We analyzed multiple-choice questions provided in test banks for introducto
ry psychology textbooks. In Study 1, we found that about 70% of students re
sponded correctly to a given item, and there was a significant but inconsis
tent difference in difficulty across chapter topics. In Study 2, a single c
lass of introductory psychology students took comprehensive exams of random
ly selected test items that varied in estimated difficulty. As before, chap
ter topic had a significant, yet inconsistent effect, but test items had se
veral desirable properties compared to Study 1. These data allow several re
commendations for instructors who want to improve the characteristics of th
e multiple-choice exams they use.