Organizers

Cynthia D’Angelo, SRI International
Cindy Hmelo-Silver, Indiana University
Marcela Borge, Pennsylvania State University
Alyssa Wise, New York University
Bodong Chen, University of Minnesota.

Do you have questions about how collaborative learning data can be used as input to instructional decision making or adaptive support? About how models of collaboration can be used to decide what data is important and technically tractable? The goal of this half-day workshop is to provide opportunities for interactive knowledge sharing and discussions on these issues. Participants will be asked to code two short datasets prior to the workshop in preparation using either their own coding approach or select from a collection that will be provided by the conference organizers. These will then be used as a context to discuss issues central to the workshop topic. We will begin with several tables that include people interested in sharing knowledge or learning about a particular topic.

A preliminary set of topics for discussion includes:
1. Visualization techniques for CSCL data
2. Coding schemes for different levels of analysis
3. Text-mining approaches
4. Opportunities and challenges of temporal analysis of CSCL
5. Using data to provide instructional support

Two rounds of discussions will provide opportunities for both sharing and https://cpeas-cscl.github.io to indicate learning within and across groups. We will then provide opportunities for whole group discussion. If you are interested, please fill out the application form at https://cpeas-cscl.github.io to indicate your interest, and which topics you are able to share knowledge about and what topics you would be interested in learning more about (as well as to suggest additional topics). The intended audience for this half-day workshop on June 19 is an interdisciplinary group of researchers, that analyze social interactions in a variety of contexts. Our aim is to examine trade-offs associated with
analysis approaches at multiple levels of scale, from individual, to small group, to massive online settings. We encourage participation from those with technical interests in manual coding of communication, interaction analysis, learning analytics, intelligent systems, and adaptive support broadly conceived.

Description of the event

To help make progress in understanding and designing for CSCL, it is important to have common data sets and tools to serve as a boundary object for discussion. We will organize tables for interactive work sessions based on prominent adaptive support themes, with concrete examples of collaborative learning based on pre-workshop preparation and coordination. Some candidate themes might include:

1. Visualization techniques for CSCL data
2. Coding schemes for different levels of analysis
3. Text-mining approaches
4. Opportunities and challenges of temporal analysis of CSCL
5. Using data to provide instructional support

The workshop will start with two 45 minute rounds of mini-tutorial knowledge exchanges surrounding each theme, with both experts and novices for each group so that all participants have opportunities to be teachers and learners. Each group will have a member of organizing group to facilitate the exchanges and raise key issues. Following these topic specialty discussions, we will create jigsaw groups that will allow cross-talk grounded in the context of datasets that have been pre-coded from multiple perspectives. The final hour will leave time for reports from the jigsaw groups and whole group discussion.

Expected outcomes and contributions

Expected outcomes of the workshop include compiling shared data sets and coding constructs for participants to think together with, networking and knowledge exchange both within and across expertise groups (e.g., qualitative discourse coding and automatic text analysis), improved strategies for and research questions about implementing adaptive support in CSCL, and an improved understanding of the challenges of doing this kind of work.

The data sets have been chosen to represent different aspects of CSCL. Data Set 1 includes middle school students working face-to-face in groups of three, with a shared computer screen that shows their short problems to solve. While there is audio associated with this data set, a transcribed version will be available for the researchers to code. Data Set 2 includes text and clickstream data from the online asynchronous discussions of undergraduate students in groups of three to six solving authentic real-world business challenges in preparation for a joint in-class presentation. Data Set 3 includes synchronous chat-based text of an online discussion about course content for 13 teams, in five sessions across a ten week period. The dataset also includes individual reflections on the quality of chat session and team sensemaking about discussion quality at each time point.

Each of the data sets will be coded with multiple coding schemes. Some will be coding schemes that were initially developed to be used with one of those particular data sets, while others are meant to be used more broadly. This will allow the group to discuss the relative strengths and weaknesses of applying each coding scheme to each data set and explore the ways in which the coding schemes would lead to different analyses and/or findings with respect to the data. This will also allow for the exploration of different ways to represent both the data and the annotations/coding results and any analysis. Participation requirements: Interest in CSCL, models of collaboration and adaptive support; willingness to engage in pre-workshop preparation activities that will involve coding two 2-3 page datasets (using one of the available coding schemes); looking to share expertise as well as learn about different aspects of analysis and support. Please fill out an application form linked on our workshop website: https://cpeas-cscl.github.io .

Important dates:

Applications due by April 28
Notification of participation: CSCL early registration deadline
Workshop: June 19 (half day)