Information Technology Reference
In-Depth Information
VMT environment. The reproduced chat room is separated by a thin line at the bot-
tom from a VCR-like interface for replaying the session (see Fig. 2.2). The session
can be replayed in real time or at any integral multiple of this speed. It can be started
and stopped at any point. An analyst can drag the pointer along the timeline to scroll
both the whiteboard history and the chat history in coordination. One can also step
through the recorded actions, including all the awareness messages. In addition,
spreadsheet logs can be automatically generated in various useful formats.
The data analyzed in the VMT Project is recorded with complete objectivity.
There is no selectivity involved in the data generation, recording or collecting
process. Furthermore, the complete recording can be made available to other
researchers as a basis for their reviews of our analyses or the conducting of their
own analyses. For instance, there have been multiple published analyses of the VMT
data by other research groups following somewhat different research questions, the-
ories and methods (Koschmann & Stahl, 2009; Stahl, 2009). While collaborative
sessions are each unique and in principle impossible to reproduce, it is quite possi-
ble to reproduce the unfolding of a given session from the persistent, comprehensive
and replayable record.
Collaborative Data Sessions (Reliability)
Interpretation of data in the VMT Project first begins with an attempt to describe
what is happening in a chat session. We usually start this process with a data session
(Jordan & Henderson, 1995) involving six to twelve researchers. A typical data
session is initiated by a researcher who is interested in having a particular segment
of a session log discussed by the group. Generally, the segment seems to be both
confusing and interesting in terms of a particular research question.
For our data sessions, we sit around a circle of tables and project an image of the
VMT Replayer onto a screen visible to everyone. Most of us have laptop computers
displaying the same Replayer, so that we can scan back and forth in the segment pri-
vately to explore details of the interaction that we may want to bring to the attention
of the group. The group might start by playing the segment once or twice in real
time to get a feel for how it unfolds. Then we typically go back to the beginning and
discuss each line of the chat sequentially in some detail.
The interpretation of a given chat line becomes a deeply collaborative process.
Generally, one person will make a first stab at proposing a hypothesis about the inter-
actional work that line is doing in the logged discourse. Others will respond with
suggested refinements or alternatives to the proposal. The group may then engage
in exploration of the timing of chat posts, references back to previous postings or
events, etc. Eventually the data analysis will move on to consider how the student
group took up the posting. An interesting interpretation may require the analysts
to return to earlier ground and revise their tentative previous understandings (Stahl,
2009, chap. 10).
The boundaries of a segment must be considered as an important part of the anal-
ysis. When does the interaction of interest really get started and when is it resolved?
