The Heart & Soul of Change: What Works in Therapy

For over four decades, the message from psychotherapy outcome research has been getting clearer: the theories and techniques of professional therapy have very little to do with therapeutic success.

Rating: 5

The Heart & Soul of Change: What Works in Therapy

Edited by , 1999. ISBN 155798557X. American Psychological Association. xxiv + 462 pages.

For over four decades, the message from psychotherapy outcome research has been getting louder and clearer: the theories and techniques of professional therapy have very little to do with therapeutic success. This edited volume of 14 papers from 24 contributors sets out to explore what factors the research tells us do make a difference to outcome — and what clinicians should do about it. As the editors put it, “The recurrent finding that theories and their associated technical operations do not significantly contribute to outcome…deserves much more notice” (p. 12).

They pull no punches in this hard-headed look at what really works, as distinct from what theories tell us ought to work, describing the book as follows:

It aims to explain what enables psychotherapy and practices in certain related fields to achieve their results. For this reason, we have attempted to put to rest the customary equivocation found in the writings of therapy researchers and clinicians. The profession has labored under the conventions of “it seems,” “it appears,” “it could also be” when describing therapy and its processes… It is easy to derive the impression that nobody wants to say, “This is it. This is fact as close as we can come to it now, and here’s what to do.” (pp. 11-12)

Four sections make up the book, covering empirical foundations, four classes of ‘common factors’ responsible for therapeutic outcome, special applications of the common factors, and their financial and practical implications for the field.

Empirical Evidence for Common Factors in Psychotherapy

The first of two chapters on empirical foundations, by Asay and Lambert, can be understood as an explanation, expansion, buttressing and application of a conclusion first suggested by Lambert (1992) — namely, that psychotherapy outcome variance is attributable to the following factors in discernible proportions:

  • 40%: client and extratherapeutic factors (such as ego strength, social support, etc.)
  • 30%: therapeutic relationship (such as empathy, warmth, and encouragement of risk-taking)
  • 15%: expectancy and placebo effects
  • 15%: techniques unique to specific therapies

In other words, the bulk of client improvement is attributable to factors common to all different types of counselling and psychotherapy (sometimes called ‘nonspecific factors’), as distinct from being attributable to factors specific to individual approaches.

In the course of their extensive review of the research literature, the authors’ notes on rates of spontaneous remission (pp. 32-33) will be especially interesting to clinicians who regard any particular therapist characteristics as somehow necessary for therapeutic change to occur: clearly no therapist characteristics can actually be necessary, if some clients improve without them! (This point returns elsewhere in the volume.)

Their review of the research on the therapeutic relationship — perhaps the most frequently studied of the common factors — highlights the fact that it is not enough to focus on therapist-provided contributions to the relationship (such as the ‘necessary and sufficient’ core conditions of the person-centred approach), but that attention should be paid to the relationship itself. No one questions the importance of core conditions like acceptance, accurate empathy and therapist genuineness; but it turns out that client perceptions of the relationship are consistently more correlated with outcome than those of objective raters. In other words, how clients experience the characteristics offered by therapists is more important than what those therapists are ‘objectively’ offering. (This point returns more strongly in a later chapter in the book’s second section.)

The authors’ subsequent review of expectancy and placebo effects is quite brief, but this class of common factor receives a great deal more attention later in the volume. In connection with techniques, the authors point to a few cases where research has demonstrated the effectiveness of specific techniques, ultimately offering the take-home message that:

There is no reason for those who are devoted to the development and testing of specific techniques to discount the obvious benefits of common factors and particularly the importance of therapist attitudes of respect, caring, understanding, and concern. By the same token, those of us who are convinced of the primary importance of the therapist, as a person, would be well served by remaining open to the likelihood that specific techniques, when offered within the safety of the therapeutic relationship, will appreciably add to the therapeutic encounter. (p. 49)

The chapter concludes with a list of eight specific implications for practice and training, ranging from a reminder that psychotherapy clearly does work (while it is clear that theories and models play little role, it is just as clear that psychotherapy is highly effective) to encouragement to learn about client variables:

Therapists should be familiar with client variables that have been shown to affect outcome and develop the skills to evaluate the suitability of a given client for the intervention offered. In addition, as a supplement to their own psychotherapeutic skills, it behooves the therapist to become familiar with the social support networks and community resources available to their clients and to help them in identifying and using these resources. (p. 43)

(Just to be nitpicky, I have to wonder: why the phrase ‘the suitability of a given client for the intervention offered’, rather than ‘the suitability of the intervention considered for the given client’?)

The second of two chapters on empirical foundations, by Maione and Chenail, focuses on qualitative research on common factors. It reads as a minimally-structured research review, recounting in series the results of study after study generally consistent with the conclusions of the previous chapter. I personally find this less useful than the previous chapter’s quantitative approach, although that probably says more about my own preferences for research writing than the quality of the review itself. The same is probably true about my slight irritation with the chapter’s profusion of TLAs (three letter acronyms) and my sense that the piece pays too much attention to defending the relevance of qualitative research at the expense of focusing on exactly what the chapter was contributing.

Exploring Therapy Common Factors in Detail

Karen Tallman and Arthur C. Bohart lead the exploration of common factors in greater detail with a chapter on clients as self-healers:

Our thesis in this chapter is that the client’s capacity for self-healing is the most potent common factor in psychotherapy. It is the “engine” that makes therapy work… Therapy facilitates naturally occurring healing aspects of clients’ lives. Therapists function as support systems and resource providers. This view contrasts sharply with most of the literature on psychotherapy. There, the therapist is the “hero” who, with potent techniques and procedures, intervenes in clients’ lives and fixes their malfunctioning machinery… (p. 91)

Referring to the now-famous ‘dodo bird verdict’, from Luborsky et al’s (1975) borrowing of Lewis Carroll’s “everybody has won, so all must have prizes”, the authors state simply: “In this chapter we argue that the most parsimonious explanation for the dodo bird verdict is that it is the client, not the therapist or technique, that makes therapy work” (p. 91) Moreover, referring to Lambert’s review cited earlier,

Considering that placebo factors are client factors (client self-healing through hope and belief), and clients contribute at least as much to the therapeutic relationship as does the therapist, Lambert’s figures also imply that the client is responsible for 70% or more of the outcome variance. (p. 95)

Not content with merely reciting the clear messages of the research literature, the authors attempt to offer some explanation of the client’s primary importance and the dodo bird verdict itself.

We believe the dodo bird verdict occurs because the client’s abilities to use whatever is offered surpass any differences that might exist in techniques or approaches… Clients utilize and tailor what each approach provides to address their problems. Even if different techniques have different specific effects, clients take these effects, individualize them to their specific purposes, and use them. Thus, for example, a client can use cognitive or interpersonal techniques…or emotional exploration procedures, or empathically based client-centered therapy…to move themselves out of depression. (p. 95, emphasis original)

Hammering home the importance of the client, the authors explore self-generated change and spontaneous recovery, and on the therapeutic relationship, they suggest:

We do not disagree with these proposed effects of the relationship, yet they are insufficient to explain the therapeutic change process. Not only do client variables account for more of the variance, but also the noted studies…clearly demonstrate that therapists are frequently not necessary for change to take place. Of course, if the therapist is unnecessary, then the relationship is unnecessary as well.

Therefore, the impact of the relationship must be reinterpreted…we believe that the relationship is yet another resource which clients utilize to mobilize personal agency and change. In other words, the relationship…is not in principle different from techniques, self-help books, or computer programs. It is a resource that facilitates, supports, or focuses clients’ self-healing efforts. (p. 102, emphasis original)

The second chapter of this section, by Bachelor and Horvath, focuses on the therapeutic relationship. One main message of this chapter is that more attention should be paid to the relationship itself, which in many cases differs from paying attention to what relationship conditions the therapist thinks he or she is offering to the client.

For example,

Effective responses are attitudes and interventions that are appropriate to the individual client… Attitudes or interventions such as warmth, support, self-disclosure, deeper exploration, and so on appear to be highly beneficial to some clients, but more or less inconsequential to others, whereas still other clients may react adversely to such responses. (pp. 146-7)

Thus, paying attention to the relationship itself requires significant flexibility and a willingness to adapt on the part of the therapist. It won’t do simply to decide in advance that one will be warm or that one will explore deeply or that one will…[fill in your choice], without regard for what actually promotes the relationship in the case of a single, individual client.

By way of example, the authors discuss Bachelor’s (1988) study indicating how differently clients may interpret even basic ‘core conditions’ such as empathy. The study in question shows that

About 44% of clients valued a cognitive-type of empathic response, another 30% valued an affective-toned communication, and the remaining clients viewed empathy to be optimally a sharing of personal information or a nurturant-like therapist response. (p. 144)

Another study by Bachelor (1995) revealed three relatively different types of relationship that were deemed to be therapeutic, depending on the individual client; in other words, whether a relationship is therapeutic is to be found — not surprisingly! — in the eye of the client and not in a specific set of relationship conditions offered by the therapist as a one-size-fits-all package.

Indeed, from my own perspective, having trained in the person-centred approach (“Person-Centred Counselling”), this chapter serves to crystallize and confirm for me a long-growing but previously unarticulated suspicion: to the extent that they focus on what the therapist provides (namely, the ‘core conditions’), person-centred theorists do not take either the relationship or the client nearly as seriously as they would typically have us believe. By contrast, this chapter focuses on a wide range of variables, both within the client and within the therapist, which contribute to the quality and effectiveness of the relationship itself.

The third chapter of this second section, by Snyder, Michael and Cheavens, offers a theoretical framework about hope within which to understand placebo and expectancy effects. They argue that the awakening or strengthening of hopefulness within the client provides an appropriate way of understanding the impact of placebo and expectancy. The theoretical framework suggests interesting echoes both with cognitive approaches (“Cognitive Therapy & CBT”), in terms of a client thinking through ways of achieving goals and in terms of addressing negative biases, and with existential approaches (“Existential Counselling”), in terms of a client identifying what is important to them. More an exploration of hope theory itself than of placebo and expectancy effects themselves, the emphasis of this chapter feels slightly out of place, but placebo effects return for more coverage in the subsequent section.

The fourth chapter of the section addresses models and techniques. By Ogles, Anderson and Lunnen, it is subtitled ‘Contradictions Between Professional Trends and Clinical Research’ — referring to the clash between the dodo bird verdict and professional trends favouring specificity.

The chapter argues once again (from empirical evidence, of course!) that models and techniques are of limited importance in therapy outcome, although the authors do cite some examples of technical approaches which have demonstrated superiority, including exposure techniques for anxiety disorders and behavioural approaches for some sexual dysfunctions. The authors conclude by urging the development of models which emphasize common factors and general relationship-building skills, rather than techniques.

The section’s final chapter explores Prochaska’s stage theory of change. He suggests that “What the field now needs most is an adequate theory of behavior change… The field needs to know how people change before therapy begins, after it ends, and when therapy never occurs” (pp. 227-228). Later, he says,

This quest began with a comparative analysis of the major systems of psychotherapy… I found that these systems had much more to say about personality and psychotherapy. That is, they are theories more about why people do not change than how people can change. They emphasize more the content of therapy — such as feelings, fantasies, thoughts, overt behaviors, and relationships — than the process of change. (p. 228)

And he goes on:

What we discovered was a phenomenon that was not contained within any of the leading theories of therapy. Ordinary people taught us that change involves progress through a series of stages. At different stages, people apply particular processes to progress to the next stage. (p. 228)

The remainder of the chapter explores the model of stages of change and describes how interventions may be targeted so as to match the stage where the client begins — and the costs in terms of efficacy if stage-mismatched interventions are chosen. Unlike other contributions in the book, the chapter is also concerned primarily with interventions designed for delivery to large swathes of the population as a whole, such as mass programmes to address smoking, rather than with individual therapy. As Prochaska puts it, commenting on a large-scale study on computer-assisted treatment for smoking:

In summary, future mental health specialists will need to know how to prescribe and provide interactive technologies to entire populations. I am convinced that such technologies will be to behavior change what medications have been to biological medicine: the most cost-effective way to bring the maximum amount of science to bear on important problems in entire populations. More costly professional therapists will be reserved for the most complicated cases that cannot be helped by the more cost-effective computers. (p. 251)

Special Applications of Psychotherapy Common Factors

The collection’s third section includes four papers on ‘special applications’ of the common factors, the first of which, by Albert W. Scovern, addresses common factors in medicine. At first glance, this chapter is a little far afield for a book focusing primarily on common factors in therapy, but in fact it is entirely appropriate. Scovern reviews a large body of research about placebo effects, the impact of the doctor/patient relationship on treatment outcomes, and various studies in psychophysiology and psychoneuroimmunology. The bottom line is that psychological factors have a huge influence on treatment outcomes even in the most technical areas of medicine.

The second chapter focuses on common psychosocial factors as they bear on pharmacotherapy and underscores “the inevitability of psychosocial influences on biological outcomes” (p. 298). This truly fascinating chapter reveals the astonishing impact of those psychosocial factors, noting, for instance, that “for conditions such as depression, the therapeutic impact of placebos approaches the level reached by psychoactive drugs” (p. 300).

The chapter is not simply psychiatric drug bashing; Roger P. Greenberg argues clearly and concisely from empirical evidence:

Even the idea that psychoactive drugs produce their effects through biological mechanisms while psychological approaches do not is found wanting. As Fisher and Greenberg…noted, all effects occur in tissue, and one type is no more biologically real than the other. Thus, one can point to evidence that placebo effects are physiologically mediated…or that successful psychotherapeutic treatment produces biochemical changes just as drugs do. For example, Baxter and his colleagues…demonstrated equivalent brain imagery changes in successful psychotherapeutic and drug treatments for obsessive-compulsive disorder. Talking treatment proved to be as facilitative of biological alteration (on brain imagery changes) as did the specific drug therapy. In short, although medications provide specificity of ingredients that cannot be matched by psychosocial approaches, treatment outcomes may overlap to the point of being indistinguishable, even on biological measures. (p. 300)

Even more strongly, Greenberg later suggests that “psychosocial factors are arguably the largest therapeutic component in most effective psychiatric medication treatments — more critical than dosage or blood levels of the ingested drug” (p. 315). Indeed, the author also cites studies showing that not only are placebo effects sometimes equivalent in strength to psychoactive drugs, but sometimes psychosocial factors are even able to reverse the expected biochemical actions of particular drugs.

The relationship takes centre stage in Greenberg’s assessment of the impact of psychosocial factors in pharmacotherapy. Commenting on a study of data from the Treatment of Depression Collaborative Research Program, comparing four different treatment outcomes for unipolar depression (cognitive-behavioural therapy, interpersonal psychotherapy, imipramine plus clinical management, and a pill-placebo plus clinical management control condition), he notes that “Overall, the authors concluded that clinical improvement was minimally related to the type of treatment received, but substantially determined by the quality of therapeutic relationship that patients experienced” (p. 305). And he goes on later, “No specific factors associated with either drug or psychotherapy treatments proved to be more important than the patient and practitioner personal qualities that interact to establish an effective therapeutic relationship” (p. 307).

There is much more to this eye-opening chapter, including notes on flaws in nearly all double-blind drug trials as well as on the problem of the representativeness of participant samples (namely, that most patients in the real world display comorbidity, while most study participants do not).

The third chapter of this section, by Sprinkle, Blow and Dickey, covers common factors in marriage and family therapy, an area which the authors paint as still more strongly influenced by individual personalities than by actual evidence. Part of their task in fact is to combat this unfortunate situation:

We hope that this chapter will contribute to diminishing the bombast and hubris of MFT theorists. Indeed, a mellowing of the field would be most desirable. Less attention needs to be paid to charismatic claims of uniqueness and more credence given to those common variables from which most change springs. (p. 351)


We believe that the “worship” of technique is so ingrained in the MFT culture that it would produce considerable cognitive dissonance for most MFTs to be persuaded that their cherished techniques are not primarily responsible for change. (p. 352)

The principle weakness of this chapter (but certainly an understandable one) is that the authors’ descriptions of potential common factors in MFT are based upon observations about methods and behaviours which MFT therapists have in common, as distinct from empirical investigations about what factors actually influence outcome. In other words, observing that MFT therapists have a given factor in common does not imply that that given ‘common factor’ has anything at all to do with therapeutic outcome. It might be that MFT therapists commonly have brown hair, but this does not imply that having brown hair is causally relevant to therapeutic outcome.

The section’s final chapter, by John J. Murphy, addresses common factors from the standpoint of change within schools, primarily from a change management perspective, and it focuses on the author’s ‘5E’ approach.

Implications of Common Factors Research

The first of two chapters in the book’s final section, by Brown, Dreis and Nace, is a delightfully two-edged sword. On the one hand, the authors argue convincingly that ‘managed behavioural health care organizations’ are focusing efforts in the wrong directions, and they paint a promising picture of how large-scale data collection efforts on the part of MBHOs could actually benefit the field as a whole and even individual practitioners by providing solid, evidence-based feedback on effectiveness. On the other hand, it is the very influence of MBHOs and their data collection to which many practitioners would attribute a steady deterioration in client care over the past several years — as a result of evidence-based downward pressure on therapeutic ‘dosage’ and the progressive erosion of individual practitioners’ clinical freedoms in terms of whom to treat and in what ways.

The authors begin by reporting significant conclusions which have already emerged from the data-collection efforts of one large MBHO:

  • The outcome of therapy is highly variable. The standard deviation of the change score tends to be approximately twice as large as the mean average change score for a large sample of clients.
  • Ensuring the application of specific guidelines and approved or ’empirically validated’ psychotherapy methods does not lead to improved treatment outcome. Differences in treatment methods, diagnoses, and even length of treatment account for less than 5% of the variance in outcome.
  • There are significant differences in effectiveness between individual providers and even entire clinical delivery systems. Such differences cannot be explained by the nature of the cases being seen or the treatment methods being used.
  • Response to treatment in the first few sessions is highly predictive of the eventual outcome. This phenomenon makes it feasible to monitor cases and predict cases at risk of a poor outcome early in the treatment process. (p. 390, emphasis original)

The ‘negative’ edge of the sword is well summarized by the following comment:

In a real and radical way, then, the very nature of the psychotherapy business is shifting from simple reimbursement for services to compensation for the clinical outcome. In the emerging environment, the outcome of the service rather than the service itself is the product that providers and payers have to market and sell. Those unable to systematically evaluate the outcome of treatment will have nothing to sell to purchasers of health care services. (p. 393)

However, there is a ‘bright side’, too:

70% of the total outcome variance is accounted for when a strong therapeutic relationship is combined with a successful incorporation of client factors in the treatment process. Such data strongly suggest that the interests of MBHOs would be better served by focusing on how well providers facilitate client change rather than on managing or dictating the treatment approaches they use — in other words, shifting the emphasis in management of behavioral health care away from identifying the “correct” treatment and toward the personal attributes and abilities of the individual provider. (p. 399)

And summarizing again, shortly thereafter:

There is no relationship between the treatment methods endorsed by providers and differences in outcomes as reported by clients. For this reason, it is safe to say that MBHOs would be better off using their resources to manage something other than provider compliance with preferred psychotherapy methods. (p. 400)

The authors of the chapter are themselves well positioned within the MBHO industry and are largely responsible for the data collection efforts they report, so there is some hope that their views on the mistaken approaches of MBHOs will have some impact. For better or worse, that this impact may be coupled with even stronger pressures to measure, report, benchmark and compete in terms of effectiveness may ultimately turn out to be a fact of life.

Finally, the editors of the volume bring it all together with a chapter aptly titled Directing Attention to What Works. They highlight and reinforce a number of points made by contributors elsewhere in the volume, for instance reiterating Bachelor and Horvath’s view of the relationship from chapter 5:

Successful therapeutic relationships are those in which the definition of the therapist-provided variables is extended to fit with the client’s own unique experience of those variables — the client’s definition of therapist-provided warmth, empathy, respect, and genuineness. For practice, therefore, clinicians stand the greatest chance of enabling the contribution of relationship factors to outcome when they purposefully tailor their provision of the core conditions to the client’s definition. (p. 417)

They continue on the following page,

To do this on a day-to-day basis…demands a higher measure of flexibility on the part of the therapist and a willingness to change one’s relational stance to fit with the client’s perceptions of what is most helpful. Some clients, for instance, will prefer a formal or professional manner over a casual or warmer one… Clearly, the one-approach-fits-all strategy is guaranteed to undermine alliance formation. (p. 417)

On models and techniques, they suggest that “their principal contribution to therapy comes about by enhancing the potency of the other common factors — client/extratherapeutic, relationship, placebo, hope and expectancy” (pp. 421-22).

A significant part of the editors’ final chapter is dedicated to specific recommendations for improving practice by applying an awareness of the common factors, including specific questions therapists may wish to ask or specific areas on which they may wish to focus attention. This all has something of a solution-focused or a brief therapy ring to it, which is understandable given the editors’ own professional backgrounds — but it is also justified by the evidence reviewed throughout the volume. Hopefully the rest of the volume will have convinced readers that the suggestions should be evaluated not as instances of solution-focused or brief therapy thinking, but as honest attempts to derive some practical insights from the abundant research evidence — honest attempts to clarify what actually works.

If all suggestions, debates and hypotheses within counselling and psychotherapy could be evaluated against that same backdrop — a backdrop of empirical evidence and concern for what actually works in therapy — I believe the field would be much better off. And so would our clients.

Please see our Review Disclosure Policy.

All clinical material on this site is peer reviewed by one or more clinical psychologists or other qualified mental health professionals. This specific article was originally published by on and was last reviewed or updated by Dr Greg Mulhauser, Managing Editor on .

3 Comments on “Hubble, Duncan and Miller on What Works in Therapy”

The comments form is currently closed, but you can click to read the comments left previously on “Hubble, Duncan and Miller on What Works in Therapy”.

Overseen by an international advisory board of distinguished academic faculty and mental health professionals with decades of clinical and research experience in the US, UK and Europe, provides peer-reviewed mental health information you can trust. Our material is not intended as a substitute for direct consultation with a qualified mental health professional. is accredited by the Health on the Net Foundation.

Copyright © 2002-2023. All Rights Reserved.