COMPARISON OF TEST ITEMS ACROSS FORMS,
Abstract
Repetitive use of a limited number of the same items in subtests of successive forms of the Airman Classification Test permitted comparison of statistics for these items over time. The difficulty level of each item, measured by percentage of examinees marking the right answer, varied with the general ability of the sample as measured by AFQT score. The precise effects of practice from previous testing on similar kinds of material is unknown, but the results suggest that use of anchor items can yield gross estimates of differences in quality of tested groups. Such estimates can best be made from the Arithmetic Reasoning and Work Knowledge subtests, since spatial-visualization and information-type items tended to be more stable across samples. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Feb 01, 1964
- Accession Number
- AD0437953
Entities
People
- William B. Lecznar