CORDTRA (Chronologically-Ordered Representation of Discourse and Tool-Related Activity) is a visualization technique proposed by Cindy Hmelo-Silver (2003) for interpreting collaborative learning processes, based on Luckin et al.’s CORDFU (Chronologically-Ordered Representation of Discourse and Features Used). This technique could provide a holistic view of learning processes by making temporal relationship among various types of learning activities visible. By presenting a few interesting case studies, Hmelo-Silver and colleagues (2003, 2008) demonstrate that CORDTRA could provide useful insights about student interactions in computer-supported collaborative learning (CSCL).
Detailed description of CORDTRA could be found in Hmelo-Silver and her colleagues’ articles. The general idea of this visualization is to first sort discourse units chronologically and put them on an X axis, and then map interested coding results (or log information) of discourse onto the axis, with each coding category represented as a horizontal line. In this way, researchers could inspect the representation visually and look for possible relations among discourse events pertinent to their research questions. Here is an example of a CORDTRA diagram:
CORDTRA Shiny App
Working on data analysis of my doctoral research, which involves various types of student activities in Knowledge Forum, I decided to give this technique a try. According to Cindy, they used Excel to create CORDTRA diagrams. They could zoom into a timeline to study events around a certain period of time in detail. This is pretty handy! But at the same time I heard complains about the complexity of the task of creating a CORDTRA diagram from colleagues who incorporated it in their research. I could imagine that customizing shapes of dots in an Excel diagram is not always easy.
Inspired by a blog post about creating CORDTRA diagrams with D3.js, I thought it might be a good idea to develop a simple R-based program that could effortlessly convert coded CSCL data into a CORDTRA diagram. So here we are, a Shiny app (based on R) for creating CORDTRA diagrams.
What this app asks for is only a spreadsheet containing sorted discourse units and their coding results. The first few lines of the spreadsheet I use for demo look like the following:
The first column of the table contains indices of discourse units. (In my case, discourse units are Knowledge Forum notes.) The other four columns contain coding results of four coding schemes, including Ways of contributing, Levels of problems, Scientific sophistication, and Epistemic complexity. They are categorical or ordinal data, with each category or level of them representing a certain interested event in my research.
With the Shiny app I’ve developed, this table can be automatically transformed to a CORDTRA diagram like below. In this diagram, each coding scheme, which maps to each table column, has a distinctive color, and each category under a coding scheme is represented by a specific shape.
This app also allows researchers to zoom in/out the diagram by selecting range of discourse units to be visualized. It also enables researchers to filter certain coding categories by deselecting them on the left panel.
There are many way to interpret a CORDTRA visualization. Beside visual interpretation, one analysis researchers may wish to do is to check co-occurrence of certain types of events. To facilitate this analysis, this Shiny app extracts a co-occurrence matrix of all coding categories, and visualize it as a force-directed map (see below). The following visualization may not be the best example to show the power of this functionality, because the only fruitfulness of it is showing that we’ve coded scientific sophistication and complexity for theorizing notes and depth of questions for questioning notes. But if you have a richer set of coding schemes, interesting patterns which are hard to inspect in CORDTRA may become visible in the visualization of co-occurrence.
Another direction I wish to take is to make this diagram more interactive. Thanks to the googleVis R package, it becomes possible to integrate Google Charts into a Shiny app. I started by playing with the Google Motion Chart—a cool visualization that could show change of objects overtime. Probably the most famous example of this visualization is Hans Rosling’s TED talk about poverty and world economy. In this app, because I don’t have enough numeric data to feed to x and y axises, it could only track accumulated occurrence of each coding category. If student/author information of each discourse unit is available, it might become possible to create aggregated information around students to visualize and compare certain aspect of discourse performance of all students.
- Hmelo-Silver, C. E. (2003). Analyzing collaborative knowledge construction. Computers & Education, 41(4), 397–420. doi:10.1016/j.compedu.2003.07.001
- 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
- Luckin, R., Plowman, L., Laurillard, D., Stratfold, M., Taylor, J., & Corben, S. (2001). Narrative evolution: learning from students’ talk about species variation. International Journal of AI in Education, 12, 100–123.