People who find talk therapy genuinely useful are having a real experience. The relationship formed across months or years of regular sessions matters. For a great many people, the life before and the life after are meaningfully different, and no honest accounting of what therapy does should start by dismissing that. The people who tell you it changed things are not confabulating.
What deserves more scrutiny is what the research actually demonstrates, and how that maps onto the confident claims made on its behalf. The gap between the two is larger than the public-facing discussion tends to acknowledge, and larger, in our reading of the literature, than the field’s public-facing discussion tends to acknowledge.
The efficacy problem
Psychotherapy research typically distinguishes between efficacy and effectiveness. Efficacy studies are randomized controlled trials conducted under controlled conditions: carefully selected participants, manualized treatment protocols, therapists trained to deliver a specific approach with high fidelity, and outcomes measured against a control group. Effectiveness studies examine what happens in ordinary clinical practice, with the full range of patients, the full range of therapist training and experience, and the conditions that actual treatment involves.
The efficacy studies tend to look impressive. The effectiveness research is more mixed. The gap between what a modality produces under trial conditions and what it produces in the real world of clinical practice is a consistent feature of the literature, and it matters considerably when the research is used to make public claims about what therapy does. The numbers that appear in press releases and mental health campaigns are typically the efficacy numbers. The effectiveness numbers, where they exist, are less frequently cited.
This is not a minor distinction. An intervention that performs well under ideal research conditions and less well under ordinary conditions is interesting in a specific way: it may be telling you that the conditions matter more than the intervention itself.
The allegiance effect
One of the more uncomfortable findings in psychotherapy research is the allegiance effect: the documented tendency for therapy outcome studies to favor whichever modality the researcher practices or has a professional investment in. Meta-analyses examining this have found that a substantial proportion of the variance in therapy outcome research is explained not by differences between therapies, but by the theoretical orientation of the researcher conducting the study. CBT researchers tend to produce findings favorable to CBT. Psychodynamic researchers tend to produce findings favorable to psychodynamic therapy.
This is not proof that the researchers are consciously biased. It likely reflects a set of design choices, measurement selections, and interpretive tendencies that operate below the level of deliberate manipulation, which makes it harder to correct for, not easier. What it means in practice is that much of the research base for the claims made about specific therapeutic modalities cannot be read as straightforwardly as it is typically presented. Read carefully, the allegiance literature is not a footnote to the evidence base. It is a structural problem with it.
The Dodo Bird Verdict
Given the allegiance problem, it is worth paying attention to the researchers who have looked across the findings rather than within them. The conclusion that emerges from several decades of comparative meta-analysis is known in the field as the Dodo Bird Verdict, a reference to the Lewis Carroll character who awards everyone a prize. Its central claim: most forms of psychotherapy, when compared against each other, produce roughly similar outcomes. Lester Luborsky and colleagues documented this finding in a widely cited 2002 paper in Clinical Psychology: Science and Practice, and the conclusion has been replicated in subsequent meta-analyses with enough regularity that it represents something close to a consensus position among researchers who study comparative outcomes, even as it remains contested by researchers committed to specific modalities.
The Dodo Bird Verdict does not mean therapy doesn’t work. It means that the specific technique matters less than the field’s marketing of specific techniques would suggest. If CBT and psychodynamic therapy and interpersonal therapy produce similar outcomes, the active ingredient is probably not the thing that distinguishes them.
What the research does support
The most robust finding in psychotherapy outcome research is also the one that receives the least attention in public discussions of which therapy to choose. It concerns the therapeutic alliance: the quality of the relationship between therapist and client. Across decades of research, including the work of Bruce Wampold documented in his book The Great Psychotherapy Debate (Lawrence Erlbaum, 2001, second edition 2015), the therapeutic alliance accounts for more of the variance in therapy outcomes than the specific technique being employed. A good therapeutic relationship predicts better outcomes more reliably than a particular protocol does.
Jonathan Shedler’s 2010 paper in American Psychologist, “The Efficacy of Psychodynamic Psychotherapy,” documented that psychodynamic therapy produces effect sizes comparable to those of therapies that receive substantially more research attention and promotional support. His broader argument was that the relationship between research investment and modality effectiveness is not a clean one: therapies that attract more research funding get studied more, produce more publications, and therefore appear better supported, regardless of whether they outperform alternatives.
What the combined literature suggests is that the most important variable in therapy is the person you are working with, not the approach they use. The person who is a good therapist for you will produce better outcomes than a poorly matched therapist using a highly credentialed technique. This is not nothing. It is, in fact, a substantive finding. It just does not translate easily into the kind of brand-level claims the field has become accustomed to making.
What an honest conversation would look like
The field of psychotherapy has a legitimate body of evidence showing that therapy, in a variety of forms, produces better outcomes than no treatment for a range of difficulties. That finding holds across the methodological critiques, and it is important. But the more specific claims, the ones that pit one modality against another, assign precise effect sizes to particular conditions, or translate efficacy trial results directly into expectations for ordinary clinical practice, deserve more skepticism than they typically receive.
So here is the question the field has not answered in public: if the active ingredient is the relationship, and the comparative outcome data is shaped by who runs the studies, why are training programs, insurance reimbursement codes, and clinical guidelines still organized around the brand names of specific modalities? Why does the consumer-facing language still promise that CBT does this and DBT does that, when the most replicated finding in the literature is that the differences between these brands are smaller than the differences between individual therapists?
An honest reform agenda would follow from the evidence. Fund therapist-level outcome tracking rather than modality-level marketing. Credential clinicians on alliance and outcomes, not on the number of weekend workshops in a branded protocol. Tell prospective clients, plainly, that finding the right person matters more than finding the right acronym. None of that requires waiting for more research. It requires the field to stop citing evidence it has already, quietly, outgrown.
