Research: Science and Education

1058 Journal of Chemical Education • Vol. 81 No. 7 July 2004 • www.JCE.DivCHED.org

edited byDiane M. Bunce

The Catholic University of AmericaWashington, D.C. 20064

Amy J. PhelpsMiddle Tennessee State University

Murfreesboro, TN 37132

In college chemistry courses, educators are using com-puter-mediated communication and computer-supportedcollaborative learning in a variety of different ways (1–10).In some of these endeavors faculty use electronic communi-cation technologies such as WebCT, Blackboard, or discus-sion boards at their own institutions (4). In other cases,faculty combine cooperative learning techniques and elec-tronic communication to form computer-supported collabo-rative learning projects (2, 6–10). Physical Chemistry OnLine(PCOL) modules have been developed and implemented bya consortium of faculty at geographically dispersed universi-ties to help students learn physical chemistry (6–8,11). Themodules use a guided-inquiry approach and rely on collabo-ration between students in their own class, as well as com-puter-supported collaboration across institutions. To date, thePCOL consortium has produced 13 modules dealing withsubjects such as the thermodynamics of inclusion complexes,iodine spectroscopy, ozone kinetics, and molecular model-ing (11).

Initially, evaluation of the modules focused on studentcomputer usage and the student perspective of the modules,but over time it shifted to understanding how students col-laborated in the online environment and the role of the fac-ulty facilitator. To perform this evaluation we needed todevelop analysis tools that would allow us to describe andanalyze the types of student collaboration taking place viacomputer-mediated communication. Simultaneously, wesought to develop guidelines or recommendations for facultyfacilitation in an online environment. Thus, the work pre-sented here focused on

1. Describing how students and faculty interact during an on-line module.

2. Developing research based guidelines for effective facultyfacilitation in an online environment.

Research MethodsThe PCOL module investigated during the study was

“How Hot Is That Flame?” (11). The goal of the modulewas to reach consensus on a method of calculating a flametemperature of a particular fuel by applying mathematics andphysics to chemistry. The participants included 101 studentsfrom seven different colleges and universities and three fac-ulty facilitators—one experienced online facilitator and twonovices. The students and faculty were divided into 10 co-

hort groups that were composed of a faculty facilitator andthree student groups from different institutions. The experi-enced faculty facilitator functioned as a “project facilitator”who welcomed students to the project and substituted forcohort facilitators if they requested assistance during an ab-sence. Each cohort communicated via a discussion boardduring the five weeks in which the module took place.

The archived discussion boards were analyzed using twomethods. First, using transcripts of the discussion boards, eachposting was analyzed and coded by message type: original, firstfollow-up, second follow-up, or third and higher follow-up.Posts were also categorized by author—student group, cohortfacilitator, or project facilitator—and whether or not the post-ing contained a question. Each discussion board was then dia-grammed using Inspiration software.1 These diagrams allowedfor better visualization and subsequent description of the stu-dent–student and facilitator–student interactions.

The second method of analyzing the discussion boardsfocused on coding each posting using a scheme that identi-fied behaviors associated with collaboration, such as givingfeedback or challenging one’s reasoning, from the work ofJohnson and Johnson (12, 13). The coding scheme used wassimilar to that developed by Curtis and Lawson (14), whoincluded codes identifying social interaction and organiza-tion of group work along with Johnson and Johnson’s scheme.Two additional codes were generated in this study to distin-guish contributions that dealt with assigned academic tasksand those regarding technology issues. The codes associatedwith collaboration were grouped into higher-level categoriesof planning, contributing, seeking input, reflection and moni-toring, and social interaction as Curtis and Lawson (14) de-scribed. This higher-level analysis in conjunction with thecodes that were not associated with collaboration allowed fordescription of more general aspects of the discussion boardinteractions. A detailed explanation of the scheme is givenin Table A1 in the Appendix, including examples from thediscussion board transcripts.

Findings and DiscussionCohorts 1, 2, and 3 were chosen to exemplify the con-

versation that occurred on the discussion boards during thisproject. These cohorts represent a continuum of interactionfrom a case of sustained discussion and collaboration to acase of little discussion and collaboration. Figure 1 comparesthe percentage of total messages by message type (original,

Online Chemistry Modules:Interaction and Effective Faculty FacilitationLaura E. SlocumUniversity High School of Indiana, Carmel, IN 46032

Marcy Hamby Towns*Department of Chemistry, Ball State University, Muncie, IN 47306-0445; *00mhtowns@bsu.edu

Theresa Julia ZielinskiDepartment of Chemistry, Medical Technology, and Physics, Monmouth University, West Long Branch, NJ 07764-1898

Chemical Education Research

Research: Science and Education

www.JCE.DivCHED.org • Vol. 81 No. 7 July 2004 • Journal of Chemical Education 1059

first follow-up, second follow-up, and third and higher fol-low-up) that appeared within each cohort.

If effective collaboration and discussions occurred withina cohort during the project, then original postings would gen-erate follow-up postings, which one might assume would pro-duce a balanced ratio between original postings and third orhigher follow-up postings. Thus, by looking at each cohortin Figure 1 we can begin to form an impression of the typeof interactions that occurred. For example, in cohort 3 whichhad a novice faculty facilitator, over 60% of the messages wereoriginal postings, but only 15% third or higher follow-uppostings. This preponderance of original postings (nearly a4 to 1 ratio), likely indicates that few of the messages gener-ated more than one or two responses. However, in cohort 2which had an experienced online faculty facilitator, 42% ofthe postings were original messages and 31% were third andhigher follow-up messages, nearly a 4 to 3 ratio. This betterbalance of original and third or higher follow-up postingsindicates that discussion did take place, that in fact quite anumber of the third or higher follow-up postings were gen-erated in response to some of the original postings. Cohort1, which had a novice faculty facilitator, had nearly a 2:1 ra-tio of original postings to third and higher follow-ups, with50% original postings and 28% third and higher follow-uppostings. It initially appears that only cohort 2, the cohortthat had an experienced online facilitator, engaged in a pat-tern of sustained discussion and collaboration.

Discussion Board DiagramsFigures 2, 3, and 4 are the discussion board diagrams

for cohorts 1–3. The legend identifies the level of the post-ing (original, first follow-up, and so forth) and who gener-ated the posting. In each figure, the discussion begins at thelower left corner, with an original message labeled with thedate and author. Time runs vertically from bottom to top onthese diagrams, thus the conversation flows up. Postings onthe same day are placed horizontally in the order that theywere posted to the discussion board. The arrows that linkfollow-up messages back to the posting that inspired a re-sponse indicate a conversation thread. The project began on10/06 and ended 11/17.

Cohort 1, depicted in Figure 2, had two distinct con-versation threads; however, group C had no postings after10/25. In cohort 2, shown in Figure 3, there were also two

distinct conversation threads and the second conversationcontinued for 22 days from 10/26 to 11/17. The diagramfor cohort 3, displayed in Figure 4, contains one conversa-tion thread; the facilitator makes no postings after 10/26.

From the discussion board diagrams shown in Figures2–4, it is evident that the facilitator must maintain an activepresence in order for conversation to develop. When the fa-cilitator did not maintain an active presence, students merelycompleted the various assignments as noted by the many origi-nal postings. They simply posted their results without read-ing, reflecting, or generating responses to previous postings.The large number of original postings was conspicuous in co-hort 3 (Figure 4), where the cohort facilitator made an initialintroduction to the group on 10/09 and then did not postagain until 10/18. The messages between these two dates wereprimarily original messages. It was only when group G askeda question on 10/16 that the first follow-up message fromgroup H was posted on 10/17. The project facilitator’s mes-sage posted on 10/17 addressed both groups H and G, andinitiated a conversation between the student groups that the

Figure 2. Discussion board diagram for cohort 1.

Message Type

Per

cent

age

of T

otal

Mes

sage

s 70

Original 1st follow-up 2nd follow-up 3rd+ follow-up

60

50

40

30

20

10

0

Cohort 1Cohort 2Cohort 3

Figure 1. Percentage of total messages versus message type forcohort 1, 2, and 3 for the module “How Hot Is That Flame?”.

CF

CF*

CF*

B*

CF*

A*

B

PF

PF

B

CF*

AA A

CF* B*

CF*

A*

A*

AB

A

B CF B C

A*

C C

BB*

CF*

10/10

10/15

10/20

10/25

10/30

11/15

11/20

CF = Cohort FacilitatorPF = Project Facilitator

PA = University A PB = University B PC = University C

P * = Posting contains question

= Original posting

= 1st follow-up

= 2nd follow-up

= 3rd follow-up

Research: Science and Education

1060 Journal of Chemical Education • Vol. 81 No. 7 July 2004 • www.JCE.DivCHED.org

Figure 3. Discussion board diagram for cohort 2.

cohort facilitator reinforced on 10/18 and 10/19, but thenthe conversation dwindled again. After the cohort facilitator’slast message on 10/26, there were only original postings andthe cohort never did reach consensus on a flame temperaturenor on how to calculate a flame temperature.

However, the facilitator’s presence alone is not enoughto maintain conversations, as revealed by the comparison ofcohort 1 and 2. In cohort 1, the cohort facilitator initiatedthe first conversation (first follow-up message, 10/13). Thefacilitator asked specific questions related to three of the fourquestions in the students’ postings for assignment four andencouraged the students to elaborate further. This facilitator’sposting generated further conversation among the studentgroups. The facilitator posted a fourth follow-up response on10/20, again encouraging the students and once more ask-

ing further questions. Following this, the facilitator postedthree messages asking for responses from the other groups,but only one of them (10/26) was answered. Another con-versation began on 10/31 via a student group question; how-ever, in the middle of this conversation, there was a twelve-daygap where no one posted a message. Cohort 1’s discussionended with two original postings from different studentgroups (11/14 and 11/16); the third student group had notposted a message since 10/25. The final posting was a ques-tion from the cohort facilitator that went unanswered. Thiscohort also did not come to consensus on a flame tempera-ture or a method for calculating a flame temperature.

In cohort 2, a student group initiated the first conversa-tion on 10/11. Within this conversation another originalposting also generated a separate short conversation. The con-versation begun on 10/26 illustrates the role the facilitatorplayed in promoting and nurturing conversation and pro-vides insight into the importance of the facilitator’s role.

Numerous original student group postings were made incohort 2 on 10/26; half of these contained questions. One of

JF*

D

J

F*

E*F*

J

E

D J*

J

D E

DF

J

F

E

J*

E

F

E D*

J

D* J* EEF*

EJ

EJ J

E*

F* E E*

F*

E

10/10

10/15

10/20

10/25

10/30

11/5

11/10

11/15

12/05

F*

E*

J

= Original posting

= 1st follow-up

= 2nd follow-up

= 3rd follow-up

F = Cohort FacilitatorD = University DE = University EJ = University J

* = Posting contains question

G

PFHI

H

H

HG*

CF*PF*H

GHH

H* I*

I

H

CF G G G

I

GG I CF*

H CF*I

PF*

G*

10/10

10/15

10/20

10/25

10/30

11/5

11/10

11/15

G*

= Original posting

= 1st follow-up

= 2nd follow-up

= 3rd follow-up

CF = Cohort FacilitatorPF = Project Facilitator

PG = University G PH = University H PI = University I

P * = Posting contains question

Figure 4. Discussion board diagram for cohort 3.

Research: Science and Education

www.JCE.DivCHED.org • Vol. 81 No. 7 July 2004 • Journal of Chemical Education 1061

these original questions generated a student response that sameday. The following day, the facilitator replied, first with anencouraging word to the entire cohort, then comments to in-dividual groups, and concluded the posting in this manner:

Third, I need some help on deciding two things. 1) Dan’sgroup said that Cp was parabolic “as expected” but later some-one said that Cp is directly proportional to T. I don’t thinkthose statements are consistent. Which is correct? (Anybodycan help out here!) 2) Zanne’s groups said they got “close” tothe same result as Dori, Hannah, Kim, but you all talked aboutLOTS of different heat capacities-CO2, propane, N2-whichheat capacities did you use and what was the final tempera-ture that each group determined? We need to know a tem-perature we agree on for propane, so that we can expand ourcalculations next week! Please help me here!

The cohort 2 facilitator posted a message or responseno more than eight days apart and once a conversation be-gan, the facilitator both encouraged the students to work to-gether and continued to ask questions to keep all the groupsengaged in the conversation. The facilitator’s comments on11/3 addressed each group specifically and asked them ques-

tions that pointed out discrepancies in the approach eachgroup developed to calculate the flame temperature and en-couraged the student groups to resolve these differences.However, as with cohort 1 and 3, this cohort did not cometo consensus on the flame temperature for a specific fuel anddiscrepancies in the calculation method persisted.

These findings point toward the need for effective fac-ulty facilitation within each cohort. Just as a classroom in-structor provides additional understanding to students duringquestion and answer sessions or office hours, the online fac-ulty facilitator provides additional support to students on thediscussion board and guides the students to gain a greaterunderstanding of the concepts under investigation.

Coding Scheme Analysis of Discussion BoardsTable 1 presents the data obtained by using a coding

scheme that identified behaviors, collaborative and otherwise,within the postings for cohorts 1–3. In order to distinguishstudent behaviors from the conversation taking place acrossthe entire discussion board, the category percentages wererecalculated without the facilitator contribution in the “with-out facilitator posts” column.

Code identities: GS = Group skills; OW = Organizing work; IA = Initiating activities; HeG = Help given; FBG = Feedback given; SKR = Sharingknowledge and resources; Ch = challenging reasoning; Ex = Explain, elaborate, or summarize; HeS = Help sought; FBS = Feedback sought; Ef =Advocating effort; ME = monitoring group effort; SI = Social interaction; A= Assignment; T = Technology

Table 1. Analysis of Postings for Cohorts 1, 2, and 3 Where n Is the Number of Coded Discussion Board Postsand n’ Is the Total Number without Facilitator Coded Discussion Board Posts

1trohoC(n ;48= ’n )95=

2trohoC(n ;801= ’n )57=

3trohoC(n ;38= ’n )36=

roivaheBseirogetaC

sedoC edoC)%(

yrogetaC)%(

tuohtiWrotatilicaF)%(stsoP

edoC)%(

yrogetaC)%(

tuohtiWrotatilicaF)%(stsoP

edoC)%(

yrogetaC)%(

tuohtiWrotatilicaF)%(stsoP

gninnalP SG 9.5 6.4 6.9

WO 2.1 1.7 7.1 8.1 4.6 4 0. 4.2 21 0. 5.9

gnitubirtnoC GeH 6.3 7.3 2.1

GBF 3.8 3.8 6 0.

RKS 2.1 0 0. 2.1

hC 3.8 5.5 8.4

xE 3.41 7.53 3.73 1.12 6.83 04 0. 3.91 5.23 8.13

tupnIgnikeeS SeH 8.4 8.2 8.4

SBF 1.7 3.8 2.7

fE 9.5 8.71 71 0. 4.6 4.71 61 0. 8.4 9.61 9.51

ronoitcelfeRgnirotinoM

EM 8.4 8.4 0 0. 3.8 3.8 4 0. 4.2 4.2 0 0.

noitcaretnIlaicoS IS 5.51 5.51 6.81 1.01 1.01 8 0. 3.91 3.91 6.02

tnemngissA A 1.31 1.31 6.81 4.71 4.71 3.52 9.61 9.61 2.22

ygolonhceT T 9.5 9.5 8.6 8.1 8.1 7.2 0 0. 0 0. 0 0.

latoT 9.99 001 0. 001 0. 001 0. 001 0. 001 0.

Research: Science and Education

1062 Journal of Chemical Education • Vol. 81 No. 7 July 2004 • www.JCE.DivCHED.org

In each cohort the largest percentage of behaviors werecontributing behaviors: statements or utterances that contrib-uted to the completion of the task at hand by either givingfeedback (FBG) or help (HeG), sharing knowledge or re-sources (SKR), challenging each other’s reasoning (Ch), orexplaining a previously posted position (Ex). For example, thestatement: “We looked at your MathCad document, and wedon’t think you can estimate the final temperature of the flamefrom a Cp graph for propene. Do you guys think you shouldaccount for the nitrogen present in air?” was coded as chal-lenging (Ch) another’s reasoning. A statement such as “Weobserved that as the number of bonds increased, the higherthe temperature of the flame was for that fuel. Also the car-bon–oxygen bond is stronger than carbon–hydrogen bondsmaking it harder to break the bonds in propanol. Therefore,propanol has a higher flame temperature.” was coded as ex-plaining, elaborating, or summarizing (Ex) a previous posi-tion. Other prominent categories of behaviors and codesincluded seeking input, social interaction (SI), and postingacademic assignments (A). In cohorts 1 and 2 the combina-tion of contributing behaviors and seeking input behaviorsaccounted for over 50% of the utterances in the entire dis-cussion board. If the postings belonging to the cohort facili-tators and project facilitators were removed from the analysis,then in cohort 1 and 2 the percentage of contributing behav-iors increased slightly while in cohort 3 it decreased slightly.

In cohorts 1 and 3, contributing behaviors such as chal-lenging reasoning (Ch), explaining, elaborating, or summa-rizing (Ex), and giving feedback (FBG) were common, as werebehaviors associated with seeking input and social interac-tion (SI). However, we noted that in neither cohort did the stu-dents engage in reflection or monitoring behaviors.

In cohort 2, contributing behaviors such as challengingreasoning (Ch), explaining, elaborating, or summarizing (Ex),giving feedback (FBG), and giving help (HeG) occurred fre-quently. One student group posted messages that are examplesof monitoring the cohort’s efforts and achievements (ME):

After reading all of our postings, it is obvious that the cohortas a whole has come to similar conclusions...

It’s nice to see that [group H] posted their final answers. How-ever, we believe the point of assignment 10 was to come to aconsensus on how to determine the flame temperature. Wewould appreciate it very much if we could get some input asto how we (as a cohort!!) should determine the flame tem-perature. We have all gotten extremely different answers, so itonly makes sense that we all need to discuss what we shoulddo.

This online monitoring action is akin to the group pro-cessing activities described by Johnson and Johnson (13). Thepurposes of these activities are to recognize progress andachievements, and to consider what actions the group needsto continue doing, stop doing, and start doing in order tosustain and increase their efforts to reach a certain goal. Theabsence of monitoring activity by students in cohorts 1 and3 suggests one possible reason why the students did not cometo consensus on either the flame temperatures for a particu-lar fuel or the methodology for calculating a flame tempera-ture—the students themselves did not monitor their cohort’sprogress and did not recognize the importance of monitor-ing their cohort’s progress.

Combining the Coding Scheme and Discussion BoardDiagram Analysis

The coding scheme was also combined with the discus-sion board diagrams to produce a more detailed view of stu-dent–student and student–facilitator interactions. It wasinteresting to look at the collaboration that took place in thelongest discussion thread in each cohort. For example, in co-hort 2, the longest thread was 17 messages long beginningon 10/26 and ending 11/17. A tabulation of the codes forthis set of messages, including only the codes feedback sought(FBS), feedback given (FBG), challenging reasoning (Ch),explaining, elaborating, or summarizing (Ex), help sought(HeS), and help given (HeG), is shown in Table 2.

Within these 17 messages, 16 include utterances identi-fied with a feedback, help, or challenge–explain code. Thus,the discussion thread illustrated here can be described as hav-ing “feedback sought–feedback given” or “challenge–explain”cycles of interaction among cohort members. These types ofinteraction patterns are documented in the cooperative learn-ing literature and promote a number of beneficial results, in-cluding greater insight into the problem under consideration(14). In this particular set of exchanges in cohort 2, the stu-dents struggled to come to consensus as a cohort on how tocalculate a flame temperature for propane and propene, thecompounds used in assignment 8 and 9, respectively. The ex-cerpts from this thread focus on the details of calculating aflame temperature for combustion of propane and propene.The discrepancy centers on whether nitrogen that is presentin the reaction mixture should be included in the calculation.

Excerpts of Messages to Groups in Cohort 2

Facilitator Posting to Group E (11/3)

I thought it was very perceptive of you to include the N2, sinceit is part of the air! Good job! But there is a disagreement inyour document as to whether the H2O should be liquid orgas. Which is it?...I also wonder if the inclusion of the N2 isenough to account for the difference in flame temp for pro-pane btw [between] you and Group J? Remember they saidthe final T was 4837 K [Group E reported 2378 K]. Can youall check that out and let me know? Thanks!

Facilitator Posting to Group E (11/8)

We have read through everyone’s postings and we have a fewthings we want to discuss... Group J, We looked at yourMathCad document, and we don’t think you can estimate thefinal temperature of the flame from a Cp graph for propene.Do you guys think you should account for the nitrogen presentin air? We re-read the scenario, and it said it was combustedin a box filled with plain air. (i.e., 20% O2 and 80% N2).Please comment on our procedure as soon as possible.

Facilitator Posting to Group J (11/9)

Also in response to Group E’s question about including ni-trogen in our calculations [for the flame temperature of pro-pene], we were wondering exactly how you did this. Ourcomputer is unable to open your posted document. Due totime constraints we have decided to follow our model whichonly considers oxygen.

Research: Science and Education

www.JCE.DivCHED.org • Vol. 81 No. 7 July 2004 • Journal of Chemical Education 1063

Group E Posting to Cohort 2 (11/10)After group J posts their final flame temperatures, group

E writes:We believe the point of assignment 10 was to come to a con-sensus on how to determine the flame temperature. We wouldappreciate it very much if we could get some input as to howwe (as a cohort!!) should determine the flame temperature.We have all gotten extremely different answers, so it onlymakes sense that we all need to discuss what we should do.

Facilitator Posting to Cohort 2 (11/13)

We suggest our cohort look at the trends between our data inorder to form a consensus. Nitrogen is an important factor totake into account. However, the trends should be the samewhether or not nitrogen is used in the calculations. I hopethat everyone can reach a consensus based on observable trendsand agree that there is more than one way to calculate thetemperature of a flame.

Facilitator Posting to Cohort 2 (11/15)Since you are having trouble...why don’t you just post someinfo here? Group E: your group reported... 2462K [for pro-pene after recalculation]. Group J: reported... propene 5207K. Those are pretty BIG differences on the propene. Let’s as-sume that you both did the calculations correctly–there mustbe something significantly different in the models you used.Can you each outline your model in a few sentences and seeif you can identify the difference? Then you can decide whichmodel you like better, OK?

Group E Response to Facilitator (11/15) Group E responded with a summary of its approach to

calculating the flame temperature of propane. Students in thisgroup explained the inclusion of nitrogen in the calculationby stating:

We noticed that the reaction took place in air. It [the vessel]had not been evacuated and filled with oxygen. Therefore thereis nitrogen present. Air is composed of approximately 80%nitrogen and 20% oxygen. Because of this factor, there is atleast 20 moles of nitrogen present at the beginning and endof the reaction... We accounted for this in our calculationssince the nitrogen would absorb some of the heat and changethe final temperature.

Additional Group E Response to Facilitator (11/15)We think nitrogen needs to be used in the calculations becauseit has such a huge effect on the final flame temperature this isdue to the fact that there is 20 moles of nitrogen present. Thisis a large amount that absorbs a HUGE amount of heat. There-fore, this cannot be excluded from the calculations. We cannotagree with the models used by the other groups in our cohort.

Facilitators Foster Effective Online Discussion ThreadsWithin these excerpts there is evidence that the cohort

members explained and elaborated on their approach whilechallenging each other’s reasoning. They also were able tosupply feedback by considering each other’s approaches andanalyses. The facilitator was skillful in focusing student at-tention on their differing approaches to calculating the flame

temperature by citing discrepancies. The facilitator also sup-ported collaboration among cohort members by weaving to-gether previous postings, pointing out inconsistencies,challenging the student’s reasoning, and requesting feedback.One way of visualizing this exchange of messages is shownin Figure 5.

The challenge–explain cycles of interaction that devel-oped in this thread also provided opportunities to generatefeedback or help cycles as shown by the dotted lines in Fig-ure 5. The overall impact of the generation of these cycles

feedbackgiven

feedbacksought

challenge

explain,elaborate,summarize

helpgiven

helpsought

Figure 5. Challenge–explain cycles of interaction can also stimu-late feedback sought–given cycles and help sought–given cycles.

Table 2. Tabulation of Codes for Messages fromCohort 2 Involved in the Thread of 10/26 to 11/17

The “Message Author” column identifies a student group D, E, J, or F,the facilitator, as the author of the posting. An “X” means an utterancewas present in the message that corresponds to that particular code.

egasseMrohtuA

egasseMetaD

sadedoCGBFroSBF

sadedoCxErohC

sadedoCGeHroSeH

D tcO62 X

J tcO62 X

E tcO62 X X

F tcO72 X X X

J tcO03 X

E voN20

D voN20 X X

F voN30 X X X

E voN60 X

J voN70 X

E voN80 X X X

J voN90 X X

E voN01 X

J voN31 X

F voN51 X X

E voN51 X X

E voN51 X X

F voN71 X

Research: Science and Education

1064 Journal of Chemical Education • Vol. 81 No. 7 July 2004 • www.JCE.DivCHED.org

within the discussion was to support and to extend the threadof discussion.

We found similar discussion–interaction patterns in thelongest threads for cohorts 1 and 3 as well. Based upon theanalysis of these longest threads, it appears that facilitatorsin each cohort responded similarly. Thus, it appears that ef-fective online facilitation involves monitoring student postingsso that responses summarize or weave together previous postings,challenging the student’s reasoning by pointing out discrepan-cies, and requesting feedback.

Implications of Research for PracticeGiven this data and analysis what can we say to chemi-

cal education researchers and to chemists interested in con-ducting research on or implementing projects which requirestudent collaboration via computer-mediated communica-tion? For chemical education researchers, the methods devel-oped in this project of analyzing discussion boards may help themexamine electronic interactions in a more meaningful manner.The discussion board diagrams may help researchers (andpractitioners) visualize interactions that have taken place overthe duration of a project. The diagrams reveal interactionsthat go undetected in the usual analysis of a discussion boardtranscript. This method of diagramming could also be ex-tended to other means of electronic communication such aschat rooms and listservs. The coding scheme can help re-searchers identify and describe the presence or absence of col-laboration over the time of the project. It can also be used tohighlight specific behaviors observed during an online col-laboration. For chemistry faculty implementing online ac-tivities that require collaboration, the coding scheme mayallow them to identify behaviors for students that encourageand sustain collaboration.

The most important finding emerging from this studyis the clear example of effective online faculty facilitation. Itis one thing to say that faculty should facilitate student learn-ing, it is entirely another to document successful implemen-tation of those words in a classroom be it a conventional oronline environment. Within the longest thread of each dis-cussion board we found challenge–explain cycles of interac-tion among the students and facilitator. From our findingswe present the following recommendations to foster effec-tive online facilitation by faculty.

• Maintain an active online presence by regularly posting mes-sages.

• Monitor students’ progress and acknowledge their achieve-ments. Encourage students to generate postings trackingtheir progress through the module.

• Summarize previous posts, weave them together, and pointout discrepancies.

• Challenge the students’ reasoning. Ask questions that re-quire the students to resolve discrepancies or address dif-ferences of position or approach.

• Request responses from students and provide students withfeedback.

• Offer students support and encouragement.

Based upon our research findings, we believe that theserecommendations will help faculty and students to collaborate

more effectively in an online environment. We do recognize,however, that previous work in cooperative learning shows stu-dents need to be prepared to work in groups (13, 16–19).Even in the face of our research-based recommendations, stu-dents may not collaborate effectively if they have not beenprepared to do so. Thus, we also recommend that faculty en-gaging in projects requiring collaboration use the recommen-dations in the literature to help their students know what toexpect when they are asked to work in groups (13, 16–19).

Acknowledgments

Partial support for this work was provided by the Na-tional Science Foundation’s Division of Undergraduate Edu-cation through grant DUE #9950809.

Note1. The Inspiration program we used is produced by Inspira-

tion Software, Inc., 7412 SW Beaverton Hillsdale Highway, Suite102, Portland, OR 97225-2167; (800) 877-4292.

Literature Cited

1. Anthony, S.; Mernitz, H.; Spencer, B.; Gutwill, J.; Kegley, S.;Molinaro, M. J. Chem. Educ. 1998, 75, 322–324.

2. Glaser, R. E.; Poole, M. J. J. Chem. Educ. 1999, 76, 699–703.

3. Holmes, C.; Warden, J. J. Chem. Educ. 1996, 73, 325–331.4. Paulissee, K. W.; Polik, W. F. J. Chem. Educ. 1999, 76, 704–

708.5. Pence, L. E. J. Chem. Educ. 1999, 76, 697–699.6. Sauder, D.; Towns, M.; Stout, R.; Long, G.; Zielinski, T. J.

Chem. Educ. 1997, 74, 269–270.7. Sauder, D. G.; Towns, M. H.; Derrick, B.; Grushow, A.;

Kahlow, M.; Long, G.; Miles, D.; Shalhoub, G.; Stout, R.;Vaksman, M.; Pfeiffer, W. F. ; Weaver, G. C.; Zielinski, T. J.The Chemical Educator [Online] 2000, 5 (2), 77–82.

8. Towns, M.; Sauder, D.; Whisnant, D.; Zielinski, T. J. J. Chem.Educ. 2001, 78, 414–415.

9. Whisnant, D.; Lever, L. S.; Howe, J. J. J. Chem. Educ. 2000,77, 199–201.

10. Whisnant, D.; Lever, L. S.; Howe, J. J. J. Chem. Educ. 2000,77, 1648–1649.

11. Long, G. The PCOL Web site. http://pcol.ch.iup.edu/ (accessedApr 2004).

12. Johnson, D. W.; Johnson, R. T.; Smith, K. A. Active Learn-ing: Cooperation in the College Classroom; Interaction BookCompany: Edina, MN, 1991.

13. Johnson, D. W.; Johnson, R. T. In Handbook of Research forEducational Communications and Technology; Jonassen, D. H.,Ed.; Simon and Schuster Macmillan: New York, 1996; pp1017–1044.

14. Curtis, D. D.; Lawson, J. J. Journal of Asynchronous LearningNetworks [Online] 2001, 5 (1), 21–34.

15. Felder, R. M.; Brent, R. Cooperative Learning in TechnicalCourses: Procedures, Pitfalls, and Payoffs; Report ED377038,ERIC Document Reproduction Service, U.S. Department ofEducation’s Institute of Education Sciences, 1994.

16. Felder, R. M.; Brent, R. College Teaching 1996, 44 (2), 43–47.17. Millis, B. J.; Cottel, P. G. Cooperative Learning for Higher

Education Faculty; American Council on Education and OryxPress: Phoenix, AZ,1998.

Research: Science and Education

www.JCE.DivCHED.org • Vol. 81 No. 7 July 2004 • Journal of Chemical Education 1065

Table A1. Coding Scheme Used To Describe Utterances on the Discussion Boards, Based on Curtis and Lawson (14)

seirogetaCroivaheB sedoC noitpircseD elpmaxE

gninnalP SG deilppaedoccirenegA:sllikspuorGpuorgegaruocnetahtsnoisserpxeot

.ssenevisehocdnaytivitca

!!syug,bojdooG

WO puorggninnalP:krowgnizinagrOdnasksatderahsgnittes,krow

.senildaed

dellacdraobehtnonoitcesnoissucsidwenasierehTteguoynehW”.serutarepmeTemalFderahS“

trohocsihtnihtiweergadna,9tnemngissahguorht?ko,erehtsrewsnaruoytsop,stluserehtno

gnitubirtnoC GeH snoitseuqotgnidnopseR:gnivigpleH.srehtomorfstseuqerdna

lecxEdnadachtaMgnitsoprofsnoitcurtsnI”gnittamrof“ehtrednuelbaliavaerastnemucod

.lenapdnah-tfelehtnonoitces

GBF kcabdeefgnidivorP:gnivigkcabdeeF.srehtomorfslasoporpno

slacidarrosnoitahtsnoitseggusruoytahtknihtI-gnitseretniyllaersisemalfnidecudorpebthgim

naevahIesuacebemotesnessekamtidnadnasratsnislacidarseidutsohwdneirfremonortsa

!tohyllaererayehtemussaI—stenalp

RKS otsecruoserdnaegdelwonkgnirahS.srebmempuorgrehtotsissa

nidradnatsdlogehtsahtahtetisbeWasahTSINtaehgnidulcni,seititnauqcimanydomrehtsuoirav

.)nevigLRU(otgnitagivanyrT.seiticapac

hC ehtgnignellahC:srehtognignellahCdnasrebmemrehtofosnoitubirtnoc

.etabedniegagneotgnikees

ewdna,tnemucoddaChtaMruoytadekooleWfoerutarepmetlanifehtetamitsenacuoyknihtt’nod

Camorfemalfeht p uoyoD.eneporprofhpargnegortinehtroftnuoccadluohsuoyknihtsyug

?rianitneserp

xE gnitroppuS:gnitarobalerogninialpxEgniwollofylbissop(noitisopnwos’eno

.)egnellahca

sdnobforebmunehtsatahtdevresboeWemalfehtfoerutarepmetehtrehgiheht,desaercnisidnobnegyxo-nobracehtoslA.leuftahtrofsaw

tignikamsdnobnegordyh-nobracnahtregnorts,eroferehT.lonaporpnisdnobehtkaerbotredrah

.erutarepmetemalfrehgihasahlonaporp

tupnIgnikeeS SeH morfecnatsissagnikeeS:gnikeespleH.srehto

?suplehesaelpenoemosnaC

SBF kcabdeefgnikeeS:gnikeeskcabdeeF.decnavdanoitisopaot

rofdohtemevobaehtgnisunodedicedpuorgruOekildluowew;emalfehtfoerutarepmetehtgnidnif

ehttahtdohtemrehtoynadnasihtnotupniruoy.detagitsevnievahspuorgrehto

fE otsrehtognigrU:troffegnitacovdA.troffepuorgehtotetubirtnoc

!tupnis’esleenoyreverofgnitiaweraeW

/noitcelfeRgnirotinoM

EM stnemmoC:troffepuorggnirotinoMdnasessecorps’puorgehttuoba

.stnemeveihca

tahtsuoivbositi,sgnitsopruofollagnidaerretfAralimisotemocsahelohwasatrohoceht

.snoisulcnoc

laicoSnoitcaretnI

IS laicosyldneirF:noitcaretnilaicoS.rettap

niebrehtarhcumdluow,)eman(,rentrapyM.wonthgirdnalyraMnignizeerfsitisaadirolF

tnemngissA A .ksatcimedacadengissA hcussksatcimedacadengissaotsesnopsertnedutSafoerutarepmetehterusaemyllatnemirepxeot“sa

arehtieesudluoceno,leufaybdecudorpemalf”.retemitlumaroelpuocomreht

ygolonhceTeussI

T .eussierawtfosrorevreS nitaborcAhcnualotelbanullitssiretupmocruO.puorgs’ennaZybtnemucodehtdaerotredro

18. Nurrenbern, S. C. Experiences in Cooperative Learning: A Col-lection for Chemistry Teachers; Institute for Chemical Education,

University of Wisconsin-Madison: Madison, WI, 1995.19. Towns, M. H. J. Chem. Educ. 1998, 75, 67–69.

Appendix