Meta-Analytic Review of Marital Therapy Outcome Research Ryan L. Dunn Medical Group Mountain Home Air Force Base Andrew I. Schwebel Department of Psychology Ohio State University ABSTRACT This meta-analytic review examines the findings of 15 methodologically rigorous marital therapy outcome studies reported in 19 journal articles. These findings were used to assess the efficacy of three treatment approaches in fostering change in spouses' relationship-related behavior, cognitions, affect, and general assessment of their relationship. Behavioral marital therapy (BMT), cognitive—behavioral marital therapy (CBMT), and insight-oriented marital therapy (IOMT) were all found to be more effective than no treatment in bringing change in spouses' behavior and in the general assessment of their relationship. IOMT was more effective than BMT or CBMT in bringing change in spouses' general relationship assessment, while CBMT was the only approach that induced significant change in spouses' posttherapy relationship-related cognitions. The role of meta-analytic reviews of marital therapy approaches is also discussed. Correspondence may be addressed to Andrew I. Schwebel, Department of Psychology, Ohio State University, 1885 Neil Avenue Mall, Columbus, Ohio, 43210. Electronic mail may be sent to [email protected]. Received: March 10, 1994 Revised: July 6, 1994 Accepted: July 6, 1994 Although most narrative reviewers of marital therapy outcome studies have concluded that this intervention is effective (e.g., Gurman, Kniskern, & Pinsoff, 1986 ; Piercy & Sprenkle, 1990 ), some authors have been more skeptical ( Bednar, Burlingame, & Masters, 1988 ), while others suggest that some approaches are effective and others are not ( Beach & O'Leary, 1985 ). One explanation for these conflicting conclusions is that each reviewer has approached the literature with a unique perspective. As Cooper (1989) noted, "traditionally, reviewers [have] interpreted data using rules of inference unknown even to them. . . . Analysis and interpretation methods were idiosyncratic . . . [and] this subjectivity in analysis and interpretation led to skepticism about the conclusions of many reviews" (p. 82). In contrast to the more-or-less "eyeball" judgments that narrative reviewers must make about a discrete group of studies ( Hahlweg & Markman, 1988 ; Rosenthal, 1984 ), meta-analytic reviewers use empirical methods that remove much subjectivity in the evaluation of a body of literature. In addition, meta-analytic reviewers are less likely to underestimate treatment efficacy than narrative reviewers who, in an effort to avoid Type I errors, are prone to make Type II errors ( Lipsey & Wilson, 1993 ). Meta-Analysis in Marital Therapy Journal of Family Psychology © 1995 by the American Psychological Association March 1995 Vol. 9, No. 1, 58-68 For personal use only--not for distribution. Page 1 of 1311/27/2000http://spider.apa.org/ftdocs/fam/1995/march/fam9158.htmlHahlweg and Markman (1988) published the first meta-analysis that addressed marital therapy outcome studies separately from family therapy outcome studies. They assessed 17 behavioral marital therapy (BMT) outcome studies and reported (a) that BMT is more effective in alleviating marital distress than no treatment in both Europeans and Americans, (b) that improvements are likely to remain stable for at least a year, (c) that self-report and observational measures indicate similar changes in couples, and (d) that despite improvement, clinical measures suggested that many couples remained distressed after BMT treatment. Plattor (1990) conducted a more compre-hensive meta-analytic review. It included 25 studies published between 1970 and 1988 that had randomly assigned participants to marital therapy approaches and that had used a no-treatment or waiting-list control group. Treatment groups were categorized following Beach and O'Leary's (1985) divisions: communication training alone, contracting alone, contracting withcommunication training, and insight-oriented marital therapy. Plattor (1990) found no differences in effectiveness among these approaches across a general population of distressed couples. Although Plattor's (1990) review represented a significant advance in methodological rigor, it did not address several important issues. First, studies from the early 1970s to the late 1980s were grouped together, which is problematic in light of the evolving nature of most marital therapies (e.g., Jacobson, 1991 ). Second, three of the four treatment group categories compared are standard components of behavioral therapy, share identical paradigms, use similar and often overlapping techniques, and have been shown to have similar effects ( Jacobson, 1984 ). Third, couples with presenting problems of chronic pain and alcoholism were combined with couples complaining of marital conflict and communication problems, thus increasing heterogeneity and reducing power. Fourth, because multiple data sets from some studies were included in the analyses, inappropriate weighting may have been given to certain samples ( Baucom, 1982 ; Baucom & Lester, 1986 ; Hahlweg, Revenstorf, & Schindler, 1982 ; Hahlweg, Schindler, Revenstorf, & Brengelmann, 1984 ). Finally, although Plattor's study reported useful comparative data, it did not address the absolute effectiveness of each approach–that is, if it was significantly more effective than no treatment and, if so, by how much more. Building on Hahlweg and Markman's (1988) and Plattor's (1990) work, the present study reviewed marital therapy outcome studies published 1980—1993 that have investigated the efficacy of BMT, cognitive—behavioral marital therapy (CBMT), and insight-oriented marital therapy (IOMT) approaches. The studies included met methodological criteria identified as important by meta-analytic researchers ( Giblin, Sprenkle, & Sheehan, 1985 ; Glass & Kliegl, 1983 ). Specifically, the studies (a) randomly assigned participants to treatment conditions, (b) included a control group, (c) used a design that had no major threats to internal validity, (d) reported sufficient data to allow meta-analytic computations, and (e) clearly described treatment procedures so that informed decisions could be made in grouping treatments. In planning the present review, we took various factors into account–including the different theoretical underpinnings of BMT, CBMT, and IOMT–and the fact that
View Full Document
Unlocking...