implicit personality theory

n. Implicit personality theory (IPT) refers to laypersons' co-occurrence expectancies for traits and behaviors. Persons expect, for example, that gregarious persons are talkative as well. Such assumptions tend to be shared by people. Thus the meaning of IPT is much more specific than the meaning of "implicit theories," which refer to lay theories concerning a wide range of phenomena.

The structure of IPT has been studied using two approaches, one direct and the other indirect. The direct approach involves letting judges estimate conditional likelihoods that persons with attribute A will have attribute B as well, for instance, that a gregarious person will also be talkative. The indirect approach relies on ratings by strangers that are based on minimal information. Strangers tend to disagree concerning the attributes of a particular person (what goes with whom), whereas within-judge correlations between ratings of different attributes (what goes with what) are similar across judges, reflecting their shared assumptions on attribute co-occurrences. For example, observers whose only information on Alice's personality is a photograph of her are likely to disagree on her nervousness. Nevertheless, judges who perceive Alice as nervous are likely to perceive her as anxious as well.

Direct and indirect approaches to IPT have yielded similar findings concerning its structure, and this structure resembles the structure of ratings by knowledgeable informants to some extent. But there are differences as well: ratings by strangers are more highly correlated than self-reports and ratings by close acquaintances, implying that the more judges know about a person, the more complex is their description of that person’s personality. Moreover, IPT does not reflect asymmetries in conditional likelihoods that result from different base rates. For instance, because friendliness is more widespread than homosexuality, it is more likely that a homosexual person will also be friendly, compared to that a friendly person will also be homosexual. But judges estimate these two conditional likelihoods as by and large the same.

Experiments on person memory show that IPT operates as a schema: If research participants do not remember all information on fictitious characters, their recall tends to be biased in IPT-consistent ways. Thus, as far as information about personality cannot be recalled, it seems that persons rely on their IPT to fill the gaps. A related phenomenon is illusory correlations, that is, that persons sometimes report correlations that do not exist. One source of illusory correlations seems to be associative relationships, another shared distinctiveness: the co-occurrence of two rare events tends to be overestimated because rare events attract more attention and are therefore more accessible in the mental networks of their observers. This might explain why stereotyping is particularly strong for members of minority groups.

From experimental findings that co-occurrences may be misperceived, it has been concluded that the correlations among ratings of personality are fundamentally flawed, and that they reflect the structure of IPT but not the structure of personality. This, however, is contentious: Proponents of the view that IPT is illusory argue that it reflects the associative relationships among personality attributes that are unrelated to their actual co-occurrences, and that IPT therefore systematically distorts the correlations among ratings of personality. By contrast, proponents of the view that IPT is accurate argue that it reflects the accurately perceived co-occurrences among traits and behaviors and therefore does not contribute to biased correlations among ratings of personality. The issue is important because correlations among ratings of personality are a major data source in various fields of psychology. But it seems that the controversy on the illusory versus veridical nature of IPT has been misconceived as that actual co-occurrences among personality attributes partly reflect their meaning relations: There is evidence that similar traits refer to overlapping sets of indicators; for example, that nervousness is correlated with anxiety partly reflects that many indicators of nervousness (unpleasant affect, physiological arousal) are indicators of anxiety as well.

- PB