Epstein (1986) has demonstrated that aggregation leads to high correlations which are not spurious but which reflect only stability. He claims that there is no distinction between a demonstration of test-retest reliability and behavioral stability. It is suggested that Epstein has confused personality theory and measurement models and has made the logical fallacy of claiming that high correlation coefficients uniquely reflect stability. Epstein has failed to distinguish clearly between reaction variables and mediating variables and between temporal stability and cross-situational consistency. While high correlation coefficients may reflect stability there is no necessary one-to-one relationship between personality theories and their measurement models nor between reaction variables and mediating variables. The interaction model recognizes that behavior is consistent in the sense of coherence, i.e. a lawful idiographically predictable pattern of behavior. This pattern involves personsituation interactions, and changes over time. Aggregation may obscure these patterns.