This paper describes a therapy analogue experiment in which an interviewer's interpretation of a client's presenting problem was systematically varied along dimensions suggested by causal attribution theory. Subjects were 64 undergraduate students with the same presenting problem of unassertiveness. They were given two different interpretations of their difficulty. One group was given a behavioristic explanation asserting that their problems were caused by an unstable, externally induced conditioning process. The other group was given a "depth" explanation which suggested that their problems were caused by a subtle internal personality trait with long and complex origins. Results indicated no differences in expectancy for change and in behavior change between the two groups. Immediately following the interview, those receiving the conditioning interpretation reported more anxiety than those receiving the trait interpretation, but a week later the differences had attenuated. These results are discussed in terms of the "Rumpelstiltskin Effect" and the equipotentiality of different insights. (Author)