This article by Kristen DiCerbo, Assessment Data Scientist at GlassLab, is cross-posted here from Pearson’s Research & Innovation Network blog.
I have previously written about my involvement in GlassLab, the Games Learning and Assessment Lab partnership between EA, Institute of Play, ETS, and the Pearson Center for Research on Digital Data, Analytics, and Adaptive Learning.
We are making good progress on game design a SimCity-based scenario. Recently, some of the GlassLab staff and a developer of the commercial SimCity game recorded an update of the latest developments.
Some of the highlights:
At 11:30, MJ (the lead GlassLab designer) demos a bit of the current game.
At 20:52, there is a nice demo of a couple of tools for teachers that are in development. One in particular allows teachers to monitor in-progress game play on a mobile device.
There is also a question about player testing. We have just wrapped up another round of testing of an early version of the game and I am digging into the log file data and playing with visualizations of activity.
One area of interest lies in identifying indicators of skill in the log files, for example: given that we introduced players to a problem, how long do they spend gathering information about the city and problem before they take some kind of action? Below is a plot of players’ (each player is a different row) actions over time (the bottom axis is seconds). Each symbol is a different activity, but the key is that information-gathering activities (I) are coded in blue and intervention activities (A) are coded in red. So, with a quick glance, you can see how far along the time scale each player was at the time of their first blue activity.
(Click for larger view.)
Now for the big question: does this amount of time spent in understanding the city prior to intervention lead to better choice of intervention? better outcomes? better understanding of the system? We know that past research on novice and expert behavior in different domains has sometimes suggested that experts spend longer on this diagnosis/problem identification stage. Is that true here? Stay tuned…
News | June 23rd, 2013