Relationships

Relationship compatibility test: what it can and cannot tell you

By Matthew Firth, founder of NexSpark · Sep 18, 2026 · 8 min read

A relationship compatibility test is a questionnaire that compares two people's stated values, habits, and preferences to estimate how well they might fit together. The uncomfortable finding underneath most of these tests is that they measure the part of a relationship that predicts the least: the largest study ever run on the topic found that individual traits and preferences, the exact material most compatibility tests are built from, explain a small share of relationship quality and almost none of whether that quality rises or falls over time.

What a compatibility test actually measures

Open any relationship compatibility test and you will find some mix of the same four ingredients. Trait similarity asks whether you are both introverts, both morning people, both spenders or savers. Values alignment asks about children, religion, money, and where you want to live in ten years. Attachment style questions ask how you behave when a partner pulls away or gets too close. And love language questions ask whether you feel most cared for through words, time, gifts, touch, or acts of service. Each of these is a real, measurable thing about a person. What none of them measure directly is the relationship itself, how the two of you actually behave together, under stress, over years.

That distinction between a trait and a relationship dynamic turns out to matter more than it sounds like it should, and it is the subject of the largest study ever conducted on the question.

What the research says predicts relationship quality instead

In 2020, psychologist Samantha Joel of Western University and Paul Eastwick of UC Davis led a team of roughly 85 researchers in pooling 43 longitudinal datasets covering 11,196 couples, then used machine learning to test which of more than 2,000 variables actually predicted relationship quality. The paper, published in the Proceedings of the National Academy of Sciences, split the predictors into two families: individual-difference variables, personality traits, attachment style, life satisfaction, depression, the kind of thing most compatibility tests measure, and relationship-specific variables, how committed you perceive your partner to be, how appreciated you feel, sexual satisfaction, and conflict.

The individual-difference variables explained as much as 21% of the variance in relationship quality at baseline, and that figure shrank to about 12% by the end of the studies. The relationship-specific variables did far better, explaining as much as 45% of the variance at baseline and 18% by study end. The single most consequential finding, though, was this: none of the variables, individual or relationship-specific, could predict whether a couple's relationship quality would rise or fall over time. Who you are, and even how you currently feel about your partner, told the researchers almost nothing about where the relationship was headed.

Put plainly, the traits a compatibility test asks about are the exact category the research found weakest, and the one thing no self-report measure could touch at all, the trajectory of the relationship, is the thing most people actually want to know before they commit.

What a compatibility test can legitimately tell you

None of this means the tests are worthless, it means they are narrower than they're marketed. A well-built compatibility test is genuinely useful for three things. First, deal-breakers: if one of you wants children and the other is certain they don't, a quiz surfaces that in two minutes instead of six months. Second, values mismatches that are easy to avoid discussing directly, spending versus saving, how much family involvement feels normal, how religious observance should factor into a shared life. Third, and probably most valuable in practice, a compatibility test is an excuse to have a conversation two people might otherwise put off. The questions give you a script for talking about things that feel awkward to bring up cold.

What it cannot tell you

A compatibility test cannot measure chemistry, the immediate, hard-to-articulate pull that either exists or doesn't and that no questionnaire captures. It cannot tell you how the two of you actually handle conflict, whether you repair after a fight or let it fester, which the research above found to be one of the strongest relationship-specific predictors and something you can only observe by actually disagreeing with someone. And it cannot measure commitment, the day-to-day choice to keep showing up, because commitment is demonstrated over time, not declared on a form. These three gaps, chemistry, conflict behavior, and demonstrated commitment, are arguably the three things that matter most, and a test cannot reach any of them.

What are some signs a relationship might be incompatible?

The signs worth trusting are behavioral, not categorical. Recurring arguments that never actually resolve, one person doing most of the emotional or household labor without acknowledgment, a gap between how much conflict each of you can tolerate before shutting down or escalating, and a felt need to perform a different version of yourself to be accepted are all stronger signals than a mismatched answer on a values quiz. A shared love of hiking or a difference in love language rarely breaks a relationship. A partner who won't engage in repair after a fight, consistently, does.

Can incompatibility be overcome?

Some of it, and some of it not. Differences in personality, daily habits, or even some values are workable when both people are willing to negotiate and repair after conflict, and that willingness matters more than the size of the original gap. What tends not to be workable is a mismatch on genuine non-negotiables, wanting children or not, monogamy, where to live, combined with one partner who will not engage in repair at all. A compatibility test can flag which category a disagreement falls into. It cannot do the negotiating, and it cannot make either of you more willing to do it.

How to use a compatibility test well

Take it together, or take it separately and compare answers out loud immediately after, rather than reading a percentage match alone and treating it as a verdict. Spend most of your time on the questions where you disagreed, not the ones where you matched, because agreement rarely needs discussion and disagreement is where the useful information lives. Treat a low score as a list of topics to talk through, not a reason to end things, and treat a high score with some skepticism, since agreeing on values says nothing about how you'll handle the next hard week. The instrument is a conversation starter. The conversation is the actual test.

Where NexSpark's own quiz fits

It's worth being direct about how this applies to the questions on NexSpark's own intake, since some of them look similar to a compatibility test on the surface. They aren't scored the same way, and they aren't meant to produce a percentage match at all. The intake exists to feed a human matchmaker and an AI model information about your values, your dating archetype, and your relationship style, which then gets weighed against a pool of other members before anyone reviews an actual introduction. Given what the research above found, that a values quiz alone can't predict how a relationship unfolds, that's a deliberate design choice rather than an oversight: the quiz narrows the field, a person makes the judgment call, and no one gets a compatibility percentage instead of an actual introduction.

Sources and methodology

Samantha Joel, Paul W. Eastwick, and approximately 85 co-authors, Machine learning uncovers the most robust self-report predictors of relationship quality across 43 longitudinal couples studies, Proceedings of the National Academy of Sciences, 2020: 43 longitudinal datasets, 11,196 couples, more than 2,000 predictor variables tested; individual-difference variables explained up to 21% of variance in relationship quality at baseline and 12% by study end; relationship-specific variables explained up to 45% at baseline and 18% by study end; no combination of self-report variables meaningfully predicted change in relationship quality over time. Madeleine A. Fugère, Ph.D., The Strongest Predictors of Romantic Relationship Quality, Psychology Today, published July 28, 2020, corroborating the same figures and top five relationship-specific predictors: perceived partner commitment, appreciation, sexual satisfaction, perceived partner satisfaction, and reduced conflict. University of Toronto, Little evidence linking five 'love languages' to healthy relationships, researchers say, summarizing a review by Emily Impett, Haeyoung Gideon Park, and Amy Muise of 10 existing studies finding no support for the claim that partners who share a love language have better relationships. No NexSpark member data was used in this article.

Frequently asked questions

What are some signs a relationship might be incompatible?
Recurring fights that never resolve, one partner consistently doing more of the emotional or household labor, a mismatch in how much conflict either of you can tolerate before shutting down, and a sense that you have to perform a different version of yourself to be accepted are stronger signs than a values checklist mismatch. A compatibility test can surface a values gap; it cannot tell you how the two of you actually handle a disagreement, which is the harder and more predictive question.

Can incompatibility be overcome?
Some of it. Differences in personality, habits, or even some values are workable when both people are willing to negotiate and repair after conflict; that willingness matters more than the size of the difference itself. What is harder to overcome is a fundamental mismatch on non-negotiables like wanting children, monogamy, or where to live, and a pattern where one partner will not engage in repair at all. A test can flag which category you are in, but it cannot do the negotiating for you.

How accurate is a relationship compatibility test?
It depends what you are asking it to predict. A test is reasonably reliable at describing where your stated values, habits, and preferences overlap or diverge, because that is a straightforward comparison of two sets of answers. It is not reliable at predicting relationship quality or longevity, because the largest study on the topic found that individual-difference variables similar to what most tests measure explained only a modest share of variance in relationship quality and almost none of the change in that quality over time.

Should couples take a compatibility test together or separately?
Together, or separately followed immediately by a shared conversation about the results. The value in a compatibility test is almost entirely in the discussion it triggers, not the score it produces. Two people comparing answers out loud, especially the ones where they disagreed, learn more in twenty minutes than either of them learns from a percentage match viewed alone.