By Víctor Silva and Marcela Gómez
If you are going to invest time, money, and difficult conversations in couples therapy, it is reasonable to ask whether it works. The brief answer is that research finds average improvements with various interventions. A fuller answer needs to specify what improved, with which treatment, compared with what, and for which couples.
That does not translate into a universal percentage of “relationships saved.” Satisfaction, communication, and other relationship outcomes are not the same as staying together. Benefits found in studies neither guarantee an outcome for a particular couple nor automatically apply to any service.
What “working” means before you look at a number
For you, it may mean no longer arguing in the same way. For your partner, it may mean recovering intimacy. Or one person wants to decide whether to continue while the other hopes to preserve the relationship. The question of effectiveness changes with the outcome you seek.
Research needs to turn outcomes into observable measures: questionnaires, observations of interaction, or follow-up assessments, among other possibilities. Those measures provide information, but they do not necessarily encompass everything a person considers valuable or difficult in their process.
It helps to distinguish two matters: which effects have been studied and which goals you want to agree on in your consultation. Citing general evidence is not enough to establish every outcome that sounds desirable.
What some relevant reviews show
Relationship satisfaction and functioning
Roddy and colleagues’ 2020 meta-analysis brought together 58 studies representing 40 unique samples. It found improvements in relationship satisfaction and other areas, including communication, intimacy, and partners’ behavior. Its analyses cover different designs and comparisons; the samples were limited to different-sex couples, restricting automatic generalization to all couples.
Improvements across several areas do not mean every couple improves in all of them or every intervention measured the same things. Nor do they establish that a change in a score amounts to resolving any important conflict.
A more recent review of marital distress
Joseph and colleagues published a systematic review and meta-analysis of twelve studies of couples interventions in 2025. They found an aggregate reduction in marital distress, alongside heterogeneity and signs of publication bias. The interventions included were varied: they do not represent a single treatment or an individual prognosis.
Heterogeneity means results are not identical across studies. Publication bias warns that published research may not represent all the results produced in a balanced way. These limitations belong alongside a favorable reading, rather than being hidden behind an impressive number.
From studies to routine practice
Bradbury and Bodenmann’s 2020 review describes benefits of couples interventions alongside difficulties in transferring them to routine practice and maintaining changes. This prevents presenting the evidence as a promise of permanent improvement.
The reviews do not answer exactly the same question or always include the same treatments, samples, and follow-up periods. Selecting the most favorable number and turning it into the general effectiveness of “couples therapy” is therefore misleading.
How to read a result without turning it into advertising
Improvement between the beginning and end of a process provides information. Attributing it to treatment also requires considering the comparison: what happened to people who waited, received another intervention, or were assigned to different conditions?
Research studies do not all use the same design. A trial, an observational study, and a case series can be useful for different questions; they do not provide the same strength of evidence for saying an intervention caused an outcome.
Reporting an average improvement also differs from counting how many people reached a specific clinical criterion. An effect size is not a success percentage, much less a percentage of couples who stay together.
If you encounter “it works in 80% of cases,” questions are missing: What does working mean? Who was included in the calculation? What happened to people who dropped out? How long was the follow-up? Which treatment was studied? Without that information, the number may appear more precise than it is.
What has been measured and what needs to be defined as a goal
Reducing distress, improving communication, and changing certain interactions are outcomes present in the research reviewed here. Trust and intimacy may also be part of the work, but you should not assume all studies assessed them in the same way.
Seeking clarity about continuing or ending a relationship is a possible consultation goal that needs to be stated explicitly. These meta-analyses do not establish equally that any therapy can produce that clarity or an appropriate decision.
At Ancestrina, the institutional goal is clarity about differences, each person’s participation, and the relationship’s possibilities. A useful result does not necessarily mean staying together. This is a position on the purposes of the work and people’s autonomy; these meta-analyses do not automatically demonstrate its effectiveness.
This distinction also avoids appropriating therapeutic models: favorable results for a studied approach do not establish that Ancestrina formally uses it or reproduces its effectiveness. The data belong to the interventions and conditions investigated.
What to ask the professional
You can ask how they understand your reason for seeking help, what goals they propose, and how the process would be reviewed. You can also ask about their training and experience with that problem, the limits they recognize, and when they would consider other care or a referral.
You do not need to administer an academic examination. You are seeking information to decide whether the work offered fits your needs, especially when you have different goals, specific sexual difficulties, or safety conditions requiring assessment.
A serious explanation should be able to acknowledge uncertainty. Promising to save the relationship or resolve a problem within a fixed number of sessions does not follow from the general evidence presented here.
Participation matters; it does not make the outcome entirely an individual responsibility
A joint consultation cannot be treated as though both people share a goal that one has not chosen. Attending, wanting to rebuild, and being willing to examine your own behavior are different commitments.
It would also be unfair to say that an unexpected outcome means someone did not work hard enough. There may be different goals, limitations of the approach, material conditions, or a relationship someone does not wish to continue. This article cannot identify from the outside which explanation applies to a particular case.
Favorable evidence justifies considering a possibility of help. It does not oblige you to keep trying indefinitely or interpret every difficulty as a reason to make more effort.
How to consider whether the process makes sense for you
It helps to return to the initial reason for consulting and your agreed goal. What can you address now that you could not before? Which changes do you recognize, and which do you not? Are important questions still without space? Does the work continue to respond to what both people have consented to?
These are not validated indicators or an assessment that replaces the professional. They are questions for discussing the process, rather than judging its usefulness only by whether a session felt pleasant or produced an insight.
Useful work can include uncomfortable conversations. Emotionally intense sessions may also fail to translate into what you seek. If something is not working, you should be able to say so and review it, rather than having it automatically explained as resistance.
If you need to consider when to seek support, When to Consider Couples Therapy helps clarify that question. If you are already looking for a process, you can learn directly about Ancestrina’s couples therapy and its conditions. Research informs that possibility; it cannot promise what will happen in your relationship.
Sources and scope of the references
- Roddy, M. K., Walsh, L. M., Rothman, K., Hatch, S. G., and Doss, B. D. (2020). Meta-analysis of couple therapy: Effects across outcomes, designs, timeframes, and other moderators. Journal of Consulting and Clinical Psychology, 88(7), 583–596. Reference. Meta-analysis: Aggregate outcomes in satisfaction, communication, intimacy, and behavior; limitations of study designs and samples composed exclusively of different-sex couples.
- Joseph, B., Joseph, V. K., and Rajan, S. K. (2025). Effectiveness of Couple Interventions in Marital Distress: A Systematic Review and Meta-Analysis. Iranian Journal of Public Health, 54(1), 112–123. Reference. Systematic review and meta-analysis: Aggregate reduction in marital distress, with heterogeneity and publication bias. Varied interventions, rather than an individual prognosis.
- Bradbury, T. N., and Bodenmann, G. (2020). Interventions for Couples. Annual Review of Clinical Psychology, 16, 99–123. Reference. Scientific review: Benefits and limits of couples interventions, transfer to routine practice, and maintenance. It is not a percentage of relationships saved.
