difficulty scales and the drawing accuracy score indicat-ing the quality of learner-generated drawings produced by studentsin the drawing group during the learning phase

The comprehension test score (pre- and posttest) for each studentwas computed by awarding 1 point for each correct answer and byadding up the points to obtain the total comprehension score (outof a total possible of 25 points) Actual scores ranged from 3 to 24points with a mean of 13 points (SD = 53) Following Schwambornet al (2010) scoring of the drawing test was carried out by count-ing the total number of correct main ideas in each learnerrsquos answeracross the three drawing items The main ideas were drawn out fromboth expert visualizations and a checklist specifying important re-lational features Students could earn a maximum of 19 points onthe drawing test Two student assistants (teacher trainees in biology)scored the quality for each of the three drawings for each studentwith an acceptable inter-rater agreement of GoodmanndashKruskalgamma of 090 Actual scores ranged from 0 to 185 points with amean of 77 points (SD = 47) Total scores of both the comprehen-sion and the drawing test were transferred into z-standardized scoresto make them comparable across studies

The drawing accuracy score (concerning drawing during learn-ing in the drawing group) was computed by using a coding schemeadapted from Schwamborn et al (2010) which was based on expertdrawings and a checklist specifying important relational featuresof the drawings Students could earn a maximum drawing-accuracyscore of 22 points Again the two student assistants scored each ofthe seven learner-generated drawings for each student with an ac-ceptable interrater agreement of GoodmanndashKruskal gamma of 92Both coding schemes were constructed by the first author and abiology teacher Actual scores ranged from 4 to 21 points with amean of 133 points (SD = 50) The total drawing accuracy score wasagain transferred into a z-standardized score

In addition the spatial ability test was scored by tallying thenumber correct out of 10 and the motivation questionnaire wasscored by tallying the nine ratings on both subscales to a total scoreof motivation Finally for comparing performance across the dif-ferent tests the proportion correct on each test was computed bydividing the studentrsquos obtained score by the total possible score

242 Are the groups equivalent on basic characteristicsBefore looking at treatment effects on the dependent variables

we analyzed whether the two groups differed on several control vari-ables A chi-square analysis indicated that there were no significantdifferences regarding gender (p = 562) Separate univariate analy-ses of variance (ANOVAs) revealed that the groups did not differsignificantly on age F lt 1 on spatial ability F lt 1 or on motiva-tion F(1 46) = 360 p = 064 However groups differed significantlyon prior knowledge F(1 46) = 3890 p lt 001 partial eta2 = 46 inthat students in the drawing group scored significantly lower onthe comprehension pretest (M = 10 SD = 15) than students in thecontrol group (M = 34 SD = 12) Thus we included studentsrsquo priorknowledge in the following analyses

243 Is there support for the generative drawing effectMean proportion correct and SDs on the comprehension and

drawing posttests for both groups are presented in Table 2 Repeat-ed measures univariate analyses of variance (ANOVA) with thecomprehension pre- and post-test scores as the within-subject factorsand group (drawing versus control) as the between-subject factorshowed a main effect over time indicating that overall partici-pants reached significant knowledge gains between thecomprehension pretest and the comprehension posttest F(146) = 9897 p lt 001 partial eta2 = 68 An interaction additionallyshowed that these knowledge gains were significantly higher forthe drawing group than for the control group F(1 46) = 4617p lt 001 partial eta2 = 50

For the drawing test a repeated measures ANOVA was not pos-sible since these items were only used in the posttest In this casea univariate analysis of covariance (ANCOVA) predicting the drawingtest score with group (drawing versus control) as the factorial in-dependent variable and prior knowledge as a covariate showed thatthe drawing group scored significantly better than the control groupon the drawing posttest F(1 45) = 1349 p = 001 partial eta2 = 231

Cohenrsquos d favoring the drawing group over the control group was085 for the comprehension posttest and 115 for the drawingposttest all of which are considered large effects Thus there is strongsupport for the generative drawing effect as predicted

Additionally results revealed that the drawing group needed sig-nificantly more learning time (M = 2108 min SD = 424) than thecontrol group (M = 1738 min SD = 333) F(1 46) = 1134 p = 002partial eta2 = 20 Thus to test whether learning time mediates thepositive effect of drawing on text comprehension additional me-diation analyses (Baron amp Kenny 1986) were calculated by includinglearning time as an additional predictor in the aforementioned linearmodel A mediation effect would be detected if in this case effectsof drawing on text comprehension would significantly decreaseResults of the mediation analyses showed that the effect of drawingon both comprehension test scores and drawing test scores was notfully mediated by learning time That is including learning time stillrevealed the interaction between group (drawing versus control)and time (pre- versus post) in that the drawing group had signifi-cantly higher knowledge gains than the control group on thecomprehension test items (p lt 001) Furthermore the drawing groupalso still outperformed the control group on the drawing posttestafter controlling for learning time (p = 009)

Furthermore results revealed that students in the drawing grouprated their invested mental effort during learning significantly higher(M = 504 SD = 112) than students in the control group (M = 396SD = 165) F(1 46) = 705 p = 011 partial eta2 = 13 There was nodifference between the two groups on the perceived difficulty item(drawing group M = 408 SD = 150 control group M = 425SD = 122 F lt 1) Thus consistent with predictions concerning thegenerative drawing effect there is partial support for the idea that

1 The exclusion of studentsrsquo prior knowledge in the reported analyses does notchange the reported pattern of results

Table 2Mean proportion correct on the comprehension test and drawing test for two groupsndash Experiment 1

Group Type of test

n Comprehension test Drawing test

M SD M SD

Drawing 24 61 20 52 27Control 24 44 20 28 11

Note Asterisk () indicates significant difference from control group at p lt 05

280 A Schmeck et alContemporary Educational Psychology 39 (2014) 275ndash286

drawing causes students to engage in more generative processingduring learning

Taken together the results suggest that the drawing strategy en-courages students to engage in generative processing during learningas is indicated by their higher learning outcomes Thus the dataprovide further evidence for the generative drawing effect consis-tent with the results of Schwamborn et al (2010) Additionallyresults indicate that students in the drawing condition seem to investmore mental effort than students in the control group without per-ceiving higher levels of difficulty

244 Is there support for the prognostic drawing effectMean proportion correct on drawing accuracy during learning

was 59 (SD = 23) A correlation analysis revealed that the drawing-accuracy score of learner-generated drawings correlated significantlywith the comprehension posttest score r = 620 p lt 001 and withthe drawing posttest score r = 623 p lt 001 Additional correla-tion analyses revealed that the drawing-accuracy score of learner-generated drawings correlated significantly negatively with theperceived difficulty score r = minus489 p = 015 There were no signif-icant correlations between the drawing accuracy score and eitherthe invested mental effort score r = minus134 p = 533 the prior knowl-edge test score r = minus004 p = 984 the spatial ability test score r = 072p = 739 or the motivation test score r = 086 p = 690 Thus as pre-dicted the data provide further evidence for the prognostic drawingeffect consistent with the results of Schwamborn et al (2010)

In sum the results of Experiment 1 are consistent with the pre-diction that students learn better from a science text when they areasked to draw illustrations representing the main ideas of the textand that the quality of the generated drawings during learning cor-relates positively with studentsrsquo text comprehension (egSchwamborn et al 2010 van Meter 2001 van Meter et al 2006)

However it might be argued that the reported results are dueto the way we supported the strategy use In other words the re-ported positive effect of the learner-generated drawing strategy mightnot be caused by studentsrsquo engagement in generative learning ac-tivities during reading (de Jong 2005 Mayer 2004 2009 Wittrock1990) but rather by the additional pictorial information given in thedrawing prompt Additionally looking at studentsrsquo learning out-comes our results indeed show positive effects of drawing howevermean scores of learning outcomes for the drawing group aremedium-sized Thus it might be argued that the way we sup-ported the strategy use was not fully sufficient In other words thereported positive effect of the learner-generated drawing strategyie the generative drawing effect might be increased by giving stu-dents instructional support in addition to the drawing prompt (vanMeter 2001 van Meter et al 2006 van Meter amp Garner 2005) Toaddress these issues we added two experimental conditions byimplementing author-generated pictures in the design in Experi-ment 2

3 Experiment 2

One possible issue with Experiment 1 is the type of control groupused In Experiment 1 following Schwamborn et al (2010) we useda reading only control group in which the control group learnedwith verbal information only In the drawing group however stu-dents not only learned with verbal information but also with pictorialinformation given by the drawing prompt Based on theories of mul-timedia learning the use of different forms of representations suchas texts and pictures can promote learning in that ldquopeople learnbetter from words and pictures than from words alonerdquo (ie mul-timedia principle Mayer 2009 p 223) because in this caseboth a (verbal) propositional representation as well as a (pictori-al) mental model are built up and are optimally integrated into oneschema that can be stored in long-term memory (Schnotz 2005)

This assumption is also in line with the dual-coding approach statedby Paivio (1986) In this regard it might be argued that the re-ported drawing effect is actually a multimedia effect that is basedon the presentation of text and picture rather than a generativedrawing effect that is based on studentsrsquo active engagement indrawing activities during reading In other words instead of askingpeople to draw pictures representing the main ideas of the textgiving them text and author-generated pictures representing themain ideas of the text might be as good or even better Thus weincluded a condition in Experiment 2 in which we added author-generated pictures to the text

An additional issue with Experiment 1 is whether the reportedgenerative drawing effect can be enhanced by using various formsof supporting the strategy First there is evidence that using adrawing prompt during learning seems to be effective in support-ing the learner-generated drawing strategy by minimizing thecreation of extraneous processing (cf Schwamborn et al 2010 seealso Exp 1) Second research has shown that instructing studentsto compare their own drawing with an author-generated picturemight be also effective in supporting the learner-generated drawingstrategy as self-monitoring processes are enhanced (cf van Meter2001) Up to now however there is no empirical evidence whetherthe combination of both ways to support the drawing strategy hasan additive effect on learning outcomes Thus we included a furthercondition in Experiment 2 in which we combined both forms ofstrategy support

The main purpose of Experiment 2 was to test the generativedrawing and prognostic drawing effects of learner-generated drawingas in Experiment 1 but this time also compared with another controlgroup (ie author-generated pictures) Additionally we were in-terested in testing whether the benefits of the learner-generateddrawing strategy can be increased when we instructionally supportstudents not only with a drawing prompt but also with an author-generated picture after the drawing process In this new treatmentwe instructed students to draw a picture of the text content andthen to compare their own drawing with an expert picture

31 Participants and design

The participants were 168 German eighth graders from highertrack secondary schools The mean age was 138 years (SD = 06)and there were 112 girls and 56 boys The study was based on a2 times 2-between-subjects design with learner-generated drawing (yesno) and author-generated picture (yesno) as factors Forty studentsserved in the drawing group 44 students served in the author-generated picture group 41 students served in the drawing + author-generated picture group and 43 students served in the control group

32 Materials

The materials were identical to those used in Experiment 1 exceptthat we used a shortened version of the comprehension pretest thatconsisted of 19 rather than 25 items (Cronbachrsquos alpha = 70) andslightly extended versions of both the comprehension posttest (28items Cronbachrsquos alpha = 84) and the drawing test (four items witha maximum score of 21 points Cronbachrsquos alpha = 78) The pretestwas shortened because the first experiment showed that the re-spective items were either much too easy or much too difficult andthus unsuitable to differentiate between successful and unsuccess-ful learners thus we deleted these items in the second experimentFurthermore we decided to add some items to the comprehen-sion posttest in the second experiment because during data analysisof the first experiment and after receiving some feedback fromexperts in the domain of biology we recognized that a few itemsassessing transfer ability could be added These transfer itemshowever would have been unsuitable to be included in the pretest

281A Schmeck et alContemporary Educational Psychology 39 (2014) 275ndash286

because they are too difficult to answer without prior training inthe topic Additionally author-generated pictures were used in thenew conditions The author-generated pictures were static func-tional pictures representing the main ideas of each paragraph andconsisted of pictorial elements identical to those provided in thedrawing prompt (as shown in Fig 2) These pictures were con-structed by the first author in cooperation with a biology teacher

The drawing version of the booklet was identical to that usedin Experiment 1 (as shown in Fig 1) The control version of the learn-ing booklet was identical to that used in Experiment 1 The author-generated picture version of the booklet consisted of seven pairsof facing pages with a text paragraph on the left page and a corre-sponding author-generated picture (such as in Fig 2) on the rightpage The drawing + author-generated picture version of the bookletcontained the material from the drawing version consisting of sevenpairs of facing pages with a text paragraph on the left page and atwo-part drawing prompt on the right page In addition attachedto each page there was an additional page that students could foldout after having generated their drawing When unfolding this ad-ditional page a picture of that text paragraph right aside the drawingprompt was provided and there was an additional instruction to

compare the learner-generated drawing with the author-generatedpicture Author-generated pictures were the same as in the author-generated picture version of the booklet

33 Procedure

The procedure was identical to that used in Experiment 1 exceptthat there were two additional groups learning with author-generated pictures Students in the author-generated picturecondition were instructed to read the text and additionally to lookat pictures representing the main ideas of each text paragraph Stu-dents in the drawing + author-generated picture version of thebooklet were instructed to read the text to draw pictures for eachtext paragraph using the drawing prompt representing the mainideas of each text paragraph and finally to compare their picturewith an author-generated picture representing main ideas of eachparagraph correctly Students in all groups learned at their own pacewhereby individual learning time was measured by the instruc-tors in the classrooms Again to ensure that studentsrsquo in both drawinggroups did not feel rushed when students in the non-drawing

Fig 2 Author-generated pictures for the seven paragraphs in the author-generated picture versions of the learning booklet Note Pictures are scaled-down from the orig-inal format

282 A Schmeck et alContemporary Educational Psychology 39 (2014) 275ndash286

group completed the task early groups were tested in separateclassrooms

34 Results

341 ScoringAll tests instruments were scored with the same procedures used

in Experiment 1 Again two student assistants (teacher trainees inbiology) scored each of the drawing test items and each of the sevenlearner-generated drawings for each student with acceptableinterrater agreements (drawing test GoodmanndashKruskal gamma of90 drawing-accuracy GoodmanndashKruskal gamma of 94) Actualscores ranged from 1 to 28 points (M = 153 points SD = 58) for thecomprehension test from zero to 21 points (M = 109 points SD = 53)for the drawing test and from 275 to 215 points (M = 141 pointsSD = 48) for drawing accuracy Again total scores of comprehen-sion drawing and accuracy were transferred into z-standardizedscores

342 Are the groups equivalent on basic characteristicsBefore looking at treatment effects on the dependent variables

we analyzed whether the four experimental groups differed onseveral control variables A chi-square analysis indicated that therewere no significant differences regarding gender (p = 097) Sepa-rate univariate analyses of variance (ANOVAs) revealed that thegroups also did not differ significantly on age F lt 1 on spatial abilityF(3 164) = 120 p = 312 or on motivation F(3 164) = 122 p = 305However groups differed significantly on prior knowledge F(3164) = 104 p = 010 partial eta2 = 07 in that students in the controlgroup scored significantly higher on the comprehension pretest(M = 25 SD = 17) than students in both (p lt 05) the author-generated picture group (M = 17 SD = 16) and the drawing + author-generated picture group (M = 15 SD = 12) the drawing group(M = 20 SD = 14) did not differ significantly from the other groupsThus we included studentsrsquo prior knowledge as a covariate in thefollowing analyses

343 Is there support for the generative drawing effectA major goal in this experiment was to determine whether asking

students to generate drawings to represent science text is a moreeffective learning strategy than asking students to learn with textalone or with text and author-generated pictures In other wordswe wanted to determine whether we could replicate and extendthe learner-generated drawing effect Additionally we were inter-ested in whether giving students an author-generated picture afterdrawing can increase the benefits of the learning strategy Mean pro-portion correct and standard deviations on the comprehension anddrawing tests for the four groups are presented in Table 3

The left portion of Table 3 summarizes the mean proportioncorrect on the comprehension test A two-factorial analysis ofcovariance (ANCOVA) predicting learning outcomes (comprehen-sion posttest score) with learner-generated drawing (yesno) and

author-generated picture (yesno) as the factorial independent vari-ables and prior knowledge as a covariate showed a significantpositive main effect of learner-generated drawing F(1 163) = 398p = 048 partial eta2 = 02 a significant interaction effect F(1163) = 626 p = 013 partial eta2 = 04 but no main effect of author-generated pictures F lt 1 In addition multiple pairwise comparisons(with p lt 05) showed that the drawing group performed signifi-cantly better than each of the three other groups which did not differsignificantly from each other Cohenrsquos d favoring the drawing groupover the author-generated picture group was 49 over the learner-generated + author-generated picture group was 57 and over thecontrol group was 52

The right portion of Table 3 summarizes the mean proportioncorrect on the drawing posttest Again a two-factorial analysis ofcovariance (ANCOVA) predicting learning outcome (drawing testscore) with learner-generated drawing (yesno) and author-generatedpicture (yesno) as the factorial independent variables and priorknowledge as a covariate showed a significant positive main effectof learner-generated drawing F(1 163) = 6260 p lt 001 partialeta2 = 28 a significant positive main effect of author-generated pic-tures F(1 163) = 1104 p = 001 partial eta2 = 06 and a significantinteraction effect F(1 163) = 1658 p lt 001 partial eta2 = 09 In ad-dition multiple pairwise comparisons (with p lt 05) showed thatboth the drawing group and the drawing + author-generated picturegroup performed significantly better than the author-generatedpicture group (d = 68 d = 59) and the control group (d = 187d = 188) In turn the author-generated picture group performed sig-nificantly better than the control group (d = 95) The drawing groupand the drawing + author-generated picture group did not differ sig-nificantly from each other (d = 15)2 Overall these results areconsistent with Experiment 1 and provide additional support forthe generative drawing effect

In accordance with Experiment 1 we were interested in whetherdifferences in learning time among the experimental groups mediatethe positive effect of drawing on text comprehension First an ANOVApredicting learning time with learner-generated drawing (yesno)and author-generated picture (yesno) as the factorial indepen-dent variables showed a significant main effect of learner-generateddrawing F(1 164) = 39226 p lt 001 partial eta2 = 71 a significantmain effect of author-generated picture F(1 164) = 1685 p lt 001partial eta2 = 09 and a significant interaction effect F(1 164) = 490p = 028 partial eta2 = 03 Linear contrasts (with p lt 05) revealedthat the drawing group (M = 1938 min SD = 380) and thedrawing + author-generated picture group (M = 2340 min SD = 551)needed significantly more learning time than the author-generatedpicture group (M = 960 min SD = 397) and the control group(M = 834 min SD = 251) Thus to test whether learning time me-diates the positive effect of drawing on text comprehensionadditional mediation analyses (Baron amp Kenny 1986) were calcu-lated by including learning time as an additional predictor in theaforementioned linear model Results of the mediation analysesshowed that the effects of drawing on the comprehension posttestand the drawing posttest scores (see multiple pairwise compari-sons) are mediated by learning time to some extent That is includinglearning time in the linear model for predicting comprehension testscores still revealed a positive effect of the drawing group com-pared with the drawing + author-generated group on thecomprehension test (p = 012) However including learning time inthe linear model for predicting comprehension posttest scoresreduced the positive effect of the drawing group compared with theauthor-generated picture group (from p = 034 to p = 281) as wellas compared with the control group (from p = 002 to p = 087) being

2 The exclusion of studentsrsquo prior knowledge in the reported analyses does notchange the overall pattern of results

Table 3Mean proportion correct on the comprehension test and drawing test for the fourgroups ndash Experiment 2

Group Type of test

n Comprehension test Drawing test

M SD M SD

Learner-generated drawing 40 63 22 66 22Author-generated picture 44 53 19 50 25Learner-generated drawing +

author-generated picture41 51 20 63 19

Control 43 52 20 30 16

Note Asterisk () indicates significant difference from control group at p lt 05

283A Schmeck et alContemporary Educational Psychology 39 (2014) 275ndash286

no longer statistically significant Regarding the drawing posttestscore including learning time does not change the reported patternof results except that the positive effect of the drawing + author-generated picture group compared with the author-generated picturegroup is no longer statistically significant (from p = 004 to p = 223)

There were neither main effects of learner-generated drawingand author-generated pictures on the mental effort item (drawinggroup M = 455 SD = 025 author-generated picture group M = 459SD = 024 drawing + author-generated picture group M = 444SD = 026 control group M = 481 SD = 024 F lt 1) nor on theperceived difficulty item (drawing group M = 363 SD = 023 author-generated picture group M = 371 SD = 022 drawing + author-generated picture group M = 395 SD = 023 control group M = 393SD = 022 F lt 1)

Taken together the drawing strategy apparently fosters stu-dents to engage in generative activities indicated by their higherlearning outcomes Thus the data provide further evidence for thegenerative drawing effect predicted by Schwamborn et al (2010)In Experiment 2 benefits of the drawing activity however are me-diated by learning time and do not involve higher mental effortAdditionally there was no increased benefit when additional drawingsupport was available in the form of author-generated pictures

344 Is there support for the prognostic drawing effectA second major goal of this study was to determine whether the

prognostic drawing effect could be extended to a new context Meanproportion correct on drawing-accuracy during learning was 60(SD = 04) for the drawing group and 68 (SD = 03) for thedrawing + author generated picture group This difference betweenthe two drawing groups is not significant F(1 79) = 252 p = 116This lack of group differences allowed us to pool the data of bothdrawing groups for subsequent correlation analyses Correlation anal-yses based on the combined data from the two drawing groupsrevealed that the drawing-accuracy score of learner-generated draw-ings correlates significantly with the comprehension posttest scorer = 470 p lt 001 as well as with the drawing posttest score r = 615p lt 001 Additional correlation analyses revealed that the drawing-accuracy score of learner-generated drawings did not correlatesignificantly with the prior knowledge test score r = 095 p = 400the spatial ability test score r = 127 p = 257 the motivation testscore r = 033 p = 769 or the mental effort test score r = 042p = 712 The correlation between the drawing-accuracy score andthe perceived difficulty score was only slightly statistical signifi-cance r = minus218 p = 053 Thus the data provide further evidencefor the prognostic drawing effect consistent with the results ofSchwamborn et al (2010)

In sum results of Experiment 2 are partly consistent with theresults of Experiment 1 in that students learn better from a sciencetext when they are asked to draw illustrations representing the mainideas of the text and the quality of the generated drawings duringlearning correlates positively with studentsrsquo text comprehension (egSchwamborn et al 2010 van Meter 2001 van Meter et al 2006)

4 Discussion

41 Empirical contributions

The present set of experiments makes three empirical contri-butions to the field First this study shows strong and consistentevidence that students who are asked to generate drawings (withsufficient support) during reading a scientific text that describes acausal sequence perform better than students who read withoutdrawing both on a comprehension test (d = 085 in Experiment 1and d = 052 in Experiment 2) and on a drawing test (d = 115 in Ex-periment 1 and d = 187 in Experiment 2) Thus the generativedrawing effect can be extended to a new domain and therefore

corresponds to Shavelson and Townersquos (2002 p 4) recommenda-tion to ldquoreplicate and generalize across studiesrdquo as one of the sixessential scientific principles of educational research

Second this study shows strong and consistent evidence that thequality of drawings that students generate during learning with ascientific text that describes a causal sequence is positively relatedto subsequent performance on tests of learning outcome includ-ing a comprehension test (r = 623 in Experiment 1 and r = 470 inExperiment 2) and a drawing test (r = 620 in Experiment 1 andr = 615 in Experiment 2) Thus the prognostic drawing effect canbe replicated and extended to a new domain consistent with stan-dards for scientific research in education prescribed by Shavelsonand Towne (2002)

Third this study shows that asking learners to draw picturesduring reading a scientific text (ie learner-generated drawing groupin Experiment 2) is more effective than simply providing draw-ings (ie author-generated picture group in Experiment 2) both ona comprehension test (d = 049) and a drawing test (d = 068) Sim-ilarly adding author-generated drawings (ie learner-generatedpictures + author-generated pictures group in Experiment 2) doesnot improve the learning outcomes of students who also draw pic-tures during learning (ie learner-generated pictures group inExperiment 2) either on a comprehension test (d = minus057) or adrawing test (d = minus015) In short the act of drawing during learn-ing (with sufficient support) improves learning beyond the simpleprovision of drawings

42 Theoretical contributions

The results are consistent with the idea that drawing during learn-ing serves as a generative activity (Mayer amp Wittrock 2006Schwamborn et al 2010 van Meter amp Garner 2005 Wittrock 1990)That is the act of drawing encourages learners to engage in gen-erative cognitive processing during learning such as organizing therelevant information into a coherent structure and integrating itwith relevant prior knowledge from long-term memory In thepresent study positive effects of drawing were indicated with a com-prehension and a drawing learning outcome test and therefore arein line with the theoretical assumption derived from the GTDC thatbenefits of drawing can be found if learning outcome tests are usedthat are sensitive to the underlying process of drawing (cf van Meteramp Garner 2005) Additionally in our study the drawing activity wassupported in a way that was intended to help learners carry out theunderlying cognitive processes of drawing (ie selecting organiz-ing and integrating) successfully In this regard results of the presentstudy might supplement the theoretical framework of learner-generated drawing by providing further evidence that benefits ofdrawing defined by van Meter and Garnerrsquos GTDC can diminish ifno instructional support is given to constrain and structure thedrawing activity However a fuller understanding of the underly-ing cognitive processes of drawing and how these processes canbe influenced via drawing support requires more direct measuresof cognitive processing during learning Additionally following theidea that metacognitive processes of monitoring and regulation areautomatically activated by drawing (van Meter amp Garner 2005) afuller understanding of the metacognitive effects of drawings is alsorequired

43 Practical contributions

The present study encourages instructional designers and in-structors to incorporate drawing activities into venues involvinglearning from text which we call the generative drawing effect Oneimportant feature of a successful drawing strategy that is presentin this study and in a previous study by Schwamborn et al (2010)is that the drawing activity was supported by providing a

284 A Schmeck et alContemporary Educational Psychology 39 (2014) 275ndash286

background scene and a legend showing how to represent eachelement to constrain and structure the drawing activity Thus animportant practical implication is that students may need supportin their drawing activity so they do not need to draw from scratch

The present study also suggests a potentially useful diagnostictool to gauge the depth of student learning namely the quality ofthe drawings created by students during learning which we referto as the prognostic drawing effect Incorporating a measure of thequality of a learnerrsquos drawing during learning can be a useful toolin developing remedial instruction to give learners individual supportIt may be important to use materials that explain a cause-and-effect process and give learners drawings of the elements they needto represent the process pictorially Asking learners to simply drawpictures of elements is unlikely to be helpful whereas asking themto generate drawings that show the relations among the elementsin a schematic form is more likely to be helpful

44 Limitations and future directions

Some limitations and future directions of our study should beaddressed As noted in the theoretical contributions subsection wedid not have direct measures of cognitive processing during learn-ing so it is not possible to pinpoint how the drawing activity affectedspecific cognitive processes such as attending to relevant informa-tion organizing it and integrating it with prior knowledge We alsodid not assess metacognitive processing during learning thus it isnot possible to pinpoint how the drawing activity affected specificmetacognitive processes such as monitoring and regulation

Furthermore results of the cognitive load rating scales (in-vested mental effort and perceived task difficulty) are inconsistentWhereas in Experiment 1 an effect on mental effort but not on per-ceived task difficulty showed up (ie students in the drawing grouprated their invested mental effort during learning significantlyhigher) no effects on mental effort and task difficulty were foundin Experiment 2 Additionally in both experiments only a nega-tive correlation of perceived task difficulty with the quality of learner-generated pictures appears but no correlation of mental effort withthe quality Following de Jong (2010) those cognitive load ratingscales might have the disadvantages that they do not give a con-current measure of cognitive load and do not measure an essentialconcept in cognitive load theory namely cognitive overload (p 125)Future studies on learner-generated drawing might also use othercognitive load measures such as physiological measures as moredirect indicators of cognitive load

As noted in the practical contributions subsection we showedthe drawing effects by using a scientific text describing how a cause-and-effect system works that is the causal steps regarding aninfection with influenza and the immune response It might be pos-sible however that for other types of text producing drawings mightharm rather than promote text comprehension Thus to test whetherthe reported drawing effects can be extended future research hasto focus on other types of text such as descriptive texts as well ason other types of relations that can be conveyed with other typesof representations such as compare and contrast relations whichcan be shown in a matrix Additionally studentsrsquo learning out-comes were tested immediately after reading thus future work isneeded to investigate the longer-term effects of generative drawingon learning outcomes

Furthermore we only compared drawing with control groupsthat received no further learning strategy instructions However en-gaging in generative learning activities such as drawing requires aconsiderable amount of time Accordingly results showed that forExperiment 2 the positive effect of the drawing group on text com-prehension compared with the author-generated picture group andto the control group was mediated by learning time To rule out thatthe effects of drawing result only from additional time on task instead

of the generative activity future research should also compare thedrawing strategy with other time demanding generative learningstrategies such as summarization (cf Leopold amp Leutner 2012)

Another point that should be noted is that students in both ex-periments received some kind of multimedia materials in that evenwhen they had to draw and did not see presented pictures they wereat least provided with the basic (visual) elements for their draw-ings which they had to do on the given background which thusalso contained information In other words when students are pre-sented with important elements of the drawings which they canuse to draw themselves they will not have to put as much effortinto summarizing visually what they have just read compared withstudents who have to draw without any instructional help Futurestudies might also compare the drawing group with a summariza-tion group in which students receive a set of verbal key terms thatare similar to the drawing elements and are asked to make a textualsummary

Additionally future research is needed to validate the prognos-tic drawing effect So far we know that the quality of learner-generated pictures is related to studentsrsquo learning outcomes (iethe higher the learning outcome the higher the drawing accuracyand vice versa) and their perceived difficulty (ie the lower the per-ceived difficulty the higher the drawing accuracy and vice versa)and that it is not related to studentsrsquo prior knowledge motivationspatial ability or mental effort However less is known about whatthis might mean That is less is known regarding the causal direc-tion of this relation or the presence of a possible further moderatorvariable Do studentsrsquo efforts to produce accurate drawings lead tobetter comprehension and lower perceived difficulty Or do stu-dents who are more adept in drawing benefit more from the strategyand thus perceive the difficulty of the learning materials as beinglower Both arguments seem convincing

Finally more work is needed to determine the level of supportthat makes the drawing strategy most effective for various kinds oflearners As noted in the empirical contribution adding author-generated drawings (ie learner-generated pictures + author-generated pictures group in Experiment 2) does not improve thelearning outcomes of students who also draw pictures during learn-ing and were supported by a drawing prompt In other words thecombination of two ways of supporting the drawing strategy (iegiving a drawing prompt during reading plus an author-generatedpicture after reading) did not improve studentsrsquo learning out-comes compared with students in the drawing group as well ascompared with students in the control and author-generated pic-tures only groups This result is inconsistent with previous research(eg van Meter 2001 van Meter et al 2006) which found that com-paring own drawings to author-generated pictures normally helpslearning van Meter and colleagues (2001 2006) however provid-ed author-generated pictures plus prompting questions after thedrawing process That is students answered prompting questionsto guide the comparison process between their self-generateddrawing and the author-generated drawing In our study studentswere only instructed to generate a drawing to inspect an author-generated one and to check whether their own drawing incomparison with the author-generated one really represented themain ideas of the text paragraph correctly In other words we didnot guide the process of comparing self-generated drawings withauthor-generated ones As a potential consequence students per-formed the intended comparison process inadequately or even notall and thus did not benefit from it One reason for this inade-quate comparison process might be that students need guidancein doing the comparison process Another reason might be the factthat students do not seriously engage in generating drawings oncethey notice that there are author-generated drawings Thus futureresearch should also use additional guidance to test whether thecombination of different ways of supporting the drawing strategy

285A Schmeck et alContemporary Educational Psychology 39 (2014) 275ndash286

(ie giving a drawing prompt during reading plus an author-generated picture after reading) helps learning as well asobservational measures of the drawing process itself (ie think aloudprotocols) to shed more light on the cognitive processes underly-ing the drawing activities

Overall drawing during learning from text appears to be a po-tentially powerful strategy for improving studentsrsquo learning fromscientific text when certain boundaries and prerequisites are takeninto account

Acknowledgments

This article is based on a research project funded by the GermanResearch Foundation (DFG LE 6459-3 as part of FOR 511) We wouldlike to thank Angela Sandmann for her assistance in developing thelearning materials

References

Ainsworth S Prain V amp Tytler R (2011) Drawing to learn in science Science 3331096ndash1097

Alesandrini K L (1981) Pictorial-verbal and analytic-holistic learning strategies inscience learning Journal of Educational Psychology 73 358ndash368

Alesandrini K L (1984) Pictures and adult learning Instructional Science 13 63ndash77Baron R M amp Kenny D A (1986) The moderator-mediator variable distinction in

social psychological research Conceptual strategic and statistical considerationsJournal of Personality and Social Psychology 51 1173ndash1182

Bruumlnken R Plass J L amp Leutner D (2003) Direct measurement of cognitive loadin multimedia learning Educational Psychologist 38 53ndash61

Carney RN amp Levin JR (2002) Pictorial illustrations still improve studentsrsquo learningfrom text Educational Psychology Review 14 5ndash26

de Jong T (2005) The guided discovery principle in multimedia learning In R EMayer (Ed) The Cambridge handbook of multimedia learning (pp 215ndash228) NewYork Cambridge University Press

de Jong T (2010) Cognitive load theory educational research and instructionaldesign Some food for thought Instructional Science 38 105ndash134

Ekstrom R B French J W amp Harman H H (1976) Manual for kit of factor-referencedcognitive tests Princeton NJ Educational Testing Service

Greene T R (1989) Childrenrsquos understanding of class inclusion hierarchies Therelationship between external representation and task performance Journal ofExperimental Child Psychology 48 62ndash89

Hall V C Bailey J amp Tillman C (1997) Can student-generated illustrations be worthten thousand words Journal of Educational Psychology 89 677ndash681

Houmlffler T N (2010) Spatial ability Its influence on Learning with visualizations ndashA meta-analytic review Educational Psychology Review 22 245ndash269

Houmlffler T N Schmeck A amp Opfermann M (2013) Static and dynamic visualrepresentations Individual differences in processing In G Schraw M TMcCrudden amp D Robinson (Eds) Learning through visual displays (pp 133ndash163)Charlotte NC Information Age Publishing

Kalyuga S Chandler P amp Sweller J (1999) Managing split-attention and redundancyin multimedia instruction Applied Cognitive Psychology 13 351ndash371

Leopold C (2009) Lernstrategien und Textverstehen [Learning strategies and textcomprehension] Muumlnster Waxmann

Leopold C amp Leutner D (2012) Science text comprehension Drawing main ideaselection and summarizing as learning strategies Learning and Instruction 2216ndash26

Lesgold A M DeGood H amp Levin J R (1977) Pictures and young childrenrsquos proselearning A supplementary report Journal of Reading Behavior 9 353ndash360

Lesgold A M Levin J R Shimron J amp Guttman J (1975) Pictures andyoung childrenrsquos learning from oral prose Journal of Educational Psychology 67636ndash642

Leutner D Leopold C amp Sumfleth E (2009) Cognitive load and science textcomprehension Effects of drawing and mentally imagining text contentComputers in Human Behavior 25 284ndash289

Mayer R E (2004) Should there be a three-strikes rule against pure discoverylearning The case for guided methods of instruction The American Psychologist59 14ndash19

Mayer R E (2005) Cognitive theory of multimedia learning In R E Mayer (Ed)The Cambridge handbook of multimedia learning (pp 31ndash48) New York CambridgeUniversity Press

Mayer R E (2009) Multimedia learning (2nd ed) New York NY CambridgeUniversity Press

Mayer R E amp Wittrock M C (2006) Problem solving In P Alexander P Winne ampG Phye (Eds) Handbook of educational psychology (pp 287ndash303) Mahwah NJErlbaum

Paas F (1992) Training strategies for attaining transfer of problem-solving skill instatisticsmdashA cognitive-load approach Journal of Educational Psychology 84429ndash434

Paas F Tuovinen J Tabbers H K amp Van Gerven P W M (2003) Cognitive loadmeasurement as a means to advance cognitive load theory EducationalPsychologist 38 63ndash71

Paivio A (1986) Mental representation A dual coding approach New York OxfordUniversity Press

Pashler H Bain P Bottage B Graesser A Koedinger K McDaniel M et al (2007)Organizing instruction and study to improve student learning Washington DCNational Center for Educational Research

Rasco R W Tennyson R D amp Boutwell R C (1975) Imagery instructions anddrawings in learning prose Journal of Educational Psychology 67 188ndash192

Rheinberg F Vollmeyer R amp Burns B D (2001) FAM Ein fragebogen zurerfassung aktueller motivation in lern- und leistungssituationen [QCM Aquestionnaire to assess current motivation in learning situations] Diagnostica47 57ndash66

Schnotz W (2005) An integrated model of text and picture comprehension In RE Mayer (Ed) The Cambridge handbook of multimedia learning (pp 49ndash70) NewYork Cambridge University Press

Schwamborn A Mayer R E Thillmann H Leopold C amp Leutner D (2010) Drawingas a generative activity and drawing as a prognostic activity Journal of EducationalPsychology 102 872ndash879

Schwamborn A Thillmann H Opfermann M amp Leutner D (2011) Cognitive loadand instructionally supported learning with provided and learner-generatedvisualizations Computers in Human Behavior 27 89ndash93

Shavelson R J amp Towne L (Eds) (2002) Scientific research in education WashingtonDC National Academy Press

Sweller J Ayres P amp Kalyuga S (2011) Cognitive Load Theory New York SpringerTirre W C Manelis L amp Leicht K (1979) The effects of imaginal and verbal strategies

on prose comprehension by adults Journal of Reading Behavior 11 99ndash106van Meter P (2001) Drawing construction as a strategy for learning from text Journal

of Educational Psychology 69 129ndash140van Meter P Aleksic M Schwartz A amp Garner J (2006) Learner-generated drawing

as a strategy for learning from content area text Contemporary EducationalPsychology 31 142ndash166

van Meter P amp Garner J (2005) The promise and practice of learner-generateddrawings Literature review and synthesis Educational Psychology Review 12261ndash312

Van Gog T amp Paas F (2008) Instructional efficiency Revisiting the original constructin educational research Educational Psychologist 43 16ndash26

Vollmeyer R amp Rheinberg F (2000) Does motivation affect learning via persistenceLearning and Instruction 4 293ndash309

Weinstein C E amp Mayer R E (1986) The teaching of learning strategies In M CWittrock (Ed) Handbook of research on teaching (pp 315ndash327) New YorkMacmillan

Wittrock M C (1990) Generative processes of comprehension EducationalPsychologist 24 345ndash376

286 A Schmeck et alContemporary Educational Psychology 39 (2014) 275ndash286

  • Drawing pictures during learning from scientific text testing the generative drawing effect and the prognostic drawing effect
  • Introduction
  • Theoretical framework for the learner-generated drawing strategy
  • Empirical framework for the learner-generated drawing strategy
  • Effectiveness of learner-generated drawings
  • Quality of learner-generated drawings
  • Overview of the experiments
  • Experiment 1
  • Participants and design
  • Materials
  • Procedure
  • Results and discussion
  • Scoring
  • Are the groups equivalent on basic characteristics
  • Is there support for the generative drawing effect
  • Is there support for the prognostic drawing effect
  • Experiment 2
  • Participants and design
  • Materials
  • Procedure
  • Results
  • Scoring
  • Are the groups equivalent on basic characteristics
  • Is there support for the generative drawing effect
  • Is there support for the prognostic drawing effect
  • Discussion
  • Empirical contributions
  • Theoretical contributions
  • Practical contributions
  • Limitations and future directions
  • Acknowledgments
  • References

drawing causes students to engage in more generative processingduring learning

Taken together the results suggest that the drawing strategy en-courages students to engage in generative processing during learningas is indicated by their higher learning outcomes Thus the dataprovide further evidence for the generative drawing effect consis-tent with the results of Schwamborn et al (2010) Additionallyresults indicate that students in the drawing condition seem to investmore mental effort than students in the control group without per-ceiving higher levels of difficulty

244 Is there support for the prognostic drawing effectMean proportion correct on drawing accuracy during learning

was 59 (SD = 23) A correlation analysis revealed that the drawing-accuracy score of learner-generated drawings correlated significantlywith the comprehension posttest score r = 620 p lt 001 and withthe drawing posttest score r = 623 p lt 001 Additional correla-tion analyses revealed that the drawing-accuracy score of learner-generated drawings correlated significantly negatively with theperceived difficulty score r = minus489 p = 015 There were no signif-icant correlations between the drawing accuracy score and eitherthe invested mental effort score r = minus134 p = 533 the prior knowl-edge test score r = minus004 p = 984 the spatial ability test score r = 072p = 739 or the motivation test score r = 086 p = 690 Thus as pre-dicted the data provide further evidence for the prognostic drawingeffect consistent with the results of Schwamborn et al (2010)

In sum the results of Experiment 1 are consistent with the pre-diction that students learn better from a science text when they areasked to draw illustrations representing the main ideas of the textand that the quality of the generated drawings during learning cor-relates positively with studentsrsquo text comprehension (egSchwamborn et al 2010 van Meter 2001 van Meter et al 2006)

However it might be argued that the reported results are dueto the way we supported the strategy use In other words the re-ported positive effect of the learner-generated drawing strategy mightnot be caused by studentsrsquo engagement in generative learning ac-tivities during reading (de Jong 2005 Mayer 2004 2009 Wittrock1990) but rather by the additional pictorial information given in thedrawing prompt Additionally looking at studentsrsquo learning out-comes our results indeed show positive effects of drawing howevermean scores of learning outcomes for the drawing group aremedium-sized Thus it might be argued that the way we sup-ported the strategy use was not fully sufficient In other words thereported positive effect of the learner-generated drawing strategyie the generative drawing effect might be increased by giving stu-dents instructional support in addition to the drawing prompt (vanMeter 2001 van Meter et al 2006 van Meter amp Garner 2005) Toaddress these issues we added two experimental conditions byimplementing author-generated pictures in the design in Experi-ment 2

3 Experiment 2

One possible issue with Experiment 1 is the type of control groupused In Experiment 1 following Schwamborn et al (2010) we useda reading only control group in which the control group learnedwith verbal information only In the drawing group however stu-dents not only learned with verbal information but also with pictorialinformation given by the drawing prompt Based on theories of mul-timedia learning the use of different forms of representations suchas texts and pictures can promote learning in that ldquopeople learnbetter from words and pictures than from words alonerdquo (ie mul-timedia principle Mayer 2009 p 223) because in this caseboth a (verbal) propositional representation as well as a (pictori-al) mental model are built up and are optimally integrated into oneschema that can be stored in long-term memory (Schnotz 2005)

This assumption is also in line with the dual-coding approach statedby Paivio (1986) In this regard it might be argued that the re-ported drawing effect is actually a multimedia effect that is basedon the presentation of text and picture rather than a generativedrawing effect that is based on studentsrsquo active engagement indrawing activities during reading In other words instead of askingpeople to draw pictures representing the main ideas of the textgiving them text and author-generated pictures representing themain ideas of the text might be as good or even better Thus weincluded a condition in Experiment 2 in which we added author-generated pictures to the text

An additional issue with Experiment 1 is whether the reportedgenerative drawing effect can be enhanced by using various formsof supporting the strategy First there is evidence that using adrawing prompt during learning seems to be effective in support-ing the learner-generated drawing strategy by minimizing thecreation of extraneous processing (cf Schwamborn et al 2010 seealso Exp 1) Second research has shown that instructing studentsto compare their own drawing with an author-generated picturemight be also effective in supporting the learner-generated drawingstrategy as self-monitoring processes are enhanced (cf van Meter2001) Up to now however there is no empirical evidence whetherthe combination of both ways to support the drawing strategy hasan additive effect on learning outcomes Thus we included a furthercondition in Experiment 2 in which we combined both forms ofstrategy support

The main purpose of Experiment 2 was to test the generativedrawing and prognostic drawing effects of learner-generated drawingas in Experiment 1 but this time also compared with another controlgroup (ie author-generated pictures) Additionally we were in-terested in testing whether the benefits of the learner-generateddrawing strategy can be increased when we instructionally supportstudents not only with a drawing prompt but also with an author-generated picture after the drawing process In this new treatmentwe instructed students to draw a picture of the text content andthen to compare their own drawing with an expert picture

31 Participants and design

The participants were 168 German eighth graders from highertrack secondary schools The mean age was 138 years (SD = 06)and there were 112 girls and 56 boys The study was based on a2 times 2-between-subjects design with learner-generated drawing (yesno) and author-generated picture (yesno) as factors Forty studentsserved in the drawing group 44 students served in the author-generated picture group 41 students served in the drawing + author-generated picture group and 43 students served in the control group

32 Materials

The materials were identical to those used in Experiment 1 exceptthat we used a shortened version of the comprehension pretest thatconsisted of 19 rather than 25 items (Cronbachrsquos alpha = 70) andslightly extended versions of both the comprehension posttest (28items Cronbachrsquos alpha = 84) and the drawing test (four items witha maximum score of 21 points Cronbachrsquos alpha = 78) The pretestwas shortened because the first experiment showed that the re-spective items were either much too easy or much too difficult andthus unsuitable to differentiate between successful and unsuccess-ful learners thus we deleted these items in the second experimentFurthermore we decided to add some items to the comprehen-sion posttest in the second experiment because during data analysisof the first experiment and after receiving some feedback fromexperts in the domain of biology we recognized that a few itemsassessing transfer ability could be added These transfer itemshowever would have been unsuitable to be included in the pretest

281A Schmeck et alContemporary Educational Psychology 39 (2014) 275ndash286

because they are too difficult to answer without prior training inthe topic Additionally author-generated pictures were used in thenew conditions The author-generated pictures were static func-tional pictures representing the main ideas of each paragraph andconsisted of pictorial elements identical to those provided in thedrawing prompt (as shown in Fig 2) These pictures were con-structed by the first author in cooperation with a biology teacher

The drawing version of the booklet was identical to that usedin Experiment 1 (as shown in Fig 1) The control version of the learn-ing booklet was identical to that used in Experiment 1 The author-generated picture version of the booklet consisted of seven pairsof facing pages with a text paragraph on the left page and a corre-sponding author-generated picture (such as in Fig 2) on the rightpage The drawing + author-generated picture version of the bookletcontained the material from the drawing version consisting of sevenpairs of facing pages with a text paragraph on the left page and atwo-part drawing prompt on the right page In addition attachedto each page there was an additional page that students could foldout after having generated their drawing When unfolding this ad-ditional page a picture of that text paragraph right aside the drawingprompt was provided and there was an additional instruction to

compare the learner-generated drawing with the author-generatedpicture Author-generated pictures were the same as in the author-generated picture version of the booklet

33 Procedure

The procedure was identical to that used in Experiment 1 exceptthat there were two additional groups learning with author-generated pictures Students in the author-generated picturecondition were instructed to read the text and additionally to lookat pictures representing the main ideas of each text paragraph Stu-dents in the drawing + author-generated picture version of thebooklet were instructed to read the text to draw pictures for eachtext paragraph using the drawing prompt representing the mainideas of each text paragraph and finally to compare their picturewith an author-generated picture representing main ideas of eachparagraph correctly Students in all groups learned at their own pacewhereby individual learning time was measured by the instruc-tors in the classrooms Again to ensure that studentsrsquo in both drawinggroups did not feel rushed when students in the non-drawing

Fig 2 Author-generated pictures for the seven paragraphs in the author-generated picture versions of the learning booklet Note Pictures are scaled-down from the orig-inal format

282 A Schmeck et alContemporary Educational Psychology 39 (2014) 275ndash286

group completed the task early groups were tested in separateclassrooms

34 Results

341 ScoringAll tests instruments were scored with the same procedures used

in Experiment 1 Again two student assistants (teacher trainees inbiology) scored each of the drawing test items and each of the sevenlearner-generated drawings for each student with acceptableinterrater agreements (drawing test GoodmanndashKruskal gamma of90 drawing-accuracy GoodmanndashKruskal gamma of 94) Actualscores ranged from 1 to 28 points (M = 153 points SD = 58) for thecomprehension test from zero to 21 points (M = 109 points SD = 53)for the drawing test and from 275 to 215 points (M = 141 pointsSD = 48) for drawing accuracy Again total scores of comprehen-sion drawing and accuracy were transferred into z-standardizedscores

342 Are the groups equivalent on basic characteristicsBefore looking at treatment effects on the dependent variables

we analyzed whether the four experimental groups differed onseveral control variables A chi-square analysis indicated that therewere no significant differences regarding gender (p = 097) Sepa-rate univariate analyses of variance (ANOVAs) revealed that thegroups also did not differ significantly on age F lt 1 on spatial abilityF(3 164) = 120 p = 312 or on motivation F(3 164) = 122 p = 305However groups differed significantly on prior knowledge F(3164) = 104 p = 010 partial eta2 = 07 in that students in the controlgroup scored significantly higher on the comprehension pretest(M = 25 SD = 17) than students in both (p lt 05) the author-generated picture group (M = 17 SD = 16) and the drawing + author-generated picture group (M = 15 SD = 12) the drawing group(M = 20 SD = 14) did not differ significantly from the other groupsThus we included studentsrsquo prior knowledge as a covariate in thefollowing analyses

343 Is there support for the generative drawing effectA major goal in this experiment was to determine whether asking

students to generate drawings to represent science text is a moreeffective learning strategy than asking students to learn with textalone or with text and author-generated pictures In other wordswe wanted to determine whether we could replicate and extendthe learner-generated drawing effect Additionally we were inter-ested in whether giving students an author-generated picture afterdrawing can increase the benefits of the learning strategy Mean pro-portion correct and standard deviations on the comprehension anddrawing tests for the four groups are presented in Table 3

The left portion of Table 3 summarizes the mean proportioncorrect on the comprehension test A two-factorial analysis ofcovariance (ANCOVA) predicting learning outcomes (comprehen-sion posttest score) with learner-generated drawing (yesno) and

author-generated picture (yesno) as the factorial independent vari-ables and prior knowledge as a covariate showed a significantpositive main effect of learner-generated drawing F(1 163) = 398p = 048 partial eta2 = 02 a significant interaction effect F(1163) = 626 p = 013 partial eta2 = 04 but no main effect of author-generated pictures F lt 1 In addition multiple pairwise comparisons(with p lt 05) showed that the drawing group performed signifi-cantly better than each of the three other groups which did not differsignificantly from each other Cohenrsquos d favoring the drawing groupover the author-generated picture group was 49 over the learner-generated + author-generated picture group was 57 and over thecontrol group was 52

The right portion of Table 3 summarizes the mean proportioncorrect on the drawing posttest Again a two-factorial analysis ofcovariance (ANCOVA) predicting learning outcome (drawing testscore) with learner-generated drawing (yesno) and author-generatedpicture (yesno) as the factorial independent variables and priorknowledge as a covariate showed a significant positive main effectof learner-generated drawing F(1 163) = 6260 p lt 001 partialeta2 = 28 a significant positive main effect of author-generated pic-tures F(1 163) = 1104 p = 001 partial eta2 = 06 and a significantinteraction effect F(1 163) = 1658 p lt 001 partial eta2 = 09 In ad-dition multiple pairwise comparisons (with p lt 05) showed thatboth the drawing group and the drawing + author-generated picturegroup performed significantly better than the author-generatedpicture group (d = 68 d = 59) and the control group (d = 187d = 188) In turn the author-generated picture group performed sig-nificantly better than the control group (d = 95) The drawing groupand the drawing + author-generated picture group did not differ sig-nificantly from each other (d = 15)2 Overall these results areconsistent with Experiment 1 and provide additional support forthe generative drawing effect

In accordance with Experiment 1 we were interested in whetherdifferences in learning time among the experimental groups mediatethe positive effect of drawing on text comprehension First an ANOVApredicting learning time with learner-generated drawing (yesno)and author-generated picture (yesno) as the factorial indepen-dent variables showed a significant main effect of learner-generateddrawing F(1 164) = 39226 p lt 001 partial eta2 = 71 a significantmain effect of author-generated picture F(1 164) = 1685 p lt 001partial eta2 = 09 and a significant interaction effect F(1 164) = 490p = 028 partial eta2 = 03 Linear contrasts (with p lt 05) revealedthat the drawing group (M = 1938 min SD = 380) and thedrawing + author-generated picture group (M = 2340 min SD = 551)needed significantly more learning time than the author-generatedpicture group (M = 960 min SD = 397) and the control group(M = 834 min SD = 251) Thus to test whether learning time me-diates the positive effect of drawing on text comprehensionadditional mediation analyses (Baron amp Kenny 1986) were calcu-lated by including learning time as an additional predictor in theaforementioned linear model Results of the mediation analysesshowed that the effects of drawing on the comprehension posttestand the drawing posttest scores (see multiple pairwise compari-sons) are mediated by learning time to some extent That is includinglearning time in the linear model for predicting comprehension testscores still revealed a positive effect of the drawing group com-pared with the drawing + author-generated group on thecomprehension test (p = 012) However including learning time inthe linear model for predicting comprehension posttest scoresreduced the positive effect of the drawing group compared with theauthor-generated picture group (from p = 034 to p = 281) as wellas compared with the control group (from p = 002 to p = 087) being

2 The exclusion of studentsrsquo prior knowledge in the reported analyses does notchange the overall pattern of results

Table 3Mean proportion correct on the comprehension test and drawing test for the fourgroups ndash Experiment 2

Group Type of test

n Comprehension test Drawing test

M SD M SD

Learner-generated drawing 40 63 22 66 22Author-generated picture 44 53 19 50 25Learner-generated drawing +

author-generated picture41 51 20 63 19

Control 43 52 20 30 16

Note Asterisk () indicates significant difference from control group at p lt 05

283A Schmeck et alContemporary Educational Psychology 39 (2014) 275ndash286

no longer statistically significant Regarding the drawing posttestscore including learning time does not change the reported patternof results except that the positive effect of the drawing + author-generated picture group compared with the author-generated picturegroup is no longer statistically significant (from p = 004 to p = 223)

There were neither main effects of learner-generated drawingand author-generated pictures on the mental effort item (drawinggroup M = 455 SD = 025 author-generated picture group M = 459SD = 024 drawing + author-generated picture group M = 444SD = 026 control group M = 481 SD = 024 F lt 1) nor on theperceived difficulty item (drawing group M = 363 SD = 023 author-generated picture group M = 371 SD = 022 drawing + author-generated picture group M = 395 SD = 023 control group M = 393SD = 022 F lt 1)

Taken together the drawing strategy apparently fosters stu-dents to engage in generative activities indicated by their higherlearning outcomes Thus the data provide further evidence for thegenerative drawing effect predicted by Schwamborn et al (2010)In Experiment 2 benefits of the drawing activity however are me-diated by learning time and do not involve higher mental effortAdditionally there was no increased benefit when additional drawingsupport was available in the form of author-generated pictures

344 Is there support for the prognostic drawing effectA second major goal of this study was to determine whether the

prognostic drawing effect could be extended to a new context Meanproportion correct on drawing-accuracy during learning was 60(SD = 04) for the drawing group and 68 (SD = 03) for thedrawing + author generated picture group This difference betweenthe two drawing groups is not significant F(1 79) = 252 p = 116This lack of group differences allowed us to pool the data of bothdrawing groups for subsequent correlation analyses Correlation anal-yses based on the combined data from the two drawing groupsrevealed that the drawing-accuracy score of learner-generated draw-ings correlates significantly with the comprehension posttest scorer = 470 p lt 001 as well as with the drawing posttest score r = 615p lt 001 Additional correlation analyses revealed that the drawing-accuracy score of learner-generated drawings did not correlatesignificantly with the prior knowledge test score r = 095 p = 400the spatial ability test score r = 127 p = 257 the motivation testscore r = 033 p = 769 or the mental effort test score r = 042p = 712 The correlation between the drawing-accuracy score andthe perceived difficulty score was only slightly statistical signifi-cance r = minus218 p = 053 Thus the data provide further evidencefor the prognostic drawing effect consistent with the results ofSchwamborn et al (2010)

In sum results of Experiment 2 are partly consistent with theresults of Experiment 1 in that students learn better from a sciencetext when they are asked to draw illustrations representing the mainideas of the text and the quality of the generated drawings duringlearning correlates positively with studentsrsquo text comprehension (egSchwamborn et al 2010 van Meter 2001 van Meter et al 2006)

4 Discussion

41 Empirical contributions

The present set of experiments makes three empirical contri-butions to the field First this study shows strong and consistentevidence that students who are asked to generate drawings (withsufficient support) during reading a scientific text that describes acausal sequence perform better than students who read withoutdrawing both on a comprehension test (d = 085 in Experiment 1and d = 052 in Experiment 2) and on a drawing test (d = 115 in Ex-periment 1 and d = 187 in Experiment 2) Thus the generativedrawing effect can be extended to a new domain and therefore

corresponds to Shavelson and Townersquos (2002 p 4) recommenda-tion to ldquoreplicate and generalize across studiesrdquo as one of the sixessential scientific principles of educational research

Second this study shows strong and consistent evidence that thequality of drawings that students generate during learning with ascientific text that describes a causal sequence is positively relatedto subsequent performance on tests of learning outcome includ-ing a comprehension test (r = 623 in Experiment 1 and r = 470 inExperiment 2) and a drawing test (r = 620 in Experiment 1 andr = 615 in Experiment 2) Thus the prognostic drawing effect canbe replicated and extended to a new domain consistent with stan-dards for scientific research in education prescribed by Shavelsonand Towne (2002)

Third this study shows that asking learners to draw picturesduring reading a scientific text (ie learner-generated drawing groupin Experiment 2) is more effective than simply providing draw-ings (ie author-generated picture group in Experiment 2) both ona comprehension test (d = 049) and a drawing test (d = 068) Sim-ilarly adding author-generated drawings (ie learner-generatedpictures + author-generated pictures group in Experiment 2) doesnot improve the learning outcomes of students who also draw pic-tures during learning (ie learner-generated pictures group inExperiment 2) either on a comprehension test (d = minus057) or adrawing test (d = minus015) In short the act of drawing during learn-ing (with sufficient support) improves learning beyond the simpleprovision of drawings

42 Theoretical contributions

The results are consistent with the idea that drawing during learn-ing serves as a generative activity (Mayer amp Wittrock 2006Schwamborn et al 2010 van Meter amp Garner 2005 Wittrock 1990)That is the act of drawing encourages learners to engage in gen-erative cognitive processing during learning such as organizing therelevant information into a coherent structure and integrating itwith relevant prior knowledge from long-term memory In thepresent study positive effects of drawing were indicated with a com-prehension and a drawing learning outcome test and therefore arein line with the theoretical assumption derived from the GTDC thatbenefits of drawing can be found if learning outcome tests are usedthat are sensitive to the underlying process of drawing (cf van Meteramp Garner 2005) Additionally in our study the drawing activity wassupported in a way that was intended to help learners carry out theunderlying cognitive processes of drawing (ie selecting organiz-ing and integrating) successfully In this regard results of the presentstudy might supplement the theoretical framework of learner-generated drawing by providing further evidence that benefits ofdrawing defined by van Meter and Garnerrsquos GTDC can diminish ifno instructional support is given to constrain and structure thedrawing activity However a fuller understanding of the underly-ing cognitive processes of drawing and how these processes canbe influenced via drawing support requires more direct measuresof cognitive processing during learning Additionally following theidea that metacognitive processes of monitoring and regulation areautomatically activated by drawing (van Meter amp Garner 2005) afuller understanding of the metacognitive effects of drawings is alsorequired

43 Practical contributions

The present study encourages instructional designers and in-structors to incorporate drawing activities into venues involvinglearning from text which we call the generative drawing effect Oneimportant feature of a successful drawing strategy that is presentin this study and in a previous study by Schwamborn et al (2010)is that the drawing activity was supported by providing a

284 A Schmeck et alContemporary Educational Psychology 39 (2014) 275ndash286

background scene and a legend showing how to represent eachelement to constrain and structure the drawing activity Thus animportant practical implication is that students may need supportin their drawing activity so they do not need to draw from scratch

The present study also suggests a potentially useful diagnostictool to gauge the depth of student learning namely the quality ofthe drawings created by students during learning which we referto as the prognostic drawing effect Incorporating a measure of thequality of a learnerrsquos drawing during learning can be a useful toolin developing remedial instruction to give learners individual supportIt may be important to use materials that explain a cause-and-effect process and give learners drawings of the elements they needto represent the process pictorially Asking learners to simply drawpictures of elements is unlikely to be helpful whereas asking themto generate drawings that show the relations among the elementsin a schematic form is more likely to be helpful

44 Limitations and future directions

Some limitations and future directions of our study should beaddressed As noted in the theoretical contributions subsection wedid not have direct measures of cognitive processing during learn-ing so it is not possible to pinpoint how the drawing activity affectedspecific cognitive processes such as attending to relevant informa-tion organizing it and integrating it with prior knowledge We alsodid not assess metacognitive processing during learning thus it isnot possible to pinpoint how the drawing activity affected specificmetacognitive processes such as monitoring and regulation

Furthermore results of the cognitive load rating scales (in-vested mental effort and perceived task difficulty) are inconsistentWhereas in Experiment 1 an effect on mental effort but not on per-ceived task difficulty showed up (ie students in the drawing grouprated their invested mental effort during learning significantlyhigher) no effects on mental effort and task difficulty were foundin Experiment 2 Additionally in both experiments only a nega-tive correlation of perceived task difficulty with the quality of learner-generated pictures appears but no correlation of mental effort withthe quality Following de Jong (2010) those cognitive load ratingscales might have the disadvantages that they do not give a con-current measure of cognitive load and do not measure an essentialconcept in cognitive load theory namely cognitive overload (p 125)Future studies on learner-generated drawing might also use othercognitive load measures such as physiological measures as moredirect indicators of cognitive load

As noted in the practical contributions subsection we showedthe drawing effects by using a scientific text describing how a cause-and-effect system works that is the causal steps regarding aninfection with influenza and the immune response It might be pos-sible however that for other types of text producing drawings mightharm rather than promote text comprehension Thus to test whetherthe reported drawing effects can be extended future research hasto focus on other types of text such as descriptive texts as well ason other types of relations that can be conveyed with other typesof representations such as compare and contrast relations whichcan be shown in a matrix Additionally studentsrsquo learning out-comes were tested immediately after reading thus future work isneeded to investigate the longer-term effects of generative drawingon learning outcomes

Furthermore we only compared drawing with control groupsthat received no further learning strategy instructions However en-gaging in generative learning activities such as drawing requires aconsiderable amount of time Accordingly results showed that forExperiment 2 the positive effect of the drawing group on text com-prehension compared with the author-generated picture group andto the control group was mediated by learning time To rule out thatthe effects of drawing result only from additional time on task instead

of the generative activity future research should also compare thedrawing strategy with other time demanding generative learningstrategies such as summarization (cf Leopold amp Leutner 2012)

Another point that should be noted is that students in both ex-periments received some kind of multimedia materials in that evenwhen they had to draw and did not see presented pictures they wereat least provided with the basic (visual) elements for their draw-ings which they had to do on the given background which thusalso contained information In other words when students are pre-sented with important elements of the drawings which they canuse to draw themselves they will not have to put as much effortinto summarizing visually what they have just read compared withstudents who have to draw without any instructional help Futurestudies might also compare the drawing group with a summariza-tion group in which students receive a set of verbal key terms thatare similar to the drawing elements and are asked to make a textualsummary

Additionally future research is needed to validate the prognos-tic drawing effect So far we know that the quality of learner-generated pictures is related to studentsrsquo learning outcomes (iethe higher the learning outcome the higher the drawing accuracyand vice versa) and their perceived difficulty (ie the lower the per-ceived difficulty the higher the drawing accuracy and vice versa)and that it is not related to studentsrsquo prior knowledge motivationspatial ability or mental effort However less is known about whatthis might mean That is less is known regarding the causal direc-tion of this relation or the presence of a possible further moderatorvariable Do studentsrsquo efforts to produce accurate drawings lead tobetter comprehension and lower perceived difficulty Or do stu-dents who are more adept in drawing benefit more from the strategyand thus perceive the difficulty of the learning materials as beinglower Both arguments seem convincing

Finally more work is needed to determine the level of supportthat makes the drawing strategy most effective for various kinds oflearners As noted in the empirical contribution adding author-generated drawings (ie learner-generated pictures + author-generated pictures group in Experiment 2) does not improve thelearning outcomes of students who also draw pictures during learn-ing and were supported by a drawing prompt In other words thecombination of two ways of supporting the drawing strategy (iegiving a drawing prompt during reading plus an author-generatedpicture after reading) did not improve studentsrsquo learning out-comes compared with students in the drawing group as well ascompared with students in the control and author-generated pic-tures only groups This result is inconsistent with previous research(eg van Meter 2001 van Meter et al 2006) which found that com-paring own drawings to author-generated pictures normally helpslearning van Meter and colleagues (2001 2006) however provid-ed author-generated pictures plus prompting questions after thedrawing process That is students answered prompting questionsto guide the comparison process between their self-generateddrawing and the author-generated drawing In our study studentswere only instructed to generate a drawing to inspect an author-generated one and to check whether their own drawing incomparison with the author-generated one really represented themain ideas of the text paragraph correctly In other words we didnot guide the process of comparing self-generated drawings withauthor-generated ones As a potential consequence students per-formed the intended comparison process inadequately or even notall and thus did not benefit from it One reason for this inade-quate comparison process might be that students need guidancein doing the comparison process Another reason might be the factthat students do not seriously engage in generating drawings oncethey notice that there are author-generated drawings Thus futureresearch should also use additional guidance to test whether thecombination of different ways of supporting the drawing strategy

285A Schmeck et alContemporary Educational Psychology 39 (2014) 275ndash286

(ie giving a drawing prompt during reading plus an author-generated picture after reading) helps learning as well asobservational measures of the drawing process itself (ie think aloudprotocols) to shed more light on the cognitive processes underly-ing the drawing activities

Overall drawing during learning from text appears to be a po-tentially powerful strategy for improving studentsrsquo learning fromscientific text when certain boundaries and prerequisites are takeninto account

Acknowledgments

This article is based on a research project funded by the GermanResearch Foundation (DFG LE 6459-3 as part of FOR 511) We wouldlike to thank Angela Sandmann for her assistance in developing thelearning materials

References

Ainsworth S Prain V amp Tytler R (2011) Drawing to learn in science Science 3331096ndash1097

Alesandrini K L (1981) Pictorial-verbal and analytic-holistic learning strategies inscience learning Journal of Educational Psychology 73 358ndash368

Alesandrini K L (1984) Pictures and adult learning Instructional Science 13 63ndash77Baron R M amp Kenny D A (1986) The moderator-mediator variable distinction in

social psychological research Conceptual strategic and statistical considerationsJournal of Personality and Social Psychology 51 1173ndash1182

Bruumlnken R Plass J L amp Leutner D (2003) Direct measurement of cognitive loadin multimedia learning Educational Psychologist 38 53ndash61

Carney RN amp Levin JR (2002) Pictorial illustrations still improve studentsrsquo learningfrom text Educational Psychology Review 14 5ndash26

de Jong T (2005) The guided discovery principle in multimedia learning In R EMayer (Ed) The Cambridge handbook of multimedia learning (pp 215ndash228) NewYork Cambridge University Press

de Jong T (2010) Cognitive load theory educational research and instructionaldesign Some food for thought Instructional Science 38 105ndash134

Ekstrom R B French J W amp Harman H H (1976) Manual for kit of factor-referencedcognitive tests Princeton NJ Educational Testing Service

Greene T R (1989) Childrenrsquos understanding of class inclusion hierarchies Therelationship between external representation and task performance Journal ofExperimental Child Psychology 48 62ndash89

Hall V C Bailey J amp Tillman C (1997) Can student-generated illustrations be worthten thousand words Journal of Educational Psychology 89 677ndash681

Houmlffler T N (2010) Spatial ability Its influence on Learning with visualizations ndashA meta-analytic review Educational Psychology Review 22 245ndash269

Houmlffler T N Schmeck A amp Opfermann M (2013) Static and dynamic visualrepresentations Individual differences in processing In G Schraw M TMcCrudden amp D Robinson (Eds) Learning through visual displays (pp 133ndash163)Charlotte NC Information Age Publishing

Kalyuga S Chandler P amp Sweller J (1999) Managing split-attention and redundancyin multimedia instruction Applied Cognitive Psychology 13 351ndash371

Leopold C (2009) Lernstrategien und Textverstehen [Learning strategies and textcomprehension] Muumlnster Waxmann

Leopold C amp Leutner D (2012) Science text comprehension Drawing main ideaselection and summarizing as learning strategies Learning and Instruction 2216ndash26

Lesgold A M DeGood H amp Levin J R (1977) Pictures and young childrenrsquos proselearning A supplementary report Journal of Reading Behavior 9 353ndash360

Lesgold A M Levin J R Shimron J amp Guttman J (1975) Pictures andyoung childrenrsquos learning from oral prose Journal of Educational Psychology 67636ndash642

Leutner D Leopold C amp Sumfleth E (2009) Cognitive load and science textcomprehension Effects of drawing and mentally imagining text contentComputers in Human Behavior 25 284ndash289

Mayer R E (2004) Should there be a three-strikes rule against pure discoverylearning The case for guided methods of instruction The American Psychologist59 14ndash19

Mayer R E (2005) Cognitive theory of multimedia learning In R E Mayer (Ed)The Cambridge handbook of multimedia learning (pp 31ndash48) New York CambridgeUniversity Press

Mayer R E (2009) Multimedia learning (2nd ed) New York NY CambridgeUniversity Press

Mayer R E amp Wittrock M C (2006) Problem solving In P Alexander P Winne ampG Phye (Eds) Handbook of educational psychology (pp 287ndash303) Mahwah NJErlbaum

Paas F (1992) Training strategies for attaining transfer of problem-solving skill instatisticsmdashA cognitive-load approach Journal of Educational Psychology 84429ndash434

Paas F Tuovinen J Tabbers H K amp Van Gerven P W M (2003) Cognitive loadmeasurement as a means to advance cognitive load theory EducationalPsychologist 38 63ndash71

Paivio A (1986) Mental representation A dual coding approach New York OxfordUniversity Press

Pashler H Bain P Bottage B Graesser A Koedinger K McDaniel M et al (2007)Organizing instruction and study to improve student learning Washington DCNational Center for Educational Research

Rasco R W Tennyson R D amp Boutwell R C (1975) Imagery instructions anddrawings in learning prose Journal of Educational Psychology 67 188ndash192

Rheinberg F Vollmeyer R amp Burns B D (2001) FAM Ein fragebogen zurerfassung aktueller motivation in lern- und leistungssituationen [QCM Aquestionnaire to assess current motivation in learning situations] Diagnostica47 57ndash66

Schnotz W (2005) An integrated model of text and picture comprehension In RE Mayer (Ed) The Cambridge handbook of multimedia learning (pp 49ndash70) NewYork Cambridge University Press

Schwamborn A Mayer R E Thillmann H Leopold C amp Leutner D (2010) Drawingas a generative activity and drawing as a prognostic activity Journal of EducationalPsychology 102 872ndash879

Schwamborn A Thillmann H Opfermann M amp Leutner D (2011) Cognitive loadand instructionally supported learning with provided and learner-generatedvisualizations Computers in Human Behavior 27 89ndash93

Shavelson R J amp Towne L (Eds) (2002) Scientific research in education WashingtonDC National Academy Press

Sweller J Ayres P amp Kalyuga S (2011) Cognitive Load Theory New York SpringerTirre W C Manelis L amp Leicht K (1979) The effects of imaginal and verbal strategies

on prose comprehension by adults Journal of Reading Behavior 11 99ndash106van Meter P (2001) Drawing construction as a strategy for learning from text Journal

of Educational Psychology 69 129ndash140van Meter P Aleksic M Schwartz A amp Garner J (2006) Learner-generated drawing

as a strategy for learning from content area text Contemporary EducationalPsychology 31 142ndash166

van Meter P amp Garner J (2005) The promise and practice of learner-generateddrawings Literature review and synthesis Educational Psychology Review 12261ndash312

Van Gog T amp Paas F (2008) Instructional efficiency Revisiting the original constructin educational research Educational Psychologist 43 16ndash26

Vollmeyer R amp Rheinberg F (2000) Does motivation affect learning via persistenceLearning and Instruction 4 293ndash309

Weinstein C E amp Mayer R E (1986) The teaching of learning strategies In M CWittrock (Ed) Handbook of research on teaching (pp 315ndash327) New YorkMacmillan

Wittrock M C (1990) Generative processes of comprehension EducationalPsychologist 24 345ndash376

286 A Schmeck et alContemporary Educational Psychology 39 (2014) 275ndash286

  • Drawing pictures during learning from scientific text testing the generative drawing effect and the prognostic drawing effect
  • Introduction
  • Theoretical framework for the learner-generated drawing strategy
  • Empirical framework for the learner-generated drawing strategy
  • Effectiveness of learner-generated drawings
  • Quality of learner-generated drawings
  • Overview of the experiments
  • Experiment 1
  • Participants and design
  • Materials
  • Procedure
  • Results and discussion
  • Scoring
  • Are the groups equivalent on basic characteristics
  • Is there support for the generative drawing effect
  • Is there support for the prognostic drawing effect
  • Experiment 2
  • Participants and design
  • Materials
  • Procedure
  • Results
  • Scoring
  • Are the groups equivalent on basic characteristics
  • Is there support for the generative drawing effect
  • Is there support for the prognostic drawing effect
  • Discussion
  • Empirical contributions
  • Theoretical contributions
  • Practical contributions
  • Limitations and future directions
  • Acknowledgments
  • References

because they are too difficult to answer without prior training inthe topic Additionally author-generated pictures were used in thenew conditions The author-generated pictures were static func-tional pictures representing the main ideas of each paragraph andconsisted of pictorial elements identical to those provided in thedrawing prompt (as shown in Fig 2) These pictures were con-structed by the first author in cooperation with a biology teacher

The drawing version of the booklet was identical to that usedin Experiment 1 (as shown in Fig 1) The control version of the learn-ing booklet was identical to that used in Experiment 1 The author-generated picture version of the booklet consisted of seven pairsof facing pages with a text paragraph on the left page and a corre-sponding author-generated picture (such as in Fig 2) on the rightpage The drawing + author-generated picture version of the bookletcontained the material from the drawing version consisting of sevenpairs of facing pages with a text paragraph on the left page and atwo-part drawing prompt on the right page In addition attachedto each page there was an additional page that students could foldout after having generated their drawing When unfolding this ad-ditional page a picture of that text paragraph right aside the drawingprompt was provided and there was an additional instruction to

compare the learner-generated drawing with the author-generatedpicture Author-generated pictures were the same as in the author-generated picture version of the booklet

33 Procedure

The procedure was identical to that used in Experiment 1 exceptthat there were two additional groups learning with author-generated pictures Students in the author-generated picturecondition were instructed to read the text and additionally to lookat pictures representing the main ideas of each text paragraph Stu-dents in the drawing + author-generated picture version of thebooklet were instructed to read the text to draw pictures for eachtext paragraph using the drawing prompt representing the mainideas of each text paragraph and finally to compare their picturewith an author-generated picture representing main ideas of eachparagraph correctly Students in all groups learned at their own pacewhereby individual learning time was measured by the instruc-tors in the classrooms Again to ensure that studentsrsquo in both drawinggroups did not feel rushed when students in the non-drawing

Fig 2 Author-generated pictures for the seven paragraphs in the author-generated picture versions of the learning booklet Note Pictures are scaled-down from the orig-inal format

282 A Schmeck et alContemporary Educational Psychology 39 (2014) 275ndash286

group completed the task early groups were tested in separateclassrooms

34 Results

341 ScoringAll tests instruments were scored with the same procedures used

in Experiment 1 Again two student assistants (teacher trainees inbiology) scored each of the drawing test items and each of the sevenlearner-generated drawings for each student with acceptableinterrater agreements (drawing test GoodmanndashKruskal gamma of90 drawing-accuracy GoodmanndashKruskal gamma of 94) Actualscores ranged from 1 to 28 points (M = 153 points SD = 58) for thecomprehension test from zero to 21 points (M = 109 points SD = 53)for the drawing test and from 275 to 215 points (M = 141 pointsSD = 48) for drawing accuracy Again total scores of comprehen-sion drawing and accuracy were transferred into z-standardizedscores

342 Are the groups equivalent on basic characteristicsBefore looking at treatment effects on the dependent variables

we analyzed whether the four experimental groups differed onseveral control variables A chi-square analysis indicated that therewere no significant differences regarding gender (p = 097) Sepa-rate univariate analyses of variance (ANOVAs) revealed that thegroups also did not differ significantly on age F lt 1 on spatial abilityF(3 164) = 120 p = 312 or on motivation F(3 164) = 122 p = 305However groups differed significantly on prior knowledge F(3164) = 104 p = 010 partial eta2 = 07 in that students in the controlgroup scored significantly higher on the comprehension pretest(M = 25 SD = 17) than students in both (p lt 05) the author-generated picture group (M = 17 SD = 16) and the drawing + author-generated picture group (M = 15 SD = 12) the drawing group(M = 20 SD = 14) did not differ significantly from the other groupsThus we included studentsrsquo prior knowledge as a covariate in thefollowing analyses

343 Is there support for the generative drawing effectA major goal in this experiment was to determine whether asking

students to generate drawings to represent science text is a moreeffective learning strategy than asking students to learn with textalone or with text and author-generated pictures In other wordswe wanted to determine whether we could replicate and extendthe learner-generated drawing effect Additionally we were inter-ested in whether giving students an author-generated picture afterdrawing can increase the benefits of the learning strategy Mean pro-portion correct and standard deviations on the comprehension anddrawing tests for the four groups are presented in Table 3

The left portion of Table 3 summarizes the mean proportioncorrect on the comprehension test A two-factorial analysis ofcovariance (ANCOVA) predicting learning outcomes (comprehen-sion posttest score) with learner-generated drawing (yesno) and

author-generated picture (yesno) as the factorial independent vari-ables and prior knowledge as a covariate showed a significantpositive main effect of learner-generated drawing F(1 163) = 398p = 048 partial eta2 = 02 a significant interaction effect F(1163) = 626 p = 013 partial eta2 = 04 but no main effect of author-generated pictures F lt 1 In addition multiple pairwise comparisons(with p lt 05) showed that the drawing group performed signifi-cantly better than each of the three other groups which did not differsignificantly from each other Cohenrsquos d favoring the drawing groupover the author-generated picture group was 49 over the learner-generated + author-generated picture group was 57 and over thecontrol group was 52

The right portion of Table 3 summarizes the mean proportioncorrect on the drawing posttest Again a two-factorial analysis ofcovariance (ANCOVA) predicting learning outcome (drawing testscore) with learner-generated drawing (yesno) and author-generatedpicture (yesno) as the factorial independent variables and priorknowledge as a covariate showed a significant positive main effectof learner-generated drawing F(1 163) = 6260 p lt 001 partialeta2 = 28 a significant positive main effect of author-generated pic-tures F(1 163) = 1104 p = 001 partial eta2 = 06 and a significantinteraction effect F(1 163) = 1658 p lt 001 partial eta2 = 09 In ad-dition multiple pairwise comparisons (with p lt 05) showed thatboth the drawing group and the drawing + author-generated picturegroup performed significantly better than the author-generatedpicture group (d = 68 d = 59) and the control group (d = 187d = 188) In turn the author-generated picture group performed sig-nificantly better than the control group (d = 95) The drawing groupand the drawing + author-generated picture group did not differ sig-nificantly from each other (d = 15)2 Overall these results areconsistent with Experiment 1 and provide additional support forthe generative drawing effect

In accordance with Experiment 1 we were interested in whetherdifferences in learning time among the experimental groups mediatethe positive effect of drawing on text comprehension First an ANOVApredicting learning time with learner-generated drawing (yesno)and author-generated picture (yesno) as the factorial indepen-dent variables showed a significant main effect of learner-generateddrawing F(1 164) = 39226 p lt 001 partial eta2 = 71 a significantmain effect of author-generated picture F(1 164) = 1685 p lt 001partial eta2 = 09 and a significant interaction effect F(1 164) = 490p = 028 partial eta2 = 03 Linear contrasts (with p lt 05) revealedthat the drawing group (M = 1938 min SD = 380) and thedrawing + author-generated picture group (M = 2340 min SD = 551)needed significantly more learning time than the author-generatedpicture group (M = 960 min SD = 397) and the control group(M = 834 min SD = 251) Thus to test whether learning time me-diates the positive effect of drawing on text comprehensionadditional mediation analyses (Baron amp Kenny 1986) were calcu-lated by including learning time as an additional predictor in theaforementioned linear model Results of the mediation analysesshowed that the effects of drawing on the comprehension posttestand the drawing posttest scores (see multiple pairwise compari-sons) are mediated by learning time to some extent That is includinglearning time in the linear model for predicting comprehension testscores still revealed a positive effect of the drawing group com-pared with the drawing + author-generated group on thecomprehension test (p = 012) However including learning time inthe linear model for predicting comprehension posttest scoresreduced the positive effect of the drawing group compared with theauthor-generated picture group (from p = 034 to p = 281) as wellas compared with the control group (from p = 002 to p = 087) being

2 The exclusion of studentsrsquo prior knowledge in the reported analyses does notchange the overall pattern of results

Table 3Mean proportion correct on the comprehension test and drawing test for the fourgroups ndash Experiment 2

Group Type of test

n Comprehension test Drawing test

M SD M SD

Learner-generated drawing 40 63 22 66 22Author-generated picture 44 53 19 50 25Learner-generated drawing +

author-generated picture41 51 20 63 19

Control 43 52 20 30 16

Note Asterisk () indicates significant difference from control group at p lt 05

283A Schmeck et alContemporary Educational Psychology 39 (2014) 275ndash286

no longer statistically significant Regarding the drawing posttestscore including learning time does not change the reported patternof results except that the positive effect of the drawing + author-generated picture group compared with the author-generated picturegroup is no longer statistically significant (from p = 004 to p = 223)

There were neither main effects of learner-generated drawingand author-generated pictures on the mental effort item (drawinggroup M = 455 SD = 025 author-generated picture group M = 459SD = 024 drawing + author-generated picture group M = 444SD = 026 control group M = 481 SD = 024 F lt 1) nor on theperceived difficulty item (drawing group M = 363 SD = 023 author-generated picture group M = 371 SD = 022 drawing + author-generated picture group M = 395 SD = 023 control group M = 393SD = 022 F lt 1)

Taken together the drawing strategy apparently fosters stu-dents to engage in generative activities indicated by their higherlearning outcomes Thus the data provide further evidence for thegenerative drawing effect predicted by Schwamborn et al (2010)In Experiment 2 benefits of the drawing activity however are me-diated by learning time and do not involve higher mental effortAdditionally there was no increased benefit when additional drawingsupport was available in the form of author-generated pictures

344 Is there support for the prognostic drawing effectA second major goal of this study was to determine whether the

prognostic drawing effect could be extended to a new context Meanproportion correct on drawing-accuracy during learning was 60(SD = 04) for the drawing group and 68 (SD = 03) for thedrawing + author generated picture group This difference betweenthe two drawing groups is not significant F(1 79) = 252 p = 116This lack of group differences allowed us to pool the data of bothdrawing groups for subsequent correlation analyses Correlation anal-yses based on the combined data from the two drawing groupsrevealed that the drawing-accuracy score of learner-generated draw-ings correlates significantly with the comprehension posttest scorer = 470 p lt 001 as well as with the drawing posttest score r = 615p lt 001 Additional correlation analyses revealed that the drawing-accuracy score of learner-generated drawings did not correlatesignificantly with the prior knowledge test score r = 095 p = 400the spatial ability test score r = 127 p = 257 the motivation testscore r = 033 p = 769 or the mental effort test score r = 042p = 712 The correlation between the drawing-accuracy score andthe perceived difficulty score was only slightly statistical signifi-cance r = minus218 p = 053 Thus the data provide further evidencefor the prognostic drawing effect consistent with the results ofSchwamborn et al (2010)

In sum results of Experiment 2 are partly consistent with theresults of Experiment 1 in that students learn better from a sciencetext when they are asked to draw illustrations representing the mainideas of the text and the quality of the generated drawings duringlearning correlates positively with studentsrsquo text comprehension (egSchwamborn et al 2010 van Meter 2001 van Meter et al 2006)

4 Discussion

41 Empirical contributions

The present set of experiments makes three empirical contri-butions to the field First this study shows strong and consistentevidence that students who are asked to generate drawings (withsufficient support) during reading a scientific text that describes acausal sequence perform better than students who read withoutdrawing both on a comprehension test (d = 085 in Experiment 1and d = 052 in Experiment 2) and on a drawing test (d = 115 in Ex-periment 1 and d = 187 in Experiment 2) Thus the generativedrawing effect can be extended to a new domain and therefore

corresponds to Shavelson and Townersquos (2002 p 4) recommenda-tion to ldquoreplicate and generalize across studiesrdquo as one of the sixessential scientific principles of educational research

Second this study shows strong and consistent evidence that thequality of drawings that students generate during learning with ascientific text that describes a causal sequence is positively relatedto subsequent performance on tests of learning outcome includ-ing a comprehension test (r = 623 in Experiment 1 and r = 470 inExperiment 2) and a drawing test (r = 620 in Experiment 1 andr = 615 in Experiment 2) Thus the prognostic drawing effect canbe replicated and extended to a new domain consistent with stan-dards for scientific research in education prescribed by Shavelsonand Towne (2002)

Third this study shows that asking learners to draw picturesduring reading a scientific text (ie learner-generated drawing groupin Experiment 2) is more effective than simply providing draw-ings (ie author-generated picture group in Experiment 2) both ona comprehension test (d = 049) and a drawing test (d = 068) Sim-ilarly adding author-generated drawings (ie learner-generatedpictures + author-generated pictures group in Experiment 2) doesnot improve the learning outcomes of students who also draw pic-tures during learning (ie learner-generated pictures group inExperiment 2) either on a comprehension test (d = minus057) or adrawing test (d = minus015) In short the act of drawing during learn-ing (with sufficient support) improves learning beyond the simpleprovision of drawings

42 Theoretical contributions

The results are consistent with the idea that drawing during learn-ing serves as a generative activity (Mayer amp Wittrock 2006Schwamborn et al 2010 van Meter amp Garner 2005 Wittrock 1990)That is the act of drawing encourages learners to engage in gen-erative cognitive processing during learning such as organizing therelevant information into a coherent structure and integrating itwith relevant prior knowledge from long-term memory In thepresent study positive effects of drawing were indicated with a com-prehension and a drawing learning outcome test and therefore arein line with the theoretical assumption derived from the GTDC thatbenefits of drawing can be found if learning outcome tests are usedthat are sensitive to the underlying process of drawing (cf van Meteramp Garner 2005) Additionally in our study the drawing activity wassupported in a way that was intended to help learners carry out theunderlying cognitive processes of drawing (ie selecting organiz-ing and integrating) successfully In this regard results of the presentstudy might supplement the theoretical framework of learner-generated drawing by providing further evidence that benefits ofdrawing defined by van Meter and Garnerrsquos GTDC can diminish ifno instructional support is given to constrain and structure thedrawing activity However a fuller understanding of the underly-ing cognitive processes of drawing and how these processes canbe influenced via drawing support requires more direct measuresof cognitive processing during learning Additionally following theidea that metacognitive processes of monitoring and regulation areautomatically activated by drawing (van Meter amp Garner 2005) afuller understanding of the metacognitive effects of drawings is alsorequired

43 Practical contributions

The present study encourages instructional designers and in-structors to incorporate drawing activities into venues involvinglearning from text which we call the generative drawing effect Oneimportant feature of a successful drawing strategy that is presentin this study and in a previous study by Schwamborn et al (2010)is that the drawing activity was supported by providing a

284 A Schmeck et alContemporary Educational Psychology 39 (2014) 275ndash286

background scene and a legend showing how to represent eachelement to constrain and structure the drawing activity Thus animportant practical implication is that students may need supportin their drawing activity so they do not need to draw from scratch

The present study also suggests a potentially useful diagnostictool to gauge the depth of student learning namely the quality ofthe drawings created by students during learning which we referto as the prognostic drawing effect Incorporating a measure of thequality of a learnerrsquos drawing during learning can be a useful toolin developing remedial instruction to give learners individual supportIt may be important to use materials that explain a cause-and-effect process and give learners drawings of the elements they needto represent the process pictorially Asking learners to simply drawpictures of elements is unlikely to be helpful whereas asking themto generate drawings that show the relations among the elementsin a schematic form is more likely to be helpful

44 Limitations and future directions

Some limitations and future directions of our study should beaddressed As noted in the theoretical contributions subsection wedid not have direct measures of cognitive processing during learn-ing so it is not possible to pinpoint how the drawing activity affectedspecific cognitive processes such as attending to relevant informa-tion organizing it and integrating it with prior knowledge We alsodid not assess metacognitive processing during learning thus it isnot possible to pinpoint how the drawing activity affected specificmetacognitive processes such as monitoring and regulation

Furthermore results of the cognitive load rating scales (in-vested mental effort and perceived task difficulty) are inconsistentWhereas in Experiment 1 an effect on mental effort but not on per-ceived task difficulty showed up (ie students in the drawing grouprated their invested mental effort during learning significantlyhigher) no effects on mental effort and task difficulty were foundin Experiment 2 Additionally in both experiments only a nega-tive correlation of perceived task difficulty with the quality of learner-generated pictures appears but no correlation of mental effort withthe quality Following de Jong (2010) those cognitive load ratingscales might have the disadvantages that they do not give a con-current measure of cognitive load and do not measure an essentialconcept in cognitive load theory namely cognitive overload (p 125)Future studies on learner-generated drawing might also use othercognitive load measures such as physiological measures as moredirect indicators of cognitive load

As noted in the practical contributions subsection we showedthe drawing effects by using a scientific text describing how a cause-and-effect system works that is the causal steps regarding aninfection with influenza and the immune response It might be pos-sible however that for other types of text producing drawings mightharm rather than promote text comprehension Thus to test whetherthe reported drawing effects can be extended future research hasto focus on other types of text such as descriptive texts as well ason other types of relations that can be conveyed with other typesof representations such as compare and contrast relations whichcan be shown in a matrix Additionally studentsrsquo learning out-comes were tested immediately after reading thus future work isneeded to investigate the longer-term effects of generative drawingon learning outcomes

Furthermore we only compared drawing with control groupsthat received no further learning strategy instructions However en-gaging in generative learning activities such as drawing requires aconsiderable amount of time Accordingly results showed that forExperiment 2 the positive effect of the drawing group on text com-prehension compared with the author-generated picture group andto the control group was mediated by learning time To rule out thatthe effects of drawing result only from additional time on task instead

of the generative activity future research should also compare thedrawing strategy with other time demanding generative learningstrategies such as summarization (cf Leopold amp Leutner 2012)

Another point that should be noted is that students in both ex-periments received some kind of multimedia materials in that evenwhen they had to draw and did not see presented pictures they wereat least provided with the basic (visual) elements for their draw-ings which they had to do on the given background which thusalso contained information In other words when students are pre-sented with important elements of the drawings which they canuse to draw themselves they will not have to put as much effortinto summarizing visually what they have just read compared withstudents who have to draw without any instructional help Futurestudies might also compare the drawing group with a summariza-tion group in which students receive a set of verbal key terms thatare similar to the drawing elements and are asked to make a textualsummary

Additionally future research is needed to validate the prognos-tic drawing effect So far we know that the quality of learner-generated pictures is related to studentsrsquo learning outcomes (iethe higher the learning outcome the higher the drawing accuracyand vice versa) and their perceived difficulty (ie the lower the per-ceived difficulty the higher the drawing accuracy and vice versa)and that it is not related to studentsrsquo prior knowledge motivationspatial ability or mental effort However less is known about whatthis might mean That is less is known regarding the causal direc-tion of this relation or the presence of a possible further moderatorvariable Do studentsrsquo efforts to produce accurate drawings lead tobetter comprehension and lower perceived difficulty Or do stu-dents who are more adept in drawing benefit more from the strategyand thus perceive the difficulty of the learning materials as beinglower Both arguments seem convincing

Finally more work is needed to determine the level of supportthat makes the drawing strategy most effective for various kinds oflearners As noted in the empirical contribution adding author-generated drawings (ie learner-generated pictures + author-generated pictures group in Experiment 2) does not improve thelearning outcomes of students who also draw pictures during learn-ing and were supported by a drawing prompt In other words thecombination of two ways of supporting the drawing strategy (iegiving a drawing prompt during reading plus an author-generatedpicture after reading) did not improve studentsrsquo learning out-comes compared with students in the drawing group as well ascompared with students in the control and author-generated pic-tures only groups This result is inconsistent with previous research(eg van Meter 2001 van Meter et al 2006) which found that com-paring own drawings to author-generated pictures normally helpslearning van Meter and colleagues (2001 2006) however provid-ed author-generated pictures plus prompting questions after thedrawing process That is students answered prompting questionsto guide the comparison process between their self-generateddrawing and the author-generated drawing In our study studentswere only instructed to generate a drawing to inspect an author-generated one and to check whether their own drawing incomparison with the author-generated one really represented themain ideas of the text paragraph correctly In other words we didnot guide the process of comparing self-generated drawings withauthor-generated ones As a potential consequence students per-formed the intended comparison process inadequately or even notall and thus did not benefit from it One reason for this inade-quate comparison process might be that students need guidancein doing the comparison process Another reason might be the factthat students do not seriously engage in generating drawings oncethey notice that there are author-generated drawings Thus futureresearch should also use additional guidance to test whether thecombination of different ways of supporting the drawing strategy

285A Schmeck et alContemporary Educational Psychology 39 (2014) 275ndash286

(ie giving a drawing prompt during reading plus an author-generated picture after reading) helps learning as well asobservational measures of the drawing process itself (ie think aloudprotocols) to shed more light on the cognitive processes underly-ing the drawing activities

Overall drawing during learning from text appears to be a po-tentially powerful strategy for improving studentsrsquo learning fromscientific text when certain boundaries and prerequisites are takeninto account

Acknowledgments

This article is based on a research project funded by the GermanResearch Foundation (DFG LE 6459-3 as part of FOR 511) We wouldlike to thank Angela Sandmann for her assistance in developing thelearning materials

References

Ainsworth S Prain V amp Tytler R (2011) Drawing to learn in science Science 3331096ndash1097

Alesandrini K L (1981) Pictorial-verbal and analytic-holistic learning strategies inscience learning Journal of Educational Psychology 73 358ndash368

Alesandrini K L (1984) Pictures and adult learning Instructional Science 13 63ndash77Baron R M amp Kenny D A (1986) The moderator-mediator variable distinction in

social psychological research Conceptual strategic and statistical considerationsJournal of Personality and Social Psychology 51 1173ndash1182

Bruumlnken R Plass J L amp Leutner D (2003) Direct measurement of cognitive loadin multimedia learning Educational Psychologist 38 53ndash61

Carney RN amp Levin JR (2002) Pictorial illustrations still improve studentsrsquo learningfrom text Educational Psychology Review 14 5ndash26

de Jong T (2005) The guided discovery principle in multimedia learning In R EMayer (Ed) The Cambridge handbook of multimedia learning (pp 215ndash228) NewYork Cambridge University Press

de Jong T (2010) Cognitive load theory educational research and instructionaldesign Some food for thought Instructional Science 38 105ndash134

Ekstrom R B French J W amp Harman H H (1976) Manual for kit of factor-referencedcognitive tests Princeton NJ Educational Testing Service

Greene T R (1989) Childrenrsquos understanding of class inclusion hierarchies Therelationship between external representation and task performance Journal ofExperimental Child Psychology 48 62ndash89

Hall V C Bailey J amp Tillman C (1997) Can student-generated illustrations be worthten thousand words Journal of Educational Psychology 89 677ndash681

Houmlffler T N (2010) Spatial ability Its influence on Learning with visualizations ndashA meta-analytic review Educational Psychology Review 22 245ndash269

Houmlffler T N Schmeck A amp Opfermann M (2013) Static and dynamic visualrepresentations Individual differences in processing In G Schraw M TMcCrudden amp D Robinson (Eds) Learning through visual displays (pp 133ndash163)Charlotte NC Information Age Publishing

Kalyuga S Chandler P amp Sweller J (1999) Managing split-attention and redundancyin multimedia instruction Applied Cognitive Psychology 13 351ndash371

Leopold C (2009) Lernstrategien und Textverstehen [Learning strategies and textcomprehension] Muumlnster Waxmann

Leopold C amp Leutner D (2012) Science text comprehension Drawing main ideaselection and summarizing as learning strategies Learning and Instruction 2216ndash26

Lesgold A M DeGood H amp Levin J R (1977) Pictures and young childrenrsquos proselearning A supplementary report Journal of Reading Behavior 9 353ndash360

Lesgold A M Levin J R Shimron J amp Guttman J (1975) Pictures andyoung childrenrsquos learning from oral prose Journal of Educational Psychology 67636ndash642

Leutner D Leopold C amp Sumfleth E (2009) Cognitive load and science textcomprehension Effects of drawing and mentally imagining text contentComputers in Human Behavior 25 284ndash289

Mayer R E (2004) Should there be a three-strikes rule against pure discoverylearning The case for guided methods of instruction The American Psychologist59 14ndash19

Mayer R E (2005) Cognitive theory of multimedia learning In R E Mayer (Ed)The Cambridge handbook of multimedia learning (pp 31ndash48) New York CambridgeUniversity Press

Mayer R E (2009) Multimedia learning (2nd ed) New York NY CambridgeUniversity Press

Mayer R E amp Wittrock M C (2006) Problem solving In P Alexander P Winne ampG Phye (Eds) Handbook of educational psychology (pp 287ndash303) Mahwah NJErlbaum

Paas F (1992) Training strategies for attaining transfer of problem-solving skill instatisticsmdashA cognitive-load approach Journal of Educational Psychology 84429ndash434

Paas F Tuovinen J Tabbers H K amp Van Gerven P W M (2003) Cognitive loadmeasurement as a means to advance cognitive load theory EducationalPsychologist 38 63ndash71

Paivio A (1986) Mental representation A dual coding approach New York OxfordUniversity Press

Pashler H Bain P Bottage B Graesser A Koedinger K McDaniel M et al (2007)Organizing instruction and study to improve student learning Washington DCNational Center for Educational Research

Rasco R W Tennyson R D amp Boutwell R C (1975) Imagery instructions anddrawings in learning prose Journal of Educational Psychology 67 188ndash192

Rheinberg F Vollmeyer R amp Burns B D (2001) FAM Ein fragebogen zurerfassung aktueller motivation in lern- und leistungssituationen [QCM Aquestionnaire to assess current motivation in learning situations] Diagnostica47 57ndash66

Schnotz W (2005) An integrated model of text and picture comprehension In RE Mayer (Ed) The Cambridge handbook of multimedia learning (pp 49ndash70) NewYork Cambridge University Press

Schwamborn A Mayer R E Thillmann H Leopold C amp Leutner D (2010) Drawingas a generative activity and drawing as a prognostic activity Journal of EducationalPsychology 102 872ndash879

Schwamborn A Thillmann H Opfermann M amp Leutner D (2011) Cognitive loadand instructionally supported learning with provided and learner-generatedvisualizations Computers in Human Behavior 27 89ndash93

Shavelson R J amp Towne L (Eds) (2002) Scientific research in education WashingtonDC National Academy Press

Sweller J Ayres P amp Kalyuga S (2011) Cognitive Load Theory New York SpringerTirre W C Manelis L amp Leicht K (1979) The effects of imaginal and verbal strategies

on prose comprehension by adults Journal of Reading Behavior 11 99ndash106van Meter P (2001) Drawing construction as a strategy for learning from text Journal

of Educational Psychology 69 129ndash140van Meter P Aleksic M Schwartz A amp Garner J (2006) Learner-generated drawing

as a strategy for learning from content area text Contemporary EducationalPsychology 31 142ndash166

van Meter P amp Garner J (2005) The promise and practice of learner-generateddrawings Literature review and synthesis Educational Psychology Review 12261ndash312

Van Gog T amp Paas F (2008) Instructional efficiency Revisiting the original constructin educational research Educational Psychologist 43 16ndash26

Vollmeyer R amp Rheinberg F (2000) Does motivation affect learning via persistenceLearning and Instruction 4 293ndash309

Weinstein C E amp Mayer R E (1986) The teaching of learning strategies In M CWittrock (Ed) Handbook of research on teaching (pp 315ndash327) New YorkMacmillan

Wittrock M C (1990) Generative processes of comprehension EducationalPsychologist 24 345ndash376

286 A Schmeck et alContemporary Educational Psychology 39 (2014) 275ndash286

  • Drawing pictures during learning from scientific text testing the generative drawing effect and the prognostic drawing effect
  • Introduction
  • Theoretical framework for the learner-generated drawing strategy
  • Empirical framework for the learner-generated drawing strategy
  • Effectiveness of learner-generated drawings
  • Quality of learner-generated drawings
  • Overview of the experiments
  • Experiment 1
  • Participants and design
  • Materials
  • Procedure
  • Results and discussion
  • Scoring
  • Are the groups equivalent on basic characteristics
  • Is there support for the generative drawing effect
  • Is there support for the prognostic drawing effect
  • Experiment 2
  • Participants and design
  • Materials
  • Procedure
  • Results
  • Scoring
  • Are the groups equivalent on basic characteristics
  • Is there support for the generative drawing effect
  • Is there support for the prognostic drawing effect
  • Discussion
  • Empirical contributions
  • Theoretical contributions
  • Practical contributions
  • Limitations and future directions
  • Acknowledgments
  • References

group completed the task early groups were tested in separateclassrooms

34 Results

341 ScoringAll tests instruments were scored with the same procedures used

in Experiment 1 Again two student assistants (teacher trainees inbiology) scored each of the drawing test items and each of the sevenlearner-generated drawings for each student with acceptableinterrater agreements (drawing test GoodmanndashKruskal gamma of90 drawing-accuracy GoodmanndashKruskal gamma of 94) Actualscores ranged from 1 to 28 points (M = 153 points SD = 58) for thecomprehension test from zero to 21 points (M = 109 points SD = 53)for the drawing test and from 275 to 215 points (M = 141 pointsSD = 48) for drawing accuracy Again total scores of comprehen-sion drawing and accuracy were transferred into z-standardizedscores

342 Are the groups equivalent on basic characteristicsBefore looking at treatment effects on the dependent variables

we analyzed whether the four experimental groups differed onseveral control variables A chi-square analysis indicated that therewere no significant differences regarding gender (p = 097) Sepa-rate univariate analyses of variance (ANOVAs) revealed that thegroups also did not differ significantly on age F lt 1 on spatial abilityF(3 164) = 120 p = 312 or on motivation F(3 164) = 122 p = 305However groups differed significantly on prior knowledge F(3164) = 104 p = 010 partial eta2 = 07 in that students in the controlgroup scored significantly higher on the comprehension pretest(M = 25 SD = 17) than students in both (p lt 05) the author-generated picture group (M = 17 SD = 16) and the drawing + author-generated picture group (M = 15 SD = 12) the drawing group(M = 20 SD = 14) did not differ significantly from the other groupsThus we included studentsrsquo prior knowledge as a covariate in thefollowing analyses

343 Is there support for the generative drawing effectA major goal in this experiment was to determine whether asking

students to generate drawings to represent science text is a moreeffective learning strategy than asking students to learn with textalone or with text and author-generated pictures In other wordswe wanted to determine whether we could replicate and extendthe learner-generated drawing effect Additionally we were inter-ested in whether giving students an author-generated picture afterdrawing can increase the benefits of the learning strategy Mean pro-portion correct and standard deviations on the comprehension anddrawing tests for the four groups are presented in Table 3

The left portion of Table 3 summarizes the mean proportioncorrect on the comprehension test A two-factorial analysis ofcovariance (ANCOVA) predicting learning outcomes (comprehen-sion posttest score) with learner-generated drawing (yesno) and

author-generated picture (yesno) as the factorial independent vari-ables and prior knowledge as a covariate showed a significantpositive main effect of learner-generated drawing F(1 163) = 398p = 048 partial eta2 = 02 a significant interaction effect F(1163) = 626 p = 013 partial eta2 = 04 but no main effect of author-generated pictures F lt 1 In addition multiple pairwise comparisons(with p lt 05) showed that the drawing group performed signifi-cantly better than each of the three other groups which did not differsignificantly from each other Cohenrsquos d favoring the drawing groupover the author-generated picture group was 49 over the learner-generated + author-generated picture group was 57 and over thecontrol group was 52

The right portion of Table 3 summarizes the mean proportioncorrect on the drawing posttest Again a two-factorial analysis ofcovariance (ANCOVA) predicting learning outcome (drawing testscore) with learner-generated drawing (yesno) and author-generatedpicture (yesno) as the factorial independent variables and priorknowledge as a covariate showed a significant positive main effectof learner-generated drawing F(1 163) = 6260 p lt 001 partialeta2 = 28 a significant positive main effect of author-generated pic-tures F(1 163) = 1104 p = 001 partial eta2 = 06 and a significantinteraction effect F(1 163) = 1658 p lt 001 partial eta2 = 09 In ad-dition multiple pairwise comparisons (with p lt 05) showed thatboth the drawing group and the drawing + author-generated picturegroup performed significantly better than the author-generatedpicture group (d = 68 d = 59) and the control group (d = 187d = 188) In turn the author-generated picture group performed sig-nificantly better than the control group (d = 95) The drawing groupand the drawing + author-generated picture group did not differ sig-nificantly from each other (d = 15)2 Overall these results areconsistent with Experiment 1 and provide additional support forthe generative drawing effect

In accordance with Experiment 1 we were interested in whetherdifferences in learning time among the experimental groups mediatethe positive effect of drawing on text comprehension First an ANOVApredicting learning time with learner-generated drawing (yesno)and author-generated picture (yesno) as the factorial indepen-dent variables showed a significant main effect of learner-generateddrawing F(1 164) = 39226 p lt 001 partial eta2 = 71 a significantmain effect of author-generated picture F(1 164) = 1685 p lt 001partial eta2 = 09 and a significant interaction effect F(1 164) = 490p = 028 partial eta2 = 03 Linear contrasts (with p lt 05) revealedthat the drawing group (M = 1938 min SD = 380) and thedrawing + author-generated picture group (M = 2340 min SD = 551)needed significantly more learning time than the author-generatedpicture group (M = 960 min SD = 397) and the control group(M = 834 min SD = 251) Thus to test whether learning time me-diates the positive effect of drawing on text comprehensionadditional mediation analyses (Baron amp Kenny 1986) were calcu-lated by including learning time as an additional predictor in theaforementioned linear model Results of the mediation analysesshowed that the effects of drawing on the comprehension posttestand the drawing posttest scores (see multiple pairwise compari-sons) are mediated by learning time to some extent That is includinglearning time in the linear model for predicting comprehension testscores still revealed a positive effect of the drawing group com-pared with the drawing + author-generated group on thecomprehension test (p = 012) However including learning time inthe linear model for predicting comprehension posttest scoresreduced the positive effect of the drawing group compared with theauthor-generated picture group (from p = 034 to p = 281) as wellas compared with the control group (from p = 002 to p = 087) being

2 The exclusion of studentsrsquo prior knowledge in the reported analyses does notchange the overall pattern of results

Table 3Mean proportion correct on the comprehension test and drawing test for the fourgroups ndash Experiment 2

Group Type of test

n Comprehension test Drawing test

M SD M SD

Learner-generated drawing 40 63 22 66 22Author-generated picture 44 53 19 50 25Learner-generated drawing +

author-generated picture41 51 20 63 19

Control 43 52 20 30 16

Note Asterisk () indicates significant difference from control group at p lt 05

283A Schmeck et alContemporary Educational Psychology 39 (2014) 275ndash286

no longer statistically significant Regarding the drawing posttestscore including learning time does not change the reported patternof results except that the positive effect of the drawing + author-generated picture group compared with the author-generated picturegroup is no longer statistically significant (from p = 004 to p = 223)

There were neither main effects of learner-generated drawingand author-generated pictures on the mental effort item (drawinggroup M = 455 SD = 025 author-generated picture group M = 459SD = 024 drawing + author-generated picture group M = 444SD = 026 control group M = 481 SD = 024 F lt 1) nor on theperceived difficulty item (drawing group M = 363 SD = 023 author-generated picture group M = 371 SD = 022 drawing + author-generated picture group M = 395 SD = 023 control group M = 393SD = 022 F lt 1)

Taken together the drawing strategy apparently fosters stu-dents to engage in generative activities indicated by their higherlearning outcomes Thus the data provide further evidence for thegenerative drawing effect predicted by Schwamborn et al (2010)In Experiment 2 benefits of the drawing activity however are me-diated by learning time and do not involve higher mental effortAdditionally there was no increased benefit when additional drawingsupport was available in the form of author-generated pictures

344 Is there support for the prognostic drawing effectA second major goal of this study was to determine whether the

prognostic drawing effect could be extended to a new context Meanproportion correct on drawing-accuracy during learning was 60(SD = 04) for the drawing group and 68 (SD = 03) for thedrawing + author generated picture group This difference betweenthe two drawing groups is not significant F(1 79) = 252 p = 116This lack of group differences allowed us to pool the data of bothdrawing groups for subsequent correlation analyses Correlation anal-yses based on the combined data from the two drawing groupsrevealed that the drawing-accuracy score of learner-generated draw-ings correlates significantly with the comprehension posttest scorer = 470 p lt 001 as well as with the drawing posttest score r = 615p lt 001 Additional correlation analyses revealed that the drawing-accuracy score of learner-generated drawings did not correlatesignificantly with the prior knowledge test score r = 095 p = 400the spatial ability test score r = 127 p = 257 the motivation testscore r = 033 p = 769 or the mental effort test score r = 042p = 712 The correlation between the drawing-accuracy score andthe perceived difficulty score was only slightly statistical signifi-cance r = minus218 p = 053 Thus the data provide further evidencefor the prognostic drawing effect consistent with the results ofSchwamborn et al (2010)

In sum results of Experiment 2 are partly consistent with theresults of Experiment 1 in that students learn better from a sciencetext when they are asked to draw illustrations representing the mainideas of the text and the quality of the generated drawings duringlearning correlates positively with studentsrsquo text comprehension (egSchwamborn et al 2010 van Meter 2001 van Meter et al 2006)

4 Discussion

41 Empirical contributions

The present set of experiments makes three empirical contri-butions to the field First this study shows strong and consistentevidence that students who are asked to generate drawings (withsufficient support) during reading a scientific text that describes acausal sequence perform better than students who read withoutdrawing both on a comprehension test (d = 085 in Experiment 1and d = 052 in Experiment 2) and on a drawing test (d = 115 in Ex-periment 1 and d = 187 in Experiment 2) Thus the generativedrawing effect can be extended to a new domain and therefore

corresponds to Shavelson and Townersquos (2002 p 4) recommenda-tion to ldquoreplicate and generalize across studiesrdquo as one of the sixessential scientific principles of educational research

Second this study shows strong and consistent evidence that thequality of drawings that students generate during learning with ascientific text that describes a causal sequence is positively relatedto subsequent performance on tests of learning outcome includ-ing a comprehension test (r = 623 in Experiment 1 and r = 470 inExperiment 2) and a drawing test (r = 620 in Experiment 1 andr = 615 in Experiment 2) Thus the prognostic drawing effect canbe replicated and extended to a new domain consistent with stan-dards for scientific research in education prescribed by Shavelsonand Towne (2002)

Third this study shows that asking learners to draw picturesduring reading a scientific text (ie learner-generated drawing groupin Experiment 2) is more effective than simply providing draw-ings (ie author-generated picture group in Experiment 2) both ona comprehension test (d = 049) and a drawing test (d = 068) Sim-ilarly adding author-generated drawings (ie learner-generatedpictures + author-generated pictures group in Experiment 2) doesnot improve the learning outcomes of students who also draw pic-tures during learning (ie learner-generated pictures group inExperiment 2) either on a comprehension test (d = minus057) or adrawing test (d = minus015) In short the act of drawing during learn-ing (with sufficient support) improves learning beyond the simpleprovision of drawings

42 Theoretical contributions

The results are consistent with the idea that drawing during learn-ing serves as a generative activity (Mayer amp Wittrock 2006Schwamborn et al 2010 van Meter amp Garner 2005 Wittrock 1990)That is the act of drawing encourages learners to engage in gen-erative cognitive processing during learning such as organizing therelevant information into a coherent structure and integrating itwith relevant prior knowledge from long-term memory In thepresent study positive effects of drawing were indicated with a com-prehension and a drawing learning outcome test and therefore arein line with the theoretical assumption derived from the GTDC thatbenefits of drawing can be found if learning outcome tests are usedthat are sensitive to the underlying process of drawing (cf van Meteramp Garner 2005) Additionally in our study the drawing activity wassupported in a way that was intended to help learners carry out theunderlying cognitive processes of drawing (ie selecting organiz-ing and integrating) successfully In this regard results of the presentstudy might supplement the theoretical framework of learner-generated drawing by providing further evidence that benefits ofdrawing defined by van Meter and Garnerrsquos GTDC can diminish ifno instructional support is given to constrain and structure thedrawing activity However a fuller understanding of the underly-ing cognitive processes of drawing and how these processes canbe influenced via drawing support requires more direct measuresof cognitive processing during learning Additionally following theidea that metacognitive processes of monitoring and regulation areautomatically activated by drawing (van Meter amp Garner 2005) afuller understanding of the metacognitive effects of drawings is alsorequired

43 Practical contributions

The present study encourages instructional designers and in-structors to incorporate drawing activities into venues involvinglearning from text which we call the generative drawing effect Oneimportant feature of a successful drawing strategy that is presentin this study and in a previous study by Schwamborn et al (2010)is that the drawing activity was supported by providing a

284 A Schmeck et alContemporary Educational Psychology 39 (2014) 275ndash286

background scene and a legend showing how to represent eachelement to constrain and structure the drawing activity Thus animportant practical implication is that students may need supportin their drawing activity so they do not need to draw from scratch

The present study also suggests a potentially useful diagnostictool to gauge the depth of student learning namely the quality ofthe drawings created by students during learning which we referto as the prognostic drawing effect Incorporating a measure of thequality of a learnerrsquos drawing during learning can be a useful toolin developing remedial instruction to give learners individual supportIt may be important to use materials that explain a cause-and-effect process and give learners drawings of the elements they needto represent the process pictorially Asking learners to simply drawpictures of elements is unlikely to be helpful whereas asking themto generate drawings that show the relations among the elementsin a schematic form is more likely to be helpful

44 Limitations and future directions

Some limitations and future directions of our study should beaddressed As noted in the theoretical contributions subsection wedid not have direct measures of cognitive processing during learn-ing so it is not possible to pinpoint how the drawing activity affectedspecific cognitive processes such as attending to relevant informa-tion organizing it and integrating it with prior knowledge We alsodid not assess metacognitive processing during learning thus it isnot possible to pinpoint how the drawing activity affected specificmetacognitive processes such as monitoring and regulation

Furthermore results of the cognitive load rating scales (in-vested mental effort and perceived task difficulty) are inconsistentWhereas in Experiment 1 an effect on mental effort but not on per-ceived task difficulty showed up (ie students in the drawing grouprated their invested mental effort during learning significantlyhigher) no effects on mental effort and task difficulty were foundin Experiment 2 Additionally in both experiments only a nega-tive correlation of perceived task difficulty with the quality of learner-generated pictures appears but no correlation of mental effort withthe quality Following de Jong (2010) those cognitive load ratingscales might have the disadvantages that they do not give a con-current measure of cognitive load and do not measure an essentialconcept in cognitive load theory namely cognitive overload (p 125)Future studies on learner-generated drawing might also use othercognitive load measures such as physiological measures as moredirect indicators of cognitive load

As noted in the practical contributions subsection we showedthe drawing effects by using a scientific text describing how a cause-and-effect system works that is the causal steps regarding aninfection with influenza and the immune response It might be pos-sible however that for other types of text producing drawings mightharm rather than promote text comprehension Thus to test whetherthe reported drawing effects can be extended future research hasto focus on other types of text such as descriptive texts as well ason other types of relations that can be conveyed with other typesof representations such as compare and contrast relations whichcan be shown in a matrix Additionally studentsrsquo learning out-comes were tested immediately after reading thus future work isneeded to investigate the longer-term effects of generative drawingon learning outcomes

Furthermore we only compared drawing with control groupsthat received no further learning strategy instructions However en-gaging in generative learning activities such as drawing requires aconsiderable amount of time Accordingly results showed that forExperiment 2 the positive effect of the drawing group on text com-prehension compared with the author-generated picture group andto the control group was mediated by learning time To rule out thatthe effects of drawing result only from additional time on task instead

of the generative activity future research should also compare thedrawing strategy with other time demanding generative learningstrategies such as summarization (cf Leopold amp Leutner 2012)

Another point that should be noted is that students in both ex-periments received some kind of multimedia materials in that evenwhen they had to draw and did not see presented pictures they wereat least provided with the basic (visual) elements for their draw-ings which they had to do on the given background which thusalso contained information In other words when students are pre-sented with important elements of the drawings which they canuse to draw themselves they will not have to put as much effortinto summarizing visually what they have just read compared withstudents who have to draw without any instructional help Futurestudies might also compare the drawing group with a summariza-tion group in which students receive a set of verbal key terms thatare similar to the drawing elements and are asked to make a textualsummary

Additionally future research is needed to validate the prognos-tic drawing effect So far we know that the quality of learner-generated pictures is related to studentsrsquo learning outcomes (iethe higher the learning outcome the higher the drawing accuracyand vice versa) and their perceived difficulty (ie the lower the per-ceived difficulty the higher the drawing accuracy and vice versa)and that it is not related to studentsrsquo prior knowledge motivationspatial ability or mental effort However less is known about whatthis might mean That is less is known regarding the causal direc-tion of this relation or the presence of a possible further moderatorvariable Do studentsrsquo efforts to produce accurate drawings lead tobetter comprehension and lower perceived difficulty Or do stu-dents who are more adept in drawing benefit more from the strategyand thus perceive the difficulty of the learning materials as beinglower Both arguments seem convincing

Finally more work is needed to determine the level of supportthat makes the drawing strategy most effective for various kinds oflearners As noted in the empirical contribution adding author-generated drawings (ie learner-generated pictures + author-generated pictures group in Experiment 2) does not improve thelearning outcomes of students who also draw pictures during learn-ing and were supported by a drawing prompt In other words thecombination of two ways of supporting the drawing strategy (iegiving a drawing prompt during reading plus an author-generatedpicture after reading) did not improve studentsrsquo learning out-comes compared with students in the drawing group as well ascompared with students in the control and author-generated pic-tures only groups This result is inconsistent with previous research(eg van Meter 2001 van Meter et al 2006) which found that com-paring own drawings to author-generated pictures normally helpslearning van Meter and colleagues (2001 2006) however provid-ed author-generated pictures plus prompting questions after thedrawing process That is students answered prompting questionsto guide the comparison process between their self-generateddrawing and the author-generated drawing In our study studentswere only instructed to generate a drawing to inspect an author-generated one and to check whether their own drawing incomparison with the author-generated one really represented themain ideas of the text paragraph correctly In other words we didnot guide the process of comparing self-generated drawings withauthor-generated ones As a potential consequence students per-formed the intended comparison process inadequately or even notall and thus did not benefit from it One reason for this inade-quate comparison process might be that students need guidancein doing the comparison process Another reason might be the factthat students do not seriously engage in generating drawings oncethey notice that there are author-generated drawings Thus futureresearch should also use additional guidance to test whether thecombination of different ways of supporting the drawing strategy

285A Schmeck et alContemporary Educational Psychology 39 (2014) 275ndash286

(ie giving a drawing prompt during reading plus an author-generated picture after reading) helps learning as well asobservational measures of the drawing process itself (ie think aloudprotocols) to shed more light on the cognitive processes underly-ing the drawing activities

Overall drawing during learning from text appears to be a po-tentially powerful strategy for improving studentsrsquo learning fromscientific text when certain boundaries and prerequisites are takeninto account

Acknowledgments

This article is based on a research project funded by the GermanResearch Foundation (DFG LE 6459-3 as part of FOR 511) We wouldlike to thank Angela Sandmann for her assistance in developing thelearning materials

References

Ainsworth S Prain V amp Tytler R (2011) Drawing to learn in science Science 3331096ndash1097

Alesandrini K L (1981) Pictorial-verbal and analytic-holistic learning strategies inscience learning Journal of Educational Psychology 73 358ndash368

Alesandrini K L (1984) Pictures and adult learning Instructional Science 13 63ndash77Baron R M amp Kenny D A (1986) The moderator-mediator variable distinction in

social psychological research Conceptual strategic and statistical considerationsJournal of Personality and Social Psychology 51 1173ndash1182

Bruumlnken R Plass J L amp Leutner D (2003) Direct measurement of cognitive loadin multimedia learning Educational Psychologist 38 53ndash61

Carney RN amp Levin JR (2002) Pictorial illustrations still improve studentsrsquo learningfrom text Educational Psychology Review 14 5ndash26

de Jong T (2005) The guided discovery principle in multimedia learning In R EMayer (Ed) The Cambridge handbook of multimedia learning (pp 215ndash228) NewYork Cambridge University Press

de Jong T (2010) Cognitive load theory educational research and instructionaldesign Some food for thought Instructional Science 38 105ndash134

Ekstrom R B French J W amp Harman H H (1976) Manual for kit of factor-referencedcognitive tests Princeton NJ Educational Testing Service

Greene T R (1989) Childrenrsquos understanding of class inclusion hierarchies Therelationship between external representation and task performance Journal ofExperimental Child Psychology 48 62ndash89

Hall V C Bailey J amp Tillman C (1997) Can student-generated illustrations be worthten thousand words Journal of Educational Psychology 89 677ndash681

Houmlffler T N (2010) Spatial ability Its influence on Learning with visualizations ndashA meta-analytic review Educational Psychology Review 22 245ndash269

Houmlffler T N Schmeck A amp Opfermann M (2013) Static and dynamic visualrepresentations Individual differences in processing In G Schraw M TMcCrudden amp D Robinson (Eds) Learning through visual displays (pp 133ndash163)Charlotte NC Information Age Publishing

Kalyuga S Chandler P amp Sweller J (1999) Managing split-attention and redundancyin multimedia instruction Applied Cognitive Psychology 13 351ndash371

Leopold C (2009) Lernstrategien und Textverstehen [Learning strategies and textcomprehension] Muumlnster Waxmann

Leopold C amp Leutner D (2012) Science text comprehension Drawing main ideaselection and summarizing as learning strategies Learning and Instruction 2216ndash26

Lesgold A M DeGood H amp Levin J R (1977) Pictures and young childrenrsquos proselearning A supplementary report Journal of Reading Behavior 9 353ndash360

Lesgold A M Levin J R Shimron J amp Guttman J (1975) Pictures andyoung childrenrsquos learning from oral prose Journal of Educational Psychology 67636ndash642

Leutner D Leopold C amp Sumfleth E (2009) Cognitive load and science textcomprehension Effects of drawing and mentally imagining text contentComputers in Human Behavior 25 284ndash289

Mayer R E (2004) Should there be a three-strikes rule against pure discoverylearning The case for guided methods of instruction The American Psychologist59 14ndash19

Mayer R E (2005) Cognitive theory of multimedia learning In R E Mayer (Ed)The Cambridge handbook of multimedia learning (pp 31ndash48) New York CambridgeUniversity Press

Mayer R E (2009) Multimedia learning (2nd ed) New York NY CambridgeUniversity Press

Mayer R E amp Wittrock M C (2006) Problem solving In P Alexander P Winne ampG Phye (Eds) Handbook of educational psychology (pp 287ndash303) Mahwah NJErlbaum

Paas F (1992) Training strategies for attaining transfer of problem-solving skill instatisticsmdashA cognitive-load approach Journal of Educational Psychology 84429ndash434

Paas F Tuovinen J Tabbers H K amp Van Gerven P W M (2003) Cognitive loadmeasurement as a means to advance cognitive load theory EducationalPsychologist 38 63ndash71

Paivio A (1986) Mental representation A dual coding approach New York OxfordUniversity Press

Pashler H Bain P Bottage B Graesser A Koedinger K McDaniel M et al (2007)Organizing instruction and study to improve student learning Washington DCNational Center for Educational Research

Rasco R W Tennyson R D amp Boutwell R C (1975) Imagery instructions anddrawings in learning prose Journal of Educational Psychology 67 188ndash192

Rheinberg F Vollmeyer R amp Burns B D (2001) FAM Ein fragebogen zurerfassung aktueller motivation in lern- und leistungssituationen [QCM Aquestionnaire to assess current motivation in learning situations] Diagnostica47 57ndash66

Schnotz W (2005) An integrated model of text and picture comprehension In RE Mayer (Ed) The Cambridge handbook of multimedia learning (pp 49ndash70) NewYork Cambridge University Press

Schwamborn A Mayer R E Thillmann H Leopold C amp Leutner D (2010) Drawingas a generative activity and drawing as a prognostic activity Journal of EducationalPsychology 102 872ndash879

Schwamborn A Thillmann H Opfermann M amp Leutner D (2011) Cognitive loadand instructionally supported learning with provided and learner-generatedvisualizations Computers in Human Behavior 27 89ndash93

Shavelson R J amp Towne L (Eds) (2002) Scientific research in education WashingtonDC National Academy Press

Sweller J Ayres P amp Kalyuga S (2011) Cognitive Load Theory New York SpringerTirre W C Manelis L amp Leicht K (1979) The effects of imaginal and verbal strategies

on prose comprehension by adults Journal of Reading Behavior 11 99ndash106van Meter P (2001) Drawing construction as a strategy for learning from text Journal

of Educational Psychology 69 129ndash140van Meter P Aleksic M Schwartz A amp Garner J (2006) Learner-generated drawing

as a strategy for learning from content area text Contemporary EducationalPsychology 31 142ndash166

van Meter P amp Garner J (2005) The promise and practice of learner-generateddrawings Literature review and synthesis Educational Psychology Review 12261ndash312

Van Gog T amp Paas F (2008) Instructional efficiency Revisiting the original constructin educational research Educational Psychologist 43 16ndash26

Vollmeyer R amp Rheinberg F (2000) Does motivation affect learning via persistenceLearning and Instruction 4 293ndash309

Weinstein C E amp Mayer R E (1986) The teaching of learning strategies In M CWittrock (Ed) Handbook of research on teaching (pp 315ndash327) New YorkMacmillan

Wittrock M C (1990) Generative processes of comprehension EducationalPsychologist 24 345ndash376

286 A Schmeck et alContemporary Educational Psychology 39 (2014) 275ndash286

  • Drawing pictures during learning from scientific text testing the generative drawing effect and the prognostic drawing effect
  • Introduction
  • Theoretical framework for the learner-generated drawing strategy
  • Empirical framework for the learner-generated drawing strategy
  • Effectiveness of learner-generated drawings
  • Quality of learner-generated drawings
  • Overview of the experiments
  • Experiment 1
  • Participants and design
  • Materials
  • Procedure
  • Results and discussion
  • Scoring
  • Are the groups equivalent on basic characteristics
  • Is there support for the generative drawing effect
  • Is there support for the prognostic drawing effect
  • Experiment 2
  • Participants and design
  • Materials
  • Procedure
  • Results
  • Scoring
  • Are the groups equivalent on basic characteristics
  • Is there support for the generative drawing effect
  • Is there support for the prognostic drawing effect
  • Discussion
  • Empirical contributions
  • Theoretical contributions
  • Practical contributions
  • Limitations and future directions
  • Acknowledgments
  • References

no longer statistically significant Regarding the drawing posttestscore including learning time does not change the reported patternof results except that the positive effect of the drawing + author-generated picture group compared with the author-generated picturegroup is no longer statistically significant (from p = 004 to p = 223)

There were neither main effects of learner-generated drawingand author-generated pictures on the mental effort item (drawinggroup M = 455 SD = 025 author-generated picture group M = 459SD = 024 drawing + author-generated picture group M = 444SD = 026 control group M = 481 SD = 024 F lt 1) nor on theperceived difficulty item (drawing group M = 363 SD = 023 author-generated picture group M = 371 SD = 022 drawing + author-generated picture group M = 395 SD = 023 control group M = 393SD = 022 F lt 1)

Taken together the drawing strategy apparently fosters stu-dents to engage in generative activities indicated by their higherlearning outcomes Thus the data provide further evidence for thegenerative drawing effect predicted by Schwamborn et al (2010)In Experiment 2 benefits of the drawing activity however are me-diated by learning time and do not involve higher mental effortAdditionally there was no increased benefit when additional drawingsupport was available in the form of author-generated pictures

344 Is there support for the prognostic drawing effectA second major goal of this study was to determine whether the

prognostic drawing effect could be extended to a new context Meanproportion correct on drawing-accuracy during learning was 60(SD = 04) for the drawing group and 68 (SD = 03) for thedrawing + author generated picture group This difference betweenthe two drawing groups is not significant F(1 79) = 252 p = 116This lack of group differences allowed us to pool the data of bothdrawing groups for subsequent correlation analyses Correlation anal-yses based on the combined data from the two drawing groupsrevealed that the drawing-accuracy score of learner-generated draw-ings correlates significantly with the comprehension posttest scorer = 470 p lt 001 as well as with the drawing posttest score r = 615p lt 001 Additional correlation analyses revealed that the drawing-accuracy score of learner-generated drawings did not correlatesignificantly with the prior knowledge test score r = 095 p = 400the spatial ability test score r = 127 p = 257 the motivation testscore r = 033 p = 769 or the mental effort test score r = 042p = 712 The correlation between the drawing-accuracy score andthe perceived difficulty score was only slightly statistical signifi-cance r = minus218 p = 053 Thus the data provide further evidencefor the prognostic drawing effect consistent with the results ofSchwamborn et al (2010)

In sum results of Experiment 2 are partly consistent with theresults of Experiment 1 in that students learn better from a sciencetext when they are asked to draw illustrations representing the mainideas of the text and the quality of the generated drawings duringlearning correlates positively with studentsrsquo text comprehension (egSchwamborn et al 2010 van Meter 2001 van Meter et al 2006)

4 Discussion

41 Empirical contributions

The present set of experiments makes three empirical contri-butions to the field First this study shows strong and consistentevidence that students who are asked to generate drawings (withsufficient support) during reading a scientific text that describes acausal sequence perform better than students who read withoutdrawing both on a comprehension test (d = 085 in Experiment 1and d = 052 in Experiment 2) and on a drawing test (d = 115 in Ex-periment 1 and d = 187 in Experiment 2) Thus the generativedrawing effect can be extended to a new domain and therefore

corresponds to Shavelson and Townersquos (2002 p 4) recommenda-tion to ldquoreplicate and generalize across studiesrdquo as one of the sixessential scientific principles of educational research

Second this study shows strong and consistent evidence that thequality of drawings that students generate during learning with ascientific text that describes a causal sequence is positively relatedto subsequent performance on tests of learning outcome includ-ing a comprehension test (r = 623 in Experiment 1 and r = 470 inExperiment 2) and a drawing test (r = 620 in Experiment 1 andr = 615 in Experiment 2) Thus the prognostic drawing effect canbe replicated and extended to a new domain consistent with stan-dards for scientific research in education prescribed by Shavelsonand Towne (2002)

Third this study shows that asking learners to draw picturesduring reading a scientific text (ie learner-generated drawing groupin Experiment 2) is more effective than simply providing draw-ings (ie author-generated picture group in Experiment 2) both ona comprehension test (d = 049) and a drawing test (d = 068) Sim-ilarly adding author-generated drawings (ie learner-generatedpictures + author-generated pictures group in Experiment 2) doesnot improve the learning outcomes of students who also draw pic-tures during learning (ie learner-generated pictures group inExperiment 2) either on a comprehension test (d = minus057) or adrawing test (d = minus015) In short the act of drawing during learn-ing (with sufficient support) improves learning beyond the simpleprovision of drawings

42 Theoretical contributions

The results are consistent with the idea that drawing during learn-ing serves as a generative activity (Mayer amp Wittrock 2006Schwamborn et al 2010 van Meter amp Garner 2005 Wittrock 1990)That is the act of drawing encourages learners to engage in gen-erative cognitive processing during learning such as organizing therelevant information into a coherent structure and integrating itwith relevant prior knowledge from long-term memory In thepresent study positive effects of drawing were indicated with a com-prehension and a drawing learning outcome test and therefore arein line with the theoretical assumption derived from the GTDC thatbenefits of drawing can be found if learning outcome tests are usedthat are sensitive to the underlying process of drawing (cf van Meteramp Garner 2005) Additionally in our study the drawing activity wassupported in a way that was intended to help learners carry out theunderlying cognitive processes of drawing (ie selecting organiz-ing and integrating) successfully In this regard results of the presentstudy might supplement the theoretical framework of learner-generated drawing by providing further evidence that benefits ofdrawing defined by van Meter and Garnerrsquos GTDC can diminish ifno instructional support is given to constrain and structure thedrawing activity However a fuller understanding of the underly-ing cognitive processes of drawing and how these processes canbe influenced via drawing support requires more direct measuresof cognitive processing during learning Additionally following theidea that metacognitive processes of monitoring and regulation areautomatically activated by drawing (van Meter amp Garner 2005) afuller understanding of the metacognitive effects of drawings is alsorequired

43 Practical contributions

The present study encourages instructional designers and in-structors to incorporate drawing activities into venues involvinglearning from text which we call the generative drawing effect Oneimportant feature of a successful drawing strategy that is presentin this study and in a previous study by Schwamborn et al (2010)is that the drawing activity was supported by providing a

284 A Schmeck et alContemporary Educational Psychology 39 (2014) 275ndash286

background scene and a legend showing how to represent eachelement to constrain and structure the drawing activity Thus animportant practical implication is that students may need supportin their drawing activity so they do not need to draw from scratch

The present study also suggests a potentially useful diagnostictool to gauge the depth of student learning namely the quality ofthe drawings created by students during learning which we referto as the prognostic drawing effect Incorporating a measure of thequality of a learnerrsquos drawing during learning can be a useful toolin developing remedial instruction to give learners individual supportIt may be important to use materials that explain a cause-and-effect process and give learners drawings of the elements they needto represent the process pictorially Asking learners to simply drawpictures of elements is unlikely to be helpful whereas asking themto generate drawings that show the relations among the elementsin a schematic form is more likely to be helpful

44 Limitations and future directions

Some limitations and future directions of our study should beaddressed As noted in the theoretical contributions subsection wedid not have direct measures of cognitive processing during learn-ing so it is not possible to pinpoint how the drawing activity affectedspecific cognitive processes such as attending to relevant informa-tion organizing it and integrating it with prior knowledge We alsodid not assess metacognitive processing during learning thus it isnot possible to pinpoint how the drawing activity affected specificmetacognitive processes such as monitoring and regulation

Furthermore results of the cognitive load rating scales (in-vested mental effort and perceived task difficulty) are inconsistentWhereas in Experiment 1 an effect on mental effort but not on per-ceived task difficulty showed up (ie students in the drawing grouprated their invested mental effort during learning significantlyhigher) no effects on mental effort and task difficulty were foundin Experiment 2 Additionally in both experiments only a nega-tive correlation of perceived task difficulty with the quality of learner-generated pictures appears but no correlation of mental effort withthe quality Following de Jong (2010) those cognitive load ratingscales might have the disadvantages that they do not give a con-current measure of cognitive load and do not measure an essentialconcept in cognitive load theory namely cognitive overload (p 125)Future studies on learner-generated drawing might also use othercognitive load measures such as physiological measures as moredirect indicators of cognitive load

As noted in the practical contributions subsection we showedthe drawing effects by using a scientific text describing how a cause-and-effect system works that is the causal steps regarding aninfection with influenza and the immune response It might be pos-sible however that for other types of text producing drawings mightharm rather than promote text comprehension Thus to test whetherthe reported drawing effects can be extended future research hasto focus on other types of text such as descriptive texts as well ason other types of relations that can be conveyed with other typesof representations such as compare and contrast relations whichcan be shown in a matrix Additionally studentsrsquo learning out-comes were tested immediately after reading thus future work isneeded to investigate the longer-term effects of generative drawingon learning outcomes

Furthermore we only compared drawing with control groupsthat received no further learning strategy instructions However en-gaging in generative learning activities such as drawing requires aconsiderable amount of time Accordingly results showed that forExperiment 2 the positive effect of the drawing group on text com-prehension compared with the author-generated picture group andto the control group was mediated by learning time To rule out thatthe effects of drawing result only from additional time on task instead

of the generative activity future research should also compare thedrawing strategy with other time demanding generative learningstrategies such as summarization (cf Leopold amp Leutner 2012)

Another point that should be noted is that students in both ex-periments received some kind of multimedia materials in that evenwhen they had to draw and did not see presented pictures they wereat least provided with the basic (visual) elements for their draw-ings which they had to do on the given background which thusalso contained information In other words when students are pre-sented with important elements of the drawings which they canuse to draw themselves they will not have to put as much effortinto summarizing visually what they have just read compared withstudents who have to draw without any instructional help Futurestudies might also compare the drawing group with a summariza-tion group in which students receive a set of verbal key terms thatare similar to the drawing elements and are asked to make a textualsummary

Additionally future research is needed to validate the prognos-tic drawing effect So far we know that the quality of learner-generated pictures is related to studentsrsquo learning outcomes (iethe higher the learning outcome the higher the drawing accuracyand vice versa) and their perceived difficulty (ie the lower the per-ceived difficulty the higher the drawing accuracy and vice versa)and that it is not related to studentsrsquo prior knowledge motivationspatial ability or mental effort However less is known about whatthis might mean That is less is known regarding the causal direc-tion of this relation or the presence of a possible further moderatorvariable Do studentsrsquo efforts to produce accurate drawings lead tobetter comprehension and lower perceived difficulty Or do stu-dents who are more adept in drawing benefit more from the strategyand thus perceive the difficulty of the learning materials as beinglower Both arguments seem convincing

Finally more work is needed to determine the level of supportthat makes the drawing strategy most effective for various kinds oflearners As noted in the empirical contribution adding author-generated drawings (ie learner-generated pictures + author-generated pictures group in Experiment 2) does not improve thelearning outcomes of students who also draw pictures during learn-ing and were supported by a drawing prompt In other words thecombination of two ways of supporting the drawing strategy (iegiving a drawing prompt during reading plus an author-generatedpicture after reading) did not improve studentsrsquo learning out-comes compared with students in the drawing group as well ascompared with students in the control and author-generated pic-tures only groups This result is inconsistent with previous research(eg van Meter 2001 van Meter et al 2006) which found that com-paring own drawings to author-generated pictures normally helpslearning van Meter and colleagues (2001 2006) however provid-ed author-generated pictures plus prompting questions after thedrawing process That is students answered prompting questionsto guide the comparison process between their self-generateddrawing and the author-generated drawing In our study studentswere only instructed to generate a drawing to inspect an author-generated one and to check whether their own drawing incomparison with the author-generated one really represented themain ideas of the text paragraph correctly In other words we didnot guide the process of comparing self-generated drawings withauthor-generated ones As a potential consequence students per-formed the intended comparison process inadequately or even notall and thus did not benefit from it One reason for this inade-quate comparison process might be that students need guidancein doing the comparison process Another reason might be the factthat students do not seriously engage in generating drawings oncethey notice that there are author-generated drawings Thus futureresearch should also use additional guidance to test whether thecombination of different ways of supporting the drawing strategy

285A Schmeck et alContemporary Educational Psychology 39 (2014) 275ndash286

(ie giving a drawing prompt during reading plus an author-generated picture after reading) helps learning as well asobservational measures of the drawing process itself (ie think aloudprotocols) to shed more light on the cognitive processes underly-ing the drawing activities

Overall drawing during learning from text appears to be a po-tentially powerful strategy for improving studentsrsquo learning fromscientific text when certain boundaries and prerequisites are takeninto account

Acknowledgments

This article is based on a research project funded by the GermanResearch Foundation (DFG LE 6459-3 as part of FOR 511) We wouldlike to thank Angela Sandmann for her assistance in developing thelearning materials

References

Ainsworth S Prain V amp Tytler R (2011) Drawing to learn in science Science 3331096ndash1097

Alesandrini K L (1981) Pictorial-verbal and analytic-holistic learning strategies inscience learning Journal of Educational Psychology 73 358ndash368

Alesandrini K L (1984) Pictures and adult learning Instructional Science 13 63ndash77Baron R M amp Kenny D A (1986) The moderator-mediator variable distinction in

social psychological research Conceptual strategic and statistical considerationsJournal of Personality and Social Psychology 51 1173ndash1182

Bruumlnken R Plass J L amp Leutner D (2003) Direct measurement of cognitive loadin multimedia learning Educational Psychologist 38 53ndash61

Carney RN amp Levin JR (2002) Pictorial illustrations still improve studentsrsquo learningfrom text Educational Psychology Review 14 5ndash26

de Jong T (2005) The guided discovery principle in multimedia learning In R EMayer (Ed) The Cambridge handbook of multimedia learning (pp 215ndash228) NewYork Cambridge University Press

de Jong T (2010) Cognitive load theory educational research and instructionaldesign Some food for thought Instructional Science 38 105ndash134

Ekstrom R B French J W amp Harman H H (1976) Manual for kit of factor-referencedcognitive tests Princeton NJ Educational Testing Service

Greene T R (1989) Childrenrsquos understanding of class inclusion hierarchies Therelationship between external representation and task performance Journal ofExperimental Child Psychology 48 62ndash89

Hall V C Bailey J amp Tillman C (1997) Can student-generated illustrations be worthten thousand words Journal of Educational Psychology 89 677ndash681

Houmlffler T N (2010) Spatial ability Its influence on Learning with visualizations ndashA meta-analytic review Educational Psychology Review 22 245ndash269

Houmlffler T N Schmeck A amp Opfermann M (2013) Static and dynamic visualrepresentations Individual differences in processing In G Schraw M TMcCrudden amp D Robinson (Eds) Learning through visual displays (pp 133ndash163)Charlotte NC Information Age Publishing

Kalyuga S Chandler P amp Sweller J (1999) Managing split-attention and redundancyin multimedia instruction Applied Cognitive Psychology 13 351ndash371

Leopold C (2009) Lernstrategien und Textverstehen [Learning strategies and textcomprehension] Muumlnster Waxmann

Leopold C amp Leutner D (2012) Science text comprehension Drawing main ideaselection and summarizing as learning strategies Learning and Instruction 2216ndash26

Lesgold A M DeGood H amp Levin J R (1977) Pictures and young childrenrsquos proselearning A supplementary report Journal of Reading Behavior 9 353ndash360

Lesgold A M Levin J R Shimron J amp Guttman J (1975) Pictures andyoung childrenrsquos learning from oral prose Journal of Educational Psychology 67636ndash642

Leutner D Leopold C amp Sumfleth E (2009) Cognitive load and science textcomprehension Effects of drawing and mentally imagining text contentComputers in Human Behavior 25 284ndash289

Mayer R E (2004) Should there be a three-strikes rule against pure discoverylearning The case for guided methods of instruction The American Psychologist59 14ndash19

Mayer R E (2005) Cognitive theory of multimedia learning In R E Mayer (Ed)The Cambridge handbook of multimedia learning (pp 31ndash48) New York CambridgeUniversity Press

Mayer R E (2009) Multimedia learning (2nd ed) New York NY CambridgeUniversity Press

Mayer R E amp Wittrock M C (2006) Problem solving In P Alexander P Winne ampG Phye (Eds) Handbook of educational psychology (pp 287ndash303) Mahwah NJErlbaum

Paas F (1992) Training strategies for attaining transfer of problem-solving skill instatisticsmdashA cognitive-load approach Journal of Educational Psychology 84429ndash434

Paas F Tuovinen J Tabbers H K amp Van Gerven P W M (2003) Cognitive loadmeasurement as a means to advance cognitive load theory EducationalPsychologist 38 63ndash71

Paivio A (1986) Mental representation A dual coding approach New York OxfordUniversity Press

Pashler H Bain P Bottage B Graesser A Koedinger K McDaniel M et al (2007)Organizing instruction and study to improve student learning Washington DCNational Center for Educational Research

Rasco R W Tennyson R D amp Boutwell R C (1975) Imagery instructions anddrawings in learning prose Journal of Educational Psychology 67 188ndash192

Rheinberg F Vollmeyer R amp Burns B D (2001) FAM Ein fragebogen zurerfassung aktueller motivation in lern- und leistungssituationen [QCM Aquestionnaire to assess current motivation in learning situations] Diagnostica47 57ndash66

Schnotz W (2005) An integrated model of text and picture comprehension In RE Mayer (Ed) The Cambridge handbook of multimedia learning (pp 49ndash70) NewYork Cambridge University Press

Schwamborn A Mayer R E Thillmann H Leopold C amp Leutner D (2010) Drawingas a generative activity and drawing as a prognostic activity Journal of EducationalPsychology 102 872ndash879

Schwamborn A Thillmann H Opfermann M amp Leutner D (2011) Cognitive loadand instructionally supported learning with provided and learner-generatedvisualizations Computers in Human Behavior 27 89ndash93

Shavelson R J amp Towne L (Eds) (2002) Scientific research in education WashingtonDC National Academy Press

Sweller J Ayres P amp Kalyuga S (2011) Cognitive Load Theory New York SpringerTirre W C Manelis L amp Leicht K (1979) The effects of imaginal and verbal strategies

on prose comprehension by adults Journal of Reading Behavior 11 99ndash106van Meter P (2001) Drawing construction as a strategy for learning from text Journal

of Educational Psychology 69 129ndash140van Meter P Aleksic M Schwartz A amp Garner J (2006) Learner-generated drawing

as a strategy for learning from content area text Contemporary EducationalPsychology 31 142ndash166

van Meter P amp Garner J (2005) The promise and practice of learner-generateddrawings Literature review and synthesis Educational Psychology Review 12261ndash312

Van Gog T amp Paas F (2008) Instructional efficiency Revisiting the original constructin educational research Educational Psychologist 43 16ndash26

Vollmeyer R amp Rheinberg F (2000) Does motivation affect learning via persistenceLearning and Instruction 4 293ndash309

Weinstein C E amp Mayer R E (1986) The teaching of learning strategies In M CWittrock (Ed) Handbook of research on teaching (pp 315ndash327) New YorkMacmillan

Wittrock M C (1990) Generative processes of comprehension EducationalPsychologist 24 345ndash376

286 A Schmeck et alContemporary Educational Psychology 39 (2014) 275ndash286

  • Drawing pictures during learning from scientific text testing the generative drawing effect and the prognostic drawing effect
  • Introduction
  • Theoretical framework for the learner-generated drawing strategy
  • Empirical framework for the learner-generated drawing strategy
  • Effectiveness of learner-generated drawings
  • Quality of learner-generated drawings
  • Overview of the experiments
  • Experiment 1
  • Participants and design
  • Materials
  • Procedure
  • Results and discussion
  • Scoring
  • Are the groups equivalent on basic characteristics
  • Is there support for the generative drawing effect
  • Is there support for the prognostic drawing effect
  • Experiment 2
  • Participants and design
  • Materials
  • Procedure
  • Results
  • Scoring
  • Are the groups equivalent on basic characteristics
  • Is there support for the generative drawing effect
  • Is there support for the prognostic drawing effect
  • Discussion
  • Empirical contributions
  • Theoretical contributions
  • Practical contributions
  • Limitations and future directions
  • Acknowledgments
  • References

background scene and a legend showing how to represent eachelement to constrain and structure the drawing activity Thus animportant practical implication is that students may need supportin their drawing activity so they do not need to draw from scratch

The present study also suggests a potentially useful diagnostictool to gauge the depth of student learning namely the quality ofthe drawings created by students during learning which we referto as the prognostic drawing effect Incorporating a measure of thequality of a learnerrsquos drawing during learning can be a useful toolin developing remedial instruction to give learners individual supportIt may be important to use materials that explain a cause-and-effect process and give learners drawings of the elements they needto represent the process pictorially Asking learners to simply drawpictures of elements is unlikely to be helpful whereas asking themto generate drawings that show the relations among the elementsin a schematic form is more likely to be helpful

44 Limitations and future directions

Some limitations and future directions of our study should beaddressed As noted in the theoretical contributions subsection wedid not have direct measures of cognitive processing during learn-ing so it is not possible to pinpoint how the drawing activity affectedspecific cognitive processes such as attending to relevant informa-tion organizing it and integrating it with prior knowledge We alsodid not assess metacognitive processing during learning thus it isnot possible to pinpoint how the drawing activity affected specificmetacognitive processes such as monitoring and regulation

Furthermore results of the cognitive load rating scales (in-vested mental effort and perceived task difficulty) are inconsistentWhereas in Experiment 1 an effect on mental effort but not on per-ceived task difficulty showed up (ie students in the drawing grouprated their invested mental effort during learning significantlyhigher) no effects on mental effort and task difficulty were foundin Experiment 2 Additionally in both experiments only a nega-tive correlation of perceived task difficulty with the quality of learner-generated pictures appears but no correlation of mental effort withthe quality Following de Jong (2010) those cognitive load ratingscales might have the disadvantages that they do not give a con-current measure of cognitive load and do not measure an essentialconcept in cognitive load theory namely cognitive overload (p 125)Future studies on learner-generated drawing might also use othercognitive load measures such as physiological measures as moredirect indicators of cognitive load

As noted in the practical contributions subsection we showedthe drawing effects by using a scientific text describing how a cause-and-effect system works that is the causal steps regarding aninfection with influenza and the immune response It might be pos-sible however that for other types of text producing drawings mightharm rather than promote text comprehension Thus to test whetherthe reported drawing effects can be extended future research hasto focus on other types of text such as descriptive texts as well ason other types of relations that can be conveyed with other typesof representations such as compare and contrast relations whichcan be shown in a matrix Additionally studentsrsquo learning out-comes were tested immediately after reading thus future work isneeded to investigate the longer-term effects of generative drawingon learning outcomes

Furthermore we only compared drawing with control groupsthat received no further learning strategy instructions However en-gaging in generative learning activities such as drawing requires aconsiderable amount of time Accordingly results showed that forExperiment 2 the positive effect of the drawing group on text com-prehension compared with the author-generated picture group andto the control group was mediated by learning time To rule out thatthe effects of drawing result only from additional time on task instead

of the generative activity future research should also compare thedrawing strategy with other time demanding generative learningstrategies such as summarization (cf Leopold amp Leutner 2012)

Another point that should be noted is that students in both ex-periments received some kind of multimedia materials in that evenwhen they had to draw and did not see presented pictures they wereat least provided with the basic (visual) elements for their draw-ings which they had to do on the given background which thusalso contained information In other words when students are pre-sented with important elements of the drawings which they canuse to draw themselves they will not have to put as much effortinto summarizing visually what they have just read compared withstudents who have to draw without any instructional help Futurestudies might also compare the drawing group with a summariza-tion group in which students receive a set of verbal key terms thatare similar to the drawing elements and are asked to make a textualsummary

Additionally future research is needed to validate the prognos-tic drawing effect So far we know that the quality of learner-generated pictures is related to studentsrsquo learning outcomes (iethe higher the learning outcome the higher the drawing accuracyand vice versa) and their perceived difficulty (ie the lower the per-ceived difficulty the higher the drawing accuracy and vice versa)and that it is not related to studentsrsquo prior knowledge motivationspatial ability or mental effort However less is known about whatthis might mean That is less is known regarding the causal direc-tion of this relation or the presence of a possible further moderatorvariable Do studentsrsquo efforts to produce accurate drawings lead tobetter comprehension and lower perceived difficulty Or do stu-dents who are more adept in drawing benefit more from the strategyand thus perceive the difficulty of the learning materials as beinglower Both arguments seem convincing

Finally more work is needed to determine the level of supportthat makes the drawing strategy most effective for various kinds oflearners As noted in the empirical contribution adding author-generated drawings (ie learner-generated pictures + author-generated pictures group in Experiment 2) does not improve thelearning outcomes of students who also draw pictures during learn-ing and were supported by a drawing prompt In other words thecombination of two ways of supporting the drawing strategy (iegiving a drawing prompt during reading plus an author-generatedpicture after reading) did not improve studentsrsquo learning out-comes compared with students in the drawing group as well ascompared with students in the control and author-generated pic-tures only groups This result is inconsistent with previous research(eg van Meter 2001 van Meter et al 2006) which found that com-paring own drawings to author-generated pictures normally helpslearning van Meter and colleagues (2001 2006) however provid-ed author-generated pictures plus prompting questions after thedrawing process That is students answered prompting questionsto guide the comparison process between their self-generateddrawing and the author-generated drawing In our study studentswere only instructed to generate a drawing to inspect an author-generated one and to check whether their own drawing incomparison with the author-generated one really represented themain ideas of the text paragraph correctly In other words we didnot guide the process of comparing self-generated drawings withauthor-generated ones As a potential consequence students per-formed the intended comparison process inadequately or even notall and thus did not benefit from it One reason for this inade-quate comparison process might be that students need guidancein doing the comparison process Another reason might be the factthat students do not seriously engage in generating drawings oncethey notice that there are author-generated drawings Thus futureresearch should also use additional guidance to test whether thecombination of different ways of supporting the drawing strategy

285A Schmeck et alContemporary Educational Psychology 39 (2014) 275ndash286

(ie giving a drawing prompt during reading plus an author-generated picture after reading) helps learning as well asobservational measures of the drawing process itself (ie think aloudprotocols) to shed more light on the cognitive processes underly-ing the drawing activities

Overall drawing during learning from text appears to be a po-tentially powerful strategy for improving studentsrsquo learning fromscientific text when certain boundaries and prerequisites are takeninto account

Acknowledgments

This article is based on a research project funded by the GermanResearch Foundation (DFG LE 6459-3 as part of FOR 511) We wouldlike to thank Angela Sandmann for her assistance in developing thelearning materials

References

Ainsworth S Prain V amp Tytler R (2011) Drawing to learn in science Science 3331096ndash1097

Alesandrini K L (1981) Pictorial-verbal and analytic-holistic learning strategies inscience learning Journal of Educational Psychology 73 358ndash368

Alesandrini K L (1984) Pictures and adult learning Instructional Science 13 63ndash77Baron R M amp Kenny D A (1986) The moderator-mediator variable distinction in

social psychological research Conceptual strategic and statistical considerationsJournal of Personality and Social Psychology 51 1173ndash1182

Bruumlnken R Plass J L amp Leutner D (2003) Direct measurement of cognitive loadin multimedia learning Educational Psychologist 38 53ndash61

Carney RN amp Levin JR (2002) Pictorial illustrations still improve studentsrsquo learningfrom text Educational Psychology Review 14 5ndash26

de Jong T (2005) The guided discovery principle in multimedia learning In R EMayer (Ed) The Cambridge handbook of multimedia learning (pp 215ndash228) NewYork Cambridge University Press

de Jong T (2010) Cognitive load theory educational research and instructionaldesign Some food for thought Instructional Science 38 105ndash134

Ekstrom R B French J W amp Harman H H (1976) Manual for kit of factor-referencedcognitive tests Princeton NJ Educational Testing Service

Greene T R (1989) Childrenrsquos understanding of class inclusion hierarchies Therelationship between external representation and task performance Journal ofExperimental Child Psychology 48 62ndash89

Hall V C Bailey J amp Tillman C (1997) Can student-generated illustrations be worthten thousand words Journal of Educational Psychology 89 677ndash681

Houmlffler T N (2010) Spatial ability Its influence on Learning with visualizations ndashA meta-analytic review Educational Psychology Review 22 245ndash269

Houmlffler T N Schmeck A amp Opfermann M (2013) Static and dynamic visualrepresentations Individual differences in processing In G Schraw M TMcCrudden amp D Robinson (Eds) Learning through visual displays (pp 133ndash163)Charlotte NC Information Age Publishing

Kalyuga S Chandler P amp Sweller J (1999) Managing split-attention and redundancyin multimedia instruction Applied Cognitive Psychology 13 351ndash371

Leopold C (2009) Lernstrategien und Textverstehen [Learning strategies and textcomprehension] Muumlnster Waxmann

Leopold C amp Leutner D (2012) Science text comprehension Drawing main ideaselection and summarizing as learning strategies Learning and Instruction 2216ndash26

Lesgold A M DeGood H amp Levin J R (1977) Pictures and young childrenrsquos proselearning A supplementary report Journal of Reading Behavior 9 353ndash360

Lesgold A M Levin J R Shimron J amp Guttman J (1975) Pictures andyoung childrenrsquos learning from oral prose Journal of Educational Psychology 67636ndash642

Leutner D Leopold C amp Sumfleth E (2009) Cognitive load and science textcomprehension Effects of drawing and mentally imagining text contentComputers in Human Behavior 25 284ndash289

Mayer R E (2004) Should there be a three-strikes rule against pure discoverylearning The case for guided methods of instruction The American Psychologist59 14ndash19

Mayer R E (2005) Cognitive theory of multimedia learning In R E Mayer (Ed)The Cambridge handbook of multimedia learning (pp 31ndash48) New York CambridgeUniversity Press

Mayer R E (2009) Multimedia learning (2nd ed) New York NY CambridgeUniversity Press

Mayer R E amp Wittrock M C (2006) Problem solving In P Alexander P Winne ampG Phye (Eds) Handbook of educational psychology (pp 287ndash303) Mahwah NJErlbaum

Paas F (1992) Training strategies for attaining transfer of problem-solving skill instatisticsmdashA cognitive-load approach Journal of Educational Psychology 84429ndash434

Paas F Tuovinen J Tabbers H K amp Van Gerven P W M (2003) Cognitive loadmeasurement as a means to advance cognitive load theory EducationalPsychologist 38 63ndash71

Paivio A (1986) Mental representation A dual coding approach New York OxfordUniversity Press

Pashler H Bain P Bottage B Graesser A Koedinger K McDaniel M et al (2007)Organizing instruction and study to improve student learning Washington DCNational Center for Educational Research

Rasco R W Tennyson R D amp Boutwell R C (1975) Imagery instructions anddrawings in learning prose Journal of Educational Psychology 67 188ndash192

Rheinberg F Vollmeyer R amp Burns B D (2001) FAM Ein fragebogen zurerfassung aktueller motivation in lern- und leistungssituationen [QCM Aquestionnaire to assess current motivation in learning situations] Diagnostica47 57ndash66

Schnotz W (2005) An integrated model of text and picture comprehension In RE Mayer (Ed) The Cambridge handbook of multimedia learning (pp 49ndash70) NewYork Cambridge University Press

Schwamborn A Mayer R E Thillmann H Leopold C amp Leutner D (2010) Drawingas a generative activity and drawing as a prognostic activity Journal of EducationalPsychology 102 872ndash879

Schwamborn A Thillmann H Opfermann M amp Leutner D (2011) Cognitive loadand instructionally supported learning with provided and learner-generatedvisualizations Computers in Human Behavior 27 89ndash93

Shavelson R J amp Towne L (Eds) (2002) Scientific research in education WashingtonDC National Academy Press

Sweller J Ayres P amp Kalyuga S (2011) Cognitive Load Theory New York SpringerTirre W C Manelis L amp Leicht K (1979) The effects of imaginal and verbal strategies

on prose comprehension by adults Journal of Reading Behavior 11 99ndash106van Meter P (2001) Drawing construction as a strategy for learning from text Journal

of Educational Psychology 69 129ndash140van Meter P Aleksic M Schwartz A amp Garner J (2006) Learner-generated drawing

as a strategy for learning from content area text Contemporary EducationalPsychology 31 142ndash166

van Meter P amp Garner J (2005) The promise and practice of learner-generateddrawings Literature review and synthesis Educational Psychology Review 12261ndash312

Van Gog T amp Paas F (2008) Instructional efficiency Revisiting the original constructin educational research Educational Psychologist 43 16ndash26

Vollmeyer R amp Rheinberg F (2000) Does motivation affect learning via persistenceLearning and Instruction 4 293ndash309

Weinstein C E amp Mayer R E (1986) The teaching of learning strategies In M CWittrock (Ed) Handbook of research on teaching (pp 315ndash327) New YorkMacmillan

Wittrock M C (1990) Generative processes of comprehension EducationalPsychologist 24 345ndash376

286 A Schmeck et alContemporary Educational Psychology 39 (2014) 275ndash286

  • Drawing pictures during learning from scientific text testing the generative drawing effect and the prognostic drawing effect
  • Introduction
  • Theoretical framework for the learner-generated drawing strategy
  • Empirical framework for the learner-generated drawing strategy
  • Effectiveness of learner-generated drawings
  • Quality of learner-generated drawings
  • Overview of the experiments
  • Experiment 1
  • Participants and design
  • Materials
  • Procedure
  • Results and discussion
  • Scoring
  • Are the groups equivalent on basic characteristics
  • Is there support for the generative drawing effect
  • Is there support for the prognostic drawing effect
  • Experiment 2
  • Participants and design
  • Materials
  • Procedure
  • Results
  • Scoring
  • Are the groups equivalent on basic characteristics
  • Is there support for the generative drawing effect
  • Is there support for the prognostic drawing effect
  • Discussion
  • Empirical contributions
  • Theoretical contributions
  • Practical contributions
  • Limitations and future directions
  • Acknowledgments
  • References

(ie giving a drawing prompt during reading plus an author-generated picture after reading) helps learning as well asobservational measures of the drawing process itself (ie think aloudprotocols) to shed more light on the cognitive processes underly-ing the drawing activities

Overall drawing during learning from text appears to be a po-tentially powerful strategy for improving studentsrsquo learning fromscientific text when certain boundaries and prerequisites are takeninto account

Acknowledgments

This article is based on a research project funded by the GermanResearch Foundation (DFG LE 6459-3 as part of FOR 511) We wouldlike to thank Angela Sandmann for her assistance in developing thelearning materials

References

Ainsworth S Prain V amp Tytler R (2011) Drawing to learn in science Science 3331096ndash1097

Alesandrini K L (1981) Pictorial-verbal and analytic-holistic learning strategies inscience learning Journal of Educational Psychology 73 358ndash368

Alesandrini K L (1984) Pictures and adult learning Instructional Science 13 63ndash77Baron R M amp Kenny D A (1986) The moderator-mediator variable distinction in

social psychological research Conceptual strategic and statistical considerationsJournal of Personality and Social Psychology 51 1173ndash1182

Bruumlnken R Plass J L amp Leutner D (2003) Direct measurement of cognitive loadin multimedia learning Educational Psychologist 38 53ndash61

Carney RN amp Levin JR (2002) Pictorial illustrations still improve studentsrsquo learningfrom text Educational Psychology Review 14 5ndash26

de Jong T (2005) The guided discovery principle in multimedia learning In R EMayer (Ed) The Cambridge handbook of multimedia learning (pp 215ndash228) NewYork Cambridge University Press

de Jong T (2010) Cognitive load theory educational research and instructionaldesign Some food for thought Instructional Science 38 105ndash134

Ekstrom R B French J W amp Harman H H (1976) Manual for kit of factor-referencedcognitive tests Princeton NJ Educational Testing Service

Greene T R (1989) Childrenrsquos understanding of class inclusion hierarchies Therelationship between external representation and task performance Journal ofExperimental Child Psychology 48 62ndash89

Hall V C Bailey J amp Tillman C (1997) Can student-generated illustrations be worthten thousand words Journal of Educational Psychology 89 677ndash681

Houmlffler T N (2010) Spatial ability Its influence on Learning with visualizations ndashA meta-analytic review Educational Psychology Review 22 245ndash269

Houmlffler T N Schmeck A amp Opfermann M (2013) Static and dynamic visualrepresentations Individual differences in processing In G Schraw M TMcCrudden amp D Robinson (Eds) Learning through visual displays (pp 133ndash163)Charlotte NC Information Age Publishing

Kalyuga S Chandler P amp Sweller J (1999) Managing split-attention and redundancyin multimedia instruction Applied Cognitive Psychology 13 351ndash371

Leopold C (2009) Lernstrategien und Textverstehen [Learning strategies and textcomprehension] Muumlnster Waxmann

Leopold C amp Leutner D (2012) Science text comprehension Drawing main ideaselection and summarizing as learning strategies Learning and Instruction 2216ndash26

Lesgold A M DeGood H amp Levin J R (1977) Pictures and young childrenrsquos proselearning A supplementary report Journal of Reading Behavior 9 353ndash360

Lesgold A M Levin J R Shimron J amp Guttman J (1975) Pictures andyoung childrenrsquos learning from oral prose Journal of Educational Psychology 67636ndash642

Leutner D Leopold C amp Sumfleth E (2009) Cognitive load and science textcomprehension Effects of drawing and mentally imagining text contentComputers in Human Behavior 25 284ndash289

Mayer R E (2004) Should there be a three-strikes rule against pure discoverylearning The case for guided methods of instruction The American Psychologist59 14ndash19

Mayer R E (2005) Cognitive theory of multimedia learning In R E Mayer (Ed)The Cambridge handbook of multimedia learning (pp 31ndash48) New York CambridgeUniversity Press

Mayer R E (2009) Multimedia learning (2nd ed) New York NY CambridgeUniversity Press

Mayer R E amp Wittrock M C (2006) Problem solving In P Alexander P Winne ampG Phye (Eds) Handbook of educational psychology (pp 287ndash303) Mahwah NJErlbaum

Paas F (1992) Training strategies for attaining transfer of problem-solving skill instatisticsmdashA cognitive-load approach Journal of Educational Psychology 84429ndash434

Paas F Tuovinen J Tabbers H K amp Van Gerven P W M (2003) Cognitive loadmeasurement as a means to advance cognitive load theory EducationalPsychologist 38 63ndash71

Paivio A (1986) Mental representation A dual coding approach New York OxfordUniversity Press

Pashler H Bain P Bottage B Graesser A Koedinger K McDaniel M et al (2007)Organizing instruction and study to improve student learning Washington DCNational Center for Educational Research

Rasco R W Tennyson R D amp Boutwell R C (1975) Imagery instructions anddrawings in learning prose Journal of Educational Psychology 67 188ndash192

Rheinberg F Vollmeyer R amp Burns B D (2001) FAM Ein fragebogen zurerfassung aktueller motivation in lern- und leistungssituationen [QCM Aquestionnaire to assess current motivation in learning situations] Diagnostica47 57ndash66

Schnotz W (2005) An integrated model of text and picture comprehension In RE Mayer (Ed) The Cambridge handbook of multimedia learning (pp 49ndash70) NewYork Cambridge University Press

Schwamborn A Mayer R E Thillmann H Leopold C amp Leutner D (2010) Drawingas a generative activity and drawing as a prognostic activity Journal of EducationalPsychology 102 872ndash879

Schwamborn A Thillmann H Opfermann M amp Leutner D (2011) Cognitive loadand instructionally supported learning with provided and learner-generatedvisualizations Computers in Human Behavior 27 89ndash93

Shavelson R J amp Towne L (Eds) (2002) Scientific research in education WashingtonDC National Academy Press

Sweller J Ayres P amp Kalyuga S (2011) Cognitive Load Theory New York SpringerTirre W C Manelis L amp Leicht K (1979) The effects of imaginal and verbal strategies

on prose comprehension by adults Journal of Reading Behavior 11 99ndash106van Meter P (2001) Drawing construction as a strategy for learning from text Journal

of Educational Psychology 69 129ndash140van Meter P Aleksic M Schwartz A amp Garner J (2006) Learner-generated drawing

as a strategy for learning from content area text Contemporary EducationalPsychology 31 142ndash166

van Meter P amp Garner J (2005) The promise and practice of learner-generateddrawings Literature review and synthesis Educational Psychology Review 12261ndash312

Van Gog T amp Paas F (2008) Instructional efficiency Revisiting the original constructin educational research Educational Psychologist 43 16ndash26

Vollmeyer R amp Rheinberg F (2000) Does motivation affect learning via persistenceLearning and Instruction 4 293ndash309

Weinstein C E amp Mayer R E (1986) The teaching of learning strategies In M CWittrock (Ed) Handbook of research on teaching (pp 315ndash327) New YorkMacmillan

Wittrock M C (1990) Generative processes of comprehension EducationalPsychologist 24 345ndash376

286 A Schmeck et alContemporary Educational Psychology 39 (2014) 275ndash286

  • Drawing pictures during learning from scientific text testing the generative drawing effect and the prognostic drawing effect
  • Introduction
  • Theoretical framework for the learner-generated drawing strategy
  • Empirical framework for the learner-generated drawing strategy
  • Effectiveness of learner-generated drawings
  • Quality of learner-generated drawings
  • Overview of the experiments
  • Experiment 1
  • Participants and design
  • Materials
  • Procedure
  • Results and discussion
  • Scoring
  • Are the groups equivalent on basic characteristics
  • Is there support for the generative drawing effect
  • Is there support for the prognostic drawing effect
  • Experiment 2
  • Participants and design
  • Materials
  • Procedure
  • Results
  • Scoring
  • Are the groups equivalent on basic characteristics
  • Is there support for the generative drawing effect
  • Is there support for the prognostic drawing effect
  • Discussion
  • Empirical contributions
  • Theoretical contributions
  • Practical contributions
  • Limitations and future directions
  • Acknowledgments
  • References