ABSTRACT
Psychological tests, including self-report scales, are often treated as tools for predicting rather remote nontest behaviors, i.e., behaviors very different from the test behavior itself. Behaviorally oriented psychologists have tended to target specific behaviors in specific situations, while psychometrically and trait-oriented psychologists have been interested in more encompassing classes of situations and responses such as those represented by personality tests. The use of Research Diagnostic Criteria is often recommended, resulting in the selection of narrowly defined samples. A restrictive selection can be desirable for testing specific and theoretically focused research hypotheses. Self-reports can have informative nontest correlates, but as in the case of other psychological measures, they can be used to much greater advantage in combination with other indicators. Discrepancies as well as congruencies between different measures can then be observed and interpreted.
