Contact Us

Use the form on the right to contact us.

You can edit the text in this area, and change where the contact form on the right submits to, by entering edit mode using the modes on the bottom right. 


123 Street Avenue, City Town, 99999

(123) 555-6789


You can set your address, phone number, email and site description in the settings tab.
Link to read me page with more information.

Preference for Facial Self-Similarity (Part I)

Preference for Facial Self-Similarity (Part I)

As of 1/13/15, this is part of ongoing research at Yale University. 



While it is often said that “beauty is in the eye of the beholder,” this claim is somewhat false. There is widespread consensus both between individuals and between cultures as to whose faces are attractive, and to what degree they are attractive. Qualities such as bilateral symmetry, averageness, and secondary sex characteristics have all been found to affect judgments of beauty (Thornhill & Gangestad 1999). If we take an evolutionary view, this is because our facial preferences are adaptive in some way and arrowed to our evolutionary history (Symon 1987).

Like other species, humans evolved adaptations to improve reproductive success (RS). Mate selection is one such adaption. In order to ensure the promotion of our genetic material, we would want to select a mate with whom we can reproduce successfully (Symon 1987). We might discriminate based on facial features because those cues reflect honest information about an individual’s health (Santos 2012a). For example, fluctuating asymmetry tends to be caused by mutations, pathogens, and toxins; therefore, our facial preference for symmetry might be related to its ability to communicate genetic health (Thornhill & Gangestad 1999).

 Brad Pitt's face, for example, is very symmetrical. 

Brad Pitt's face, for example, is very symmetrical. 

Facial resemblance is another helpful cue in mate selection because it signals genetic relatedness and kinship. The importance of this kinship signal is twofold. First, when judging a potential mate, we tend to avoid kin because genetic diversity is evolutionarily advantageous, and there are negative consequences to inbreeding (Bittles & Neel 1994). The Westermarck effect, for example, claims that children reared together tend to see each other as unappealing (Wolf 1995). But kinship is also important in non-sexual contexts: our relatives are often members of our social in-group and so we tend to like them more. People are more likely to trust partners in economic games when the image of the partner’s face is subtly manipulated to resemble their own (DeBruine 2002).   


(A) Effect of facial similarity on attractiveness judgments

How does facial similarity affect attractiveness judgments? Repeated exposure to a stimulus increases positive affect towards that stimulus. Theories of generalized exposure would predict that we would find self-similar faces more attractive, as our exposure to our own face is very high (Rhones, Halberstady, & Brajkovich, 2001). Debruine (2004) tests this hypothesis using computer graphic image manipulation to investiage the effect of facial similarity on attractiveness ratings. She argued that our attractiveness ratings would be moderated by evolutionary reactions to kinship markers. Debruine first created composite faces for different sexes and phenotypic categories (for example, European, East Asian, West Asian). Participants were photographed, and their faces were morphed 50% with the composite faces. Later, the participants were given a series of two-alternative forced-choice decisions between (composite morphed 50% with themselves) and facially dissimilar (composite morphed 50% with a different face) photographs. Using these methods, Debruine found that facial resemblance increased attractiveness ratings in same-sex faces more than in other-sex faces, an effect that implies that our reactions to facial resemblance are not simply a generalized exposure effect (Debruine et al. 2004).

Taken from Debruine (2004). Column 1: Participant Face; Column 2: Same-Sex Composite Face; Column 3: Same-Sex Self-Similar; Column 4: Other-Sex Composite Face; Column 5: Other-Sex Self-Similar

DeBruine explains that “attractiveness judgments of other-sex faces are more likely to include an evaluation of desirability to the judge as a sexual or romantic partner, while judgments of same-sex faces are more likely to be evaluated on imagined desirability to other-sex people or on non-sexual, general positive regard, such as likeability.” And that may very well be true for heterosexual individuals. But for homosexual people, same-sex relationships are not necessarily non-sexual prosocial ones, and other-sex relationships do not necessarily contain a sexual evaluation. In the present research, we intend to test the effect of facial resemblance on attractiveness ratings for homosexual people and compare it to that of heterosexual people. The kin-related evolutionary predictions for homosexual individuals depend on how homosexuality might have evolved.


(B) Theories of homosexuality

Examples of homosexuality can be found throughout human written history: Wilson (1978) argues that homosexuality is normal in human societies, found historically and across cultures at relatively low levels (estimates range from 1 to 10% of the population). There is a universality of homosexuality that seems counterintuitive to our understanding of evolutionary history. The main tenant of human evolution is that we develop traits that enhance our RS, but whereas heterosexual erotic behavior has conceptive benefits, homosexual erotic behavior does not.[1] Natural selection predicts that the extent to which variations of genes persist in the population depends on the effects on fitness, and by that logic, we might expect genes for homosexuality to be selected out and prevalence to be low. If we assume homosexual individuals have lower RS than heterosexual individuals, how can homosexuality be such a common trait?

1.     Kin Selection Hypothesis

According to legend, when evolutionary biologist John Haldane was asked, “Would you lay down your life to save your brother?” he responded, “No, but I would to save two brothers or eight cousins.” (Clark 1969). Haldane wasn’t being selfish. Instead, he was making a genetic argument. Brothers share, on average, half of your genetic material; cousins share about one-eighth (Santos 2012a). Thus Haldane is stating that he would give up his own life (and, presumably RS) in exchange for a genetically equal amount.

The premise of inclusive fitness is the same – for our genes to continue to the next generation, a person can either reproduce or promote the RS of their relatives (Hamilton 1964). The kin-selection hypothesis, strongly tied to Hamilton’s idea of inclusive fitness, states that homosexuals can forego their individual reproduction altruistically to assist their kin in reproduction. However, while this hypothesis explains the lack of sexual behavior, it fails to explain homosexual behavior specifically. Moreover, there is little evidence to suggest that homosexual individuals actively assist kin RS (Santos 2012b).

2.     Balanced Polymorphism hypothesis

Edward Miller of the University of New Orleans argues that the persistence of homosexuality can be explained if sexual orientation is a polygenetic trait (Miller 1999). The model for homosexuality is akin to the Polymorphism model of mental illness (Polimeni & Reiss 2003). Perhaps homosexual behaviors themselves are not helpful for direct fitness, but are correlated with behaviors that assist in direct fitness. Imagine that a number of genes influence our sexual orientation; inheritance of all of the alleles produces homosexuality and a low RS. Having none of the alleles results in heterosexuality with moderately low RS. Having some, but not all, of the alleles is the successful middle ground – a heterosexual individual with a higher RS. Miller (1999) predicts, for example, that homosexual men would be more feminine than their heterosexual counterparts. Heterosexual men with some of the genes predisposing one to homosexuality would be slightly more feminine than heterosexual men without those genes. That femininity could manifest itself in the form of decreased aggression and greater empathy, traits that females might find attractive in a mate (Miller 1999).

 Evidence for this theory is limited, but does exist. One study finds that more psychologically masculine females and feminine males are more likely to be homosexual but, when heterosexual, have more other-sex sexual partners. The same study looks at twin pairs, finding that heterosexuals with homosexual twins have, on average, more other-sex sexual partners than heterosexual twin pairs (Zietsch et al. 2008). This data would suggest that genes predisposing one to homosexuality might confer a reproductive benefit to heterosexuals, a theory that helps to explain the propagation of homosexuality across generations. However, the specific genes or traits involved in this mechanism are unknown.  

3. Alliance Maintenance hypothesis

The alliance maintenance hypothesis proposes that homosexual behavior could have evolved because it strengthens same-sex bonds that aid and contribute to reproduction and survival through resource sharing and mutual defense against predators (Muscarella et al. 2005). Homosexual behavior might increase your social status and same-sex coalitional alliances, which might in turn improve fitness. In this model, engaging in homosexual behavior is simply an exaptation of pro-sociality. However, there are many ways to promote alliance without sexual behavior; if this theory is true, why homosexual behavior specifically? 


C) Previous research on homosexual preferences and their implication for homosexual facial-similarity preferences

Prior research has shown sexual preferences moderates judgments of facial attractiveness. For homosexual men, sexual desire is positively correlated with preferences for exaggerated sex-typical shape cues in same-sex but not other-sex individuals (Welling et al. 2013). In the same way that heterosexual males tend to look for exaggerated feminine cues in women, homosexual males preferred exaggerated masculine cues in men.

One study has also found that homosexual preference is further moderated by sexual self-labels – “top,” meaning a man who prefers the penetrative role, “bottom,” meaning a man who prefers the receptive role, and “versatile,” meaning a man willing to perform either role. Contrary to the prediction of the Balanced Polymorphism model, where homosexuality and feminine preferences might be linked, Zheng, Hart, and Zheng (2012) found that “tops” prefer more masculine faces, “bottoms” prefer more feminized faces, and “versatiles” did not show a significant preference.

Teuscher & Teuscher (2007) find that homosexual and heterosexual males both exhibited a youth bias, but only when judging a sexual partner. In other words, homosexual males preferred youthful traits in male faces but not in female faces, and heterosexual males showed the opposite trend, preferring youthful traits in female faces but not in male faces. Even in the absence of conceptive benefits of sex, homosexuals seem to be influenced by the same evolutionary concerns when choosing a sexual partner. This finding suggests that homosexuals might show a preference for a dissimilar same-sex faces, a result akin to what would be predicted by the Balanced Polymorphism model.

However, most previous research focuses on homosexual preference for masculinity, femininity, and sexually dimorphic characteristics. The present research is unique in looking at the effect of sexual preference on self-similarity biases to gain insight into which evolutionary theories of same-sex sexual behavior make the most sense. Specifically, the present experiment tests whether homosexual self-similarity preferences are similar to heterosexual ones, and if self-similar preferences exist at all when sexual behavior does not have conceptive benefits. In doing so, this research will add to the body of literature about how evolution shapes our facial preferences.



Bittles, A.H., Neel, J.V. (1994). The costs of human inbreeding and their implications for variation at the DNA level. Nature Genetics. Vol. 8(8): 117–121.

Clark (1968). JBS: The Life and Work of J. B. S. Haldane. (New York: Coward-McCann).

DeBruine, L.M. (2002). Facial resemblance enhances trust. Proc. R. Soc. Lond. B, 269:      1307-1312.  

Debruine L.M. (2004). Facial resemblance increases the attractiveness of same-sex faces more than other sex faces. Proc. R. Soc. Lond. B. 271: 2085-2090.

Devlin, S.J. Dong, H.K., & Brown, M. (1993). Selecting a scale for measuring quality. Marketing Research. Vo1. 5(3): 12-17.

Miller, E.M. (1999). Homosexuality, birth order, and evolution: towards a equilibrium reproductive economics of homosexuality. Department of Economics and Finance Working Papers1991-2006.Paper 19.

 Muscarella et al. (2005). The alliance theory of homosexual behavior and the perception of social status and reproductive opportunities. Neuro Endocrinol Lett. 26(6), 771-774.   

Polemini, J., & Reiss, J.P. (2003). Evolutionary perspectives on Schizophrenia. Can J Psychiatry, Vol 48(1): 34-39.

Santos, L. (2012a). What sexual behavior is for. Sex, Evolution, and Human Nature. Lecture conducted from Yale University, New Haven, CT.

 Santos, L. (2012b). Why should we be nice? Sex, Evolution, and Human Nature. Lecture conducted from Yale University, New Haven, CT.

Symon, D. (1987). If we’re all Darwinians, what’s the fuss about? Crawford, C., Smith, M., Krebs, D. (Eds.), Sociobioogy and Psychology: Ideas, Issues and Applications, 121-148.

Thornhill, R. & Gangestad, S.W. (1999). Facial attractiveness. Trends in Cognitive Science. 3, 452-460.

Wilson, E.O. (1978). On Human Nature. Cambridge: Harvard University Press.

Welling L.L.M. et al. (2013). Self-reported sexual desire in homosexual men and women predicts preferences for sexually dimorphic facial cues. Arch Sex Behav, 42: 785-791.

Zietsch et al. (2008). Genetic factors predisposing to homosexuality may increase mating success in heterosexuals. Evolution and Human Behavior, 29(6), 424-433.

[1] It is true, however, that homosexual individuals often have children. The famous homosexual poet, Oscar Wilde, for example, had two children. However, we assume that homosexual men and women have fewer children, on average, than heterosexual men and women; thus in comparison, genes for homosexuality should fare poorly.