The Honest Truth About How Therapy Actually Works

By Felix Murad, M.Ed., LPC-S, Licensed Professional Counselor-Supervisor (Texas & Washington)

What the outcome research shows new clinicians, and what grad school quietly left out.

Graduate school hands you a clean story. Choose the evidence-based protocol, deliver it with fidelity, and the client improves. Match the diagnosis to the manual and the rest is execution.

The room does not read the manual.

I’ve spent fourteen years in the chair and a good chunk of that supervising people new to it, and I’d rather you hear the honest version now than reverse-engineer it around hour 2,000. None of what follows is cynical. Most of it is in the literature you were assigned and somehow never quite got told. Taken together, it changes what you should be working on this year.

It’s the relationship, more than you were led to believe

The single most replicated finding in psychotherapy research is unglamorous: the quality of the alliance between client and clinician predicts outcome, across orientations, fairly consistently. The largest synthesis to date, 295 samples, more than 30,000 patients, puts the alliance-outcome correlation at roughly r =.28, accounting for around 8% of the variance in whether people get better (Flückiger et al., 2018). That is a fact, and by the standards of this field it is a large, durable effect.

Here’s where new clinicians misread it. "The relationship matters most" is not permission to be warm and aimless. The alliance, as it’s actually measured, isn’t rapport or being liked. Bordin (1979) defined it as three things working at once: agreement on the goals, agreement on the tasks, and the bond that holds while you do hard work together. Two of those three are about competent direction. The clients who feel most safely held are usually the ones who can tell you know where you’re going. Warmth without a map doesn’t build alliance. It builds a pleasant stall.

Who you are matters more than which brand you practice

This is the finding that reorganizes how you should think about your own development, so sit with it.

When researchers measure how much outcome variance is explained by which specific treatment a client receives versus which therapist delivers it, the therapist usually wins. Across the literature, therapist effects average about 5% of outcome variance, and the most effective therapists are roughly twice as effective as the least effective (Baldwin & Imel, 2013; Johns et al., 2019). In one well-known reanalysis of a major depression trial, about 8% of the variance was attributable to the therapist and roughly 0% to the particular treatment delivered (Baldwin & Imel, 2013).

Read that again, because it has a direct consequence for you: you are the instrument. Not the protocol you memorized, you, and how skillfully you run it. The brand on the door of your training ("I’m a CBT therapist," "I’m EMDR-trained") tells a client far less than they think about whether they’ll improve. Which means the hours you’re about to log aren’t the point. What you do with them is. Two Associates can complete identical 3,000-hour timelines and come out clinicians of very different quality, and the difference won’t be the modality on their LinkedIn.

And yet, technique is not interchangeable, and pretending otherwise is how people get hurt

Now the counterweight, because the common-factors literature gets badly abused, and your future clients pay for the misreading.

"The relationship explains a lot of the variance" does not mean "any approach works as well as any other." That conclusion is averaged across mixed presentations, and the average hides the cases where method is decisive. For obsessive-compulsive disorder, exposure and response prevention, built on an inhibitory learning framework, reliably outperforms supportive, insight-oriented talk (Craske et al., 2014). For PTSD, trauma-focused protocols. For specific phobia, exposure. In these conditions the technique is not a flavor preference; it’s the active ingredient.

Here’s the part that should make you uncomfortable, in the useful way. If you take an OCD client and offer them a warm, validating, relationship-only space, reassuring them, exploring the feelings, avoiding the exposures because they’re distressing, you are not being kind. You are very likely reinforcing the avoidance that maintains the disorder. Good alliance plus the wrong method can still harm. The honest synthesis isn’t "relationship or technique." It’s that the relationship is the vehicle and the right method is the road, and you need both pointed the same direction.

This is, incidentally, why "I just meet the client where they are" is not a treatment plan. With that said, which method, run with enough skill to actually work, is a question you’ll spend your whole career answering. That’s the craft.

Therapy fails more often than the brochures admit

You will not be told this at orientation, so I’ll tell you. Roughly one in five clients drops out of treatment prematurely, the meta-analytic rate is 19.7% across 669 studies and nearly 84,000 clients (Swift & Greenberg, 2012). Separately, a meaningful minority don’t just fail to improve, they deteriorate; Lambert’s work puts that figure around 5-10%, and a substantial share of clients in trials simply don’t improve at all (Lambert, 2013).

New clinicians do one of two unhelpful things with these numbers. They either personalize all of it ("I’m failing everyone") or none of it ("clients just aren’t ready"). The honest stance is in between and harder: some of the failure is your developing skill, some is fit, and some is the genuine limit of the work. You cannot tell which is which by vibes. You find out by measuring, routine outcome monitoring catches the client who’s quietly getting worse while nodding politely, and it catches it early enough to change course (Lambert, 2013). Most Associates don’t measure. Be the one who does.

So what actually moves the needle, for you, this year

Pulling it together into things you can act on, with the clear caveat that this section is my clinical judgment built on the evidence above, not a citation for each sentence:

Build the alliance on purpose, not by accident. Get explicit agreement on goals and tasks early. When a client disengages, treat it as a repairable rupture, not a verdict on your likability.

Train yourself like the instrument you are. Therapist effects don’t shrink with raw hours, clinicians don’t reliably improve just by accumulating caseload. They improve through deliberate practice: isolating a specific weak skill, drilling it, getting feedback, repeating. Bring recordings to supervision. Ask to be corrected on the thing you’re worst at, not praised for the thing you’re already good at.

Match the method to the disorder, then run it competently. Know which conditions have a decisive evidence-based treatment and don’t improvise around them to spare anyone discomfort, including your own.

Tolerate your own distress. The reason new clinicians avoid exposures, skip the hard feedback, and soften interventions is usually their own anxiety, not the client’s fragility. The same exposure logic you’ll use with clients applies to you: you get more capable by approaching what you’d rather avoid.

Measure, so reality can correct you. Feedback you don’t collect can’t make you better.

Why grad school told you the cleaner version

Not a conspiracy, a limitation. Protocols are teachable, testable, and licensable; you can put "fidelity to the manual" on a rubric. The messier truths, that the clinician is the active ingredient, that alliance is a skill, that some of this work fails, are harder to manualize and harder to grade, so they get compressed into a footnote. Programs optimize for what they can certify. The rest, you mostly learn in the years right after, ideally next to someone who’ll tell you the truth. That’s most of what supervision is for.

Frequently asked questions

Does this mean the modality doesn’t matter?

No. Averaged across mixed presentations, who delivers treatment often explains more variance than which treatment it is (Baldwin & Imel, 2013), but for specific conditions like OCD, PTSD, and phobias, the right protocol is decisive (Craske et al., 2014). Both are true. Don’t use the first to excuse skipping the second.

How do I actually get better instead of just accruing hours?

Deliberate practice plus feedback plus supervision that pushes on your weak spots. Raw caseload volume is not a development plan; reps on hard skills, observed and corrected, is.

Should I track outcomes as an Associate?

Yes. Routine outcome monitoring is one of the few low-cost things shown to improve results, largely by flagging the clients who are deteriorating before it’s too late to adjust (Lambert, 2013).

Before you take your next session, sit with these

See how I actually supervise

If this is the lens you want on your development, supervision that treats you as the instrument worth sharpening, drills the relationship and the method instead of just signing your hours, and uses real feedback instead of vibes, that’s the whole basis of how I work with Associates.

New to all this? Start with what to know before you begin LPC Associate supervision, then how to choose the right supervisor.

Felix Murad, M.Ed., LPC-S, LMHC, CMHC, NCC, Licensed Professional Counselor-Supervisor. Licensed by the Texas Behavioral Health Executive Council / Texas State Board of Examiners of Professional Counselors. Licensed in Texas, Washington, New Hampshire, and Florida (telehealth). This article is general professional education for current and prospective clinicians and is not clinical or legal advice.

References

Baldwin, S. A., & Imel, Z. E. (2013). Therapist effects: Findings and methods. In M. J. Lambert (Ed.), Bergin and Garfield’s handbook of psychotherapy and behavior change (6th ed., pp. 258-297). Wiley.

Bordin, E. S. (1979). The generalizability of the psychoanalytic concept of the working alliance. Psychotherapy: Theory, Research & Practice, 16(3), 252-260.

Craske, M. G., Treanor, M., Conway, C. C., Zbozinek, T., & Vervliet, B. (2014). Maximizing exposure therapy: An inhibitory learning approach. Behaviour Research and Therapy, 58, 10-23.

Flückiger, C., Del Re, A. C., Wampold, B. E., & Horvath, A. O. (2018). The alliance in adult psychotherapy: A meta-analytic synthesis. Psychotherapy, 55(4), 316-340.

Johns, R. G., Barkham, M., Kellett, S., & Saxon, D. (2019). A systematic review of therapist effects: A critical narrative update and refinement to Baldwin and Imel’s (2013) review. Clinical Psychology Review, 67, 78-93.

Lambert, M. J. (2013). The efficacy and effectiveness of psychotherapy. In M. J. Lambert (Ed.), Bergin and Garfield’s handbook of psychotherapy and behavior change (6th ed., pp. 169-218). Wiley.

Swift, J. K., & Greenberg, R. P. (2012). Premature discontinuation in adult psychotherapy: A meta-analysis. Journal of Consulting and Clinical Psychology, 80(4), 547-559.

Wampold, B. E., & Imel, Z. E. (2015). The great psychotherapy debate: The evidence for what makes psychotherapy work (2nd ed.). Routledge.