WeBWorK Main Forum

statistics for "get a new version" and attempts

statistics for "get a new version" and attempts

by Joel Trussell -
Number of replies: 0
I have several statistical analyses that I'd like to do on the ECE Webwork classes. I have a program that reads the answers.log file and extracts the correct indicator for each problem and counts the number of attempts for each. My goal is to correlate the performance of the students in Webwork (accuracy of answers and number of attempts) with their test performance.

problem 1 - I'd like to determine if students actually use the "get a new version" button to review for tests. If the student works on a "closed" problem set, the time is logged and the problem set and problem number are recorded in the log. But the answers, if s/he entered any, are not recorded and the answer indicators are all zeros. Thus I can't tell what the student did. Did he try the problem, check a solution, hit the "get a new version" button multiple times to see the range of parameters or range of solutions.

One instructor creates "practice problems" that are similar to the actual problems, but are closed so the students can see the answers. They might do this but they might use the practice for other reasons. it would be nice to be able to determine what the student has done. This would allow us to determine the effectiveness of giving such "practice problems"

There was a comment on the forum back in 2014 or so that it is possible to allow students to see the answers before the due date. This would allow the student answers to the "practice problems" to be recorded. I tried resetting the due date and problem answer date in this mode, but Webwork would not allow it - even though I have changed the permissions on view answers to "student" from the default "ta".

Problem 2 - I can check the accuracy of the answers recorded in the answer.log by the binary string in the file. However, counting the number of attempts is done by counting the responses to a given problem in a given problem set. Since it is possible, but not likely - from experience - that the student has used the problem for practice after the close date, my count for number of attempts for solving the actual problem and using the practice mode can be confused. is there a way to distinguish practice attempts from actual attempts?

I note the time of the action is recorded in the answer.log but the closing time is not recorded there.