SOME CONSEQUENCES OF USING RAW‐SCORE REPORTS OF VOCATIONAL INTERESTS

Abstract

The use of norms in deriving meaning from assessments of human traits (as distinguished from proficiency or achievement) is a well-established practice. According to psychometric theory, raw scores merely tally observations for a particular sample of behavior. The results of this tally cannot be taken at face value, and comparisons of raw scores across scales are not justified if inferences concerning relative strength of psychological traits or characteristics are to be drawn (see, e.g., Angoff, 1971; Thorndike & Hagen, 1969). In past, norms have been used to give raw scores psychological meaning. However, raw-score reports of psychological constructs (e.g., personality traits, career interests) have become increasingly popular. For example, counselee's profile for Career Maturity Inventory displays raw scores for each of six scales, and counselors are encouraged to attribute meaning to these scores. According to manual (Crites, 1973), the individual's degree of career development can be appraised by raw scores obtained on [six] parts (p. 36). The Ohio Vocational Interest Survey reports raw scores, in conjunction with normed scores, for each of 24 interest scales. In 1973-74, American College Testing Program reported raw scores in conjunction with normed scores to 900,000 college-bound students who took interest inventory included in ACT. Perhaps most vocal advocates of use of raw scores in assessing and comparing strength of interest and personality traits have been John Holland and his co-workers (see, e.g., Gottfredson & Holland, 1975; Gottfredson, Holland, & Gottfredson, 1975; Holland, Gottfredson, & Gottfredson, 1975). Holland uses raw scores from his instruments like a physicist uses a ruler (Holland, 1976, p. 353). In his Self-Directed Search (SDS), Holland (1972) ranks and weights raw scores on each of three sets of six scales in order to obtain high-point codes used to describe a person's personality pattern and to identify occupations (environments) for exploration (Holland, 1973). Holland does not discuss psychometric basis for this use of raw scores either in latest statement of his theory of careers (Holland, 1973) or in SDS manual (Holland, 1972). However, he and his co-workers maintain that when interest score reports are based on norms, rather than raw scores, they will have lowered validity (Gottfredson et al., 1975; Gottfredson & Holland, 1975; Holland, 1975). Widely used interest inventories (e.g., Kuder Preference Record, Strong-Campbell Interest Inventory) and personality inventories (e.g., MMPI, Edwards Personal Preference Schedule, California Psychological Inventory, Personality Research Form) base reports of psychological traits on norms, typically same-sex norms. Hence, recent trend toward use of raw score reports of psychological traits represents a major departure from previous common practice. From a practical standpoint, trend is desirable. The use of raw score reports avoids enormous expenses often involved in selecting and testing a representative norm group. Self-scoring is facilitated, thus reducing need for machine scoring while making immediate feedback of results possible. Finally, there are no complicated statistics (e.g., percentile ranks, standard