How to find a couples therapist who was trained for it

Two sentences from a professor of family social science explain most of what goes wrong here, and they are twenty years old. Nothing that ranks for this question repeats them.

Surveys indicate that about 80 percent of therapists in private practice do couples therapy.

William Doherty, “Bad Couples Therapy”, Psychotherapy Networker

Most therapists practicing today never took a course in couples therapy and never did their internships under supervision from someone who’d mastered the art.

the same article

Read them together. Almost everyone in private practice offers this, and most of them were trained to sit with one person. Doherty also notes that no licensing profession in the United States — psychology, social work, counselling, psychiatry — requires a course in marital therapy to qualify.

Which means the thing you would naturally use to choose — the letters after the name — tells you almost nothing about the question you are actually asking. A psychologist, a clinical social worker and a licensed marriage and family therapist can all be excellent with two people in a room, and all three can have arrived there with no training in it at all. The one exception is worth knowing: LMFT is the only one of the common licences whose training is relational by design, and even that is not a guarantee.

The two directories that mean something, and the one that does not

There are exactly two free, public directories that tell you a therapist was trained specifically to work with couples.

• The Gottman Referral Network. Everyone listed has completed at least Level 2 training in the method — a four-day programme with Level 1 as a prerequisite. Therapists who have gone all the way through certification are marked in gold on the directory, which gives you a checkable label rather than a claim.
• The ICEEFT directory, for Emotionally Focused Therapy. The floor is completion of an EFT externship. ICEEFT also says something the others do not: its directory is not a credential check, and you should verify the licence itself with the state board. That is good advice about every directory, including its own.

And then there is Psychology Today, which is where nearly everyone starts. It is worth being clear about what a listing there is: a paid advertisement, around thirty dollars a month, taken out by the therapist. The verification is real but narrow — it confirms that the person holds a valid licence with no restrictions on it. The specialties on the profile, including “Relationship Issues”, are a list the therapist ticks for themselves.

What it costs

Advertised rates rather than measured data, gathered from practice and price-guide pages in 2026. Two things push the number up that people do not expect: certification in a named method, and length — couples sessions often run ninety minutes rather than the fifty-minute individual hour.

The training clinic line is the one worth pausing on. It looks like the budget option and it is, but supervision is a real quality mechanism and not merely a discount: a graduate clinician being watched closely by someone experienced is a different proposition from a licensed therapist working alone on a kind of case they were never taught.

And the genuinely free route almost nobody uses: your employer’s assistance programme. These typically carry three to eight free sessions per issue per year, and relationship counselling usually sits inside the same allowance. It costs one email to find out.

Why insurance will not pay, spelled out

Every practice page has an “Insurance & Rates” link and almost none of them answers this. The mechanism is not complicated, and it has a consequence you should see before you agree to anything.

1. The session is billable. There is a code for couples and family psychotherapy with the patient present.
2. But payment turns on the diagnosis code, not the session code.
3. Relationship distress has a code — Z63.0, “problems in relationship with spouse or partner”. It is a Z-code, which describes a circumstance of life rather than an illness.
4. Insurers will not accept a Z-code as the primary diagnosis. They want a mental disorder listed first.
5. And there is no billable entity that corresponds to us. A couple is not a patient.

So for a claim to be paid, one of you has to be made the patient. One partner receives a psychiatric diagnosis, and the couples work is billed as part of treating that person’s disorder.

Does it work — the honest version

Yes, in trials. The numbers you will see quoted, usually something around seventy per cent, come from efficacy studies of Emotionally Focused Therapy and the Gottman Method, and those are real studies. Two things about them are usually left off.

They were mostly authored by the people who developed the methods. And they ran long — twenty-five sessions and up — with couples who were screened in.

What happens in ordinary practice is different, and there is a paper about exactly this gap.

That is not an argument against going. It is an argument for going in with your eyes open, and three things follow from it that nobody tells you.

• You are allowed to check. Most couples who will not benefit are detectable by about the fourth session, by the simple fact that nothing has changed. Therapists who use a standard measure and review it with you catch deterioration far earlier than those who go on impression.
• Arriving in worse shape is not the problem you think. In the meta-analytic work, greater distress at the start predicted larger gains — the reverse of the urgency you are usually sold.
• And you have probably not left it too late. The widely repeated line that the average couple waits six years is not well supported. When Doherty’s group actually measured it, the average was 2.68 years, and the great majority arrived within two.

It can also do harm, and the harm has been counted

This is the part that does not appear anywhere in the advice about choosing a therapist, and it is the best reason to ask the questions below rather than just booking someone with a nice photograph.

Doherty’s group surveyed 270 couple-therapy clients about five specific things a therapist might have said: that the two of them were incompatible, that the therapist could not help them, that the relationship was beyond repair, that divorce was the best or most realistic option, or that one partner had a personality problem. Each individual statement was reported by between ten and twenty-eight per cent of clients. Forty per cent reported at least one. Undermining statements were associated with poorer outcomes and with therapy ending sooner.

Doherty’s own explanation is not that these therapists are unkind. It is that some of them lack a relational framework and the skills that go with it, and so become frustrated and pessimistic when a couple is difficult. Which is the same sentence this page started with, arriving from the other end.

The twelve questions

Say these on the free consultation call. They take about ten minutes and they are the highest-value ten minutes in the whole process.

1. Did you have formal coursework and supervised training specifically in couples therapy — not individual?The one that matters most. No licence in the United States requires it.
2. What share of your current caseload is couples?Under about a third and you are the side project.
3. Which model do you work in, and where did you train in it?Then go and check the answer in the directory yourself.
4. What happens in session one, and what do the first four look like?Too little structure is the first named failure. If there is no shape, sessions become the same argument with an audience.
5. How do you keep one of us from feeling ganged up on?Listen for a mechanism — separate intakes, turn-taking, interruption rules — not for reassurance.
6. How will we know at eight weeks whether this is working? Do you use a measure and go through it with us?Most couples who will not benefit are visible by session four. Measuring and feeding it back cuts deterioration sharply.
7. Do you screen each of us separately for safety before we start?A yes is a marker of proper training. Fewer than four per cent of surveyed couple therapists did this consistently.
8. Do you see us individually too, and what do you do with what one of us tells you alone?Secrets policies vary enormously and are a common way this blows up. Good clinicians state theirs upfront.
9. If one of us turns out to be unsure about staying, do you do discernment counselling, or refer?In community practice at least one partner in around a third of couples is there to decide, not to improve things.
10. What is your own position on trying to save a relationship, and when would you tell a couple to stop?They have one. Better to hear it in minute three than in month six.
11. Do you bill insurance — and if so, what diagnosis goes on the claim, and whose name is on it?The most uncomfortable question here, and the one with consequences that outlast the therapy.
12. How long do you expect this to take, and how many sessions?The trials behind the good numbers ran twenty-five sessions or more. “Three or four” is a different product.

Two framings worth borrowing: ask for an explicit trial period, and give it eight to ten sessions before you judge it — that is the Gottman Institute’s own advice, and it is a reasonable middle between quitting at two and drifting for a year.

When this is the wrong thing to book

• Where there is violence, intimidation or coercive control. Joint sessions are not the right setting and can raise the risk, because what is said in the room can be answered for afterwards. The National Domestic Violence Hotline is explicit that it does not encourage anyone in an abusive relationship to seek counselling with their partner. In the United States: 1-800-799-7233, or text START to 88788. Our page on telling difficult from dangerous goes through this properly.
• Where one of you is already deciding whether to leave. This is far more common than the format assumes, and standard couples therapy is not built for it — it is built for two people who both want to be there. The thing to ask for by name is discernment counselling, which is designed for exactly this and is short.
• Because you keep having the same argument. That, on its own, is not a diagnosis. A large share of what couples disagree about is perpetual rather than solvable — it is about temperament and what each of you is like — and recurrence is the normal condition rather than evidence that something is broken.

A note on the online platforms

If you search for this you will land on ReGain, which is the couples service run by BetterHelp. Two things are worth knowing before you sign up. The account is joint: written messages are visible to both partners and the therapist, which is a different proposition from a private hour, and it matters if there is anything one of you is not ready to say in front of the other. And in 2023 the Federal Trade Commission ordered BetterHelp to pay $7.8 million over disclosing consumers’ email addresses, IP addresses and health-questionnaire answers to Facebook, Snapchat, Criteo and Pinterest for advertising, after promising not to use health data that way. The final order bans the practice and requires a privacy programme; around 800,000 people were sent refund notices in 2024.

Where inbetween fits, and where it does not

We are not therapy, we do not employ clinicians, and this page exists partly to help you find one. So the honest statement is narrow.

The evidence for structured self-help is real, and it is for mild to moderate distress — which, judging by who searches for this, is the band most people reading are actually in. It is also worth saying that the best trials of guided self-help for couples all included calls with a human coach, and we do not have those.

What inbetween does is narrower still and useful while you are waiting, which you will be. You talk to it, they talk to it, and it is the only one who hears both sides, so nothing you say reaches them unless you allow it. Arriving at a first session already knowing what each of you actually wants is not wasted preparation — and if what is between you is bigger than that, it will say so and keep saying it.

If you have not booked anything yet, what actually happens in a first session is the companion to this page.