This paper provides a review of previous literature reviews on the effectiveness ofmarital therapy and evaluates the long­term efficacy of couples therapy inpreventing marital separation and divorce. There is general agreement thatmarital therapy can be effective in reducing marital conflict and promoting maritalsatisfaction, at least in the short term. Research examining the long­term efficacyof couples therapy for the prevention of marital separation and divorce is sparsebut promising with respect to its effectiveness in promoting marital stability.Predictors of successful marital therapy outcomes, the clinical significance ofmarital therapy outcomes, and the cost effectiveness of marital therapy are alsodiscussed.It is estimated that one half to two thirds of all marriages in the United States willexperience disruption due to separation or divorce (Castro Martin & Bumpass,1989; Norton & Moorman, 1987). Although divorce and marital conflict are notalways viewed as negative (Gottman, 1993), they can have a major impact on thehealth and well­being of all family members (Bray & Hetherington, 1993).Separated and divorced adults have the highest rates of acute medical problems,chronic medical conditions, and disability (Verbrugge, 1979). Divorced men are atincreased risk for suicide, admission to mental hospitals, vulnerability to physicalillness, and becoming victims of violence; separated and divorced women are atincreased risk for depression and increased utilization of medical services(Bloom, Asher, & White, 1978; Dorian et al., 1982; Zeiss, Zeiss, & Johnson,1980). Children in families characterized by marital conflict or divorce are atgreater risk for a variety of behavioral and emotional problems, includingoppositional behavior, aggression, and symptoms of depression and anxiety(Bray & Hetherington, 1993; Emery, 1982; Jouriles, Farris, & McDonald, 1991). Insum, divorce and marital conflict are associated with multiple family problemsthat negatively affect the well­being of family members.Reduction of marital conflict and the prevention of divorce should represent highpriorities for modern families. A number of important reviews of the maritaltherapy literature have already appeared (Baucom & Hoffman, 1986; Bradbury &Fincham, 1990; Dunn & Schwebel, 1995; Gurman & Kniskern, 1981; Gurman,Kniskern, & Pinsof, 1986; Hahlweg & Markman, 1988; Jacobson & Addis, 1993;O’Leary & Smith, 1991; Raffa, Sypek, & Vogel, 1990; Shadish et al., 1993;Williams & Miller, 1981). This paper reviews previous reviews on the effectiveness of marital therapy in reducing marital conflict. This paper alsoreviews in detail our knowledge of the long­term effectiveness of couples therapyin preventing marital separation and divorce.It should be noted at the outset that divorce is not always a negative outcome ofmarital therapy. Experienced clinicians can attest to the fact that some spousespresenting for marital therapy have already made up their minds that they want adivorce and seek marital therapy for reasons other than improving maritalsatisfaction (e.g., to assuage a partner in a last­ditch effort to salvage therelationship). It is also debatable whether spouses should remain in relationshipsin which there are irreconcilable differences (of the type that are likely to preventgeneral happiness) or whether marital therapy sometimes helps spouses becomeaware of such differences. Children often adjust better in a stable, divorced homethan in an unhappy, highly conflictual intact home (Emery, 1982; Hetherington,Cox, & Cox, 1982). In addition, following divorce, some women develop betterself­esteem, higher levels of competence, and career achievements(Hetherington et al., 1982). This is not to imply that divorce is positive orrecommended for troubled marriages. Most research indicates that being marriedis usually associated with better adjustment and fewer health problems thanbeing divorced or single(Bloom et al., 1978). However, helping couples divorce isa legitimate part of marital therapy, and the therapy may help the adults have abetter postdivorce adjustment. Unfortunately, this aspect of marital therapy israrely considered or evaluated in outcome research.EFFECTIVENESS OF MARITAL THERAPYReviewers of the marital therapy literature agree that marital therapy can beeffective in reducing conflict and increasing marital satisfaction, at least in theshort term, when compared to no­treatment controls (Baucom & Hoffman, 1986;Bradbury & Fincham, 1990; Dunn & Schwebel, 1995; Gurman & Kniskem, 1981;Gurman et al., 1986; Hahlweg & Markman, 1988; Jacobson & Addis, 1993;O’Leary & Smith, ]991; Shadish et al., 1993; Williams & Miller, 1981). Thisconclusion has been reached by reviewers of varying theoretical perspectives(e.g., behavioral or systems) and by reviewers who have used very differentmethods of evaluating the literature. Moreover, this conclusion is based on aliterature that includes dozens of studies in peer­reviewed journals and many other studies reported in other sources.Behavioral marital therapy (BMT) is the most researched form of marital therapy,and its effectiveness has repeatedly been demonstrated both in the United Statesand in other countries (Hahlweg & Markman, 1988). BMT is a skills­orientedtreatment that includes training in communication and problem­solving skills andbehavior exchange principles. This approach is based, to some degree, onempirical research on functional and distressed marriages (Gottman, 1994;Gottman, Notarius, Gonso, & Markman, 1976; Weiss, 1981). Although othertherapy approaches, such as systems approaches or insight­oriented maritaltherapy, have been given less attention by marital therapy researchers, data alsosupport the efficacy of these approaches when compared to no­treatmentconditions.Reviewers generally use three methods for evaluating research on theeffectiveness of marital therapy. First, in a narrative approach reviewers evaluatestudies on a case­by­case basis using a nonempirical approach to critique thestudies (e.g., Bradbury & Fincham, 1990; Gurman & Kniskern, 1981; Gurman etal., 1986; O’Leary & Smith, 1991; Williams & Miller, 1981). A second method,meta­analysis (Smith, Glass, & Miller, 1980), is a statistical technique that allowscomparison of the effect sizes of various treatments across studies (e.g., Dunn &Schwebel, 1995; Hahlweg & Markman, 1988; Shadish et al., 1993). Effect size(ES) statistics reflect the magnitude of effect that a given treatment has incomparison to a control group. The larger the ES, the more effective thetreatment. Third, evaluation of clinical significance is a method of comparingstudies to determine the clinical relevance of outcomes (e.g., Hahlweg &Markman, 1988; Jacobson, Follette, & Revenstorf, 1984; Shadish et al., 1993).Clinical significance is usually measured by comparing posttreatment outcomesto normative data, such as on measures that discriminate distressed andnondistressed couples. The clinical effectiveness of a treatment is assessed bythe percentage of couples who are in the nondistressed range following therapy.There are advantages and disadvantages inherent in each of these approaches.For example, the descriptive approach is subject to potential reviewer biases; themeta­analytic approach summarizes over a larger number of outcomes andmethods. It is important to reiterate, however, that reviews based on each methodhave concluded that marital therapy is somewhat effective in reducing marital conflict.There are three published meta­analyses of marital therapy outcome literature(Dunn & Schwebel, 1995; Hahlweg & Markman, 1988; Shadish et al., 1993).These reviews indicate that the average posttherapy ES for marital therapiesrange from .51 to .95. The ES varied from .00 to 1.37 depending on the type oftherapy evaluated, the type of dependent variables, when ES are calculated(posttherapy or follow­up), and the range of studies included (published onlyversus published and dissertations). Inclusion of dissertations reduced the overallES. Shadish et al. (1993) and Dunn and Schwebel (1995) replicated theposttherapy findings of Hahlweg and Markman (1988) for BMT with overall ESof .95 and .78, respectively. These ES suggest that the probability ofimprovement at posttreatment is 40% greater for BMT than for no­treatmentcontrols (Hahlweg & Markman, 1988) and that there is a 60% chance that atreatment couple will be better following marital therapy than a couple who did notreceive therapy (Shadish et al., 1993). These ES are generally comparable to ESof other kinds of psychotherapy (Smith et al., 1980). Of note is that in the Shadishreview only one study reported a negative ES and Dunn and Schwebel (1995)reported no negative ES, providing little evidence for significant negative effectsfrom treatment.An important issue to consider in evaluating the results of meta­analysesconcerns the outcome criteria used in marital therapy studies. Dunn andSchwebel (1995) reported different posttherapy ES by outcome variables;behavioral measures ES =.76, cognitive measures ES = .61, affective measuresES = .52, and general relationship measures ES = .90. Most of the outcomestudies include multiple measures of outcome, and these measures arecombined in determining ES. Thus, the ES are computed from measuresassessing different aspects of marital functioning, not just marital conflict.Although reviewers agree that marital therapy is effective in reducing maritalconflict, there is less support for the clinical significance of these outcomes.Jacobson and Addis (1993) conclude that "most tested treatments report nobetter than 50% success" (p. 86). Hahlweg and Markman (1988) found that theclinical significance of change in marital satisfaction and adjustment in BMT wasan average ES of .15, indicating that a substantial number of couples still report unsatisfying and distressed marriages after treatment. Using a broader sample ofmarital therapy studies, Shadish et al. (1993) found that 41 % of couples inmarital therapy moved from distressed to nondistressed status followingtreatment. These findings are sobering. They indicate that existing treatments formarital discord and distress need substantial improvement.One potential confound, particularly in measuring outcomes, concerns whethercouples felt helped by marital therapy if they decided to terminate their marriage.None of the outcome studies have investigated this issue (O’Leary & Smith,1991). As noted earlier, separation or divorce may be a desired outcome fromtherapy, and researchers need to assess whether this was the case either in thebeginning of therapy or posttherapy to determine whether therapy was successfulin meeting the clients’ goals. The recent research on marital assessment andpredictors of divorce may help therapists work more effectively with couples todetermine their purpose for seeking marital therapy (Gottman, 1994; Williams &Jurich, 1995). However, as basic research on marital process indicates, there aremany long­term, stable couples who are quite different from one another in termsof patterns of interactions, desires and expectations, and methods of handlingdiscord (Gottman, 1994). Research on marital therapy outcomes also needs torecognize this and factor in these possibilities in evaluating treatment efficacy.LONG­TERM EFFECTS OF MARITAL THERAPY IN THE PREVENTION OFDIVORCEVery little is known about the long­term effectiveness of couples therapy withrespect to the prevention of marital separation and divorce. Until recently, follow­up data on participants in studies of treatments for marital distress have beenlimited to several months following treatment. In their review of the maritaltherapy literature, Dunn and Schwebel (1995) reported that follow­up periods inmarital therapy research ranged from 1 to 48 months after termination, with amean of 9.2 months posttherapy. These authors report follow­up ES for generalassessments of the relationship ranging from .54 to 1.04. Overall, it is unclearhow effective marital therapy is in promoting long­term marital stability. Althoughrecent reviews and commentaries suggest that a sizable number of couples whoimprove during marital therapy relapse before long­term follow­up assessments(Jacobson & Addis, 1993), only three studies were identified that examined the long­term effectiveness of marital therapy for distressed couples (at least 18months posttreatment), were published in peer­reviewed journals, had randomlyassigned couples to treatment and comparison conditions, and reported resultsin a manner that allowed for the computation of marital separation and divorcerates.Crowe (1978) examined the effectiveness of three conjoint marital therapyapproaches (a directive approach, an interpretive approach, and a supportive­control approach) over 18 months. The directive approach emphasized theincrease of positive interpersonal behavior, the interpretive approach centeredaround interpretation of feelings and conflict, and the supportive­control approachencouraged couples to talk with each other, avoiding both directives andinterpretations. At the 18­month follow­up assessment, the directive groupshowed significantly more improvement on marital adjustment than thesupportive­control group. Seven of the 42 couples in this study had separated bythe 18­month follow­up: 1 from the directive, 2 from the interpretive, and 4 fromthe supportive­control conditions. These differences across groups were notstatistically significant. Combination of the data from the directive and interpretiveapproaches indicated that 11% of the couples who received a marital therapyseparated by the 18­month follow­up assessment, in comparison to 29% of thecouples who received the supportive­control approach. Again, the difference didnot reach traditional levels of statistical significance.In a component analysis of BMT, Jacobson, Schmaling, and Holtzworth­Munroe(1987) compared a complete BMT treatment package (CBMT) to twocomponents of BMT: behavior exchange (BE) and communication and problem­solving training (CO). The BE intervention emphasized homework assignmentsdesigned to instigate positive behavior changes in each spouse. The COintervention focused on teaching couples skills with which to deal effectively withrelationship conflicts. The CBMT combined key elements of BE and CO. Thisstudy might be conceptualized as a comparison between a marital interventionand two theoretically less potent marital interventions. At a 2­year follow­upassessment, there were no statistically significant differences between groups onmeasures of marital adjustment and presenting problems. However, couples inthe CBMT condition were more likely to be "happily married" than couples in theother two conditions. Moreover, only 9% of couples who received CBMT separated or divorced, in comparison to 45% of those who received one of thecomponents of BMT, a difference which is statistically significant.Snyder, Wills, and Grady­Fletcher (1991) collected 4­year follow­up data oncouples participating in a study comparing behavioral and insight­oriented maritaltherapies. The BMT included behavioral exchange and communication andproblem­solving training. The insight­oriented marital therapy centered aroundinterpretations of both intrapsychic and interpersonal dynamics contributing tomarital conflict. At the 4­year follow­up assessment, 3% of the couples whoreceived insight­oriented marital therapy and 38% of the couples who receivedBMT had separated or divorced, a difference which is also statistically significant.Although meager, the literature on the long­term outcome of marital therapy maybe construed as encouraging from the perspective of preventing maritalseparation and divorce. The two outcome studies that incorporated comparisonconditions, interventions theoretically less potent than complete marital therapies(i.e., Crowe, 1978; Jacobson et al., 1987), resulted in fewer separations anddivorces among couples receiving the complete marital interventions. However,only the Jacobson et al. (1987) study yielded statistically significant differences.When data from these two studies were combined (40 families received acomplete marital intervention and 36 received a less potent intervention), 90% ofthe couples who received a complete marital intervention were still together 18 to24 months later, in comparison to 61% of the couples who received a less potentintervention (x sup 2 (1) = 23.16, p < .0l). It is also interesting to note that this90% "stability" rate is based primarily on forms of BMT (26 of the 40 familiesreceived a complete marital intervention based on BMT). Somewhat lessencouraging, however, is the finding from Jacobson et al. (1987) which suggeststhat many of the couples who remain together following marital therapy are nothappy in their marriages. Specifically, only 50% (6 of 12) of couples whoparticipated in their complete marital therapy intervention were classified as"happily married" 2 years posttreatment. In addition, there is an absence of dataon the long­term effects of marital therapy in reducing the frequency or intensityof marital conflict.The results of the Snyder et al. (1991) study suggest that an insight­orientedmodel might be even more effective than a behavioral model in fostering long­ term stability. However, it should be noted that Jacobson argues that the insight­oriented model evaluated by Snyder is really a modern version of BMT, and theBMT used in that study was an early version that was known to be less effectivethan later versions of BMT (Jacobson, 1991; Jacobson & Addis, 1993). In sum,the long­term effectiveness of marital therapy is not a glowing success story, butfindings do suggest that such therapies are promising in promoting maritalstability. Gaping holes exist in the research literature which need to be filled byfuture research efforts that focus on different types of long­term outcomes acrossthe couple life cycle.A relevant question in evaluating the long­term effectiveness of marital therapy,or any psychotherapy, concerns whether it is fair to assume that one series oftherapy is or should be all that is needed during a couple’s lifetime. Alternatively,is it reasonable to assume that couples may need therapy at different points intime? Consequently, if a couple needs additional therapy it does not mean thatthe first series was not effective. Some therapy models assume that people mayneed brief psychotherapy at various points across the life cycle (Cummings,1986). Most studies do not examine this possibility and there is very little writtenabout this in the literature.PREVENTION OF MARITAL CONFLICT AND DIVORCEIn contrast to more clinical, treatment­oriented conceptualizations of maritaltherapy in which interventions are initiated with families already experiencingmarital distress (often at very severe levels), prevention programs target couplescharacterized by little or no distress. The overarching goal of most of theseprograms is to help couples maintain satisfactory levels of marital functioning andto prevent the development of marital dysfunction and divorce. Detaileddescriptions of several of these programs are available in the literature (seeGuerney, Guerney, & Cooney, 1985; Levant, 1986, for overviews). The existingprevention programs are diverse, with programs differing widely on manypotentially important clinical dimensions, for example, the theoretical/scientificbasis for the intervention, the phase of the relationship in which the intervention isinitiated, and the length and format of the intervention (Bradbury & Fincham,1990). Evidence from controlled outcome studies for the efficacy of most of theseprograms in preventing divorce is lacking. However, outcome data do suggest that a number of these programs produce short­term gains in communicationskills and relationship satisfaction. Furthermore, Hahlweg and Markman’s (1988)meta­analysis reported that behaviorally oriented prevention programs had apositive effect in the short term.A notable exception to the absence of controlled outcome studies in theprevention area is the work of Markman and colleagues in the development andevaluation of their Premarital Relationship Enrichment Program (PREP; Markman& Hahlweg, 1993; Markman, Floyd, Stanley, & Lewis, 1986; Markman, Floyd,Stanley, & Storaasli, 1988; Markman, Renick, Floyd, Stanley, & Clements, 1993;Renick, Blumberg, & Markman, 1992). PREP is based on a relatively largetheoretical and empirical literature linking communication and problem­solvingskills to effective marital functioning (Gottman et al., 1976). PREP borrowstechniques and methods from other intervention packages, such as relationshipenhancement therapy (Guerney et al., 1985) and BMT (Jacobson & Margolin,1979; Weiss, 1981).Results of a 5­year follow­up evaluation of PREP have recently been published(Markman et al., 1993). This evaluation was conducted with 114 couples planningmarriage for the first time. Couples were matched on four variables (engaged vs.planning marriage, relationship satisfaction, confidence in getting married, andratings of communication impact) and assigned to either the intervention orcontrol condition. Twenty­five couples comprised the intervention condition, 47couples comprised the control group, and 42 couples made up a second controlcondition­­those who were offered the intervention but declined to participate.PREP proved to be effective in preventing relationship dissolution prior tomarriage: 4% of the intervention couples had ended their relationships prior tomarriage, compared to 25.5% of the control couples and 26.2% of the declinecouples. Although a similar pattern emerged for divorce/marital separation (8.3%of the intervention couples, 16.1% of the control couples, and 8% of the declinecouples had divorced or separated following their marriage), the groupdifferences were not statistically significant. PREP was found to enhancerelationship satisfaction among husbands, increase couples’ use ofcommunication skills, and reduce occurrences of physical marital violence.The work of Markman and colleagues suggests that prevention efforts can indeed have long­term beneficial effects for couples. However, two points shouldbe noted when interpreting their findings. First, only 25 couples received theintervention in this evaluation effort, and second, many of the couples originallyoffered treatment elected not to participate in the intervention. These factorsmake it difficult to draw broad conclusions about the effectiveness of PREPacross different types of couples. Although there are similarities between PREPand other prevention programs (e.g., relationship enhancement therapy), PREP’seffectiveness does not imply that these other prevention programs are similarlyeffective. As previously noted, great diversity exists among prevention programsin this area, and we have not yet empirically discerned the important ingredientsof an effective prevention program. It is also interesting to note that the long­termfollow­up results of this study were much different from the results immediatelyposttreatment. These findings highlight the need for long­term evaluation effortsin prevention research. To our knowledge, the work of Markman and colleaguesrepresents the only empirically sound, long­term evaluation of a preventionprogram for marital conflict and relationship dissolution. Additional preventionstudies that develop and examine the effects of other promising programs shouldbe a high priority for future research.PREDICTORS OF SUCCESSFUL OUTCOMES FROM MARITAL THERAPYAs noted earlier, not all couples respond equally well to marital therapy. Jacobsonand Addis (1993) note that couples who respond better to marital therapy do nothave premature closure in their problem­solving attempts and are le…