Models and meanings: Therapist effects and the stories we tell

Two studies, same data set, opposite conclusions. Therapist effects are either vital or virtually nonexistent. Previous research either supports the importance of therapist effects or provides little or no evidence that these effects exist. What kind of science is this? The question the articles in this special section address is an important one for psychotherapy research: Is there major variation in outcomes between therapists? Since the pioneering work of Crits-Christoph et al. (Crits-Christoph, Baranackie, Kurcias, & Beck, 1991; Crits-Christoph & Mintz, 1991), psychotherapy researchers have been aware that therapist effects can be a potential confound in studies of treatment effectiveness. Leaving these effects out of analyses can result in incorrect parameter estimation and statistical tests and the drawing of incorrect inferences. More practically, the existence of substantial therapist differences in effectiveness could lead to the examination of the characteristics and practices of the more effective therapists as contrasted with those of the least effective therapists; the resultant knowledge could help improve psychotherapy practice. Both articles in this special section, by Elkin, Falconnier, Martinorich, and Mahoney, and Kim, Wampold and Bolt, examine the existence of therapist