Story of Psychology - Morton Hunt [206]
Ratings by observers: Researchers often ask an individual’s friends or acquaintances to rate him or her on a number of specified traits. To achieve precision, the researchers direct respondents to weigh each trait on a scale that runs from zero to five or perhaps one to ten—essentially what Thomasius suggested in 1692. But the method has many difficulties. Raters have their own styles of rating (some avoid extremes, others favor them); subjects are not necessarily consistent when asked the same questions at different times; and ratings are subject to the “halo effect” (a subject rated high for one trait tends to be rated high for others).21
In general, then, ratings are considered neither especially reliable nor especially valid. (A reliable method yields consistent answers time after time; a valid method measures what it is supposed to be measuring.) Still, under certain conditions ratings can be both reliable and valid. Raymond Cattell, a leading trait researcher who relied on them in his own work, used only data from raters who saw the subject under many circumstances and over a long time (a year, if possible) and gathered ratings on only one trait at a time to reduce the halo effect. Such conditions improve both reliability and validity but make the method prohibitively costly, time-consuming, and nearly impossible to use anywhere except in an institution, where the population is relatively fixed and always visible.22
The questionnaire: This is by far the most commonly used tool for personality assessment. As we have seen, the method quickly expanded beyond simple self-evaluation to quasi-objective techniques, such as presenting real-life situations and asking respondents how they would most likely behave in them. Other early tests continued to present questions about the respondent’s attitudes and feelings rather than probable behavior but were worded in ways that made the respondent less likely to prettify his self-portrait than did the questions in the Personal Data Sheet. Most offered as possible answers “yes-no” or “true-false” options, but some included a “don’t know” middle ground.
The famous Minnesota Multiphasic Personality Inventory (MMPI), developed in the late 1930s by Starke Hathaway, a psychologist, and J. C. McKinley, a psychiatrist, both of the University of Minnesota, is of the latter type. It contained 550 statements, among them:
I am happy most of the time.
I enjoy social gatherings just to be with people. I am certainly lacking in self-confidence.
I believe I am a condemned person.
The respondent answers “yes,” “no,” or “?” (uncertain) to each question. The questions were grouped into ten scales that measured hypochondriasis, depression, hysteria, psychopathic deviancy, masculinity-femininity, paranoia, psychasthenia, schizophrenia, hypomania, and social introversion. These names convey the impression that the MMPI was concerned chiefly with mental illness; it did measure mental illness but also traits of normal personality. Those who, for instance, answered “false” to “I am happy most of the time” and most other questions in the same scale were said to be shrewd, guarded, and worrisome; those who answered “true” to “I enjoy social gatherings just to be with people” and related questions were rated sociable, colorful, and ambitious; and those who answered “false” were rated modest, shy, and self-effacing.23
Such interpretations were based not on intuition or common sense but on empirical evidence. In constructing the MMPI, Hathaway and McKinley tried a large number of questions on people hospitalized with neurosis or mental illness and on the presumably normal people who came to visit them; the MMPI was made up only of those items