Despite a widespread tendency among people to describe themselves and others in traitlike terms (intelligent, friendly, aggressive, domineering, and so forth), research has shown that a person's behavior cannot be predicted from a few broadly generalized personality traits. This does not mean that behavior is totally inconsistent, but that dispositions alone are insufficient to explain consistency and that dispositional, as well as situational, variables need to be taken into account for a complete analysis.
To separate the effects of person and situation variables on behavior, Mischel and his colleagues conducted a series of experiments. In one study, the experimenters assessed adolescents' dispositions toward success or failure. Weeks later they had them solve skill-related tasks and, regardless of their actual performance, gave one group success, a second group failure, and a third group no feedback on their performance. Then the adolescents had to choose between a less desirable reward, one for which attainment was independent of performance on similar tasks, and a preferred reward, for which attainment was performance-dependent. In both bogus feedback conditions, the situational variables had a powerful effect and completely overrode preexisting dispositions toward success or failure. Adolescents who believed they had failed the tasks more often selected the noncon-tingent reward, while those who believed they had succeeded chose the contingent reward. For subjects in the no-feedback condition, however, the preexisting expectancy scores were highly accurate predictors of their reward choices. This study illustrates how dispositions emerge under weak situa-tional cues but play a trivial role when the setting provides strong cues for behavior. Therefore, Mischel (1973) considers it more meaningful to analyze "behavior-contingency units" that link specific behavior patterns to those conditions in which they are likely to occur, rather than looking only at behavior. In other words, instead of labeling people "aggressive," it would be more useful to specify under what conditions these people display aggressive behaviors. Such precise specifications would guard against an oversimplified trait approach and highlight the complexities and idiosyncrasies of behavior as well as its interdependence with specific stimulus conditions.
Was this article helpful?