Quiz

Ok, in the tracker issue versions 2.7 and 2.8 are mentioned, that's why I thought this has changed or been improved as far as the UI goes.

I have a test category of three questions. I'm showing two of them in a Quiz at random. Now when I go to the Quiz statistics page, I'm shown discrimination index et cetera for "question #1" and for "question #2". Both links go to a similar drill down for each question. But neither of them has any indication as to what question out of the three in the category it is. How is the discrimination calculated, as it needs to be question specific, not "slot specific"?

We recently updated to 3.4, and something does appear to have broken for our quiz statistics as a result.

Specifically, most of our exams are designed to pull random questions from categories. In both the online statistics report view, and the downloaded files however, statistics for the first category of random questions no longer get displayed, but instead, just the fact that random questions are pulled from that category X # of times. This is the case whether 1 question is to be randomly pulled from a category, or 10+.

Screenshots as examples:

In the below, the quiz is set to pull 1 question from a bank of 4 for Q1, and 1 question from a bank of 4 for Q2. For Q2, results display as expected, but for Q1, no breakdown of results is accessible.

In the below, the quiz is set to pull 13 questions from a bank of 25 for Unit 1, and 7 questions from a bank of 15 for Unit 2. Based on previous report versions, I would have expected to see the beginning of the actual statistics for Unit 1 right after the Q13 line. The report now skips over any/all data for the first category that gets questions randomly assigned, and then immediately provides data for subsequent categories after every question slot. 

Question slots 14 and 15 both pull from the same category (Unit 2). Individual item response data is the same for all 15 questions, and that data gets repeated 7 times in the downloaded report (once for each number of questions randomly pulled from that category). The only unique characteristics for the 2nd, 3rd... categories is that it calculates the statistics for each question slot individually (for instance, students answered Q14 correctly 60% of the time, and Q15 65% of the time).

The end result is that we now both lack any real useful data on whatever questions happen to be in the first randomly assigned category (the 4 questions from example 1 above, and the 25 questions from example 2), and in downloaded reports, have useful, but quite a bit of redundant data for all remaining randomly pulled from categories.

On quizzes where Q1 is not randomly assigned, the statistics report does pull data for that question accurately.