The Oldest Algorithm: What the Science of Age-Gap Attraction Actually Says, and What It Does Not Excuse

On the reproductive imperatives encoded in every testosterone cascade, the data from 130 countries that simultaneously confirms and complicates the popular narrative, and what three decades of examining digital evidence in exploitation cases teaches about the distance between what biology explains and what it permits.

In the course of my work examining digital evidence in sexual exploitation cases, I have more than once encountered a particular species of defense exhibit: a stack of evolutionary psychology literature, carefully bound, arguing that the defendant’s attraction to significantly younger women was not a pathology but a feature of the human male brain shaped over millions of years of selection pressure, and that one therefore cannot speak of criminal intent in the usual sense when biology itself is the architect. I have read these arguments carefully, because they are not entirely wrong, which is precisely what makes them interesting, and precisely what makes them dangerous. The evolutionary biology is sound. The inference from the biology to the legal conclusion is not. Understanding why requires spending some time with both the data and the gap between what data explains and what it excuses.

The Algorithm Is Real: What 130 Countries Actually Found

Let us begin where the science actually begins, which is not with courtroom exhibits but with one of the most replicated findings in cross-cultural psychology. Buss, in a 1989 study that remains a reference point in the field, recruited 10,047 participants across 37 cultures to examine mate preferences, and found that across every single one of those cultures, men preferred younger mates while women preferred older mates with higher social status (Buss, D.M., 1989, “Sex Differences in Human Mate Preferences: Evolutionary Hypotheses Tested in 37 Cultures”, Behavioral and Brain Sciences, 12(1), 1-49). This is not a contested result. It has been replicated, extended, and subjected to extensive criticism over the subsequent 35 years, and it has survived all of it as a finding, though the interpretation of the finding remains the subject of legitimate scientific debate.

A more recent study with an even larger dataset confirmed the directional consistency. Ausubel and colleagues, analyzing data from 130 countries in a 2022 paper in Population Studies, found that across every religious group and every continent surveyed, men were on average older than their female partners (Ausubel, J., Kramer, S., Shi, A.F., & Hackett, C., 2022, “Measuring age differences among different-sex couples: Across religions and 130 countries, men are older than their female partners”, Population Studies, 76(3), 465-476). The worldwide average age gap was 4.2 years, with men older. Regional variation was substantial: North America reported an average gap of 2.2 years, Europe 2.7 years, while sub-Saharan African countries in polygamous traditions reported gaps exceeding 12 years, with Gambia recording the highest average at 14.5 years. In monogamous societies the average was 2.8 years, in polygamous societies 6.4 years.

A European-focused longitudinal study by Gottfried and colleagues, published in Personal Relationships in 2024 and using data from 35,996 participants surveyed in the Survey of Health, Ageing, and Retirement in Europe, found a telling dynamic: for men, the projected age of their partner decreased by approximately 1 year for every 5 years of their own age, meaning a 50-year-old man in the European sample was partnered with a woman roughly 10 years younger on average than a man who first partnered at 25 (Gottfried et al., 2024, “Couples age discrepancies in a large-scale European sample: Evolutionary and sociocultural perspectives”, Personal Relationships, 31(4), 987-1000). Women showed a similar but smaller effect.

88 percent of heterosexual married couples worldwide have the husband older than the wife by at least 1 year. 30 percent of marriages globally feature an age gap of 5 or more years with the male partner older. These are facts of demography, not of ideology, and they demand explanation.

Trivers, Parental Investment, and What the Theory Actually Predicts

The standard evolutionary explanation for this pattern begins with Robert Trivers’ 1972 paper on parental investment theory, one of the most influential papers in 20th century biology (Trivers, R.L., 1972, “Parental investment and sexual selection”, in Campbell, B.G. (Ed.), Sexual Selection and the Reproductive Competition in Mammals, 136-179, Academic Press). Trivers’ insight was that the sex which invests more in reproduction, in terms of time, energy, and biological cost, will be the more selective of the two when choosing partners, while the sex which invests less will compete more intensely for access to the higher-investing sex.

In humans, this asymmetry is pronounced and physiologically fixed. A woman’s reproductive potential is bounded by the number of eggs she produces over a lifetime, declines measurably in her early thirties, and terminates entirely at menopause. A man’s reproductive potential, which depends on continuous sperm production, declines more gradually and extends considerably further in life, though sperm quality deteriorates with age in ways that carry their own clinical implications. From this asymmetry, evolutionary theory predicts that men will have evolved psychological mechanisms oriented toward cues of female fertility, and that those cues will tend to correlate with youth: clear skin indicating hormonal health, facial symmetry correlating with developmental stability, waist-to-hip ratio indicating reproductive anatomy. These are not preferences men consciously select; they are, in the language of evolutionary psychology, adaptations, preferences that increased reproductive success often enough over deep evolutionary time to become characteristic of the species.

Kenrick and Keefe confirmed in a 1992 analysis that male age preferences in actual partner selection, as opposed to stated preferences, remain consistently oriented toward women in the peak fertility range even as the men themselves age, while women’s preferences shift with their own age, tending toward men who are somewhat older but tracking fairly close to their own cohort (Kenrick, D.T. & Keefe, R.C., 1992, “Age preferences in mates reflect sex differences in human reproductive strategies”, Behavioral and Brain Sciences, 15(1), 75-91). This asymmetry between how age preference shifts for men versus women has been documented across multiple datasets and is not plausibly explained by social conditioning alone.

The criticism of this framework from within evolutionary biology itself is not that the pattern is wrong but that the causal story is incomplete. Modern contraception has decoupled sex from reproduction for roughly 70 years in developed countries, meaning that the fertility-tracking mechanisms, if they exist, are now producing preferences that operate in a reproductive environment entirely unlike the one in which they evolved. This matters because it is possible for an evolved algorithm to produce outputs that are suboptimal or actively harmful in a novel environment while still being a genuine feature of the evolved psychology. The fact that male brains consistently orient toward youthful cues does not require those brains to be actively calculating fertility; it simply requires the orientation to have been selected for in environments where youthful cues reliably predicted fertility. Whether that orientation is appropriate in a 21st-century European context where most such partnerships will not produce children is a question evolutionary theory is not equipped to answer, because evolutionary theory describes what is, not what ought to be.

What the Brain Is Actually Reading, and Why It Cannot Stop

The neuroscientific basis of these preferences is well documented. Testosterone, the primary androgen in both male and female physiology though in dramatically different concentrations, correlates strongly with male orientation toward fertility cues. Roney and Simmons documented in 2008 that men’s testosterone levels respond measurably to interaction with women, with the response modulated by the perceived attractiveness and age of the women involved (Roney, J.R. & Simmons, Z.L., 2008, “Men’s facial masculinity predicts changes in testosterone levels when interacting with women”, Evolution and Human Behavior, 29(6), 365-369). The neuroendocrine loop here is not trivial: testosterone simultaneously amplifies the perception of attractiveness cues and is amplified by exposure to them, which is not a self-correcting system.

The dopamine dimension is equally important. Exposure to attractive young individuals activates the ventral tegmental area and the nucleus accumbens, the core of the brain’s reward circuitry, producing the neurochemical signature of anticipatory pleasure that also characterizes food anticipation, gambling, and, in elevated concentrations, drug use. This is not a metaphor. The machinery is the same. What this means practically is that the subjective experience of attraction to young women in an older man is, at the neural level, genuinely difficult to disengage from by an act of will, in the same way that the subjective experience of wanting to eat is not easily resolved by information about nutritional content. The brain does not process the desire and the counter-argument through the same circuits.

This is where Terror Management Theory, developed by Greenberg, Pyszczynski, and Solomon in the 1980s, contributes a dimension that pure reproductive psychology misses (Greenberg, J., Pyszczynski, T., & Solomon, S., 1986, “The causes and consequences of a need for self-esteem: A terror management theory”, in Baumeister, R. (Ed.), Public Self and Private Self, 189-212, Springer). The theory proposes that humans develop psychological defense mechanisms in response to awareness of their own mortality, and that one of these mechanisms is affiliating with entities that symbolize youth, vitality, and temporal continuity. An older man’s attraction to a substantially younger woman can carry, in this framework, an additional layer that has nothing to do with reproduction: the psychological relief of proximity to someone whose horizon is further away than his. This is not a conscious calculation. It surfaces as feeling, and it is genuinely difficult to distinguish from other kinds of feeling, which is part of why self-report data on this subject needs to be read with some caution.

The prefrontal cortex complicates this picture from the other side of the age gap. The prefrontal cortex, which governs impulse control, long-term planning, and the weighing of consequences, completes its myelination and functional development not in adolescence but at around age 25, a fact established rigorously by Casey, Jones, and Hare among others (Casey, B.J., Jones, R.M., & Hare, T.A., 2008, “The adolescent brain: Self-control and social decision-making in adolescence versus adulthood”, Annals of the New York Academy of Sciences, 1124(1), 111-126). A 19-year-old making decisions about a relationship with a 45-year-old is doing so with neural hardware that is factually not yet fully commissioned for the purpose of evaluating long-term relational consequences under conditions of significant resource and experience asymmetry. This is not an excuse and not a determinism. It is a datum that should be in the room during any honest discussion of consent in large age-gap partnerships.

The Halo Effect, Confirmation Bias, and the Cognitive Architecture of Desire

Several well-documented cognitive biases compound the neurobiological picture in ways that are worth naming precisely, because they help explain why the preferences described above persist and amplify even in environments where the reproductive logic that supposedly generated them is entirely absent.

The halo effect, first documented by Thorndike in 1920 in his studies of how military officers rated soldiers, describes the tendency to generalize positive assessments from 1 salient attribute to the whole person (Thorndike, E.L., 1920, “A constant error in psychological rating”, Journal of Applied Psychology, 4, 25-29). In attraction contexts this means that physical markers of youth, which are visually salient, activate generalized positive evaluations of personality, intelligence, and character for which there is no evidential basis. The young woman at the adjacent table in the pizzeria, noted in passing, is not being evaluated by the observing older man on the basis of her personality, which is unknown, or her values, which are similarly unknown, but on the basis of physical characteristics that trigger a cascade of generalized positive attribution. This is not a character flaw; it is the halo effect operating as it always operates, and being aware of it does not make it stop.

Zajonc’s mere exposure effect, established in 1968, documents that repeated exposure to a stimulus increases preference for it regardless of the intrinsic properties of the stimulus (Zajonc, R.B., 1968, “Attitudinal effects of mere exposure”, Journal of Personality and Social Psychology, 9, 22-27). In cultures saturated with media images pairing young women with older, powerful men in romantic contexts, this means that the pairing itself acquires familiarity and thus preference-enhancement through sheer repetition. The brain habituates to seeing this configuration and encodes it as normal, which is not the same as it being natural in any biologically meaningful sense, and is most certainly not the same as it being ideal by any evaluative standard.

Confirmation bias, documented by Nickerson among many others (Nickerson, R.S., 1998, “Confirmation bias: A ubiquitous phenomenon in many guises”, Review of General Psychology, 2(2), 175-220), ensures that once a preference is established, information supporting it is weighted more heavily than information contradicting it. A man whose preference for significantly younger women is strong will tend to discount incompatibility evidence, reframe generational difference as charming rather than problematic, and interpret the younger partner’s deference as evidence of personality compatibility rather than as a possible artifact of power differential.

37 Cultures Agreed. Economic Equality Is Changing the Terms.

The cross-cultural universality of the pattern is real, but the magnitude of the pattern is not fixed, and the evidence for what drives that magnitude is instructive. The gap is smallest in the most economically equal societies and largest where women have the least economic independence. This is not a confound; it is signal. In Northern European countries where women’s labor force participation, educational attainment, and earnings approach parity with men, the average spousal age gap is under 2 years. In sub-Saharan regions with severe female economic dependence and high fertility pressures, it exceeds 10 years. The Pew Research analysis of data from 130 countries found that religiously unaffiliated couples had the smallest average gap at 2.3 years, while Muslim-majority communities had the largest at 6.6 years, a pattern that tracks closely with indicators of female educational and economic participation (Pew Research Center, 2020, “Globally, women are younger than their male partners, more likely to age alone”).

This tells us that the evolutionary biology, whatever its role, is not operating in a vacuum. Cultural structures that constrain women’s economic options increase the advantage of partnering with a wealthier older man, and decrease women’s ability to exercise preferences independent of those constraints. This is the part of the story that evolutionary psychology, focused on universal mechanisms, underweights, and that sociology and feminist theory, focused on structural inequalities, overweights by treating the whole phenomenon as reducible to patriarchal control. Both are partially right and insufficiently complete.

The concept of hypergamy, the statistical tendency of women to partner with men of equal or higher social status, is documented cross-culturally, though its interpretation is contested. Resource-based mate selection theory holds that this preference, like male fertility-orientation, is an evolved strategy reflecting the conditions under which female reproductive success was historically maximized. Social structural theory holds that it reflects the rational response of women with limited economic options to a market in which male resources provide security that female earnings cannot always match. The honest answer, given the data, is probably that both processes operate, that they are difficult to disentangle empirically, and that the relative contribution of each shifts as economic equality changes.

The Forensic Record: What the Science Does Not Excuse

The evolutionary biology, the neuroscience, and the cross-cultural data all speak to tendencies, distributions, and averages in populations. They say nothing at all about whether any given relationship is consensual, equitable, or legal, and this is where I want to be very direct, because in my experience of forensic practice the transition from population-level statistics to individual-level claims is where arguments go wrong in both directions.

The defense exhibit argument, that the evolutionary evidence diminishes culpability because the preference is natural, fails for a reason that would be obvious if we applied the same logic elsewhere. Aggression is also an evolved trait with extensive cross-cultural documentation. So is the tendency to prefer in-group members over out-group members in resource allocation. The documented universality of a behavioral tendency does not constitute a defense against having acted on that tendency in a context where doing so was harmful, non-consensual, or illegal. Evolution produces appetites; law and ethics govern what we do with them.

The cases that have been through forensic examination in connection with age-gap exploitation, and there are many, share structural features that have nothing to do with evolutionary biology and everything to do with the systematic exploitation of asymmetric power. Wealth differential, social status differential, emotional dependence cultivated through incremental trust-building, the controlled management of a younger person’s information environment and social relationships, and the use of material resources to create conditions in which exit from the relationship becomes practically difficult: these are the mechanisms. They are not natural; they are deliberate. And they frequently operate within relationships that are technically legal, which is why the legal question and the ethical question need to be held separately.

The prefrontal cortex data matters here in a specific way. An 18-year-old is legally an adult in most jurisdictions. An 18-year-old whose prefrontal cortex will not complete development for another 7 years is also making decisions about a relationship with a 55-year-old with incomplete neural hardware for assessing long-term relational consequences. This does not make the relationship illegal. It does mean that the informed-consent framework appropriate to 2 adults of similar age and experience does not map cleanly onto relationships where the developmental and experiential asymmetry is this large. Legal thresholds are bright lines drawn for reasons of administrative clarity; they do not resolve the ethics, and a forensic practitioner who has seen what these asymmetries look like in the evidence record is not inclined to treat the age of majority as a complete ethical answer.

The Feminist Critique: The Parts That Are Right and the Parts That Are Not

The feminist critique of age-gap relationships is not monolithic, and it is worth distinguishing the parts of it that are empirically well-grounded from the parts that are not.

The well-grounded parts are these: the systematic devaluation of aging women as partners and the simultaneous elevation of aging men as distinguished is a cultural artifact that has no biological necessity and substantial documented harm. The fact that older women pursuing younger men are socially coded as pathetic or predatory while older men pursuing younger women are socially coded as successful is not a product of evolutionary biology; it is a product of cultural encoding of a double standard. The media’s role in sustaining and amplifying this encoding is real and well-documented. The economic structures that historically and in many regions contemporarily make older, wealthier men attractive partners for women who lack independent economic options are structural constraints, not expressions of free preference, and treating them as if they were equivalent to purely voluntary choice is epistemically wrong.

The parts of the critique that are empirically weaker are the claims that any substantial age gap is necessarily a power imbalance and therefore necessarily ethically problematic, or that women who enter large age-gap relationships are by definition failing to exercise genuine agency. The empirical record includes relationships with large age gaps that are by any measure functional, mutual, and equitable, alongside relationships with negligible age gaps that are coercive and harmful. Age gap is a correlate of certain risk factors, not a determinant of relational quality. The divorce risk data is worth noting in this context: a 10-year age gap is associated with 39 percent higher divorce risk compared to same-age couples, and a 20-year gap with 95 percent higher risk. These are statistical headwinds, not destinies.

Ultimately, dismissing all large age-gap relationships as inherently exploitative does exactly what the evolutionary psychology defense does in the opposite direction: it treats a population-level statistical tendency as if it were a statement about every individual case. The ethics of a specific relationship depends on the specific conditions of that relationship, on the degree of genuine mutuality, the freedom of exit, the presence or absence of economic coercion, and the psychological dynamics between the individuals involved. These are empirical questions, not questions that can be answered by knowing the age difference alone.

A Word Before the Closing

I want to briefly note that this topic produces discomfort at both ends of the ideological spectrum, and that this discomfort is not itself a reason to distort the evidence in either direction. The reader who believes evolutionary psychology is an elaborate rationalization for male bad behavior will find the cross-cultural fertility-preference data uncomfortable. The reader who believes that evolutionary arguments settle the ethical question will find the discussion of structural coercion and the forensic record uncomfortable. Both of them are probably sitting in a restaurant right now, looking at the table across from them where an age gap is visible, and drawing conclusions that the data, properly understood, does not fully support.

The algorithm is real. The algorithm does not justify its outputs in all contexts. These 2 statements are simultaneously true, and holding them simultaneously without collapsing them into a single oversimplified verdict is precisely the analytical demand that this subject makes.

Closing

There are 4 things the science of age-gap attraction tells us with reasonable confidence. First, the preference of older men for younger women is cross-culturally documented, neurobiologically grounded, and not reducible to mere social conditioning. Second, the magnitude of that preference in any given society is substantially modulated by how much economic independence women have, which means the evolutionary tendency is real but its expression is not biologically fixed. Third, the biological explanation of a preference is not a legal or moral defense for acting on that preference in ways that cause harm or involve coercion, and this is not a complicated point. Fourth, the statistical risks associated with large age gaps, including higher divorce rates, greater power asymmetries, and the documented developmental incompleteness of the younger partner’s decision-making apparatus, do not make such relationships impossible to conduct well, but they do mean that the burden of demonstrating genuine mutuality falls more heavily on those relationships than it does on partnerships between developmental and experiential peers.

The evolutionary algorithm that orients male attention toward youth is older than Homo sapiens and will not be revised by this article. What can be revised is the quality of the conversation we have about it, specifically by insisting that explaining a tendency and excusing its misuse are 2 entirely different intellectual operations, and that confusing them serves neither truth nor the people most likely to be harmed by the confusion.

References

• Ausubel, J., Kramer, S., Shi, A.F., & Hackett, C. (2022). Measuring age differences among different-sex couples: Across religions and 130 countries, men are older than their female partners. Population Studies, 76(3), 465-476. https://doi.org/10.1080/00324728.2022.2094452
• Buss, D.M. (1989). Sex differences in human mate preferences: Evolutionary hypotheses tested in 37 cultures. Behavioral and Brain Sciences, 12(1), 1-49. https://doi.org/10.1017/S0140525X00023992
• Buss, D.M., & Schmitt, D.P. (1993). Sexual strategies theory: An evolutionary perspective on human mating behavior. Psychological Review, 100(2), 204-232.
• Casey, B.J., Jones, R.M., & Hare, T.A. (2008). The adolescent brain: Self-control and social decision-making in adolescence versus adulthood. Annals of the New York Academy of Sciences, 1124(1), 111-126. https://doi.org/10.1196/annals.1440.010
• Gottfried, E., et al. (2024). Couples age discrepancies in a large-scale European sample: Evolutionary and sociocultural perspectives. Personal Relationships, 31(4), 987-1000. https://doi.org/10.1111/pere.12579
• Greenberg, J., Pyszczynski, T., & Solomon, S. (1986). The causes and consequences of a need for self-esteem: A terror management theory. In Baumeister, R. (Ed.), Public Self and Private Self (pp. 189-212). Springer.
• Kenrick, D.T., & Keefe, R.C. (1992). Age preferences in mates reflect sex differences in human reproductive strategies. Behavioral and Brain Sciences, 15(1), 75-91.
• Nickerson, R.S. (1998). Confirmation bias: A ubiquitous phenomenon in many guises. Review of General Psychology, 2(2), 175-220.
• Pew Research Center. (2020). Globally, women are younger than their male partners, more likely to age alone. https://www.pewresearch.org/short-reads/2020/01/03/globally-women-are-younger-than-their-male-partners-more-likely-to-age-alone/
• Roney, J.R., & Simmons, Z.L. (2008). Men’s facial masculinity predicts changes in testosterone levels when interacting with women. Evolution and Human Behavior, 29(6), 365-369.
• Singh, D., & Young, R.K. (1995). Body weight, waist-to-hip ratio, breasts, and hips: Role in judgments of female attractiveness and desirability for relationships. Ethology and Sociobiology, 16(6), 483-507.
• Thorndike, E.L. (1920). A constant error in psychological rating. Journal of Applied Psychology, 4, 25-29.
• Trivers, R.L. (1972). Parental investment and sexual selection. In Campbell, B.G. (Ed.), Sexual Selection and the Reproductive Competition in Mammals (pp. 136-179). Academic Press.
• Zajonc, R.B. (1968). Attitudinal effects of mere exposure. Journal of Personality and Social Psychology, 9, 22-27.