Abstract
Objective: The purpose of this study was to investigate the reliability of change indices in exposure therapy. Method: We analyzed data from 119 adults with elevated social anxiety who participated in studies of public speaking exposures. Participants reported their distress using the Subjective Units of Distress Scale (SUDS) at regular intervals. The reliability of within-session habituation, between-session habituation, and expectancy change measures were estimated. Results: Commonly used measures of between-session habituation (BSH) typically demonstrated only moderate reliability, and commonly used measures of within-session habituation (WSH) typically demonstrated poor reliability. Exploratory tests in expectancy change indices suggests these findings apply to other measures of change during exposure. Conclusions: Our results indicate that measures underlying the study of habituation may be inadequate because of low reliability. This threatens the foundation of clinical and translational research, such that no substantive conclusions can be drawn from them.