Interaction Analysis of a Biology Chat

Abstract

Abstract. This is an analysis of data from a first attempt to combine (a) VMT technology, (b) helping agents, (c) collaborative small groups, and (d) accountable-talk prompting in order to scaffold biology student online chats about videotaped results of a biology experiment. Analysis of the response structure of the chat log of one of the student groups reveals characteristics of their interactions in terms of building collaborative knowledge. In particular, the mediation by the VMT technology, helping agents and accountable-talk training is analyzed to determine their influences in promoting productive learning-oriented interaction. A design-based research analytic perspective provides suggestions for redesign of the socio-technical approach based on the findings of the interaction analysis.

Towards Careful Practices for Automated Linguistic Analysis of Group Learning

“…There were two qualitative analyses, a network analytic approach, and our own SouFLé analysis. In this set of analyses, both one of the qualitative analyses (Stahl, 2013) and the network analytic approach (Goggins & Dyke, 2013) adopted a network-like representation. Another qualitative analysis (Cress & Kimmerle, 2013) took a purely descriptive approach.…”

“…Another qualitative analysis (Cress & Kimmerle, 2013) took a purely descriptive approach. The issue of social positioning was the focus of the SouFLé analysis (Howley, Mayfield, Rosé, & Strijbos, 2013) as well as one qualitative analysis (Stahl, 2013), and the network approach (Goggins & Dyke, 2013;Stahl, 2013). The main contrast was in terms of the focus of inquiry.…”

Computer-Supported Academically Productive Discourse

“…In experiments investigating the use of software agents in the VMT environment to scaffold and guide group cognition, we have seen how problematic accountable-talk agents can be (Stahl, 2013a). Agents were sometimes distracting, confusing, disruptive.…”

Web-based conversational agents for collaborative learning support

“…The above view appears to be consistent with Michaels et al (2008) work indicating that the effectiveness of the various APT facilitation strategies may depend on various elements such as the authority of the teacher or the background of the students. Additionally, Stahl (2013a) underlines the need to reduce the verbosity level of the APT interventions in order to minimize the interruption effect emerging from the interference of the conversational agent.…”

“…Next, we calculated the 'explicit response ratio' (ERR) metric introduced in our previous study in order to measure how many explicit contributions were triggered on average by each agent intervention. After examining the content of all agent-induced contributions, we expanded our analysis to investigate any patterns arising from the sequentiality of students' contributions (Stahl, 2013a).…”