PERSONALITY PROCESSES AND INDIVIDUAL DIFFERENCES Does the Defining Issues Test Measure Psychological Phenomena Distinct From Verbal Ability?: An Examination of Lykken's Query Cheryl E. Sanders, David Lubinski, and Camilla Persson Benbow Iowa State University This study examined the incremental validity of the Denning Issues Test (DIT), a test purporting to measure moral reasoning ability relative to verbal ability and other major markers of the construct of general intelligence (g). Across 2 independent studies of intellectually precocious adolescents (top 0.5%), results obtained with the DIT revealed that gifted individuals earned significantly higher moral reasoning scores than did their average-ability peers; they also scored higher than college freshmen, who were 4 to 5 years older. The relative standing of the intellectually gifted adolescents on moral reasoning, however, appears to be due to their superior level of verbal ability as opposed to any of a number of the other psychological variables examined here. The hypothesis that the DIT is conceptually distinct from con- ventional measures of verbal ability was not confirmed. Investigators conducting subsequent studies involving the assessment of moral reasoning are advised to incorporate measures of verbal ability into their designs, thereby enabling them to ascertain whether moral reasoning measures are indeed capturing systematic sources of individual differences distinct from verbal ability. This idea also is relevant to other concepts and measures purporting to assess optimal forms of human functioning more generally (e.g., creativity, ego development, and self-actualization). In the social sciences, measures do not always assess what they purport to measure, and the causal determinants of our most favorite constructs and outcomes do not always fit prior expectations. All too frequently in social science research, the- oretically appealing constructs are assessed and studied without attending to competing variables that might be causally linked to their status as well as the criteria they predict. Socioeconomic status (SES), for example, is a variable that social scientists fre- quently assume to be causally related to a host of psychologi- cally important phenomena (see the writings of Humphreys, 1991, and Meehl, 1970,1971 a, 1971 b, on this topic). In school and work settings, for example, SES is presumed to exert a pro- found causal role in determining the outcomes on conventional achievement criteria (cf. Humphreys, 1991). More powerful variables exist, however, for predicting academic and vocational criteria (e.g., human abilities; Humphreys, 1992; Lubinski & Dawis, 1992; Schmidt, 1994). Yet, they are seldom studied con- comitantly with SES (Humphreys, 1991). Indeed, the practice Cheryl E. Sanders, David Lubinski, and Camilla Persson Benbow, De- partment of Psychology, Iowa State University. This article was based on a dissertation submitted by Cheryl E. Sand- ers to Iowa State University in partial fulfillment of the Doctor of Phi- losophy degree. An earlier version of this article profited from com- ments and suggestions from Lloyd G. Humphreys, Paul E. Meehl, and Julian C. Stanley. Correspondence concerning this article should be addressed to David Lubinski, Department of Psychology, Wl 12 Lagomarcino Hall, Iowa State University, Ames, Iowa 50011 -3180. of inferring causal links between SES—the "master variable" of sociological inquiry (Gordon, 1987)—and its many correlates, without evaluating competing correlated factors (such as ability), has been referred to as the sociologist'sfallacy (Jensen, 1973). This common methodological shortcoming is an exam- ple of underdetermined causal modeling—and the social sci- ences are dotted with several others. One of the more striking examples of this in contemporary psychological research is the tendency for researchers to evalu- ate the importance of students' self-efficacy or its manipulation for choosing to embark on conceptually demanding educational or vocational paths, without simultaneously assessing relevant ability requirements necessary for performing competently in the targeted disciplines (for a number of examples, see Betz & Fitzgerald, 1993). Another example can be found in research on parent perceptions of their sons' and daughters' strengths and weaknesses. Gender differentiating expectations are fre- quently interpreted as a function of sex role stereotyping (e.g., Jacobs & Eccles, 1992); although this may be so, more defini- tive conclusions would be obtained [{objective measures of the rated skill domain under analysis were concomitantly assessed (say, for example, ratees' mathematical reasoning ability), on which parents' perceptions and ratings are at least partly based. We would then be in a position to determine whether genuine gender differentiating competencies were actually observed and rated with precision, or whether parents' ratings were indeed moderated by sex role stereotypes and, as such, are systemati- cally biased. Erroneous suppositions regarding presumed causal paths are Journal of Personality and Social Psychology, 1995, Vol. 69, No. 3,498-504 Copyright 1995 by the American Psychological Association, Inc. 0022-3514/95/53.00 498 This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.