The meaning of personality test scores.

The process of reexamining the methodological and metatheoretical assumptions of personality psychology over the past two decades has been useful for both critics and practitioners of personality research. Although the field has progressed substantially, some critics continue to raise 1960s-vintage complaints, and some researchers perpetuate earlier abuses. We believe that a single issue--construct validity--underlies the perceived and actual shortcomings of current assessment-based personality research. Unfortunately, many psychologists seem unaware of the extensive literature on construct validity, This article reviews five major contributions to our understanding of construct validity and discusses their importance for evaluating new personality measures. This review is intended as a guide for practitioners as well as an answer to questions raised by critics. Because the problem of construct validity is generic to our discipline, these issues are significant not only for personality researchers but also for psychologists in other domains. Walter Mischel's 1968 book stimulated 20 years of careful examination of the methodological and metatheoretical assumptions of personality psychology. This reappraisal has been useful for both critics and practitioners of measurement-based personality research. Within the personality research establishment, things seem to have returned to normal; people are comfortable once again with the notions that (a) personality assessment is an appropriate methodology in many areas of personality, clinical, and industrial psychology, and (b) there is stability to personality descriptors over time and occasions. Nonetheless, plus (a change, plus c'est la m~me chose: On one hand, 1960s-vintage criticisms continue to come up (i.e., test scores are contaminated by social desirability; validity coefficients are modest; there is no stable core to personality). On the other hand, some personality researchers continue to invite reproof by perpetuating abuses about which critics legitimately complain. Landy (1986) documented the confusion that still surrounds the testing enterprise despite 20 years of soul searching. We believe that a single issue underlies the perceived and actual shortcomings of current assessmentbased personality research. When critics complain and when personality researchers fail, construct validity is usually at the heart of the matter. A review of the notion of construct validity may serve as a guide for practitioners as well as an answer to many of the questions that seem to bother critics of measurement-based personality research (cf. Kagan, 1988). Drawing on Landy's essay, in this article we argue that (a) all validity is construct validity; (b) the process of test validation is hypothesis testing; and (c) measurement-based research is formally identical with any other type of legitimate scientific in-