COMMON FACTORS AND OUR SACRED MODELS
Sprenkle, Douglas H
In this article we argue that much of what makes one treatment effective is common to other forms of effective treatment-both in psychotherapy generally and in marital and family therapy (MFT) specifically. Yet MFT has largely ignored the research on common factors. In this article we present a moderate view of common factors that, while repudiating the extreme position that there is no difference among treatment models, stresses that there are common factors and mechanisms of change that undergird most forms of successful treatment. These common mechanisms of change should be given more attention in our field, which has tended to emphasize the uniqueness of our sacred models. We delineate some of the major common factors, review the empirical evidence for them, and discuss implications of adapting a common factors informed approach to family therapy.
In the conclusions and implications chapter of their groundbreaking book length meta-analysis of the psychotherapy research literature (N = 475 studies), Smith, Glass, and Miller (1980) made the following rather shocking statement:
We did not expect that the demonstrable benefits of quite different types of psychotherapy would be so little different. It is the most startling and intriguing finding we came across. All the psychotherapy researchers should be prompted to ask how it can be so. If it is truly so that major differences in technique, count for so little in terms of benefits, then what is to be made of volumes devoted to the careful drawing of distinctions among styles of psychotherapy? And what is to be made of the deep divisions and animosities among different psychotherapy schools? (p. 185)
In the most recent and, arguably, the most careful and comprehensive meta-analysis of the current psychotherapy literature, Wampold (2001) drew a similar conclusion, “As has been shown throughout this book the evidence is overwhelmingly unsupportive . . . of the specificity of unique ingredients of any therapy” (p. 210).
Interestingly, both the first and the most recent meta-analyses argued forcefully that psychotherapy is effective. What their data throw into serious question, however, is whether different models of psychotherapy are differentially effective relative to each other, especially when confounding variables are controlled. As will be noted below, the major meta-analyses of MFT outcome research literature reach the same conclusion-few meaningful differences in efficacy among models (Shadish & Baldwin, 2002; Shadish, Ragsdale, Glaser, & Montgomery, 1995).
There are probably three ways to make sense of these findings. First, perhaps the various models achieve similar efficacy but through different change mechanisms (Lambert & Ogles, 2004). We cannot rule this out given that, although there is considerable evidence that psychotherapies work, we have only modest evidence regarding why they work (Johnson, 2002; Pinsof & Wynne, 2000; Sprenkle, 2002). However, as we believe we will demonstrate, there is reasonable evidence to conclude that many of the change mechanisms in the various models are overlapping (Henggeler & Sheidow, 2002; McFarlane, Dixon, Lukens, & Lucksted, 2002).
Second, perhaps there are meaningful differences in outcome among models, but they have not been teased out-either because the right research questions are not being asked, or because research methods are not sufficiently fine grained to make these distinctions (Sprenkle, 2002). We believe that there is some truth to this assertion, and we are fairly confident that there are some meaningful differences in outcome for some treatments, for some clients, by some therapists, under certain circumstances-to paraphrase Paul’s famous query (Paul, 1967). However, we also believe that these differences are unlikely to be strong enough for any model to claim overall superiority and that these differences are highly unlikely to be as powerful as the ways in which effective psychotherapies are similar.
The third explanation, which we believe is most likely, is that common mechanisms of change, which cut across all effective psychotherapy approaches, explain the finding of minimal differences between models. These common mechanisms account for much more of the outcome variance among effective psychotherapies than unique aspects of treatment models (Hubble, Duncan, & Miller, 1999; Wampold, 2001).
As a result, we argue for a different emphasis in how we think about what causes change in family therapy. Throughout most of its history, family therapists have tended to explain change through the lenses of sacred models (Sprenkle, Blow, & Dickey, 1999). Even with the recent shift toward evidence-based practices, we argue that the field is still too model focused when it comes to understanding the why of therapeutic change.
Our argument is largely based on a long tradition within psychotherapy-the common factors approach (Frank, 1976; Garfield, 1992; Hubble et al., 1999; Lambert, 1992; Luborsky, Singer, & Luborsky, 1975; Rosenzweig, 1936). Within this tradition, common factors (sometimes also called general factors) are those variables that contribute to change in psychotherapy that are not the province of any specific theoretical approach or model. Common factors have been both narrowly and broadly defined. The narrow view (Lambert, 1992) describes them only in terms of nonspecific aspects of treatment models (e.g., creating changes in meaning, which is found in many models under different names, such as refraining or externalizing problems). The broad conceptualization (Hubble et al., 1999; Tallman & Bohart, 1999) suggests that common factors include other dimensions of the treatment setting-including client, therapist, relationship, and expectancy variables. This latter understanding will be the approach taken in this article. Common factors can be contrasted with specific factors-those variables that contribute to therapeutic change that are unique to a particular theoretical approach or model (Waltz, Addis, Koerner, & Jacobson, 1993; Wampold, 2001).
We argue in this article that the conclusion regarding psychotherapy in general reached by Lambert and Ogles (2004), in the definitive fifth edition of the Handbook of Psychotherapy and Behavior Change, also applies to family therapy, “It is clear that much of what makes one treatment effective is common to other forms of treatment” (p. 179). We first describe how our moderate approach differs from some common interpretations (or misinterpretations) of the common factors lens of psychotherapy. Second, we discuss why family therapy has been slow to recognize the importance of common factors. Third, we provide a brief conceptual history of the common factors movement. Fourth, we describe some of the empirical controversies that have a bearing on the common factors challenge, and review results of key studies and meta-analyses. Fifth, we describe, more explicitly, what some of the major common factors are both in psychotherapy in general, and in MFT in particular. Finally, we review what the role of our sacred models is in a common factors context and discuss implications of the adoption of the common factors conceptualization for the MFT field.
A MODERATE COMMON FACTORS APPROACH
Our moderate approach is different from more extreme interpretations of the common factors position that we believe are unfortunate or misleading. It also represents the evolution of our own thinking about common factors, and this article represents more accurately our current thinking than our previous work (Blow & Sprenkle, 2001; Sprenkle et al., 1999). We do not support the literal belief that just any approach is as good as another. We think it unfortunate that the “dodo bird verdict” catchphrase-taken from the bird’s conclusion at the end of a race in Alice in Wonderland that “Everybody has won and all must have prizes” (Luborsky et al., 1975; Rosenzweig, 1936)-has been used by many common factors proponents, ourselves included (Sprenkle et al., 1999) to describe the finding of no or small differences among treatment outcomes across models (Duncan & Miller, 2000). Whether intentional or not, the dodo bird verdict connotes that it does not matter what one does in therapy. If taken at face value, the dodo bird verdict puts an impressive, empirically validated model like emotion focused therapy (EFT; Johnson, 1996, 2003) on the same level as tarot cards and Ouija boards. In contrast, our moderate common factors approach argues that among effective psychotherapies there are relatively small overall differences in treatment outcome, particularly when key confounding variables are controlled.
An unfortunate closely related misunderstanding of the common factors movement is that it, by definition, disparages treatment models. Our argument is not with models per se, but rather with assumptions regarding why they work. We hypothesize that models work largely (but not exclusively) because they are the vehicles through which common factors operate. The models activate or potentiate common mechanisms of change.
Yet another typical interpretation of common factors is that if models mean little, the therapeutic relationship means everything. Some scholars interpret the common factors approach to mean that the therapeutic relationship is the healer-it is both necessary and sufficient (see Patterson, 1984). Although we do believe that the therapeutic relationship is a highly significant factor in treatment outcome, our approach argues that there is a much larger array of common mechanisms of change that should be studied as well.
Some proponents of common factors, whose opinions we otherwise respect (Duncan & Miller, 2000; Wampold, 2001) take a dimmer view of clinical trials research than we do. Although we agree that this research has its limitations, and typically ignores common factors, we also believe that there is nothing inherent in this methodology that would prevent common factors from being researched. Our moderate common factors position takes a less adversarial stance toward traditional psychotherapy research methodology (although we will note some significant concerns below).
Finally, we do not believe that that common factors versus specific factors has to be couched as a rigid either/or. As noted above, there are probably types of problems, clients, circumstances, and therapists for which a particular model is especially well suited. A moderate common factors position allows for some added benefit from specific factors but argues that the common factors have not been accorded their rightful place within family therapy.
RESISTANCE TO COMMON FACTORS IN FAMILY THERAPY
With few exceptions-especially the work of Duncan and his colleagues (Duncan & Miller, 2000; Hubble et al., 1999; Miller, Duncan, & Hubble, 1997) family therapy has paid little attention to common factors. We agree with Wampler’s (1997) assertion, “Outcome research in marriage and family therapy has largely ignored the research literature on common factors underlying effective psychotherapy” (p. 10).
There have been several forces at work to emphasize the distinctiveness, rather than commonalities, within family therapy models, especially in the first three decades of family therapy’s 60-year history. Family therapy began as a maverick discipline, which was oppositionally defiant to the prevailing psychotherapeutic zeitgeist. There was pressure to accentuate differences from mainstream psychotherapy (Sprenkle et al, 1999). There was also competitive pressure among the various theoretical camps to accentuate their differences and alleged superiority vis-a-vis each other. This emphasis on distinctiveness was made easier because the field was not particularly influenced by research and, for most of the early decades, the growth of the field depended more on its intuitive or emotional appeal, rather than on solid research evidence (Nichols & Schwartz, 2001). In addition, the field has historically sought out solutions to difficult cases. This may have contributed to the belief that the unique models and methods that were devised to treat these cases were crucial to the successful outcome. We speculate that, because the field has welcomed innovation, it may attract people who have a greater than average need to believe that what they are doing is unique and distinctive. In addition, given the fact that many family therapists have expended great energy mastering their preferred model, and are often quite emotionally invested in it, there may be too much cognitive dissonance to admit that it may not be superior after all. Finally, a common factors approach is unlikely to generate as much excitement as the latest approach by a charismatic/confident presenter on the workshop circuit. As Frank (1976) put it, “Little glory derives from showing that the particular method one has mastered with such effort may be indistinguishable from other methods in its effects” (p. 47).
Even as family therapy has become more evidence based (Sprenkle, 2002), a movement that we strongly applaud, family therapy research has followed the lead of general psychotherapy research in largely emphasizing outcome research of treatment models at the cost of ignoring other variables that impact the effectiveness of psychotherapy (Pinsof & Wynne, 2000). As Beutler, Malik, and Alimohamed (2004) concluded in their research review in the Handbook of Psychotherapy and Behavior Change:
The strongest impression with which we are left at the conclusion of this review is that over the last two decades there has been a precipitous decline of interest in research in areas that are not associated with specific effects of treatment and its implementation, (p. 289)
Beutler et al. (2004) suggest that this current research emphasis may be driven more “by funding patterns and political agenda than by true promise for improving psychotherapy” (p. 291).
A BRIEF CONCEPTUAL HISTORY OF THE COMMON FACTORS MOVEMENT
The common factors movement goes back almost 70 years. Only some of the highlights will be noted here. For more details, the reader is referred to sources such as Hubble et al. (1999), Norcross and Goldfried (1992), and Wampold (2001). Saul Rosenzweig is typically credited with the first mention of the concept that the effectiveness of different therapies has more to do with common elements than the more specific tenants on which they are supposedly based. With regard to treatment models he noted that “besides the intentionally utilized methods and their consciously held theoretical foundations, there are inevitably certain unrecognized factors in any therapeutic situation-factors that may be even more important than those being purposely employed” (Rosenzweig, 1936, p. 412).
In the first edition of Persuasion and Healing Jerome Frank (1961) began to describe ways in which psychotherapy shared commonalities with other forms of healing. By the third edition, he and his daughter (Frank & Frank, 1991) argued that all psychotherapies shared four basic components: (a) An emotionally charged confiding relationship with a helping person; (b) a setting that is judged to be therapeutic, in which the client believes the professional can be trusted to provide help on his or her behalf; (c) a therapist who offers a credible rationale or plausible theoretical scheme for understanding the patient’s symptoms; and (d) a therapist who offers a credible ritual or procedure for addressing the symptoms. In a contemporary, empirically supported defense of common factors, Wampold (2001) uses these four dimensions as a basis for what he calls a contextual model of psychotherapy, which he juxtaposes with the medical model, and in so doing convincingly argues that it is these four components of therapy that can explain a great deal of outcome variance in therapy.
Frank and Frank (1991) also discussed six elements that were common to the rituals and procedures just mentioned: (a) The therapist combats the client’s demoralization and alienation by establishing a strong relationship; (b) the therapist links hope for improvement to the process of therapy, which heightens the patient’s expectation; (c) the therapist offers new learning experiences; (d) the client’s emotions are aroused and reprocessed; (e) the therapist facilitates a sense of mastery or self-efficacy; and (f) the therapist offers opportunities for the client to practice new behaviors. Far from disparaging models, Frank believed that specific models and associated techniques were necessary both to provide a coherent structure for treatment that therapists could believe in and also to provide an experience that seemed credible to clients.
Weinberger (1995) has noted that Frank’s work stood virtually alone from 1961 until about 1980 when an outpouring of works, views, and lists about common factors began to appear (Hubble et al., 1999). Although Rosenzweig (1936) first used the term, Luborsky et al. (1975) are often given the credit for first applying the dodo bird verdict to the finding of minimal differences among models. In so doing, they began the debate related to the comparative efficacy of models of psychotherapy.
The movement toward psychotherapy integration also spurred common factors. The publication of the Handbook of Psychotherapy Integration (Norcross & Goldfried, 1992) was a signal that contemporary psychotherapists had widely come to look across models to find the keys to efficacy (a point also stressed more recently by Lebow, 1997, with regard to the integration movement within MFT).
In this same volume, Michael Lambert (1992) wrote a highly influential paper in which he proposed a four-factor model of change based on common elements of change among theories. He based his model on his review of empirical studies of outcome research and estimated the percentage of variance in outcome contributed by each factor. Unfortunately, although these percentages were just educated estimates, they can easily be misinterpreted as mathematically derived; and they are frequently and unfortunately cited in the literature as factual statements of the percentage of the variance accounted for by the four factors (Hubble etal., 1999).
Family therapists Duncan, Miller, and colleagues (Hubble et al., 1999; Miller et al., 1997) have modified this model and used it in their work. In Lambert’s original model, common factors were a separate component. Miller et al. (1997) included all of the factors under the umbrella of common factors (Blow & Sprenkle, 2001). Their modified four factors, which will be explained below, were: (a) Client and extratherapeutic factors (said to account for 40% of change), (b) relationship factors (30%), (c) model/technique factors (15%), and (d) placebo, hope, and expectancy factors (15%). Wampler (1997) and Sprenkle et al. (1999) were the first scholars to write specifically about common factors that might be unique to MFT. These will also be addressed in more detail below.
SOME EMPIRICAL CONTROVERSIES SURROUNDING COMMON FACTORS
Opponents of common factors can point to hundreds of individual studies that show differences among treatment approaches. The recent review of controlled research on MFT published by the American Association for Marriage and Family Therapy (AAMFT) highlights many of these investigations (Sprenkle, 2002). On the surface, it might appear that this research contradicts common factors. There are, however, compelling reasons not to reject blithely the common factors argument on the basis of this evidence.
First, a fairly high percentage of psychotherapy outcome studies (especially in family therapy) are carried out by the founders of models and their students or colleagues, leading to high allegiance effects in the results (Wampold, 2001). Graduate students of the first author were able to find approximately 45 original empirical articles related to three of the most widely researched models in family therapy-emotionally focused therapy (EFT), multisystemic therapy (MST), and functional family therapy (FFT). Of the currently published papers, fewer than 10% appear to have been done by independent investigators. Meta-analysis has demonstrated that investigator allegiance is strongly associated with positive outcome (Shadish & Baldwin, 2002; Wampold, 2001). For whatever reasons (e.g., subtle pressures to show that their favored models work better), when investigators have allegiance to models they are investigating, metaanalysis shows that effect sizes are consistently larger than if the investigators have no such allegiance. Until comparable results can be shown by investigators with less allegiance one cannot have the same confidence that the model itself is producing the result.
Second, and closely related to the first point, investigators (especially those with an allegiance) frequently compare their preferred model to some treatment as usual condition or to other control conditions that are not equally valued (Wampold, 2001). Even if the alternatives are manualized and carried out well, they are frequently done by people who are less enthusiastic or whose context does not support them to the same extent as the therapists in the preferred treatment. These subtle influences occur even when investigators are highly ethical and seek to be objective (Heppner, Kivlighan, & Wampold, 1999; Wampold, 2001).
The reader will remember that our moderate common factors approach only states that among effective psychotherapies there are only modest overall outcome differences. We would expect differences when less valued treatments of unknown effectiveness are compared to strong treatment packages. We could not find a single study in family therapy, however, by investigators with no particular agenda to discredit another model, that was a head-to-head comparison of treatments previously demonstrated to be effective, in which the studies were carried out by neutral although enthusiastic experts in supportive contexts-that is in which all treatments were equally valued. We hasten to add that such studies are difficult and expensive to carry out. Interestingly, a few such model studies have been carried out in non-MFT psychotherapy research, and the results are highly supportive of a common factors interpretation. We will now describe two strong examples of high quality, relatively unbiased psychotherapy research.
Sloane, Staples, Cristol, Yorkston, and Whipple (1975) completed an exemplary study comparing short-term, analytically oriented psychotherapy and behavioral therapy (and a minimal intervention control group). Lambert and Bergin (1994) asserted that this study “established a standard and methodological sophistication that few subsequent studies have surpassed” (p. 158). Wolpe (1975) referred to this investigation as “unmatched-in the history of psychotherapy” (p. xix). In addition to the typical methodological strengths that characterize most strong research designs, the therapists were very experienced and respected proponents of the respective approaches, and evaluations were done by an independent assessor. Both long-and short-term results demonstrated minimal differences among the outcomes of the active treatment groups.
Perhaps even more telling are the results of the famous National Institute of Mental Health (NIMH) Collaborative Depression Study (Elkin et al., 1989; Imber et al., 1990). This was arguably the most carefully done and most expensive study in the history of psychotherapy. The investigators randomly assigned 250 patients to four different treatment groups in three cities. An antidepressant plus clinical management were contrasted (along with a drug placebo and clinical management) with cognitive behavioral therapy (CBT) and interpersonal psychotherapy (IPT). The last two approaches had been shown, independently, to be effective with depression; therefore, this project offered a head-to-head opportunity to look at the differential effectiveness of two previously established approaches. Therapists were carefully trained and monitored and unusual efforts were made to ensure treatment fidelity and fairness.
Although the results of this investigation are very complex, and space prevents presenting the details, basically few differences were found between IPT and CBT on any of the major outcome variables-including those that might reasonably be assumed to favor one approach versus another. However, differences in outcome were strongly associated with several important variables unrelated to the models that were investigated (Blatt, Sanislow, Zuroff, & Pilkonis, 1996).
Because one should not overgeneralize from the results of two studies (however well done and unbiased), any more than the detractors of common factors can generalize from isolated studies believed to contradict the perspective, we turn now to the results of meta-analysis. The advantage of meta-analysis is that it enables scholars to draw conclusions based on aggregate data from a large number of studies (Shadish & Baldwin, 2002).
Furthermore, in meta-analysis, by using regression techniques, it is possible to determine to what extent certain variables contribute to the effect size (the typical standardized measure of outcome of treatments; Shadish & Baldwin, 2002; Wampold, 2001). This enables investigators to look at variables that may confound results, such as the aforementioned allegiance of the investigator. Another confounding variable that can be controlled in meta-analysis is the use of reactive measures. Some therapies may look better simply because they use measures that are more reactive to change. That is, perhaps the use of a particular measure rather than the intervention itself leads to a greater effect size. Indeed, the first author and colleagues discovered doing a meta-analysis of all enrichment program related outcome research (Giblin, Sprenkle, & Sheehan, 1985) that using a behavioral measure rather than a self-report measure contributed more to an enrichment model getting a high effect size than the intervention method itself. This confound is important to examine because there are few head-to-head comparisons in family therapy research of models using the same instruments. One approach claiming high effect sizes (relative to a competing approach) may owe this superiority mostly to the choice of measures used to evaluate the models (Shadish & Baldwin, 2002). Another common confounding factor in general psychotherapy research is the common use of analog designs or milder cases with certain treatments (Lambert & Ogles, 2004).
Some major meta-analyses have found treatment differences. For example, Shapiro and Shapiro (1982) found greater effect sizes among cognitive and behavioral treatments. However, they attributed these results to the wide use of analog studies and mild cases, along with highly reactive measures in the cognitive and behavioral literature. Although there are some minor differences, we believe the major meta-analyses have produced results generally consistent with a common factors explanation (Lambert & Ogles, 2004).
The aforementioned meta-analysis by Wampold (2001) is important because he only included studies that directly compared two or more bona fide treatments, thus eliminating the confounding clement of comparing treatments that were carried out in different studies. Wampold (2001) presents evidence that at most 8% of the outcome variance in psychotherapy is due to the unique contributions of models-even including popular models like CBT (the reader will note that this figure is considerably less than the Lambert, 1992, estimate of 15%). General factors shared by all successful psychotherapies accounted for about 70% of outcome variance and about 22% of the variance was unexplained (Wampold, 2001). Wampold also reviewed the results of component analyses, in which key components of treatments were added and subtracted, and found that these changes had little impact on the core effects of treatments.
The lack of differential effectiveness has received strong empirical support in the MFT literature in the major meta-analyses reported by William Shadish and his colleagues (Shadish & Baldwin, 2002; Shadish et al., 1995). These data support the statement that, although MFT approaches offer clear evidence of effectiveness, there is little evidence that they are differentially effective relative to each other. Shadish et al. put it this way, “Despite some superficial evidence apparently favoring some orientations over others, no orientation is yet demonstrably superior to any other. This finding parallels the psychotherapy literature generally” (p. 348). These investigators further stated that whatever modest theoretical orientation differences do show up seem to be the result of being confounded with other variables (such as allegiance, the use of reactive measures, etc.). When the authors entered these confounding variables into a regression equation, all orientation differences were wiped out. Shadish and Baldwin concluded, “There is little evidence for differential efficacy among the various approaches to marriage and family interventions, particularly if mediating and moderating variables are controlled” (p. 365).
We wish to conclude this section by reminding the reader that the finding of no meaningful differences does not rule out the possibility that there may be meaningful differences in the efficacy of models. As we know from elementary statistics, failure to reject the null hypothesis (of no difference) is not the same as proving the null hypothesis. Meta-analysis is no panacea if the original studies are not asking the right questions or using the best methods. As previously noted, our moderate common factors approach recognizes that some approaches may be more helpful for certain types of clients, with certain types of problems, by certain therapists, under certain sets of circumstances. There is some evidence, for example, within the general field of psychotherapy, that even though there are few data to support the superiority of any model, for the general moderate-difficult outpatient psychotherapy population, behavioral (Emmelkamp, 2004) and cognitive methods (Hollon & Beck, 2004) appear to add value for the treatment of a number of difficult problems like panic, phobias, and compulsions (Lambert & Ogles, 2004). Thus, a caveat to the common factors approach may be that for some highly specific difficult problems specific factors may make enough of a difference to permit the claim of differential efficacy.
THE COMPONENTS OF A COMMON FACTORS APPROACH
As previously noted, when conceived broadly, common factors are dimensions of the treatment setting that include client, therapist, relationship, expectancy, and treatment variables that are not specific to a particular model. Because therapy is inherently an interactional process, distinctions among these variables are somewhat artificial. It is hard to disentangle, for example, the effect of a treatment per se from the therapist’s belief in the treatment, or the client’s assessment of its credibility, or the client’s attitude toward the person who is delivering the treatment. It also appears that certain client variables like gender and race have little predictive value independent of the therapist’s response to these client characteristics (Clarkin & Levy, 2004). Beutler, Harwood, Alimohamed, and Malik (2002) have also made an impressive case for the interaction of client type and treatment. Clients who are more self-reflective, introspective and/or introverted are more likely to benefit from insight-oriented procedures. Clients who are more impulsive and aggressive do better with skill building and symptom-focused methods. Finally, some variables in the literature that are classified as client variables-for example, the client’s level of motivation in therapy-can be treated as a client-independent variable if motivation is seen as something a client brings to therapy. Alternatively, however, enhancing client motivation can also be conceptualized as a goal of treatment and therefore as a treatment related dependent variable (Prochaska, 1999). Functional family therapy, for example, focuses the first stage of treatment around enhancing engagement and motivation (Sexton & Alexander, 2003). With these caveats in mind, we will proceed to examine the basic components of a common factors lens regarding what brings about change in therapy.
The Client as a Common Factor in Psychotherapy
“If we look at factors contributing to the success of treatments, it is not the clinician or treatment procedure that is key, but the motivation, awareness, expectations, and preparation of the patient or client” (Beutler, Bongar, & Shurkin, 1998, p. 8). In our judgment, one of the most significant contributions of the common factors movement has been to highlight the truth that the client (not the model or even the therapist) is the ultimate hero (Duncan & Miller, 2000) of psychotherapy. It is so easy to focus on our models, techniques, and skills (what Tallman & Bohart, 1999, call “professional centrism”) that we sometimes forget that therapy only works to the degree that it activates the natural healing propensities of clients.
Tallman and Bohart (1999) offer a metaphor to make this point. Many people go to health clubs to achieve the goal of cardiovascular fitness (think, good psychotherapeutic effects). The health club offers them a variety of pieces of equipment to achieve this end-elliptical trainers, treadmills, stationary bicycles, and stair-stepper machines (think, therapy models). If the goal of cardiovascular fitness is successful, it matters considerably more that the health club clients motivate themselves to get out of bed, get to the club, use the machines, and stick with the program, than it matters which particular machine they choose to achieve the goal. However, this does not mean that the machines are not helping achieve cardiovascular fitness. It is important to remember that there are many pathways to cardiovascular health, just as there are many therapy models that can be used to achieve desired therapeutic effects.
Tallman and Bohart (1999) speculate that the reason that a number of models work equally well is because the client’s ability to use whatever is offered overwhelms whatever differences there are in the techniques or approaches. They argue further that even if the techniques were to have specific effects, clients would individualize them for their own purposes. This finding is borne out by Helmeke and Sprenkle’s (2000) research on pivotal moments in couples therapy. These investigators found that clients often reconstructed interventions for their own purposes so that what they thought was pivotal was different from the therapist’s perception of what was pivotal in the same session. Research on early responders-clients who respond very quickly to treatment-also supports the common factors argument in that change occurs before the specific factors of models have a chance to be employed (Haas, Hill, Lambert, & Morrell, 2002).
Michael Lambert first drew widespread attention to the role of the client in accounting for a substantial portion of the outcome variance in psychotherapy (Lambert, 1992). As previously noted, he used the more narrow definition of common factors, restricting the term to common elements among treatment models. Nonetheless, he believed that client factors operated independently of treatment models. Hubble et al. (1999) see the contributions of the clients to be an essential component of change and as a result, they modified Lambert’s model to include client variables as a common factor.
Lambert (1992) called this major source of the variance in psychotherapy client and extra-therapeutic factors. Client factors are characteristics or qualities of the client (such as level of motivation and commitment to change, inner strength, and religious faith) and extra-therapeutic factors are ingredients in life and environment of the client that impact change (such as social support, community involvement, and stressful events). Lambert (1992) stated that this factor accounts for about 40% of psychotherapy outcome, although, as previously noted, these figures have unfortunately been reified to be more objective than they are. Whatever the exact percentage, several scholars (Beutler et al., 1998; Tallman & Bohart, 1999) share the viewpoint of Miller et al. (1997) that “the research literature makes it clear that the client is actually the single, most potent contributor to outcome in psychotherapy” (pp. 25-26).
In an earlier publication that described client-related factors in MFT (Sprenkle et al., 1999), we indicated that there has been some MFT research related to static characteristics of individuals, such as age, gender, race, and sexual orientation, and less research on nonstatic characteristics of individuals, couples, and families, such as individual learning style, level of couple commitment, family cohesion, and family expressed emotion, that are potentially malleable in therapy but which probably bias outcome from the onset of intervention. There has been little research, however, on motivational characteristics that relate directly to the clients’ engagement in treatment like perseverance, willingness to participate in treatment, and cooperation with homework assignments. The study by Holtzworth-Munroe, Jacobson, DeKlyen, and Whisman (1989) is one of few studies on these themes. This is unfortunate, because we think these client variables are likely to be strongly related to outcome. Although the research of Prochaska (1999) on client motivation/readiness for change is well known in the individual psychotherapy literature, we know of no published MFT applications of Prochaska’s model.
Marriage and family therapy seems to have followed the lead of psychotherapy research generally, which has tended to focus on the client’s diagnosis (typically homogenous for a DSM IV category) “while ignoring the idiosyncratic aspects of the client that are even more salient in predicting change and guiding treatment decisions” (Clarkin & Levy, 2004, p. 195). Although random clinical trials do not inherently produce this bias, they typically consider client individual differences as a source of error or noise, rather than a fertile ground for discovery. There are, however, some examples where family therapy model developers have presented evidence regarding the types of clients for whom their model may be most appropriate and effective, for example, Johnson and Talitman (1997) for EFT and Jacobson and Christensen (1996) for behavioral marital therapy (BMT). We applaud this research.
Therapist Effects as a Common Factor
“Good therapists are good therapists, from whatever theoretical province they derive their skills. The power of these particular therapists is greater than any therapeutic contribution that may stem from any of their theories” (Beutler et al., 1998, p. 117). On an intuitive basis, it may be obvious that therapist effects (therapist characteristics and skills that contribute to outcome that are not the province of a particular theory) should loom large as factors in successful psychotherapy. For example, we believe that most of us, when we refer a close friend or relative to a therapist, consider the personal qualities and virtues of that therapist more than their theoretical allegiance. We also intuitively know that certain therapists seem to be unusually effective, whereas others (even if they are proponents of the same model) seem to have a hard time holding on to clients and are probably not effective even with a majority of them.
Surprisingly, other than research related to the therapist’s contribution to the therapeutic relationship or alliance, there is very little research directed toward identifying those characteristics that differentiate more competent from less competent therapists (Wampold, 2001). There is some evidence that better therapists (aside from creating a strong alliance) offer a level of activity that is consistent with the client’s expectations and preferences, are creative in perceiving new ways of approaching problems, suggest credible new ways of learning adaptive skills, and are personally well integrated (Beutler et al., 1998, 2004). In the research on competence, high competence often simply gets operationalized as getting better outcomes (Wampold, 2001).
Perhaps more surprisingly, there has been less attention to therapist factors as the empirically validated treatment movement has gained ascendancy. As Beutler et al. (2004) put it recently, “Over the past two decades the emphasis on randomized clinical trials and specific therapy models has resulted in decreasing attention given to discrete therapist factors” (p. 227). In most clinical trials, efforts have been made to control therapist factors through the use of treatment manuals and adherence checks. This is even true of research in more real world clinical settings (Wampold, 2001). Designs ignore the idiosyncrasies of unique contributions of therapists and view treatment as though it is separate from the persons who deliver it. Therapist effects are treated as sources of error rather than sources of variance.
However, there is considerable evidence that variability among therapists, even in manualized projects, continues to be the rule rather than the exception. As Beutler et al. (2004) explained, “Unfortunately, standardizing the treatment has not eliminated the influence of the individual therapist on outcomes” (p. 245). Earlier, we referred to the extremely careful and expensive NIMH Collaborative Depression Study. Although there were minimal differences among treatment models, there were serious differences among the therapists in their abilities to implement the models, even after careful screening of the therapists and eliminating some who were not effective. When Blatt et al. (1996) divided therapists into less effective, moderately effective, and more effective based on composite outcome scores for their clients, they concluded that their analyses “indicate that significant differences exist in the therapeutic efficacy among therapists, even with the experienced and well-trained therapists in the [NIMH study]” (p. 1281). They furthermore said that these differences were independent of treatment model used, the research site, or even the therapists’ general level of experience.
In his comprehensive meta-analysis of the psychotherapy research literature, Wampold (2001) devotes an entire chapter to therapist effects. He offers strong statistical evidence that therapists often contribute much more to outcome than the particular therapy they happen to use. The first author once participated in a study in which the order of effectiveness of three treatment approaches were reversed depending on whether the data from one therapist were included or excluded.
The Therapeutic Relationship as a Common Factor
Although early research placed primary emphasis on therapist characteristics (such as accurate empathy, nonpossessive warmth, etc.) that contribute to the therapeutic relationship, attention has shifted to the therapeutic alliance, which, by definition, is the joint product of the therapist and client together focusing on the work of therapy. This focus includes the client perceptions of the alliance and is important because therapist and client ratings of the relationship are often different and client ratings are clearly superior in predicting outcome (Bachelor & Horvath, 1999; Orlinsky, Grawe, & Parks, 1994).
Bordin (1979) suggested that the alliance is composed of three elements: Bonds (the affective quality of the client-therapist relationship that includes dimensions such as trust, caring, and involvement); tasks (the extent to which the client and therapists are both comfortable with the major activities in therapy and the client finds them credible; and goals (the extent to which the client and therapist are working toward compatible goals). Research on the alliance includes data on these three variables. They are also reflected in the family and couple therapy alliance scales (Pinsof & Catherall, 1986).
Lambert (1992) estimated that the therapeutic relationship accounted for about 30% of the outcome in psychotherapy, second only to the client related variables. Although other scholars (Clarkin & Levy, 2004) believe that the alliance accounts for less of the variance than Lambert (1992), almost all psychotherapy researchers believe that it is a potent variable (Bachelor & Horvath, 1999). It also strengthens the argument for common factors, given that early alliance ratings can predict outcome before specific therapy procedures are applied (Lambert & Ogles, 2004; Martin, Garske, & Davis, 2000).
The therapeutic relationship is the common factor most studied in MFT research. In the largest single MFT outcome study ever completed (3956 cases), Beck and Jones (1973) used regression techniques to analyze 11 factors that contributed to positive therapy outcome. What they wrote set the pace for what would be said in MFT research reports for the next 30 years:
Of the findings with specific implications for practice, probably the most important in the present study is that of the marked association of the counselor-client relationship with outcomes. . . .This factor was found to be twice as powerful a predictor of outcomes as any other client or service characteristic covered by the study and more powerful than all client characteristics combined. An unsatisfactory relationship was found to be highly associated with client-initiated disengagement and with negative explanations by the client of his reason for terminating, (p. 8)
Sprenkle et al. (1999) offer a number of examples from the MFT literature that describe the potency of the client-therapist relationship (Green & Herget, 1991; Kuehl, Newfield, & Joanning, 1990; Stolk & Perlesz, 1990). Johnson and Talitman (1997) researched predictors of success and EFT, and found that alliance scores accounted for 22% of the variance in post treatment satisfaction and 29% of the variance at follow-up. Additional analyses suggested that the task subscale of the Family and Couple Therapy Alliance Scale (Pinsof & Catherall, 1986) accounted for most of this variance. That is, in this study, whether the activities of therapy seemed credible and meaningful to the clients presenting concerns was the most important dimension of the alliance.
Although we do not believe that relationship skills are sufficient to explain family therapy outcome, we are pleased to note that they are a widely valued component of common factors. Among the MFT models with the strongest empirical support, like EFT (Johnson, 2003), FFT (Sexton & Alexander, 2003), and MST (Sheidow, Henggeler, & Schoenwald, 2003), building a strong therapeutic alliance is greatly emphasized.
Expectancy (Placebo) as a Common Factor
Lambert (1992) stated that expectancy and placebo factors were the portion of improvement that resulted from the client’s knowledge of being in treatment, becoming hopeful, and believing in the credibility of the treatment. The reader will remember that this is also a significant aspect of the Frank model (Frank & Frank, 1991). Lambert believed that expectancy accounted for about 15% of improvement in psychotherapy.
We concur with other scholars (Lambert & Ogles, 2004) that the term placebo is probably a misnomer in psychotherapy, given that all treatments, including the most rigorously empirically validated ones, rely on expectancy and the creation of hope. Howard, Moras, Brill, Martinovich, and Lutz (1996) wrote that moving from demoralization to re-moralization is the critical first stage of therapy, and although they employ different terms, re-moralization is clearly built into the first stage of empirically validated MFT models (e.g., Johnson, 2003; Sexton & Alexander, 2003).
There is very little research in MFT specifically on the hope aspect of expectancy (Blow & Sprenkle, 2001; Sprenkle et al., 1999). There is some evidence regarding the credibility aspect in that at least three studies (Crane, Griffin, & Hill, 1986; Kuehl et al., 1990; Johnson & Talitman, 1997) found that the therapist’s ability to present treatment in a way that is consistent and congruent with client expectations clearly contributes to successful outcome.
Nonspecific Treatment Variables as a Common Factor
There are a number of conceptualizations and lists of nonspecific treatment variables that are found in the literature (Grencavage & Norcross, 1990). Karasu (1986) offered one of the most influential and parsimonious. He argued that nonspecific treatment variables could be subsumed under three dimensions: Behavioral regulation, cognitive mastery, and affective experiencing. We will briefly discuss each of these as applied to MFT.
Behavioral regulation (changing the doing). This occurs when therapists facilitate clients’ changing interactional patterns or dysfunctional sequences, modifying boundaries and changing family structures, learning new skills, becoming more supportive of each other, and learning to empower self and others. Examples might include disrupting dysfunctional behavioral sequences within strategic therapy (Haley, 1987), marking boundaries within structural therapy (Minuchin, 1974), interrupting patterns of behavioral interaction within EFT (Johnson, 1996), teaching new communication skills within FFT (Sexton & Alexander, 2003), and facilitating a dialogue between the self and parts in the internal family systems (IFS) model (Schwartz, 1995).
Cognitive mastery (changing the viewing). This occurs when therapists facilitate clients gaining new perspectives (new meanings) about interactional processes within themselves and the family, between the family and other systems, and across generations (Sprenkle et al., 1999; Wampler, 1997). Therapists facilitate cognitive mastery by offering reframes, interpretations, explanations, or rationales. They also offer information. Examples include reframing the sequence in EFT in the language of attachment (Johnson, 1996), developing an empowering alternative story within narrative therapy (White & Epston, 1990), or exonerating parents within contextual family therapy by helping clients see that their parents did the best they could with what they had (Boszormenyi-Nagy, 1987).
Emotional experiencing (affective experiencing/regulation). This occurs when therapists facilitate clients regulating or experiencing emotions and making emotional connections with themselves, the therapist, and (most importantly) each other. Examples include getting couples to express the primary emotions underlying angry/defensive exchanges within EFT (Johnson, 1996), and accessing wounded parts within IFS (Schwartz, 1995).
We acknowledge that it is sometimes difficult to distinguish what is a predominantly behavioral, cognitive, or affective intervention. If, for example, the therapist gets a couple to relate to each other in a novel way, the behavioral shift could be experienced first as a new way of thinking about each other or in terms of the powerful felt sense of emotional closeness.
A developmental sequence of common factors. Before leaving the section, we note that Lambert and Ogles’ (2004, p. 173) present a typology of common factors in which they group common factors into support (e.g., trust, reassurance), learning (e.g., advice, insight), and action (e.g., taking risks, modeling) categories. They believe that these major categories represent a sequence that operates in many psychotherapies and that the sequence (whereby supportive functions generally precede changes in beliefs and attitudes, which precede therapists’ attempts to encourage new patient action) is often mediated by the common factors that they list.
Together they provide for a cooperative working endeavor in which the patient’s increased sense of trust, security, and safety; along with decreased tension, threat, and anxiety; lead to changes in conceptualizing his or her problems and ultimately in acting differently by refraining fears, taking risk, and working through problems and interpersonal relationships. (Lambert & Ogles, 2004, p. 173)
COMMON FACTORS UNIQUE TO MFT
We believe there are only three common factors that are truly unique to MFT. The other common factors mentioned above (like behavioral regulation), are only unique to MFT to the extent that they operate through these mechanisms (Sprenkle et al., 1999).
Relational conceptualization is the translation of human difficulties into relational terms. While not denying the role of biology or intrapsychic causation, most MFTs, for example, would conceptualize a depressed person’s malady within the context of his social network, being mindful of the complex web of reciprocal influences involved in the depression. Furthermore, this conceptualization leads MFTs to keep the whole system (or systems) in mind when interacting with a part of a system (Wampler, 1997). While focusing on the depressed person’s relationship with his employer, for instance, the therapist might also be paying attention to patterns or expectations from his family of origin. Moreover, MFTs typically attempt to relate in a positive way to all parts of the system(s) regardless of who happens to be in the treatment room (Wampler, 1997). So, if the depressed person’s spouse refused treatment, he or she might still be very much present in the session, as might the person’s colleagues at work. This conceptualization holds true for most MFT orientations and may be a common explanation for some of the efficacy of MFT (Sprenkle et al, 1999).
The Expanded Direct Treatment System
Many family therapists tend to push to involve more people than the identified patient (or in some cases the willing participant) directly in treatment. Pinsof (1995) calls this the direct patient system as opposed to the indirect patient system (persons not physically present in treatment, but who may affect the problem treated in important ways, and who may, in turn, be significantly affected by the therapy). Therapists who take pains to expand the direct system, even in the face of resistance, believe that the power at the heart of family treatment resides in affecting live systems. Both Napier (1997) and Minuchin (1998) have criticized newer social-constructionist therapies for not being truly family therapy, because they rely too much on changing inner meaning and fail to capitalize on the immediacy only made possible by in-the-room interaction. Regrettably, little research is available that sheds light on the relative efficacy of approaches that directly target family interaction versus those that merely rely on a systemic conceptualization of family problems. (See Sprenkle et al., 1999, for more details.)
The Expanded Therapeutic Alliance
When more than one person is involved in the direct treatment system, the expanded therapeutic alliance may be a common factor that is unique to MFT. The therapist forms an alliance not only with each member of the family individually, but also with certain subsystems, and with the family as a whole (Pinsof, 1995; Sprenkle et al., 1999). Unfortunately, there is little research to determine whether these multiple alliances are an asset or a liability. There is some evidence, for example, that a split alliance is deleterious and that having a balanced alliance with various family members is more important than the strength of the alliances (Hollander-Goldfein, 1989; Pinsof, 1995).
THE ROLE OF OUR SACRED MODELS AND IMPLICATIONS FOR THE FIELD
This article began with a quote that asked the question, “And what is to be made of the deep divisions and animosities among different psychotherapy schools?” (Smith et al., 1980, p. 185). If nothing else we hope that this article will contribute to the demise of whatever parochialism, triumphalism, and divisiveness remains in family therapy on the basis of our sacred models.
There is some encouraging evidence that model developers are beginning to recognize the commonalities among models. After describing the body of research about three major empirically validated models for treating conduct disorder and delinquency (FFT, MST, and the Oregon Treatment Foster Care, Chamberlain & Mihalic, 1998), Henggeler and Sheidow (2002) devoted a section of their chapter to similarities among models. They noted that, “Although each of the aforementioned evidenced-based models was developed independently, they share several commonalities in their conceptualization, delivery, and procedures” (p. 41). They noted that similarities in these programs will have heightened interest as the programs are moved more into real world community settings. In a chapter in the same volume (Sprenkle, 2002), McFarlane et al. (2002) described similarities among the major family approaches to the management of major mental illness. McFarlane et al. (2002) noted that at a meeting of The World Schizophrenia Fellowship model developers were challenged to develop a consensus regarding core ingredients that cut across approaches. The authors noted that “this process led to convergence of concept rather than the usual process of the field splitting into competing schools” (McFarlane et al., 2002, p. 257). These authors indicated that when these core ingredients are present, the disparate methods work about equally well.
We believe that therapists and trainees should be more modest about their cherished models. They should think less about their superiority even if they happen to be empirically validated. Although we strongly encourage empirical validation, it has not yet offered convincing evidence of differential results relative to other effective treatments. The reader is reminded what happened in the NIMH Collaborative Depression Study as well as Shadish’s meta-analyses of the MFT literature. It is also possible that certain models that have not been empirically validated, such as the IFS (Schwartz, 1995) may prove to be just as potent as their empirically validated counterparts. Although we strongly encourage models that have not been researched to demonstrate their efficacy, it strikes us as somewhat arrogant to assume automatically that empirically validated models are more effective or superior.
As we hope this article has made clear, the common factors that we highlighted are not the province of particular models. In fact, we find it highly significant that the newer, empirically validated models like EFT, FFT, and MST are themselves highly integrative. The first author’s preferred model, EFT, for example, is a creative amalgam of experiential/gestalt, client-centered, interactional/family systems, structural, social-constructionist, and feminist models (Johnson & Denton, 2002). These newer models apparently recognize that there are truths that “transcend scholastic boundaries” (Lebow, 1997, p. 3).
As we have stated repeatedly, it may turn out that certain models offer added value in certain circumstances. Again, to use EFT as an example, what seems to be unique about it is the use of emotion in breaking destructive cycles of interaction (Johnson & Denton, 2002. p. 223). The approach also focuses strongly on creating secure attachments and repairing attachment injuries. As previously mentioned, Johnson and Talitman (1997) offer some evidence that the approach works best for those who can readily conceptualize their problems within an attachment perspective and for whom EFT tasks seem especially credible. We applaud the efforts of model developers to look at mediators and moderators of effectiveness, and believe that this will be a much more productive avenue of research than examining main effects of treatments. Each model seems to be like a magnifying glass, which brings clarity and a special emphasis to a particular aspect of the process of change. Future research will hopefully determine the circumstances when and if this special focus adds unique value.
For this reason, we reiterate that common factors versus specific factors does not need to be couched as an either/or phenomenon. Although clinical trials research is likely to remain the dominant research modality in the foreseeable future (and this method tends to highlight treatment differences), we do not believe that there is anything inherent in this methodology that would prevent more examination of the common factors we have described. It is probably important, for example, for all clinical trials to look at therapist effects. We believe that more research attention to the variables described in this article might contribute greatly to our knowledge of effective psychotherapy.
Even when models do not offer unique benefits, we still believe that they are important. Therapy is much more than a pleasant conversation, and even developing a strong therapeutic alliance alone is not sufficient to bring about the same results as effective psychotherapies (Bachelor & Horvath, 1999). The models are important because they are the vehicles though which the common factors do their work. Many models offer strategies that are often highly effective in activating these common factors.
However, we think that a common factors lens will force us to think more about therapeutic effects than therapeutic strategies or techniques. Therapeutic effects are what take place in a therapeutic system that leads to a positive outcome. For example, if looking at one’s problems in new way (what we referred to above as changing the viewing) is a therapeutic effect, there are probably hundreds of ways to achieve the same effect-including refraining, positive connotation, prescribing of the symptom, externalizing the problem, rewriting one’s life story, and so on. Indeed, the emphasis on strategies and techniques as opposed to effects has probably lead to the proliferation of therapy models and the mistaken assumption that specific models, strategies, and techniques are primarily responsible for therapeutic change. Hopefully, if common factors are taken seriously, it will no longer be necessary to continue inventing models! As Sol Garfield (1987) once quipped, “I am inclined to predict that sometime in the next century there will be one form of psychotherapy for every adult in the Western world” (p. 98). We hope that therapists and model developers will reflect much more on what it is that unites them.
Finally, if a common factors approach were taken seriously, it would have major implications for the ways in which MFTs are trained. The current emphasis on learning models would be decreased and more attention would be paid to learning about the common factors and mechanisms of change. Although both of us were graduates of AAMFT-accredited programs, neither of us receive much information about client variables, therapist variables, the role of expectancy, and the like, and the significant ways in which common factors interact. Marriage and family therapy education would probably place more emphasis on core therapeutic skills, including empathic listening and responding, developing a working alliance, and engaging and motivating clients, before stressing particular approaches. We are currently articulating more thoroughly a common factors-informed agenda for training.
What we have said in this article runs counter to the culture of family therapy, which has and continues to emphasize its sacred models. We hope that this article has stimulated the reader to think in new ways about what really causes change in family therapy.
Bachelor, A., & Horvath, A. (1999). The therapeutic relationship. In S. D. Miller (Ed.), The heart and soul of change: What works in therapy (pp. 133-178). Washington, DC: American Psychological Association.
Beck, D. F., & Jones, M. A. (1973). Progress on family problems: A nationwide study of clients’ and counselors’ views on family agency services. New York: Family Service Association of America.
Beutler, L. E., Bongar, B., & Shurkin, J. N. (1998). A consumer’s guide to psychotherapy. New York: Oxford University Press.
Beutler, L. E., Harwood, T. M., Alimohamed, S., & Malik, M. (2002). Functional impairment and coping style. In J. Norcross (Ed.), Psychotherapy relationships that work: Therapist contributions and responsiveness to patient needs (pp. 145-170). New York: Oxford University Press.
Beutler, L. E., Malik, M. L., & Alimohamed, S. (2004). Therapist variables. In M. J. Lambert (Ed.), Bergin and Garfield’s handbook of psychotherapy and behavior change (5th ed., pp. 227-306). New York: Wiley.
Blatt, S. J., Sanislow, C. A., Zuroff, D. C., & Pilkonis, P. A. (1996). Characteristics of effective therapists: Further analyses of data from the National Institute of Mental Health treatment of depression collaborative research program. Journal of Consulting and Clinical Psychology, 64, 1276-1284.
Blow, A. J., & Sprenkle, D. H. (2001). Common factors across theories of marriage and family therapy: A modified Delphi study. Journal of Marital and Family Therapy, 27, 385-402.
Bordin, E. S. (1979). The generalizabilily of the psychoanalytic concept of the working alliance. Psychotherapy: Theory, Research, and Practice, 16, 252-260.
Boszormenyi-Nagy, I. (1987). Foundations of contextual therapy: Collected papers of Ivan Boszormenyi-Nagy. New York: Brunner/Mazel.
Chamberlain, P., & Mihalic, S. (1998). Blueprints for violence prevention, hook eight: Multidimensional treatment foster care. Boulder, CO: Center for the Study and Prevention of Violence.
Clarkin, J. R, & Levy, K. N. (2004). Research on client variables in psychotherapy. In M. J. Lambert (Ed.), Bergin and Garfield’s handbook of psychotherapy and behavior change (5th ed., pp. 194-226). New York: Wiley.
Crane, D. R., Griffin, W., & Hill, R. D. (1986). The influence of therapist skills on client perception of marriage and family therapy outcome: Implications for supervision. Journal of Marital and Family Therapy, 12, 91-96.
Duncan, B. L., & Miller, S. D. (2000). The heroic client: Doing client directed, outcome-informed therapy. San Francisco: Jossey-Bass.
Elkin, I., Shea, T., Watkins, J. T., Imber, S. D., Sotsky, S. M., Collins, J. F., et al. (1989). National Institute of Mental Health treatment of depression collaborative research program: General effectiveness of treatments. Archives of General Psychiatry, 46, 971-982.
Emmelkainp, P.M.G. (2004). Behavior therapy with adults. In M. J. Lambert (Ed.), Bergin and Garfield’s handbook of psychotherapy and behavior change (5th ed., pp. 393-446). New York: Wiley.
Frank, J. D. (1961). Persuasion and healing: A comparative study of psychotherapy. Baltimore: Johns Hopkins University Press.
Frank, J. D. (1976). Psychotherapy and the sense of mastery. In R. L. Spitzer & D. F. Klein (Eds.), Evaluation of psychotherapies: Behavior therapies, drug therapies, and their interactions (pp. 47-56). Baltimore: Johns Hopkins University Press.
Frank, J. D., & Frank, J. B. (1991). Persuasion and healing: A comparative study of psychotherapy. (3rd ed.). Baltimore: Johns Hopkins University Press.
Garfield, S. L. (1987). Towards a scientifically oriented eclecticism. Scandinavian Journal of Behavior Therapy, 16, 95-109.
Garfield, S. L. (1992). Eclectic psychotherapy: A common factors approach. In J. C. Norcross & M. Goldfried (Eds.), Handbook of psychotherapy integration (pp. 169-201). New York: Basic Books.
Giblin, P., Sprenkle, D. H., & Sheehan, R. (1985). Enrichment outcome research: A meta-analysis of premarital, marital, and family interventions. Journal of Marital and Family Therapy, 11, 257-271.
Green, R. J., & Herget, M. (1991). Outcomes of systemic/strategic team consultation: III. The importance of therapeutic warmth and active structuring. Family Process, 30, 321-336.
Grencavage, L. M., & Norcross, J. C. (1990). Where are the commonalities among the therapeutic common factors? Professional Psychology Research and Practice, 21, 372-378.
Haas, E., Hill, R. D., Lambert, M. J., & Morrell, B. (2002). Do early responders to psychotherapy maintain treatment gains? Journal, of Clinical, Psychology, 58, 1157-1172.
Haley, J. (1987). Problem solving therapy (2nd ed.). San Francisco: Jossey Bass/Pfeiffer.
Helmeke, K. B., & Sprenkle, D. H. (2000). Clients’ perceptions of pivotal moments in couple’s therapy: A qualitative study of change in therapy. Journal of Marital and Family Therapy, 26, 469-484.
Henggeler, S. W., & Sheidow, A. J. (2002). Conduct disorder and delinquency. In D. H. Sprenkle (Ed.), Effectiveness research in marriage and family therapy (pp. 27-52). Alexandria, VA: American Association for Marriage and Family Therapy.
Heppner, P. P., Kivlighan, D. M., & Wampold, B. E. ( 1999). Research design in counseling (2nd ed.). Belmont, CA: Wadsworth.
Hollander-Goldfein, B. (1989). Basic principles: Structural elements of the intersystem approach. In G. R. Weeks (Ed.), Treating couples: The intersystem model of the Marriage Council of Philadelphia (pp. 38-69). New York: Brunner Mazel.
Hollon, S. T., & Beck, A. T. (2004). Cognitive and cognitive behavioral therapies. In M. J. Lambert (Ed.), Bergin and Garfield’s handbook of psychotherapy and behavior change (5th ed., pp. 447-492). New York: Wiley.
Holtzworth-Munroe, A., Jacobson, N. S., DeKlyen, M., & Whisman, M. A. (1989). Relationship between behavioral marital therapy outcome and process variables. Journal of Consulting and Clinical Psychology, 57, 658-662.
Howard, K. I., Moras, K., Brill, P. L., Martinovich, Z., & Lutz, W. (1996) The evaluation of psychotherapy: Efficacy, effectiveness, and patient progress. American Psychologist, 51, 1059-1064.
Bubble, M. A., Duncan, B. L., & Miller, S. ( 1999). The heart and soul of change: What works in therapy. Washington, DC: The American Psychological Association.
Imber, S. D., Pilkonis, P. A., Sotsky, S. M., Elkin, I., Watkins, J. T., Collins, J. F., et al. (1990). Mode-specific effects among three treatments for depression. Journal of Consulting and Clinical Psychology, 64, 352-359.
Jacobson, N. S., & Christensen, A. (1996). Integrative couple therapy: Promoting acceptance and change. New York: Norton.
Johnson, S. M. (1996). Creating connection: The practice of emotionally focused marital therapy. New York: Brunner/Mazel.
Johnson, S. M. (2002). Marital problems. In D. H. Sprenkle (Ed.), Effectiveness research in marriage and family therapy (pp. 163-190). Alexandria, VA: American Association for Marriage and Family Therapy.
Johnson, S. M. (2003). Emotionally focused couples therapy: Empiricism and art. In T. L. Sexton, G. R. Weeks, & M. S. Robbins (Eds.), Handbook of family therapy: The science and practice of working with families and couples (pp. 263-280). New York: Brunner-Routledge.
Johnson, S. M., & Denton, W. (2002). Emotionally focused couples therapy: Creating connection. In A. S. Gurnian (Ed.), The clinical handbook of couple therapy (3rd ed., pp. 221-250). New York: Guilford Press.
Johnson, S. M., & Talitman, E. (1997). Predictors of success in emotionally focused marital therapy. Journal of Marital and Family Therapy, 23, 135-152.
Karasu, T. B. (1986). The specificity versus nonspecificity dilemma: Toward identifying therapeutic change agents. American Journal of Psychiatry, 143, 687-695.
Kuehl, B. P., Newfield, N. A., & Joanning, H. (1990). A client-based description of family therapy. Journal of Family Psychology, 3, 310-321.
Lambert, M. J. (1992) Psychotherapy outcome research: Implications for integrative and eclectic therapists. In J. C. Norcross & M. R. Goldfried (Eds.), Handbook of psychotherapy integration (pp. 94-129). New York: Basic Books.
Lambert, M. J., & Bergin, A. E. (1994). The effectiveness of psychotherapy. In A. E. Bergin & S. L. Garfield (Eds.), Handbook of psychotherapy and behavior change (4th ed., pp. 143-189). New York: Wiley.
Lambert, M. J., & Ogles, B. M. (2004). The efficacy and effectiveness of psychotherapy. In M. J. Lambert (Ed.), Bergin and Garfield’s handbook of psychotherapy and behavior change (5th ed., pp. 139-193). New York: Wiley.
Lebow, J. (1997). The integrative revolution in couple and family therapy. Family Process, 36, 1-17.
Luborsky, L., Singer, B., & Luborsky, L. (1975). Comparative studies of psychotherapy: Is it true that “Everyone has won and all must have prizes?” Archives of General Psychiatry, 32, 995-1008.
Martin, D. J., Garske, J. P., & Davis, M. K. (2000). Relationship of the therapeutic alliance with outcome and other variables: A meta-analytic review. Journal of Consulting and Clinical Psychology, 68, 438-450.
McFarlane, W. R., Dixon, L., Lukens, E., & Lucksted, A. (2002). Severe mental illness. In D. H. Sprenkle (Ed.), Effectiveness research in marriage and family therapy (pp. 225-288). Alexandria, VA: American Association for Marriage and Family Therapy.
Miller, S. D., Duncan, B. L., & Hubble, M. A. (1997). Escape from Babel: Toward a unifying language for psychotherapy practice. New York: Norton.
Minuchin, S. (1974). Families and family therapy. Cambridge, MA: Harvard University Press.
Minuchin, S. (1998). Where is the family in narrative family therapy? Journal of Marital and Family Therapy, 24, 397-404.
Napier, A. (1997, September). Called by families: The journey of the therapist. Plenary address presented at the Annual Meeting of the American Association for Marriage and Family Therapy, Atlanta, GA.
Nichols, M. P., & Schwartz, R. C. (2001). Family therapy: Concepts and methods (5th ed.). Boston: Allyn & Bacon.
Norcross, J. C., & Goldfried, M. (Eds.). (1992). Handbook of psychotherapy integration. New York: Basic Books.
Orlinsky, D. E., Grawe, K., & Parks, B. K. (1994). Process and outcome in psychotherapy-Noch einmal. In A. E. Bergin & S. L. Garfield (Eds.), Handbook of psychotherapy and behavior change (4th ed., pp. 270-378). New York: Wiley.
Palterson, C. H. (1984). Empathy, warmth, and genuineness in psychotherapy: A review of reviews. Psychotherapy: Theory, Research, Practice, Training, 21, 431-438.
Paul, G. (1967). Strategy of outcome research in psychotherapy. Journal of Consulting Psychology, 31, 109-118.
Pinsof, W. M. (1995). Integrative problem-centered therapy. New York: Basic Books.
Pinsof, W. M., & Catherall, D. (1986). The integrative psychotherapy alliance: Family, couple and individual therapy scales. Journal of Marital and Family Therapy, 12, 137-152.
Pinsof, W. M., & Wynne, L. C. (2000). Toward progress research: Closing the gap between family therapy practice and research. Journal of Marital and Family Therapy, 26, 1-8.
Prochaska, J. O. (1999). How do people change, and how can we change to help many more people? In M. A. Hubble, B. L. Duncan, & S. Miller (Eds.), The heart and soul of change: What works in therapy (pp. 227-255). Washington, DC: American Psychological Association.
Rosenzweig, S. (1936). Some implicit common factors in diverse methods in psychotherapy. Journal of Orthopsychiatry, 6, 412-415.
Schwartz, R. C. (1995). Internal family systems therapy. New York: Guilford Press.
Sexton, T. L., & Alexander, J. F. (2003). Functional family therapy: A mature clinical model for working with at-risk adolescents and their families. In T. L. Sexton, G. R. Weeks, & M. S. Robbins (Eds.), Handbook of family therapy: The science and practice of working with families and couples (pp. 323-350). New York: Brunner-Routledge.
Shadish, W. R., & Baldwin, S. A. (2002). Meta-analysis of MFT interventions. In D. H. Sprenkle (Ed.), Effectiveness research in marriage and family therapy (pp. 339-370). Alexandria, VA: American Association for Marriage and Family Therapy.
Shadish, W. R., Ragsdale, K., Glaser, R. R., & Montgomery, L. M. (1995). The efficacy and effectiveness of marital and family therapy: A perspective from meta-analysis. Journal of Marital and Family Therapy, 21, 345-360.
Shapiro, D. A., & Shapiro, D. (1982). Meta-analysis of comparative therapy outcome studies: A replication and refinement. Psychological Bulletin, 92, 581-604.
Sheidow, A. J., Henggeler, S. W., & Schoenwald, S. K. (2003). Multisystemic therapy. In T. L. Sexton, G. R. Weeks, & M. S. Robbins (Eds.), Handbook of family therapy: The science and practice of working with families and couples (pp. 303-321). New York: Brunner-Routledge.
Sloane, R. B., Staples, F. R., Cristol, A. H., Yorkston, N. J., & Whipple, K. (1975). Psychotherapy versus behavior therapy. Cambridge, MA: Harvard University Press.
Smith, M. L., Glass, G. V., & Miller, T. I. (1980). The benefits of psychotherapy. Baltimore: Johns Hopkins University Press.
Sprenkle, D. H. (Ed.). (2002). Effectiveness research in marriage and family therapy. Alexandria, VA: American Association for Marriage and Family Therapy.
Sprenkle, D. H., Blow, A. J., & Dickey, M. H. (1999). Common factors and other nontechnique variables in marriage and family therapy. In M. A. Hubble, B. L. Duncan, & S. D. Miller (Eds.), The heart and soul of change: What works in therapy (pp. 329-360). Washington, DC: American Psychological Association.
Stolk, Y., & Perlesz, A.J. (1990). Do better trainees make worse therapists? A follow-up study of client families, family Process, 29, 45-58.
Tallman, K., & Bohart, A. C. (1999). The client as a common factor: Clients as self-healers. In M. A. Hubble, B. L. Duncan, & S. D. Miller (Eds.), The heart and soul of change: What works in therapy (pp. 91-131). Washington, DC: American Psychological Association.
Waltz, J., Addis, M. E., Koerner, K., & Jacobson, N. S. (1993). Testing the integrity of a psychotherapy protocol: Assessment of adherence and competence. Journal of Consulting and Clinical Psychology, 61, 620-630.
Wampler, K. (1997). Systems theory and outpatient mental health treatment: Twelve priorities for MFT research. Paper presented at the Inaugural AAMFT Research Conference, Santa Fe, NM.
Wampold, B. E. (2001). The great psychotherapy debate: Models, methods, and findings. Mahwah, NJ: Erlbaum.
Weinberger, J. (1995). Common factors aren’t so common: The common factors dilemma. Clinical Psychology-Science and Practice, 2, 45-69.
White, M., & Epston, D. (1990). Narrative means to therapeutic ends. New York: Norton.
Wolpe (1975). Forward. In B. Sloane, F. Staples, A. Cristol, N. Yorkston, & K. Whipple (Eds.), Psychotherapy versus behavior therapy. Cambridge, MA: Harvard University Press.
Douglas H. Sprenkle
Adrian J. Blow
Saint Louis University
Douglas H. Sprenkle, PhD, Marriage and Family Therapy Program, Department of Child Development and Family Studies, Purdue University; Adrian J. Blow, PhD, Department of Counseling and Family Therapy, Saint Louis University.
Correspondence regarding this article should be addressed to Douglas H. Sprenkle, PhD, Doctoral Program in Marriage and Family Therapy, Purdue University, Fowler House, 1200 W. State Street, West Lafayette, Indiana, 47907-2055. E-mail: firstname.lastname@example.org
Copyright American Association for Marriage and Family Therapy Apr 2004
Provided by ProQuest Information and Learning Company. All rights Reserved