Personality compatibility tests: what predicts a good match

What do personality compatibility tests claim?

Most of them make one of three claims. The first is that certain pairings work: a type here belongs with a type there. The second is that similarity is the ingredient, so the closer two profiles look, the better the prospects. The third is that difference is the ingredient, and two people balance each other. The second and the third cannot both be right.

Behind all three sits a single number, usually a percentage, presented as a property of the pair. That number is the part worth examining, because research has looked directly at how much of a relationship's quality belongs to the pair rather than to the two people in it.

What actually predicts how satisfied you are?

Dyrenforth and colleagues put the question to three nationally representative samples, from Australia, Britain and Germany, with tens of thousands of couples between them. They separated the effect of your own personality on your satisfaction (the actor effect), the effect of your partner's personality (the partner effect), and the effect of similarity between the two profiles (Dyrenforth et al., 2010).

The ordering was consistent across all three countries. Your own traits did the most work, accounting for several times more of the differences in relationship satisfaction than your partner's traits did. Similarity between the profiles added almost nothing once the other two were accounted for, well under one percent of the variance (Dyrenforth et al., 2010).

That is a strange result if you believe in matching and a reassuring one otherwise. The question a compatibility test is built around, how well two profiles fit, turns out to be close to the least informative question you could ask, and the leverage sits where you can reach it.

Which of a partner's traits matter, and how much?

The partner effect is small but not nothing, and it has a clear shape. A meta-analysis of couples found that partners of people lower in Neuroticism reported more satisfaction, with smaller links for Agreeableness, Conscientiousness and Extraversion, and none for Openness to Experience (Malouff et al., 2010). Emotional reactivity, not warmth or sociability, shows up first.

A second body of work predicts relationship quality a little better than broad traits do. Attachment measures ask what you do when a bond feels uncertain, and score two dimensions, anxiety and avoidance. Those two predicted relationship quality better than the Big Five and continued to predict it once the Big Five were accounted for (Noftle & Shaver, 2006). Attachment dimensions and trait measures are consistently related without either making the other redundant, so the two literatures point at much the same ground from different angles.

None of this yields a matching rule. Knowing that lower Neuroticism tends to be easier to live with does not tell you which two people belong together.

What can a compatibility result tell you, and what can it not?

A comparison of two properly measured profiles is genuinely useful, for a smaller purpose than the marketing suggests.

One caution applies to every result of this kind. People accept general personality descriptions as accurate portraits of themselves, which Forer demonstrated by giving a class the same sketch and watching them rate it as highly accurate (Forer, 1949). A compatibility report in that register feels insightful to almost any pair, so recognition is no evidence that anything was measured.

How should two people read two profiles together?

Start from a rule: a difference is not an incompatibility. A gap describes where translation will be needed, not where the relationship is weak. What causes trouble is rarely the gap itself, it is the interpretation each person puts on it. One partner plans the holiday in March and reads the other's improvising as not caring; the other reads the planning as pressure. The story about the difference is the problem.

A useful conversation with two profiles in front of you has three moves. Name the gap plainly. Say what it costs each of you in a recent situation. Then agree an arrangement rather than a winner: alternate rather than compromise, split by domain rather than splitting each task, decide in advance what a pause means.

The failure mode is easy to spot and hard to resist: using a profile to explain a partner instead of asking them something, or to excuse yourself. A score starts a question. It does not settle an argument.

Does any of this apply to friendship?

It is the same question with lower stakes. Friendships are chosen more freely and carry less of the shared logistics that make trait differences expensive. Research on close relationships is mostly done on couples, so treat the transfer as reasonable rather than proven.

What carries across is the ordering. Someone's own tendencies, how quickly they take offence, how reliably they show up, how easily they repair, shape their friendships more than any match between two profiles. The value of getting this right is not small: across many studies, people with stronger social relationships had a substantially greater likelihood of survival over follow-up (Holt-Lunstad et al., 2010).

So a friendship test is best read as a description of two people rather than a score for the pair. Where two friends differ in sociability or in how directly they raise a problem, the same three moves apply: name it, cost it, arrange it.

How can two people use Wellington?

Wellington is a science-based personality assessment from Therabot Labs LLC. It measures personality as continuous traits, built on the Big Five and HEXACO models, and writes the results back as a personal report rather than a type.

Each person takes their own free Snapshot: 60 questions, about 7 minutes, covering six dimensions, Extraversion, Agreeableness, Conscientiousness, Emotional Stability, Openness to Experience, Honesty-Humility. The first ten questions are answered before anything is asked; after question ten you sign in with an email so answers are saved and results can be shown. There is no card. Each of you gets a percentile and a written page per dimension, on norms that are provisional until the norming sample is complete, and a portrait: the combination of bands has a name and a motto, one of 243 in the portrait gallery. The bands and the Portrait names exist to make the result easier to talk about; the percentile is the measurement, and it is always shown alongside.

From a portrait you can share a card with one other person. When they take their own Snapshot, the shared page fills in both sides of the six dimensions, with a lens for reading it as friends or as a match. Nobody is scored by their partner, and there is no compatibility percentage, because we do not believe one could be honestly calculated.

The Wellington Membership, $4.99 a month, adds 336 further questions to reach 90 traits, including every Big Five facet, where a domain score stops flattening the differences that come up between two people, then continues toward 255 traits across eight domains, including relationship traits such as Trusting Disposition, Conflict Repair Capacity, Boundary Clarity and Romantic Engagement. Wellington is a wellness tool for reflection, not a medical or psychological diagnosis, and it does not replace care from a qualified professional. It is not couples counselling, not a test of whether a relationship should continue, and not designed for hiring. You can export or delete your data at any time.

Questions people ask

Do personality compatibility tests work?

Not as match scores. When actor, partner and similarity effects were separated in three nationally representative samples, a person's own traits predicted their relationship satisfaction by a wide margin, a partner's traits predicted less, and similarity added almost nothing (Dyrenforth et al., 2010). A trait-by-trait comparison is still useful, as a conversation rather than a verdict.

Which personality traits matter most in a relationship?

In a meta-analysis of couples, partners of people lower in Neuroticism reported the most satisfaction, with smaller links for Agreeableness, Conscientiousness and Extraversion and no reliable link for Openness to Experience (Malouff et al., 2010). Attachment anxiety and avoidance predict relationship quality a little better again (Noftle & Shaver, 2006), and the two overlap.

Can my partner and I take a personality test for couples together?

Each of you should answer for yourself, then read the two results side by side. Answering on someone else's behalf measures your view of them rather than their own report. In Wellington, both people take their own free Snapshot and a shared card fills in both sides, with no match score attached.

Is there such a thing as a friendship test?

Not in the sense of a validated score for a pair. Most relationship research is done on couples, and its clearest finding is about individuals: a person's own tendencies shape their relationships more than any match between two profiles. Comparing two honest trait profiles is a good way to name a recurring difference, which is what friendship tests are reaching for.