Methods for Equating Mental Tests.
Abstract
The technology for test equating has arisen from the need to make new tests comparable to old ones. The equating of military tests has two objectives: (a) to make scores on different tests forms and on different composites of test forms comparable, and at the same time (b) to solve the norming problem by relating all scores on new tests and composites back to a large sample of talent indicative of an anticipated population of military enlistees. In this study, simulated and actual Air Force test data were used to compare the different procedures for equating mental tests and to delineate those conditions under which each equating procedure performed best. Specific testing-condition manipulations included variations in test length, item difficulty, sample size, and examinee ability distributions. Coventional (equipercentile and linear), Item Response Theory (IRT), and Strong True Score Theory (STST) methods comprised the equating procedures that were studied. The data collection designs that were used included the single-group, equivalent-groups, and anchor-test designs. Equating transformations were evaluated by comparing equated scores with true scores using bias and root-mean-squared-error indices. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Nov 01, 1984
- Accession Number
- ADA149544
Entities
People
- C. D. Vale
- K. A. Gialluca
- L. I. Crichton