7
The learning kit project: Software tools for supporting and researching regulation of collaborative learning Philip H. Winne a, * , Allyson Fiona Hadwin b , Carmen Gress a a Faculty of Education, Simon Fraser University, Burnaby, British Columbia, Canada V5A 1S6 b University of Victoria, Canada article info Available online xxxx Keywords: Collaborative learning environment gStudy Self-regulated learning Tracing abstract Computer-supported collaborative learning (CSCL) is a dynamic and varied area of research. Ideally, tools for CSCL support and encourage solo and group learning processes and products. However, most CSCL research does not focus on supporting and sustaining the co-construction of knowledge. We identify four reasons for this situation and identify three critical resources every collaborator brings to collaborations that are underutilized in CSCL research: (a) prior knowledge, (b) information not yet transformed into knowledge that is judged relevant to the task(s) addressed in collaboration, and (c) cognitive processes used to construct these informational resources. Finally, we introduce gStudy, a software tool designed to advance research in the learning sciences. gStudy helps learners manage cognitive load so they can re-assign cognitive resources to self-, co-, and shared regulation; and it automatically and unobtrusively traces each user 0 s engagement with content and the means chosen for cognitively processing content, thus generating real-time performance data about processes of collaborative learning. Ó 2007 Elsevier Ltd. All rights reserved. 1. Introduction Computer-supported collaborative learning (CSCL) is a dynamic area of research involving an assortment of methodologies, various theoretical and operational definitions, and several technological tools for investigating multiple collaborative structures (Gress, Hadwin, Page, & Church, this issue). Overall, CSCL environments aim to advance research on models of collaborative learning and facilitate learners 0 co-construction of knowledge (Koschmann, 2001; Salovaara & Järvelä, 2003). More specifically, CSCL interactive tools aim to encourage, support, and sustain solo and group regula- tion of collaboration, learning processes and products by prompt- ing, coaching and providing interactive feedback (Kirschner, 2004). Collaboration typically is operationalized as student-centered small group activities in which learners are supposed to develop skills for sharing the responsibility to be active, critical, creative co-constructors of learning processes, and products. Conditions that facilitate effective collaborative processes include, for exam- ple, positive interdependence, positive social interaction, individ- ual and group accountability, interpersonal and group social skills, and group processing (Johnson & Johnson, 1989; Johnson & Johnson, 1999; Kreijns, Kirschner, & Jochems, 2003). Some CSCL software tools attempt to support these kinds of engagements. Examples are awareness tools designed to support positive social interaction (Carroll, Neale, Isenhour, Rosson, & McCrickard, 2003) and negotiation tools designed to support group social skills and discussions (Beers, Boshuizen, Kirschner, & Gijselaers, 2005). De- spite much activity in the CSCL field, there is relatively little re- search on how types of tools support and sustain productive collaboration (Gress, Hadwin, Page, & Church, this issue; Hadwin, Gress, Page, & Ross, 2005). We identify four reasons for this situation. First, much research in CSCL focuses on developing and testing technologically-based tools (e.g., text chat tools, conferencing tools, email systems, and so forth) for sharing information (Gress, Fior, Hadwin, & Winne, this issue). These tools provide a means for collaborating online, but facilitating research about how and why collaboration takes the shape(s) it does and has the effect(s) it does is a much less the goal of these projects. Second, a review of the CSCL literature (Gress, Hadwin, Page, & Church, this issue) uncovered multiple ‘‘modes” for collaboration. Some describe students working asyn- chronously on individual contributions towards one document, others portray students working asynchronously on one document and reflecting on their collaborators’ contributions, and still others describe students contributing to a shared document in a synchro- nous environment. It may be that each model of collaboration is best suited to a particular type of task and pedagogical approach, though this is not demonstrated. Notwithstanding, tools for collab- orating do (and probably should) differ dramatically depending upon instructional goals, tasks, and tools available to learners. Third, the extensive educational literature on cooperative and 0747-5632/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.chb.2007.09.009 * Corresponding author. Fax: +1 778 782 4203. E-mail address: [email protected] (P.H. Winne). Computers in Human Behavior xxx (2008) xxx–xxx Contents lists available at ScienceDirect Computers in Human Behavior journal homepage: www.elsevier.com/locate/comphumbeh ARTICLE IN PRESS Please cite this article in press as: Winne, P. H. et al., The learning kit project: Software tools for supporting and researching ..., Computers in Human Behavior (2008), doi:10.1016/j.chb.2007.09.009

The Learning Kit Project Software Tools For Collaborative Learning

Embed Size (px)

DESCRIPTION

The learning kit project- Software tools for ... collaborative learning

Citation preview

Page 1: The Learning Kit Project  Software Tools For     Collaborative Learning

Computers in Human Behavior xxx (2008) xxx–xxx

ARTICLE IN PRESS

Contents lists available at ScienceDirect

Computers in Human Behavior

journal homepage: www.elsevier .com/locate /comphumbeh

The learning kit project: Software tools for supporting and researchingregulation of collaborative learning

Philip H. Winne a,*, Allyson Fiona Hadwin b, Carmen Gress a

a Faculty of Education, Simon Fraser University, Burnaby, British Columbia, Canada V5A 1S6b University of Victoria, Canada

a r t i c l e i n f o

Available online xxxx

Keywords:Collaborative learning environmentgStudySelf-regulated learningTracing

0747-5632/$ - see front matter � 2007 Elsevier Ltd. Adoi:10.1016/j.chb.2007.09.009

* Corresponding author. Fax: +1 778 782 4203.E-mail address: [email protected] (P.H. Winne).

Please cite this article in press as: Winne, Pin Human Behavior (2008), doi:10.1016/j.c

a b s t r a c t

Computer-supported collaborative learning (CSCL) is a dynamic and varied area of research. Ideally, toolsfor CSCL support and encourage solo and group learning processes and products. However, most CSCLresearch does not focus on supporting and sustaining the co-construction of knowledge. We identify fourreasons for this situation and identify three critical resources every collaborator brings to collaborationsthat are underutilized in CSCL research: (a) prior knowledge, (b) information not yet transformed intoknowledge that is judged relevant to the task(s) addressed in collaboration, and (c) cognitive processesused to construct these informational resources. Finally, we introduce gStudy, a software tool designedto advance research in the learning sciences. gStudy helps learners manage cognitive load so they canre-assign cognitive resources to self-, co-, and shared regulation; and it automatically and unobtrusivelytraces each user0s engagement with content and the means chosen for cognitively processing content,thus generating real-time performance data about processes of collaborative learning.

� 2007 Elsevier Ltd. All rights reserved.

1. Introduction

Computer-supported collaborative learning (CSCL) is a dynamicarea of research involving an assortment of methodologies, varioustheoretical and operational definitions, and several technologicaltools for investigating multiple collaborative structures (Gress,Hadwin, Page, & Church, this issue). Overall, CSCL environmentsaim to advance research on models of collaborative learning andfacilitate learners0 co-construction of knowledge (Koschmann,2001; Salovaara & Järvelä, 2003). More specifically, CSCL interactivetools aim to encourage, support, and sustain solo and group regula-tion of collaboration, learning processes and products by prompt-ing, coaching and providing interactive feedback (Kirschner, 2004).

Collaboration typically is operationalized as student-centeredsmall group activities in which learners are supposed to developskills for sharing the responsibility to be active, critical, creativeco-constructors of learning processes, and products. Conditionsthat facilitate effective collaborative processes include, for exam-ple, positive interdependence, positive social interaction, individ-ual and group accountability, interpersonal and group socialskills, and group processing (Johnson & Johnson, 1989; Johnson &Johnson, 1999; Kreijns, Kirschner, & Jochems, 2003). Some CSCLsoftware tools attempt to support these kinds of engagements.Examples are awareness tools designed to support positive social

ll rights reserved.

. H. et al., The learning kit phb.2007.09.009

interaction (Carroll, Neale, Isenhour, Rosson, & McCrickard, 2003)and negotiation tools designed to support group social skills anddiscussions (Beers, Boshuizen, Kirschner, & Gijselaers, 2005). De-spite much activity in the CSCL field, there is relatively little re-search on how types of tools support and sustain productivecollaboration (Gress, Hadwin, Page, & Church, this issue; Hadwin,Gress, Page, & Ross, 2005).

We identify four reasons for this situation. First, much researchin CSCL focuses on developing and testing technologically-basedtools (e.g., text chat tools, conferencing tools, email systems, andso forth) for sharing information (Gress, Fior, Hadwin, & Winne,this issue). These tools provide a means for collaborating online,but facilitating research about how and why collaboration takesthe shape(s) it does and has the effect(s) it does is a much lessthe goal of these projects. Second, a review of the CSCL literature(Gress, Hadwin, Page, & Church, this issue) uncovered multiple‘‘modes” for collaboration. Some describe students working asyn-chronously on individual contributions towards one document,others portray students working asynchronously on one documentand reflecting on their collaborators’ contributions, and still othersdescribe students contributing to a shared document in a synchro-nous environment. It may be that each model of collaboration isbest suited to a particular type of task and pedagogical approach,though this is not demonstrated. Notwithstanding, tools for collab-orating do (and probably should) differ dramatically dependingupon instructional goals, tasks, and tools available to learners.Third, the extensive educational literature on cooperative and

roject: Software tools for supporting and researching ..., Computers

Page 2: The Learning Kit Project  Software Tools For     Collaborative Learning

2 P.H. Winne et al. / Computers in Human Behavior xxx (2008) xxx–xxx

ARTICLE IN PRESS

collaborative leaning (e.g., see Abrami, Lou, Chambers, Poulsen, &Spence, 2000; O0Donnell & King, 1999; Slavin, 1999) is often ig-nored in designing CSCL environments, tools, and research. As a re-sult, research about CSCL and research about collaboration ingeneral share too little ground, particularly regarding supports orinstruction. Fourth, researching how individuals and groups learnto collaborate and support collaborators poses significant chal-lenges for measurement and evaluation (Gress, Fior, Hadwin, &Winne, this issue).

2. Supporting software tools for researching collaborativelearning

One goal of the Learning Kit Project is to research how studentsadopt and adapt strategies for solo and collaborative learning. Ourwork extends along a continuum from conventional models of self-regulated learning (SRL) as (a) a solo activity to (b) learning in agroup to (c) shared-collaboration of learning that emphasizes col-lective negotiation and regulation of task understanding, goal set-ting, planning, and enacting strategies.

The bulk of research on collaborative learning in classrooms andas supported by software technologies has appropriately focusedon features of collaboration per se; that is, the nature and patternsof exchanges of information among collaborators (see O0Donnell &King, 1999). These efforts have revealed a great deal. Nonetheless,we posit that models of collaboration are misspecified. Specifically,little and sometimes no attention has been focused on three criticalresources every collaborator brings to collaborations: (a) priorknowledge, (b) information not yet transformed into knowledgethat is judged relevant to the task(s) addressed in collaboration,and (c) cognitive processes used to construct these informationalresources. The logic upon which our conjecture rests is as follows.

1. In research on solo learning, measures of prior knowledge areoften the most potent variables affecting outcomes. While col-laboration may do much to fill gaps in an individual0s knowl-edge and to stimulate recall of an individual0s knowledge thatotherwise would not be brought to bear, one collaborator0s gainin this respect often depends on another0s knowledge. That is,collaborators0 knowledge as a group is almost certainly greaterthan any one collaborator0s knowledge. Those with less knowl-edge about a particular sector of the collaborative task benefitfrom group mates0 prior knowledge. That knowledge may beabout the task, the content, or the collaborative process itself.

2. Information that a collaborator can access but which is not yetanyone0s knowledge—information that can be contributed tothe group but is only partly understood by the contributor—may be a powerful resource in collaboration. This is becauseeach member0s partial knowledge may, in concert with contri-butions of such information made by others in the group, gen-erate a synergy that boosts the group0s productivity past acritical threshold. Through collaboration, information that isinitially no one0s knowledge may become knowledge forgedwithin the group.

3. How learners learn—the tactics and strategies learners knowand apply to transform information into knowledge—tends tobe stable. The simplest demonstration of this is the considerableeffort that must be spent to teach learners new tactics andstrategies for learning and, once these are learned, the addi-tional effort that must be spent to coax learners to use thosenewly acquired tactics and strategies. Consequently, the pro-cesses that each learner typically uses to learn very likely arecarried over to the collaborative setting. Furthermore, therepresumably are tactics and strategies that learners collabora-tively develop and refine to engage in collaboration itself.

Please cite this article in press as: Winne, P. H. et al., The learning kit pin Human Behavior (2008), doi:10.1016/j.chb.2007.09.009

While our first claim is empirically justified, we acknowledgethe second and third are speculations, though they have consider-able collateral support. What would be required to test thesepropositions in the context of CSCL and lend empirical addressto the possibility that models of collaborative learning are mis-specified? In this article, we offer a partial answer by describingan advanced software learning environment, called gStudy, thatwe and colleagues are developing (see Winne, Hadwin et al.,2006; and also Winne, Nesbit, et al., 2006; http://www.learning-kit.sfu.ca). gStudy software harnesses a platform for supportingsolo SRL to support collaborative learning. Our explicit goal indesigning collaborative tools and structures has been to supportstudents in learning to regulate collaborative learning activitiesand tasks.

Since this paper overviews and introduces software tools dis-cussed more deeply within other papers in this special issue, weprovide illustrations of how each tool can be used to facilitate col-laborative activity. In depth discussion of those collaborativeenterprises is found in the other papers in the special issue.

3. gStudy

3.1. Overview

gStudy (Winne, Hadwin, et al., 2006) is a state-of-the-art, cross-platform software system that puts into practice proposals Winne(1992) made about using software to substantially extend researchin the learning sciences. gStudy is a shell in which a learner or aninstructional designer can create or import content about almostany topic. (Topics defined by enacting physical skills, such as play-ing a piano or dissecting a frog, are excluded.) Information abouttopics is rendered using the hypertext markup language (HTML)in forms including text, diagrams, photos, charts, tables, audioand video clips—that is, the information formats common to hard-copy library resources and on the Internet. A unified collection ofthese materials is called a learning kit.

gStudy provides cognitive tools for learners to create, share, andexchange information objects. Every information object is linked toa file, data the learner selects within a file, or data the learner se-lects within a remote web site outside the learning kit. Each toolhas been designed, as much as possible, to instantiate research thatdemonstrates using the tool will positively influence solo and col-laborative learning and problem solving. In this article, we high-light features designed to promote collaborative learning inparticular but these features are seamlessly interwoven with fea-tures designed to promote solo learning. Hadwin, Oshige, Gress,and Winne (this issue) describe some specific models of collabora-tion that can be supported in the gStudy software, based on threedifferent views of regulation of learning.

Fig. 1 shows one of gStudy0s views, the browser view, onto alearning kit. The kit may belong to a learner, to a collection oflearners, or to a learner and instructor. In Fig. 1, all of gStudy0s pan-els are exposed to show search, concept maps, the catalog of kits(including the one being viewed), the selected kit0s table of con-tents, and a panel that identifies information objects that arelinked to data in the section of the kit that is in view. In practice,it would be rare that a student exposed all of gStudy’s panelssimultaneously.

3.2. Information objects

In gStudy, each information object is characterized by metadatathat specify: author of the object, date created, and date modified.This allows identifying each information object by one or several ofthese characteristics. It means a student can share objects and

roject: Software tools for supporting and researching ..., Computers

Page 3: The Learning Kit Project  Software Tools For     Collaborative Learning

Fig. 1. gStudy main window several panels displayed.

P.H. Winne et al. / Computers in Human Behavior xxx (2008) xxx–xxx 3

ARTICLE IN PRESS

information about those objects with one another. It also meansthat students who are co-constructing a collaborative learning kitcan review information about objects added to the kit, who addedthe object and when it was added. Types of information objectsimplemented in gStudy and methods for creating them are de-scribed in the following section.

3.3. Notes

Learners can elaborate on any source information by creating anote. The primary method for creating a note is to select (click anddrag the cursor across) text, a region within a graphical display, ora frame in an audio or video clip; then right-click (in the Windowsenvironment or control-click under Apple OS X) to display a con-textual menu and choose ‘‘Link to new note”.

Fig. 2. A note wi

Please cite this article in press as: Winne, P. H. et al., The learning kit pin Human Behavior (2008), doi:10.1016/j.chb.2007.09.009

Notes record information in a template that is a schema for var-ious forms of information (see Fig. 2). At a minimum, a templateconsists of a text field that ‘‘titles” the note. Beyond this, a templatecan include (a) fields for recording text, (b) sliders for rating fea-tures of interest, (c) checkboxes that allow the learner to enumer-ate one or more items in a list, (d) radio buttons that provide forselecting one and only one attribute within a set, (e) a field wherethe learner can attach files, such as a text file or a picture file, and(f) instructions or labels, (e.g., ‘‘Rate the importance of this infor-mation by dragging the slider”).

Templates can be designed by someone else—the author of alearning kit, a teacher, or a collaborator. As well, a learner can con-struct new templates to suit a particular need. For example, collab-orators may develop a template for sharing information andunderstandings about a task (task analysis note) or a template

th template.

roject: Software tools for supporting and researching ..., Computers

Page 4: The Learning Kit Project  Software Tools For     Collaborative Learning

4 P.H. Winne et al. / Computers in Human Behavior xxx (2008) xxx–xxx

ARTICLE IN PRESS

for providing feedback on each other0s writing (peer feedbacknote). In a shared kit, students can add to and revise each other0snote templates as well as the content of notes. Students using indi-vidual learning kits can exchange note templates and instantiatednote objects with one another. For example, in a Women0s studiescourse, students might develop a template for critiquing currentmedia representations from a Feminist perspective. That templatecan be shared amongst a group working on a class project to iden-tify themes across a range of media presentations. (See Hadwin,et al., this issue, for further discussions on collaborative works.)

3.4. Glossary entries

Glossary entries are information objects that record data aboutfundamental elements of the information in a domain of knowl-edge. Each glossary entry is recorded according to a template (sim-ilar to a note template). For example, a glossary entry templatemight be designed to record essential data about an element inthe periodic table of the elements such as its symbol, atomic num-ber, atomic mass, density, group, series, and so on. Another glos-sary entry template might be designed to record data about thediscovery of each element: who discovered it, the date of discov-ery, method of discovery, etc. The method for creating a glossaryentry is the same as for a note. Similar to notes and other objects,students can co-construct a glossary in a shared kit, or exchangetheir interpretations of terms and concepts by sending each otherspecific glossary notes.

3.5. Strategy library

The strategy library is a set of pre-stocked notes and note tem-plates providing information about a range of learning strategies(see Fig. 3). Each template describes the strategy, explains whento use it, why it helps, and provides examples. Students can editeach strategy note, add new strategy notes, and delete ones theydeem ineffective. Collectively, students can co-construct a libraryof strategy notes they judge to work really well for them in the areawhere they are working. Alternatively, students can share strategy

Fig. 3. Strateg

Please cite this article in press as: Winne, P. H. et al., The learning kit pin Human Behavior (2008), doi:10.1016/j.chb.2007.09.009

objects they create solo and review strategies peers have tried andthen choose whether to add that strategy to their individual strat-egy library.

3.6. Labels

Learners often categorize information according to variousattributes of a personal or task-related nature, e.g., ‘‘confusing”,‘‘surprising”, ‘‘needs more study!” and the like. gStudy provides alabel tool to serve this function. Labels are created in the sameway as notes (drag over content and right/control click) and orga-nized in a tree (outline). When students collaborate on a task,labeling can become a primary tool for organizing and assigningspecific roles and tasks amongst members in the group. For exam-ple, after completing a text-based chat discussion about a collabo-rative project, students can review the chat log, highlight specifictasks and goals that were discussed, and label them with the nameof the group member who will follow up on that item.

3.7. Search

gStudy provides a sophisticated search tool for locating infor-mation in one or multiple learning kits (see Fig. 4). Searches canbe simple, such as identifying where every occurrence of a termappears in one learning kit. Searches also can be complex Booleanqueries that seek data within specific kits and examine a particularkind of information object, such as notes.

To search for information, the learner clicks a button in gStudy0stoolbar. This opens a window in which the learner titles the searchquery and designs the search to be carried out. On clicking‘‘Search”, gStudy constructs a table to display every occurrence ofinformation that satisfies the search query. Various metadata thatdescribe each ‘‘hit” are identified in the rows of this table. Clickingon a row displays the result in context. Fig. 4 shows a search queryand a table of results. In collaborative work, searches can, forexample, identify objects authored by each group member andthereby help the group monitor and review contributions to thecollective project.

y Library.

roject: Software tools for supporting and researching ..., Computers

Page 5: The Learning Kit Project  Software Tools For     Collaborative Learning

Fig. 4. Search panel.

P.H. Winne et al. / Computers in Human Behavior xxx (2008) xxx–xxx 5

ARTICLE IN PRESS

3.8. My documents

This tool is for learners to write essays, lab reports, and othercompositions. It is a basic HTML editor disguised as a word proces-sor called ‘‘My Documents”. Once a document is finalized, the lear-ner can save these information objects in the learning kit. Likeother gStudy objects, students can share documents and use themto build a collaborative studying kit or project kit.

3.9. Chat

Using a near synchronous or true synchronous chat tool, learn-ers collaborate online. As learners chat, they construct a record thatsupports later review and reflection about their conversation andinformation objects they shared. Our chat tool (see Fig. 5) can beconfigured to provide prompts and roles to guide learners in their

Fig. 5. Chat w

Please cite this article in press as: Winne, P. H. et al., The learning kit pin Human Behavior (2008), doi:10.1016/j.chb.2007.09.009

collaborative work (see Morris, Church, Hadwin, Gress, & Winne,this issue). For example, chat roles, prompts and scripts might tar-get kinds of text processing activities found in reciprocal teaching(summarizer, questioner, clarifier, predictor). Alternatively, roles,scripts and prompts might emphasize different aspects of theself-regulatory cycle such as task understanding, goal setting/plan-ning, enacting the task, and reviewing, adapting, and revising pro-cesses. Chat can also be used to guide students in applying peerreview collaboration strategies by prompting students to construc-tively comment on different aspects of a writing project andprompting the author to direct responses toward clarificationquestions and elaborations rather than reactive responses. It alsoprovides a channel for learners to share information objects, suchas notes, glossaries, and essays, with collaborators. The learnerdrags an information object to the text entry field and drops it todistribute it to all participants in the chat.

indow.

roject: Software tools for supporting and researching ..., Computers

Page 6: The Learning Kit Project  Software Tools For     Collaborative Learning

6 P.H. Winne et al. / Computers in Human Behavior xxx (2008) xxx–xxx

ARTICLE IN PRESS

3.10. Concept map

Every information object created in gStudy is linked to some-thing, either the kit as a whole, another information object, or asubset of information within an information object. When the lear-ner wishes, information objects and links among them can be dis-played as a concept map, i.e., a node-link graph (see Fig. 6). Aconcept map can represent all the information objects in a learningkit or be filtered to display a subset of them. As well as showing al-ready existing information objects, learners can construct informa-tion objects in the concept map tool by identifying the type ofinformation object in a palette, then drawing in the concept mapwhere they want to place that object in the concept map space.Links between information objects also can be created in the con-cept map. Concept mapping offers a platform for students to sharewith one another complex representations and perceived relation-ships amongst concepts (glossary notes) or strategies (strategynotes), for example. Furthermore, concept maps can be harnessedby collaborative groups as a tool for reflecting upon collaborativeprogress and individual contributions to collective projects.

3.11. Kits

Kits may be pre-configured by an author (instructional de-signer, teacher, commercial author) to conform to a particularmodel of information presentation. When students engage in sololearning activities, they use a kit and supplement it with their owninformation objects. gStudy also allows learners to create theirown kits and, through the Kit Management System, students canshare entire kits with one another by checking them into a repos-itory and taking turns to check out collaborators0 kits. Sharingmight constitute (1) taking turns with a single shared kit that in-cludes contributions from each collaborator or (2) making an indi-vidual0s kit accessible to other collaborators (see Hadwin, et al., thisissue).

3.12. Analysis

An additional benefit gStudy offers learners and researchers isits ability to automatically and unobtrusively trace each user0sengagement with the content as well as the methods learnerschoose for cognitively processing content. gStudy records log dataabout every selection and modification made in information ob-jects (e.g., notes, strategies, glossary notes, labels), changes in alearner0s ‘‘view” onto the information in a learning kit (e.g., select-

Fig. 6. Conce

Please cite this article in press as: Winne, P. H. et al., The learning kit pin Human Behavior (2008), doi:10.1016/j.chb.2007.09.009

ing the strategy library or the chat tool), specific tools selected,scrolling actions, editing actions (e.g., adding new content to anote) and links a learner creates between content and other infor-mation objects. In addition, gStudy logs, for example, options cho-sen in menus, buttons clicked, kits selected, windows opened andclosed, and chat discussion logs viewed. Every logged event istime-stamped.

The data are formatted according to an XML (extensible markuplanguage) schema that supports detailed analyses of occurrence,frequency, sequence, pattern, and other qualities of how learnersstudy (Winne, Gupta, & Nesbit, 1994). Targets of analyses can be,for example, (a) processes an individual uses to learn new skillsor information, (b) content learners select for further operationor ignore, (c) information they choose when asked or assigned topick a side and debate, and (d) learners0 beliefs about the credibil-ity of information presented in various ways. When students col-laborate, gStudy collects data about which objects were sharedand added by individuals and how information objects are inte-grated in the kit. These kinds of data about which content learnersstudy and how they study it is key to corroborating models aboutwhat goes on when learners learn solo and collaboratively. To ourknowledge, research has not attempted to examine the weave ofindividual and collaborative activity in collaborative learning pro-jects. Data collected in gStudy affords opportunities to examinethis interplay of SRL and social activity across time and a rangeof academic tasks.

To analyze traces, our team designed Log Analyzer (Hadwin,Winne, Nesbit, & Joulovian, 2005) which inspects gStudy0s XMLlogs of a learner0s or multiple learners0 studying session(s). LogAnalyzer computes statistics such as (a) frequencies of events(e.g., how many labels were made), (b) properties of event se-quences (e.g., length, duration, density of information), and (c)properties of patterns of learning events (strategies) by analyzingtransition matrices of traces in terms of graph theoretic methods(e.g., cohesion within a pattern, structural equivalence of two pat-terns) (Winne et al., 1994).

Trace data are proximal indicators of particular informationprocessing and can be counted to generate ratio scales of cognitiveactivities. Trace-based methodology, now emerging in someempirical studies, does not interrupt cognitive processing like athink-aloud can and does not rely on learners0 fallible memoriesof how they studied or summations of how they study in general.Thus, trace data can provide a powerful complement to otherforms of data that characterize how learners learn and self-regu-late collaborative learning. Real-time analysis of traces (now being

pt map.

roject: Software tools for supporting and researching ..., Computers

Page 7: The Learning Kit Project  Software Tools For     Collaborative Learning

P.H. Winne et al. / Computers in Human Behavior xxx (2008) xxx–xxx 7

ARTICLE IN PRESS

implemented) would provide a basis for characterizing individualtactics, collaborative activity, and patterns of individual versus col-laborative engagement. Feedback could be provided immediatelyand in context about whether tactics are applied, how long it takesto apply each, and conditional probabilities that the learnerchooses some follow-on tactics over others conditional on a priortactic having been used. Scores can be presented in terms of fre-quency of use of tactics, the conditional probability that a tacticis used when a cue signals it should be applied, whether tacticsare applied randomly or according to a pattern, and the degree towhich the pattern of study tactics is similar (or appropriately dis-similar) across tasks and studying episodes (Winne et al., 1994).

In addition to logging events within individual or collaborativekits, gStudy also records logs of students0 text-based chat discus-sions (chat logs). Chat logs provide information regarding idea ex-change and formulation, use of collaborative scaffolds, such asroles and prompts, and a record of objects that were sharedthrough chat. Analyzing chat logs provides opportunities to tracepatterns of discourse and transitions in idea as collaborators at-tempt to co-construct shared meanings, plans, and reflections withrespect to a given task.

4. How gStudy contributes to research on collaboration

A challenge in researching collaborative learning has been gath-ering data essential to modeling cognitive and motivational vari-ables that generate collaborative processes and characterize howthese processes support collaborators as they construct productsachieved by collaboration. Unless collaborators are physically adja-cent, all their work—solo and collaborative—is rarely captured.gStudy can fill this gap. When data that gStudy gathers are mergedwith data collected using other instrumentation, such as videorecordings and self-reports gathered at various times (before, dur-ing and after collaboration), gStudy fine-grained, time-stampedindicators of these variables offers researchers opportunities to tri-angulate their interpretations of collaboration.

Other opportunities are afforded because gStudy0s records ofchats are made available as resources. These chat records can bemined by researchers and collaborators alike. This adds opportuni-ties to investigate issues such as whether previously generated col-laborative products overly anchor future collaborations, andwhether groups can augment collaborative effectiveness by re-vis-iting previous collaborative episodes.

Because collaborators can share an information objects withmembers of their group, gStudy also affords opportunities to inves-tigate the roles of information resources in shaping collaborationand the effects of sharing different types of information resourceson the flow and effectiveness of collaborative activities. Marryingdata that trace key features of collaboration with, for example, col-laborators0 perceptions of flow and effectiveness can better reflectrhythms in collaboration and the ways collaborators orchestratetheir process.

In sum, the kinds of and extent of data gStudy provides abouthow learners collaborate, how they interleave solo work withincollaborative episodes, and the very information that is the focusof collaboration will help researchers observe variance in collabo-ration and identify factors that influence that variance. Subsequentarticles in this special issue elaborate on these themes.

Please cite this article in press as: Winne, P. H. et al., The learning kit pin Human Behavior (2008), doi:10.1016/j.chb.2007.09.009

Acknowledgement

Support for this work was provided by grants to Philip H. Winnefrom the Social Sciences and Humanities Research Council of Can-ada (410-2002-1787; 512-2003-1012, R. Azevedo, A. F. Hadwin, S.Lajoie, J. Nesbit, & V. Kumar, -Co-Investigator), the Canada ResearchChair program, and Simon Fraser University; and to Allyson Had-win from the Social Sciences and Humanities Research Council ofCanada (410-2001-1263).

References

Abrami, P. C., Lou, Y., Chambers, B., Poulsen, C., & Spence, J. C. (2000). Why shouldwe group students within-class for learning? Educational Research & Evaluation,6, 158.

Beers, P. J., Boshuizen, H. P. A. E., Kirschner, P. A., & Gijselaers, W. H. (2005).Computer support for knowledge construction in collaborative learningenvironments. Computers in Human Behavior, 21, 623–643.

Carroll, J. M., Neale, D. C., Isenhour, P. L., Rosson, M. B., & McCrickard, D. S. (2003).Notification and awareness: Synchronizing task-oriented collaborative activity.International Journal of Human–Computer Studies, 58, 605.

Gress, C. L. Z., Fior, M., Hadwin, A. F., & Winne, P. H. (this issue). Measurement andassessment in computer supported collaborative learning. Computers & HumanBehavior.

Gress, C. L. Z., Hadwin, A. F., Page, J., & Church, H. (this issue). A review of computer-supported collaborative learning: Informing standards for reporting CSCLresearch. Computers in Human Behavior, doi:10.1016/j.chb.2007.05.012.

Hadwin, A. F., Gress, C. L. Z., Page, J., & Ross, S. P. (2005). Computer supportedcollaborative work: A review of the research 1999–2004. Paper presented at theannual meeting of the Canadian society for the study of education, London, ON.

Hadwin, A.F., Oshige, M., Gress, C.L.Z., & Winne, P.H. (this issue). Innovative ways forusing gStudy to orchestrate and research social aspects of self-regulatedlearning. Computers in Human Behaviour, doi:10.1016/j.chb.2007.06.007.

Hadwin, A. F., Winne, P. H., Nesbit, J. C., & Joulovian, T. (2005). Log analyzer: Atoolkit for analyzing gStudy log data and computing transition metrics (version2.0). University of Victoria, Victoria, British Columbia, Canada.

Johnson, D. W., & Johnson, R. T. (1989). Cooperation and learning: Theory andresearch. Edina, MN: Interaction Book Company.

Johnson, D. W., & Johnson, R. T. (1999). Learning together and alone: Cooperativecompetitive and individualistic learning (5th ed.). Boston, MA: Allyn & Bacon.

Kirschner, P. A. (2004). Design, development, and implementation of electroniclearning environments for collaborative learning. Educational TechnologyResearch & Development, 52, 39–46.

Koschmann, T. (2001). Revisiting the paradigms of instructional technology. Paperpresented at the annual conference of the Australasian society for computers inlearning in tertiary education (ASCILITE 2001), Melbourne, Australia.

Kreijns, K., Kirschner, P. A., & Jochems, W. (2003). Identifying the pitfalls for socialinteraction in computer-supported collaborative learning environments: Areview of the research. Computers in Human Behavior, 19, 335–353.

Morris, R., Church, H., Hadwin, A. F., Gress, C. L. Z., & Winne, P. H. (this issue). Theuse of roles, scripts, and prompts to support CSCL in gStudy. Computers inHuman Behavior.

O0Donnell, A. M., & King, A. (1999). Cognitive perspectives on peer learning. Mahwah,NJ: Lawrence Erlbaum.

Salovaara, H., & Järvelä, S. (2003). Students0 strategic actions in computer-supportedcollaborative learning. Learning Environments Research, 6, 267–284.

Slavin, R. E. (1999). Comprehensive approaches to cooperative learning. Theory intoPractice, 38, 74–79.

Winne, P. H. (1992). State-of-the-art instructional computing systems that affordinstruction and bootstrap research. In M. Jones & P. H. Winne (Eds.), Adaptivelearning environments: Foundations and frontiers (pp. 349–380). Berlin,Germany: Springer-Verlag.

Winne, P. H., Gupta, L., & Nesbit, J. C. (1994). Exploring individual differences instudying strategies using graph theoretic statistics. The Alberta Journal ofEducational Research, XL, 177–193.

Winne, P. H., Hadwin, A. F., Nesbit, J. C., Leacock, T., Kumar, V., & Beaudoin, L. (2006).gStudy: A toolkit for developing computer-supported tutorials and researchinglearning strategies and instruction (Version 3.1). Simon Fraser University,Burnaby, British Columbia, Canada.

Winne, P. H., Nesbit, J. C., Kumar, V., Hadwin, A. F., Lajoie, S. P., Azevedo, R. A., et al.(2006). Supporting self-regulated learning with gStudy software: The learningkit project. Technology, Instruction, Cognition and Learning, 3, 105–113.

roject: Software tools for supporting and researching ..., Computers