Skip to main content
PMC home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2012 Sep 17.
Published in final edited form as: Am Nat. 2011 May;177(5):605–616. doi: 10.1086/659629

Variation in human mate choice: Simultaneously investigating heritability, parental influence, sexual imprinting, and assortative mating

Brendan P Zietsch 1,2,*, Karin J H Verweij 2,1,*, Andrew C Heath 3, Nicholas G Martin 2
PMCID: PMC3444291  NIHMSID: NIHMS396401  PMID: 21508607

Abstract

Human mate choice is central to individuals’ lives and to the evolution of the species, but the basis of variation in mate choice is not well understood. Here we look at a large community-based sample of twins and their partners and parents (N > 20,000 individuals) to test for genetic and family environmental influences on mate choice, with and without controlling for the effects of assortative mating. Key traits are analyzed, including height, body mass index, age, education, income, personality, social attitudes, and religiosity. This revealed near-zero genetic influences on male and female mate choice over all traits and no significant genetic influences on mate choice for any specific trait. A significant family environmental influence was found for the age and income of females’ mate choices, possibly reflecting parental influence over mating decisions. We also tested for evidence of sexual imprinting, where individuals acquire mate-choice criteria during development by using their opposite-sex parent as the template of a desirable mate; there was no such effect for any trait. The main discernable pattern to mate choice was assortative mating; we found that partner similarity was due to initial choice rather than convergence and also due at least in part to phenotypic matching.

Keywords: mate choice, mate preferences, behaviour genetics, evolutionary psychology, sexual imprinting, assortative mating

Introduction

The choice of a relationship partner (mate choice) often determines who will contribute half the genome and care of our offspring, the vessels that may or may not carry our genes into future generations. The sexual choices of our ancestors over hundreds of thousands of years have helped shape our evolution (Miller 2000), yet the factors governing human mate choice are poorly understood.

One pervasive finding is that romantic partners are similar to each other in many respects. Romantic partners correlate positively and strongly on age, social attitudes, and religiosity; correlate moderately on general intelligence, education, and physical attractiveness; and correlate weakly on height, weight, and personality traits (Price and Vandenberg 1980; Martin et al. 1986; Feingold 1988; Watson et al. 2004; Koenig, McGue, and Iacono 2009; Hatemi et al. 2010). Several lines of evidence indicate that similarity between partners is due to initial choice (assortative mating), rather than convergence during the relationship (Mascie-Taylor 1989; Caspi, Herbener, and Ozer 1992; Watson et al. 2004). Family studies suggest that assortative mating is at least partly due to similarity in social backgrounds (i.e. social homogamy), and is probably also due to individuals preferring or being more likely to encounter one another because they have similar phenotypes (i.e. phenotypic matching; Nagoshi, Johnson, and Ahern 1987; Reynolds, Baker, and Pedersen 1996, 2000).

Beyond matching on similarity, little is known about why we choose who we choose. It seems intuitively obvious that individuals choose a partners on the basis of more than just their level of self-resemblance, and many studies have established the general tendency of men and women to prefer certain physical and behavioral characteristics over others (Buss and Barnes 1986; Buss 1989; Gangestad and Simpson 2000). There has been much research into mate preferences (e.g. self-reported preferences, factors influencing attractiveness), but it is less understood how these preferences relate to the formation of long-term pair bonds. Some studies show that women find appealing the smell of men who have dissimilar immune-related (MHC) genes (Havlicek and Roberts 2009), but actual couples have no more similar or dissimilar MHC genes than expected by chance (Abdellaoui et al., unpublished manuscript). Similarly, speed dating studies show surprisingly low concordance between self-reported mate preferences and actual choices (Todd et al. 2007; Eastwick and Finkel 2008). Other research shows that self-reported preferences do predict whom an individual finds attractive (Wood and Brumbaugh 2009), but the constraints of the mating market must considerably decrease the concordance between mate preferences and actual choices, especially for exclusive long-term relationships. For example, most women might prefer a husband with an above average income (Buss and Barnes 1986), but not all women can actually have a husband with above-average income. Although population-wide mate preferences are very important in that they reveal what is evolutionarily valued in a partner (e.g. indicators of genetic quality, parental investment), it is equally important to understand the basis of variation in actual mate choices because most human offspring are born to pair-bonded couples (Anderson 2006). These pair-bond mate choices necessarily vary greatly between individuals (since individuals themselves vary greatly), but the sources of this variation are not well understood.

Broadly, the factors underlying variation in mate choice could be genetic or nongenetic. One possible nongenetic factor is the influence of parents on their children’s mate decisions. Anthropological evidence shows that in many cultures throughout history mate choice has been heavily influenced by an individual’s parents (see Buunk, Park, and Dubbs (2008) for a review). It has been proposed that this is because parents have an evolutionary interest in ensuring their daughters mate with higher-investing (e.g. older and higher income) men (Buunk, Park, and Dubbs 2008), and this may take subtle forms even in Western populations today. To the extent that parental influence over a child’s mate decisions varies between families, it should be revealed in a twin study as a family environmental effect on females for mate choice regarding investment-related traits. On the other hand, Bereczkei et al. (2004) argue for sexual imprinting, a nongenetic mechanism whereby individuals acquire mate-choice criteria during development by using their opposite-sex parent as the template of a desirable mate. This has been observed in animals (Oetting, Prove, and Bischof 1995;Kendrick et al. 1998); though not for continuous traits (Schielzeth et al. 2008), but its role in human mating is controversial.

The contribution of genetic influences on variation in mate choice is also unclear. The handful of animal studies quantifying genetic influences on female mate choice (Schielzeth et al. 2008) have mostly found low or zero heritability (the proportion of variation due to genetic variation). Only two human studies have investigated the heritability of mate choice, and those have conflicting conclusions. The first measured various personality-style traits of 269 pairs of identical (monozygotic; MZ) and nonidentical (dizygotic; DZ) twins and their spouses and found mostly weak correlations between the twins’ spouses, regardless of whether the twins were MZ or DZ (Lykken and Tellegen 1993). The authors concluded that mate choice in humans is determined by romantic infatuation, and that this phenomenon is inherently random. The second study (Rushton and Ann Bons 2005) also investigated twins (N = 123 pairs) and their spouses, on traits including height, weight, income, education, social attitudes, and Eysenck’s personality scales. In this study, there were higher correlations between the spouses of MZ twins than between the spouses of DZ twins, and the authors reported that “both correlational and model fitting analyses showed that at least 10 to 30% of the variance in partner choice is due to genetic factors, at least 10% is due to shared environmental factors, and the remaining 60% is due to unique or chance environmental events” (Rushton and Bons 2005, p. 558). Thus, the results and interpretation of the two existing quantitative genetic studies of human mate choice are strongly at odds.

Although not acknowledged in these two studies, there must necessarily be some familiarity (i.e. genetic and/or shared environmental components) to mate choice whenever there is assortative mating on familial traits (e.g. height, intelligence; Heath and Eaves 1985). That is, because twins are of similar intelligence (for example) and there is assortative mating on intelligence, the intelligence of twins’ spouses will automatically be correlated. Of interest is whether there are any familial influences on mate choice beyond that created by assortative mating, but no studies have assessed this (in any species, to our knowledge).

In this article, we use genetic modeling to quantify genetic and environmental influences on mate choice, using data on a large community-based sample of identical and nonidentical twins (N = 11,357 individuals), their partners (N = 7,266 individuals), and their parents (N = 4,238 individuals). Key traits are analyzed, including height, body mass index (BMI), age, education, income, Eysenck’s and Cloninger’s personality traits, social attitudes, and religiosity (some of these measures were available for only a subset of the full sample). The larger sample size compared with previous studies affords greater power and more detailed analyses, including separate estimation of effects in males and females (cross-culturally, men and women differ in the traits they value in a potential partner; (Buss 1989)). In contrast to previous studies, we also control for the twins’ own traits to eliminate the effects of assortative mating on the genetic and family environmental components of mate choice. With availability of twin and parental data for the same traits, we can also test the sexual imprinting hypothesis of human mate choice (Bereczkei, Gyuris, and Weisfeld 2004). On average, each nonidentical twin is equally genetically similar to each of their parents and his or her co-twin; if sexual imprinting plays a role in mate choice, then a twin’s partner should be more similar to the twin’s opposite-sex parent than to their co-twin or same-sex parent, controlling for age. Further, we check that partner similarity is not due to convergence over time, and we test whether assortative mating can be explained by social homogamy or whether phenotypic matching plays a role. This is one of the largest and most comprehensive studies on the genetic and environmental influences on mate choice in any species.

Methods

Participants

Health and lifestyle questionnaires were sent to two cohorts of adult Australian twins and their families (parents, children, partners and siblings), the first in 1988 and for a second time in 1990. The total number of participants was >27,000. For this study, we use data from the twins and their parents and partners, including 22,861 individuals from 6,105 independent families (see Table 1 for sample details). The number of available data varied widely per measured variable. Data collection was approved by the Queensland Institute of Medical Research (QIMR) Ethics Committee, and informed consent was obtained from all participants. Further details about the sample, data collection and zygosity determination can be found elsewhere (Heath, Cloninger, and Martin 1994).

Table 1.

Sample description.

Twins Partner
Male Female Father Mother Male Female
No. individuals 4388 6969 2037 2201 4835 2431
Age (SD) 32.0 (12.4) 34.4 (13.6) 62.4 (9.9) 59.6 (9.7) 40.9 (13.9) 35.8 (12.6)
Open in a new tab

Note: All values for number of individuals include partner-reported data where self-report data were missing. The average age of twins is lower than that of partners because many younger twins do not have partners but are included in the study. For actual couples, men are, on average, 2.7 years older than their partners.

Measures

Education was analyzed as the highest level of education completed, using the following categories: (1) primary school/high school (8–10 years of education), (2) high school (11–12 years of education), (3) apprentice/diploma etc, (4) technical/college, (5) university degree, and (6) university postgraduate. Data were available for 1,794 complete sets of twins and both twins’ partners (hereafter referred to as complete sets), but all available data contributed to estimation of means, variances, and covariances using full information maximum likelihood.

Individuals’ yearly income was assessed with the response options: (1) none, (2) <5,000, (3) $5,000–$10,000, (4) $10,000–$15,000, (5) $15,000–$25,000, (6) $25,000–$35,000, (7) $35,000–$50,000, (8) >$50,000. The average full-time annual income at the time of the questionnaire was ~$25,000. Income of twin’s parents was not assessed. Data for this category were available for a total of 1,197 complete sets.

Religiosity was measured as frequency of church attendance (or other observances), with the response options: (1) rarely, (2) once or twice a year, (3) every month or so, (4) once a week, and (5) more than once a week. Data for this category were available for 1,765 complete sets.

Social attitudes were scored from responses to a list of topics (e.g., birth control, private schooling, casual sex, Asian immigration). Participants were asked to indicate whether or not they agreed with each topic (with a score of 0 or 2, depending on direction) or whether they were uncertain (with a score of 1). On the basis of an exploratory factor analysis, we combined 23 items (those with absolute factor loadings >0.3) into a scale ranging from 0 (most conservative) to 46 (most liberal); see Table A1 for details. Missing scores were set to the sample mean unless there were more than three, in which case the scale score was treated as missing. Data for this category were available for 441 complete sets.

Personality was measured using the short versions of Eysenck’s Personality Questionnaire (EPQ-R; Eysenck, Eysenck, and Barrett 1985) and the revised Tridimensional Personality Questionnaire (TPQ). The EPQ includes the personality scales psychoticism (12 items), neuroticism (12 items), and extraversion (12 items), and the revised TPQ (see Cloninger 1994) includes harm avoidance (18 items), novelty seeking (19 items), reward dependence (12 items), and persistence (5 items). Items were true/false format, and responses were summed. Missing responses were replaced with the sample mean, unless >25% of items on a particular scale were missing, in which case the scale score was treated as missing. In order to minimize departures from normality, scores were converted to a proportional scale before transformation into arcsine values (Freeman and Tukey 1950). Data for this category were available for 439–451 complete sets.

Height and age were self-reported. Data for these categories were available for 439 and 2,195 complete sets, respectively.

Body Mass Index (BMI) was calculated from self-reported height and weight values and log transformed to normalize the distribution. Data for this category were available for 426 complete sets.

Length of relationship with current partner was self-reported. For twin’s parents, this variable was only used if the current partners were the twin’s parents. This information was available for 7,723 couples.

For religiosity, education, income, and age, participants were also asked to give information about their partner. We used this information when the partners did not participate or when there were missing data on any of these variables. Over the whole sample, this increased the available data for these measures by 23%–33%, but it increased data specifically for twins’ partners’ by 110%–180% for these variables. For participants with both sets of information, correlations between self-reported and partner-reported data were 0.87 (religiosity), 0.84 (education), 0.74 (income), and 0.99 (age).

Statistical Analyses

Data preparation was performed in PASW-Statistics 17.0. Statistical analyses employed full-information maximum-likelihood modeling procedures using the statistical package Mx (Neale et al. 2006). Mx accounts for the non-independence of family members by producing maximum-likelihood estimates that are corrected for kinship by explicitly modeling the family relationships within the covariance structure. In a maximum-likelihood model, the goodness of fit of a model to the observed data is distributed as chi-square (χ2). By testing the change in chi-square (Δχ2) against the change in degrees of freedom (Δdf), we can test whether dropping or equating specific model parameters (e.g., the MZ and DZ twin pair correlations) significantly worsens the model fit. In this way, we can test hypotheses regarding those parameters. For all analyses, mean effects of sex, age, and sex by age were controlled for (separately for males and females) by regressing age and standardizing each trait.

Correlations between different family members (twin pairs, partner pairs, parent-offspring, etc) on each trait were determined in Mx. For each trait we investigated whether social homogamy (i.e. partner similarity in social backgrounds) could account for all of the similarity between partners. Heath and Eaves (1985) mathematically show that the social homogamy model makes a number of predictions regarding the correlations between various family members and partners, but power simulations showed that the most statistically powerful prediction is that the correlation between a twin and the co-twin’s partner is equal for MZ and DZ twin pairs and is equal to the correlation between a twin and that twin’s own partner (these correlations are indicated with the symbol † in Table 2). If these correlations are significantly different (i.e. the prediction from social homogamy is violated) then phenotypic matching must play a role (as it is the only alternative). See the study by Heath and Eaves (1985) for further details.

Table 2.

Correlations between family members for each trait after controlling for age and sex effects.

BMI Height Education Income Religiosity Attitudes Neuroticism Extraversion Psychoticism Harm
Avoidance		 Novelty
Seeking		 Reward
Dependence		 Persistence
MZF twins   .69 (.01) .84 (.01) .64 (.01) .32 (.03) .62 (.01) .64 (.01)   .42 (.02)   .45 (.02) .35 (.02)   .42 (.02)   .40 (.02)   .37 (.02)   .35 (.02)
MZM twins   .71 (.02) .87 (.01) .69 (.02) .47 (.04) .62 (.02) .59 (.02)   .36 (.03)   .51 (.03) .37 (.03)   .44 (.03)   .40 (.03)   .37 (.03)   .35 (.03)
DZF twins   .37 (.03) .48 (.02) .47 (.02) .32 (.05) .50 (.02) .51 (.02)   .17 (.03)   .17 (.03) .23 (.03)   .14 (.03)   .12 (.03)   .16 (.03)   .14 (.03)
DZM twins   .41 (.04) .41 (.04) .48 (.03) .28 (.06) .50 (.04) .48 (.04)   .10 (.05)   .17 (.05) .19 (.05)   .15 (.05)   .12 (.05)   .13 (.05)   .11 (.05)
DOS twins   .25 (.03) .47 (.02) .41 (.02) .10 (.04) .39 (.03) .40 (.03)   .12 (.03)   .14 (.03) .09 (.03)   .11 (.03)   .14 (.03)   .08 (.03)   .10 (.03)
Parents   .18 (.03) .22 (.03) .44 (.02) … .70 (.01) .62 (.02)   .10 (.03)   .09 (.03) .14 (.03)   .06 (.03)   .09 (.03)   .08 (.03)   .08 (.03)
Partners †   .12 (.02) .20 (.02) .45 (.01) .08 (.02) .73 (.01) .62 (.01)   .06 (.02)   .04 (.02) .16 (.02)   .03 (.02)   .04 (.02)   .02 (.02)   .03 (.02)
Father-daughter   .20 (.03) .39 (.02) .26 (.02) … .46 (.02) .39 (.02)   .14 (.03)   .19 (.02) .04 (.03)   .15 (.03)   .21 (.03)   .14 (.03)   .09 (.03)
Mother-daughter   .20 (.02) .46 (.02) .25 (.02) … .46 (.02) .43 (.02)   .16 (.02)   .18 (.02) .11 (.02)   .14 (.02)   .13 (.02)   .16 (.02)   .08 (.02)
Father-son   .15 (.04) .42 (.03) .26 (.03) … .47 (.02) .41 (.02)   .15 (.03)   .17 (.03) .13 (.03)   .11 (.03)   .16 (.03)   .12 (.03)   .11 (.03)
Mother-son   .17 (.03) .44 (.02) .19 (.03) … .43 (.02) .39 (.02)   .10 (.03)   .16 (.03) .07 (.03)   .08 (.03)   .14 (.03)   .10 (.03)   .09 (.03)
Father-son’s partner −.02 (.07) .15 (.04) .15 (.03) … .34 (.03) .24 (.04)   .07 (.06)   .07 (.05) .09 (.05)   .03 (.06) −.01 (.05) −.01 (.05)   .10 (.06)
Mother-son’s partner   .05 (.05) .10 (.05) .15 (.03) … .32 (.03) .24 (.04)   .06 (.06) −.06 (.05) .11 (.05)   .01 (.05)   .00 (.05)   .06 (.05)   .03 (.05)
Father-daughter’s partner   .06 (.04) .13 (.04) .22 (.03) … .40 (.02) .21 (.03) −.03 (.05)   .00 (.04) .01 (.04)   .02 (.05) −.01 (.04)   .01 (.05) −.03 (.04)
Mother- daughter’s partner   .06 (.03) .11 (.03) .19 (.03) … .35 (.02) .23 (.03)   .01 (.04)   .03 (.04) .07 (.04) −.02 (.04)   .01 (.04)   .04 (.04)   .02 (.04)
MZF twin-cotwin’s M partner †   .09 (.03) .18 (.03) .38 (.02) .05 (.03) .45 (.02) .37 (.02)   .01 (.03)   .04 (.03) .07 (.03)   .04 (.03)   .07 (.03)   .01 (.03)   .03 (.03)
MZM twin-cotwin’s F partner †   .10 (.04) .18 (.03) .33 (.02) .06 (.04) .43 (.03) .42 (.03)   .03 (.05)   .03 (.05) .09 (.05)   .02 (.05)   .09 (.05)   .01 (.05)   .03 (.05)
DZF twin-cotwin’s M partner †   .04 (.05) .11 (.04) .27 (.03) .06 (.04) .39 (.02) .32 (.03) −.01 (.05) −.03 (.05) .02 (.05) −.01 (.05) −.03 (.05) −.10 (.05) −.04 (.05)
DZM twin-cotwin’s F partner † −.03 (.06) .23 (.07) .31 (.03) .08 (.05) .40 (.04) .32 (.05)   .15 (.07) −.03 (.06) .02 (.07)   .00 (.07) −.12 (.06)   .00 (.07)   .08 (.06)
DZM twin-OS cotwin’s M partner   .02 (.08) .21 (.06) .31 (.04) .06 (.05) .34 (.04) .34 (.05) −.03 (.07) −.06 (.06) .04 (.07)   .02 (.07) −.09 (.07) −.05 (.07)   .06 (.07)
DZF twin-OS cotwin’s F partner   .11 (.06) .10 (.05) .32 (.03) .02 (.06) .26 (.03) .34 (.05)   .06 (.07) −.09 (.06) .18 (.06)   .10 (.06)   .01 (.06) −.04 (.06)   .03 (.06)
Open in a new tab

Note: Values in parentheses are standard errors. Correlations between partner of twin 1 and partner of twin 2 are presented in Table 4.

M = male, F = female, OS = opposite-sex, MZ = monozygotic, DZ = dizygotic.

†

These correlations are used in the test of social homogamy.

We tested whether sexual imprinting plays a role in mate choice by testing whether twins’ partners were more similar to their opposite-sex parent than others who are equally genetically similar to the twin (i.e. same-sex parent, DZ co-twin). This “genetic control” is necessary because of the tendency for assortative mating in humans: twins’ partners might be similar to their opposite-sex parent simply because they tend to partner with someone similar to themselves (and therefore similar to their genetic relatives). Note that this analysis does not account for the possibility of imprinting on same-sex parent or siblings.

To test for a genetic effect on mate choice for each trait, we tested whether MZ twin pairs’ partners (i.e. mate choices) resemble each other more than DZ twin pairs’ partners. For each trait, we also performed the same test after regressing the twin’s own trait values to control for the effects of assortative mating (which increases resemblance between partners of twin pairs, as twins themselves are similar). To test for an “overall” genetic effect on mate choice, combining data from all traits, we included all traits in a single model and constrained the twin pair correlations to be equal across traits, effectively creating a weighted average of twin correlations across traits.

Estimating genetic and environmental influences on traits and mate choice for those traits

We used the classical twin design, in which variation in traits and mate choice is partitioned into that due to genetic sources (additive and nonadditive), that due to family environmental sources (environmental factors shared by twin pairs), and that due to residual sources (Neale and Cardon 1992). Additive genetic variation results from the sum of allelic effects within and across genes. Nonadditive genetic effects include dominance and epistasis, which are allelic interactions within and across genes, respectively. Family environmental factors may include shared home environment, parental style, and uterine environment. Residual variation includes that due to environmental factors not shared by twin pairs (e.g. idiosyncratic experiences), that due to stochastic biological effects, and that due to measurement error. Trait variances are standardized to equal 1, so the genetic, family environmental, and residual parameters equal the proportion of variance accounted for by each source. Note that the proportion of trait variance accounted for by additive genetic factors is the narrow-sense heritability (h2) of the trait and the proportion of trait variance accounted for by all genetic factors is the broad-sense heritability (H2) of the trait.

Partitioning of phenotypic variance into genetic, shared environmental, and residual components can be achieved because MZ twins share all their genes, while DZ twins share on average only one-half of their segregating genes. Thus, if additive genetic influences were the sole source of variance in a trait, we would expect a twin correlations of 1.0 for MZ pairs and 0.5 for DZ pairs. If nonadditive genetic influences were the sole source of variance in a trait, we would expect twin correlations of 1.0 for MZ pairs and, at most, 0.25 for DZ pairs (see Posthuma et al (2003) for an explanation). By definition, if family environment were the sole source of variance in a trait, we would expect a twin correlation of 1 for both MZ and DZ pairs; and conversely, if residual factors were the sole source of variance in a trait we would expect a twin correlation of 0 for both MZ and DZ pairs.

For most traits, observed MZ and DZ twin correlations reflect a combination of these genetic and environmental influences, and (using the expected correlations above) structural equation modeling allows us to determine the combination that best matches the observed data. A limitation of the classical twin design is that there is little statistical power to resolve nonadditive from additive genetic effects and, further, nonadditive genetic effects are confounded with family environmental effects. Preliminary analyses showed little evidence for nonadditivity (in no cases were DZ correlations significantly less than one-half of the MZ correlations), so we estimated only additive genetic variation with the caveat that this estimate includes any nonadditive effects. (When nonadditive effects are not modeled, all nonadditive genetic variance is absorbed into the estimate of additive genetic variance. This is because the expected pattern of twin correlations for nonadditive genetic variation is similar to that for additive genetic variation but very different from that of family environmental and residual variation (Neale and Cardon 1992)).

Results

Partner similarity

Table 3 shows that partners correlated very weakly on some traits (e.g., income and personality) but strongly on others (e.g., religiosity and attitudes). As a first step, we examined whether partner similarity could be accounted for by convergence during the relationship. We did this (cross-sectionally) by testing each trait for a correlation between relationship duration (<1 year to 60 years) and absolute difference scores of couples (see Table A2). With the large sample, there were significant correlations in both directions (i.e., convergence and divergence), but all were of negligible magnitude (R2 < 0.01), also when limiting to relationships <2 years in length. These results suggest that, in general, partner similarity is due to initial choice and not convergence over time.

Table 3.

Means (± SD), partner correlations (including parents), and broad-sense heritability estimates (uncorrected and corrected for assortative mating) for each trait.

Mean ± SD Partner correlation Heritability
Trait Males Females r (95% CI) No. pairs
forming basis		 % (95% CI) Corrected
for AM
Body mass index 24.4 ± 3.3 23.2 ± 4.1 .14 (.11, .17) 3,479 68 (64, 70) 68
Height (cm) 178 ± 7.0 164 ± 6.8 .20 (.17, .24) 3,544 75 (69, 82) 85
Education 4.1 ± 1.3 3.8 ± 1.4 .45 (.42, .147) 7,320 43 (36, 49) 58
Income 2.9 ± 1.6 2.5 ± 1.5 .08 (.04, .11) 4,150 34 (19, 40) 34
Religiosity 6.1 ± 1.5 4.1 ± 1.9 .72 (.69, .75) 7,389 30 (22, 38) 39
Attitudes 25.0 ± 9.1 23.3 ± 8.8 .61 (.58, .64) 3,469 32 (25, 40) 38
Neuroticism 2.5 ± 1.7 1.8 ± 1.5 .07 (.04, .11) 3,586 39 (35, 42) 39
Extraversion 6.7 ± 3.6 7.0 ± 3.5 .05 (.02, .08) 3,545 46 (42, 49) 46
Psychoticism 4.0 ± 3.1 5.1 ± 3.1 .16 (.12, .19) 3,595 36 (28, 39) 36
Harm Avoidance 6.2 ± 4.2 8.0 ± 4.3 .04 (.01, .07) 3,480 40 (37, 44) 40
Novelty Seeking 7.8 ± 3.9 7.8 ± 3.7 .06 (.02, .09) 3,474 38 (34, 42) 38
Reward Dependence 6.5 ± 2.7 8.2 ± 2.5 .04 (.01, .07) 3,477 35 (30, 38) 35
Persistence 3.0 ± 1.5 2.9 ± 1.5 .05 (.02, .08) 3,516 33 (29, 37) 33
Age 41.7 ± 17.0 39.8 ± 16.2 .97 (.95, .99) 8,315 … …
Open in a new tab

Note: AM = assortative mating, CI = confidence interval.

Social homogamy could not completely account for assortative mating on education, attitudes, religiosity, and age (p<.001), indicating that phenotypic matching must play a role in partner similarity. The relevant correlations can be found in Table 2, indicated with the symbol †. The other traits exhibited much weaker assortative mating, so in those cases there was likely too little statistical power to distinguish phenotypic matching from social homogamy (Heath and Eaves 1985).

Trait heritability

MZ and DZ twin pair correlations for each trait can be found in Table 2, and Table 3 shows heritability estimates for each trait. For each trait, twin pair correlations were significantly greater for MZ pairs than DZ pairs (p<.001), and heritability estimates as obtained from univariate twin models ranged between 28% for religiosity to 68% for BMI.

Parental assortative mating (assuming it is due to phenotypic matching) on heritable traits increases the genetic correlation between DZ twins but not MZ twins (who are genetically identical). Therefore, the difference between MZ and DZ correlations is lower and heritability will be underestimated. It is possible to correct heritability estimates for this potential underestimation by using formulas detailed by Martin, Eaves, and Fulker (1978); doing so increased the heritability estimates for several of the traits in this study (see the last column in Table 3). Note that the above issue is unrelated to that described below, where we control for the effect of assortative mating inflating familial resemblance in mate choice.

Genetic and environmental influences on mate choice

Table 4 shows the correlations between twin pair partners (i.e., twin pair correlations for mate choice) on each trait. Correlation between twin pair partners was not significant for BMI, height, and all personality scales, but there were small but significant correlations between twin pair partners for education, income, religiosity, attitudes, and age. For the traits with significant correlation between twin pair partners, we tested if the correlation between MZ twin pair partners was significantly higher than the correlation between DZ twin pair partners. This was not the case for any trait, or over all the traits equated, indicating no significant genetic influence on mate choice. Over all the traits, the nonsignificant heritability estimate for mate choice was 4% and 6% for males and females respectively. After controlling for assortative mating (i.e., regressing the twin’s own trait values), the correlations between twin pair partners decreased, particularly for those variables with stronger assortative mating; again, correlations between twin pair partners did not differ for MZ twins versus DZ twins over all the traits equated or for any specific trait. Over all the traits, the nonsignificant heritability estimate for mate choice after controlling for assortative mating was 5% for males and 8% for females. The relative magnitude of familial (including genetic) effects is limited by the MZ twin correlation; given the confidence intervals in Table 4, we can be confident that familial (including genetic) influences account for <13% of the variation in male and female mate choice averaged across all measured traits, after controlling for assortative mating.

Table 4.

Correlations between twin pair partners (i.e., twin pair correlations for mate choice), uncorrected and corrected for assortative mating.

Mate
Choice		 BMI Height Education Income Religiosity Attitudes Neuroticism Extra-
version		 Psychoticism Harm
avoidance		 Novelty
seeking		 Reward
dependence		 Persistence Age Overall
(95%CI)
Raw data:
MZF   .02 (.07)   .15 (.07) .35 (.03)**   .30 (.04)**   .38 (.02)**   .27 (.05)**   .00 (.08) −.05 (.07)   .15 (.07) −.01 (.07) −.07 (.07)   .05 (.07)   .04 (.07)   .93 (.01)**   .18 (.14, .21)**
MZM   .28 (.11)   .15 (.12) .26 (.05)**   .12 (.06)   .30 (.04)**   .31 (.06)**   .07 (.11)   .10 (.11)   .05 (.12) −.08 (.12) −.02 (.11) −.16 (.10)   .06 (.11)   .92 (.02)**   .14 (.09, .19)**
DZF −.24 (.12)   .01 (.11) .23 (.04)**   .37 (.05)**   .32 (.03)**   .18 (.08)   .04 (.12) −.13 (.12) −.15 (.13)   .03 (.11)   .11 (.10) −.12 (.11) −.01 (.11)   .90 (.02)**   .15 (.10, .20)**
DZM   .06 (.18) −.26 (.15) .18 (.06)*   .24 (.08)*   .36 (.05)**   .30 (.09)*   .32 (.17) −.29 (.14)   .07 (.17) −.33 (.14) −.07 (.14)   .09 (.16)   .09 (.15)   .92 (.03)**   .12 (.05, .19)**
DZOS −.20 (.12) −.07 (.10) .22 (.04)**   .06 (.06)   .26 (.04)**   .40 (.09)** −.30 (.12)   .06 (.10) −.24 (.09) −.02 (.11)   .13 (.10) −.02 (.10)   .08 (.12)   .90 (.02)**   .09 (.04, .13)**
Controlling for assortative mating:
MZF   .03 (.07)   .11 (.08) .20 (.04)**   .24 (.05)**   .12 (.03)**   .07 (.07)   .00 (.07) −.08 (.07)   .14 (.07) −.01 (.07) −.09 (.07)   .03 (.07)   .04 (.07)   .19 (.03)**   .10 (.07, .13)**
MZM   .26 (.11)   .09 (.12) .12 (.06)   .07 (.06) −.01 (.05)   .09 (.10)   .05 (.10)   .09 (.11)   .02 (.12) −.07 (.11) −.06 (.11) −.17 (.10)   .04 (.11)   .09 (.06)   .05 (.00, .06)
DZF −.31 (.13) −.06 (.12) .08 (.05)   .30 (.06)**   .03 (.05) −.03 (.12)   .03 (.11) −.10 (.12) −.17 (.13)   .05 (.11)   .11 (.10) −.12 (.10) −.01 (.11)   .18 (.05)**   .06 (.02, .11)*
DZM   .08 (.17) −.29 (.15) .04 (.08)   .21 (.09)   .09 (.08)   .17 (.14)   .25 (.18) −.33 (.13)   .00 (.19) −.36 (.13) −.09 (.14)   .09 (.17)   .07 (.16) −.02 (.04)   .01 (−.04, .07)
DZOS −.21 (.11) −.09 (.10) .02 (.05) −.09 (.07)   .06 (.06)   .22 (.14) −.23 (.11)   .06 (.10) −.25 (.10) −.02 (.10)   .11 (.11) −.03 (.11)   .09 (.12)   .00 (.07) −.02 (−.06, .03)
Open in a new tab

Note: Values in parentheses are standard deviations, except where indicated. “Overall” correlations (and 95% confidence intervals) are estimated by constraining twin correlations to be equal across all traits (except for age in the raw data).

MZF = monozygotic female pairs, MZM = monozygotic male pairs, DZF = dizygotic female pairs, DZM = monozygotic male pairs, DZOS = dizygotic opposite-sex pairs.

*

Significant at P < .01.

**

Significant at P < .001.

However, specifically for income and age, correlations between female MZ and DZ twin pair partners were highly significant and similar in size, indicating a genuine family environmental influence on women’s mate choice for these traits even after controlling for assortative mating. Modeling showed significant family environmental influences accounting for 25% and 18% of the variance in mate choice for income and age, respectively.

On the other hand, there was no evidence for the sexual imprinting hypothesis. Twins’ partners were not significantly more similar in any trait to the twin’s opposite-sex parent than to the twin’s same-sex parent or a DZ co-twin of either sex, nor was there even a trend in this direction (see Table 2 for the relevant correlations).

Discussion

Using a large sample of twins and their parents and partners, we found that genetic variation accounts for very little individual variation in human mate choice. Over all the measured traits, twins’ partners correlated very weakly (except for age), and MZ twins’ partners did not correlate significantly more than DZ twins’ partners (mate choice heritability = 4% for males and 6% for females). After correcting for the twin’s own trait values (i.e. correcting for assortative mating), the correlation between the twins’ partners over all traits was even lower, and was again not significantly different between MZ and DZ pairs (mate choice heritability = 5% for males and 8% for females, respectively). Because of the large sample size and tight confidence intervals, we can be confident that familial effects (including genetic effects) accounted for <13% of the variance in male and female mate choice across all of the traits. Furthermore, there were no significant genetic effects on mate choice in either males or females for any individual trait. Nor were there any sexual imprinting effects, whereby the opposite-sex parent is used as the template of a desirable mate. That is, for males and females in every measured trait, twins’ partners were not more similar to the twins’ opposite-sex parent than to others with the same level of genetic sharing with the twin (i.e., same-sex parent, DZ co-twin). There was, however, a genuine influence of family environment on female mate choice in terms of age and income, supporting previous evolutionary predictions of such an effect (Buunk, Park, and Dubbs 2008).

The finding of near-zero heritability of human mate choice accords with the findings of Lykken and Tellegen (1993), and extends those results with a broader range of traits, a larger sample, sex-specific analyses, and control of the effects of assortative mating. On the other hand, these results contradict the conclusions of Rushton and Bons (2005), who report a substantial genetic component (~30%) to mate choice. However, their reported correlations between twins’ partners actually suggest a lower genetic estimate (18%), and this genetic component is nonsignificant. Thus, their larger estimate of heritability may reflect greater sampling error due to the much smaller sample size in that study.

Given that variation in practically every behavioral trait studied has been revealed to be under substantial genetic influence (Bouchard and McGue 2003), it is remarkable that a choice behaviour so central to individuals’ lives (Headey, Muffels, and Wagner 2010) exhibits a near-zero genetic component. A difference between mate choice and many other choice behaviours (e.g., smoking (Vink, Willemsen, and Boomsma 2005), food (Teucher et al. 2007): ~50% heritability) is that choice of a mate requires reciprocity: person A cannot “choose” person B in this sense unless person B also chooses person A. Therefore, the choice of partner may only partially reflect underlying mate preferences, and the extent to which choice reflects preferences will differ between individuals according to characteristics such as their mate value and mating strategies (Penke et al. 2007). Thus, the formation of a couple depends on a complex interaction between the characteristics of both members, characteristics that may themselves be heritable. Components of mate value (e.g., height in men, intelligence, BMI: 50%–90% heritable) and mating strategies (sociosexuality, number of sexual partners: ~50%–60% heritable) are known to be substantially influenced by genes (Bailey et al. 2000; Zietsch et al. 2008), but the heritability of mate preferences has not been investigated in humans. It is thought that much variation in mate choice behavior (e.g., preferences and strategies) is due to facultative calibration to local reproductive conditions (e.g., pathogen prevalence, resource availability) and availability of mates (e.g., operational sex ratio; (Schmitt 2005; DeBruine et al. 2010)), which is consistent with the high levels of nongenetic variation we see here. Ideally, we would measure mating preferences, strategies, and behavior; mate value (attractiveness, fitness traits); and relevant environmental factors in a large sample of twins and their partners. Genetic and/or environmental variation, covariation, and interaction between these factors, and their association with mate choice, would be very revealing in terms of how sexual selection operates in humans and perhaps in other pair-bonding species. Knowledge of the relationships between these factors could also allow the development of more sophisticated twin/family models that take into account the reciprocity of human mate choice; for example, knowledge of mate value and mate preferences could be used to estimate the extent to which each individual had free choice of a mate, which could in turn alter expected covariances between partners of the twins.

One positive finding in our results is an influence of family environment on female mate choice in terms of the income and age of their partner, even after controlling for the effect of assortative mating. For both traits, female twins’ spouses were correlated modestly but highly significantly, to the same degree, in DZ twins and MZ twins. This family environment effect, which explains 18% and 25% of the variance in mate choice for age and income, respectively, may reflect parental influence over mating decisions. On the basis of evolutionary theory and anthropological evidence, this parental influence is expected to primarily involve pressure on a daughter to mate with a high-investing (e.g., older and higher-income) man (Buunk, Park, and Dubbs 2008). There was little evidence for substantial familial effects on other aspects of mate choice. In particular, the sexual imprinting hypothesis was not supported: twins’ spouses showed no more similarity to the twin’s opposite-sex parent than would be expected given assortative mating and genetic relatedness between the twin and his or her parent. A strong previous demonstration of sexual imprinting was subsequently found to contain serious errors (Bereczkei, Hegedus, and Hajnal 2009). The distinct lack of an effect in the current large sample using appropriate genetic controls indicates that sexual imprinting does not play a role in human mate choice, at least in terms of the range of physical and psychological traits investigated here.

Given that assortative mating remains the primary detectable pattern to mate pairings in humans, it is worth briefly describing how our understanding of it has been furthered by this study. First, we extended previous findings for intelligence, personality, and social attitudes that similarity between partners is not due to becoming more similar over time (Mascie-Taylor 1989; Caspi, Herbener, and Ozer 1992), since partner similarity was not substantively related to relationship length for any trait. Second, we showed that strong assortative mating (i.e. on age, social attitudes, education, religiosity) was not solely due to social homogamy, so phenotypic matching must play a significant role. (Weak assortative mating (e.g. height, BMI) afforded little statistical power to detect phenotypic matching.) Given its effect of increasing the genetic variance of traits in the population, phenotypic matching is an area worth exploring further with regard to the maintenance of genetic variation in fitness-related traits (Keller and Miller 2006; Penke, Denissen, and Miller 2007). For example, a possible cause of phenotypic matching could be a genetic correlation between preference for a trait and the trait itself, which could reflect runaway (Fisherian) sexual selection (Andersson 1994).

Despite the large, genetically informative sample and broad range of traits, there are some important limitations to this study. One is that we may not have measured the “right” traits; that is, unmeasured traits such as pheromones or facial characteristics could potentially inform mating decisions more than the traits we investigate here. However, countless human mating studies support the importance of preferences for intelligence (correlated with education), age, earning capacity (men’s), height (men’s), BMI (especially women’s), and personality traits (Buss 1989; Kurzban and Weeden 2005; Fink et al. 2007; Todd et al. 2007; Lukaszewski and Roney 2010). Therefore, although we have not measured all of the traits that are important to human mate selection, we are confident we have measured some of them. Another limitation is that the mate pairings sampled are biased towards established couples. Although most children today are born to pair-bonded couples, a substantial portion result from short-term sexual liaisons and this was probably more common before the recent availability of contraception. Evolutionary psychology shows that humans have psychological adaptations to short-term as well as long-term mating (Gangestad and Simpson 2000; Gangestad and Thornhill 2008), indicating the importance of short-term sexual liaisons in the evolutionary past. Although the relationships sampled in this study range in length from one month to 60 years, the bias towards established couples limits evolutionary interpretations of the findings. Another issue is that the potential for non-random missingness of data cannot be discounted, particularly as fewer partners than twins participated. However, there were no large differences in the mean values of the twins and those of their partners, suggesting that that the partners’ lower rate of participation of partners was not substantively related to any traits of interest. Finally, although our sample is large compared with most previous genetic studies of mating, our power is still limited and there may be extremely low heritability in mate choice that we are not detecting. A recent large study of mate choice in birds (N = 8,500) found a significant heritability of <3% (Qvarnstrom, Brommer, and Gustafsson 2006). Our study could not detect genetic contributions to mate choice of this magnitude, and these could still be important over evolutionary time.

In summary, we found near-zero genetic effects and no sexual imprinting effects on human mate choice on a range of physical, psychological, and demographic traits. Despite being one of the most important choices in human life, variation in partner choice followed no apparent order aside from a small family environmental influence on the age and income of females’ mate choices and the similarity of partners on some traits. Partner similarity was due at least partly to matching on phenotype, and the causes of phenotypic matching should be further investigated. Our results are consistent with theory and data suggesting that much variation in mate choice behavior arises from facultative calibration to local reproductive conditions and mating markets; however, an integrated evolutionary genetic understanding of human mating and pair bonding calls for examining mating preferences, strategies, and behavior, fitness traits, and relevant environmental factors in a large sample of twins and their families and partners.

Supplementary Material

01

Acknowledgements

We greatly thank the twins analyzed in this study for their participation. BPZ is supported by a University of Queensland Postdoctoral Fellowship. KJHV is supported by an ANZ Trustees PhD scholarship in Medical Research. This study was supported by National Institutes of Health grants AA07535, AA07728, AA10249, AA11998, MH31302, and National Health and Medical Research Council (Australia) grants 941177 and 971232. We thank Lars Penke and an anonymous reviewer for invaluable comments on an earlier version of this manuscript.

Footnotes

Declaration of interest: None

The authors declare no conflict of interests.

Contributor Information

Karin J. H. Verweij, Email: Karin.Verweij@qimr.edu.au.

Andrew C. Heath, Email: aheath@wustl.edu.

Nicholas G. Martin, Email: Nick.Martin@qimr.edu.au.

References

  1. Anderson KG. How well does paternity confidence match actual paternity? Evidence from worldwide nonpaternity rates. Current Anthropology. 2006;47(3):513–520. [Google Scholar]
  2. Andersson M. Sexual selection. New Jersey: Princeton University Press; 1994. [Google Scholar]
  3. Bailey JM, Kirk KM, Zhu G, Dunne MP, Martin NG. Do individual differences in sociosexuality represent genetic or environmentally contingent strategies? Evidence from the Australian twin registry. Journal of Personality and Social Psychology. 2000;78(3):537–545. doi: 10.1037//0022-3514.78.3.537. [DOI] [PubMed] [Google Scholar]
  4. Bereczkei T, Gyuris P, Weisfeld GE. Sexual imprinting in human mate choice. Proceedings of the Royal Society of London Series B-Biological Sciences. 2004;271(1544):1129–1134. doi: 10.1098/rspb.2003.2672. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Bereczkei T, Hegedus G, Hajnal G. Facialmetric similarities mediate mate choice: sexual imprinting on opposite-sex parents (Retraction of vol 276, pg 91, 2009) Proceedings of the Royal Society B-Biological Sciences. 2009;276(1659):1199. doi: 10.1098/rspb.2008.1774. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Bouchard TJ, McGue M. Genetic and environmental influences on human psychological differences. Journal of Neurobiology. 2003;54(1):4–45. doi: 10.1002/neu.10160. [DOI] [PubMed] [Google Scholar]
  7. Buss DM. Sex-differences in human mate preferences: Evolutionary hypotheses tested in 37 cultures. Behavioral and Brain Sciences. 1989;12(1):1–14. [Google Scholar]
  8. Buss DM, Barnes M. Preferences in human mate selection. Journal of Personality and Social Psychology. 1986;50(3):559–570. [Google Scholar]
  9. Buunk AP, Park JH, Dubbs SL. Parent-offspring conflict in mate preferences. Review of General Psychology. 2008;12(1):47–62. [Google Scholar]
  10. Caspi A, Herbener ES, Ozer DJ. Shared experiences and the similarity of personalities: A longitudinal study of married couples. Journal of Personality and Social Psychology. 1992;62(2):281–291. doi: 10.1037//0022-3514.62.2.281. [DOI] [PubMed] [Google Scholar]
  11. Cloninger CR. The Temperament and Character Inventory (TCI): A guide to its development and use. St Louis, Missouri: Washington University; 1994. [Google Scholar]
  12. DeBruine LM, Jones BC, Crawford JR, Welling LLM, Little AC. The health of a nation predicts their mate preferences: cross-cultural variation in women's preferences for masculinized male faces. Proceedings of the Royal Society B-Biological Sciences. 2010;277(1692):2405–2410. doi: 10.1098/rspb.2009.2184. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Eastwick PW, Finkel EJ. Sex differences in mate preferences revisited: Do people know what they initially desire in a romantic partner? Journal of Personality and Social Psychology. 2008;94(2):245–264. doi: 10.1037/0022-3514.94.2.245. [DOI] [PubMed] [Google Scholar]
  14. Eysenck SBG, Eysenck HJ, Barrett P. A revised version of the Psychoticism Scale. Personality and Individual Differences. 1985;6(1):21–29. [Google Scholar]
  15. Feingold A. Matching for attractiveness in romantic partners and same sex friends: A meta-analysis and theoretical critique. Psychological Bulletin. 1988;104(2):226–235. [Google Scholar]
  16. Fink B, Neave N, Brewer G, Pawlowski B. Variable preferences for sexual dimorphism in stature (SDS): Further evidence for an adjustment in relation to own height. Personality and Individual Differences. 2007;43(8):2249–2257. [Google Scholar]
  17. Freeman MF, Tukey JW. Transformations related to the angular and the square root. Annals of Mathematical Statistics. 1950;21(4):607–611. [Google Scholar]
  18. Gangestad SW, Simpson JA. The evolution of human mating: Trade-offs and strategic pluralism. Behavioral and Brain Sciences. 2000;23(4):573–587. doi: 10.1017/s0140525x0000337x. [DOI] [PubMed] [Google Scholar]
  19. Gangestad SW, Thornhill R. Human oestrus. Proceedings of the Royal Society B-Biological Sciences. 2008;275(1638):991–1000. doi: 10.1098/rspb.2007.1425. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Hatemi PK, Hibbing JR, Medland SE, Keller MC, Alford JR, Smith KB, Martin NG, Eaves LJ. Not by twins alone: Using the extended twin family design to investigate the genetic basis of political beliefs. American Journal of Political Science. 2010;54(3):798–814. [Google Scholar]
  21. Havlicek J, Roberts SC. MHC-correlated mate choice in humans: A review. Psychoneuroendocrinology. 2009;34(4):497–512. doi: 10.1016/j.psyneuen.2008.10.007. [DOI] [PubMed] [Google Scholar]
  22. Headey B, Muffels R, Wagner GG. Long-running German panel survey shows that personal and economic choices, not just genes, matter for happiness. Proceedings of the National Academy of Sciences of the United States of America. 2010 doi: 10.1073/pnas.1008612107. electronic publication ahead of print. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Heath AC, Cloninger CR, Martin NG. Testing a model for the genetic-structure of personality - a comparison of the personality systems of Cloninger and Eysenck. Journal of Personality and Social Psychology. 1994;66(4):762–775. doi: 10.1037//0022-3514.66.4.762. [DOI] [PubMed] [Google Scholar]
  24. Heath AC, Eaves LJ. Resolving the effects of phenotype and social background on mate selection. Behavior Genetics. 1985;15(1):15–30. doi: 10.1007/BF01071929. [DOI] [PubMed] [Google Scholar]
  25. Keller MC, Miller G. Resolving the paradox of common, harmful, heritable mental disorders: Which evolutionary genetic models work best? Behavioral and Brain Sciences. 2006;29(4):385–404. doi: 10.1017/S0140525X06009095. [DOI] [PubMed] [Google Scholar]
  26. Kendrick KM, Hinton MR, Atkins K, Haupt MA, Skinner JD. Mothers determine sexual preferences. Nature. 1998;395(6699):229–230. doi: 10.1038/26129. [DOI] [PubMed] [Google Scholar]
  27. Koenig LB, McGue M, Iacono WG. Rearing environmental influences on religiousness: An investigation of adolescent adoptees. Personality and Individual Differences. 2009;47(6):652–656. doi: 10.1016/j.paid.2009.06.003. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Kurzban R, Weeden J. HurryDate: Mate preferences in action. Evolution and Human Behavior. 2005;26(3):227–244. [Google Scholar]
  29. Lukaszewski AW, Roney JR. Kind toward whom? Mate preferences for personality traits are target specific. Evolution and Human Behavior. 2010;31(1):29–38. [Google Scholar]
  30. Lykken DT, Tellegen A. Is human mating adventitious or the result of lawful choice: A twin study of mate selection. Journal of Personality and Social Psychology. 1993;65(1):56–68. doi: 10.1037//0022-3514.65.1.56. [DOI] [PubMed] [Google Scholar]
  31. Martin NG, Eaves LJ, Fulker DW. Twin Research: Psychology and Methodology. New York: Alan R Liss, Inc.; 1978. Genetics of social and sexual attitudes. [Google Scholar]
  32. Martin NG, Eaves LJ, Heath AC, Jardine R, Feingold LM, Eysenck HJ. Transmission of social attitudes. Proceedings of the National Academy of Sciences of the United States of America. 1986;83(12):4364–4368. doi: 10.1073/pnas.83.12.4364. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Mascie-Taylor CGN. Spouse similarity for IQ and personality and convergence. Behavior Genetics. 1989;19(2):223–227. doi: 10.1007/BF01065906. [DOI] [PubMed] [Google Scholar]
  34. Miller GF. The mating mind: How sexual choice shaped the evolution of human nature. New York: Doubleday; 2000. [Google Scholar]
  35. Nagoshi CT, Johnson RC, Ahern FM. Phenotypic assortative mating vs social homogamy among Japanese and Chinese parents in the Hawaii family study of cognition. Behavior Genetics. 1987;17(5):477–485. doi: 10.1007/BF01073114. [DOI] [PubMed] [Google Scholar]
  36. Neale MC, Boker SM, Xie G, Maes HH. Mx: Statistical modeling. 7th ed. VCU Box 900126, Richmond, VA 23298: Department of Psychiatry; 2006. [Google Scholar]
  37. Neale MC, Cardon LR. Methodology for genetic studies of twins and families. Boston: Kluwer; 1992. [Google Scholar]
  38. Oetting S, Prove E, Bischof HJ. Sexual imprinting as a two-stage process: Mechanisms of information storage and stabilization. Animal Behaviour. 1995;50:393–403. [Google Scholar]
  39. Penke L, Denissen J, Miller G. The evolutionary genetics of personality. European Journal of Personality. 2007;21:549–587. [Google Scholar]
  40. Penke L, Todd PM, Lenton AP, Fasolo B. How self-assessments can guide human mating decisions. In: Geher G, Miller GF, editors. Mating intelligence: Sex, relationships, and the mind's reproductive system. Mahwah, NJ: Lawrence Erlbaum; 2007. [Google Scholar]
  41. Posthuma D, Beem AL, de Geus EJC, van Baal GCM, von Hjelmborg JB, Lachine I, Boomsma DI. Theory and practice in quantitative genetics. Twin Research. 2003;6(5):361–376. doi: 10.1375/136905203770326367. [DOI] [PubMed] [Google Scholar]
  42. Price RA, Vandenberg SG. Spouse similarity in American and Swedish couples. Behavior Genetics. 1980;10(1):59–71. doi: 10.1007/BF01067319. [DOI] [PubMed] [Google Scholar]
  43. Qvarnstrom A, Brommer JE, Gustafsson L. Testing the genetics underlying the co-evolution of mate choice and ornament in the wild. Nature. 2006;441(7089):84–86. doi: 10.1038/nature04564. [DOI] [PubMed] [Google Scholar]
  44. Reynolds CA, Baker LA, Pedersen NL. Models of spouse similarity: Applications to fluid ability measured in twins and their spouses. Behavior Genetics. 1996;26(2):73–88. doi: 10.1007/BF02359886. [DOI] [PubMed] [Google Scholar]
  45. Reynolds CA, Baker LA, Pedersen NL. Multivariate models of mixed assortment: Phenotypic assortment and social homogamy for education and fluid ability. Behavior Genetics. 2000;30(6):455–476. doi: 10.1023/a:1010250818089. [DOI] [PubMed] [Google Scholar]
  46. Rushton JP, Ann Bons T. Mate choice and friendship in twins - Evidence for genetic similarity. Psychological Science. 2005;16(7):555–559. doi: 10.1111/j.0956-7976.2005.01574.x. [DOI] [PubMed] [Google Scholar]
  47. Schielzeth H, Burger C, Bolund E, Forstmeier W. Sexual imprinting on continuous variation: do female zebra finches prefer or avoid unfamiliar sons of their foster parents? Journal of Evolutionary Biology. 2008;21(5):1274–1280. doi: 10.1111/j.1420-9101.2008.01568.x. [DOI] [PubMed] [Google Scholar]
  48. Schmitt DP. Sociosexuality from Argentina to Zimbabwe: A 48-nation study of sex, culture, and strategies of human mating. Behavioral and Brain Sciences. 2005;28(2):247–275. doi: 10.1017/s0140525x05000051. [DOI] [PubMed] [Google Scholar]
  49. Teucher B, Skinner J, Skidmore PML, Cassidy A, Fairweather-Tait SJ, Hooper L, Roe MA, Foxall R, Oyston SL, Cherkas LF, Perks UC, Spector TD, MacGregor AJ. Dietary patterns and heritability of food choice in a UK female twin cohort. Twin Research and Human Genetics. 2007;10(5):734–748. doi: 10.1375/twin.10.5.734. [DOI] [PubMed] [Google Scholar]
  50. Todd PM, Penke L, Fasolo B, Lenton AP. Different cognitive processes underlie human mate choices and mate preferences. Proceedings of the National Academy of Sciences of the United States of America. 2007;104(38):15011–15016. doi: 10.1073/pnas.0705290104. [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Vink J, Willemsen G, Boomsma D. Heritability of smoking initiation and nicotine dependence. Behavior Genetics. 2005;35(4):397–406. doi: 10.1007/s10519-004-1327-8. [DOI] [PubMed] [Google Scholar]
  52. Watson D, Klohnen EC, Casillas A, Simms EN, Haig J, Berry DS. Match makers and deal breakers: Analyses of assortative mating in newlywed couples. Journal of Personality. 2004;72(5):1029–1068. doi: 10.1111/j.0022-3506.2004.00289.x. [DOI] [PubMed] [Google Scholar]
  53. Wood D, Brumbaugh CC. Using revealed mate preferences to evaluate market force and differential preference explanations for mate selection. Journal of Personality and Social Psychology. 2009;96(6):1226–1244. doi: 10.1037/a0015300. [DOI] [PubMed] [Google Scholar]
  54. Zietsch BP, Morley KI, Shekar SN, Verweij KJH, Keller MC, Macgregor S, Wright MJ, Bailey JM, Martin NG. Genetic factors predisposing to homosexuality may increase mating success in heterosexuals. Evolution and Human Behavior. 2008;29:424–433. [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

01
NIHMS396401-supplement-01.pdf (760.8KB, pdf)

RESOURCES