Skip to content

The Role of the Allegiance Effect in Psychotherapy Research: Controlling and Accounting for It

by Neil 5 Jacobson, Neil S Jacobson
()

Abstract

The allegiance effect is by now a well-established phe-nomenon in psychotherapy research. It is of interest both because of its potential for contaminating between-group comparisons and because no definitive explanation for it has been generally accepted by the field. This commentary underscores the need to control for allegiance effects but also speculates on ways to account for it. Because of difficulties in eliminating the allegiance effect, we might consider it a phenomenon worthy of study in its own right, so we can at least understand it. The innovative documentation of this effect by Luborsky et al. (this issue) serves as another reminder of just how cautious we must be when evaluating the results ofpsychotherapies based on one study. Ths is espe-cially true when the first clinical trial evaluating a new treatment is conducted in the laboratory of the innovator who developed the treatment. Not only does the alle-giance effect teach us that it is especially important to rep-licate, but also the replications must be truly independent and not create new and conflicting confounds with counterallegiance. Such replications are easier said than done. Let me take two examples to illustrate both the need for replication and the potential Pandora's box to which we are subjected when we do replicate. First, the alle-giance effect may have rendered uninterpretable the most expensive and influential study ever done on treatment for depression: the Treatment of Depression Collaborative Research Program (TDCRP; Elkin et al., 1989; Jacob-son & Hollon, 1996a, 1996b). Because the study was con-ducted at three sites, with differing allegiances at the sites, the site differences create a possibility that cannot be refuted Particular psychosocial treatments for depression, especially cognitive therapy (CT), were not adequately tested at one or more of the sites (Jacobson & Hollon, 1996b). The site differences, especially in the performance ofCT, preclude any conclusions about the relative efficacy of pharmacotherapy/clinical management and cognitive therapy with severely depressed outpatients. The fact that the site differences were not statistically significant is irrel-evant: The TDCRP clearly lacked adequate statistical power for detecting such differences, and the size of the treatment X site interaction effects was substantial. I think it is hard to avoid the conclusion that C T was compro-mised at two of the sites and only fairly tested at the one site where the theoretical orientation was already compat-ible with CT. At this latter site (presumably in Oklahoma), CT performed as well as both imipramine/clinical man-agement and the other psychosocial treatment, interper-sonal psychotherapy. The second example involves a study comparing insight-oriented marital therapy (IOMT) to behavioral marital therapy (BMT; Snyder & Wills, 1989). There were several methodological strengths in this study, but among its shortcomings was a subtle but fatal example of the allegiance effect in action. Although the principal investigator (Snyder) had no allegiance to either approach and was truly attempting a fair test of the two treatments, the co-investigator (Wills), who wrote the manuals and supervised therapists in both conditions, was psychody-namically trained and practiced psychodynamic couple therapy. Moreover, he was never trained in BMT. Although I have no doubt that both investigators were well intentioned and implemented this study with the utmost integrity, BMT was compromised (Jacobson, 1989). When the investigators reported a significantly lower divorce rate at the 4-year follow-up favoring IOMT (Snyder, Wills, & Grady-Fletcher, 1991), the interpreta-tion of these comparisons is hopelessly compromised by the possibility that BMT was not adequately tested. Although independent ratings of the competence demon-strated by therapists conducting BMT may have helped to unravel the confound created by the allegiance of the co-0 1999 AMERICAN PSYCHOLOGICAL ASSOCIATION D 1 2 116

Cite this document (BETA)

Readership Statistics

1 Reader on Mendeley
by Discipline
 
100% Psychology
by Academic Status
 
100% Student > Master

Sign up today - FREE

Mendeley saves you time finding and organizing research. Learn more

  • All your research in one place
  • Add and import papers easily
  • Access it anywhere, anytime

Start using Mendeley in seconds!

Sign up & Download

Already have an account? Sign in