TopoMath in CPI360 Analyses

[jonwetzel]

• [ ] 1) Which problems are too easy? Too hard?
• [ ] Identify and count "Lost cause" students. 3 conditions: low pre-test, getting a lot of yellow without slacking, and low on post-test
• [ ] Find the "superstar" students who are getting mostly stars or green and rerun the difficulty analysis without them, looking for converging evidence.
• [ ] 2) Amount of time CPI students spent doing topomath problems. Min, max, average times. Histogram maybe?
• [ ] 3) Look at solver post-test, and label students as "topomath method user" or "non-topomath method user" Then look at which type of student (never used, could have opted out, mandatory user) was in each category
• [ ] 4) Compare post-test scores of students who finished vs students who didn't. Cross this by mandatory/not mandatory.

We can intervene on slacking, not finishing the training, and not using the method on the exam.

[jonwetzel]

One difficulty analysis is based on the yellows (already done).

Another is time/# of nodes. A fast rate could correlate with slacking.
Could we detect cases where people struggle a little and then start slacking through the rest?

If the malleable middle is spending a moderate amount of time, that's good.

[jonwetzel]

How many problems can we get done on Saturday? What's the max and minimum time to completion? Histogram.