Optimal Averaging in Performance Tests.
Abstract
The purpose of this research was to develop a methodology for optimizing the temporal stability and predictive validity of performance tests and to apply that methodology to the Project-A, computer-administered tests. In the present research, a performance test is treated as a task to be practiced, and tests are analyzed as individual differences in skill acquisitions and retention. Classical test theory is also used. The predictive validity of the Project-A, computer-administered tests for a simulated anti-aircraft criterion task was studied over a 4-month interval in a sample of 102 college students; the 4-month temporal stability of the tests was studied concurrently in the same sample. Three of the 10 Project-A tests (Choice Reaction, Target Tracking 2, and Cannon Shoot) show a forward stability optimum. Cannon Shoot also has high predictive validity (.59). It could have the highest predictive validity of any test in the Project-A battery if its temporal stability could be improved. In none of these tests, however, can temporal stability be improved by lengthening the tests. (AN)
Document Details
- Document Type
- Technical Report
- Publication Date
- Jan 01, 1995
- Accession Number
- ADA298836
Entities
People
- Marshall B. Jones
Organizations
- Penn State Milton S. Hershey Medical Center