1. Introduction 

Studies of human attractiveness have usually focused on physical traits of those being judged (Thornhill & Gangestad 1999). Information on the mate choice of conspecifics may also play an important role in determining preferences. Using the judgment of others may allow an individual to assess a potential mate quickly and efficiently. In nonhuman animal species, mate-choice copying has been observed among females of various species (Brown & Fawcett, 2005, Dugatkin, 2000, Galef & Laland, 2005, White, 2004). Most commonly studied are fish (Dugatkin & Godin, 1992, Dugatkin & Godin, 1993, Godin et al., 2005, Witte & Ryan, 2002) and bird species (Galef & White, 1998, Swaddle et al., 2005, White & Galef, 2000). Mate-choice copying has also been observed in males (Schlupp & Ryan, 1997, Witte & Ryan, 2002).

There are a greater number of studies conducted on female choice, and these studies have generally shown that when females observe another female paired with one of two males, they are subsequently more likely to prefer the paired male over the unpaired male. In a classic early example using stuffed dummies, Hoglund, Alatalo, Gibson and Lundberg (1995) showed that female grouse mated preferentially with the male that appeared to have other females in his territory. Such studies have generally been conducted in the laboratory, but the same effects have been found when experiments are conducted in the wild (e.g., Witte & Ryan 2002).

Several researchers have noted that mate-choice copying may influence human mate preferences (e.g., Brown & Fawcett, 2005, Dugatkin, 2000, Uller & Johansson, 2003), though as yet there is only limited evidence. Uller and Johansson (2003) found that the presence or absence of wedding rings on men, an indicator that the man has been chosen by another female, did not influence women's preference.

A wedding ring, however, is indicative only of partnership status and provides no information about partner value or partner attitude (Jones, DeBruine, Little, Burriss, & Feinberg, 2007). Jones et al. (2007) have shown that the valence of attitude is an important determinate of whether female attention increases male attractiveness. They found that observing other women with smiling (i.e., positive) expressions who were looking at male faces increased female observers' preferences for those men to a greater extent than did observing women with neutral (i.e., relatively negative) expressions looking at such faces. Women then do appear to mimic the attitude of other women to specific men.

Alongside valence of attitude, the traits of the observed “chooser” may also play a role in social transmission. For example, the mate choice of older (more experienced) female guppies is copied more than the mate choice of younger female guppies (Dugatkin & Godin 1993). If attractive individuals are better able to attract high-quality mates, the preferences of attractive individuals may be copied to a greater extent than unattractive individuals. Sigall and Landy (1973) have shown using real individuals that positive characteristics are attributed more frequently to men when paired with attractive than with unattractive women. Such a phenomenon is suggestive of mate-choice copying, whereby women use the quality of partner a man can himself attract to judge his attractiveness.

People are sensitive to mating context in preferences; for example, women prefer more masculine male faces for short- than for long-term relationships. The relationship between attractiveness and male facial masculinity is not clear-cut; some findings show attraction to masculinity (Cunningham et al., 1990, DeBruine et al., 2006, Grammer & Thornhill, 1994) and others show attraction to femininity (Berry & McArthur, 1985, Cunningham et al., 1990, Little & Hancock, 2002, Perrett et al., 1998, Rhodes et al., 2000). Human males bring two factors to a parenting relationship: a level of paternal investment and potential heritable benefits (e.g., genes for high-quality immune systems). The owners of masculine faces are perceived to have high dominance and lower levels of cooperation, pointing to lower paternal investment (Perrett et al., 1998). In the context of a short-term relationship, the perceived cues to high paternal investment in the feminine-faced male are of little value to a female. Females may therefore seek to maximize the genetic fitness of potential offspring if they are not extracting any other benefits from their mates. In long-term relationships, better parenting and increased cooperation may outweigh the benefits of genetic fitness, thereby enhancing the attractiveness of feminine-faced males. Indeed, studies have shown that masculinity in male faces is preferred more in short-term than in long-term contexts (Little et al., 2002, Scheib, 2001). For copying-like behavior, we might expect that copying will be more prominent in long-term decisions if it functions to guide choice of individuals who will make good investing partners. Alternatively, we would expect that copying would be more prominent in short-term decisions if it functions to lead women to choose more indirect, or genetic, benefits.

Although most of the research into mate-choice copying has focused on the behavior of females (e.g., Uller & Johansson 2003), males might also mate-choice copy. Indeed, in a species where males have been tested, sailfin mollies, males are found to follow the apparent preferences expressed by other males (Schlupp & Ryan, 1997, Witte & Ryan, 2002).

Here we examine the use of information about the attractiveness of paired same-sex images in attractiveness judgments of opposite-sex images for both long- and short-term judgments of attractiveness for both men and women.

We paired target faces with attractive and unattractive faces presented as their partners to examine if individuals were being influenced by this information. We then predicted a mate-choice copying-like effect whereby individuals would use the attractiveness of partners to infer information about the quality of the target and find faces more attractive if paired with an attractive partner than an unattractive partner. Many of the characteristics that individuals prefer in short-term mates are readily discernable, such as physical attractiveness (Buss & Schmitt 1993), and this is less true for characteristics preferred in long-term mates, where desired traits, such as intelligence (Buss & Schmitt 1993), are not usually open to rapid visual assessment. We therefore predicted a copying-like effect to occur more in long-term than short-term contexts. Target faces were also presented in attractive and unattractive forms to address whether target attractiveness influenced attractiveness here.

2. Methods 

2.2. Stimuli 

We used masculinity in male faces and femininity in female faces as attractive traits because a previous study using faces from this stimulus set revealed that, overall, women preferred masculine male faces and men preferred feminine female faces (Little, Cohen, Jones, & Belsky, 2007). All stimuli were constructed using established techniques (Little et al., 2001, Little et al., 2002, Penton-Voak et al., 1999, Perrett et al., 1998, Rowland & Perrett, 1995, Tiddeman et al., 2001) for manipulating the appearance of face images in an objective systematic manner. (For technical details including mathematical algorithms, see Rowland & Perrett, 1995, Tiddeman et al., 2001.)

Stimuli were composite images of male and female faces. Original images were 50 young adult male and 50 young adult female photographs taken under standard lighting conditions and with a neutral expression. Twenty-five images of each sex were randomly allocated to five sets per sex (five images in each set). From each set of images, a single composite face was produced, giving five male and five female composites. The composite faces were created using specially designed software. Key locations (174 points) were manually marked around the main features (e.g., points outline, eyes, nose, and mouth) and the outline of each face (e.g., jaw line, hair line). The average location of each point in the five faces in each set was then calculated. The features of the individual faces were then morphed to the relevant average shape before superimposing the images to produce a photographic quality result. All images were standardized on interpupillary distance and made perfectly symmetrical by averaging each image with its horizontally flipped version prior to transformation.

Each face was transformed on a sexual dimorphism dimension using the linear difference between a composite of all 50 males and all 50 females following the technique reported in Perrett et al. (1998). Using the shape difference between male and female, the vector of sexual dimorphism can be parameterized allowing manipulation along the vector, described here as a percentage of the distance between male and female. Transforms represented 50% plus or minus the difference between these two composites. Transformations resulted in two images for each base face: one a masculinized version and the other a feminized version. Examples of transformed images can be seen in Fig. 1.

Fig. 1. Example face pairs shown to male participants—feminine female/feminine male (A), feminine female/masculine male (B), masculine female/feminine male (C), masculine female/masculine male (A). Female participants saw the same pairs but with male faces on the left.

3. Results 

All tests are presented two tailed. ηp2 denotes partial Eta2.

3.1. Women 

A repeated measures analysis of variance (ANOVA) was carried out with “male masculinity” (masculine/feminine), “paired female masculinity” (masculine/feminine), and “term” (short/long) as within-participant factors. This revealed a significant effect of term (F1,50=18.92, p<.001, ηp2=0.989), a significant effect of target male masculinity (F1,50=13.56, p<.001, ηp2=0.951), and a significant interaction between paired female masculinity and term (F1,50=4.35, p=.042, ηp2=0.534). Other effects and interactions were not significant (all F1,50<2.19, p>.145, ηp2<0.306). Women rated the men as less attractive for short- than long-term decisions (short-term=2.12, long-term=2.46) and masculine men as more attractive than feminine men (masculine=2.41, feminine=2.18). The interaction reflected that women were only influenced by the pictured women for long-term decisions (described further below). Mean rating scores can be seen in Fig. 2. Rerunning the ANOVA with the order of rating (long-term first/short-term first) as a between-participants factor revealed no significant effect or interactions with the order (all F1,49<2.04, p>.159, ηp2<0.288).

Fig. 2. Ratings of attractiveness for the target faces by female (top) and male (bottom) judges. Scores are split by term and the masculinity of the opposite-sex face paired with it (feminine/masculine). The scale is the same between graphs, but the range changes.

Using only those responses made by women in the long-term condition, a repeated measures ANOVA was carried out with “male masculinity” (masculine/feminine) and “paired female masculinity” (masculine/feminine) as within-participant factors. This revealed a significant effect of target male masculinity (F1,50=6.07, p=.017, ηp2=0.676), a significant effect of paired female masculinity (F1,50=6.35, p=.015, ηp2=0.695), and no interaction between the two variables (F1,50=1.38, p=.246, ηp2=0.210). Masculine male faces were rated as more attractive than feminine male faces, and men paired with feminine female faces were rated more highly than men paired with masculine female faces.

For short-term decisions in women, a repeated measures ANOVA with factors defined as above revealed a significant effect of target male masculinity (F1,50=13.23, p=.001, ηp2=0.946), no significant effect of paired female masculinity (F1,50=0.38, p=.541, ηp2=0.093), and no interaction between the two variables (F1,50=0.71, p=.404, ηp2=0.131). Masculine male faces were rated as more attractive than feminine male faces, but men paired with feminine female faces were not rated more highly than men paired with masculine female faces.

4. Discussion 

The data presented here demonstrate that both men and women are influenced in their attractiveness judgments by the attractiveness of a target's partner for long-term decisions but not for short-term decisions. Observers moderate their preferences for specific individuals by using information about the attractiveness of that person's partner. The effect is not strictly copying in this case because all images are presented as partnered, but does show humans use information on others' choices in guiding their own judgments. Our data are then in line with previous studies of social transmission of preference (Jones et al., 2007). We found no significant effect of target attractiveness (whether the face paired with a fictitious partner was masculine or feminine) on the copying-like effect, suggesting that transmission of preference influences both attractive and unattractive targets equally. Although potentially, there could be contrast effects in judgments of attractiveness, with, for example, pairing of a feminine face to a target making the target appear more masculine, the finding that attractiveness judgments changed only for long-term decisions suggests this basic mechanism cannot account for the current data. The copying-like effects here depended on individuals following the choices of attractive over unattractive individuals. This is a more sophisticated form of copying than simply being attracted to those who have vs. those who do not have partners. Indeed, for humans, where only a few individuals remain unpartnered throughout their lives, copying of mate-choice that does not take account of partner quality appears unlikely to be a useful mechanism for identifying high-quality partners. Previous studies have shown that being partnered in the absence of other information is not sufficient to generate copying-like behavior (Uller & Johansson 2003). We did not find a sex-difference in the copying-like effect, and although men may have been expected to copy less as they are more influenced by physical attractiveness when choosing a partner, this finding is in line with studies demonstrating that both men and women highly value positive personality traits in long-term partners (Buss & Schmitt 1993).

Mate-choice copying has been proposed to be adaptive when there is a cost, such as energy, to evaluating the quality of potential mates or when discriminating between the quality of potential mates is difficult (Wade & Pruett-Jones 1990). In this way, social transmission may allow individuals to assess a potential mate quickly and efficiently, and perhaps teaches individuals what to look for in a mate. In humans, there are many aspects to a partner other than their physical traits, and potentially, the choices of others can be used to infer positive or negative traits, such as behavior, resources, or intelligence, which are difficult to infer just from physical appearance.

Specificity to long-term preferences implies that social influence is being used to determine nonphysical traits that make a target a good long-term partner. Studies have shown that individuals value physical attractiveness in short-term contexts over other attractive traits such as pleasant personalities (Buss & Schmitt 1993). Judges may then be able to acquire the physical information from a photograph to judge physical attractiveness for short-term contexts, and hence, the extra information from the paired partner is of little relevance. Humans bring two factors to a parenting relationship: a level of parental investment and potential heritable benefits (e.g., genes for high-quality immune systems). Social information may be more useful for judging the former given such information is less readily discernable. Of course, in species without parental care, mate-choice copying likely occurs because individuals are able to acquire information about the genetic quality of a prospective mate (Witte & Ryan 2002), and potentially, this is also true in humans despite our finding of specificity to long-term judgments.

It is important to note that there may be other issues here. We presented our targets as partnered, and individuals may, consciously or unconsciously, assume that when judging attractiveness for long-term relationships, the target would be leaving their current partner, whereas short-term judgments may imply targets would be cheating on their partner but remaining with them in the long-term. Judges may have taken into account potential retaliation of the current partner (though this is also an issue for the long-term condition) or that individuals with attractive partners may have less reason or desire to cheat (Gangestad, Thornhill, & Garver, 2005). Factors such as these may then also detract from using social information in short-term contexts.

An effect in humans influencing long- but not short-term decisions is consistent with recent findings in mice. Kavaliers et al. (2006) found that female mice are influenced by the smell of other females in guiding their interest in males, a finding suggestive of mate-choice copying, and that female mice with deletions of the oxytocin (OT) gene do not show this effect. This finding suggests that OT appears to mediate mate-choice copying-like effects, at least in mice. Given OT's role in pair bonding (Ferguson et al., 2001, Pedersen & Boccia, 2002), it is possible that mice with intact OT genes were more interested in pair bonding and show preferences more akin to a long-term mating decision for humans.

Models of mate-choice have suggested that social transmission of mate preferences can contribute to sexual selection for male traits (Brown & Fawcett, 2005, Galef & Laland, 2005, Kirkpatrick & Dugatkin, 1994, Laland, 1994). Potentially, our study adds a caveat to some assumptions in these models—individuals may select their long-term partners based on copying-like behavior, but because their short-term decisions remain less affected, the potential for copying to lead to rapid spread in certain genes may be more complicated.

In summary, the studies presented here extend work on the social transmission of mate preference in humans, demonstrating that both men and women use information about the attractiveness of a partner to inform their long-term but not their short-term preferences. Although our experiment is somewhat unnatural, we note that face preferences have been found to correlate with perceived partner characteristics (DeBruine et al., 2006), which suggests that findings from preference studies might reflect choices outside of the laboratory. An effect akin to mate-choice copying in humans may then partly determine selection of long-term partners, whereas other physical factors determine short-term choices.