Some Problems with the Criterion-Keying Approach to Occupational Interest Scale Development

Abstract

A traditionally used method of keying interest inventories, referred to as criterion-keying, involves selecting items which best distinguish a group of incumbents in a particular occupation from another group intended to represent the population of interest (men-in-general). Some theoretical results are presented which show that in occupations where explicit selection is operating items may be selected for a key merely because they correlate with the explicit selectors. Data are presented for a military sample which suggest that criterion-keyed scales tend to replicate the assignment of dissatisfied incumbents into the same fields. The results are discussed in terms of their implications for the usefulness of criterion-keyed scales and further research.