Profiting from controversy. Lessons from the person-situation debate.

For the past two decades the person-situation debate has dominated personality psychology and had important repercussions in clinical, social, and organizational psychology. This controversy strikes to the heart of each of these disciplines because it puts on trial the central assumption that internal dispositions have an important influence on behavior. According to emerging views of scientific progress, controversy serves the useful function of narrowing the field of competing hypotheses. In this light, we examine seven hypotheses that arose during the course of the person-situation debate, ranging from most to least pessimistic about the existence of consensual, discriminative personality traits. The accumulated evidence fails to support the hypotheses that personality traits are simply (a) in the eye of the beholder, (b) semantic illusions, (c) artifacts of base-rate accuracy, (d) artifacts of shared stereotypes, (e) due to discussion between observers (who ignore behavior in favor of verbal self-presentation or reputation), or (f) mere by-products of situational consistencies. Evidence also fails to support the hypothesis (g) that although traits are related to behavior, the relationship is too small to be important. Yet we have not simply come full circle to a reacceptance of traits as they were understood 20 years ago. Research generated by these hypotheses has allowed us to better specify the circumstances under which personality assessments will be valid. Whether we are acting as professional psychologists, as academic psychologists, or simply as lay psychologists engaging in everyday gossip, the assumption that people have "traits" (or enduring cross-situational consistencies in their behavior) provides a basis for many of our decisions. When a clinical or counseling psychologist uses a standard assessment battery, he or she assumes that there is some degree of trait-like consistency in pathological behavior to be measured. When an organizational psychologist designs a personnel selection procedure, he or she assumes that consistent individual differences between the applicants are there to be found. When an academic psychologist teaches a course in personality, he or she must either assume some consistency in behavior or else face a bit of existential absurdity for at least three hours a week. Likewise, a good portion of our courses on clinical and developmental psychology would be unimaginable unless we assumed some cross-situational consistency. Even in everyday lay psychology, our attempts to analyze the behaviors of our friends, relatives, and co-workers are riddled with assumptions about personality traits. Despite the wide appeal of the trait assumption, personality psychologists have been entangled for some time in a debate about whether it might be based more on illusion than reality (e.g., Alker, 1972; Allport, 1966; Argyle & Little, 1972; Bem, 1972; Block, 1968, 1977; Bowers, 1973; Epstein, 1977, 1979, 1980; Fiske, 1974; Gormly & Edelberg, 1974; Hogan, DeSoto, & Solano, 1977; Hunt, 1965; Magnusson & Endler, 1977; Mischel, 1968, 1983; West, 1983). Murmurs of the current debate could be heard more than 40 years ago (Ichheisser, 1943), but the volume increased markedly after Mischel's (1968) critique, and things have not quieted down yet (Bem, 1983; Epstein, 1983; Funder, 1983; Kenrick, 1986; Mischel, 1983; Mischel & Peake, 1982, 1983). Of late, discussants have begun to express yearning to end what some see as an endless cycle of repeating the same arguments. Mischel and Peake (1982) and Bem (1983), for instance, both use the term ddjd vu in the titles of recent contributions, suggesting that they feel as if they have been here before. Other commentators maintain that the debate has been a "pseudo-controversy" (Carlson, 1984; Endler, 1973)that never should have occurred in the first place. However fatiguing it may now seem to some of its erstwhile protagonists, the debate over the alleged inconsistency of personality has been more than an exercise in sophistry. In the course of the nearly two decades since Mischers (1968) critique, a number of provocative hypotheses have been put forward, along with a host of studies to evaluate them. Platt (1964) and Popper (1959), among others, maintained that science typically progresses through the accumulation of negative informat iontha t is, by eliminating hypotheses that data suggest are no longer tenable. From this perspective, it may be worth taking a look back at the hypotheses suggested during the consistency controversy, this time in the improved light shed by two decades of research. In this light, the debate can be seen as an intellectually stimulating chapter in the history of the discipline, replete with useful lessons for professionals who include assessment in their repertoire. The "Pure Trait" Mode l and Its Alternatives Discussions of the "person versus situation" debate traditionally begin with the "pure trait" model (Alston, 1975; January 1988 9 American Psychologist Copyright 1988 by the Amtxican Psychological Association. Inc. 0003-066X/88/$00.75 Vol. 43, No. 1, 23-34 23 Argyle & Little, 1972; Mischel, 1968): that people show powerful, unmodulated consistencies in their behavior across time and diverse situations. This position has been attacked frequently over the years. However, it is really just a "straw man," and even traditional personality researchers find it unacceptable (see, e.g., Allport, 1931, 1966; Block, 1977; Hogan et al., 1977; Jackson, 1983; Wiggins, 1973; Zuroff, 1986). Complete invariance in behavior is associated more with severe psychopathology than with "normal" behavior. If the consensus rejects the "pure trait" position, then what can replace it? Several alternative hypotheses have been advanced over the years. These hypotheses differ with regard to four issues, which can be arranged into a logical hierarchy: 1. Consensus versus solipsism. Are traits merely idiosyncratic constructs that reside solely inside the heads of individual observers, or can observers reach agreement in applying trait terms? 2. Discriminativeness versus generality. I f observers can agree with one another in ascribing traits to targets, is it simply because they apply a nondiscriminative "one size fits all" approach? 3. Behavior versus labeling. I f observers can agree with one another, and can also differentiate between who is low or high on a given trait, does this occur because they really observe behavior? Or do they merely provide their judgments based on superficial stereotypes, targets' self-presentations, or other socially assigned labels? 4. Internal versus external locus of causal explanation. If observers can agree with one another and can distinguish individual differences on the basis of actual behavior of the people they are observing, are the causes ofthese consistencies located within each person or within his or her situation and role? Each of these issues depends on the resolution of those earlier in the list. For instance, if observers cannot agree with one another about who has which traits, there is no point in going on to debate whether traits have a behavioral basis. Ultimately, assumptions about traits must pass the tests of consensus, discriminativeness, behavioral foundation, and internality. We will discuss seven hypotheses that assume that traits fail one or more of these tests. In Table 1, we list the hypotheses in terms of the four hierarchical issues just discussed. As can be seen, the hypotheses can be arranged more or less in order of their pessimism regarding the existence of (consensually verifiable, discriminative, internal) trait-like consistencies. I We will consider each hypothesis in its purest form and, for the moment, disregard the various qualifications that have sometimes been attached to each. Placing each hypothesis in bold relief allows us to assess it most clearly, and philosophers of science tell us that we learn most when hypotheses are stated in such a way as to allow disproof(e.g., Platt, 1964; Popper, 1959). Moreover, each of these hypotheses has, at some time, actually been stated in its bold form. In 1968, for instance, one social psychologist argued that T a b l e 1 Hierarchy of Hypotheses From the Person-Situation Controversy, Arranged From Most to Least Pessimistic Cdtical assumptions Hypotheses