I have been asked to look at some item analysis statistics from Canvas. The discrimination numbers make no sense to me. For example, the class is composed of 74 students, and 99% of the class answer the question correctly, but the DI is given as 0.65.  In another question (same class), 99% of the students answer correctly but now the DI is given as 0.1.  Can anyone explain to me where these numbers are coming from?  I know how to calculate both point biserial and DI manually, but have not done either yet.

Thanks!