C(lnrca[PsychologyRealrw, Vol 13, pp. 169-186, 1993 Printed in the USA. All rights reserved. 0272.7358/93 $6.00 + .oo Copyright 0 1993 Pergamon Press Ltd. zyxwvutsr A META - A NA LYSIS O F SELF- HELP TREA TMENT A PPRO A C HES Robert A. Could George A. C/urn Virginia Polytechnic institute and State University ABSTRACT. zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA We conducted a meta-analysis of40 self-help studies examining 61 treatments which used as control groups no-treatment, wait-list, or phuebo comparisons. W e found an overall treatment effect size for self-help interventions of 0.76 at posttreatment and zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONM an effect size of 0.53 atfollow-up. Studies that used a more stringent control group -placebo-had a lower mean effect size than studies utilizing a no-treatment control. No differences were foundfor unadulterated self-help treatments and those with minimal contact from a therapist or which were in fact therapist-assisted, Some target problems were more amenable to self-help approaches, including skills deficits and diagnostic problems, stub as fears, depression, he&he, and sleep disturbance. Habit disturbances such as smoking drinking, and overeating, were less amenable to self-help treatments. Compliance with the treatment regimen was found to improve treatment effects. W hile effect size was negatively related to duration of treatment, this effect was largely a function of duration being confounded with type of problem. Results were examinedfrom the perspective of problems that are targetable with self- help approaches. Recommendations for improvements in research design for future studies are also made. It has been 10 years (Glasgow & Rosen, 1982) since a published review of the self-help literature has appeared. At that time Glasgow and Rosen were somewhat sanguine about the future for psychologically oriented self-help products, noting that a number of them successfully demonstrated their validity in controlled clinical studies. This sentiment echoed the conclusions of an earlier review (Glasgow & Rosen, 1978). By 1987, however, Rosen, while noting the potential benefits of self-help materials, was much less hopeful and, indeed, was focusing on the many risks of such materials. The present review sought to bring the basic question of the effectiveness of self-help materials up to date. It w as further interested in applying meta-analytic methodology to this task. Such an approach permitted a much more comprehensive analysis of the benefits and limitations of this field. The previous reviews of this literature left a number of unanswered questions. First was the issue of whether self-help materials were effective. Because of the vast differences in methodological rigor among studies reviewed, it was impossible for Glasgow and Rosen to Correspondence should be addressed to George A. Clum, Virginia Tech, Department of Psycholo- gy, Blacksburg, VA 24061-0436. I69