A study across 45 countries with 14,399 people confirmed: men and women say they want different things in a partner, with the same patterns showing up almost everywhere (Walter et al., 2020).
Then a study of 9,000+ blind dates found those stated differences mostly vanish when people actually meet (Eastwick et al., 2025).
This note is about that gap.
The Working Theory
Why might preferences differ at all? The dominant explanation comes from Robert Trivers’ parental investment theory (1972).
In mammals, the sex that invests more in offspring becomes the choosier one:
- Female humans invest enormously: pregnancy, lactation, years of dependency
- Male humans can biologically invest the bare minimum: minutes
- Different selection pressures over hundreds of thousands of years → different psychological adaptations
This is the field’s working hypothesis. Not gospel. Increasingly contested. But it’s the lens most evolutionary psychology runs through.
A note before going further: this entire chapter describes statistical tendencies, not moral entitlements. Online communities, especially the so-called “manosphere,” have weaponized this research to justify misogyny (Manne et al., 2023). The findings describe averages, not who deserves what.
What People Say They Want
When asked directly, men and women across cultures give consistently different answers:
- Men, on average, rate physical attractiveness and youth higher
- Women, on average, rate financial prospects, ambition, and slightly older partners higher
David Buss’s original 1989 study covered 37 cultures. The 2020 replication by Walter and colleagues covered 45 countries with preregistered methods. The patterns held.
The chart shows the direction of effect for one trait (physical attractiveness preference) across all 45 countries in Walter et al. 2020. Almost every dot lands on the same side of zero, meaning men consistently rate it higher than women, regardless of culture.
The cross-cultural consistency is real. When you ask people what they want, the gendered patterns show up almost everywhere.
What People Actually Choose
Here’s where the field has shifted dramatically in the last decade.
Stated preferences are weak predictors of actual romantic interest. When people meet face to face, those clean gender differences mostly disappear.
The biggest recent finding: Eastwick et al. (2025) analyzed 9,084 blind dates from a matchmaking service. The classic Buss finding says women prefer older partners. The actual data: women showed slight attraction to younger partners on dates, just like men did.
This isn’t isolated:
- Eastwick & Finkel (2008) and a 2014 meta-analysis showed that stated preferences predict actual romantic outcomes only weakly
- Online dating data (Bruch & Newman, 2018): both men and women weight attractiveness almost identically in initial decisions. The stereotype that “only men care about looks” doesn’t hold up in real behavior.
- A 1 standard deviation increase in stated intelligence raises match odds about 2%. Looks have roughly 10x the impact, for both sexes.
Two takeaways: What you say you want and who you actually choose are usually different people.
The biggest “gender differences” are in what people say, not what they do.
The One Asymmetry That Holds Up: Sexual vs Emotional Jealousy
If most stated preferences fade in real behavior, what’s left? One specific finding about jealousy stays remarkably stable.
The research isn’t asking whether men or women get more jealous overall (that’s a different question). It’s asking a very specific one. Imagine your partner cheats. Picture two scenarios:
- A: They have passionate sex with someone else, but feel nothing for them
- B: They form a deep emotional bond with someone else, but never sleep together
Which would upset you more?
When you ask thousands of people across many cultures, the answer splits sharply by gender:
- Men more often pick A as the worse betrayal (the sexual infidelity)
- Women more often pick B (the emotional infidelity)
This pattern, sometimes called the “sexual vs emotional jealousy asymmetry”, replicates across:
- Different measurement methods (forced-choice, continuous scales)
- Physiological measures (heart rate, sweat response, skin conductance)
- Multiple cultures
- Decades of replications, including in gender-egalitarian societies like Norway, where the effect is if anything larger
Most recent confirmations: Buss (2018) review and Kennair et al. (2021) community-sample study.
The evolutionary interpretation:
- Men face paternity uncertainty. A man can never be 100% sure a child is genetically his without a test. Sexual infidelity is the direct threat to that certainty.
- Women face resource uncertainty. A woman is always certain a child is hers, but not certain her partner will stay and provide. Emotional bonding with someone else is the direct threat to that.
The mechanism (whether evolutionary or socially conditioned) is still debated. But the pattern itself, in which type of infidelity triggers the bigger reaction, is one of the most robust findings in the entire field.
Famous Claims That Have Weakened
Pop psychology loves to repeat findings that seem universal. Two famous examples have aged poorly:
The “0.7 Waist-to-Hip Ratio” Universal
Devendra Singh claimed in 1993 that men universally prefer a WHR around 0.7. This became one of the most cited findings in evolutionary psychology and got passed around uncritically for decades.
Bovet’s 2019 systematic review of 104 studies concluded the universality claim is poorly supported:
- Different cultures show significantly different preferred WHR values
- Body mass index is a stronger predictor of attractiveness than WHR
- Many WHR effects disappear when BMI is controlled
The “universal 0.7” turned out to be largely a Western finding generalized too far.
“Gender Equality Erases All Differences”
Eagly and Wood (1999) proposed an alternative biosocial model: gender differences in mate preferences should shrink as societies become more egalitarian. The idea was that the differences are mostly cultural conditioning, not deep biology.
The evidence is mixed:
- The 45-country Walter et al. (2020) study did not find that gender equality robustly predicted the magnitude of sex differences. A blow to the strong version of the biosocial model.
- BUT a 2025 PNAS study (Conroy-Beam et al., 2025) found that resource preferences specifically do shift with economic conditions. When women have more economic power, they care less about a partner’s resources. Age and attractiveness preferences don’t shift as much.
The modern picture: some preferences are flexible (resources). Some are stable (jealousy, attractiveness weighting). It’s not “all biology” or “all culture.”
The Field Is Moving On
The traditional frame was “males compete, females choose.” That’s increasingly being replaced by Mutual Mate Choice (MMC) models: both sexes are choosy, both sexes compete, and the differences are smaller than the classic Buss framing implied.
The honest current summary:
- Men and women differ in what they say they want (robust across cultures)
- Men and women differ less in what they actually do (especially in real-world behavior)
- Some preferences are evolutionarily stable (jealousy patterns)
- Some preferences flex with economic context (resource preferences)
- Some claims that became famous don’t hold up (universal WHR)
Why This Matters (Self-Defense)
Three concrete things to take from this chapter:
1. Don’t fully trust your stated preferences.
What you say you want and what you actually click with diverge. The “type” you’re convinced you have? Probably less reliable than you think.
2. Attraction systems were calibrated for ancestral environments, not modern life.
Your brain still scans for fertility cues, status markers, resource indicators. None of these were designed for 21st-century compatibility. The most attractive person can be a terrible long-term match.
3. Manipulation exploits these systems on purpose.
- Dating apps are weaponized visual attraction
- Advertising hijacks status signals constantly
- “Negging” exploits female status assessment
- Conspicuous wealth signals trigger resource-detection
- Beauty filters amplify ancestral fitness cues
Knowing the hardware doesn’t disable it. But it gives you a moment of pause when the hardware fires: “Is this attraction telling me about a person, or about a 200,000-year-old reflex?”