Comparative credibility of treatment rationales: three tests of expectancy theory

Previous research has challenged the adequacy of conventional "placebo' controls, by demonstrating that active treatments, such as systematic desensitization, are more credible than supposedly inert control conditions. Such findings give support to the hypothesis that the arousal of positive expectancies is the key to therapeutic change. Three experiments are reported, in which Borkovec's credibility rating method was applied to treatment comparisons modelled upon the findings of outcome research. In the first experiment, systematic desensitization was more credible than rational-emotive and client-centered therapy (P less than 0.001), consistent with an expectancy-arousal explanation of Di Loreto's (1971) comparative outcome study. In the second experiment, Paul's (1966) attention placebo condition was less credible than three active treatments (P less than 0.001). In the third experiment, in vivo desensitization was in some respects more credible than imaginal desensitization, which was in turn more credible in some respects than relaxation alone (P less than 0.001). The results of the latter two experiments were broadly consistent with an expectancy-arousal interpretation of much outcome literature. These results emphasize the insufficiency of merely including a designated "placebo control condition' in outcome research (Kazdin & Wilcoxon, 1976). Credibility differences did not mirror precisely the results of the outcome research on which these studies were modelled, however. Methodological limitations are discussed, and further clinical research is advocated, in which the persuasiveness of treatment rationales is studied in its own right as contribution to treatment efficacy.