read

References

Citekey: @hmelo-silver2008b

Hmelo-Silver, C. E., Chernobilsky, E., & Jordan, R. (2008). Understanding collaborative learning processes in new learning environments. Instructional Science, 36(5-6), 409–430. doi:10.1007/s11251-008-9063-8

Summary

This paper builds on the 2003 paper, further explaining how CORDTRA diagrams help us interpret learning.

What I found very useful in this paper in particular is the coding schemes they used to analyze collaborative learning. The coding scheme of “complexity” and “monitoring” could be especially useful for me to analyze use of promisingness tool by students.

Complexity (Bereiter and Scardamalia, 1987; Chernobilsky et al. 2004) Telling Elaborated telling Transforming

Monitoring (Hmelo-Silver 2004) Individual monitoring Group monitoring Self-directed learning Other monitoring

Raw notes

Page: 1

Page: 1 Author: bodong Subject: Highlight Date: 22/05/2013, 2:17:05 PM Frequency analyses provided an easily interpreted snapshot of each group’s activity. The CORDTRA analyses provide a more dynamic view that helps researchers and teachers better under- stand how collaborative learning unfolds.

Page: 1 Author: bodong Subject: Highlight Date: 22/05/2013, 2:16:54 PM The use of Chronologically-Ordered Representations of Discourse and Tool-Related Activity (CORDTRA) diagrams is one way of achieving this under- standing.

Page: 2

Page: 2 Author: bodong Subject: Highlight Date: 22/05/2013, 2:19:19 PM We frame our work in sociocultural and social constructivist theories which (1) argue that learning is inherently social and situated (Palincsar 1998) and (2) emphasize the critical role of tools in mediating learning (Cole 1996).

Page: 2 Author: bodong Subject: Highlight Date: 22/05/2013, 2:21:35 PM We view the learning environment as an activity system (Engestro¨m 1999). In under- standing this system, we need to investigate the cultural tools that are used—the language used in discourse and the computer technology that embodies a particular pedagogical model.

Page: 2 Author: bodong Subject: Highlight Date: 22/05/2013, 2:21:08 PM This requires going beyond coding and counting of individual speech acts to developing a holistic means by which we can view learning events as they unfold.

Page: 2 Author: bodong Subject: Highlight Date: 22/05/2013, 2:17:45 PM the general rubric of constructivist learning environments, which emphasize collaborative engagement in authentic problems (Gijbels et al. 2006).

Page: 3

Page: 3 Author: bodong Subject: Highlight Date: 22/05/2013, 2:24:07 PM We argue that CORDTRA (Chronologically-Ordered Representation of Discourse and Tool-Related Activity) can foster holistic visualization of the data while enabling the fine grain coding, which proved valuable in some of the studies illustrated above.

Page: 3 Author: bodong Subject: Highlight Date: 22/05/2013, 2:25:56 PM Other approaches to analyzing CSCL data include experimental methods (e.g., Rummel and Spada 2005; Su- thers and Hundhausen 2003), multilevel analysis (Cress 2008; Janssen et al. 2007) and social network analysis (e.g., Nurmela et al. 1999). Finally, others have used interaction analysis (Jordan and Henderson 1995; Hmelo-Silver et al. 2007b; Lantz-Andersson et al. in press) and conversation analysis (e.g., Stahl 2006a, b).

Page: 3 Author: bodong Subject: Highlight Date: 22/05/2013, 2:27:11 PM As Larkin and Simon (1987) noted, a diagram is often much easier to interpret than verbal presentations of the same material and can make it easier to interpret complex patterns.

Page: 3 Author: bodong Subject: Highlight Date: 22/05/2013, 2:25:47 PM Analyzing the complexity of collaborations is often conducted using intricate coding schemes that quantify different types of utterances

Page: 3 Author: bodong Subject: Highlight Date: 22/05/2013, 2:23:45 PM Some researchers have integrated time into their analyses in a variety of ways. De Laat and Lally (2003) examined how discourse changed over time in an online discussion using a combination of critical event recall and content analysis over different discrete phases of an online discussion. Zemel et al. (2007), developed an emergent coding scheme based on critical events as they unfolded during sequential online chats. Using multivariate statistics and chronologically oriented sequences to visualize the data, these individuals were able to draw conclusions about group interactions. Schu¨mmer et al. (2005) combined discourse with log data. In their paper, however, they addressed the difficulty of representing these data for analysis. Luckin et al. (2001) and Luckin (2003) presented a means to visualize their data through CORDFU (Chronologically-Ordered Representation of Discourse and Features Used).

Page: 4

Page: 4 Author: bodong Subject: Highlight Date: 22/05/2013, 2:31:07 PM Other researchers have developed analytic tools that focus on how a concrete artifact is constructed. In particular, Avouris et al. (2003) developed the OCAF (Object-oriented Collaboration Analysis Framework) to work backward from a collaboratively created diagrammatic artifact and study the history that led to the construction of that artifact.

Page: 4 Author: bodong Subject: Highlight Date: 22/05/2013, 2:38:35 PM CORDTRA diagrams

Page: 4 Author: bodong Subject: Highlight Date: 22/05/2013, 2:30:49 PM Strom et al. (2001) used several directed graphs to map the semantic space of the instructional discourse as students coordinated conceptual and procedural knowledge. In a small group CSCL environment, Suthers (2006) developed uptake graphs to examine how ideas flowed during collaborative knowledge construction as pairs of students engaged in online discussions.

Page: 5

Page: 5 Author: bodong Subject: Sticky Note Date: 22/05/2013, 2:39:33 PM is there a tool available to create such diagrams?

Page: 5 Author: bodong Subject: Highlight Date: 22/05/2013, 2:40:16 PM For this analysis, we compare and contrast two groups: one that collaborated extremely well and another that was less successful in their collaboration.

Page: 5 Author: bodong Subject: Highlight Date: 22/05/2013, 2:40:29 PM based on students’ self reports of what they learned from the interaction at the end of each problem.

Page: 5 Author: bodong Subject: Highlight Date: 22/05/2013, 2:40:05 PM Method

Page: 5 Author: bodong Subject: Highlight Date: 22/05/2013, 2:39:00 PM We use one more effective and one less effective group (defined below) that collaborated in the STELLAR environment to contrast the CORDTRA analysis with a ‘‘code and count’’ content analysis.

Page: 6

Page: 6 Author: bodong Subject: Highlight Date: 22/05/2013, 2:45:04 PM Coding

Page: 6 Author: bodong Subject: Highlight Date: 22/05/2013, 2:45:41 PM A post could be coded for multiple categories but only once within each category.

Page: 6 Author: bodong Subject: Highlight Date: 22/05/2013, 2:47:26 PM The conversations which occurred during the three-week unit were coded for content, collaboration, complexity, questioning, justification, and monitoring.

Page: 6 Author: bodong Subject: Highlight Date: 22/05/2013, 2:45:13 PM When possible, discourse was coded at the unit of a post

Page: 6 Author: bodong Subject: Highlight Date: 22/05/2013, 2:41:55 PM STELLAR (Socio-Technical Environment for Learning and Learning Activity Research) is an integrated online PBL environment for preservice teachers (Derry 2006; Derry et al. 2006) that includes a learning sciences hypermedia (the Knowledge Web), a library of videocases, and an online activity structure that gives students and instructors access to a suite of individual and collaborative tools.

Page: 7

Page: 7 Author: bodong Subject: Highlight Date: 22/05/2013, 2:47:48 PM Collaboration was coded by looking at the extent to which students worked with the ideas that were circulating in the group.

Page: 8

Page: 8 Author: bodong Subject: Highlight Date: 22/05/2013, 2:52:59 PM Questions can be indicators of cognitive processing, so we examined whether students asked for factual information, detailed explanations, or were metacognitive in nature.

Page: 8 Author: bodong Subject: Highlight Date: 22/05/2013, 2:53:33 PM examine the metacognitive activity that groups and individuals engaged in.

Page: 8 Author: bodong Subject: Highlight Date: 22/05/2013, 2:53:07 PM The nature of the evidence students used to justify their ideas was coded to examine the basis for the claims they made.

Page: 8 Author: bodong Subject: Highlight Date: 22/05/2013, 2:49:02 PM To address issues of sophistication or complexity of ideas, the Bereiter and Scardamalia (1987) scale was adapted. The adapted scale included three categories: telling, elaborated telling and transforming (Chernobilsky et al. 2004).

Page: 9

Page: 9 Author: bodong Subject: Highlight Date: 22/05/2013, 2:54:44 PM Complexity

Page: 10

Page: 10 Author: bodong Subject: Highlight Date: 22/05/2013, 2:54:36 PM Monitoring

Page: 11

Page: 11 Author: bodong Subject: Highlight Date: 22/05/2013, 3:28:50 PM Frequency counts of content analysis

Page: 11 Author: bodong Subject: Highlight Date: 22/05/2013, 2:57:26 PM Results

Page: 12

Page: 12 Author: bodong Subject: Highlight Date: 22/05/2013, 3:32:26 PM CORDTRA analyses

Page: 15

Page: 15 Author: bodong Subject: Highlight Date: 22/05/2013, 3:35:59 PM In contrast, in Group 2, the facilitators joined in much later and made infrequent contributions in the form of questions. This is because in this group the students themselves asked a number of questions throughout the duration of the problem.

Page: 15 Author: bodong Subject: Highlight Date: 22/05/2013, 3:39:33 PM Group 2 engaged in group monitoring throughout the activity whereas Group 1 did not engage in this kind of monitoring until relatively late in the activity.

Page: 15 Author: bodong Subject: Highlight Date: 22/05/2013, 3:37:38 PM Another aspect of how the CORDTRA diagrams help us distinguish the collaborative activity between two groups is by showing the overall relation between the discourse and the tool use.

Page: 15 Author: bodong Subject: Highlight Date: 22/05/2013, 3:35:44 PM One point distinguishing the two groups from the outset is how the groups engaged the facilitator (a TA) and the course instructor (CHS, who also helped facilitate). In Group 1, the facilitators were involved early and fairly frequently, and asked most of the explanatory and metacognitive questions.

Page: 17

Page: 17 Author: bodong Subject: Highlight Date: 22/05/2013, 3:40:43 PM CORDTRA diagrams also allow zooming in and out on different parts of the activity.

Page: 19

Page: 19 Author: bodong Subject: Highlight Date: 22/05/2013, 3:48:01 PM In contrast, Group 2 asked many questions and used all the resources available to help with the development of their instructional plan.

Page: 19 Author: bodong Subject: Highlight Date: 22/05/2013, 3:49:12 PM Both representations have their strengths and weaknesses.

Page: 19 Author: bodong Subject: Highlight Date: 22/05/2013, 3:48:57 PM they are not trivial to construct and are not easily interpreted in a glance as the histograms used in the frequency analysis.

Page: 19 Author: bodong Subject: Highlight Date: 22/05/2013, 3:47:50 PM Group 1 tended to approach the task very sequentially—they used the video and the Knowledge Web, and then they developed their proposal.

Page: 20

Page: 20 Author: bodong Subject: Highlight Date: 22/05/2013, 3:52:18 PM This can also help both the facilitators and researchers to focus on the appropriate parts of the activity that need to be improved either while teaching or during the post-teaching analyses. Being able to actually accomplish this would require auto- mating construction of CORDTRA diagrams so that they can be available in a timely fashion.

Blog Logo

Bodong Chen


Published

Image

Crisscross Landscapes

Bodong Chen, University of Minnesota

Back to Home