What Is The Success Rate Of Marriage Counselling

There is no single success rate, and the confident figures you have already read mostly do not survive being looked up. We went and looked all of them up.

We are Richard and Helena Boyd and we have run a couples practice in Perth since 2005. This page gives the honest numbers, explains what researchers actually mean by success, and traces each of the circulating statistics back to where it came from. Some of them are real and misdescribed. At least one is an inverted figure that means the opposite of what it is used to say.

The Short Answer

Across controlled trials, roughly 60 to 72 per cent of people in couples therapy show reliable improvement or recovery by the end of treatment, and roughly a third to a half move all the way into the non-distressed range.

The average effect is large. The typical treated couple ends up better off than about 80 per cent of comparable couples who received nothing.

Everyday practice is more modest. In a nationwide study of 554 couples in ordinary counselling settings, 55 per cent improved or were at least stabilised, and 45 per cent did not benefit.

Those three sentences are the honest answer. Everything below explains why the number you were given was different, and why several of the figures in circulation cannot be reconciled with each other no matter how generously you read them.

None of this means the research is weak. Couples therapy has one of the better evidence bases in applied psychology, and it is measured more rigorously than most fields manage. The problem is not the science. It is what happens to a number after it leaves the paper it came from.

Why This Question Is Harder Than It Looks

Success rate sounds like one thing and is at least four.

It might mean the proportion who feel better. Or the proportion who cross from distressed into non-distressed on a standardised measure. Or the proportion still together some years later. Or the proportion who got what they came for, which sometimes includes separating well.

Those produce very different percentages from the same data. A page that quotes a single figure without saying which one it means is not necessarily lying, but it is not telling you anything either.

What Success Actually Means In The Research

Researchers use four categories, and the difference between them explains most of the confusion in the public numbers.

Recovered

The person’s score improved by more than measurement error could account for, and their post-treatment score crossed from the distressed range into the non-distressed range. This is the strict definition, and it is what most people picture when they hear success.

Reliably Improved

The score improved by more than measurement error, but the person is still in the distressed range at the end. Measurably better, not yet well. This category is real progress and it is not recovery.

Unchanged

Any movement was within the margin of error of the questionnaire. Not worse, not better.

Deteriorated

The score moved reliably in the wrong direction. This category exists, it is not small, and almost nobody publishes it.

Once you know these four exist, the contradictory figures make sense. Seventy per cent and thirty five per cent can both be accurate descriptions of the same trial, because one counts recovered plus reliably improved, and the other counts recovered only.

How It Is Measured

Most of this research runs on a questionnaire called the Dyadic Adjustment Scale.

It has 32 items, scores between 0 and 151, and a score around 97 or below is conventionally treated as relationship distress. Its internal consistency is high, around .95 for the total score, which is why it has survived since 1976 and appears throughout the marital and family therapy literature.

The clinical significance test applied to it has two parts and both must be met to count as recovery. First, the change has to exceed what measurement error could produce, which is a statistical threshold rather than a judgement. Second, the final score has to cross the distressed to non-distressed line. Passing only the first test makes you reliably improved. Passing neither makes you unchanged.

This is more rigorous than most fields manage, and it is worth knowing that the numbers rest on something real rather than on therapists’ impressions.

The Difference Between A Person And A Couple

Here is a distinction almost no page makes, and it changes how you should read every percentage on this subject.

Most published figures classify individuals rather than couples. So a 70 per cent success rate usually means 70 per cent of the people in the study, not 70 per cent of the relationships.

The couple level figure is lower. In only about half of treated couples do both partners show significant improvement in relationship satisfaction by the end of treatment. If you are asking whether therapy will work for your marriage rather than for one of you, that halved figure is the relevant one.

The Headline Numbers

With those definitions in place, here is what the best synthesis of the research reports.

At the end of treatment, about 48 per cent of couples have recovered, around 18 per cent have reliably improved without recovering, roughly 21 per cent are unchanged, and about 13 per cent are worse than when they started. The average effect size is around 0.84.

Taking the strict recovery definition alone, the figure across measures runs from about 33 to 46 per cent. Taking recovery plus reliable improvement, it runs from about 59 to 72 per cent. Both are correct. They answer different questions.

Trials Versus Everyday Practice

The gap between these two is the single most under-reported thing in this literature.

In research trials the therapy is manualised, the therapists are supervised and monitored, and the couples are screened for suitability. In an ordinary private practice none of that applies.

In a nationwide study of 554 couples across 30 counselling centres, the effect on relationship distress was 0.38, roughly half the trial figure. Around 40.5 per cent moved into the non-distressed range, 55 per cent improved or stabilised, and 29 per cent had separated or were heading that way by follow up.

A large United Kingdom service study of 877 clients found 36.6 per cent achieved clinically significant change on psychological distress and 12.1 per cent on relationship quality, with 3.6 and 7.8 per cent respectively deteriorating.

Neither of those is a bad result. They are simply what happens when you measure real clients rather than trial participants, and they are the numbers a private practice should quote about itself.

What Improvement Actually Looks Like

Percentages are abstract, so it is worth describing what a recovered score corresponds to in a real week.

The first change is almost always in conflict. The same arguments still start, and they get shorter and recover faster. Communication breakdowns that used to cost three days cost an afternoon. This is the least romantic form of progress and the most reliable, and it usually arrives before anything feels warmer.

Then the harmful patterns become visible while they are running rather than afterwards. Couples describe catching themselves mid sentence, which sounds small and is the mechanism by which the score moves. Different communication styles stop being read as character defects. Conflict resolution becomes possible because both people can stay in the room.

Emotional connection follows rather than leads. Where trust has been damaged, this is the stage at which partners rebuild trust, and it is slower than the rest of it. Emotional closeness and physical intimacy generally return last of all, which matters practically, because couples who expect that order reversed often stop at the point the work is about to pay off.

A meaningful improvement in the research sense usually corresponds to something modest in life. Not transformation. Two people who can raise a difficult subject, communicate effectively enough to be understood, and repair afterwards. Higher relationship satisfaction is what that produces rather than what causes it.

What Couples Bring, And Whether It Changes The Odds

The presenting problem varies enormously and the underlying work varies less than people expect.

Most couples arrive with recurring conflict, emotional distance, or both. Underneath sit the things that were never raised directly, which in long relationships is commonly money, differing expectations about children, or one of the major life transitions that quietly rewrites the arrangement. A birth, a bereavement, illness, redundancy or reduced income all put load on a relationship that was previously coping.

Financial stress deserves particular mention because it is both a common presenting problem and a compounding one, since it limits the options available for dealing with everything else.

Sexual difficulties are among the least raised and among the more responsive, though where there is pain or a physical change a medical assessment comes first. Physical health problems and chronic illness change relationship dynamics substantially and are rarely what a couple books about.

Does any of this change the success rate. Honestly, less than you would think. The evidence on pretreatment characteristics is that they predict outcome poorly at the individual level. What predicts well is whether things start moving early, which is a fact about the process rather than about the problem you walked in with.

The Circulating Figures, Traced

We looked for the source of each statistic that appears on competing pages. Here is what we found.

The Seventy To Seventy Five Per Cent Figure

This traces to a 1999 meta analysis of emotionally focused therapy covering four trials, three of which were run by the model’s own developers. Four small studies of one branded approach, generalised into a claim about marriage counselling as a whole.

An independent review of twelve emotionally focused therapy studies produced an effect size of 0.73 rather than the 1.31 reported by the developers, which is the size of the allegiance effect in this field.

The Ninety Per Cent Figure

You will read that around 90 per cent of clients report improved emotional health after therapy. This is not an outcome study.

It descends from a 1996 American survey in which clients were recruited by their own treating therapists, with no control group, no standardised outcome measure and no follow up. It measures whether people liked their therapist. That is worth something, and it is not a success rate.

The Seventy Per Cent Cognitive Behavioural Figure

The claim that cognitive behavioural couple therapy has a 70 per cent improvement rate has no primary source we could find. It appears to be three separate numbers that have merged.

One is a 2012 review stating that couple therapy positively affects 70 per cent of couples receiving treatment, which is a statement about couples therapy in general. One is the emotionally focused therapy figure above. One is the 70.4 per cent reported for integrative behavioural couple therapy in a large trial.

The actual figure for cognitive behavioural couple therapy is an effect size of 0.53, from an independent meta analysis of 33 studies covering 2,730 participants. Worth adding that a component analysis found the specifically cognitive elements associated with smaller effects, while communication and problem solving skills training predicted larger ones.

The Fifty Per Cent Stay Together Figure

This one is not merely unsupported. It means the opposite of what it is used to say.

It traces to a 1991 study reporting that 38 per cent of couples in behavioural marital therapy had divorced at four year follow up. That divorce figure was inverted into a claim about couples staying together, then rounded upward, then repeated.

The real numbers are considerably better. Roughly 74 per cent of couples in the largest long term trial were still together five years after therapy.

The Six Year Wait

Couples wait six years before seeking help is one of the most repeated lines in this industry and it is not supported. When researchers measured it properly across 270 people entering therapy, the average was 2.68 years, and most came within two years.

Success Rates By Approach

People reasonably want to know which method has the best numbers. The answer is less satisfying than the question.

Emotionally Focused Therapy

Emotionally focused couples therapy works on the emotional bond and the cycle between partners. It has the best known figures and the largest gap between developer led and independent research, at 1.31 versus 0.73.

Behavioural And Integrative Behavioural Couple Therapy

The behavioural family has the largest and most replicated evidence base, anchored by a trial of 134 seriously distressed couples followed for five years. In that trial 71 per cent of integrative behavioural couple therapy participants reliably improved or recovered at the end of treatment against 59 per cent for the traditional version, 69 versus 60 per cent at two years, and 50 versus 46 per cent at five years.

Cognitive Behavioural Couple Therapy

Not cleanly separable from behavioural couple therapy in the outcome literature, since reviews use the terms interchangeably. Effect size 0.53, and the cognitive component has not been shown to add anything.

The Gottman Method

Built on a genuinely impressive body of observational research about how couples interact. As a treatment it has no published independent randomised trial, which is a different statement from saying the underlying ideas are wrong.

Why No Approach Comes Out On Top

In the largest independent head to head comparison, 33 studies and 2,730 participants, cognitive behavioural and emotionally focused approaches produced 0.53 and 0.73 at post treatment and 0.35 and 0.66 at six months, with no statistically significant difference between them.

That pattern repeats across the field. Evidence based approaches perform comparably, and the developers of each approach are usually the people testing it. What this means practically is that choosing a skilled couples therapist you can both be honest with matters more than choosing a brand.

What Happens At Two Years

Gains erode, and the honest figures are rarely published.

At two years, roughly 37 per cent of behavioural couple therapy couples remained recovered and about 25 per cent had deteriorated or separated. For emotionally focused therapy the figures were about 39 per cent recovered and 32 per cent deteriorated or separated.

What Happens At Five Years

At five years, roughly a third remain recovered across approaches.

On relationship status specifically, about a quarter had separated or divorced, at 25.7 per cent for integrative behavioural couple therapy and 27.9 per cent for the traditional version. Older studies reported higher rates, including 38 per cent divorced at four years in one 1991 trial and 43.6 per cent at five years in a 1980 study.

Put simply, of the couples who reliably improve by the end of therapy, only about half are still in that position five years later.

Deterioration, The Number Nobody Publishes

About 13 per cent of couples are worse at the end of treatment. Within two to five years, roughly a quarter to a third have deteriorated or separated.

You will see 35 to 50 per cent quoted for this. That is inflated. It appears to take the top of an older, broader satisfaction decline band and present it as a separation rate.

What is known about why couples deteriorate is correlational rather than settled. The factors that recur are divergent goals between the partners, a split alliance where one person connects with the therapist and the other does not, very high distress at intake, and an affair that has not been disclosed.

How Long It Takes

You will read that marriage counselling typically lasts from three months to a year. We could not find a source for that range.

What the research reports is around 20 sessions over about six months in trial and review contexts, and a mean of 23.3 sessions in a large naturalistic study, with a range from two to 150. Some couples do what they need in four. Others work for well over a year.

What Predicts Success For You

General success rates matter less than the factors that apply to your situation, and these are better evidenced than the headline percentages.

Goal alignment is the clearest. In a study of 249 couples, 7.8 per cent had separated six months later where both partners came wanting to improve the relationship, against 45 per cent where their goals differed and 56 per cent where both came to work out whether to continue.

The therapeutic alliance matters, and its imbalance matters more. Across 48 studies the alliance was associated with outcome at around 0.30, and a split alliance predicted poorer outcomes at a similar strength.

Early change predicts late outcome. Most measurable improvement happens across roughly the first four sessions and then plateaus, about a third of couples are already off track by session three, and change by the end of therapy correctly classified couples as stable or separated 89 per cent of the time.

Timing helps. Lower distress at the outset is associated with better results, which is the practical case for early intervention rather than a moral one.

Active participation between sessions is the part you control. Real growth tends to come from applying things outside the room rather than from the hour itself.

Success Rates In Specific Situations

The overall figures conceal wide variation, and some situations have their own evidence.

After infidelity, couples arrive more distressed, improve faster, and finish treatment at a comparable level to other couples. At five years they separate at roughly twice the rate. Whether the affair was disclosed rather than concealed is the factor that separates recovery from deterioration.

Where one partner has depression, behavioural couples therapy is recommended in United Kingdom clinical guidelines when relationship problems may be contributing, delivered across 15 to 20 sessions.

Anxiety disorders are thinner than people assume. For panic disorder with agoraphobia, partner assisted exposure produced clinically significant improvement in 23 to 45 per cent of cases alone, rising to 82 to 86 per cent combined with cognitive therapy and communication training, but that rests on twelve small studies published between 1977 and 1993. For obsessive compulsive disorder there are five trials plus a pilot. For generalised anxiety disorder there are no controlled trials of couple based treatment at all.

For substance use, behavioural couples therapy has good evidence as a treatment for the substance use itself, and it is a structured protocol delivered with a non using partner rather than general relationship counselling.

Online therapy performs comparably. A randomised trial run in Perth found couples therapy delivered by videoconference produced outcomes no different from face to face, though it involved 30 couples.

Relationship Education Compared

Relationship education is a different product with different numbers, and the two get conflated.

Across 117 studies, relationship education produced effects of about 0.36 on relationship quality and 0.44 on communication skills, largely maintained at follow up. Those are small to moderate, against the large effects for therapy.

That is not evidence that education works less well. Education is delivered to couples who are not clinically distressed, so there is less room to improve. Comparing the two numbers directly is a category error.

The Australian Picture

Almost all the research above comes from overseas, which is worth saying plainly.

There is one peer reviewed Australian routine practice outcome study of couple counselling, conducted within a large not for profit provider and published in 2014 in the Australian and New Zealand Journal of Family Therapy. Relationships Australia has published a national outcomes framework and program level data, but no couples specific success rate, so any figure attributed to them is unsourced.

One Australian finding is worth knowing. Relationship counselling clients here report roughly four times the rate of very high psychological distress found in the general population, which says something about who actually walks through the door.

Who This Research Was Done On

A limitation the field acknowledges and most pages ignore.

The outcome evidence comes overwhelmingly from Western, predominantly white samples. The most ethnically diverse landmark trial in the field had 22 per cent non white participants, which is the high water mark rather than the norm. We could locate no published analysis testing whether success rates differ across cultural contexts.

There is therefore no basis for claiming marriage counselling works equally well across cultural groups, and equally none for claiming it does not. It is an open question, and a page telling you otherwise is filling a gap with confidence.

What You Can Do To Improve Your Own Odds

Most of the variance is not in the method, which means a reasonable amount of it sits with you.

Attend consistently. Irregular attendance is one of the more obvious ways a course of therapy stops working, and the therapeutic process depends on continuity between sessions rather than on the intensity of any one of them.

Be actively involved rather than present. Couples who engage honestly, including about the things that are embarrassing, do better than couples managing their own presentation. A safe space only produces value if somebody uses it.

Do the work between appointments. Most couples find that real change happens applying something on a Tuesday rather than in the room on a Thursday, and active participation of that kind is what separates couples who consolidate gains from those who lose them.

Say what you actually want early. If one of you is undergoing therapy to decide rather than to repair, that changes the work entirely and is better named in session one.

Raise it when it is not working. Ask your marriage counsellor directly whether they think anything is shifting. A practitioner who cannot answer that question is not tracking your outcome, and professional support without measurement is guesswork.

How To Read A Practice’s Claims

Since you are likely reading several of these pages, here is how to assess what you find.

Be sceptical of any specific success rate attached to a particular practice. Unless they have measured their own outcomes, which almost nobody in private practice does, they are quoting research figures as though they described their results. Australian consumer law and health advertising rules both prohibit claims that create unreasonable expectations of benefit.

Notice whether the page distinguishes recovery from improvement. If it does not, the figure is uninterpretable regardless of whether it is accurate.

Notice whether negative outcomes appear anywhere. Deterioration is roughly 13 per cent at the end of treatment, and a page describing only positive outcomes has decided what to leave out.

Treat evidence based as a claim to check rather than a reassurance. Evidence based methods in couples work means the behavioural family, emotionally focused therapy and a small number of others, and research supports them at roughly comparable levels. Where research suggests something more modest than the page claims, that gap tells you what the page is for.

Ask what training the practitioner has specifically in couples work, and whether they also offer individual therapy or work as a family therapist, since general counselling training does not cover couples work and the disciplines are not interchangeable. Experienced therapists in individual work are not automatically good at this.

Does Staying Together Equal Success

Only if that was what you came for.

The research counts separation as a poor outcome, which is a limitation of the research rather than a description of anyone’s life. Counselling that produces a clear, well managed separation has done something valuable, particularly where children are involved, and couples who reach that point with professional guidance generally handle it better than those who reach it alone.

Success can reasonably mean healthier communication, deeper insights into what kept happening, a healthier relationship, relationship repair after a specific injury, significant benefits to each person’s mental health and emotional well being, or an amicable ending. Nothing about the research says survival equals success. Deciding which one you are measuring is your job rather than the researcher’s.

What The Success Rates Do Not Capture

Every figure on this page measures relationship satisfaction. Several things that matter are not in the numbers at all.

The first is individual benefit. In the largest naturalistic study, psychological distress improved more than relationship quality did, at 36.6 per cent achieving clinically significant change against 12.1 per cent. Couples counselling frequently produces significant improvements in how each person is functioning even where the relationship measure barely moves, and a success rate built on relationship scores records that as a partial failure.

The second is children. Conflict between parents is the part of a difficult marriage that reaches a child most directly. Reducing it changes a household whether or not the relationship score crosses a threshold, and no outcome measure in this literature captures it.

The third is transferable skill. What people learn about their own communication patterns tends to follow them into work, parenting and friendships. Successful therapy in that sense outlasts the presenting marriage problems.

The fourth is clarity. Couples who attend therapy and reach a clear decision, including to separate, have got something real, and the research counts it as a poor outcome.

The fifth is the relationship issues that never surface as complaints. Plenty of couples arrive over one thing and find the therapy process spends most of its time somewhere else entirely, which is difficult to score and often the most valuable part.

None of this is an argument that the numbers are wrong. It is an argument that evidence based therapy is being assessed on one dimension because that dimension is measurable, and that you should hold the resulting percentages a little loosely.

What This Looks Like In Practice

Three composites below, drawn from recurring patterns rather than particular couples, with details changed.

A couple married sixteen years would count in the research as reliably improved rather than recovered. Their score moved substantially and stayed inside the distressed range. Both described the year afterwards as the best in a decade.

A couple married four years recovered on every measure by session ten and separated eighteen months later. The therapy worked and the marriage did not, and those are genuinely different questions.

A couple in their sixties came for eight sessions after forty years. Nothing dramatic changed. They interrupted each other less, which they said was the whole of what they wanted.

Frequently Asked Questions

The questions people ask about the numbers themselves.

So What Is The Actual Success Rate?

Between about 60 and 72 per cent of people reliably improve or recover by the end of treatment, and about a third to a half fully recover. In everyday practice roughly 55 per cent improve or stabilise. Any single figure quoted without that context is doing something other than informing you.

Why Do Different Websites Give Different Numbers?

Because they are measuring different things, usually without saying so, and because several of the circulating figures were copied from sources that do not support them. Recovery, improvement, satisfaction and staying together produce four different percentages from identical data.

Does A High Success Rate Mean It Will Work For Us?

Not directly. Average figures describe populations. What predicts your outcome is whether you both want the same thing, whether both of you feel the therapist is yours, and whether anything shifts in the first few sessions.

Is A More Experienced Therapist More Likely To Succeed?

Less than you would expect. Therapists account for about 8 per cent of the variation in outcomes, and within that the biggest factor is the quality of alliance they typically form rather than years in practice. Experience specifically with couples contributes, general seniority much less.

How Many Sessions Before You Know?

Sooner than most people expect. Measurable change concentrates in roughly the first four sessions and then plateaus, and about a third of couples are already off track by session three. If nothing has shifted by then it is worth raising directly rather than assuming more time will fix it.

What Success Rate Should A Therapist Claim?

None, unless they have measured their own. A practice quoting a research figure as though it described their own results is misusing it, and Australian advertising rules prohibit claims that create unreasonable expectations of benefit.

Getting Started In Perth

If you take one thing from this page, take the reason the numbers vary. Success in this research means something precise, most published percentages describe individuals rather than couples, and the gap between trial conditions and real practice is roughly a factor of two.

The honest version is that marriage counselling produces large average improvements, that roughly half of couples recover, that a meaningful minority do not and some get worse, and that gains need maintaining. That is a good enough return to be worth doing and not good enough to promise anybody.

What actually moves your odds is not the method. It is arriving while there is still goodwill, both of you wanting the same thing, nothing being concealed, and both of you willing to look at your own half.

We see couples at our Inglewood rooms and by video across Western Australia, with evening and Saturday morning appointments and no referral needed. Phone 1300 956 227 to book, or read does marriage counselling work for the shorter version of this argument.

General information rather than clinical advice. Research figures describe averages across study populations and cannot predict an individual outcome. Lifeline is on 13 11 14, and emergency services on 000.