The basic objective of the study was to determine the validity of four new indices of item quality. Three of these were based on analyses of differential, empirical weights for item choices, and the fourth was designed to measure the relative attractiveness of distracters. A secondary objective was to ascertain the validity of the conventional discrimination indices. To attain these objectives, multiple-choice items designed to vary in quality with respect to nine common item-writing principles were prepared. The quality of each item was rated independently by three judges, and the average of their ratings was used as the criterion to determine the validity of the indices. The special test items were administered to a sample of college undergraduates, and the five indices were computed on the basis of their responses. The data were analyzed, and the conventional discrimination index was found to be a moderately valid measure of item quality. The weighted combination of the new indices also appeared to be valid. Because all of the new indices did not operate in the way expected, however, it is suggested that further research on them is necessary before they are considered for practical use in test-construction projects. (Author)