Going quasi: the premature disclosure effect in a randomized clinical trial

Abstract

This paper describes a randomized clinical trial investigating a stress management program for women with breast cancer, which inadvertently turned quasi-experimental. Due to logistical considerations, group assignment was disclosed to participants (n = 63) prior to baseline assessment. Analyses of baseline measures unexpectedly revealed statistically significant differences between groups on psychological functioning. We suggest that what appears to be failed randomization may in fact point toward an important phenomenon which we have termed premature disclosure effect (PDE). A hierarchical regression model, developed to help explain the PDE, accounted for 47% of the variance. The findings indicate the importance of considering participant belief, preferences, and attributes when designing research protocols and interventions. Potential implications of PDE for clinical research in behavioral medicine are discussed and specific statistical methodologies suggested.