Psychology by BogdanovaThe Encyclopedia of Psychology for the Artificial Era

RESEARCH DOSSIER

Human Mate Choice and Sexual Selection

Abstract. Human mate choice is the situated process through which people evaluate, pursue, accept, reject, and form relationships with possible partners. Sexual selection concerns evolutionary changes produced by differential access to mates and reproductive success, while contemporary human partner formation is jointly structured by individual preferences, mutual consent, exposure, demographic opportunity, norms, and institutions. This research monograph reconstructs Darwinian foundations, parental investment and sexual strategies models, social-role explanations, and interactionist accounts. It evaluates cross-cultural stated preferences, speed-dating decisions, ideal-preference matching, budget trade-offs, assortative mating, online dating, facial-attractiveness judgments, reproductive-hormone hypotheses, sexual orientation, and social inequality. A critical narrative synthesis of experimental, observational, longitudinal, meta-analytic, and cross-national evidence distinguishes population-average sex differences from individual prediction and separates expressed ideals from observed choice. It concludes that human mating is a reciprocally constrained, culturally organized and developmentally conditional process; identifying evolved mechanisms requires causal, historically informed and cross-population tests. The analysis also specifies methodological priorities for the Artificial Era of algorithmically mediated encounters.

Keywords: human mate choice; sexual selection; mate preferences; sexual strategies; mutual mate choice; assortative mating; speed dating; ideal partner preferences; parental investment; gender roles; dating markets; sexual orientation; cross-cultural psychology; evolutionary psychology; relationship formation; Artificial Era

The Scientific Object of Human Mate Choice

Human mate choice is a process of selective evaluation and reciprocal relationship formation in which possible partners are encountered, compared, approached, accepted, rejected, and sometimes chosen for continued intimacy or reproduction. The process unfolds through at least five analytically separable components: preferences for attributes, access to a pool of potential partners, behavioral decisions in actual encounters, mutual agreement to form a relationship, and the subsequent maintenance or dissolution of that relationship. Observing a couple provides information about the result of these interacting processes; asking an unpartnered person about desired qualities measures only part of the antecedent process. A scientific account must therefore specify whether its outcome is an expressed ideal, initial attraction, contact initiation, partner selection, cohabitation, marital sorting, sexual behavior, or reproductive success Eastwick et al. (2014) Schwartz (2013).

Sexual selection refers to the evolutionary consequences of differences in mating success and reproductive success associated with heritable traits. Evolutionary explanations of mate choice may concern long historical intervals, whereas psychology measures proximate cognition and behavior in contemporary environments. Claims that a person values attractiveness, kindness, earning potential, or social similarity are descriptions at one level; claims that a preference evolved because it affected ancestral reproductive success require additional comparative, genetic, historical, and fitness-related evidence. Darwin established the conceptual distinction between selection through mate competition and choice, but modern humans also confront culturally generated opportunities, contraception, institutions, and variable intentions regarding children Darwin (1871) Trivers (1972) Buss (2023).

The research object is accordingly a coupled system of individual evaluations, social opportunities, and joint decisions. The same trait may influence the probability of attracting attention, the likelihood of a reciprocal yes in a brief meeting, and the prospects of a stable partnership by different mechanisms. Preference measures alone cannot reveal how these stages are connected. This monograph uses critical narrative synthesis of foundational theory, original experiments, repeated cross-cultural surveys, meta-analysis, demographic records, and newer registered research. It asks what mate choice means empirically, what theoretical mechanisms plausibly explain recurrent patterns, when preferences predict behavior, how culture and ecology shape feasible decisions, and what uncertainty remains when evidence is transferred across populations and historical settings Fisman et al. (2006) Eastwick et al. (2025).

Defining Preference, Choice, Mating Strategy, and Sexual Selection

A mate preference is a relative evaluation of an attribute or combination of attributes in prospective partners. A preference can be expressed as an abstract rating, a ranking among alternatives, a budget allocation under constraint, an observed approach decision, or a difference in willingness to continue an interaction. These operationalizations need not measure an identical latent variable. An attribute can be judged desirable when considered alone and exert little effect in a realistic choice set where other attributes covary; a preference can also be strong but behaviorally unexpressed when the preferred partners are unavailable or uninterested. The critical distinction is between a marginal stated evaluation and a conditional behavioral decision Li et al. (2002) Todd et al. (2007) Zhao et al. (2025).

Mate choice is typically mutual in the human case. Each participant may express attraction while retaining the capacity to decline, negotiate, reconsider, or terminate a relationship. Competitive behavior, courtship effort, self-presentation, social endorsement, and accessibility influence the encounter before either person decides. The phrase mating strategy refers to recurrent, condition-sensitive patterns of attention, effort, selectivity, and relationship behavior rather than an immutable personality category. Strategic pluralism holds that people can alter their allocation of courtship and parenting effort as opportunities, constraints, relationships, and personal characteristics change; its hypotheses must be judged against observed decisions and outcomes Gangestad & Simpson (2000) Buss (2023).

Assortative mating describes nonrandom similarity or dissimilarity among established partners. Educational homogamy, similarity in religion or values, and age-related matching can result from active selection, structural segregation, shared institutions, or social networks that make some encounters more likely than others. A partner correlation measures a pattern of completed matching and cannot on its own identify which of those pathways produced it. Sexual conflict, in turn, concerns divergence between participants' reproductive or relationship interests under particular circumstances; disagreement does not establish a universal adversarial relationship between sexes. Both concepts are indispensable to an account in which paired outcomes emerge through selection on two sides Schwartz (2013) Buss & Schmitt (1993).

Darwinian Origins and the Evolutionary Problem

Charles Darwin's The Descent of Man, and Selection in Relation to Sex, published in 1871, provided a framework for explaining traits that could be costly for survival yet advantageous in competition for mates. Darwin distinguished contests among rivals from preferential responses by potential mates and proposed that the history of these interactions could alter appearance and behavior over generations. His work established explanatory questions about ornaments, displays, rivalry, and choosiness that have remained productive across evolutionary biology. The social and biological assumptions embedded in nineteenth-century descriptions, however, require critical historical interpretation rather than unexamined translation into present-day human categories Darwin (1871) Buss (2023).

Later theory connected sexual selection to differences in the relation between mating opportunity and reproductive success. Bateman's 1948 Drosophila experiments became influential for describing variation in mating and reproductive outcomes, and Trivers's 1972 parental-investment analysis emphasized the potentially different consequences of investment in offspring. Yet a theoretical model derives its predictions from assumptions about reproductive constraints, encounter rates, costs of choice, and the available routes to reproductive success. Those assumptions must be measured in each ecological setting. The importance of a classical study for the history of a field cannot substitute for an assessment of its design, computations, and sampling Bateman (1948) Trivers (1972).

Historical revision is part of cumulative science. Snyder and Gowaty's reassessment of Bateman's original experiment identified problems including sampling bias, pseudoreplication, statistical treatment, and the interpretation of fitness variation. Their critique weakens confidence in treating the original data as a uniquely decisive empirical law, while leaving open the study of sex-specific selection gradients with stronger methods. Human research also requires distinguishing effects of mating behavior from effects of contraception, parental arrangements, institutional power, and differential survival. An evolutionary narrative becomes scientifically informative when it yields measurable predictions that discriminate between causal alternatives Snyder & Gowaty (2007) Eagly & Wood (1999).

Parental Investment, Mutual Choice, and Reproductive Trade-Offs

Parental investment theory begins with a functional question: how does expenditure on current offspring alter a parent's capacity to invest in other offspring? Where reproductive investment differs systematically, mate competition and selectivity may differ in predicted ways. The human case requires specifying pregnancy, lactation, provisioning, cooperative care, kin assistance, reproductive age, partner commitment, and the varied costs of reproduction. Both prospective parents can invest substantially over time, and both may gain from selectivity. The resulting predictions concern conditional averages in particular ecological situations rather than categorical prescriptions about individual relationships Trivers (1972) Buss (2023).

A simple expectation that one sex chooses and the other competes fails to capture the full range of human courtship. Men and women may compete for desirable partners, exercise selective preferences, respond to partner resources or affection, and assess prospective investment. Selectivity may rise when candidates are numerous, when the risks of partnership are high, or when remaining unpartnered has relatively low costs. It may decline when alternatives are scarce or search is expensive. Evolutionary models therefore need parameters for the availability and quality of alternatives, relative searching costs, and the consequences of a joint relationship rather than a one-sided theory of preference Fisman et al. (2006) Gangestad & Simpson (2000).

Reproductive trade-offs are also affected by relationship horizons. A trait valued in an anticipated brief encounter may carry a different weight in co-parenting, where reliability, care, shared goals, and mutual responsiveness become consequential. The terminology of short- and long-term strategies can organize questions about context, but it does not by itself classify all forms of intimacy or determine their reproductive purposes. Many people form relationships without intending to reproduce, and observed romantic partnerships can involve same-sex partners, infertility, postponed childbearing, or deliberate childlessness. Psychological models must measure the intended relationship function instead of assuming it from participants' biological sex Buss & Schmitt (1993) Takayanagi et al. (2024).

Evolutionary Strategies and Social-Role Explanations

Sexual strategies theory proposed by Buss and Schmitt in 1993 describes possible sex-linked and context-linked adaptations in mating goals and selective criteria. Gangestad and Simpson's strategic-pluralism account further emphasizes costs, benefits, ecological variation, and trade-offs in reproductive investment. These perspectives generate hypotheses about what people value, how short- and long-term decisions differ, and why the same person might display multiple patterns over time. Their strongest empirical use is to identify predictions with observable boundary conditions, including potential competitors and mates, perceived resources, willingness to invest, and the stability of environmental cues Buss & Schmitt (1993) Gangestad & Simpson (2000).

Social-role and biosocial approaches analyze how divisions of labor, economic dependence, social expectations, and unequal opportunities can generate systematic differences in expressed partner preferences. Eagly and Wood's 1999 account advanced social-role mechanisms as an alternative to attributing all contemporary sex differences to fixed ancestral dispositions. The approaches need not be assessed solely by rhetorical plausibility. A useful test asks whether preferences change when gender inequality, employment opportunities, social norms, and economic security change; another asks which cross-national relationships persist after testing measurement equivalence and confounding by country-level variables Eagly & Wood (1999) Zentner & Mitura (2012).

These theoretical traditions explain different portions of the causal chain. A population-average preference difference may have a historical evolutionary component, a contemporaneous normative component, and an individual learning history. A strong narrative explanation supplies no quantitative estimate of their relative weights. The required comparison is therefore between formal models that state identifiable mechanisms and generate divergent predictions. Cross-cultural persistence can constrain an exclusively local socialization account, while reproducible change with institutions constrains an account of entirely invariant behavioral expression. Neither observation alone establishes the evolutionary origin of a psychological mechanism Walter et al. (2020) Eagly & Wood (1999).

Cross-Cultural Foundations of Mate Preference Research

Cross-cultural surveys offered a major empirical advance by moving beyond single-university samples and measuring the relative desirability of partner attributes in many social settings. Buss's 1989 investigation combined 37 samples from 33 countries, with a total of 10,047 participants, and reported recurring sex differences in average preferences concerning resource acquisition, age, and characteristics interpreted as reproductive cues. Its influence rests both on breadth of sampling and on translating evolutionary proposals into explicit survey predictions. Cross-national coverage nevertheless does not make every national sample population-representative, and translation, response style, and local interpretations of marriage can alter the measured construct. The appropriate inference is evidence of geographically recurrent preference patterns within the sampled groups Buss (1989).

A subsequent large-scale replication by Walter and colleagues examined 14,399 participants in 45 countries. It found recognizable average differences in ratings of attractiveness, age, and financial prospects, while investigating whether national context altered these associations. Replicating a direction of average differences is meaningful, particularly when source instruments and analysis methods differ, yet it cannot establish that all individuals conform or that the sizes of these differences remain invariant across historical periods. Statistical tests of moderation require attention to population composition, measurement reliability, multiple comparisons, and the ecological fallacy, which arises when associations between countries are ascribed to persons without sufficient evidence Walter et al. (2020).

The 2025 cross-cultural dataset assembled by Kowal and hundreds of collaborators contains 117,293 records from 175 countries, collected in 45 languages. This is an important resource for examining romantic experience, partner preferences, and associated social conditions. Its publication as a data descriptor establishes the existence, scope, and technical documentation of a dataset, rather than a single causal conclusion about why any preference occurs. Participation was not designed to yield representative national probability samples. Analysis must therefore record the recruitment context, completion patterns, country-level sample sizes, translation quality, and the exact predictor and outcome being tested before asserting universal tendencies Kowal et al. (2025).

Economic Prospects, Status, and Social Resources

Interest in a potential partner's economic prospects can reflect assessments of future security, lifestyle compatibility, perceived competence, social position, and capacity to contribute to a shared household. These interpretations should be separated because the same survey item can represent different psychological constructs to participants facing different institutions and life opportunities. Buss's cross-cultural questionnaire identified higher average ratings for financial prospects among women in its samples, while Walter and colleagues later replicated related sex differences across a larger set of countries. Neither study observes a randomized manipulation of lifetime earning potential or reproductive outcomes; both are principally studies of stated evaluation Buss (1989) Walter et al. (2020).

Social-role explanations predict that when earnings, education, and access to independent resources become more equal, partner preferences may change because the practical benefits and constraints attached to economic dependence change. Zentner and Mitura's cross-national analysis explored such moderation using indicators of gender gaps and mate preferences. These associations are pertinent but vulnerable to country-level confounding, sampling differences, historically correlated institutions, and ambiguity about the scale at which a predictor acts. A robust causal account would need policy variation, longitudinal evidence, or comparable natural experiments that distinguish changes in opportunity from concurrent shifts in values Eagly & Wood (1999) Zentner & Mitura (2012).

Behavioral evidence complicates the interpretation of a survey response as an exact decision rule. Eastwick and colleagues' meta-analysis found that declared priorities do not consistently reproduce the familiar sex-differentiated pattern when predicting observed romantic evaluations. In the worldwide registered report published in 2025, tests of earning potential showed systematic discrepancies between what participants said they valued and how ratings related to partners' attributes. Those findings do not establish that money never matters; instead they locate its effect within a network of observed attributes, feasible choices, and relationship judgments and imply that the design of preference instruments changes the question answered Eastwick et al. (2014) Eastwick et al. (2025).

Physical Attractiveness, Appearance, and Perceptual Judgment

Appearance contributes to initial attention and romantic evaluations through multiple distinguishable channels: features of the face and body, movement, grooming, expression, contextual presentation, and culturally learned aesthetic standards. The psychological task is to establish which signals participants perceive and how they weight them under a particular judgment. A judgment of facial appeal is different from choosing a dating partner after a conversation; a standardized photograph removes voice, humor, mutual interest, and conversational coordination. Meta-analytic work by Langlois and colleagues and Rhodes's review established broad research programs on agreement in beauty judgments, face perception, and possible developmental and evolutionary sources of attractiveness responses Langlois et al. (2000) Rhodes (2006).

Cross-cultural studies and laboratory evaluations repeatedly report that physical attractiveness matters to many participants, though the magnitude of its relevance depends on context, the attractiveness variance within a choice pool, and the available measurement scale. An abstract comparison between attractiveness and kindness may produce an order that changes when people experience those qualities simultaneously. The meta-analysis by Eastwick and colleagues found substantial associations of physically attractive qualities with romantic evaluation across male and female participants in observed-relationship studies, a pattern that complicates strong predictions about sharply different sex-specific revealed preferences. Accuracy also depends on whether appearance is rated by independent observers, by the individual making a choice, or by the eventual partner Eastwick et al. (2014) Eastwick & Finkel (2008).

Attractiveness judgments should not be treated as transparent biological tests of fertility, heritable fitness, or health. Some appearance cues may correlate with physiological characteristics in some populations, while social learning, familiarity, perceptual fluency, and cultural practices also alter evaluation. If a hypothesis proposes that a visual cue predicts reproductive fitness, it needs direct measures of the proposed mediator and consequential outcome, not only a preference for a digitally modified photograph. Face-manipulation paradigms are useful for detecting perceptual preferences; their ecological validity must be assessed separately from the causal interpretation of what the manipulated feature purportedly signals Rhodes (2006) Jones et al. (2018).

Kindness, Intelligence, Compatibility, and Partner Ideals

Ratings of kindness, intelligence, honesty, emotional support, and shared values describe evaluations of qualities that can contribute to relationship functioning as well as immediate attraction. Some are difficult to judge reliably in a brief encounter because participants infer enduring dispositions from sparse behavioral evidence. Consequently, stated importance may be high while immediate choices rely on readily observable cues such as appearance, warmth, conversational responsiveness, and self-presentation. A preference for kindness also requires an operational definition: hypothetical descriptions, observed cooperative behavior, and perceived partner support each contain different information and different potential measurement biases Li et al. (2002) Eastwick et al. (2025).

Budget-allocation procedures introduce explicit scarcity. Li and colleagues' necessities-and-luxuries paradigm asks participants to allocate constrained resources to potential attributes, thereby probing whether people prioritize minimum levels of particular characteristics before spending remaining resources elsewhere. Such tasks illuminate trade-offs hidden by questionnaires in which every positive trait can receive a high rating. The allocation, however, is still hypothetical and depends on the supplied trait list, points available, and the respondent's understanding of the budget analogy. Correspondence with actual partner decisions requires independent validation in natural encounters or relationship records Li et al. (2002) Conroy-Beam & Buss (2016).

A study of 778 Brazilian participants spanning different sexes and sexual orientations found that intelligence and kindness ranked prominently in budgeted partner preferences, with variation across groups and relationship contexts. Takayanagi and colleagues measured heterosexual, bisexual, and homosexual respondents and reported that familiar average sex differences were concentrated especially in heterosexual groups in their sample. This design demonstrates why sexual orientation and desired relationship horizon are substantively relevant variables. The sample and hypothetical instrument limit claims about behavioral selection or universal population orderings. Nevertheless, it challenges simple models that equate sex category with an invariant preference profile Takayanagi et al. (2024).

Sex Differences, Individual Variation, and Gendered Contexts

Population-average differences in partner preference distributions are compatible with considerable overlap between individuals. The statistical difference between group means does not describe the preference structure of every member of either group, nor determine the relative importance of the trait when people choose among real partners. Variation within groups may depend on age, socioeconomic resources, relationship goals, social norms, sexual orientation, prior relationship experience, and the composition of the accessible partner pool. Cross-national replication provides evidence that some aggregate patterns recur, while investigations of actual choice ask how strongly those patterns predict individual behavior Walter et al. (2020) Eastwick et al. (2014).

The sex and gender variables used by researchers carry different operational meanings. Reproductive anatomy, sex assigned at birth, gender identity, social gender role, sexual orientation, and the gender of desired partners may be correlated in a given dataset but cannot be substituted freely in analysis. Older binary questionnaires commonly restrict who can participate or how they can describe a partner. A modern research design should distinguish these variables according to the hypothesis and identify where its observations are drawn from mostly heterosexual different-sex participants. Research involving non-heterosexual and gender-diverse respondents can test whether a proposed mechanism concerns reproduction, social position, attraction targets, or culturally specific relationship scripts Takayanagi et al. (2024) Eagly & Wood (1999).

Cross-cultural analyses can simultaneously observe convergence and divergence. A recurrent ranking of attributes does not imply that societies supply the same incentives, marriage institutions, partner access, or meaning of romance. The large multi-country resources now available increase statistical power and descriptive coverage, but unequal recruitment, measurement translation, and limited within-country representativeness mean that broad geographic reach alone is not a causal or universal test. The productive question is which associations reproduce under standardized measurement, which vary with identified contexts, and which survive explicit adjustment for sample composition and social structure Kowal et al. (2025) Zentner & Mitura (2012).

Stated Preferences and Revealed Attraction

The stated–revealed distinction organizes one of the most important methodological findings in modern mate-choice psychology. Stated preferences are responses to hypothetical questions about desirable attributes, often elicited outside an encounter with a candidate. Revealed attraction is measured through evaluations, decisions, or actions directed toward an identifiable person. The two can correspond imperfectly because people have limited access to the determinants of attraction, because a single characteristic is evaluated differently in combinations, and because situations constrain what can be expressed. Weak correspondence does not establish that preferences are fictional; it establishes that any particular questionnaire must be validated against the behavior it purports to forecast Todd et al. (2007) Eastwick & Finkel (2008).

Eastwick and Finkel tested how people report ideal qualities and respond to actual partners in speed-dating settings. Their findings challenged the assumption that familiar sex differences in self-reported ideals necessarily predict similarly differentiated attraction after face-to-face meetings. In the 2014 meta-analysis by Eastwick and colleagues, observed romantic evaluations were related to several desirable partner attributes but customary sex-difference predictions for revealed evaluations lacked consistent support. The relevant outcome is an association between a measured attribute and romantic judgment under particular conditions, not a demonstration that everyone selects the same partners Eastwick & Finkel (2008) Eastwick et al. (2014).

The central inference concerns levels of measurement. Group-average stated preferences may reproduce reliably while individual differences in those preferences offer only small incremental prediction when used to explain who likes whom. If nearly every respondent values kindness, a questionnaire item may contain little discriminating variance even if kindness is consequential. Likewise, the restricted range of available partners can suppress an observable association. Psychometric assessment must model baseline desirability of attributes, the variance of attributes in a choice pool, each chooser's idiosyncratic weighting, and the extent to which partner attributes were assessed independently of later attraction Eastwick et al. (2025) Zhao et al. (2025).

Speed-Dating Experiments and Encounter-Level Decisions

Speed-dating studies create structured sequences of brief encounters and collect immediate ratings, invitations to meet again, and in some designs mutual matches. These studies improve on purely hypothetical instruments by observing behavior directed toward actual candidates. Researchers can vary encounter order, the number and distribution of available partners, or assignment to sessions. Yet the studies remain samples of people willing and able to participate, with limited acquaintance time and context-specific social pressures. Their main contribution is the measurement of initial attraction and reciprocal contact decisions under controlled exposure, rather than the final outcome of long-term mate choice Fisman et al. (2006) Todd et al. (2007).

Fisman and colleagues' 2006 speed-dating experiment introduced controlled variation in the pool of potential partners and reported gender-associated differences in evaluations of physical attractiveness, intelligence, socioeconomic background, and selectivity. Because participants were randomly introduced to candidates within events, the design supports stronger inference about how encounter-level characteristics relate to ratings than a retrospective survey of married partners. Nevertheless, the local social market, sample composition, self-selection into participation, and the measured time horizon limit generalization. A result from an experimentally organized meeting cannot simply be transferred to family-arranged partnerships, online dating, or established couples Fisman et al. (2006).

Todd and colleagues combined stated ratings with observed choices in speed dating and found that the processes governing the two outcomes could differ. Such discrepancies have more than one possible explanation. Participants may lack introspective access to the relevant computations, may use a simplifying rule under time pressure, or may be unable to evaluate stated traits accurately in a short exchange. An essential distinction is between failure of an underlying preference to influence choice and failure of a design to detect that influence. Larger samples, repeated encounters, independent measurement of candidate traits, and preregistered analytic specifications are needed to discriminate these possibilities Todd et al. (2007) Zhao et al. (2025).

Preference Matching and the Worldwide Registered Report

Preference matching asks whether a person evaluates a partner more favorably when that partner's attributes resemble the person's ideals. The question differs from whether a desirable trait is generally attractive to many people. Normative preference matching captures similarity to commonly valued trait profiles, while distinctive preference matching captures the additional effect of an individual's departures from collective standards. If this distinction is ignored, a high correlation between an ideal and a partner may arise because most respondents endorse broadly desirable traits rather than because a particular individual's weighting is uniquely predictive Eastwick et al. (2014) Eastwick et al. (2025).

Eastwick and an international collaboration published a registered report in 2025 assessing 10,358 participants from 43 countries and 22 languages. Using a corrected pattern metric across 35 traits, the analysis reported an initial standardized association of approximately beta .19 between matching and evaluations, reduced to beta .11 when modeled with normative matching. Average specific-trait interaction effects were much smaller, about beta .04. These values refer to particular statistical estimands and should not be interpreted as the percentage of partnerships caused by preferences. The registered design, international implementation, and explicit comparison of metrics strengthen evidential confidence while preserving the observational nature of much of the relationship material Eastwick et al. (2025).

The study also observed discrepancies between declared and revealed valuations of attractiveness and earning potential and found broadly similar effects among partnered and single respondents. These findings support an intermediate conclusion: partner ideals possess some predictive information when modeled appropriately, but individual-specific prediction is modest and dependent on how the match is defined. A prospective report of ideals before a relationship improves temporal ordering, yet the eventual partnering process continues to involve mutual consent, candidate exposure, feedback, and correlated traits. To identify causal effects of preferences on partner formation, further research must manipulate information, opportunity, or choice environments while observing consequences Eastwick et al. (2025) Conroy-Beam & Buss (2016).

Simulation, Statistical Power, and the Stated–Revealed Gap

The apparent inconsistency between questionnaires and observed romantic choices may partly reflect an inferential problem. If many candidate traits independently contribute to overall attraction, any one trait's correlation with a short-term evaluation will be diluted by variation in the other traits. When candidate attributes are measured with noise or few people are encountered, an actual attribute-specific preference can produce a small and unstable observed association. Simulations can quantify this attenuation under explicit assumptions, though their usefulness depends on whether the assumptions about dimensionality, measurement error, and candidate distributions resemble real encounters Zhao et al. (2025).

Zhao and colleagues reported a speed-dating study with 1,145 participants in which four of nine evaluated traits showed evidence of correspondence between stated and revealed preferences. A companion simulation demonstrated how a design involving multiple influential independent traits can yield weak isolated trait associations even when preferences play a causal role. This result helps reconcile apparent contradictions between classic survey research and encounter studies. It also calls for restraint: simulated effects conditional on a model are not direct evidence that the model is the unique mechanism operating in real-world partnerships Zhao et al. (2025) Eastwick et al. (2014).

Future analyses should model choice as a multivariate function of actual partner traits, subjective perceptions, opportunity sets, and measurement uncertainty. Merely increasing sample size while retaining invalid indicators will not repair poor construct validity. Transparent reporting should identify whether analyses are at the dyad, individual, encounter, session, or country level and account for the repeated observations generated when each participant meets several candidates. Cross-validation can distinguish an explanatory model that fits existing ratings from a prediction rule that generalizes to new persons and dating contexts Zhao et al. (2025) Eastwick et al. (2025).

Reciprocal Acceptance and the Formation of Actual Couples

Romantic attraction is directed, but a relationship requires at least two people who can express compatible intentions. A person may prefer a particular candidate who does not reciprocate, while a less idealized match develops through repeated positive interaction. This creates a selection mechanism at the dyadic level: the people observed in established partnerships have passed through two sets of preferences and opportunity constraints. Analyses of already formed couples condition on this mutual-selection process and risk selection bias when attributing partner similarity exclusively to an individual preference. A rigorous model needs separate probabilities for exposure, approach, response, reciprocal acceptance, and persistence Schwartz (2013) Conroy-Beam & Buss (2016).

Conroy-Beam and Buss combined simulations and studies of existing couples to ask whether preference fulfillment relates to partner selection. Their work demonstrates how computational representations can clarify theoretical predictions about matching and choice, provided the algorithm's assumptions are disclosed. The existence of compatibility patterns among couples does not establish that a single optimization procedure guided the original formation. The sample excludes pairs that never formed and may underrepresent dissolved relationships; current evaluations can also be affected by adaptation to the existing partner. Retrospective assessments consequently require triangulation with prospectively measured ideals and longitudinal records Conroy-Beam & Buss (2016) Eastwick et al. (2025).

Dyadic selection can be investigated by recording both persons' initial values, opportunities, and subsequent choices before the outcome is known. The theoretical analysis should allow unilateral interest, mutual interest, continued contact, and relationship commitment to appear as distinct endpoints. Comparing accepted and rejected candidates within the same individuals can control some stable preferences but cannot automatically account for unobserved candidate qualities or conversation dynamics. Better estimates require crossed random effects for chooser and candidate, attention to the scale of each decision, and transparency about attrition after the first meeting Fisman et al. (2006) Todd et al. (2007).

Conditional Strategies and Relationship Time Horizons

The concept of a conditional mating strategy addresses how preferences and behavior can vary with personal goals and circumstances. Researchers often distinguish interest in brief sexual encounters from the search for long-term partnership, but these categories measure intentions and expectations rather than discrete biological types. A single person may value physical appeal for initial attraction, kindness for extended involvement, and shared priorities for durable cooperation. Short-term intentions and long-term attachment can also coexist, and decisions evolve as interactions accumulate. Models should therefore register the participant's actual relationship goals and the time horizon of each judgment, rather than infer them from the survey context alone Buss & Schmitt (1993) Gangestad & Simpson (2000).

Strategic-pluralism models predict that choice can depend on ecological uncertainty, prospective parental investment, attractiveness, access to alternatives, and anticipated costs. Empirical designs can test these predictions with experimental variations in relationship descriptions or by following people whose social and economic conditions change. Theoretical flexibility is valuable only if the analyst specifies in advance which cues should change which preference, in what direction, and under what circumstances. Otherwise any outcome can be redescribed after the fact as an adaptive adjustment, leaving the theory weakly testable Gangestad & Simpson (2000) Conroy-Beam & Buss (2016).

Trade-offs also operate within a single desired relationship. A potential partner can be financially secure yet emotionally unavailable, attractive yet geographically distant, or compatible in values yet uninterested in parenthood. Hypothetical trait budgets make such constraints salient, but actual decisions involve unknown attributes, imperfect information, unequal probabilities of reciprocal interest, and a changing pool of options. The effects of time horizon should be evaluated in combinations rather than as an exhaustive two-category typology. Multiattribute choice models, observed meetings, and longitudinal relationship data can test whether people revise priorities with experience Li et al. (2002) Zhao et al. (2025).

Sexual Conflict, Competition, and Cooperation

Sexual conflict arises when the reproductive interests or preferred outcomes of partners differ in particular conditions. Evolutionary theory can illuminate tensions surrounding investment, exclusivity, mating opportunities, or distribution of effort, but conflict must be inferred from actual divergence in interests and relevant outcomes. Competition for partners may involve efforts to increase visibility, social status, persuasive communication, or appeal relative to rivals. The same social environment can also reward cooperation, honest signaling, mutual support, and long-term coordination. Human courtship contains intersexual negotiation and intrasexual competition, with mechanisms shaped by personal values and institutions Trivers (1972) Buss (2023).

An empirical study of competition must distinguish a person's intention to attract a partner, the reactions of prospective partners, and eventual reproductive or relationship consequences. A highly visible courtship tactic may increase initial contact while undermining later trust, illustrating that selection at one stage does not determine success at another. Economic and social inequalities can make rivalry more consequential by altering access to safe meeting environments and the perceived costs of rejection. Measures of jealousy, rivalry, and mate guarding should specify their context and their behavioral consequences; such experiences vary widely and do not by themselves establish an adaptive evolutionary explanation Buss & Schmitt (1993) Eagly & Wood (1999).

Consent and autonomy are fundamental components of the observed human process. A model of selection that describes only which competitor prevails omits the decision-making rights and responses of each person. Norms and legal protections determine which forms of pursuit are permissible, and experiences of coercion cannot be treated as ordinary preference fulfillment. Scientific studies should separate competition from coercion and direct attention to partner safety, mutuality, and the long-term outcomes of chosen relationships. This makes the psychological mechanisms of negotiation and reciprocal acceptance central explanatory targets Schwartz (2013) Buss (2023).

Hormonal-State Hypotheses and Reproducibility

One influential research program proposed that preferences for certain masculine or other appearance cues vary predictably across the ovulatory cycle. These hypotheses attempted to connect within-person hormonal changes to context-dependent mate choice and predicted that a repeated-measures design could detect reliable changes in judgments across cycles. The theoretical importance is clear: evidence for such shifts would provide a proximate mechanism through which reproductive physiology affects evaluations. Yet measuring the relevant hormonal state, confirming ovulation, selecting stimuli, specifying relationship contexts, and separating within-person change from sampling fluctuation are all essential for testing the prediction Gildersleeve et al. (2014).

Gildersleeve, Haselton, and Fales conducted a 2014 meta-analytic review of cycle-related preference research and reported a pattern they interpreted as supportive of shifts in some mate preferences. Meta-analysis aggregates existing observations but inherits potential weaknesses in the underlying measurement and selective publication. Studies relying on calendar estimates or broad fertile-window classifications may misclassify physiological state; flexible analytic decisions and small samples can increase apparent effects. A credible causal explanation needs repeated hormonal measurement and preregistered tests of the predicted within-person association, rather than post hoc assignment of cycle phase Gildersleeve et al. (2014) Jones et al. (2018).

Jones and colleagues' 2018 longitudinal study directly measured salivary hormones in 584 women and found no compelling evidence that changes in facial-masculinity preferences tracked hormone levels. The contrast illustrates why later, more precisely measured evidence can change the interpretation of an earlier literature even when earlier meta-analytic estimates appeared positive. The study's result addresses the specified facial-cue hypothesis and does not imply that hormones never influence sexual desire, behavior, or other outcomes. Separate claims about desire, attraction, partner preferences, and reproductive decision making require separate operational measures and independent replication Jones et al. (2018) Gildersleeve et al. (2014).

Sexual Orientation, Relationship Diversity, and Selection Models

Human mate-choice research historically concentrated on heterosexual couples seeking marriage or reproduction. That sampling choice limits the constructs studied and the populations to which the conclusions can be generalized. Romantic attraction can occur among same-sex and different-sex partners, as well as in relationships whose goals include companionship, intimacy, or cohabitation without childbearing. The scientific analysis should identify whether a particular model seeks to explain general partner evaluation, sexual attraction, reproductive outcomes, or a specific culturally organized relationship form. These domains overlap but have distinct explanatory targets Takayanagi et al. (2024) Eastwick et al. (2014).

Takayanagi and colleagues' Brazilian study examined preferences among heterosexual, bisexual, and homosexual participants using budgeted choice tasks and found both common desirability patterns and variation associated with sex, orientation, and relationship goals. Its contribution is the inclusion of groups underrepresented in classic mate-preference studies and direct measurement of several preference dimensions. It remains a study of expressed valuations within one national sample. The relevant scientific inference is that a model developed for a heterosexual reproductive setting requires new tests before it is applied to all romantic selection Takayanagi et al. (2024).

Research can improve by separating gender identity, sex-related biological variables when scientifically necessary, gender of desired partners, sexual and romantic attraction, and relationship intentions. Those variables may have distinct associations with age preferences, perceived compatibility, expected social acceptance, and anticipated family support. Asexually identifying respondents, for example, may distinguish romantic partnership goals from sexual attraction in ways that standard mate-choice questionnaires obscure. Inclusive measurement expands construct validity by matching the instrument to the psychological phenomenon. It also identifies the conditions under which a model of sexual selection is an explanatory framework for a specific outcome rather than an assumption about every relationship Takayanagi et al. (2024) Eagly & Wood (1999).

Assortative Mating, Educational Similarity, and Population Structure

Assortative mating is usually measured by the association between partners' characteristics after relationships form. Education, age, religion, language, and social background may be more similar within couples than between two randomly paired people in the same population. The existence of resemblance does not establish preference for resemblance: shared schools, residential separation, family networks, labor markets, and available encounter times all organize whom people are likely to meet. The reference population and the construction of a plausible counterfactual matching distribution are therefore central to interpreting assortative mating. When these structural components are ignored, outcomes of social segregation can be misread as freely expressed psychological preferences Schwartz (2013).

Schwartz's 2013 review surveyed demographic accounts of assortative mating, emphasizing historical variation, inequality, and the structure of partner search. A more recent analysis by Hirschl, Schwartz, and Boschetti used U.S. census and American Community Survey data spanning 1940 through 2020. They reported that increases in educational homogamy stalled around 1990 and reversed in the 2000s, in part as women increasingly partnered with men who had less formal education. This finding is important because it shows how historical social change can reverse a pattern sometimes treated as a stable behavioral regularity. It also illustrates how shifts in the relative educational composition of men and women affect matching outcomes even when preferences are not directly measured Hirschl et al. (2024) Schwartz (2013).

Assortment can influence household resources, intergenerational opportunity, and social networks, but causal decomposition is demanding. Household similarity may reflect shared institutions before partnership, changes within the relationship, or the selection and retention of couples over time. Comparing marriages in different decades also requires accounting for education expansion and changes in the marriage rate. A strong empirical strategy uses cohorts, exposure data, demographic matching simulations, and measured partner preferences to assess these mechanisms separately. Its unit of analysis may be the couple, a local marriage market, or a cohort, and the conclusions should not confuse effects at these levels Hirschl et al. (2024) Schwartz (2013).

Online Dating, Algorithms, and Search Markets

Digital dating systems transform the encounter stage by reorganizing the visibility, ordering, and perceived availability of possible partners. Profile photographs, text, search filters, location, mutual-match interfaces, and recommendation systems govern what attributes can be seen before direct interaction. People may express preferences through likes, messages, responses, and the time invested in exchanges, but these measures represent distinct stages of selection. An algorithm can increase a person's probability of being displayed without changing another person's fundamental valuation of their attributes. The behavioral effects of a platform thus arise jointly from user preferences and platform-mediated exposure, and neither can be inferred accurately from message counts alone Bruch & Newman (2018) Schwartz (2013).

Bruch and Newman analyzed online-dating contact patterns and described aspirational pursuit in the sampled markets, including uneven patterns of sending messages across inferred desirability rankings. These are observed behavioral data, an advantage over surveys, yet the platform's participants are a selected population, and contact attempts are not marriages, mutual attraction, or durable partnerships. A message's absence may indicate limited exposure, search costs, platform sorting, or competing priorities. Investigating matching success requires linking the initial approach to reciprocal responses, later contact, and relationship outcomes when participants consent to such data use Bruch & Newman (2018).

Algorithmic environments also alter what researchers mean by the available choice set. Interface defaults can emphasize particular ages, distances, or profile signals while suppressing others. Recommender models often learn from recorded interactions that are themselves shaped by the previous ranking policy; observed preference estimates can therefore contain feedback loops. Research should report whether candidate exposure was randomized, how often recommendations were shown, whether filters were user-selected or imposed by the interface, and how deleted, inactive, or unresponsive profiles were handled. Experiments comparing exposure regimes can estimate platform effects without conflating them with latent preference change Bruch & Newman (2018) Eastwick et al. (2025).

Cultural Ecology, Opportunity, and the Social Organization of Choice

Mate choice operates inside locally structured opportunities. Schools and universities, occupations, mobility, kin networks, housing constraints, religion, gender norms, and migration patterns determine who is likely to meet and which relationships are feasible or socially recognized. These structures shape both expressed desires and observed pair formation. A preference for a shared educational background can reflect valued common experiences or the practical fact that educational institutions concentrate particular populations in the same spaces. Ecological analysis must distinguish choice among encountered candidates from the prior social process that determines which candidates become encounterable Schwartz (2013) Hirschl et al. (2024).

Changes in social and economic security can affect how people evaluate potential partners. Employment prospects, gender inequality, family policy, and the cost of housing may modify the perceived relevance of resources and partnership stability. Cross-country differences in preference ratings are informative, but they generally bundle multiple correlated dimensions and seldom isolate one institutional variable experimentally. The social-role tradition poses precise causal questions about opportunity and norms; evolutionary perspectives ask whether recurring patterns survive changes in those environments. Both require measurements at the individual level and independent verification of country-level indicators Eagly & Wood (1999) Zentner & Mitura (2012).

Demographic processes can also change the meaning of a preference distribution. In a region with a large imbalance in numbers of eligible partners, a highly selective individual may confront different realistic opportunities than an otherwise similar person elsewhere. Age structure, delayed marriage, educational expansion, and geographic separation shape the price of search and the composition of mutually available candidates. Researchers should model the relevant local market, define which people could actually pair, and attend to matching rules and reciprocal agency. Comparisons of partners across historical cohorts can then identify whether observed changes are best explained by changing preferences, exposure, composition, or social constraints Fisman et al. (2006) Hirschl et al. (2024).

Measurement, Experimental Design, and Evidence Appraisal

A reliable science of human mate choice requires a precise mapping from construct to method. Direct ratings elicit desirability in the abstract; rank-order surveys measure relative priority; budget tasks elicit trade-offs; speed dating measures brief interpersonal attraction; online logs record platform-specific contact; partner reports document established relationships; demographic data reveal aggregate pairing patterns. Each measure has a different observation window and causal status. A reported association between a trait and a rating cannot serve as the same kind of evidence as a randomized change in available information or a prospective record of relationship formation Li et al. (2002) Fisman et al. (2006) Schwartz (2013).

Psychometric evaluation should assess internal consistency where relevant, test–retest stability, cross-language comparability, measurement invariance, and the distinction between evaluative desirability and behaviorally applied choice weights. For multi-country surveys, the assumption that the same word denotes the same psychological quantity must be examined rather than assumed. For speed-dating designs, clustered data and reciprocal dyads require crossed random or fixed effects, and selective participation complicates external validity. For repeated hormonal designs, precision of biological measurement and prospectively defined time windows are decisive. This methodological diversity makes a single universal effect estimate inappropriate for the full research object Kowal et al. (2025) Zhao et al. (2025) Jones et al. (2018).

Source appraisal also depends on the strength of the inferential design. Large cross-national samples provide precision in estimating descriptive patterns among participants; they do not make ecological correlations causal. A meta-analysis synthesizes compatible effect definitions but is vulnerable to measurement heterogeneity and publication bias. A randomized encounter or platform-exposure manipulation identifies an intervention effect under its particular implementation, while longitudinal observation improves temporal ordering but does not eliminate time-varying confounding. Registered reports and transparent multiverse analyses strengthen confidence by limiting selective analysis choices. Evidence should therefore be evaluated for the claim at issue, rather than ranked solely by sample size or publication prestige Eastwick et al. (2025) Eastwick et al. (2014) Walter et al. (2020).

Critical Synthesis of Competing Explanations

The evidence supports several conclusions at different levels of generality. Repeated cross-cultural surveys establish that some average partner preferences differ by sex category in samples from many countries. Encounter-level studies show that these stated differences do not map simply or consistently onto immediate attraction. A registered international study finds positive, generally modest preference-matching effects, with much smaller effects attributable to particular individualized trait priorities. Demographic records demonstrate that the similarity of actual partners changes over historical time. These findings concern compatible but distinct outcomes, and a complete theory must explain their coexistence rather than use a result from one measure to invalidate another Buss (1989) Walter et al. (2020) Eastwick et al. (2025) Hirschl et al. (2024).

Explanations that emphasize evolved reproductive trade-offs provide a historical basis for some questions about selectivity, age preferences, and parental investment. Social-role accounts specify how resource distribution and institutions reshape incentives. Cognitive and decision-theoretic accounts explain why hypothetical ideals can differ from real-time decisions, while demographic approaches explain how nonrandom encounters generate sorting. The sharpest unresolved issues concern the identification of causal pathways and the division of variance between processes. Simply labeling a pattern evolutionary, social, or cognitive leaves its causal contribution unspecified; the next step is to test models in which the proposed pathways generate differentiated predictions Trivers (1972) Eagly & Wood (1999) Zhao et al. (2025) Schwartz (2013).

Reproducibility becomes especially consequential when a theory is supported by a chain of indirect inferences. The history of Bateman's experiment and the disputed menstrual-cycle literature demonstrate how influential claims can change when sampling, measurement, and analysis are revisited. Conversely, replication of a descriptive sex difference across independent populations strengthens confidence in that descriptive observation while leaving its ancestral origin open. A defensible synthesis reports which effect replicates, which measurement class is involved, and which explanation remains conjectural. The standard is explanatory precision within established uncertainty, not the ambition to fit every possible human relationship into one theory Snyder & Gowaty (2007) Jones et al. (2018) Walter et al. (2020).

Mate Choice in the Artificial Era

The Artificial Era introduces new settings in which artificial systems mediate the information, opportunities, and decisions involved in human relationships. Search engines, matching algorithms, recommender models, automated profile assistance, and conversational agents can change which partners people encounter and the cues available for evaluation. Their empirically tractable role is to alter exposure, interpretation, communication, and the perceived cost of search. This raises research questions about whether mediated first impressions predict later face-to-face compatibility and whether ranking systems amplify historically accumulated inequalities in participation. Existing online-dating analyses provide starting points for these questions but cannot establish the consequences of future artificial systems without direct study Bruch & Newman (2018) Eastwick et al. (2025).

Artificial systems can also become participants in a person's relational environment as advisors, collaborators, or conversational companions. That development requires careful specification of the phenomenon under investigation. A user may consult an artificial assistant about prospective human partners, outsource profile writing, or use automated tools to filter and compare options. Each case involves measurable changes in human attention and decisions. The empirical question is how such interactions influence agency, confidence, trust, privacy, and relationship outcomes. Philosophical classifications of artificial persons belong to a distinct level of inquiry and do not on their own supply evidence about human mate-choice effects Todd et al. (2007) Bruch & Newman (2018).

A cumulative research program should measure the counterfactual encounter opportunities created by algorithmic systems and evaluate the effects of automated recommendations on both sides of a prospective match. Randomized platform experiments, informed consent, fairness assessment across demographic groups, and long-term follow-up can distinguish effects on clicks from effects on stable, mutually desired relationships. Models should account for feedback loops as users learn the platform and the platform learns from their actions. The central scientific continuity is the mutuality of selection: even when artificial mediation alters candidate exposure, human partner formation remains a process involving distinct agents and situated commitments Fisman et al. (2006) Schwartz (2013).

Conclusion

Human mate choice is a reciprocally organized process through which preferences, access to potential partners, interpersonal evaluations, and relationship decisions produce selective pair formation. Sexual selection provides the evolutionary framework for asking how mate competition and differential reproductive success could have shaped relevant capacities and preferences. Modern psychological evidence shows that these capacities are expressed through changing social institutions, cultural norms, perceptual judgments, learning histories, and personal goals. The scientific unit of analysis must match the question: a stated ideal, an attraction rating, a mutual dating decision, a formed couple, and a reproductive outcome are related phenomena with different measurement requirements Darwin (1871) Buss (2023) Eastwick et al. (2025).

Historical theory and contemporary evidence are strongest when read together critically. Classic cross-cultural research and large-scale replication establish recurrent average preference patterns; controlled speed-dating work shows how they can change when real people interact; newer registered international evidence demonstrates measurable but often modest individual-specific matching effects. Population and demographic analyses reveal how institutional environments and available choice sets influence the partners people actually form relationships with. Challenges to early sexual-selection experiments and to some hormonal-cycle hypotheses reinforce the value of independent replication, direct measurement, and clearly specified estimands Buss (1989) Todd et al. (2007) Eastwick et al. (2025) Hirschl et al. (2024) Snyder & Gowaty (2007) Jones et al. (2018).

The next stage of research requires integrated, prospectively measured dyadic and population studies. They should collect both partners' preferences before encounters, document realistic candidate exposure, track reciprocal decisions and relationship development, include diverse cultural and sexual-orientation groups, and distinguish technological mediation from human evaluation. Such studies can adjudicate competing evolutionary, social-role, cognitive, and demographic pathways while identifying stable findings and meaningful context effects. The resulting scientific picture is of human partner formation as a conditional, multilevel process with measurable regularities and historically variable mechanisms, increasingly reorganized by the institutions and technologies of the Artificial Era Kowal et al. (2025) Zhao et al. (2025) Schwartz (2013).

References

  • Bateman, A. J. (1948). Intra-sexual selection in Drosophila. Heredity, 2, 349–368. https://doi.org/10.1038/hdy.1948.21
  • Bruch, E. E., & Newman, M. E. J. (2018). Aspirational pursuit of mates in online dating markets. Science Advances, 4(8), eaap9815. https://doi.org/10.1126/sciadv.aap9815
  • Buss, D. M. (1989). Sex differences in human mate preferences: Evolutionary hypotheses tested in 37 cultures. Behavioral and Brain Sciences, 12(1), 1–14. https://doi.org/10.1017/S0140525X00023992
  • Buss, D. M. (2023). The sexual selection of human mating strategies: Mate preferences and competition tactics. In The Oxford handbook of evolutionary psychology and romantic relationships (pp. 15–41). Oxford University Press. https://doi.org/10.1093/oxfordhb/9780197524718.013.1
  • Buss, D. M., & Schmitt, D. P. (1993). Sexual strategies theory: An evolutionary perspective on human mating. Psychological Review, 100(2), 204–232. https://doi.org/10.1037/0033-295X.100.2.204
  • Conroy-Beam, D., & Buss, D. M. (2016). Do mate preferences influence actual mating decisions? Evidence from computer simulations and three studies of mated couples. Journal of Personality and Social Psychology, 111(1), 53–66. https://doi.org/10.1037/pspi0000054
  • Darwin, C. (1871). The descent of man, and selection in relation to sex. John Murray. https://www.gutenberg.org/ebooks/2300
  • Eagly, A. H., & Wood, W. (1999). The origins of sex differences in human behavior: Evolved dispositions versus social roles. American Psychologist, 54(6), 408–423. https://doi.org/10.1037/0003-066X.54.6.408
  • Eastwick, P. W., & Finkel, E. J. (2008). Sex differences in mate preferences revisited: Do people know what they initially desire in a romantic partner? Journal of Personality and Social Psychology, 94(2), 245–264. https://doi.org/10.1037/0022-3514.94.2.245
  • Eastwick, P. W., Luchies, L. B., Finkel, E. J., & Hunt, L. L. (2014). The predictive validity of ideal partner preferences: A review and meta-analysis. Psychological Bulletin, 140(3), 623–665. https://doi.org/10.1037/a0032432
  • Eastwick, P. W., Sparks, J., Finkel, E. J., Meza, E. M., Adamkovič, M., Adu, P., Ai, T., Akintola, A. A., Al-Shawaf, L., Apriliawati, D., Arriaga, P., Aubert-Teillaud, B., Baník, G., Barzykowski, K., Batres, C., Baucom, K. J., Beaulieu, E. Z., Behnke, M., Butcher, N., ... Coles, N. A. (2025). A worldwide test of the predictive validity of ideal partner preference matching. Journal of Personality and Social Psychology, 128(1), 123–146. https://doi.org/10.1037/pspp0000524
  • Fisman, R., Iyengar, S. S., Kamenica, E., & Simonson, I. (2006). Gender differences in mate selection: Evidence from a speed dating experiment. Quarterly Journal of Economics, 121(2), 673–697. https://doi.org/10.1162/qjec.2006.121.2.673
  • Gangestad, S. W., & Simpson, J. A. (2000). The evolution of human mating: Trade-offs and strategic pluralism. Behavioral and Brain Sciences, 23(4), 573–587. https://doi.org/10.1017/S0140525X0000337X
  • Gildersleeve, K., Haselton, M. G., & Fales, M. R. (2014). Do women's mate preferences change across the ovulatory cycle? A meta-analytic review. Psychological Bulletin, 140(5), 1205–1259. https://doi.org/10.1037/a0035438
  • Hirschl, N., Schwartz, C. R., & Boschetti, E. (2024). Eight decades of educational assortative mating: A research note. Demography, 61(5), 1293–1307. https://doi.org/10.1215/00703370-11558914
  • Jones, B. C., Hahn, A. C., Fisher, C. I., Wang, H., Kandrik, M., Han, C., Fasolt, V., Morrison, D., Lee, A. J., Holzleitner, I. J., O'Shea, K. J., Roberts, S. C., Little, A. C., & DeBruine, L. M. (2018). No compelling evidence that preferences for facial masculinity track changes in women's hormonal status. Psychological Science, 29(6), 996–1005. https://doi.org/10.1177/0956797618760197
  • Kowal, M., Sorokowski, P., Gjoneska, B., Pisanski, K., Pfuhl, G., Aguilar, L., Janssen, S. M. J., Gelbart, B., Arriaga, P., Antfolk, J., Zvončáková, K., Lidborg, L. H., Contreras-Garduño, J., Kozlov, M. V., Milfont, T. L., Varella, M. A. C., Križanić, V., Boussena, M., Kavčič, T., ... Prazeres, F. (2025). Cross-cultural data on romantic love and mate preferences from 117,293 participants across 175 countries. Scientific Data, 12, Article 1103. https://doi.org/10.1038/s41597-025-05365-2
  • Langlois, J. H., Kalakanis, L., Rubenstein, A. J., Larson, A., Hallam, M., & Smoot, M. (2000). Maxims or myths of beauty? A meta-analytic and theoretical review. Psychological Bulletin, 126(3), 390–423. https://doi.org/10.1037/0033-2909.126.3.390
  • Li, N. P., Bailey, J. M., Kenrick, D. T., & Linsenmeier, J. A. W. (2002). The necessities and luxuries of mate preferences: Testing the tradeoffs. Journal of Personality and Social Psychology, 82(6), 947–955. https://pubmed.ncbi.nlm.nih.gov/12051582/
  • Rhodes, G. (2006). The evolutionary psychology of facial beauty. Annual Review of Psychology, 57, 199–226. https://doi.org/10.1146/annurev.psych.57.102904.190208
  • Schwartz, C. R. (2013). Trends and variation in assortative mating: Causes and consequences. Annual Review of Sociology, 39, 451–470. https://doi.org/10.1146/annurev-soc-071312-145544
  • Snyder, B. F., & Gowaty, P. A. (2007). A reappraisal of Bateman's classic study of intrasexual selection. Evolution, 61(11), 2457–2468. https://doi.org/10.1111/j.1558-5646.2007.00212.x
  • Takayanagi, J. F. G. B., Siqueira, J. de O., Silveira, P. S. P., & Valentova, J. V. (2024). What do different people look for in a partner? Effects of sex, sexual orientation, and mating strategies on partner preferences. Archives of Sexual Behavior, 53(3), 981–1000. https://doi.org/10.1007/s10508-023-02767-4
  • Todd, P. M., Penke, L., Fasolo, B., & Lenton, A. P. (2007). Different cognitive processes underlie human mate choices and mate preferences. Proceedings of the National Academy of Sciences, 104(38), 15011–15016. https://doi.org/10.1073/pnas.0705290104
  • Trivers, R. L. (1972). Parental investment and sexual selection. In B. Campbell (Ed.), Sexual selection and the descent of man, 1871–1971 (pp. 136–179). Aldine. https://www2.nau.edu/~shuster/shustercourses/BIO%20698/Literature/Trivers%201972.pdf
  • Walter, K. V., Conroy-Beam, D., Buss, D. M., Asao, K., Sorokowska, A., Sorokowski, P., Aavik, T., Akello, G., Alhabahba, M. M., Alm, C., Amjad, N., Anjum, A., Atama, C. S., Atamtürk Duyar, D., Ayebare, R., Batres, C., Bendixen, M., Bensafia, A., Bizumic, B., ... Zupančič, M. (2020). Sex differences in mate preferences across 45 countries: A large-scale replication. Psychological Science, 31(4), 408–423. https://doi.org/10.1177/0956797620904154
  • Zentner, M., & Mitura, K. (2012). Stepping out of the caveman's shadow: Nations' gender gap predicts degree of sex differentiation in mate preferences. Psychological Science, 23(10), 1176–1185. https://doi.org/10.1177/0956797612441004
  • Zhao, A. A. Z., Sidari, M. J., Lee, A. J., & Zietsch, B. P. (2025). Speed-dating and simulation data explain the discrepancy between stated and revealed mate preferences. European Journal of Personality, 39(4), 535–557. https://doi.org/10.1177/08902070241286254