...
Purpose
This report compares the highest and lowest performance on each question. It selects the upper 27% and lower 27% of students who have selected each option in a question. (The percentage can be overridden in the cohort dropdown on the previous screen). A question's difficulty p is calculated as (average score)/(maximum achievable score). It ranges from 0.0 (nobody got it right) to 1.0 (everybody got it right). Values of p<0.25 or p>0.75 indicated questions with limited differentiating power.
...