Psychodynamic Therapy vs CBT: How to Choose

Somebody has told you to get therapy, or you have decided it yourself, and within ten minutes of searching you have run into an acronym problem. CBT. DBT. ACT. EMDR. Psychodynamic. Psychoanalytic. Integrative. Every directory listing seems to claim several at once, and none of the descriptions tell you what would actually happen if you sat down in the room.

I am a board-certified psychoanalyst in Washington, DC, and I have practiced and taught this work for over fifty years. I am not a neutral party — my own practice is psychoanalytic, and you should read what follows knowing that. But the honest position is not the one you might expect from a psychoanalyst: cognitive behavioural therapy is a serious treatment that helps a great many people, and for certain problems I would send you to a CBT therapist before I would take you on myself.

What follows is what I would tell a friend who asked me how to choose. Not which therapy is better — that question has no answer — but what each is aiming at, which problems each is shaped for, and the questions worth asking before you commit your money and your Tuesday evenings.

Written by: John J. Hartman, Ph.D., board-certified psychoanalyst (FABP), licensed psychologist in DC, Florida and Michigan · Last reviewed: 10 September 2026

This article is general information, not medical advice. Speak with a physician or qualified clinician about your own situation.

What you are actually choosing between

Start with the thing the directories obscure. Psychotherapy is not a single product sold in different flavours. The various schools disagree about something substantive: what a symptom is.

To a cognitive behavioural therapist, a symptom is a maintained pattern. Something set it off — that part is not the main interest — and it now persists because of a loop running in the present. You avoid the thing you fear, the avoidance brings relief, the relief teaches you that avoidance was necessary, and the fear is fed rather than starved. Break the loop and the symptom loses its supply.

To a psychodynamic therapist, a symptom is a compromise. It is doing something for you, at a cost you did not consciously agree to. The panic that arrives before a promotion may be protecting you from a success that would mean surpassing someone. The flatness that settles over a good relationship may be holding down something you are not permitted to feel about it. On this account the loop is real but downstream: interrupt it and something else tends to take its place, because the thing generating it has not been addressed.

Those are genuinely different theories, and they lead to genuinely different rooms. It matters which room you are walking into, because the two ask different things of you and count different things as progress.

A note on the words themselves

Psychodynamic therapy and psychoanalytic psychotherapy are, in ordinary use, the same treatment: individual therapy, usually once or twice a week, working on the conflicts underneath a symptom. Some clinicians reserve “psychoanalytic” for work by someone with formal analytic training — worth asking about, though the method in the room is continuous.

Psychoanalysis proper means higher frequency — usually three or more sessions a week — with a trained analyst, over years rather than months. It is the same method applied with enough intensity that patterns start to happen inside the treatment rather than only being reported to it. I describe how that works on my page about online psychoanalysis in Washington, DC.

CBT is itself a family, not one thing. Classical cognitive therapy, behavioural activation for depression, exposure and response prevention for obsessive-compulsive difficulties and the third-wave treatments such as acceptance and commitment therapy all sit under the heading and differ considerably. “I do CBT” is about as specific as “I cook European food.”

What CBT actually is, in the room

A well-run course of cognitive behavioural therapy is structured, collaborative and finite, and those three features are deliberate.

In the first session or two the therapist takes a history and then does something distinctive: together you build a formulation, a working diagram of your particular problem. What situations set it off, what goes through your mind, what you feel in your body, what you then do, and what that doing achieves in the short term and costs in the long term. The formulation is written down, shown to you, and open to your correction.

From there the work is organised around targets. If the difficulty is panic, you might spend sessions deliberately provoking the physical sensations you fear, to learn from experience that they are not the catastrophe you take them for. If it is depression, you might start with behavioural activation: scheduling small activities before you feel like doing them, on the principle that motivation follows action rather than preceding it. If it is a set of harsh beliefs about yourself, you might spend weeks catching those thoughts as they occur and testing them against evidence.

Between sessions there is homework, and it is not optional decoration. In most forms of CBT the between-session work is where the change happens; the session is where it is planned and reviewed. A course typically runs somewhere in the range of eight to twenty sessions, with a planned ending, and progress is often tracked with brief questionnaires.

What this is good at is hard to overstate. It is transparent, teachable, and fast relative to the alternatives. And for a class of problems it is very well matched indeed: specific phobias, panic disorder, obsessive-compulsive difficulties, and a good deal of straightforward anxiety and depression. I have written separately about what those conditions involve in my complete guide to anxiety, its symptoms and treatment and my guide to what actually helps with depression.

Where CBT tends to run into trouble

The limits show up in predictable places, and any honest CBT therapist will name them.

The first is the person who does the work impeccably and does not get better. They complete the thought records. The evidence against the belief is overwhelming and they can recite it. They still feel worthless. That usually indicates the belief is not really a belief: it is a position being held for a reason, and the reason has not been touched.

The second is the problem that will not sit still long enough to be targeted. “Nothing is wrong. I have a good job and a good marriage and I feel like I am watching my life from behind glass.” There is no loop to diagram there yet.

The third is the recurring pattern in relationships. Structured treatment can address how you behave in an argument; it has a harder time with why you keep choosing the argument, or the person — a difficulty I see often enough to have given it a page of its own on therapy for relationship conflicts.

The fourth is difficulty that is characterological rather than symptomatic — trouble that is not something you have but something you seem to be. Structured protocols exist for some of this, but the timescale stops being twelve weeks.

What psychodynamic therapy actually is, in the room

The first thing people notice is the absence of an agenda, and it is disconcerting. No formulation handed across the desk, no worksheet, no obvious plan. You are asked to say what comes to mind and to try not to edit it into a report. It is worth understanding what the apparent aimlessness is for.

Deliberate reflection filters. When you decide in advance what is relevant, you have already excluded the material that does not fit your account of yourself — which is precisely the material that would revise it. The instruction to speak without pre-selecting is a way of getting underneath your own editing. What surfaces is often trivial-seeming, and the triviality is the point: an irritation at a colleague, a fragment of a dream, an odd blankness about a subject that ought to matter. Left alone, these accumulate into a shape.

The therapist's job is to listen for that shape, and particularly for the thing being circled without being said. Over months, themes recur. The same structure turns up in a memory, in an account of your marriage, in something that happened at work on Tuesday, and eventually in how you are with the therapist — the appointment you forget, the compliment you deflect, the sense that you must be an interesting patient or you will be dropped.

That last one deserves a plain explanation, because it is the part of the method most often caricatured. The relationship with the therapist becomes, over time, a live specimen of the pattern causing you trouble elsewhere. Not a metaphor for it — an instance of it, occurring where it can be examined while it is happening rather than reconstructed afterwards from an unreliable account. If you are somebody who expects to be a disappointment, you will eventually expect it of the therapist too, and then there is something in the room to work with rather than a story about your father.

The therapist's other main tool is interpretation, which in practice is far more modest than the word implies. It is usually a small observation offered tentatively, at a moment when you are close to seeing it yourself: you have changed the subject twice now, both times just after mentioning your brother. The value is not the therapist's cleverness. It is your being able to see, from outside, a move you make constantly and have never watched yourself make.

Where psychodynamic therapy tends to run into trouble

The limits here are equally real.

It is slower, and there is no honest way around that. Symptom relief often arrives earlier than people expect — frequently within the first months — but the change that makes the relief durable takes considerably longer.

It is harder to evaluate from inside. Without symptom scores and session targets, it can be difficult to tell deep work in progress from a treatment that has quietly become a habit. The remedy is structural: a periodic, explicit review of whether this is still the right instrument. A therapist unwilling to have that conversation has told you something.

It asks for a tolerance of ambiguity that not everyone has to spare, particularly in acute distress. Somebody who cannot leave the house needs to be able to leave the house; insight into why can come later. And it is not the right first move for certain presentations — active suicidal crisis, an eating disorder at a dangerous weight, untreated psychosis, substance dependence still in full flight all need stabilisation first, from someone equipped to provide it.

The differences that matter

Most comparison tables you will find online list surface features. These are the differences that actually change your experience of the treatment.

The first two rows cause the most avoidable disappointment. People arriving in psychodynamic therapy from structured treatment often spend a month waiting to be told what to do, and read the absence of instruction as the therapist not doing their job; people arriving in CBT from open-ended work sometimes feel the agenda is being managed rather than heard. Both reactions are about the frame rather than the therapist's competence, and both are worth saying out loud early.

What the research can and cannot tell you

Here I want to be careful, because this is where the internet is least trustworthy in both directions. Enthusiasts on either side will tell you the evidence settles it. It does not, and understanding why is more useful to you than any figure I could quote.

What can be said with reasonable confidence: both cognitive behavioural and psychodynamic psychotherapies are established treatments with a body of research behind them, and both appear in mainstream summaries of the psychotherapies for common conditions such as anxiety and depression. The National Institute of Mental Health overview of the psychotherapies and the American Psychological Association's introduction to psychotherapy are sober starting points.

What cannot be said is that a trial result tells you what will happen to you. Here is why.

• Trials measure averages, and you are not an average. A finding that treatment A beats treatment B by a modest margin across two hundred people is entirely compatible with B being much better for a substantial minority, including possibly you.
• Trials measure what is measurable in the time available. A study running twelve weeks cannot in principle detect a benefit that emerges at eighteen months. That is a limitation of the instrument, and it cuts in a specific direction.
• Trial conditions are not clinic conditions. Participants are screened for a single clean diagnosis. The people who arrive in my consulting room usually have several difficulties at once, plus a marriage and a job and a history.

And the therapy is not the only active ingredient. One of the more robust findings across the whole field is that the quality of the working relationship accounts for a meaningful part of the outcome regardless of school. That is a finding you can use: your sense, in the first few sessions, that this person grasps what you are telling them is arguably better evidence about your prospects than the modality label.

If you must choose on evidence, the most useful evidence available to you is not a published effect size. It is whether, after three or four sessions, you find yourself saying things in that room you have not said elsewhere.

Which approach fits which problem

With those caveats standing, some matching is sensible. This is not a diagnosis and cannot substitute for a conversation with a clinician who has heard your history — but if you are trying to orient yourself, this is roughly how I think about it.

Where I would point you towards structured, skills-based treatment

• A specific phobia. Flying, needles, dogs, lifts. Graded exposure is the treatment of choice and often remarkably efficient. Understanding the origin of the fear is interesting and unnecessary.
• Panic attacks as the central problem. When the difficulty has become a fear of the panic itself, the loop is doing most of the work and breaking it is the priority.
• Obsessive-compulsive difficulties. Exposure and response prevention is a specialised technique that should be delivered by someone properly trained in it. This is a case where I would refer out without hesitation.
• Insomnia as a standalone problem. There is a dedicated cognitive behavioural protocol for insomnia, and it is a good use of six sessions.
• A first, situationally comprehensible episode. Low mood or anxiety that started with something identifiable — a bereavement, a redundancy, a move — in someone whose life has otherwise gone reasonably well.
• When you need to function next month. A dissertation to finish, a court date, a job to hold. Stabilise first; the deeper question keeps.

Where I would point you towards psychodynamic work

• The same difficulty has come back after previous treatment. This is the single clearest indication. Something helped, and then the same mood or the same impasse returned. That is not a failure of effort; it usually means the treatment reached the surface and not what was generating it.
• You understand your problem perfectly and remain inside it. You can name the pattern, trace it to its origins, predict your next move — and you make it anyway. Insight at that level has already been tried.
• The pattern is relational and it follows you. The fourth relationship with the same arc; the fourth boss unreasonable in the same specific way. When the constant across situations is you, the work needs to look at that without it becoming an accusation.
• The symptom is vague, diffuse or hard to name. Emptiness, meaninglessness, a life that looks correct from outside and feels borrowed from inside. There is no target to aim structured treatment at, and the vagueness is itself the material.
• Self-esteem that does not respond to evidence — achievements that are real and unavailable to you, a critical inner voice that survives every disconfirmation. More on therapy for low self-esteem.
• Long-standing difficulties that feel like character rather than illness — patterns of relating that have been present as long as you can remember. See treatment for personality disorders.
• Sexual difficulties with no medical explanation, where desire, inhibition and the meaning of intimacy are entangled. See therapy for sexual issues.
• Work and school trouble that does not match your capability — procrastination that is not laziness, achievement that never registers, ambition that stalls short of arriving. More on therapy for work and school problems.

Where the honest answer is “it depends more on the person than the problem”

Ordinary anxiety and ordinary depression — most of what brings people to therapy — sit in a genuinely open middle. Both approaches treat both conditions. If your anxiety is recent, well-defined and behaving like a loop, I would try the structured route first; if it is long-standing, diffuse and has already survived a course of treatment, I would not. If neither description fits, the deciding factor is reasonably which room you can imagine speaking freely in — not a soft consideration, given what the research says about the working relationship.

And if what you are dealing with is a relationship rather than a symptom, the choice is different again: whether the work should be individual at all, or joint. I have written about that decision in couples therapy vs marriage counseling, including when a joint session is the wrong first step.

The question almost nobody asks: how often

Prospective patients research modalities exhaustively and then accept whatever frequency they are offered, usually weekly, as though it were a law of nature. It is not, and it may be the most consequential variable in the whole decision.

Consider what a week does. You arrive, spend fifteen minutes reporting the week's events, reach something live at minute forty, and stop. Then seven days pass and whatever had been stirred up settles. For many problems this rhythm is perfectly adequate, and for some it is preferable — distance can be exactly what a person needs.

But for a certain kind of difficulty, weekly frequency structurally prevents the work. If the pattern you need to examine only becomes visible when it is happening in the room, and the room resets to formality every seven days, you may spend years circling something you never get close enough to see. Twice weekly changes this noticeably; three or more times weekly — which is what psychoanalysis means — changes it categorically, because continuity stops being something you maintain by effort.

This is why I offer both once- or twice-weekly psychoanalytic psychotherapy and full analysis at higher frequency rather than treating them as different products. For a great many people weekly work is the better trade: it reaches far enough and it is sustainable, which a treatment nobody can keep up is not.

What matters is that frequency gets discussed rather than defaulted. It is a fair question to ask any therapist: given what I have described, is once a week enough for this?

When you have already tried therapy and it did not hold

A large share of the people who contact me have been in therapy before, often more than once. They tend to arrive apologetic, as though the failure were theirs, and with a private theory that they are simply not somebody therapy works on.

That conclusion is almost always premature, because “it did not work” covers at least five different situations with different remedies.

• The wrong instrument. A short, structured treatment was applied to a long-standing characterological difficulty, or open-ended exploratory work to an acute phobia. Each would have been right for the other problem. This is the commonest version and the most fixable.
• The wrong dose. The approach was right and the frequency too low to reach what needed reaching, so the treatment became a pleasant weekly conversation that did not move. Sometimes the same therapist at twice the frequency is the whole answer.
• The wrong person. You did not feel understood by that particular clinician and stayed anyway out of politeness — a common and entirely human mistake. Fit is not a luxury, and it is legitimate to change.
• The wrong moment. You went because someone else wanted you to, or during a period when your energy was consumed by surviving. The same treatment at a different point in your life is a different treatment.
• It did work, and then life continued. The treatment did what it was for, held for years, and new circumstances reactivated an old vulnerability. That is not failure; it is how most chronic difficulties behave, and returning is not starting from zero.

Whichever applies, it is worth reconstructing with any new clinician: what you tried, for how long, at what frequency, and what happened when it stopped helping. That history is diagnostic information, and a good clinician will want it in detail.

Combining approaches, and where medication fits

The framing of this article — one versus the other — is a convenience for thinking, not a description of how care actually proceeds. Three qualifications matter.

Most experienced clinicians are less pure than their labels

A psychodynamic therapist whose patient cannot get out of bed will talk about getting out of bed. A cognitive behavioural therapist working with someone whose self-criticism turns out to be in their mother's voice will not pretend not to have noticed. What distinguishes practitioners is the primary lens they think through, not a refusal to use anything else.

Sequencing is often better than mixing

Running two therapies at once with two therapists more often diffuses both, and gives you somewhere to take material away from wherever it is inconvenient. Sequencing is cleaner: stabilise an acute symptom with focused work, then, if the underlying question is still standing, take it up in open-ended treatment. Many people come to analytic work by that route.

Medication is a separate axis

Whether to take an antidepressant or an anxiolytic is a question for a physician or psychiatrist, and it is independent of which psychotherapy you choose. Medication does not disqualify you from either treatment, and neither obliges you to take or avoid it. My guide to dealing with depression covers the landscape of treatment options in more detail; decisions about your own prescription belong with the clinician who can examine you.

Six questions to ask before you book

Directory profiles are close to useless here, because the modality checkboxes are self-selected and cost nothing to tick. Almost all clinicians will speak to you briefly first. These are the questions worth spending that call on.

1. What were you actually trained in, where, and for how long?

This separates a checkbox from a competence. Formal training in either tradition is substantial — a psychoanalytic training runs for years and includes supervised cases and, in most institutes, the trainee's own analysis; proper CBT training involves supervised practice, not a weekend course. The willingness to answer plainly is itself informative.

2. Given what I have described, is this the right treatment for it?

Listen for whether the answer contains any conditions. A clinician who tells you their approach suits everybody has told you something, though not what they intended.

3. How often would you suggest we meet, and why that often?

The reasoning matters more than the number. “Weekly is what I do” is a scheduling answer. “Weekly to start, and if we spend each session catching up rather than working, we should discuss twice” is a clinical answer.

4. How will we know whether this is working, and when will we check?

Structured treatment will point to targets and measures; open-ended treatment should point to a periodic review and be able to say what early change tends to look like. What you do not want is no answer at all.

5. What happens between sessions?

Homework every week, or nothing at all, are both legitimate — but mismatched expectation here is a common reason people leave treatments that would have suited them.

6. If this is not working in a few months, what then?

The answer you want indicates willingness to reconsider, including a referral elsewhere. Anyone who cannot imagine the treatment failing has not been doing it long enough.

What progress looks like in each

Knowing what you are watching for prevents two opposite errors: abandoning something that is working because it does not feel like working, and continuing something that is not.

In structured treatment

Early progress is usually visible and often measured. The panic attacks reduce in frequency. You do the thing you were avoiding. The questionnaire score moves. Its limitation is worth knowing too: gains made this way can fade after treatment ends, which is why relapse-prevention planning is a standard part of a well-run course rather than an afterthought. If your therapist has not raised it by the closing sessions, raise it yourself.

In psychodynamic treatment

Progress arrives in a different order and is easier to miss because it does not announce itself.

• First, the symptom eases — often earlier than people expect, sometimes before anything has been understood. Sleep improves; the background dread thins.
• Then reactions become legible. Feelings that used to arrive from nowhere acquire a history and a trigger you can see coming.
• Then the gap widens between feeling and acting. The impulse to send the email arrives and there is now a space before the sending. This is often the first change other people notice.
• Then choices stop feeling compelled. The familiar argument, the familiar retreat, the familiar kind of partner stop presenting themselves as the only available move. This is the change that outlasts treatment, and the slowest to arrive.

Throughout, what you can say in the room expands. If, six months in, you are saying things you could not have said at the start — including critical things about the therapy itself — the treatment is doing what it is supposed to do, whatever the symptom score says.

Signs worth taking seriously in either treatment

• Several months in with no change of any kind, and no review having been proposed.
• Sessions that have become companionable and comfortable, with nothing difficult in them for a long stretch.
• A sense that you are managing the therapist — supplying good material, avoiding subjects that would trouble them.
• Feeling consistently worse in a way that is not the temporary sharpening that sometimes accompanies real work, and being unable to say so.

Every one of these is something to raise in the session itself. In psychodynamic work particularly, raising it is the work.

What each costs, in money and in time

The arithmetic is not incidental to the choice, and vague answers here do nobody any favours.

A focused cognitive behavioural course is a bounded commitment: a defined number of weekly sessions with an ending in view, which makes the total calculable in advance. Because these treatments were built around measurable targets, they also fit insurance reimbursement more comfortably than open-ended work does.

Open-ended psychodynamic treatment cannot be costed the same way, and anyone who quotes you a total is guessing. What can be said honestly is the per-session fee, the likely frequency, and that the timescale is months to years rather than weeks. In my own practice, sessions are fifty minutes at $325, the first fifteen-minute consultation is free, and a sliding scale exists — it is worth asking whether you are eligible. I do not bill insurance directly, though if your plan covers out-of-network mental health care you may be able to submit for reimbursement; the fees and insurance section sets out the details.

Two things are worth weighing against the sticker figure. A treatment that does not reach the problem is not cheap merely because it was short; several such courses across a decade add up. And the reverse: open-ended treatment continued out of habit is an expensive way to have a weekly conversation. Both errors are avoided by the same discipline of periodic, explicit review.

Frequently asked questions

Conclusion

The choice between cognitive behavioural and psychodynamic therapy is real, but it is a different shape from the one the internet describes. It is not a contest between a modern treatment and an old one, and it is not settled by an effect size. It is a question of what your difficulty actually is: a loop running in the present that can be interrupted, or a compromise being maintained for reasons you have not yet had access to.

If your problem is recent, well-defined and behaving like a loop, look for someone properly trained in structured treatment and start there. If it is long-standing, if it has already come back after therapy that helped for a while, or if you can explain it perfectly and cannot move — that is a different problem, needing an instrument that reaches further. Whichever you choose, the two things most worth protecting are your willingness to say plainly when something is not working, and your clinician's willingness to hear it.

If you would like to talk through which kind of help fits your situation, you can book a free fifteen-minute consultation or call (813) 787-2048. I work online with adults and adolescents, and I will tell you honestly if I think someone else is the better fit. If you would rather read more first, my pages on weekly psychoanalytic psychotherapy and psychoanalysis describe how each treatment actually goes, and the rest of my writing is collected in Insights.

---

About the author

John J. Hartman, Ph.D. is a board-certified psychoanalyst (FABP) with over 50 years of clinical experience. He holds a B.A. from Harvard College and a Ph.D. in clinical psychology from the University of Michigan, with psychoanalytic training at the Michigan Psychoanalytic Institute. A former tenured Associate Professor of Psychiatry at the University of Michigan, he is a training and supervising analyst and has published six books and numerous peer-reviewed papers. He practices online and is licensed in DC, Florida and Michigan, with PsyPACT authorization across more than 40 states.

References and further reading

• National Institute of Mental Health. Psychotherapies.
• American Psychological Association. Understanding psychotherapy and how it works.
• National Institute of Mental Health. Help for Mental Illnesses.