Development of a Pattern Analysis Technique for Use in the Selection of Predictors.
Abstract
The study describes a computerized item-selection program called PAIN that uses a pattern-analysis approach to select a most-valid subset of items from a set. The results of this study indicate that PAIN is capable of selecting a small subset of items which, when scored by pattern analysis, has greater validity than the original set. It appears that, as well as reducing the sizes of standard tests without losing predictive value, PAIN may also be of value in selecting biographical items of information for use as predictors. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Jun 01, 1973
- Accession Number
- AD0764484
Entities
People
- Burton Fredrick Folce Jr
Organizations
- Naval Postgraduate School