Forget the proverbial wisdom: Opposites don’t really attract, study finds

We've all heard the common folk wisdom that when it comes to forming romantic partnerships, opposites attract. Researchers at the University of colourado, Boulder, contend that this proverbial wisdom is largely false, based on the findings of their sweeping September study, published in the journal Nature Human Behavior. The saying, "birds of a feather flock together," is a more apt summation of how we choose our partners.

“These findings suggest that even in situations where we feel like we have a choice about our relationships, there may be mechanisms happening behind the scenes of which we aren't fully aware,” said co-author Tanya Horwitz, a psychology and neuroscience graduate student at UCB. “We’re hoping people can use this data to do their own analyses and learn more about how and why people end up in the relationships they do.”

Horwitz et al. conducted a systematic review of peer-reviewed studies in the English language involving comparisons of the same or similar complex traits in partners, all published before August 17, 2022, with the oldest dated 1903. They excluded same-sex/gender partners, maintaining that these partnerships warranted a separate analysis since the patterns could differ significantly. The meta-analysis focused on 22 distinct traits. The team also conducted a raw data analysis of an additional 133 traits, drawing from the UK's Biobank dataset, one of the largest and most detailed in the world for health-related information on more than 500,000 people. All told, the study encompassed millions of couples spanning over a century: co-parents, engaged pairs, married pairs, and cohabitating pairs.

The personality traits included were based on the so-called Big Five basic personality traits: neuroticism, extraversion, openness, agreeableness, and conscientiousness. (The Big Five is currently the professional standard for social psychologists who study personality. Here's a good summary of what those traits mean to psychologists.) The other traits studied included such things as educational attainment, IQ score, political values, religiosity, problematic alcohol use, drinking, quitting smoking, starting smoking, quantity of smoking, smoker status, substance use disorder, BMI, height, waist-to-hip ratio, depression, diabetes, generalized anxiety, whether they were breastfed as a child, and age of first intercourse, among others.

The meta-analysis and Biobank analysis revealed that the strongest correlations for couples were for birth year and traits like political and religious attitudes, educational attainment, and certain IQ measures. Couples tend to be similar when it comes to their substance use, too: heavy drinkers tend to be with other heavy drinkers, and teetotalers tend to pair with fellow teetotalers. There were a handful of traits among the Biobank couples where opposites did seem to attract, most notably whether one is a morning person or a night owl, tendency to worry, and hearing difficulty.

The weakest correlations were for traits like height, weight, medical conditions, and personality traits, although these were still mostly positive, apart from extroversion, which somewhat surprisingly showed almost no correlation. “People have all these theories that extroverts like introverts or extroverts like other extroverts, but the fact of the matter is that it’s about like flipping a coin," said Horwitz. "Extroverts are similarly likely to end up with extroverts as with introverts."

Horwitz et al. cautioned that even the strongest correlations they found were still fairly modest. As for why couples show such striking similarities, the authors write that there could be many reasons. Some people might just be attracted to similar sorts, or couples might become more similar over time. (The study also found that the strength of the correlations changed over time.) Perhaps two people who grow up in the same geographical area or a similar home environment might naturally find themselves drawn to each other.

The authors were careful to note several limitations to their meta-analysis. Most notably, most of those partners sampled came from Europe and the United States, with only a handful coming from East and South Asia, Africa, Latin America, and the Caribbean. Furthermore, all participants in the UK Biobank dataset were between the ages of 40 and 69 when they were originally recruited, all of whom were less likely to smoke, be socioeconomically deprived, or drink daily. The studies included in the meta-analysis also varied widely regarding sample sizes used to draw correlations across traits. For these reasons, the authors caution that their findings "are unlikely to be generalizable to all human populations and time periods."

Nature Human Behavior, 2023. DOI: 10.1038/s41562-023-01672-z  (About DOIs).