The Need for New Constructs

One could not ask for a clearer opposition of views on continuity than the articles by Michael Lewis (this issue), on the one hand, and Avshalom Caspi and Brent Roberts (this issue), on the other. The former scholar questions whether any personality traits are preserved, whereas the latter pair celebrates preservation. I suspect that most scientists in this field regard a middle position as the one that matches current data best. There is some preservation, but as Lewis noted, it is modest. I suggest that deeper insights in this domain will not occur if psychologists continue to rely only on self-report instruments to measure personality. Most of the work cited by Caspi and Roberts (this issue) use questionnaires to assess personality categories or dimensions. However, different personality types would emerge if the thousands of people who filled out those questionnaires had been filmed for 24 hr over 6-month period in different contexts and scientists factor analyzed those observations. Advocates of the popular personality constructs assume, without any detailed argument, that a person's semantic judgments about their behaviors, beliefs, and moods are the most valid bases for inferring personality types. This premise is surprising for no biologist would use the reports of informants, even sophisticated ones, to decide on the basic human diseases. No economist would rely only on interviews with consumers to discover the fundamental economic activities of a society. No cognitive psychologist would record participants' careful descriptions of their perceptions, memories, and ways of solving problems to infer the basic cognitive competencies. Self-report instruments have several serious problems. The first flaw is based on the semantic constraints on judgments about self. A woman who affirms on a questionnaire that she likes meeting new people is biased to respond affirmatively to all questions semantically related to that statement to maintain consistency. Second, if a human trait does not have a popular semantic name, personality questionnaires do not contain relevant items. More important, most personality categories describe relatively superficial evaluative dispositions like sociality, conscientiousness, nervousness, and openness to ideas. Each of the four is regarded as either a good or bad trait by most Americans and all four are likely to be characteristics one would ask a friend about when a stranger moved into the neighborhood. By contrast, constructs like chronic feeling tone, ambivalence over goals, nodes of uncertainty, energy level, emotional reactions to individuals with higher status, and intensity of sexual arousal are difficult to measure with questionnaires and may represent sources of variation with significant influences on behavior and major life choices. Finally, answers to questionnaires are influenced by a comparison of past with present, and the accuracy with which the past is retrieved. Sadly, adults often construct judgments about their moods and behaviors on the spot using information that is most accessible. A woman's answers to questions about her sociability represent her sense of meaning of extroversion or introversion, but not the referential meaning. It is the latter that determint s whether the proposition is true or false.