Evaluation of Optimal Appropriateness Measurement for Use in Practical Settings
Final technical paper Oct 1987-Jun 1989
ILLINOIS UNIV AT URBANA MODEL BASED MEASUREMENT LAB
Pagination or Media Count:
Optimal appropriateness indices were evaluated for use in applied settings. The first study demonstrated the adaption of existing computer software to the problem of identifying cheaters with scores in specific test score ranges. It is shown that the Levine and Drasgow 1988 algorithm can be applied directly to this problem. Then simulated and real data were used to determine the rates of detection of cheating on 5, 10, or 15 items on each of two tests and obtaining a total test score in one of two total test score ranges. As expected, low rates of identification were obtained for cheating on only 5 items per test and reasonably high rates were obtained for cheating on 15 items. Moderately lower identification rates were obtained with real data than with simulation data. The extent of difficulties that are likely to occur when the assumptions of optimal indices are violated were evaluated in the four robustness experiments that constituted Study Two. Keywords Armed services vocational aptitude battery, Appropriateness measurement, Aptitude test.
- Computer Programming and Software