Therapists often develop a coherent explanation for why a client improved. We may conclude that the attachment relationship became more secure, the client learned to regulate emotion, a protective part relaxed, or an avoided feeling was finally experienced. These explanations may be clinically useful and sometimes correct. Yet psychotherapy research continues to remind us that a plausible explanation is not the same as a demonstrated mechanism of change.
This distinction matters because our theories shape what we notice. When therapists believe that attachment repair produces change, we are likely to interpret improvement through the language of attachment. When we believe emotion regulation is central, we may describe the same improvement as increased regulatory capacity. The client may indeed have become more secure and better able to tolerate feeling—but those changes could be causes of recovery, consequences of recovery, or mutually reinforcing aspects of a more complicated process.
Several recent studies offer a useful corrective. Together, they suggest that therapists should pay closer attention to what actually happens in sessions, the sequence in which change occurs, and the difference between a meaningful clinical formulation and a verified causal explanation.
The First Session Is Already Treatment
Malins and colleagues examined 103 first-session transcripts from a pragmatic clinical trial involving people with severe depression. Sixty-three clients received person-centered experiential therapy and 40 received cognitive behavioral therapy. Trained observers coded the kinds of interactions that occurred and examined whether those interactions predicted improvement over the course of treatment.
Across both therapies, clients who became actively engaged in therapy-related action planning during the first session tended to improve more. By contrast, outcomes were poorer when larger portions of the first session were consumed by the administrative completion of measures rather than by discussing what those measures meant. Within person-centered experiential therapy, clients also improved more when the first session allowed substantial space to describe their difficulties and the contexts in which those difficulties had developed.
These findings do not prove that action planning causes recovery. A client’s ability to participate in planning may reflect hope, agency, alliance, readiness, therapist skill, or lower levels of impairment. Nevertheless, the study challenges the assumption that an initial meeting is primarily an information-gathering encounter. The first session is already shaping the client’s experience of therapy: whether the person feels studied or met, processed or understood, instructed or invited into meaningful participation.
For experiential therapists, “action planning” should not be reduced to assigning homework or moving quickly toward solutions. An action may be as modest as noticing what happens in the body when the client anticipates disappointing someone, pausing before automatically reassuring a partner, or returning next week with curiosity about a recurring emotional sequence. The clinically important feature is authorship: the client participates in choosing what will be explored rather than passively receiving the therapist’s plan.
This has direct implications for supervision. We should ask whether an intake session created a living therapeutic encounter or remained an interview about the client’s life. Did the therapist use an assessment to deepen shared understanding, or merely complete it? Did the session end with greater agency and curiosity, or with the client feeling that the therapist had gathered information without yet joining them?
What If Emotion Regulation Follows Improvement?
A second study complicates another common assumption. Egli and colleagues studied 198 people with acute and severe major depression receiving intensive inpatient or day-clinic schema therapy or cognitive behavioral therapy. The researchers repeatedly measured depression, attachment anxiety, attachment avoidance, cognitive reappraisal, and expressive suppression.
All four proposed mechanisms improved over the course of treatment. At first glance, that might appear to confirm the familiar conclusion that reduced attachment insecurity and improved emotion regulation produced the reduction in depression. More rigorous temporal modeling, however, did not support that conclusion. Changes in attachment and emotion regulation did not predict subsequent symptom improvement and did not mediate the effect of treatment on depression.
For cognitive reappraisal and expressive suppression, the temporal pattern pointed in the reverse direction. Improvements in depression preceded later improvements in these emotion-regulation capacities. The researchers therefore raised the possibility that people may become more capable of regulating emotion because depression has begun to lift, rather than depression lifting because they first learned to regulate emotion.
This does not make emotion-regulation work unimportant. It does mean that therapists should be cautious about requiring regulation, insight, or attachment security as prerequisites for recovery. A deeply depressed person may initially benefit from structure, behavioral movement, medication, relational hope, relief from environmental stress, or simply the experience that change is possible. As energy and flexibility return, the person may then become more capable of reflection, emotional differentiation, connection, and regulation.
The finding also creates useful epistemic friction for experiential therapists. When a client becomes more emotionally available after several weeks of improvement, we should not automatically conclude that affective deepening caused the improvement. The client may now be able to engage more deeply because something else has already shifted. Therapy may involve multiple reciprocal processes without any one of them functioning as the singular mechanism.
In supervision, this calls for greater precision. “The client improved as attachment security increased” is an observation about co-occurring change. “The client improved because we repaired attachment” is a causal claim. The second statement requires evidence that is rarely available in an individual clinical case.
Attachment Is More Useful as Behavior Than as a Label
A third study examined emotionally focused couple therapy with Spanish-speaking couples from Argentina, Costa Rica, Guatemala, Mexico, and Spain. Seventy couples were randomized to EFCT or a waiting list, with 68 couples entering the final analysis. The treatment averaged almost 19 sessions and measured accessibility, responsiveness, and engagement—the behavioral qualities sometimes summarized as secure attachment.
The results were meaningful but mixed. Participants receiving EFCT showed greater session-by-session improvement in how they perceived their partners’ secure attachment behavior, but not in how they rated their own behavior. In the pre–post analysis, the reliable treatment effect emerged only in women’s perceptions of their male partners. Improvements in perceived secure behavior were associated with improvements in relationship adjustment, but this association did not establish that attachment change caused the relationship improvement.
This is an important restraint on overly global descriptions of couple change. One partner’s experience that the other has become more accessible does not necessarily mean that both partners have changed equally, that each experiences the relationship similarly, or that the couple has developed a globally “secure attachment.” Attachment language can organize clinical attention, but it can also obscure asymmetry.
A more useful clinical inquiry may be concrete:
- Can I reach you when I am distressed?
- Do you remain emotionally present when my experience differs from yours?
- Does your response help me feel understood, or does it move too quickly toward correction?
- Can each of us remain engaged without requiring the other to surrender psychological authority?
- Who experiences the relationship as safer, and whose experience has not changed?
These questions shift attachment from a personality label to an observable relational process. They also help therapists distinguish mutual influence from equal responsibility. Both partners participate in a relational system, but they may not contribute equally to a particular injury, withdrawal, escalation, or failure of repair.
From Theoretical Allegiance to Observable Practice
Taken together, these studies do not discredit attachment-based or experiential therapy. They ask therapists to become more disciplined about the level of certainty attached to our explanations. Attachment, regulation, emotional processing, therapeutic alliance, client agency, and symptom relief may develop together while influencing one another in different directions at different phases of treatment.
The practical question is therefore not only, “What theory explains this client?” It is also, “What can we observe happening now?” A therapist may describe a session as attachment-focused, but did the client actually risk reaching for connection? The therapist may believe that affect was deepened, but did the client become more able to identify, tolerate, express, and make meaning from the emotion? The therapist may intend to promote agency, but did the client experience genuine choice?
This is particularly important in clinical supervision. Therapists are usually able to provide thoughtful accounts of what they intended to do. Intention, however, is not intervention, and intervention is not impact. Reviewing recordings, transcripts, or carefully reconstructed moments can reveal whether the therapist’s formulation became an observable clinical act—and how the client responded.
Supervision can therefore hold several questions at once:
- What did the therapist intend?
- What did the therapist observably do?
- What happened in the client immediately afterward?
- What alternative explanations fit the same sequence?
- What does the longer course of therapy suggest—but not prove—about the process of change?
This kind of inquiry is not designed to undermine therapeutic intuition. It gives intuition a more accountable foundation.
How Brian Kennedy Learning Approaches Change
Brian Kennedy Learning was developed around a similar distinction between understanding something conceptually and experiencing something differently. BKL is a clinically informed educational environment, not psychotherapy and not a substitute for psychotherapy. Its pathways translate attachment, experiential, systemic, parts-oriented, emotion-regulation, and depth perspectives into structured opportunities for reflection and experimentation.
The process begins with ordinary life rather than with a diagnostic conclusion. A learner notices a recurring moment: becoming defensive when a partner is disappointed, disappearing internally when someone becomes angry, taking responsibility for another person’s emotional state, or pushing for understanding until the other person retreats. The first task is observation—slowing the sequence enough to see what happens before deciding what it means.
BKL then offers a provisional framework for understanding the pattern. The stance is compassionate without becoming exculpatory: Of course this makes sense—and it may also be costing you something. Protective responses are dignified as adaptations before they are challenged. Understanding the childhood, relational, or cultural origins of a pattern does not remove responsibility for its present impact.
From there, the learner chooses a bounded experiment. The aim is not to perform the “correct” response but to discover what happens when one element of the sequence changes. A person might notice bodily activation before explaining, name a feeling before defending a position, ask a partner what was received rather than repeating what was intended, or remain present for several seconds longer without immediately fixing the other person’s distress.
The experiment is followed by actual experience and revised understanding. Did the feared outcome occur? Did the protection soften, intensify, or change form? Did the other person respond differently? Did the new behavior create greater freedom, or did it expose a limit that had previously been obscured?
This sequence—encounter, understanding, experiment, experience, and revision—reflects what the current research makes increasingly clear: change cannot be assumed from theory alone. A pathway may be informed by attachment research without claiming to repair attachment. An exercise may invite emotional awareness without claiming that emotion regulation is the mechanism producing recovery. A relational experiment may increase accessibility or responsiveness without proving that the relationship has become secure.
BKL’s assessments are therefore intended as reflection tools rather than diagnostic instruments. They help learners generate questions, recognize patterns, and identify possible experiments. They have not been independently validated as clinical measures, and the effectiveness of BKL itself remains to be demonstrated through future outcome research. That limitation is important because clinical credibility depends partly on saying clearly what we do not yet know.
The Therapist’s Task Is Larger Than Defending a Model
Every psychotherapy model offers a selective way of seeing. A useful model brings neglected experience into view; a dangerous model begins to explain everything. The therapist’s task is not to abandon theory but to remain able to observe when the client’s experience exceeds it.
The recent research suggests several practical commitments. Make the first session a meaningful encounter. Treat agency as participation rather than compliance. Do not assume that emotion regulation or attachment change necessarily precedes symptom relief. Assess secure functioning through concrete behavior and through each person’s separate experience. In supervision, distinguish intention, intervention, impact, and causal explanation.
Most importantly, remain curious about the sequence of change. Therapy may help people understand their adaptations, encounter previously avoided experience, act with greater freedom, and build more responsive relationships. We can value all of these developments without pretending that we know more about their causal order than the evidence allows.
That stance is not uncertainty for its own sake. It is a form of clinical respect—for the complexity of change, for the limits of our models, and for the client’s right to remain larger than our explanation.
Clinical Foundations and Further Reading
Andrade-Urbina, D., Rodríguez-González, M., Cueli-Naranjo, M. A., Osorio, A., & Sandberg, J. G. (2026). Attachment behavior change following emotionally focused couple therapy: A randomized controlled trial with Spanish-speaking couples. Journal of Marital & Family Therapy, 52(4), e70173. https://doi.org/10.1111/jmft.70173
Egli, S., Badenbach, T., van Emmerik, A., Stefan, A., OPTIMA Study Group, & Kopf-Beck, J. (2026). Mechanisms of change in the treatment of depression: On the role of attachment, reappraisal and suppression. Clinical Psychology & Psychotherapy, 33(5), e70345. https://doi.org/10.1002/cpp.70345
Malins, S., Saxon, D., Figueredo, G., Clos, J., Trimble, T., Fatehi, K., Waldram, D., Boycott, N., Frost, A., Sales, C., Horton, K., Hardy, G. E., Barkham, M., Morriss, R., Schröder, T., & Moghaddam, N. (2026). Predicting patient outcomes from initial in-session interactions within cognitive behavioral therapy and person-centered experiential therapy for severe depression. Behaviour Research and Therapy, 205, 105152. https://doi.org/10.1016/j.brat.2026.105152
Evidence note
The studies discussed here support careful clinical hypotheses, not definitive conclusions about how psychotherapy works. The first-session study was an observational process analysis within a randomized trial; the EFCT study used random assignment but did not establish attachment behavior as a causal mediator; and the depression-mechanism study was a secondary analysis conducted in an intensive multimodal setting. None establishes a specific neural mechanism for experiential or attachment-oriented psychotherapy.