Upload
others
View
5
Download
0
Embed Size (px)
Citation preview
Seediscussions,stats,andauthorprofilesforthispublicationat:https://www.researchgate.net/publication/282589365
Sensor-AugmentedVirtualLabs:UsingPhysicalInteractionswithScienceSimulationstoPromoteUnderstandingofGas...
ArticleinJournalofScienceEducationandTechnology·July2015
DOI:10.1007/s10956-015-9574-4
CITATIONS
3
READS
206
4authors,including:
Someoftheauthorsofthispublicationarealsoworkingontheserelatedprojects:
BigData:LargeScaleResearchonEngineeringDesignBasedonBigLearnerDatabyaCADToolView
project
NextStepLearning:BridgingScienceEducationandCleantechIndustrywithInnovativeTechnologies
Viewproject
JieChao
TheConcordConsortium
15PUBLICATIONS47CITATIONS
SEEPROFILE
JenniferL.Chiu
UniversityofVirginia
38PUBLICATIONS362CITATIONS
SEEPROFILE
AllcontentfollowingthispagewasuploadedbyJenniferL.Chiuon15October2015.
Theuserhasrequestedenhancementofthedownloadedfile.
Sensor-Augmented Virtual Labs: Using Physical Interactionswith Science Simulations to Promote Understanding of GasBehavior
Jie Chao1 • Jennifer L. Chiu2 • Crystal J. DeJaegher3 • Edward A. Pan4
! Springer Science+Business Media New York 2015
Abstract Deep learning of science involves integrationof existing knowledge and normative science concepts.
Past research demonstrates that combining physical and
virtual labs sequentially or side by side can take advantageof the unique affordances each provides for helping stu-
dents learn science concepts. However, providing simul-
taneously connected physical and virtual experiences hasthe potential to promote connections among ideas. This
paper explores the effect of augmenting a virtual lab with
physical controls on high school chemistry students’understanding of gas laws. We compared students using the
augmented virtual lab to students using a similar sensor-
based physical lab with teacher-led discussions. Resultsdemonstrate that students in the augmented virtual lab
condition made significant gains from pretest and posttest
and outperformed traditional students on some but not allconcepts. Results provide insight into incorporating mixed-
reality technologies into authentic classroom settings.
Keywords Science simulation ! Probeware ! Mixed-reality ! Gas laws ! Kinetic molecular theory ! Tangibleuser interface
Introduction
Science lab experiences in K-12 school settings aim to give
students direct access to natural phenomena and help stu-dents learn the practices of scientists (Lunetta et al. 2007;
National Research Council 2006). Despite extensive use of
physical labs in K-12 science classrooms, researchdemonstrates that students using physical labs alone often
have difficulty developing understanding of complex con-
cepts (Hofstein and Lunetta 2004). Hands-on, physical labsdo not typically provide visualizations or representations of
phenomena, existing at scales too large or small to be
directly observed, which can contribute to student misun-derstanding (Finkelstein et al. 2005; Johnstone 1991). For
example, students may investigate gas laws with a physical
lab and understand that as pressure increases, volumedecreases, but fail to understand the molecular behaviors
that explain why this happens on the molecular level (Liu
2006).Virtual labs that include visualizations and simulations
of scientific phenomena can help students develop richconceptual understanding (Honey and Hilton 2011), espe-
cially for understanding molecular-level phenomena
(Kozma and Russell 1997; Levy and Wilensky 2009).However, research demonstrates that students can focus on
superficial instead of relevant parts of simulations (Lowe
2004) and overestimate their understanding with visual-izations (Chiu and Linn 2012). In addition, a growing body
of research suggests that physical interactions provide
important and consequential contributions to cognition(e.g., Abrahamson et al. 2012; Tolentino et al. 2009).
Combining physical and virtual labs has the potential to
leverage benefits of both kinds of experiences to optimizescience learning. Prior research has explored combining
physical labs and virtual labs in sequence (Olympiou and
Electronic supplementary material The online version of thisarticle (doi:10.1007/s10956-015-9574-4) contains supplementarymaterial, which is available to authorized users.
& Jie [email protected]
1 The Concord Consortium, 25 Love Ln, Concord, MA 01742,USA
2 University of Virginia, Charlottesville, VA, USA
3 University of Notre Dame, Notre Dame, IN, USA
4 Immersion Consulting, Annapolis, MD 21401, USA
123
J Sci Educ Technol
DOI 10.1007/s10956-015-9574-4
Zacharia 2010; Trundle and Bell 2010), side by side
(Blikstein 2014; Jaakkola et al. 2011), embedding virtualphenomena in real, physical space and time (Lui and Slotta
2013; Novellis and Moher 2011), augmenting virtual labs
with haptic technologies (Bivall et al. 2011; Han and Black2011), and enhancing virtual interactions with tangible user
interfaces (Price et al. 2009; Schneider et al. 2013). These
mixed-reality approaches can augment learning by pro-viding multimodal experiences and multiple representa-
tions as well as leveraging students’ haptic interactionskills and spatial reasoning skills. However, relatively few
studies investigate how these kinds of mixed-reality
experiences can be used in authentic classroom settings(Lindgren and Johnson-Glenberg 2013), as many of these
approaches have extensive costs, setup times, or space
requirements that can hinder widespread adoption intoclassrooms.
This paper describes exploratory work using the Frame
(Xie 2012), which provides a streamlined, flexible, andscalable mixed-reality science learning solution using
probeware and sensors already available in many schools.
The Frame allows students to physically interact withcomputer simulations through a variety of sensors (i.e.,
temperature, force, and pressure sensors) that are located
such that students’ actions appear to directly impact thevirtual world. Instead of clicking or touching icons on
computer screen, students interact with the Frame through
physical actions with everyday objects. For example, stu-dents use hot jars or hair dryers to control temperature,
push on a spring to control the force applied to the system,
and a pump to control the number of molecules in thesystem. The purpose of this study is to explore how the
Frame can support learning of complex science topics in
authentic classroom settings and in particular, how theFrame can help students connect molecular-level behaviors
to observable phenomena.
Theoretical Framework
Students bring diverse and rich existing ideas to scienceclassrooms, including intuitive knowledge developed from
everyday experiences, sociocultural beliefs, epistemolo-
gies, and ideas from previous instruction (Bransford et al.2000). Students’ everyday experiences can be used pro-
ductively to build understanding (e.g., Hammer 2000), but
instruction that fails to build upon students’ existing ideasoften results in isolated and incoherent understanding of
science (Linn and Eylon 2011). Complex understanding of
science relies on linking observable phenomena to under-lying molecular-level behaviors (Ozmen 2013), and these
connections are fundamental to many chemistry topics
(Gabel et al. 1987; Snir et al. 2003) as well as physics, life,and earth science concepts (Benson et al. 1993; Bouwma-
Gearhart et al. 2009; Lee et al. 1993; Noh and Scharmann
1997). However, students of all ages have difficulty makingthese links (Bodner 1991; Lin and Cheng 2000; Nakhleh
1992; Novick and Nussbaum 1981). Students often con-
found molecular and macroscopic characteristics and makeincorrect inferences based on everyday macroscopic
experiences (e.g., Ben-Zvi et al. 1986). For robust, deep,
and transferrable learning, new knowledge needs to beintegrated and connected to existing knowledge (Bransford
et al. 2000).To help students link molecular and macroscopic levels,
this study leverages a knowledge integration (KI) learning
perspective (Linn and Eylon 2006). A KI perspective val-ues everyday experiences that students bring to classrooms
and encourages instructional strategies that support stu-
dents to elicit existing ideas, add normative ideas, andprovide opportunities to experiment and sort out their ideas
(Clark 2006; Linn and Eylon 2011). This entails an
effortful and iterative process, in which students developscientific criteria to evaluate their existing ideas and new
scientific ideas, followed by sorting and connecting ideas
into networks that can be effectively utilized in future,novel situations (Linn et al. 2004). Much research
demonstrates how KI-based instruction can help students
develop connected understanding of science (Chiu andLinn 2014; Linn et al. 2004; McElhaney and Linn 2011;
Zhang and Linn 2011).
Combining physical and virtual labs provides novelopportunities to support knowledge integration and science
learning (de Jong et al. 2013). Physical labs provide
effective ways to elicit existing student ideas by physicallyexperiencing the phenomena of interest (Marshall and
Young 2006) and promote mindful scientific inquiry
practices (Chen et al. 2014; Chien et al. 2015). Virtual labscan help students add normative ideas by providing visu-
alizations of abstract and unobservable concepts and pro-
cesses, which are particularly useful for dynamic andmolecular processes (Padilla 2009). Opportunities to
experiment in both physical and virtual settings can help
students sort out their ideas, as going back and forthbetween physical and virtual representations can help stu-
dents discover conflicts between existing and new ideas
and refine understanding of science concepts (Blikstein2014).
Combined Physical and Virtual Labs
To leverage the unique affordances of physical and virtual
labs, researchers have explored various ways to combinephysical and virtual labs to optimize learning. Researchers
have sequentially combined labs for the same concepts
(Akpan and Andre 2000; Chini et al. 2012; Sarabando et al.2014; Trundle and Bell 2010), or different but related
J Sci Educ Technol
123
concepts (e.g., Olympiou and Zacharia 2010), or blended
physical and virtual parts of labs with selected affordancesassociated with certain learning objectives (e.g., Olympiou
and Zacharia 2012), or run labs in parallel such that virtual
and physical labs are available at the same time (e.g.,Jaakkola et al. 2011), or connected physical and virtual labs
that are compared side by side (Blikstein 2014). Overall,
these studies suggest a trend that combining physical andvirtual labs helps students gain greater understanding of
certain science topics compared to either virtual or physicallabs alone (de Jong et al. 2013). In particular, advantages of
combined labs may be more prominent when targeting
complex phenomena because virtual labs help them visu-alize abstract processes and gain appropriate conceptual
mental models (e.g., Zacharia and de Jong 2014).
However, research also suggests that students havedifficulties in making connections between physical and
virtual labs in these combined conditions, particularly for
subjects that involve abstract and molecular concepts. Forexample, McBride et al. (2010) found that undergraduate
physics students were unable to identify correspondence
between a physical magnetic system and its virtual coun-terpart, which was intentionally designed to match the
physical system. Similarly, Liu (2006) found that high
school chemistry students had difficulty explaining gaslaws at the molecular level after completing a sequentially
combined lab that explicitly guided them through molec-
ular visualizations. Additionally, the effects of the orderingof physical and virtual experiences on student outcomes is
relatively mixed, with some suggesting that physical first
has most benefit (e.g., Smith and Puntambekar 2010),whereas others suggest virtual first has more benefit
(Zacharia and de Jong 2014). These mixed results call for
further investigation into how physical and virtual experi-ences can be combined to help students learn complex
science concepts (e.g., de Jong et al. 2013).
Augmenting Virtual Labs
Instead of sequentially or simultaneously combiningphysical labs and virtual labs, researchers have explored
various mixed-reality approaches that augment virtual labs
and simulations. Different than augmented reality, wherereal phenomena are augmented with virtual objects (Mil-
gram and Kishino 1994), augmented virtual experiences
enable students to physically manipulate and interact withrepresentations of very large, small, or abstract phenomena.
Augmented virtual experiences simultaneously connect
physical and virtual experiences and have potential tofacilitate connections among scientific ideas (Lindgren and
Johnson-Glenberg 2013). Gestures or physical manipula-
tion can elicit and build upon students’ everyday knowl-edge (Abrahamson and Lindgren 2014), whereas the
simulation and visualization can help students add and
develop normative understanding of phenomena. Researchsuggests that providing real-world anchors or links even
within purely virtual approaches can benefit science
learning by eliciting students’ tactile experiences. Forexample, Clark and Jorde (2004) used a virtual represen-
tation of a hand touching something hot to help students
understand heat transfer and thermal conductivity. Studentssignificantly benefitted from the ‘‘tactile’’ version of the
simulation as opposed to the simulation without the rep-resentation of the hand. Another successful approach is the
use of ‘‘intuitive metaphors’’ within simulations that use
icons to build upon students’ prior experiences. Forexample, PhET simulations use intuitive representations
such as faucets or switches for variable controls to
implicitly cue students’ existing ideas (Podolefsky et al.2010). These studies provide evidence that actual physical
augmentation of virtual labs holds promise, as physical
objects may elicit an even broader network of existingideas and help students refine connections among everyday
macroscopic ideas and molecular ideas presented in
simulations.One existing approach of augmenting virtual experi-
ences is through haptic technologies. Haptic technologies
allow learners to physically manipulate virtual objects andfeel force and kinesthetic feedback generated by a com-
puter. Haptic augmentation is generally well received by
students and teachers at various educational levels (Joneset al. 2006; Okamura et al. 2002; Tanhua-piiroinen et al.
2010; Williams et al. 2003). Haptic augmentation can also
promote conceptual understanding of microscopic phe-nomena, such as viruses (Minogue and Jones 2009) and
biomolecular binding processes (Bivall et al. 2011), and
force-related topics, such as the Coriolis effect (Kim et al.2011) and simple machines (Han and Black 2011). For
example, in Han and Black’s study, students manipulated a
simulation of gear system through a joystick, which pro-vided force and kinesthetic feedback similar to a physical
gear system.
Similar to haptic augmentation, tangible augmentationuses objects or human body movements as input devices
for simulations. Instead of using a mouse and keyboard,
students explore simulated scientific phenomena bymanipulating physical objects, gesturing, and whole-body
movements (Price et al. 2008). For example, with a tan-
gible tabletop of optics, students arrange a flashlight andcolored blocks on a semitransparent surface that displays
the behaviors of light in the arrangement (Price and Falcao
2011). Although many claims have been made about thepotential advantages of tangible user interfaces to enhance
science learning (Antle and Wise 2013), only a few studies
have compared these interfaces against alternative learningmedia. For example, learners using a tangible interface to
J Sci Educ Technol
123
study sustainable infrastructure design converged on solu-
tions faster and collaborated more compared to learnersusing a mouse-based interface (Shelley et al. 2011). A
recent crossover experiment showed that students learned
more with a tangible tabletop of the human optical systemthan with text of the same topic, and the advantages of the
tangible tabletop persisted after students were exposed to
both learning media (Schneider et al. 2013). Results sug-gest that tangible augmentation of simulations can benefit
learning.Othermixed-reality approaches use physical augmentation
of large, classroom-sized simulations to embed or immerse
studentswithin phenomena (e.g., Tolentino et al. 2009).Thesekinds of embedded (Moher 2006) or embodied (Johnson-
Glenberg et al. 2014) approaches leverage physical time and
space to distribute simulations across a classroom. Studentsinteract with the simulations through gestures and movement
or can experience and take observations of phenomena that
happen in their classroom over a period of time (Moher 2006).Research suggests that students in these kinds of embedded
approaches outperform students with regular instruction
(Johnson-Glenberg et al. 2014) as well as non-embeddedvirtual approaches (Moher et al. 2010).
Although these studies demonstrate the potential of
augmented virtual approaches in classrooms, existingresearch provides mixed results for haptic augmentation
(e.g., Minogue and Borland 2012; Wiebe et al. 2009). Most
haptic or tangible augmentation experiences only allowmanipulation of and feedback from one or two variables,
which limits the ways in which students can physically
interact with the simulation. For example, students arecommonly limited to changing the position and applying
varying amount of force to the simulated objects (Bivall
et al. 2011; Han and Black 2011). Accordingly, due tothese technological limitations, research with haptic devi-
ces focuses on relatively simple conceptual domains—ei-
ther macroscopic phenomena or microscopic phenomenaalone, but rarely a combination of both. Additionally, many
of the immersive or embedded augmentation approaches
require relatively extensive setup for regular classrooms.Investigating different ways of augmenting virtual experi-
ences can potentially help students make connections
across levels of complex phenomena and provide moreaccessible approaches for teachers.
Frame Technology: Sensor-Augmented Virtual Labs
The Frame technology (Xie 2012) offers another way to
physically augment virtual simulations. Different from thetypical location-sensor-based tangible user interface, the
Frame utilizes a computer simulation and multiple fast-re-
sponse sensors that constantly collect and send data to the
simulation to change its parameters. Thus, students interact
with the computer simulations from the edges of the monitorthat makes it look like they directly impact the virtual world.
The use of sensors enables students to interactwith the Frame
through a variety of physical objects, such as hair dryers, hotjars, or pushing on springs (Fig. 1).
The Frame technology differs from existing applications
and past research in several ways. First, instead of usinghaptic devices such as a glove or joystick or visual images
of a flame or pump (e.g., PhET), the Frame uses actualphysical objects to augment students’ experiences with a
virtual lab. Instead of a virtual icon of a pump, students add
molecules to a simulation through a physical, real-worldpump, or increase the temperature of the simulation
through the use of a hot jar or a hair dryer. Thus, interac-
tions with these existing everyday objects may help bringforth existing, experiential ideas. Second, the Frame allows
students to physically manipulate multiple variables of the
system instead of only one single variable typicallyallowed in haptic simulations. For example, the Gas Frame
enables students to change the temperature, number of
molecules, and force on a container of gas and see resultantbehavior of gas molecules. Third, the Frame uses sensors
that secondary science teachers typically use for existing
physical labs, connects similarly as existing physical labs(through a USB), and freely available simulations making
adoption potentially easier for teachers.
This paper investigates the potential of the Frame to helpstudents in authentic classrooms connect molecular-level
behaviors to observable phenomena. The Frame differs from
past mixed-reality approaches by simultaneously connectingsimulations and real-world physical controls along several
variables. Past studies have demonstrated that the Frame can
help middle school students make connections among ideasand refine alternative ideas about gases (Chiu et al. 2015).
However, this study only explored student changes from
pretest to posttest without any kind of comparison group.This study compares students using the Frame to students
learning about the same topic using a traditional sensor-
based physical lab with accompanying instruction to inves-tigate what, if any, benefits the sensor-augmented approach
may have for student learning. Specifically, we addressed the
following research questions:
1. Do students working with the Frame improve their
understanding of gas behavior?2. Do students working with the Frame outperform
students learning with traditional instruction and a
physical lab on measures assessing understanding ofgas behavior?
3. If they do, what do students working with the Gas
Frame learn in addition to what they might havelearned from traditional instruction?
J Sci Educ Technol
123
Materials and Methods
The Gas Frame
The Gas Frame was designed to help students learn the gas
laws and kinetic molecular theory (KMT), as these are cen-
tral topics to physical science (NGSS Lead States 2013). Thegas laws describe the relationships between quantity, vol-
ume, temperature, and pressure of a gas; KMT explains these
relationships based on behaviors of ideal gas molecules.Students typically have difficulty understanding these con-
cepts (e.g., Ozmen 2013). Students can make incorrect
inferences based on existing knowledge. For example, manystudents believe that gas molecules in a fixed container will
get closer together when cooled (Nurrenbem and Pickering
1987; Sanger and Phelps 2007)—amisapplication of the ideathat hot things expand and cold things shrink. Also, a deep
understanding of gas laws and KMT requires complex con-
nections among macroscopic properties and molecularproperties (Fig. 2). Students typically do not make these
links, resulting in heavy reliance on algorithmic techniques
rather than qualitative reasoning to solve problems (Sangeret al. 2000). Students often develop isolated ideas about gas
laws in high school (Robins et al. 2009) that persist through
undergraduate and graduate studies (Gabel et al. 1984).The Frame technology provides a novel approach to
help students build upon their existing ideas and make
connections between molecular and macroscopic levels.The physical interactions with real-world objects can
effectively bring students’ existing ideas forward. The
visual simulation not only provides visualizations of
associated molecular behavior but also incorporates mul-tiple concepts in a single system and dynamically presents
the relationships between them. In the Gas Frame (Fig. 1),
the simulation displays molecules of an ideal gas enclosedin a container with a frictionless piston. A temperature
sensor, a pressure sensor, and a force sensor are connected
to the USB port of the tablet computer and housed underthe tablet in a custom-made box. The temperature sensor
registers the temperature of its surroundings, and the sim-
ulation uses this data as the temperature of the simulatedgas. Students can move a jar of hot water to touch the
temperature sensor tip, creating a scenario as if energy
transfers from the hot jar to the gas molecules in the virtualworld. The pressure sensor is connected to a syringe
(typically used in sensor-based labs), which acts like an air
pump. As students press or pull the syringe plunger, thesimulation responds to the detected pressure increase or
decrease with injecting or removing gas molecules through
the nozzle on the left side of the screen. The force sensor isconnected with a spring, which acts like an extension of the
piston pole in the simulation. As students press or pull the
spring, the simulation uses the force data to adjust theexternal pressure on the virtual piston and make it move
correspondingly. Meanwhile, numerical outputs of these
data (i.e., temperature, pressure, volume, and quantity ofthe gas) are displayed at the bottom right corner of the
screen.
Fig. 1 Frame apparatus (topleft) with a molecule ‘‘pump’’on the left, spring pistoncontroller on the right, and a jarfor holding hot/cold waterplaced next to the temperaturesensor. Students interact withthe visualization (top right)through physical controls(bottom)
J Sci Educ Technol
123
Research Design
We employed a quasi-experimental design to compare how
the Gas Frame activity and traditional instruction with a sen-
sor-based lab affected high school chemistry students’understanding of the gas laws and kinetic molecular theory.
Because this was an exploratory pilot study focused on the
implementation of Frame technology and curriculum in realclassroom settings, we made several research design choices
to ensure ecologically valid feedback for the development
team. First, wewanted to compare students using the Frame tohigh-quality instruction using sensor-based physical labs.We
chose a teacher with known experience conducting successful
sensor-based physical labs in her classroom. Second, we drewupon research-based best practices with both physical and
virtual labs to create a curriculum to accompany the Frame
technology, such as guided inquiry (Hmelo-Silver et al. 2007);instructional patterns such as predict–observe–explain (White
andGunstone 1992); and helping students to observe relevant,instead of superficial, details of simulations (Lowe 2004).
Thus, our research design aimed to compare high-quality
traditional lab instruction to our best guess at high-qualityaugmented virtual lab instruction. Third, to maintain ecolog-
ical validity and minimize disruption to the classes, we
assigned students to the two conditions at the class levelinstead of at the individual level. Fourth, two researchers were
present during the Gas Frame lab to provide technical assis-
tance and conduct field observations.
Participants and Context
A total of thirty students were recruited from two chemistry
classes in a public high school in a mid-Atlantic state. All
students and their parents signed consent forms, and they
reserved the right to terminate their participation at any
time. Most of the students were in 10th or 11th grade. Theirdemographic background was representative of the school:
49.2 % female, 30.4 % ethnic minority. The teacher had
over 10 years of teaching experience, with advanceddegrees in both chemistry and education. Both classes were
honor classes. One class with 16 students was assigned to
the Frame group, and the other class with 14 students to thetraditional group.
Interventions
Both groups studied the same content (Table 1), with the
teacher’s guidance, in two consecutive 90-min class peri-ods as an introduction to the 2-week gas laws unit. In the
lab sessions, both conditions worked in small groups (two
or three students per group) and the teacher circulated toprovide instructional support.
The Frame group used the Gas Frame with a paper-based
guide (ESM Appendix 1) to help students investigate thebehaviors of gas molecules, gas temperature, gas pressure,
Boyle’s law,Charles’ law, the relationship betweenmolecular
mass and pressure, and partial pressure. The lab curriculumused a predict–observe–explain–reflect instructional pattern
(e.g., Tien et al. 2007) to support students through the processof knowledge integration (Linn et al. 2004). The Frame lab
also featured a progression from structured inquiry to guided
inquiry, gradually removing scaffolds for experimental designand making observations as students progressed to more
advanced investigations. The first author and another
researcher provided technical support and observed theclassroom implementation.
Fig. 2 Conceptual network ofkinetic molecular theory and gaslaws
J Sci Educ Technol
123
The traditional group used the first class period to
investigate Boyles’ law using a sensor-based physical labthat included similar physical object such as an enclosed
gas syringe, a gas pressure sensor, a computer, and a
computer interface for data collection and data analysis.The physical lab guide (ESM Appendix 2) engaged stu-
dents in a structured inquiry of Boyles’ law involving an
introduction to the computer interface, setting up and cal-ibrating the sensors, data collection and analysis proce-
dures with an emphasis on the mathematical relationship
between variables, drawing conclusions, and making fur-ther predictions. In the second class period, the teacher
guided students to explain various common gas phenomena
using KMT in a whole-class discussion and accompanying
worksheets using various real-life contexts as anchors
(ESM Appendix 3). No researcher was present in theclassroom.
Data Collection
Pre-/post-assessments (ESM Appendix 4) were tailor-made
for this particular study due to the lack of standardized teststhat assess integrated understanding of the gas laws and
KMT. Typical standardized test items only assess the gas
laws and KMT separately (see Appendix 5 in ESM forsample test items from two representative states). For
example, students are asked to apply the gas laws to make
predictions for a gas system (e.g., ESM Appendix 5,
Table 1 Comparison of the Gas Frame activity and traditional instruction
Gas Frame activity Traditional instruction
Contents
Behaviors of gasmolecules
Observe collisions, velocity, and energy of gas molecules (Covered within the context of the other activities)following exercises
Temperature Heat or cool the gas and observe how gas molecules behavein different temperatures. Relate your findings to gasmolecules in cold or hot rooms
Class discussion: How does a hot air balloon work?
Worksheet: In a mixture of gases, which gas has greatestkinetic energy? Explain your answer
In a mixture of gases, which gas effuses with the greatestvelocity? Explain your answer
Pressure n–P Add molecules and hold volume constant, observe hownumber of gas molecules affects pressure
Class discussion: How does drinking through a strawwork?
V–P (Boyle’slaw)
Follow given method, collect data, draw graph, explainfinding at the molecular level
Sensor-based lab: Boyle’s law
Class discussion: How does breathing work? What effectdoes altitude have on baking cakes? What differences arethere in tennis balls in Colorado versus Virginia?
T–V (Charles’law)
Create your own method, collect data, draw graph, explainfinding at the molecular level
Class discussion: How does a hot air balloon work?
ma–P (Doesmolecular massaffect pressure?)
Predict and explain how pressure of heavy and light gasmolecules compare, create method, collect data, analyzedata, explain finding at the molecular level
Worksheet: In a mixture of gases, which gas has greatestkinetic energy? Explain your answer
In a mixture of gases, which gas effuses with the greatestvelocity? Explain your answer
In a mixture of gases, which gas exerts greatest pressure?Explain your answer
Partial pressurelaw
Evaluate a statement about partial pressure and explain yourjudgment. Collect method, collect data, analyze data,explain finding at the molecular level
Worksheet: Calculate the partial pressure of each gas in amixture of gases
Context
Time on task Two 90-min periods Two 90-min periods
Mode of learning Lab: self-directed, small group, collaborative Lab: self-directed, small group, collaborative
Other: individual practice with worksheet, whole-classdiscussion
Role of teacher Monitoring lab safety and progress Monitoring lab safety and progress
Guiding students to integrate the gas laws and KMT byinteracting with the Gas Frame
Modeling how to integrate the gas laws and KMT bydiscussing real-life examples
The Gas Frame activity is in sequential order, whereas the Traditional instruction did the sensor-based V–P lab on the first day and had the classdiscussion with explanation prompts on the second day
J Sci Educ Technol
123
Question 9), or they are asked to identify the statement that
correctly describes molecules of an ideal gas (e.g., ESMAppendix 5, Question 3). In test items that do attempt to
assess the integrated understanding, the assessed knowl-
edge is often limited to a single connection between amacroscopic property and a molecular property (e.g., ESM
Appendix 5, Question 7). Such existing test items were
insufficient to assess students’ ability to connect multipleconcepts at both the macroscopic and molecular levels.
In order to capture integrated understanding of the gaslaws and KMT, we created an assessment that consisted of
nine questions asking students to make predictions and/or
provide explanations for phenomena related to gas pres-sure, gas temperature, Gay-Lussac’s law (T–P), the rela-
tionship between quantity and pressure (n–P), Boyle’s law
(V–P), Avogadro’s law (n–V), the relationship betweenmolecular mass and pressure (ma–P), and the partial pres-
sure law. All of these questions gauged students’ ability to
connect multiple concepts at both macroscopic andmicroscopic levels. For example, Question 3 asks students
to provide a molecular explanation for why the pressure of
a gas in a fixed container would increase as its temperatureincreases. The proper explanation requires students to
make connections among the temperature of gas, the
kinetic energy and velocity of gas molecules, the frequencyand impulse of the collisions between gas molecules and
the walls of the container, and the pressure of gas.
Some of these questions were selected or adapted fromexperimental assessments in previous research (e.g., Lin
and Cheng 2000; Sanger and Phelps 2007), and others were
designed by the authors of this article. All questions werereviewed by subject matter experts, chemistry teachers, and
educational measurement specialists and had been itera-
tively validated using a sample of over 200 high schoolchemistry students in the same region. The average com-
pletion time was 25 min.
All students took the pretest 3 days before the inter-vention and the posttest within a week after the interven-
tion. The tests were in a paper-and-pencil format and
administrated by the teacher during class sessions.
Data Analysis
Two raters independently scored students’ test responses
using a knowledge integration assessment framework (Liu
et al. 2008) to capture the connections among students’ideas concerning gas laws and kinetic molecular theory.
Students’ responses were first analyzed to identify and
categorize discrete ideas as (1) normative, which is bothscientifically correct and contributes to the ideal response,
(2) alternative, which is incorrect as judged by scientific
norms or does not contribute to the ideal response, and (3)irrelevant ideas including vague statements, violation of
given conditions, misunderstanding the questions, etc.
Then, scores (0–5) were assigned to the responses based onthe number of normative, alternative, and irrelevant ideas
and connections among them (Table 2). Cohen’s Kappa
ranged from 0.689 to 0.951 for coding discrete ideas andfrom 0.695 to 0.919 for assigned scores, both indicating
substantial interrater reliability (Cohen 1968).
Paired-sample t tests were performed for the Framegroup to evaluate overall and item-level gains. ANCOVA
was performed to compare the effects of the two conditionson posttest scores controlling for pretest scores. For each
normative idea, proportions of students using it in pretest
and posttest were calculated for each group, and McNe-mar’s tests were performed to determine whether the pro-
portions increased for each group.
Results
Research Question 1: Do Students Workingwith the Frame Improve Their Understandingof Gas Behavior?
Paired-sampled t tests (Table 3) indicated that the Frame
group significantly improved their overall test performancewith a large effect size. Their performance on all the
individual questions significantly improved with medium
to large effect sizes, except for question 6 and question 9.In general, the results suggest that the Gas Frame helped
students improve their understanding of gas laws and
kinetic molecular theory as evidenced by significantincreases in scores on items that assessed understanding of
pressure, temperature, the relationship between tempera-
ture and pressure, the number of molecules and pressure,and the relationship between mass of particles and
pressure.
Further analysis across questions suggested that ques-tions 6 and 9 might have failed to elicit desired aspects of
students’ scientific knowledge. Question 6 was designed to
assess students’ understanding of the relationship betweenquantity and pressure and the relationship between volume
and pressure. A desirable answer would use the frequency
of collisions between gas molecules and container walls toexplain the two relationships. Analysis of the posttest
responses revealed that only 1 of the 16 students included
collision frequency in her or his response. Instead, most ofthem attributed pressure change either directly to density
change or some unscientific reasons such as ‘‘more gas
molecules compete for the same space,’’ ‘‘gas moleculeshave nowhere to go,’’ or ‘‘gas molecules move less freely
in a smaller space.’’ However, their responses to other
pressure-related test items showed understanding of therelationship between pressure and collisions. For instance,
J Sci Educ Technol
123
the majority of the students (11 out of 16) stated collisions
as the cause of pressure in their responses to question 1.
Over half (9 out of 16) used the concepts of collision
frequency and/or collision impulse to explain the rela-
tionship between temperature and pressure in their
responses to question 3. Similarly, half (8 out of 16) used
Table 2 Knowledge integration rubric adapted from Liu, Lee, Hofstetter, and Linn (2008) and sample student responses
KI level Score Response characteristics Sample student responses for Q3, whichasks students to explain the Gay-Lussac’s law at the molecular level
Scoring criteria for Q3
Complexlink
5 Students understand how more thantwo science concepts interact in agiven context
When the container was hot, theparticles had a faster average speed.So when the particles bounced againstthe walls they exerted more force, andbecause they’re moving faster, theybounce against the walls morefrequently. This more forceful andfrequent bouncing of the particlesincreases the pressure
Normative links between the gastemperature, the kinetic energy orvelocity of gas molecules, thefrequency AND impulse of collisionsbetween gas molecules and containerwalls, and the gas pressure
Full link 4 Students understand how twoscientific concepts interact in a givencontext
The gas molecules got more energywhen it got warmer so they pushedwith more force against the wallscreating more pressure
Normative links between the gastemperature, the kinetic energy orvelocity of gas molecules, thefrequency OR impulse of collisionsbetween gas molecules and containerwalls, and the gas pressure
Partiallink
3 Students recognize potentialconnections between concepts butcannot elaborate the nature of theconnections specific to a givencontext
When the gas inside the sealedcontainer got hotter the gas moleculeswere moving faster and thereforeexerting more pressure on the sides ofthe container
Normative links between the gastemperature, the velocity of gasmolecules, and the gas pressure
No link 2 Students have non-normative ideasand/or make scientifically invalidlinks in a given context
As the temperature heated throughoutthe afternoon, the gas moleculesbegan to expand, causing them topush outward, resulting in increasedpressure
Non-normative idea linking the gastemperature and gas pressure
Off task 1 Students make statements about non-scientific situations
The gas molecules start to increase thespeed at which they are moving whichmakes it feel warmer
The response does not explain the Gay-Lussac’s law
Noanswer
0 Blank or ‘‘I don’t know’’ I don’t know Blank or ‘‘I don’t know’’
Table 3 Paired-sample t tests between pretest and posttest scores of students working with the Frame
Question PretestMean (SD)
PosttestMean (SD)
DifferenceMean (SD)
t One-tailedp value
Effect size (Cohen’s d)
1. Pressure 2.19 (0.403) 3.31 (0.873) 1.13 (0.885) -5.084 \0.001** 1.27
2. Temperature 2.31 (0.602) 2.69 (0.602) 0.38 (0.619) -2.423 0.015* 0.61
3. T–P (1) 2.69 (0.946) 4.19 (0.981) 1.50 (1.317) -4.557 \0.001** 1.14
4. T–P (2) 2.38 (1.088) 3.00 (1.265) 0.63 (1.258) -1.987 0.033* 0.50
5. T–P (3) 2.00 (1.095) 2.62 (0.957) 0.63 (0.806) -3.101 0.004* 0.78
6. N–P and/or V–P 2.00 (0.000) 2.12 (0.500) 0.13 (0.500) -1.000 0.167 0.25
7. n–V 2.06 (0.680) 2.81 (0.655) 0.75 (0.775) -3.873 0.001* 0.97
8. ma–P 2.00 (1.095) 2.81 (1.328) 0.81 (1.424) -2.282 0.019* 0.57
9. Partial pressure 1.50 (1.095) 1.56 (1.094) 0.06 (1.692) -0.148 0.442 0.04
Total 19.13 (4.209) 25.13 (3.538) 6 (3.864) -6.211 \0.001** 1.55
* p\ 0.05; ** p\ 0.001
J Sci Educ Technol
123
these concepts to explain why molecular mass had no
effect on pressure in their responses to question 8. Thus, itis possible that the students understood the molecular
nature of pressure, but question 6 failed to elicit this piece
of knowledge. Their responses reflected an intuitive asso-ciation between the pressure and density of gas, which
could have hindered the activation and application of the
scientific concepts.Question 9 was designed to assess students’ under-
standing of partial pressure. The desirable response wouldexplain that the total pressure of a mixture of gases is the
sum of the individual gases and make connections to
molecular gas behavior. Analysis of the posttest responsesshowed that 4 out of the 16 students did not know how to
determine the partial pressure of component gas given the
total pressure and mole fractions. Three of them were in thesame lab group, and they incorrectly concluded that partial
pressure is not in proportion to mole fraction based on
incorrect data. Six other students used the change in overallquantity of gas to obtain the partial pressure of component
gas—an imprecise way to represent the partial pressure law
thus was not given credit. It is possible that the studentsunderstood the partial pressure law but did not feel the need
to formally apply it in this situation.
Research Question 2: Do Students Workingwith the Frame Outperform Students ReceivingTraditional Instruction on Measures AssessingUnderstanding of Gas Behavior?
As shown in (Table 4, ANCOVA indicated no significantdifference between the Frame group and the traditional
group in their posttest total scores, controlling for their
pretest total scores. While the two groups performedequally well on six of the nine questions, the Frame group
significantly outperformed the traditional group on the
other three questions with medium to large effect sizes,controlling for their pretest scores. These three questions
assessed students understanding of pressure (Q1), the
relationship between temperature and pressure (Q3), andthe relationship between the number of molecules and
volume (Q7), respectively.
It should be noted that the two groups performed equallywell on two other questions also targeting a molecular
explanation of the relationship between temperature andpressure (Q4 and Q5). This inconsistency across items may
be attributed to different requirements of the questions.
Questions 4 and 5 asked students to predict and explainchanges in pressure given certain changes in temperature
without specifically asking for molecular explanations as in
question 3. As a result, students in both groups only usedmacroscopic, law-based explanations (e.g., ‘‘as temperature
increases, pressure will also increase’’) to make their pre-
dictions and explanations without discussing the behaviorsof the gas molecules.
Research Question 3: If They Do, What Do TheyLearn in Addition to What They Might HaveLearned from Traditional Instruction?
The Frame group outperformed the traditional group on
questions 1, 3, and 7 on the posttest, controlling for pretest
performance. Below are analyses of normative ideas shownin pre- and post-responses, by experimental condition.
Question 1 assessed understanding of the molecular
nature of pressure by asking for an explanation of how gasmolecules in an enclosed container (e.g., an air mattress)
cause pressure. The desirable explanation involves two
connected ideas: (1) the constant motion of gas moleculesresults in collisions with the container walls (motion–
Table 4 ANCOVA comparing the Gas Frame lab (Frame) to sensor-based physical labs (Traditional) on posttest scores controlling for pretestscores
Questions Traditional adjusted mean (SE) Frame adjusted mean (SE) F statistic p value Effect size (x2)
1. Pressure 2.68 (0.235) 3.41 (0.220) 5.055 0.033* 0.100
2. T-distribution 2.80 (0.156) 2.68 (0.146) 0.342 0.564 -0.022
3. T–P (1) 3.28 (0.301) 4.25 (0.281) 5.500 0.027* 0.113
4. T–P (2) 2.77 (0.345) 2.89 (0.322) 0.065 0.800 -0.026
5. T–P (3) 2.94 (0.268) 2.55 (0.251) 1.142 0.295 0.003
6. n–P and/or V–P 2.13 (0.201) 2.13 (0.188) 0.000 1.000 -0.035
7. n–V 2.04 (0.233) 2.71 (0.217) 4.277 0.048* 0.079
8. ma–P 1.86 (0.343) 2.62 (0.319) 2.380 0.135 0.036
9. Partial pressure 1.57 (0.308) 1.50 (0.287) 0.026 0.873 -0.033
Total 22.33 (1.111) 24.34 (1.038) 1.707 0.202 0.012
N 14 16
* p\ 0.05; ** p\ 0.001
J Sci Educ Technol
123
collision); (2) pressure is the sum of the force of the
molecular collisions per unit area (collision–pressure).
McNemar’s tests (Table 5) indicated significant increasesin the proportions of students using the motion–collision
idea and the collision–pressure idea in the Frame group but
not in the traditional group.Question 3 assessed understanding of the Gay-Lussac’s
law by asking students to explain why gas pressureincreases as gas temperature increases at constant volume
in an enclosed container. A desirable explanation would
involve several connected ideas: as temperature increases,kinetic energy increases (temperature–kinetic energy) and
gas molecules move faster (temperature–velocity), result-
ing in more frequent collisions with container walls (ve-locity–collision frequency) and more force per collision
(velocity–collision impulse), which results in increased gas
pressure (collision frequency–pressure, collision impulse–pressure). McNemar’s tests (Table 6) indicated significant
increases in the proportions of students using the temper-
ature–velocity, velocity–collision impulse, and collisionimpulse–pressure ideas in the Frame group but not in the
traditional group. However, the proportions of students
using the temperature–kinetic energy, velocity–collisionfrequency, and collision frequency–pressure ideas stayed
the same in both groups.
Question 7 assessed the relationship between the quan-tity and volume of a gas at constant temperature and
pressure. It asks students to choose among four possible
causes for why helium balloons shrink overnight given the
same temperature and atmospheric pressure and asks for an
explanation of their choice. The correct choice is that some
of the helium molecules escaped through pores in the latex.The desirable explanation involves the idea of gas mole-
cules being in constant motion and an application of the
ideal gas law: Gas volume is proportional to the quantity ofgas at the same temperature and pressure. Although
McNemar’s tests (Table 7) showed no significant change inthe proportions of students using these two ideas in both
groups, the Frame group appeared to have an increasing
trend for both ideas, which together might have contributedto its significantly better KI score than the traditional
group.
In sum, these results showed that significantly morestudents in the Frame group responded in ways reflecting
connections among gas temperature, velocity of gas
molecules, collision impulse between gas molecules andcontainer walls, and gas pressure, while no such change
was found in the traditional group. As illustrated in Fig. 3,
these connections relate to an increase in student under-standing between physical and molecular properties of gas
laws.
Discussion
Augmenting virtual labs with physical interactions can
potentially optimize science learning by integrating the
virtual and physical experiences that are essential to deep
Table 5 McNemar’s tests forchange in proportions ofstudents using normative ideasfor question 1 in each condition
Ideas Traditional Frame
Pretest Posttest v2 p value Pretest Posttest v2 p value
Motion–collision 4 (29 %) 4 (29 %) 0.500 0.240 0 (0 %) 7 (44 %) 5.143 0.012*
Collision–pressure 7 (50 %) 9 (64 %) 0.250 0.309 3 (19 %) 11 (69 %) 4.900 0.013*
N 14 16
* p\ 0.05; ** p\ 0.001
Table 6 McNemar’s tests for change in proportions of students using normative ideas for question 3 in each condition
Ideas Traditional Frame
Pretest Posttest v2 p value Pretest Posttest v2 p value
Temperature–kinetic energy 4 (29 %) 2 (14 %) 0.250 0.309 0 (0 %) 0 (0 %) N/A N/A
Temperature–velocity 6 (43 %) 9 (64 %) 0.571 0.225 9 (56 %) 16 (100 %) 5.143 0.012*
Velocity–collision frequency 1 (7 %) 2 (14 %) 0.000 0.500 1 (6 %) 3 (19 %) 0.250 0.309
Velocity–collision impulse 1 (7 %) 1 (7 %) N/A N/A 1 (6 %) 6 (38 %) 3.200 0.037*
Collision frequency–pressure 2 (14 %) 3 (21 %) 0.000 0.500 0 (0 %) 3 (19 %) 1.333 0.124
Collision impulse–pressure 3 (21 %) 3 (21 %) 0.500 0.240 1 (6 %) 6 (38 %) 3.200 0.037*
N 14 16
* p\ 0.05; ** p\ 0.001
J Sci Educ Technol
123
understanding of complex science concepts. This studydemonstrated a novel approach to create such a mixed-
reality learning environment. Using sensors to connecting
physical objects and computer simulations, the Frametechnology seamlessly connects the physical world and the
virtual world, presents concepts, and processes across themacroscopic and microscopic levels in one stream of flow.
This study aimed to explore how augmenting virtual
labs with physical interactions affect student learning ofcomplex science concepts. We compared the effects of a
Frame lab and a traditional combination of sensor-based
lab and classroom discussions on high school students’understanding of the gas laws and kinetic molecular theory.
Results show that the Frame lab helped students develop
understanding of most targeted concepts, with largeimprovement on understanding of gas pressure and Gay-
Lussac’s Law, and moderate improvement on gas tem-
perature, Avogadro’s law, and the relationship betweenmolecular mass and gas pressure. Compared to traditional
instruction with sensor-based labs, the Frame lab did not
show significant advantage judged by the students’ overallperformance. However, item-level analysis revealed that
the Frame did help students learn certain concepts that
connect the physical and molecular properties of gas betterthan traditional instruction. We outline a few potential
explanations for these nuanced findings below.
These improved understandings in the Gas Frame groupare aligned with the unique affordances that the Gas Frame
provides—presenting tangible, macroscopic events andunderlying molecular behavior simultaneously. For
instance, when students touch a hot jar on the temperature
sensor tip, the molecular visualization immediately showsthe gas molecules speeding up, colliding with the container
walls more frequently. Similarly, when they press the vir-
tual piston through the physical spring, the molecularvisualization immediately shows the molecule–piston col-
lisions, represented by arrows, becoming more frequent.
When students put the hot jar to the Frame and try to keepthe volume of the virtual piston constant, they can feel the
extra force they need to impart to the physical spring. The
improved performance of the Frame group on select itemscould potentially be due to having the simultaneous, real-
time connection between the macroscopic phenomena and
the molecular simulation.The observed differences between the two groups may
also be attributed to the pedagogical differences that
Fig. 3 Improved explanationcomponents (bolded lines)among the physical propertiesand molecular properties of gasfor students using the GasFrame lab
Table 7 McNemar’s tests for change in proportions of students using normative ideas for question 7 in each condition
Ideas Traditional Frame
Pretest Posttest v2 p value Pretest Posttest v2 p value
Temperature/external pressure–volume 2 (14 %) 3 (21 %) 0.000 1.000 3 (21 %) 7 (44 %) 1.500 0.221
Normative ideas related to molecular motion 3 (21 %) 3 (21 %) 0.500 0.480 1 (6 %) 3 (19 %) 0.500 0.480
N 14 16
* p\ 0.05; ** p\ 0.001
J Sci Educ Technol
123
emerged due to the different affordances provided by the
Frame technology and traditional lab/discussion methods.The Frame required minimal procedural overhead such as
instrument calibration and provided visualizations that
dynamically showed the underlying molecular mechanismsof the gas laws. The traditional lab apparatus allowed
students to collect reliable data and conveniently conduct
mathematical analysis, but it required extra effort to set upand calibrate instruments, familiarize oneself with the
computer interface, and perform data collection and anal-ysis. Thus, in the same time period, the Frame lab allowed
students to investigate multiple gas laws on a conceptual
level, while the traditional lab supported students toinvestigate only one gas law in depth and covered other gas
laws through whole-class discussions.
The two groups learned the molecular explanations ofthe gas laws in two different modes: guided inquiry (Col-
burn 2000) and direct instruction. Guided inquiry has been
shown to be slightly more effective than direct instructiondue to factors such as self-explanation effect and genera-
tion effect (Alfieri et al. 2011). However, in specific cases
students may be limited by their own abilities to constructknowledge from interacting with computer simulations
(Chiu 2010). For example, students’ lab notes showed that
many students only generated partial molecular explana-tions for their findings, especially for those gas phenomena
that involved complex molecular processes. Students
receiving the traditional instruction, however, had access tocomplete molecular explanations provided by the teacher.
Therefore, it is difficult to determine whether different
learning modes led to the different learning outcomes in thetwo groups. The advantage of guided inquiry might have
been comprised by students’ abilities to handle complex
domains.The other major difference was the representation of
molecular processes. While students in the Frame lab used
the molecular visualizations to independently generatemolecular explanations, students receiving the traditional
instruction relied on the teacher’s verbal description to
mentally simulate the molecular behavior of gases.Research shows that science visualizations with proper
scaffolding promote conceptual learning (e.g., McElhaney
et al. 2015). It is possible that the molecular visualizationalone accounted for the better learning outcomes in the
Frame group. However, as a growing number of studies
showed the benefits of augmenting virtual labs with hapticcontrols and feedback (Bivall et al. 2011; Han and Black
2011; Kim et al. 2011; Minogue and Jones 2009), it is
reasonable to suspect that the physical controls and feed-back in the Frame played important roles in promoting
conceptual learning. A further study comparing the Frame
against an equivalent simulation with a conventional
graphical user interface would shed light on the effects of
haptic augmentation for virtual labs.
Implications for Theories
The findings of this study are consistent with previous
studies showing the benefits of integrating physical andvirtual learning experiences, such as combining physical
and virtual labs (e.g., Zacharia 2007), running both labs in
parallel (e.g., Jaakkola et al. 2011) or side by side (e.g.,Blikstein 2014), and augmenting virtual labs with haptic
devices (e.g., Han and Black 2011; Minogue et al. 2006).Results also align with studies, showing that physical
manipulation of real-life objects has unique affordances for
science learning (e.g., Chen et al. 2014; Lazonder andEhrenhard 2014). Thus, this study strengthens the general
argument that interacting with physical and virtual mate-
rials provides valuable and important ways to understandscience concepts.
Specifically, the findings of this study suggest that
simultaneously connecting physical and virtual learningexperiences can help students make connections between
macroscopic and molecular concepts and processes. This
particular function has not been documented in the literatureof physical and virtual labs.Most of the previous studies only
focused on physics topics at the macroscopic level (e.g.,
Chini et al. 2012; Han and Black 2011; Jaakkola et al. 2011;Zacharia et al. 2008; Zacharia 2007). The studies that did
involve chemistry or biology topics were often conducted
with a focus on the macroscopic level (e.g., Chen et al. 2014;Chien et al. 2015) or the microscopic/submicroscopic level
only (e.g., Bivall et al. 2011; Minogue and Jones 2009). The
present study contributes one way to conceptualize how todesign learning environments to support understanding of
connections among submicroscopic andmacroscopic levels.
Given the importance of connecting various levels of rep-resentations in domains such as chemistry and biology, more
research along this line of simultaneously connecting phys-
ical and virtual worlds (e.g., Lindgren and Johnson-Glenberg2013) is warranted.
In particular, contrasting this study to Liu’s (2006) study
provides insights into the potential benefits of the Framecompared to a sequential combination of physical and
virtual experiences. Liu (2006) compared high school
students’ learning of the Gay-Lussac’s Law in a physicallab, a virtual lab featuring molecular visualization, and a
sequential combination of both. Students in Liu’s study
spent 90 min (two 45-min class periods) to study the Gay-Lussac’s Law and its underlying molecular process in a
physical lab followed by a virtual lab, or in the reverse
order, while the students in the present study spent
J Sci Educ Technol
123
approximately 30 min to study the same content. The
students in Liu’s study were asked to describe the rela-tionship between temperature and pressure of a gas in a
sealed container and explain this relationship at the
molecular level. This question is similar to question 3 inthe present study, which asked students to explain why the
pressure of a gas in a sealed container would increase as its
temperature increased. In Liu’s study, there was no statis-tically significant change in the percentage of students who
indicated the velocity of gas molecules (72.7 % in pretest,60.6 % in first posttest, and 63.6 % in second posttest) and
the collision with walls by gas molecules (0 % in pretest,
6.1 % in first posttest, and 3.0 % in second posttest). Whilethe traditional group in the present study had the similar
learning outcomes, the Frame group had significant
increases in the percentages of students citing velocity(54 % in pretest to 100 % in posttest) and collision impulse
(6 % in pretest to 38 % in posttest) to explain for the same
phenomenon. The Frame lab appears to have targeted themacroscopic–molecular connections more effectively than
the combined lab studied by Liu. The main difference
between Liu’s combined lab and the Frame lab is thesimultaneity of the physical and virtual experiences. It is
reasonable to suspect that the better learning outcomes in
the Frame group in the present study may be attributed tothe careful integration of physical and virtual experiences
supported by the Frame technology.
Implications for Practices
The Frame technology provides a novel approach to inte-
grate physical and virtual lab experiences. It seamlessly
connects physical and virtual worlds that are typicallyasynchronous in various combinations of physical and vir-
tual labs, and it supports complex simulation with multiple
sets of physical controls and feedback that most hapticdevices are unable to support with relatively minimal setup
of technologies. In addition, the Frame technology combines
the haptic controls and feedback supported by haptic devices(Han and Black 2011) and the authenticity and familiarity of
object manipulation supported by tangible interfaces (Antle
andWise 2013), making it possible to utilize learners’ hapticmodality as well as draw upon sensorimotor skills to opti-
mize learning. This approach promotes the degree of
embodiment by providing a high level of congruencybetween control gestures and science concepts (Johnson-
Glenberg et al. 2014). For example, students can naturally
make connections between ‘‘pumping gas’’ through a syr-inge and ‘‘adding gas molecules’’ into a virtual chamber.
These unique affordances can support challenging
learning objectives called for by the Next Generation Sci-ence Standards (NGSS Lead States 2013). According to
NGSS, students should develop scientific understanding
that connects the macroscopic and microscopic/submicro-scopic levels. For instance, the chemistry standards require
students to use the kinetic molecular theory to explain the
behaviors of gases and relationship between properties ofgas. Computer simulations are effective tools to support
students to develop cross-level understanding. However,
completely replacing physical labs with simulations cantake away valuable sensory experiences from students, and
adding simulations onto physical labs inevitably doublesthe workload for both students and teachers without a
guaranteed desirable outcome. The Frame technology
configures the essential components from the physicalworld and the virtual world into a seamlessly connected
mixed-reality world, supporting an unobstructed explo-
ration across the macroscopic and molecular levels withoutadding unnecessary workload.
The Frame also provides a novel solution to providing
high-quality mixed-reality experiences for students with acombination of relatively ordinary technologies. As many
teachers are already well versed with setting up and trou-
bleshooting sensor-based labs, the Frame technology maybe relatively accessible to teachers compared to other
embedded or haptic/tangible approaches. Although the
Frame does require building a physical box per lab group,the other setup requirements are similar, if not easier, than
existing sensor-based labs. Additionally, many teachers
have these temperature, pressure, and force sensors alreadyavailable. The Frame provides an example of how effective
mixed-reality technologies can be designed for accessibil-
ity and use in authentic classroom settings.This study demonstrates how the Frame technology can
support students to make connections between macroscopic
and submicroscopic levels in an authentic classroom setting.Specifically, the findings suggest that sensor-augmented vir-
tual labs may offer similar benefits as traditional instruction
for most aspects of the targeted domain. However, whenaddressing more difficult concepts that require students to
connect molecular behaviors to observable phenomena, the
sensor-augmented virtual lab may provide additional benefitsto traditional instruction. Thus, we suggest that instructors
consider using sensor-augmented virtual labs to target specific
learning objectives and also weight this new approach againstwhat may already work in a classroom setting.
As more mixed-reality technologies like the Frame
emerge in the market, more creative solutions would bepossible to address challenging instructional goals. How-
ever, effective instruction requires knowledge of content,
pedagogy, and technology, and more importantly howthese three areas of knowledge intersect (e.g., Koehler and
Mishra 2009). Having technologically knowledgeable
instructors can potentially make difficult concepts moreaccessible through informed use of technologies. The
J Sci Educ Technol
123
Frame technology, as demonstrated in this study, made it
possible for students to understand, through a guided-in-quiry process, how an ideal gas behaves is explained by the
behaviors of the tiny gas molecules. Teacher education and
professional development programs can provide opportu-nities for teachers to explore mixed-reality technologies
such as the Frame. In-service and pre-service teachers
should be made aware of the existence and efficacy of thesemixed-reality technologies, so that they can begin to adopt
and generate solutions for increasingly challenging learn-ing standards.
Limitations
As described in the method section, we intentionally chosea small and convenient sample with particular character-
istics. These research design choices inevitably limit gen-
eralizability. The results of this study apply to advancedchemistry classes taught by high-performing teachers in
college-oriented high schools. Also, to ensure quality
implementation and feedback, two researchers providedtechnical support and conducted field observations in the
Frame lab. Although both groups were aware of their
participation in the study, students in the Frame group werelikely to be more alerted due to the presence of researchers.
Additionally, as discussed previously, some pedagogical
differences existed between the two interventions owing todifferent technology affordances. The observed advantages
of the Frame lab might be partially attributed to its inquiry-
based approach and access to molecular visualizations.
Future Research
To improve the generalizability of our research on sensor-
augmented virtual labs, we are currently conducting a studywith a large sample of students with diverse demographic
and academic background. Also, to control for potential
confounding factors such as differences in pedagogy andrepresentation of science concepts, we are comparing the
Frame lab against a virtual lab with the same visual sim-
ulation but traditional graphical user interface. This ongo-ing study will more precisely identify the effects of
augmenting virtual labs with physical interactions, and its
findings will be generalizable to a larger population.
Conclusions
In this study, we investigated the effectiveness of a sensor-
augmented virtual lab (i.e., the Gas Frame) that usesphysical controls connected to molecular visualizations to
promote students’ understanding of complex science con-
cepts. The results suggest that the Gas Frame helped stu-dents develop understanding of almost all aspects of gas
laws and kinetic molecular theory. Compared to traditional
instruction, the Gas Frame was more effective in helpingstudents develop connections between molecular properties
and physical properties of gas (e.g., the relationship
between gas temperature and velocity of gas molecules).These findings demonstrate the advantages of sensor-aug-
mented virtual labs in promoting understanding of scienceconcepts.
Acknowledgments This research was supported by the NationalScience Foundation under grant IIS-1123868. Any opinions, findings,and conclusions or recommendations expressed in this material arethose of the authors and do not necessarily reflect the views of theNational Science Foundation. The authors would like to thank theteachers and students who participated in this project. Special thanksto Charles Xie and Edmund Hazzard at the Concord Consortium fordesign and development of the Frame technology and curriculumused in this study.
Conflict of interest The authors declare that they have no conflictof interest.
References
Abrahamson D, Lindgren R (2014) Embodiment and embodieddesign. In: Sawyer RK (ed) The Cambridge handbook of thelearning sciences, 2nd edn. Cambridge University Press,Cambridge
Abrahamson D, Gutierrez J, Charoenying T, Negrete A, Bumbacher E(2012) Fostering hooks and shifts: tutorial tactics for guidedmathematical discovery. Technol Knowl Learn 17(1–2):61–86.doi:10.1007/s10758-012-9192-7
Akpan JP, Andre T (2000) Using a computer simulation beforedissection to help students learn anatomy. J Comput Math SciTeach 19(3):297–313
Alfieri L, Brooks PJ, Aldrich NJ, Tenenbaum HR (2011) Doesdiscovery-based instruction enhance learning? J Educ Psychol103(1):1–18
Antle AN, Wise AF (2013) Getting down to details: using theories ofcognition and learning to inform Tangible User Interface design.Interact Comput 25(1):1–20. doi:10.1093/iwc/iws007
Benson DL, Wittrock MC, Baur ME (1993) Students’ preconceptionsof the nature of gases. J Res Sci Teach 30(6):587–597
Ben-Zvi R, Eylon B, Silbemein J (1986) Is an atom of coppermalleable? J Chem Educ 63(1):64–66
Bivall P, Ainsworth S, Tibell LAE (2011) Do haptic representationshelp complex molecular learning? Sci Educ 95(4):700–719.doi:10.1002/sce.20439
Blikstein P (2014). Bifocal modeling: promoting authentic scientificinquiry through exploring and comparing real and ideal systemslinked in real-time. Playful User Interfaces, pp 317–352
Bodner GM (1991) I have found you an argument: the conceptualknowledge of beginning chemistry graduate students. J ChemEduc 68(5):385–388
Bouwma-Gearhart J, Stewart J, Brown K (2009) Student misappli-cation of a gas-like model to explain particle movement inheated solids: implications for curriculum and instructiontowards students’ creation and revision of accurate explanatory
J Sci Educ Technol
123
models. Int J Sci Educ 31(9):1157–1174. doi:10.1080/09500690902736325
Bransford JD, Brown AL, Cocking RR (2000) How people learn:brain, mind, experience, and school. National Academy Press,Washington
Chen S, Chang WH, Lai CH, Tsai CY (2014) A comparison ofstudents’ approaches to inquiry, conceptual learning, andattitudes in simulation-based and microcomputer-based labora-tories. Sci Educ 98(5):905–935. doi:10.1002/sce.21126
Chien K, Tsai C, Chen H, Chang W, Chen S (2015) Learningdifferences and eye fixation patterns in virtual and physicalscience laboratories. Comput Educ 82:191–201. doi:10.1016/j.compedu.2014.11.023
Chini JJ, Madsen A, Gire E, Rebello NS, Puntambekar S (2012)Exploration of factors that affect the comparative effectivenessof physical and virtual manipulatives in an undergraduatelaboratory. Phys Rev Special Topics Phys Educ Res 8(1):1–12.doi:10.1103/PhysRevSTPER.8.010113
Chiu JL (2010). Promoting links and developing students’ criteria forvisualizations by prompting judgments of fidelity. In ICLS’10Proceedings of the 9th international conference of the learningsciences. Chicago, IL, vol 2, pp 230–230
Chiu JL, Linn MC (2012) Supporting self-monitoring with dynamicvisualizations. In: Dori J, Zohar A (eds) Metacognition andscience education. Lawrence Erlbaum, Mahwah, pp 133–164
Chiu JL, Linn MC (2014) Supporting knowledge integration inchemistry with a visualization-enhanced inquiry unit. J Sci EducTechnol 23(1):37–58. doi:10.1007/s10956-013-9449-5
Chiu JL, DeJaegher CJ, Chao J (2015) The effects of augmentedvirtual science laboratories on middle school students’ under-standing of gas properties. Comput Educ 85:59–73. doi:10.1016/j.compedu.2015.02.007
Clark DB (2006) Longitudinal conceptual change in students’understanding of thermal equilibrium: an examination of theprocess of conceptual restructuring. Cognit Inst 24(4):467–563
Clark DB, Jorde D (2004) Helping students revise disruptiveexperientially supported ideas about thermodynamics: computervisualizations and tactile models. J Res Sci Teach 41(1):1–23.doi:10.1002/tea.10097
Cohen J (1968) Weighed kappa: nominal scale agreement withprovision for scaled disagreement or partial credit. Psychol Bull70(4):213–220
Colburn A (2000) An inquiry primer. Sci Scope 23(6):42–44De Jong T, Linn MC, Zacharia ZC (2013) Physical and virtual
laboratories in science and engineering education. Science (NewYork, NY) 340(6130):305–308. doi:10.1126/science.1230579
Finkelstein N, Adams W, Keller C, Kohl P, Perkins KK, PodolefskyN, LeMaster R (2005) When learning about the real world isbetter done virtually: a study of substituting computer simula-tions for laboratory equipment. Phys Rev Spec Topics Phys EducRes 1(1):1–8. doi:10.1103/PhysRevSTPER.1.010103
Gabel DL, Sherwood RD, Enochs L (1984) Problem-solving skills ofhigh school chemistry students. J Res Sci Teach 21(2):221–233.doi:10.1002/tea.3660210212
Gabel DL, Samuel KV, Hunn D (1987) Understanding the particulatenature of matter. J Chem Educ 64(8):695. doi:10.1021/ed064p695
Hammer D (2000) Student resources for learning introductoryphysics. Am J Phys 68(S1):S52. doi:10.1119/1.19520
Han I, Black JB (2011) Incorporating haptic feedback in simulationfor learning physics. Comput Educ 57(4):2281–2290. doi:10.1016/j.compedu.2011.06.012
Hmelo-Silver CE, Duncan RG, Chinn CA (2007) Scaffolding andachievement in problem-based and inquiry learning: a responseto Kirschner, Sweller, and Clark (2006). Educ Psycholog42(2):99–107. doi:10.1080/00461520701263368
Hofstein A, Lunetta VN (2004) The laboratory in science education:foundations for the twenty-first century. Sci Educ 88(1):28–54.doi:10.1002/sce.10106
Honey MA, Hilton M (2011) Learning science through computergames and simulations. National Academies Press, Washington
Jaakkola T, Nurmi S, Veermans K (2011) A comparison of students’conceptual understanding of electric circuits in simulation onlyand simulation-laboratory contexts. J Res Sci Teach 48(1):71–93. doi:10.1002/tea.20386
Johnson-Glenberg MC, Birchfield D, Tolentino L, Koziupa T (2014)Collaborative embodied learning in mixed reality motion-capture environments: two science studies. J Educ Psychol106(1):86–104. doi:10.1037/a0034008
Johnstone AH (1991) Why is science difficult to learn? Things areseldom what they seem. J Comput Assist Learn 7(2):75–83
Jones MG, Minogue J, Tretter TR, Negishi A, Taylor RM (2006)Haptic augmentation of science instruction: does touch matter?Sci Educ 90(1):111–123. doi:10.1002/sce.20086
Kim Y, Park S, Kim H, Jeong H, Ryu J (2011) Effects of differenthaptic modalities on students’ understanding of physical phe-nomena. In: 2011 IEEE World haptics conference (pp. 379–384).IEEE. doi:10.1109/WHC.2011.5945516
Koehler MJ, Mishra P (2009) What is technological pedagogicalcontent knowledge (TPACK)? Contemp Issues Technol TeachEduc 9(1):60–70. doi:10.1016/j.compedu.2010.07.009
Kozma RB, Russell J (1997) Multimedia and understanding: expertand novice responses to different representations of chemicalphenomena. J Res Sci Teach 34(9):949–968. doi:10.1002/(SICI)1098-2736(199711)34:9\949:AID-TEA7[3.0.CO;2-U
Lazonder AW, Ehrenhard S (2014) Relative effectiveness of physicaland virtual manipulatives for conceptual change in science: Howfalling objects fall. J Comput Assist Learn 30(2):110–120.doi:10.1111/jcal.12024
Lee O, Eichinger DC, Anderson CW, Berkheimer GD, Blakeslee TD(1993) Changing middle school students’ conceptions of matterand molecules. J Res Sci Teach 30(3):249–270
Levy ST, Wilensky U (2009) Students’ learning with the connectedchemistry (CC1) curriculum: navigating the complexities of theparticulate world. J Sci Educ Technol 18(3):243–254. doi:10.1007/s10956-009-9145-7
Lin H, Cheng H (2000) The assessment of students and teachers’understanding of gas laws. J Chem Educ 77(2):235–238
Lindgren R, Johnson-Glenberg MC (2013) Emboldened by embod-iment: six precepts for research on embodied learning and mixedreality. Educ Res 42(8):445–452. doi:10.3102/0013189X13511661
Linn MC, Eylon BS (2006) Science education: integrating views oflearning and instruction. In: Alexander PA, Winne PH (eds)Handbook of educational psychology, 2nd edn. LawrenceErlbaum Associates, Mahwah, pp 511–544
Linn MC, Eylon BS (2011) Science learning and instruction: takingadvantage of technology to promote knowledge integration.Routledge, New York
Linn MC, Davis EA, Eylon B (2004) The scaffolded knowledgeintegration framework for instruction. In: Linn MC, Davis EA,Bell P (eds) Internet environments for science education.Lawrence Erlbaum Associates Inc, Mahwah, pp 47–72
Liu X (2006) Effects of combined hands-on laboratory and computermodeling on student learning of gas laws: a quasi-experimentalstudy. J Sci Educ Technol 15(1):89–100. doi:10.1007/s10956-006-0359-7
Liu OL, Lee H-S, Hofstetter C, Linn MC (2008) Assessing knowledgeintegration in science: Construct, measures, and evidence. EducAssessm 13(1): 33–55. Retrieved from http://www.tandfonline.com/doi/abs/10.1080/10627190801968224
J Sci Educ Technol
123
Lowe R (2004) Interrogation of a dynamic visualization duringlearning. Learn Inst 14(3):257–274. doi:10.1016/j.learninstruc.2004.06.003
Lui M, Slotta JD (2013) Immersive simulations for smart classrooms:exploring evolutionary concepts in secondary science. TechnolPedagogy Educ. doi:10.1080/1475939X.2013.838452
Lunetta VN, Hofstein A, Clough MP (2007) Learning and teaching inthe school science laboratory: an analysis of research, theory,and practice. In: Abell SK, Lederman NH (eds) Handbook ofresearch on science education. Lawrence Erlbaum Associates,Mahwah, pp 393–441
Marshall J, Young ES (2006) Preservice teachers’ theory develop-ment in physical and simulated environments. J Res Sci Teach43(9):907–937. doi:10.1002/tea.20124
McBride DL, Murphy S, Zollman DA (2010) Student understandingof the correlation between hands-on activities and computervisualizations of NMR/MRI. In AIP conference proceedings, vol225, pp 225–228. doi:10.1063/1.3515207
McElhaney KW, Linn MC (2011) Investigations of a complex,realistic task: intentional, unsystematic, and exhaustive experi-menters. J Res Sci Teach 48(7):745–770
McElhaney KW, Chang H-Y, Chiu JL, Linn MC (2015) Evidence foreffective uses of dynamic visualisations in science curriculummaterials. Stud Sci Educ 51(1):49–85
Milgram P, Kishino F (1994) A taxonomy of mixed reality visualdisplays. IEICE Trans Inform Syst E77-D(12):1321–1329
Minogue J, Borland D (2012). Investigating students’ ideas aboutbuoyancy and the influence of haptic feedback. In: Paperpresented at the annual international conference of the NationalAssociation for Research in Science Teaching, Indianapolis, IN,pp 1–12
Minogue J, Jones MG (2009) Measuring the impact of hapticfeedback using the SOLO taxonomy. Int J Sci Educ 31(10):1359–1378. doi:10.1080/09500690801992862
Minogue J, Jones MG, Broadwell B, Oppewall T (2006) The impactof haptic augmentation on middle school students’ conceptionsof the animal cell. Virtual Real 10(3–4):293–305. doi:10.1007/s10055-006-0052-4
Moher, T. (2006). Embedded phenomena: Supporting science learn-ing with classroom-sized distributed simulations. In Proceedingsof the SIGCHI conference on human factors in computingsystems—CHI’06 (pp. 691–700)
Moher T, Wiley J, Jaeger A, Silva BL, Novellis F (2010) Spatial andtemporal embedding for science inquiry: An empirical study ofstudent learning. In ICLS’10 proceedings of the 9th internationalconference of the learning sciences—volume 1, pp 826–833)
Nakhleh MB (1992) Why some students don’t learn chemistry:chemical misconceptions. J Chem Educ 69(3):191–196. doi:10.1021/ed069p191
National Research Council (2006) America’s lab report: investiga-tions in high school science. National Academy Press,Washington
NGSS Lead States (2013) Next generation science standards: forstates, by states. The National Academies Press, Washington
Noh T, Scharmann LC (1997) Instructional influence of a molecular-level pictorial presentation of matter on students’ conceptionsand problem-solving ability. J Res Sci Teach 34(2):199–217.doi:10.1002/(SICI)1098-2736(199702)34:2\199:AID-TEA6[3.0.CO;2-O
Novellis, F., & Moher, T. (2011). How real is ‘‘real enough’’?Designing artifacts and procedures for embodied simulations ofscience practices. In Proceedings of the 10th internationalconference on interaction design and children—IDC’11,pp 90–98
Novick S, Nussbaum J (1981) Pupils’ understanding of the particulatenature of matter: a cross-age study. Sci Educ 65(2):187–196
Nurrenbem SC, Pickering M (1987) Concept learning versus problemsolving: Is there a difference? J Chem Educ 64(6):508–510
Okamura AM, Richard C, Cutkosky MR (2002) Feeling is believing:using a force-feedback joystick to teach dynamic systems. J EngEduc 91(3):345–349
Olympiou G, Zacharia ZC (2010) Comparing the use of virtual andphysical manipulatives in physics education. In Paper presentedat the international conference on education, training andinformatics. Orlando, FL
Olympiou G, Zacharia ZC (2012) Blending physical and virtualmanipulatives: an effort to improve students’ conceptual under-standing through science laboratory experimentation. Sci Educ96(1):21–47. doi:10.1002/sce.20463
Ozmen H (2013) A cross-national review of the studies on theparticulate nature of matter and related concepts. Eurasian J PhysChem Educ 5(2):81–110
Padilla K (2009) Visualization: theory and practice in scienceeducation. Int J Sci Educ 31(10):1417–1420. doi:10.1080/09500690802673768
Podolefsky NS, Perkins KK, Adams WK (2010) Factors promotingengaged exploration with computer simulations. Phys Rev SpecTopics Phys Educ Res 6(2):020117. doi:10.1103/PhysRevSTPER.6.020117
Price S, Falcao TP (2011) Where the attention is: discovery learningin novel tangible environments. Interact Comput 23(5):499–512.doi:10.1016/j.intcom.2011.06.003
Price S, Sheridan JG, Falcao TP, Roussos G (2008) Towards aframework for investigating tangible environments for learning.Int J Arts Technol 1(3/4):351. doi:10.1504/IJART.2008.022367
Price, S., Falcao, T. P., Sheridan, J. G., & Roussos, G. (2009). Theeffect of representation location on interaction in a tangiblelearning environment. Proceedings of the 3rd internationalconference on Tangible and embedded interaction—TEI’09,p 85. doi:10.1145/1517664.1517689
Robins LI, Villagomez G, Dockter D, Christopher E, Ortiz C,Passmore C, Smith MH (2009) Teacher research: challengingour assumptions–An investigation into student understanding ofthe gas laws. Sci Teach (September), 35–40
Sanger MJ, Phelps AJ (2007) What are students thinking when theypick their answer? A content analysis of students’ explanationsof gas properties. J Chem Educ 84(5):870–874
Sanger MJ, Phelps AJ, Fienhold J (2000) Using a computer animationto improve students’ conceptual understanding of a can-crushingdemonstration. J Chem Educ 77(11):1517–1520
Sarabando C, Cravino JP, Soares A (2014) Contribution of acomputer simulation to students’ learning of the physicsconcepts of weight and mass. Proc Technol 13:112–121.doi:10.1016/j.protcy.2014.02.015
Schneider B, Wallace J, Blikstein P, Pea R (2013) Preparing for futurelearning with a Tangible User interface: the case of neuro-science. IEEE Trans Learn Technol 6(2):117–129. doi:10.1109/TLT.2013.15
Shelley T, Lyons L, Zellner M, Minor E (2011). Evaluating theembodiment benefits of a paper-based tui for educationalsimulations. In Proceedings of the 2011 annual conferenceextended abstracts on Human factors in computing systems—CHI EA’11. ACM Press, New York, pp 1375–1380. doi:10.1145/1979742.1979777
Smith, G. W., & Puntambekar, S. (2010). Examining the combinationof physical and virtual experiments in an inquiry scienceclassroom. In Proceedings of the conference on computer basedlearning in science. Warsaw, Poland
Snir J, Smith CL, Raz G (2003) Linking phenomena with competingunderlying models: a software tool for introducing students tothe particulate model of matter. Sci Educ 87(6):794–830. doi:10.1002/sce.10069
J Sci Educ Technol
123
Tanhua-piiroinen, E., Pystynen, J., & Raisamo, R. (2010). Hapticapplications as physics teaching tools. In IEEE internationalsymposium on haptic audio-visual environments and games,Phoenix, AZ, pp 1–6
Tien LT, Teichert MA, Rickey D (2007) Effectiveness of a MORElaboratory module in prompting students to revise their molec-ular-level ideas about solutions. J Chem Educ 84(1):175. doi:10.1021/ed084p175
Tolentino L, Birchfield D, Megowan-Romanowicz C, Johnson-Glenberg MC, Kelliher A, Martinez C (2009) Teaching andlearning in the mixed-reality science classroom. J Sci EducTechnol 18(6):501–517. doi:10.1007/s10956-009-9166-2
Trundle KC, Bell RL (2010) The use of a computer simulation topromote conceptual change: a quasi-experimental study. ComputEduc 54(4):1078–1088. doi:10.1016/j.compedu.2009.10.012
White RT, Gunstone RF (1992) Probing understanding. Falmer,London
Wiebe EN, Minogue J, Jones MG, Cowley J, Krebs D (2009) Hapticfeedback and students’ learning about levers: unraveling theeffect of simulated touch. Comput Educ 53(3):667–676. doi:10.1016/j.compedu.2009.04.004
Williams RL, Chen M, Seaton JM (2003) Haptics-augmented simple-machine educational tools. J Sci Educ Technol 12(1):1–12
Xie, C. (2012). Framing mixed-reality labs. Retrieved from http://concord.org/publications/newsletter/2012-spring/framing-mixed-reality-labs
Zacharia ZC (2007) Comparing and combining real and virtualexperimentation: an effort to enhance students’ conceptual under-standing of electric circuits. J Comput Assist Learn 23(2):120–132
Zacharia ZC, de Jong T (2014) The effects on students’ conceptualunderstanding of electric circuits of introducing virtual manip-ulatives within a physical manipulatives-oriented curriculum.Cognit Inst 32(2):101–158
Zacharia ZC, Olympiou G, Papaevripidou M (2008) Effects ofexperimenting with physical and virtual manipulatives onstudents’ conceptual understanding in heat and temperature.J Res Sci Teach 45(9):1021–1035. doi:10.1002/tea.20260
Zhang ZH, Linn MC (2011) Can generating representations enhancelearning with dynamic visualizations? J Res Sci Teach 48(10):1177–1198. doi:10.1002/tea.20443
J Sci Educ Technol
123
View publication statsView publication stats