Short answer. People in a couple are more likely to resemble each other across many measured traits than would be expected if people paired up at random. The overlap is especially noticeable in education level, political and religious views, and some substance-use habits. For most personality traits and physical measures, the association is weaker. That is the result of a study published in Nature Human Behaviour in 2023: the authors combined data from 199 published studies on 22 traits and separately analyzed 133 traits among couples from UK Biobank.
This result does not mean that everyone is looking for a copy of themselves, that differences hurt relationships, or that similarity necessarily caused the relationship to form. The study measured correlations within already existing couples. Those can be affected by partner selection, shared environment, age, social circles, and changes over years of living together. Let’s break down the numbers, methods, limitations, and how to read this scientific result without popular but inaccurate simplifications.
What exactly did the researchers study
The study was conducted by Tanya Horwitz, Jared Balbona, Katie Paulich, and Matthew Keller. The researchers addressed two related questions: how similar are partners on the same trait, and does the strength of similarity vary across traits? The subjects were men and women in couples: spouses, cohabiting partners, engaged couples, and parents of a shared child. That definition matters: this is not about all dates, crushes, or potential partners, and the findings from the main study cannot be automatically extended to couples with different characteristics. In the original source the inclusion criteria are described in detail.
The study had two parts.
- Systematic review and meta-analysis. The authors identified 199 peer-reviewed publications released no later than August 16, 2022. From those, they extracted 480 estimates of similarity for 22 traits. For each trait, they calculated a pooled correlation accounting for differences across samples.
- UK Biobank analysis. The authors identified up to 79,074 presumed couples among UK Biobank participants. Initially, they considered 140 traits; for 133 there was enough data for the main analysis. The number of couples varied by trait and was not equal to 79,074 in every row of the results.
These are two lines of evidence, not one giant survey in which all people were measured on all 155 parameters. The meta-analysis synthesizes studies from different eras and populations; the biobank allows comparison of a wide range of measures within one large dataset. Reading them together makes the overall picture more convincing, but it does not remove the limitations of either part.
What is partner correlation
For each trait, the researchers compared one person’s value with the value of the same trait in their partner. In simple terms: if, in a sample, people with more education more often pair with people with more education, the correlation is positive. If “morning people” more often pair with “night owls” under the chosen chronotype coding, it may be negative. If values within a couple are barely related, the estimate is close to zero.
The correlation coefficient r ranges from −1 to +1. A positive sign indicates statistical similarity on the measure, a negative sign indicates statistical difference, and the magnitude shows the strength of the association in a specific sample and under a specific measurement approach. For example, r = 0.08 and r = 0.58 are both positive values, but the first indicates a much weaker association. Neither means “8%” or “58% identical people.” You also cannot convert r into the probability that two people will like each other.
There is an additional nuance: for continuous, ordinal, and binary measures, the authors used different types of correlation — Pearson, Spearman, and tetrachoric. So comparing the numbers requires attention to how the trait was defined and measured. The tables in the paper show sample size and estimate uncertainty; a single coefficient without that context does not. Methods and the results table are available in the full paper.
Main takeaway: similarity is more common than difference
In the meta-analysis, 18 of the 22 average correlations were positive and statistically significant after correction for multiple testing. For the remaining four traits, each estimate was based on only four samples, and the uncertainty was higher. In the UK Biobank analysis, 118 of 133 correlations were significant after the same correction: 115 positive and three negative. In other words, the phrase “partners are similar on 89% of traits” describes the share of tested measures with a statistically significant association in this sample, not the share of matching traits in any specific couple. The authors present the distribution of results in the UK Biobank section.
The average correlation for the 133 traits in UK Biobank was 0.19. That hides a wide range: some measures were strongly associated, others only faintly, and a few in the opposite direction. So it is more accurate to say “similarity is widespread, but its strength depends on the trait” than “people choose their doubles.”
Where similarity is strongest
In the pooled results, the following traits stood out in particular:
| Trait in the meta-analysis | Average correlation between partners | How to read the result |
|---|---|---|
| Political views | 0.58 | One of the strongest associations among the traits studied |
| Religiosity | 0.56 | Clear similarity, but with a wide uncertainty interval |
| Education level | 0.55 | Substantial similarity across samples |
| IQ score | 0.44 | The association was estimated from fewer couples than education |
| Smoking status | 0.49 | Noticeable overlap in the habit |
| Height | 0.24 | Positive, but moderate or weak association |
| Extraversion | 0.08 | Weak positive association |
These are rounded estimates from Table 1 of the original study. They refer to different sets of studies; they should not be read as a ranking of “what matters in love.” A stronger correlation does not prove that people consciously look for that exact trait when they meet.
Views and education
Political and religious beliefs matter for everyday decisions: who to spend time with, where to live, how to spend free time, and which rules to treat as important. People often meet in environments already filtered by those traits: at school, in professional circles, in local communities, or in religious organizations. So the similarity observed may arise both from preferences and from the structure of social encounters. The original meta-analysis does not separate the contribution of these causes for each specific trait.
Education is a special case. Couples with similar education levels may form because people meet while studying, work in similar fields, live in the same neighborhoods, or build a shared life around similar plans. Direct preferences are also likely. A correlation of 0.55 does not let us pick one explanation and declare it the main one. It shows a stable statistical similarity across the studies collected.
Habits: smoking and alcohol use
The correlation by smoking status in the meta-analysis was 0.49. The correlation for actually starting to smoke was 0.38. For a number of alcohol and other substance use measures, the estimates were also positive, but some of the relevant meta-analyses were small or had wide confidence intervals. The authors specifically note: not all estimates in this group survived the correction for multiple comparisonsdespite relatively high point estimates. See the results and discussion.
A couple’s shared lifestyle can gradually strengthen similarity in habits, or it may simply reflect people selecting from the same social environment. Observing a link does not show who influenced whom. And it certainly does not prove that a shared habit is inevitable: group statistics describe a trend, not the behavior of every individual participant.
Personality: there is similarity, but it is weaker
The five well-known personality dimensions in the meta-analysis did not show as strong a match as education or attitudes. For extraversion, the average correlation was 0.08; for neuroticism and agreeableness, 0.11 each; for conscientiousness, 0.16; and for openness to experience, 0.21. This is an important clarification to the intuitive phrase, “we are very similar in character.” People may feel close in values or lifestyle while still differing in temperament. All five estimates are given in the original source.
A weak correlation does not mean personality is unimportant in relationships. The study answers a different question: how similar already existing partners are on a measured trait. It did not test which traits support relationship satisfaction, help a couple get through conflict, or predict a breakup. Transferring the extraversion number to relationship quality would be a logical error.
Appearance and health
For height, the average correlation in the meta-analysis was 0.24; for body mass index, 0.16; and for waist-to-hip ratio, 0.17. For depression and diabetes, the corresponding estimates were 0.15 each. These links are smaller than those for attitudes and education, but in large datasets they are often different from zero. In the UK Biobank, the results for height, BMI, and waist-to-hip ratio were also positive: 0.27, 0.25, and 0.19, respectively. Comparing the two parts of the study shows an overall trend, but not a complete match for every estimate.
For health, special caution is needed. If a certain condition is more common in both partners, that may reflect shared age, living conditions, social environment, mutual influence, selection effects, or a combination of factors. You cannot conclude that an illness was “passed” from one partner to the other, and you cannot use such correlations to diagnose a couple. A meta-analysis describes population-level associations, not medical causation.
Do opposites really attract?
The popular saying is too broad to be tested honestly with a single number. If taken literally — as though stable couples are formed, on average, between people with opposite values on most traits — the study does not support that picture. Positive links were far more common than negative ones. But if the point is that two specific people may like their differences, the data do not rule that out.
A couple may differ in extraversion, daily routine, hobbies, or communication style and still share important views. The opposite can also be true: people may answer similarly on a questionnaire but behave differently in a relationship. Group-level correlation does not capture the full complexity of individual compatibility. More importantly, statistical similarity is not the same as attractiveness: the study covered established couples, not an experiment in which people were randomly shown potential partners and their interest in meeting was measured.
There is another subtle point here. If partners are similar on a certain trait, that is not necessarily the result of a preference like “I want someone with the same trait.” For example, people may have met in the same professional circle, and similarity in income, schedule, and attitudes may have emerged as a result of that shared environment. For that reason, the scientific concept of assortative mating should not automatically be applied to every correlation that is found.
Exception: chronotype and other negative links
The most notable negative correlation in the UK Biobank analysis concerned chronotype — the tendency to be a “morning person” or an “evening person.” In the coding method chosen by the authors, it was −0.18. The researchers identified two more statistically significant negative links among 133 measures, while 115 significant links were positive. These figures are published in the biobank analysis results.
A negative sign does not justify advising people to choose a partner with the opposite sleep schedule. It also does not prove that having contrasting routines helps relationships. A correlation can have several explanations, and the measurement method and sample composition can change the estimate. In everyday life, a different schedule will be convenient for some couples and inconvenient for others. The study did not measure that convenience.
The very strong positive correlation of 0.87 referred to the year of birth. It looks impressive, but its meaning is quite simple: partners are often close in age. This measure should not be grouped with “similarity in character” as evidence of near-total psychological match.
Why partners may become similar
The same outcome — similar measures for two people — can arise in several ways. The authors of the original source describe mechanisms that may act at the same time.
Direct selection on a trait
A person may prefer a partner with similar views, habits, or life goals. Such initial matching is called phenotypic assortative mating: similarity appears already when the couple is formed. However, from data on an existing couple, it is usually impossible to tell how much was due to deliberate preference and how much to the circumstances of meeting.
Shared social environment
People do not meet a random sample of the entire population. Their social circles are shaped by age, place of residence, education, work, language, interests, and income. That structure itself creates statistical similarity, even if a person never stated any requirements for a partner’s education or occupation. This mechanism is often discussed as social homogamy. It helps explain why a correlation does not have to reflect a romantic “selection criterion.”
Similarity through a related trait
Sometimes partners are similar on a measure indirectly: both entered the same environment for another reason. For example, a shared habit may go along with similarity in age, education, or living conditions. The observed link between partners on the first characteristic is then partly driven by the second. Understanding the causal structure requires special studies, not just a list of correlations.
Convergence after the relationship begins
Living together, people can change their habits, routines, and the way they answer questions. They are influenced by shared circumstances and mutual influence. This is called convergence. It differs from the original choice: at the start of a relationship, partners may have been more different. A snapshot of data at one point in time does not reliably separate convergence from similarity that existed before the couple met.
A separate 2023 paper by Syaarda and Kutalik based on UK Biobank data examined possible reasons for similarity and showed the role of selection, shared factors, and, for some measures, within-couple dynamics. But the authors also emphasize the limits of data on time spent living together: without measurements before the couple formed, it is difficult to clearly separate the mechanisms. This is consistent with the cautious interpretation of the large meta-analysis.
What questions the study does not answer
The scale of the work is impressive, but it does not turn correlations into universal laws. To apply the findings sensibly, it helps to see several boundaries.
It cannot predict the compatibility of specific people. An average association by education or views does not produce an individual “compatibility score.” Two people can build a good relationship despite differences, and conflict despite many similarities. The study under discussion did not build or test a couple-matching algorithm.
It cannot establish the direction of causality. We see similarity among existing partners, but this study does not show the full path from meeting to the moment of measurement. Some characteristics may have existed before the union, others may have changed within it, and others may be linked to shared circumstances.
It cannot claim that similarity improves relationships. That requires separate data on satisfaction, relationship stability, communication quality, and other pre-specified outcomes. The authors discuss the literature on these questions, but their main analysis measures similarity in traits, not the benefits of similarity.
It cannot treat all results as equally precise. In meta-analyses across different traits, the number of couples and studies varied a lot. Some confidence intervals are wide. An estimate based on a few samples requires more caution than a result backed by many independent data sets.
It cannot automatically be generalized to all countries and generations. Most of the included studies involved participants from Europe and the United States. Representation from other regions was much smaller. UK Biobank reflects a specific group of volunteers and particular age cohorts. Even if the general tendency toward similarity holds up, the strength of a given association may be different in another society or historical period. The limitations are listed in detail by the authors.
Why a large sample still calls for caution
It is easy for readers to think that if a study includes tens of thousands of couples, the result must be definitive. A larger sample reduces random error within the selected group, but by itself it does not correct systematic differences between the group and the broader population. UK Biobank recruited volunteers of a certain age; overall, participants differed from the UK population in a number of social, behavioral, and medical characteristics. This is shown in the study on UK Biobank representativeness.
In the analyzed study, couples in the biobank were identified from indirect data. These were presumed partners, not all couples in the country. Moreover, both people had to be among the biobank participants. This selection scheme can affect the estimates. The researchers explicitly discuss this risk.
The meta-analysis, in turn, combined different time periods, societies, and measurement methods. The median I² heterogeneity statistic across 22 traits was 0.915, indicating substantial differences between samples. In plain language, the average correlation for a trait is not a universal constant. It summarizes results that can vary considerably from place to place. The authors also note that some of the heterogeneity is related to the high precision of large studies and to measurement features; a single I² value is not enough to explain the reasons for the differences.
Why “statistically significant” does not mean “strong”
In large samples, a researcher can detect even a small association with high confidence. That is why a significant correlation of 0.08 for extraversion and a significant correlation of 0.58 for political views answer two very different practical questions. The first suggests that the estimate is probably not zero under the analysis assumptions. The second additionally shows much stronger similarity. Mixing up these meanings creates the false impression that all 115 positive significant associations in UK Biobank are equally important.
The authors tested many measures at once. The more tests you run, the higher the chance of getting a “significant” result by chance. The paper used Bonferroni correction for confidence intervals and p-values. That makes the criterion stricter. But the correction does not answer the question of causality and does not turn a small coefficient into an effect that matters in everyday life. The statistical methods are described in the publication.
Why confidence intervals matter
A point estimate is the best calculation based on the available data, but it is not exact knowledge about all possible couples. A confidence interval shows the range of values compatible with the data under the statistical model. For example, the meta-analytic estimate for extraversion was 0.08 with an adjusted 95% interval from 0.03 to 0.13. For political views, the estimate was 0.58 with an interval from 0.50 to 0.64. These intervals help show not only the direction but also the degree of uncertainty. The numbers are taken from Table 1.
It is especially important to read carefully the indicators derived from a small number of samples. A large overall study size does not mean the same level of precision for each of the 22 meta-analyses. For some traits, the data are scarcer and the intervals wider. That is one reason not to build a flashy ranking based only on point estimates.
Why correlations cannot be turned into a “formula for love”
Suppose a reader wants to choose a partner based on the three highest rows in the table: political views, religiosity, and education. Such an algorithm is flawed for several reasons. First, these characteristics may be related to one another and to the social environment, so their effects cannot simply be added together. Second, the study did not test whether matches improve relationships. Third, personal preferences, life circumstances, and ways of negotiating are not reduced to measured scales. The table shows which couples were more common in the data, not which couples should be formed.
Nor should one draw the opposite conclusion: if similarity is weak for most psychological traits, then character does not matter. It may be that what matters is not identical extraversion, but a person’s attitude toward differences, communication rules, and support in difficult situations. But these are separate hypotheses that need to be tested in other studies. The meta-analysis discussed here does not directly confirm or refute them.
What changed in research after publication
The main 2023 article focused on mixed-sex couples. In in a separate study by the same researchers, published in 2025, trends in similarity were compared in presumed same-sex and different-sex UK Biobank pairs. This broadens the scope of the research, but it does not make the earlier sample “inclusive of everyone”: each result needs to be read within its own design and pair-definition criteria.
For readers, this is a reason to be more cautious about broad statements like “everyone chooses similar partners.” The scientific question is not just whether an average correlation exists. What matters is for which people, measured in what way, in what era, and under what social conditions it appears.
How to use the findings in everyday life
The study does not provide a ready-made list of must-have matches for a happy relationship. It can help you ask better questions about yourself and your relationship.
- Separate values from personality traits. Similar views and education are more common among partners than the same level of extraversion. So the phrase “we have different personalities” by itself says nothing about whether life goals align.
- Consider the context in which you met. People who met within the same circle may already have shared circumstances. Similarity after meeting does not always mean that each person was intentionally looking for someone alike.
- Don’t turn statistics into a prescription. If a couple has different sleep schedules or one person is more outgoing, that is not a scientific diagnosis of incompatibility. It is more important to discuss specific friction points and agreements.
- Don’t confuse similarity with relationship quality. Having shared habits can make some decisions easier and others more difficult. Data on actual relationship outcomes are needed to judge the benefit of any specific match.
- Look at the size of the differences. A moderate correlation across thousands of couples still leaves a huge amount of room for variety in individual relationships.
This is not a psychological test or a clinical recommendation. The study works better as a correction to popular ideas: couples are not random combinations of people, but their similarity is more complex than sayings about “opposites” or “two halves.”
Frequently asked questions
Is it true that partners match on 89% of traits?
No. In UK Biobank, 118 of the 133 studied correlations were statistically significant after correcting for multiple comparisons — that is 88.72% of the indicators tested. Most significant associations were positive. That share does not show how similar any single couple is or how many specific answers they matched on. The original source gives both the share and the coefficients themselves.
Which traits are most likely to match between spouses?
In the data reviewed, the strongest statistical similarity was seen for political and religious views, education, and some substance-use characteristics. Height and other physical measures also more often showed positive associations, but usually weaker ones. For Big Five personality traits, the average correlations were relatively small.
Does this mean opposites can’t be together?
No. The study describes average patterns in groups of established couples. It does not impose a ban on relationships between people with different traits, and it does not measure every form of attraction. Strong statistical similarity on one trait does not rule out differences on another.
Why are partners similar: do they choose each other, or do they change together?
Both mechanisms are possible. In addition, a shared social circle, place of residence, education, and other circumstances also play a role. The main meta-analysis cannot precisely separate the observed correlation into these causes. That would require data on people before they met and follow-up over time.
Can correlation be used to estimate the chance of a successful marriage?
No. Correlation between partners on a trait and predicting relationship quality are different tasks. This study did not develop an individual forecast of satisfaction or relationship duration. A table of scientific coefficients should not be turned into a “perfect partner” questionnaire.
Conclusion
A large review of studies and the UK Biobank analysis present a consistent overall picture: partners are more often similar than opposite on many traits, especially education, views, and some habits. But the strength of similarity varies widely, depends on the trait and the sample, and the correlation itself does not reveal the mechanism behind it. The most accurate reading of the study is that similarity within couples is common, but it is not a rule for every couple and it is not a measure of relationship quality.
Main source: Horwitz T. B. et al., Evidence of correlations between human partners based on systematic reviews and meta-analyses of 22 traits and UK Biobank analysis of 133 traits, Nature Human Behaviour, 2023, DOI: 10.1038/s41562-023-01672-z. The data and wording in the article were verified as of September 24, 2026.