Nonindependent Mate Choice

In the early 1990’s, biologists began to document cases of nonindependent mate choice (also termed “mate choice copying”) in several species of animal, including guppies and sage grouse [1]. In these species, females preferentially select males that are seen in proximity with other females (and thus have likely mated with those females). Two reasons are generally offered to explain this phenomenon: cost avoidance and improved discrimination. The first posits that females use public mating information to select a mate without bearing the costs of active mate search and choice. The second posits that females use public information, augmenting their own assessments with the assessments of other females, to improve the quality of their mate selection [4].

Vakirtzis and Roberts [4], [5], [6] added important theoretical refinements to this literature. They noted that mate copying, or simply imitating the decision of any female in the population, would be unlikely to arise in a relatively monogamous species; rather, what would arise is a form of nonindependent mate choice that they termed “mate quality bias.” Mate quality bias is a subset of nonindependent mate choice where females assess the quality of a male’s mate and use this information in making their own mate choice decisions. In other words, assortative mating results in a reliable correlation between the quality of a male and the quality of his partner, and this correlation can be used to enhance mate choice discrimination. Because humans are relatively monogamous, Vakirtzis and Roberts proposed they should exhibit mate quality bias, not mate choice copying, i.e., women will be influenced by the quality of a man’s partner not simply the fact that he has a partner [4], [5].

Modern studies, using various methods, indicate that a man and/or woman’s putative partner affects observers perceptions of his or her traits [9], [10], [11]. Sigall and Landy [7], for example, found that both men and women rated a man who was associated with an attractive, opposite sex confederate more favorably than when he was not associated with an attractive confederate. Observers rated a man as least favorable when he was associated with an unattractive confederate. Strane and Watts [8] found that women who were paired with attractive men in photographs were rated more favorably than women who were paired with unattractive men. Waynforth [12] and Vakirzis and Roberts [13] explicitly demonstrated that the variable of importance in these cases is the relative attractiveness of the putative partner. That is, observers rate a male more favorably if he is with or desired by an attractive putative partner than if he is with or desired an unattractive putative partner, supporting Vakirtzis and Roberts’ contention that humans should exhibit mate quality bias, not indiscriminate mate choice copying.

Although mate quality bias better explains the existing data, it fails to integrate such effects into a broader, more parsimonious theory. We believe that applying signaling theory to these data has the potential to provide this integration.

Signaling and Social Information

Signaling theory has been successfully used to understand the evolution and expression of costly physical and behavioral traits, such as the elaborate plumages of peacocks, the beautifully wrought bowers of bowerbirds, and the sprightly stotting of gazelles [14], [15]. The basic tenets of signaling theory, as applied in evolutionary biology, are straightforward. Trait quality varies among individuals of all species. These traits are not always manifested or easily perceivable (e.g., dominance, intelligence, immunocompetence) but can, in principle, be reliably signaled by other traits. Thus perceivers benefit by attending and reacting to the signals. However, signalers and perceivers often have competing interests. The signalers can potentially exploit the perceivers by enhancing their signals without changing underlying traits; therefore, perceivers must remain vigilant against deceptive signals. A solution to this potential arms race is the development of signals that are difficult to fake and therefore reliable indicators of the underlying traits [16]. This, in turn, explains why many animal signals, especially ones sent between animals with competing interests, are extravagant: such signals are costly and thus cheating is difficult. However, there are other mechanisms and constraints that help ensure honesty, and it has become clear in recent years that Zahavi’s contention that signals require exorbitant costs is incorrect [17], [18]. Cost, in other words, is only one mechanism that deters deceptive signals.

Although humans have developed the most abstract and sophisticated signaling space among extant species, theorists have argued that the basic principles are the same. These theorists, influenced by Veblen’s [19] early work on conspicuous consumption, have used those principles to illuminate otherwise puzzling aspects of human social behavior. Miller [20], for example, argues that one reason the human brain is so unique is because it is a signal sending organ, and that many of the cultural products that it produces–sublime sculptures, gripping novels, eloquent poems–are signals of the producer’s underlying mate quality. Other scholars, applying signaling theory to archaeological data from the Great Basin, have argued that men shifted to large game hunting because they were competing to send reliable signals of their underlying qualities, such as strength, intelligence, and generosity [21]. More recently, Miller [22] extended costly signaling logic to consumerism, arguing that consumer products are used to advertise underlying traits. Although there is debate over the precise signaling functions of consumer products, most researchers agree that prestige goods such as Porsches, Rolexes, or rare and aged scotches, are reliable signals of status and wealth and perhaps other traits [22], [23] because obtaining them is difficult for lower status individuals. Consistent with the proposal that they function as honest signals of wealth, status, and other desirable traits, once a prestige good is easily copied or accessible, its value declines and it is replaced by another prestige good or an elaboration of the old good [24]. Whether such prestige goods function like biological signals is controversial, and researchers are skeptical that biological signals and prestige good signals adhere to the same underlying logic [25]. Cronk [26] noted that it is times more useful to use the term “hard-to-fake” signal than “costly” signal to avoid the possible pitfalls of misleading terminology. What is important for the purposes of this article is that mates function as an honest signal of a person’s underlying traits.

In what follows, we build upon these general principles (with noted caveats) and extend them to nonindependent mate choice by asserting that mates can function, in part, as social signals; that is, mates, like other luxury goods, convey social information about their partners’ traits or qualities that are not readily observable.

Social Signaling and Nonindependent Mate Choice

Humans are motivated to display signs of social and cultural status [22], [27], [28], not only to obtain desirable mates but also to gain social influence and through this access to other types of resources. One fundamental, but often overlooked, signal of a person’s status is his or her mate, because highly desirable mates are scarce, coveted, difficult to obtain, and require more investment than less desirable mates [29]; or, put more generally, desirable mates are honest indicators of their partners’ positive traits, including status, wealth, and, arguably, genetic quality. If so, the principles of signaling theory should generalize to mating, and humans should either flaunt or conceal their partners depending upon the partners’ relative mate value–a prediction Vakirtzis and Roberts [5] derived from the mate quality bias perspective. By “flaunt” we mean actively display, show off, or boast about a mate and by “conceal” we mean actively hide or remain silent about a mate. Although this signaling system is predicted to be reliable, there is a potential for deception. Perceivers should be motivated to determine the nature of an ambiguous relationship, and thus detect “cheats,” using social information such as gossip, and honest displayers should desire to make their signals unambiguous by using socially accepted advertisements of partnership (e.g., holding hands, kissing in public, exchanging jewelry, facebook status) [30]. Perceivers evaluating the relationship should also attend to disparities in affection between partners. Other things equal, individuals of high mate value or social status are predicted to expend less effort on winning and sustaining the affections of an attractive mate; therefore, perceivers should impute more status and mate value to flaunters who appear less affectionate than their partners, a fact that is often parlayed to enhance (or make explicit) the status and mate value of fictional characters like James Bond or Scarlett O’hara.

The extension of signaling theory to mates straightforwardly predicts that the quality of a person’s mate will affect the perceptions and decisions of other potential mates (nonindependent mate choice), consistent with existing data [3], [7]. The extension is also consistent with a number of hypotheses that are not readily deducible from existing mate quality bias literature. Two of the most important and easily testable predictions are that both men and women: 1) will flaunt (tested here) [31], and 2) flaunt to same sex as well as opposite sex peers (a prediction we have recently confirmed, unpublished data); flaunting to same sex individuals is predicted to enhance social influence. On the basis of signaling theory, both sexes possess traits that contribute to their social influence and mate value but are not easily perceivable [20]; therefore, both men and women should publicly advertise hard-to-fake perceptible signals (including mates) that broadcast possession of these hidden traits. As noted, previous research shows that both males and females who are paired with attractive partners are more positively evaluated by both same and opposite sex observers [7], [8]. Although the proposed extension of signaling theory, consistent with other research [32], predicts that increased (or primed) mating motives may augment an individual’s propensity to flaunt, it also predicts that an individual will flaunt or conceal regardless of his or her relationship status. This for two reasons: First, humans seldom remain in a single romantic relationship for life and, therefore, they should remain concerned with their status and mate value even when in a secure relationship. And, second, the status that accrues to humans who flaunt valued mates can be used to increase fitness through enhanced social influence generally, independently of securing other (presumably higher valued) mates. The most direct way is through passing acquired resources and status to offspring [33].

The Current Study

We derived and tested five hypotheses from our extension of signaling theory. First, men will flaunt and conceal. Second, women will flaunt and conceal. Third, men will flaunt and conceal a partner’s attractiveness more than women. The latter prediction is based on the design of our study. In it, we manipulated only the attractiveness of a participant’s putative partner. Men’s mate choices are more strongly influenced by physical attractiveness than women’s choices [28] and thus physically attractive mates have higher signal value for males than females, therefore this manipulation should result in a stronger motivation to flaunt or conceal the partner in men than women. Fourth, flaunting and concealing will not be moderated by the relationship status of the participants. Because this is a null prediction, it is not strong and the results should be interpreted with caution. Fifth, the relationship between partner attractiveness and flaunting/concealing will be mediated by expected impact on the flaunters’ status and desirability. This final prediction stems from our proposal that flaunters signal to increase or broadcast their status and mate value; we used “desirability” as a substitute for “mate value,” which is unclear to participants. Although we do not believe that flaunters or concealers need to be conscious of why they are flaunting or concealing, we do believe that, as with the display of other status signals, they are conscious of the consequences [34].

Ethics Statement

All participants completed written informed consent forms before participating in the study. Consent forms were collected and stored in a locked cabinet. After completing the study, participants were first orally debriefed and then given a written debriefing form documenting the nature of our experiment. This study and the consent and debriefing procedure were reviewed and approved by the University of Missouri Institutional Review Board.

Participants

Undergraduates at the University of Missouri participated for partial course credit. The sample consisted of 64 men (mean age = 20. 27 SD = 4.86) and 75 women (mean age = 19.10, SD = 1.87). Of these, 21 men and 28 women reported being in a relationship [Men mean relationship satisfaction = 4.23, SD = 1.26; women mean = 4.28, SD = 1.18; see Supplementary Information S1).

Design and Procedure

Participants were given pamphlets (see Pamphlet 1 in Supplementary Information S3 and Pamphlet 2 in Supplementary Information S4) dealing with higher education surveys. They were told that they would collect survey data about attitudes toward higher education and educational funding from one of two areas on campus. Participants were told we were interested in response rates and how the characteristics of the surveyors affected respondent’s answers. Participants in the experimental condition were told they would be collecting their data with a partner and were to act as if they and their assigned partner were in a happy romantic relationship (participants were told this had been shown to affect response rates). After completing basic demographic questions, participants filled out the Ten Item Personality Measure [35]. This measure has low reliability and therefore was included as a filler (rather than being used in subsequent analyses) so that participants would not become suspicious about the veracity of our cover story.

Participants were then told that pictures of their partners had been collected so they could be identified when they met to collect survey data. To ensure realism, participants were told that participants who did not want their pictures shown were removed from the study and given credit for participating in a separate study. Participants then turned the page of the pamphlet and looked at the approximately three by three inch black and white photographs of their putative partners; each participant had only one randomly assigned putative partner. The photographs were taken from a pre-rated (raters were eight men and eight women) set of 247 photographs (117 men and 130 women); in this larger set, photographs that did not appear natural (e.g. were idiosyncratically lit) were removed. From that pre-rated set, we chose three women and three men from the top decile of attractiveness and three women and three men from the bottom decile, ensuring that the attractive putative partners were very attractive and that the unattractive putative partners were very unattractive (see Supplementary Information S2).

Underneath their picture, each putative partner had a name and a schedule of open times during which they could meet in the following two weeks to conduct the surveys. Participants were asked to mark the number of hours they were available to meet. Next, participants read about two locations that they could choose to conduct the surveys. One location was described as full of undergraduate students and one was described as full of administrators (see pamphlet S3). Participants were then asked to pick a location preference on a 1 (strongly prefer the administrative location) to 7 (strongly prefer the undergraduate location) Likert scale. Finally, participants filled out six questions on a four point scale about how they would feel as they collected the survey data and seven questions on a five point scale about how they thought they would be perceived as they collected the survey data (e.g. “Other people will view me as having status.”; 1 = strongly disagree and 5 = strongly agree). After participants completed the pamphlet, participants were told that the experiment was over and were orally debriefed (see debriefing in S1). The control group was run in a similar manner. They received pamphlets and were told that they would administer surveys in one of two locations (see pamphlet S4). They were then told that participants would give out the surveys with partners, but that they would be alone; therefore, their pamphlets did not have pictures of putative partners. They then filled out the same questions as the experimental groups, were told that the experiment was over, and were orally debriefed.

Operationalization of Flaunting and Concealing

Because the undergraduate location consists of peers whose judgments are likely to be salient and relevant to the participants, it was considered the flaunting location, and the administrative location, because it consists of people outside of the participants’ relevant peer group, was considered the concealing location [36]. Flaunting is operationalized as a significant preference for the undergraduate location compared to the preference of the same-sex control group; and concealing is operationalized as a significant preference for the administrative location compared to the preference of the same-sex control group. Because we were also interested in sex differences in flaunting/concealing, we first conducted a 2 (sex) by 3 (partner attractiveness: attractive, unattractive, and control) ANOVA to test whether location preference differed across conditions and to test the sex by condition interaction. Planned comparisons using the Mean Square Residual of the ANOVA were then conducted to investigate each specific hypothesis separately for each sex (e.g., for the hypothesis “men will flaunt,” we compared men in the attractive condition to men in the control condition) and across each sex in the flaunting and concealing conditions (e.g., for the hypothesis “men will flaunt to a greater degree than women,” we compared men in the attractive condition to women in the attractive condition). Because these tests were explicitly based on a priori hypotheses, we did not control for multiple comparisons.