THE FORMATION OF STATUS HIERARCHIES IN LEADERLESS GROUPS The Role of Male Waist-to-Hip Ratio

Lorne C a m p b e l l

Simon Fraser University J e f f r y A. S i m p s o n , M a r k Stewart, and J o h n G. M a n n i n g

Texas A&M University

Two studies examined the link between social dominance and male waistto-hip ratio (WHR). Groups of four men interacted in a leaderless group discussion. In both studies, men with higher WHRs (associated with current and long-term health status) were rated by other group members as behaving more leader-like when an observer was present, and rated themselves as being more assertive. In Study 2, men with higher WHRs were rated by independent observers as behaving more dominantly, but only when the evaluator was present. These results are discussed in terms of evolutionary models of health, attraction, and intrasexual competition. KEYWORDS: Evolutionary psychology; Intrasexual competition; Leadership; Social dominance; Waist-to-hip ratio

Social d o m i n a n c e hierarchies form quickly in m a n y species, particularly a m o n g males (de Waal 1982; Ellis 1995; Fisek and Ofshe 1970). Moreover, p e o p l e can ascertain their future social status within groups quite accurately following even brief interactions (Kalma 1991). Although status striving is p r e s u m e d to be a basic h u m a n m o t i v e (Barkow 1989; Maslow 1937; S y m o n s 1979), r e m a r k a b l y little is k n o w n about w h a t kinds of personal attributes are associated with the display of d o m i n a n c e and the Received April 9, 2001; accepted June 1, 2001

Address all correspondence to Dr. Lorne Campbell, Department of Psychology, University of Western Ontario, London, Ontario, Canada, N6A 5C2. Email: [email protected] Copyright 2002by Walter de Gruyter, Inc., New York Human Nature, Vol. 13, No. 3, pp. 345-362. 345

1045-6767/01/$1.00+.10

346

Human Nature, Vol. 13, No. 3, 2002

emergence of leadership in small groups. Guided by principles of "good genes" sexual selection, the present research examined whether a heritable marker of current and long-term health--waist-to-hip ratio (WHR)-is systematically related to social dominance and emergent leadership in groups of men.

THEORIES OF INTRASEXUAL COMPETITION Intrasexual competition occurs when members of one sex compete for status among themselves or vie for the attention of the other sex. It is influenced by two processes (Darwin 1871): intrasexual selection (in which members of one sex directly or indirectly compete for mates) and intersexual selection (in which members of one sex exert "choice" by selecting mates with certain desirable attributes). Recent extensions of Trivers's (1972) sexual selection theory (e.g., sexual strategies theory: Buss and Schmitt 1993) propose that women should be attracted to men who display the motivation and ability to acquire resources that, in ancestral environments, might have been necessary to raise offspring. Other evolutionary models propose that women should be attracted to socially dominant men, either to obtain immediate resources from them (Hrdy 1997) or to pass on the "good genes" of these men to their offspring (Gangestad and Simpson 2000). Virtually all of these models suggest that socially dominant men should have attracted more mates through successful intrasexual competition, and empirical data support this premise (see Sadalla et al. 1987; P6russe 1993). 1

SOCIAL D O M I N A N C E A N D STATUS HIERARCHIES Relatively little is known about how status hierarchies form in small groups. Most previous research has simply asked people to list the behaviors and tactics they use to compete with others for mates (see, for example, Buss 1981, 1988; Schmitt and Buss 1996; Waiters and Crawford 1994). Only a handful of behavioral observation studies have documented what people actually do when competing with same-sex rivals (e.g., Simpson et al. 1999). This is a significant empirical void considering the importance of group interaction in our cultural and evolutionary heritage (Caporeal and Baron 1997; Tooby and Cosmides 1992). We also know little about how rivals perceive each other's behavior and qualities in relation to their o w n during social interactions. Organisms should have evolved to evaluate and compare the qualities of rivals against their own attributes (Dawkins 1989). This process should facilitate the development of group-based social hierarchies, allowing group mere-

WHR and Social Dominance

347

bers to avoid the high costs of continual, futile, or dangerous confrontations with healthier, more dominant, or higher status rivals.

G O O D GENES, HEALTH, AND STATUS STRIVING Several personal attributes might be associated with success at ascending dominance hierarchies. "Good genes" models of sexual selection (see Gangestad and Thornhill 1997; Gangestad and Simpson 2000) suggest that an individual's degree of genetic fitness might be inferred from specific phenotypic markers. Individuals who have these markers should, on average, be more successful in intrasexual competitions because these traits tend to be valued by members of the opposite sex (Zahavi 1975). Such individuals should also be more willing to compare themselves directly with same-sex rivals to "advertise" their valued traits to both rivals (with w h o m they compete for status) and desirable mates (Andersson 1994; Miller 2000). Such comparisons might communicate important information about an individual's health, genetic fitness, and ability to defend resources or protect mates (Gangestad and Simpson 2000). We are not suggesting that these traits function as visual "cues"; rather, they should be associated with certain traits or behaviors that honestly "advertise" (signal) greater genetic viability.

INDICATORS OF HEALTH AND VIABILITY One marker of health and underlying fitness in humans should be the pattern of body fat distribution (Singh 1993, 1995). The distribution of body fat (rather than the total amount of fat) determines body shape. Higher levels of testosterone stimulate fat deposits in the abdominal region and inhibit fat deposits in the gluteofemoral region (i.e., the buttocks and thighs; Bjorntorp 1987, 1988, 1991a, 1991b; Rebuffe-Scrive 1987a, 1987b, 1988, 1991), resulting in an android fat distribution (Rebuffe-Scrive et al. 1991). Body fat distribution is measured by computing a ratio (the waistto-hip ratio; WHR) between the circumference of the waist and the circumference of the hips. WHR is a reliable measure of the distribution of fat (Singh 1993). The typical range of WHR in young men is .80-.95 (Marti et al. 1991). Large inter-individual differences in WHR exist even after statistically controlling for total body fat (Bouchard 1988). Several lines of evidence suggest that WHR might be one indicator of genetic viability. First, WHR is a testosterone-dependent secondary sexual characteristic, and higher levels of testosterone may signal greater immunological competence. According to Andersson, "testosterone level

348

Human Nature, Vol. 13, No. 3, 2002

seems to act as a link relating the expression of secondary sex traits to male condition" (1994:70). Males with properly developed secondary sexual traits are more likely to be in good health and have less disease or parasites. In fact, men with higher levels of testosterone have a smaller ratio between the second to fourth digit length in both hands (Manning et al. 1998). This ratio is fixed b y 2 years of age. Smaller ratios are related to higher sperm count and reduced germ cell failure in men, and higher ratios are also predictive of immune dysfunction (Manning and Bundred 2000). Higher WHRs in young men, therefore, may signify better immunological competence and current health status. Second, independent of body mass, WHR predicts a variety of health risk factors, Including the incidence of diabetes, heart attack, stroke, cardiovascular disease, hypertension, poor liver functioning, and premature mortality (Bjorntorp 1988, 1991b; Leibel et al. 1989). Men with lower WHRs (those closer to the typical female range) are perceived as more feminine and less healthy (Singh 1995). The ideal W H R for men--the ratio associated with the best long-term health outcomes--is .90-.95, and men with WHRs in this "desirable" range are judged as more attractive and more leader-like (Singh 1995). WHR, therefore, may also be an honest marker of current and long-term health status. According to good-genes models, men with WHRs in the desirable range (.90-.95) should have fared well when competing for mates during evolutionary history. The W H R conveys valid information about an individual's current and long-term health and contains approximately 25% additive genetic variance (Bouchard 1988; Bouchard et al. 1990; Donahue, Prineas, Gomez, and Hong 1990), suggesting that the health and reproductive advantages of mates with desirable WHRs could have been genetically transmitted to offspring. During evolutionary history, therefore, women may have preferred men with more "desirable" WHRs (intersexual selection), and men with more "desirable" WHRs may have outcompeted their rivals for mates (intrasexual selection). Accordingly, men with more desirable WHRs should, relative to other men, display behavioral tactics that "advertise" their superior attributes in newly formed, unstructured groups, exhibiting greater leadership and social dominance. Only a few studies have explored how men's W H R might be implicated in mating. Singh (1995) had people rate line drawings depicting men with different WHRs and body sizes. Men with WHRs near the desirable range (.90) were consistently rated as the healthiest and most attractive mates. They were also rated as being more intelligent and having better leadership qualities, attributes that should be associated with more successful intrasexual competition. Men with markedly lower or higher WHRs (lower than .90 or higher than .95), in contrast, were rated as less healthy, less attractive, and as having less-desirable personality characteristics. This basic

WHR and Social Dominance

349

pattern of results has recently been replicated (Furnham et al. 1997). Although these studies suggest systematic links between WHR and dominance/leadership in groups, they have some limitations. First, ratings were made of two-dimensional line drawings rather than actual people. Second, these studies did not examine the behavioral tactics that men with different WHRs actually use to attract persons of the opposite sex. If W H R is an honest indicator of health/viability, it is important to determine whether people with more "desirable" WHRs actually display behaviors indicative of greater leadership and whether rivals actually recognize their elevated status in intrasexual competitions.

THE PRESENT RESEARCH The present research was conducted to test whether men with more "desirable" WHRs behave in a more leader-like, socially dominant fashion when interacting with other men in a group context. Heterosexual men were randomly assigned to four-person, all-male groups and asked to engage in an unstructured, emergent leadership discussion task. Following the discussion, each individual's level of social dominance and leader-like behavior was assessed with both self-report and peer-report measures.

STUDY 1 Study I involved peer-ratings and self-ratings of emergent leadership in 17 four-person groups of men. For each group, a male observer sat in the room and "evaluated" (but did not participate in) the group discussion to make salient the fact that each individual's behavior was being compared to the behavior of other members in the group.

Participants Sixty-eight male undergraduates enrolled in Introductory Psychology at Texas A&M University participated in the study in exchange for partial fulfillment of a course requirement.

Procedure Participants arrived at the laboratory in groups of 5-10 for preliminary assessment. To calculate WHR, the measure of each participant's waist was divided by the measure of his hips. Two research assistants measured each participant. The two measures were highly correlated (r = .97, p < .001), indicating that W H R was reliably assessed. The mean W H R

Phase 1.

350

Human Nature, Vol. 13, No. 3, 2002

in the sample was .82 (s.d. = .04). 2 This distribution is typical of WHRs in young men (Marti et al. 1991). Each participant's height and weight were also measured. Participants then completed questionnaires that inquired about any prior training or experiences leading groups. For exploratory purposes, participants also completed a measure of the Big Five personality traits (the Berkley Personality Inventory; Harary and Donahue 1994), a measure of shyness (Cheek and Buss 1981), and a measure of self-esteem (Rosenberg 1965). After Phase 1, each participant was randomly assigned to a four-person discussion group that was scheduled to meet the following week. Participants did not know anyone in their group. Once placed in a group, they were randomly assigned a letter from A to D.

Phase 2. Upon arrival, participants were seated in desks in a semicircle in a large room. They were seated according to their randomly assigned letter, with people having the letter A sitting in the first seat, B in the second seat, and so forth. Each participant was given a set of instructions, which the experimenter read and then explained. Each group was asked to discuss a list of fifteen teacher attributes (provided by the experimenters), remove three attributes, and then rank (as a group) the remaining twelve attributes in order of importance in terms of assessing teacher effectiveness. Groups were also asked to generate three additional teacher attributes that were not on the original list. Agreement from each group member was required to remove an item, assign a group-based rank to an item, or add a new item to the list. Participants were led to believe that their group responses would be used by the University to improve teaching evaluations. Participants were also told a person (a male) would observe and privately evaluate their group discussion. The observer took notes during the discussion but did not participate. Each discussion lasted 30-35 minutes. The format of the group interaction was a leaderless group discussion (LGD) in which no one was given an assigned role. The LGD requires participants to provide structure to an ambiguous situation, solve problems in a limited amount of time, and earn esteem from other group members (Bass 1954). It is a reliable and valid indicator of managerial job performance (Cascio 1987; Thornton and Cleveland 1990) and is ideally suited to observing and assessing social dominance and leadership emergence. Following the group discussion, individuals rated (a) each group member's performance during the discussion task and (b) his own task performance on several behaviors diagnostic of leadership. Participants completed the following measures: 1. Peer Ratings of Group Performance. Participants first rated each group member's leadership behaviors on 7-point scales (anchored 1 = not at all; 7 = very much/extremely). The seven rated items were: assertive,

WHR and Social Dominance

351

persuasive, intelligent, innovative, helpful, originator of ideas, and a motivator of others. These items were averaged to form a peer-rated leadership index (~ = .60). Higher scores indicated the display of greater leadership behavior in the group. 2. Self-Ratings. Participants next rated themselves on the seven items listed above in addition to three questions that assessed how motivated they were to lead the group and voice their opinions. Responses were made on 7-point scales (anchored 1 = not at all; 7 = very much/extremely). A se~:-reported assertiveness index was then calculated by averaging all 10 items ((z = .86). Higher scores indicated greater motivation to lead and direct the group. Results and Discussion

W H R was not significantly correlated with the self-report measures of prior leadership training or experience. Therefore, any relations between W H R and leadership behavior in the group discussions are not due to differences in leadership experience or training in leadership effectiveness. The Social Relations Model (SRM: Kenny and La Voie 1984) was used to analyze the round-robin peer-ratings on the leadership index. According to the SRM, an individual's perception of another person can be partitioned into three components: a perceiver effect (how people view group members in general), a target effect (how a person is generally perceived by other group members), and a unique dyadic (relationship) effect. The present study focused on the correlation of individuals' WHR with the target effect of leadership: that is, do group members agree about who was more versus less leader-like in their group, and is this agreement correlated with the target's WHR? Target variance accounted for 27% of the total variance in peer leadership ratings, t (16) = 3.28, p < .01, indicating that group members tended to agree on how leader-like each group member was during the discussion. W H R was positively and significantly correlated with this target variance, r = .37, t (50) = 2.24, p < .05, Cohen's d = .79. Thus, relative to other group members, men with higher WHRs (in this sample, men with more "desirable" WHRs) were perceived by other group members as behaving more leader-like during the group discussion. The SRM is not well suited to testing the relation between W H R and self-reported ratings of assertiveness. Thus, the self-reported assertiveness index was regressed on WHR. Discussion group was entered as a class variable in the analysis to (a) control for the statistical interdependence that exists within each group and (b) adjust the degrees of freedom to match the number of groups (not individuals) in the sample. A significant relationship emerged between W H R and self-reported assertiveness: b = .35, F (1,16) = 10.18, p < .01, Cohen's d = .80. Specifically, individuals with

352

Human Nature, Vol. 13, No. 3, 2002

higher WHRs rated themselves as being more assertive during the group discussion than did men with lower WHRs. The association between WHR and self-reported assertiveness remained significant even when individuals' Body Mass Index (i.e., their weight in kilograms divided by their height in squared meters) and their scores on the Big Five personality traits, shyness, and self-esteem were statistically controlled. In sum, the prediction that men with higher (more "desirable") WHRs would behave in a more socially dominant and leader-like fashion during the group discussions was supported, even after controlling for several potential confounds. Two sources of data supported this hypothesis. First, peer-ratings of leadership indicated that other group members perceived the behavior of men with higher WHRs as more leader-like. Second, individuals with higher WHRs rated themselves as having stronger motivation to lead and direct their groups.

STUDY 2 Study 2 was a conceptual replication and extension of Study 1. It differed from Study I in three ways. First, independent observers rated the emergent leadership behavior of each man in each discussion group. Second, a female "evaluator" observed (but once again did not participant in) each group so we could test whether the relation between W H R and leadership/dominance was contingent on the gender of the evaluator. Third, a control condition was added in which no observer was present. We included this condition in Study 2 because men with higher WHRs should be more motivated to ascend social hierarchies when it is clear that their performance relative to others is being observed and evaluated (i.e., when an "evaluator" is present). Accordingly, we predicted that the link between WHR and social dominance/leadership should be context dependent (i.e., it should be stronger when an observer is present versus not present).

Participants One-hundred sixty male undergraduates enrolled in Introductory Psychology at Texas A&M University participated in the study in exchange for partial fulfillment of a course requirement.

Procedure The procedures for Study 2 were very similar to those for Study 1. However, in Study 2, 22 groups were randomly assigned to the observer con-

WHR and Social Dominance

353

dition and were told that a person (a woman) would observe and evaluate their group discussion. Eighteen groups were randomly assigned to a noobserver condition and proceeded directly to the group discussion task. The group discussions were unobtrusively videotaped. The mean WHR in the sample was .84 (s.d. = .05).

Peer and Self-Ratings of Group Performance.

After the discussion task, participants rated each other on the same seven items described in Study 1, and a peer-rated leadership index was calculated by averaging the scores from each item ((x = .72). In addition, participants rated their own assertiveness during the group discussion on a 17-item scale that assessed how motivated they were to lead the group, present their ideas, and direct the group task. Responses were made on 7-point scales (1 = not at all; 7 = very much/extremely). A self-reported assertiveness index was then calculated by averaging all 17 items (~t = .86). Higher scores indicate greater motivation to lead and direct the group.

Observer Ratings. Two sets of independent raters viewed the videotaped interactions of each group discussion and rated each participant's leadership behavior. All ratings were made on 7-point scales (1 = not at all; 7 = very much/extremely). Raters were blind to the experimental hypotheses and conditions because the observer ("evaluator") was not visible on the videotape. One set of 10 coders rated the same behaviors on which group members rated each other (see above). Rater reliability was good (mean ct = .77, range = .71-.80). Thus, raters' scores were averaged for each participant on each item, and a rated leadership index was calculated for each participant by averaging the items (a = .98). Higher scores indicated higher rated leadership. A second set of nine coders rated verbal and nonverbal behaviors reflecting interpersonal and social dominance (Argyle 1994). Rater reliability was good for each rated item (mean ~t = .80, range = .70-.88). The items that formed this factor were: "How much did this person interrupt others while they were talking? .... How much did this person maintain eye contact with others while he was talking? .... To what degree did this person use hand gestures while speaking? .... How much was this person looked at by other group members while he talked?' "How much did this person talk compared to others in the group? .... How much attention did this person receive compared to other group members? .... How much respect did group members give this person compared to other group members?" "How interested were other group members in what this person had to say?" and "How much did this person 'get his way' compared to others in the group?" (i.e., how much were his ideas accepted relative to other group members?). Rater's scores were averaged for each item, and a single score was calculated for each participant by averaging the items.

354

Human Nature, Vol. 13, No. 3, 2002

Higher scores reflect higher rated dominance. The scale has high internal consistency ((x = .96).

Results and Discussion Means and standard deviations for self-reported assertiveness a n d the two observer-rated factors (Leadership and Dominance) are presented in Table 1. Scores for the assertiveness scale and both rated factors were rescaled to range from I to 7. N o n e of the means was significantly different between the two experimental conditions. Also, as in Study 1, W H R was not significantly correlated with the self-report measures of prior leadership training or experience. The SRM was again used to analyze the interpersonal perception data. Target variance accounted for 44% of the total variance in the observerpresent condition, t (21) = 4.89, p < .001, as well as 35% of the variance in the observer-absent condition, t (17) = 5.20, p < .001. A marginally significant correlation between W H R with this target variance was found in the observer-present condition, r =.27, t (65) = 1.92, p = .07, Cohen's d = .56, but not in the observer-absent condition, r = -.07, n.s. These correlations were marginally significantly different from each other, t (38) = 1.91, p = .07, Cohen's d = .30. The peer-ratings from Study I were then combined with the peer-ratings of the groups that had an observer present in Study 2 to increase the power of the test comparing the link between W H R and peerratings of leadership in the observer-present versus -absent conditions. Overall, people rated m e n with higher WHRs as more leader-like in the observer-present condition than in the observer-absent condition, t (55) = 2.08, p < .05, Cohen's d = .33. The correlations between self-reported assertiveness and the two observer-rated indexes are s h o w n in Table 2. All three measures were positively and highly correlated. Therefore, each index was standardized and aggregated to form a single social dominance measure. Individuals' social

Table 1. Means and Standard Deviations for Self-Reported Assertiveness and Observer-Rated Factors Variable

Self-Reported Assertiveness Observer-Rated Leadership Observer-Rated Dominance

Evaluator-Present Mean (s.d.)

Evaluator-Absent Mean (s.d.)

4.72 (.81) 3.90 (.76) 3.95 (.71)

4.66 (.90) 4.04 (.67) 4.08 (.66)

t .34 -1.29 -1.26

355

WHR and Social Dominance Table 2. Correlations between Self-Reported Assertiveness and Observer-Rated Factors Variable

1. Self-Reported Assertiveness 2. Observer-Rated Leadership 3. Observer-Rated Dominance

1

2

3

-.35* .46*

.48* -.75*

.43* .84" --

N o t e : Correlations below the diagonal are for groups in which the evaluator was not present; those above the diagonal are for groups in which the evaluator was present. * p < .01.

d o m i n a n c e scores w e r e then regressed on their W H R s , e x p e r i m e n t a l condition (which w a s d u m m y coded), a n d the interaction of W H R a n d condition. Discussion g r o u p w a s treated as a class variable in this analysis. The interaction w a s significant, F (1, 37) = 8.86, p < .01, C o h e n ' s d = .47. The s i m p l e slope of W H R on social d o m i n a n c e w a s significant in the observerp r e s e n t condition, b = .90, t (21) = 2.95, p < .01, b u t not in the observerabsent condition, b = - . 4 2 , t (17) = -1.69, n.s. (Figure 1). Once again, these results r e m a i n e d significant w h e n controlling for BMI, the Big Five personality traits, shyness, a n d self-esteem. In s u m , m e n w i t h higher (more "desirable") W H R s w e r e rated b y their peers, b y themselves, and b y indep e n d e n t o b s e r v e r s as being m o r e socially d o m i n a n t a n d leader-like, b u t o n l y w h e n b e i n g o b s e r v e d b y an evaluator.

2 1.5 9

0 C

1

r= 0.5 o

[

~

0

evaluator present i

- - - evaluator absent

I

-a -0.5 0 0

-1.5

-2

T

-2 sd

-1 sd

i

0

i

+1 sd

+2 sd

W H R (z-scores)

Figure 1. The interaction of WHR and experimental condition, predicting social dominance.

356

Human Nature, Vol. 13, No. 3, 2002

GENERAL DISCUSSION The present research investigated the emergence of social dominance and leadership in small groups of unacquainted men. Across two studies, three sources of data--self-report, peer-report, and behavioral ratings---converged to support our predictions that men with relatively higher WHRs (indicative of better health) should display more socially dominant behaviors and emerge as leaders in small, all-male groups, particularly in an evaluative context. Because the WHR measure shares virtually no method variance with the self-report, peer-report, or behavioral ratings; the effect sizes are reasonably large; and all effects remained significant after controlling for several potential confounds, these results appear to be robust.

Viability Indicators As is true of "good provider" models of sexual selection, mounting evidence suggests that "good genes" sexual selection may also have occurred in evolutionary history (see Gangestad and Simpson 2000). "Good genes" models posit that individuals should be attracted to potential mates who possess honest (i.e., difficult-to-fake or "handicapped") indicators of good health/viability, qualities that could have been passed on to offspring genetically (Zahavi 1975). A handicapped trait is one that indicates greater genetic fitness but is costly to develop or maintain. Selection, therefore, should have favored people who preferentially mated with individuals who had such traits. During evolutionary history, individuals who possessed these traits should have been more willing and able to compete for mates, especially in competitive situations that allowed them to "advertise" their viability to rivals and potential mates (see Simpson et al. 1999). Such advertisements might have been particularly effective in brief, unstructured interactions between unacquainted individuals because behaviors diagnostic of greater viability (i.e., intrasexual social dominance) could have been displayed quickly and directly. In contrast, it should have been more difficult for individuals to convey their ability or willingness to be good "investors" in brief encounters because extended periods of time are needed to fully demonstrate long-term investment. Although several "good genes" models have been advanced, only a few potential indicators of viability in humans have been identified. The most widely studied indicator has been fluctuating asymmetry (FA). It has been suggested that more symmetrical individuals may have stronger immune systems, rendering them more resistant to parasites and many diseases (Thornhill and Gangestad 1993). Indeed, men who are more symmetrical do tend to be more attractive to women (especially as short-term or extrapair mates), and such men are more likely to compare themselves directly

WHR and Social Dominance

357

with and derogate rivals when competing for mates (Gangestad and Simpson 2000). WHR also predicts both current and long-term health status, but it is not correlated with FA (Gangestad and Thornhill, unpublished data). As discussed earlier, WHR is a testosterone-dependent secondary sexual characteristic and, as such, may signal greater immunological competence. In addition, it m a y be related to masculinized facial features, which also are testosterone dependent. Recent research has shown that women are most attracted to masculine faces during the follicular phase of their menstrual cycle (when they are ovulating; Penton-Voak et al. 1999). Higher levels of testosterone can compromise the immune system, suggesting that features regulated by testosterone may have undergone "good genes" sexual selection. Moreover, people with higher (more "desirable") WHRs are less likely to develop health problems commonly associated with aging, such as diabetes, cardiovascular diseases, certain forms of cancer, and strokes (Singh 1995). Given that higher WHR is reliably associated with better health and greater longevity, it may "honestly" signal that an individual is likely to survive and provide resources for an extended period of time.

CAVEATS AND CONCLUSIONS Although this research provides converging support for the role that fat distribution plays in the formation of status hierarchies in all-male leaderless groups, it has some limitations. Measures other than WHR (e.g., using calipers to measure abdominal fat directly) can also be used to assess fat distribution. Future research should include different measures of body fat distribution to test whether WHR in particular explains these emergent leadership results. Additionally, factors other than testosterone (e.g., bone structure, developmental history) may also influence fat distribution during development. Physical maturation processes, for instance, affect the rate and nature of body fat distribution in both sexes. Because participants in the present studies were young adult males, the results may not generalize to older males. Future research should consider additional factors that may also be implicated in the development of body fat and test the present hypotheses with men of different age groups. Future research also needs to identify which components of viability different indicators represent, how each one is behaviorally "advertised," how the various indicators interrelate, and what specific roles they play in attracting short-term versus long-term mates. Furthermore, research must document the role that possible indicators of viability assume in the maintenance of long-term relationships (e.g., relationship stability, fertility).

358

Human Nature, Vol. 13, No. 3, 2002

In conclusion, this is the first set of studies to d e m o n s t r a t e that W H R predicts social d o m i n a n c e a n d l e a d e r s h i p behavior in men. It suggests that W H R , a p r o b a b l e m a r k e r of genetic viability, is associated with e m e r g e n t leadership a n d intrasexual c o m p e t i t i o n in a c o n t e x t - d e p e n d e n t fashion. We thank Deborah A. Kashy and Jennifer Boldry for their help with the SRM analyses. We also thank Steven Gangestad for his helpful comments and information regarding an unpublished data set, Trey Greenwood for serving as the male evaluator, and Deborah Stanush for serving as the female evaluator. Thanks are also extended to the many research assistants who helped us collect these data, and to Wayne Hibbs. Lorne Campbell received his Ph.D. in psychology at Texas A&M University in 2001. He is an assistant professor of psychology, formerly at Simon Fraser University and now at the University of Western Ontario. His research interests include the study of romantic relationships and group processes from an evolutionary perspective. Jeffry A. Simpson, a professor of psychology at Texas A&M University, received his Ph.D. from the University of Minnesota in 1986. Most of his research focuses on interpersonal relationships, particularly mating, attachment, and relationship maintenance processes. Mark Stewart received his Ph.D. from Texas A&M University in 2001. He is now a lecturer in the Department of Psychology at Southern Methodist University. John G. Manning received his DVM and M.S. from Texas A&M University. He is currently a lecturer in the Poultry Science Department at Texas A&M University.

NOTES 1. Sociocultural models (e.g., Eagly and Wood 1999) make similar predictions regarding male social dominance and why women may be attracted to socially dominant men. These models, however, propose different underlying causes for these phenomena, particularly the adoption of traditional male and female sex roles. Moreover, they do not predict that indicators of male viability should be related to the emergence of social dominance, whereas "good genes" models clearly do. For the sake of brevity, we do not elaborate on sociocultural models. 2. Of the 228 participants in both studies, only four had WHRs of 1.00 or greater, and each of these people was in a different group. WHRs this large predict a very high incidence of long-term health problems. However, due to the small number of participants in this range, we could not test for curvilinear effects (i.e., whether WHRs in the .90-.95 range predicted increased leadership and social dominance than those that were higher or lower). The present study, therefore, is limited to testing linear relationships, with higher WHRs usually being in the more "desirable" range. For the SRM results, the four participants with WHRs of 1.00 or

WHR and Social Dominance

359

greater were not removed because doing so would necessitate removing their entire group from the analyses. Removing these groups, however, did not change the pattern of reported results. For all other analyses, these four participants were removed. REFERENCES

Andersson, M. 1994 Sexual Selection. Princeton: Princeton University Press. Argyle, M. 1994 The Psychology of Social Class. New York: Routledge. Barkow, J. 1989 Darwin, Sex, and Status: Biological Approaches to Mind and Culture. Toronto: University of Toronto Press. Bass, B. M.

1954 The Leaderless Group Discussion. Psychological Bulletin 51:465-492. Bjorntorp, P. 1987 Fat Cell Distribution and Metabolism. In Human Obesity, R. J. Wurtman and J. J. Wurtman, eds. Pp. 66-72. Albany: New York Academy of Sciences. 1988 The Association between Obesity, Adipose Tissue Distribution and Disease. Acta Medica Scandinavica 723 (Suppl.):121-134. 1991a Adipose Tissue Distribution and Functions. International Journal of Obesity 15:67-87. 1991b Visceral Fat Accumulation: The Missing Link between Psychological Factors and Cardiovascular Disease? Journal of Internal Medicine 230:195-200. Bouchard, C. 1988 Genetic Factors in the Regulation of Adipose Tissue Distribution. Acta Medica Scandinavica 723 (Suppl.):135-141. Bouchard, C., J. P. Despres, A. Nadeau, P. J. Lupien, G. Theriault, J. H. Dussalt, S. Moorjani, S. Pinault, and G. Fournier 1990 The Response to Long-term Overfeeding in Identical Twins. New England Journal of Medicine 322:1477-1482. Buss, D. M.

1981 Sex Differences in the Evaluation and Performance of Dominant Acts. Journal of Personality and Social Psychology 40:147-154. 1988 The Evolution of Human Intrasexual Competition: Tactics of Mate Attraction. Journal of Personality and Social Psychology 54:616-626. Buss, D. M., and D. P. Schmitt 1993 Sexual Strategies Theory: A Contextual Evolutionary Analysis of Human Mating. Psychological Review 100:204-232. Caporeal, L. R., and R. M. Baron 1997 Groups as the Mind's Natural Environment. In Evolutionary Social Psychology, J. A. Simpson and D. T. Kenrick, eds. Pp. 317-344. Mahwah, New Jersey: Erlbaum. Cascio, W. E 1987 Applied Psychology in Personnel Management, third ed. Englewood Cliffs, New Jersey: Prentice-Hall.

360

Human Nature, Vol. 13, No. 3, 2002

Cheek, J. M., and A. H. Buss 1981 Shyness and Sociability. Journal of Personality and Social Psychology 41:330-339. Darwin, C. 1871 The Descent af Man and Selection in Relation to Sex. London: Murray. Dawkins, R. 1989 The Se~sh Gene, revised ed. New York: Oxford University Press. de Waal, F. 1982 Chimpanzee Politics: Sex and Power among Apes. Baltimore: Johns Hopkins University Press. Donahue, R. P., R. J. Prineas, O. Gomez, and C. Hong 1990 Familial Resemblance of Body Fat Distribution: The Minneapolis Children's Blood Pressure Study. International Journal of Obesity 16:161-167. Eagl35 A. H., and W. Wood 1999 The Origins of Sex Differences in Human Behavior: Evolved Dispositions Versus Social Roles. American Psychologist 54:408-423. Ellis, L. 1995 Dominance and Reproductive Success among Nonhuman Animals: A Cross-species Comparison. Ethology and Sociobiology 16:257-333. Fisek, M. H., and R. Ofshe 1970 The Process of Status Evolution. Sociometry 33:327-346. Furnham, A., T. Tan, and C. McManus 1997 Waist-to-hip Ratio Preferences for Body Shape: A Replication and Extension. Personality and Individual D~frrences 22:539-549. Gangestad, S. W., and J. A. Simpson 2000 The Evolution of Human Mating: Trade-offs and Strategic Pluralism. Behavior and Brain Sciences 23:573-587. Gangestad, S. W., and R. Thornhill 1997 Human Sexual Selection and Developmental Instability. In Evolutionary Social Psychology, J. A. Simpson and D. T. Kenrick, eds. Pp. 169-195. Mahwah, New Jersey: Erlbaum. Harary, K., and E. Donahue 1994 Who Do You Think You Are? Explore Your Many-sided Se~:with the Berkeley Personality Profile. San Francisco: Harper. Hrdy, S. B. 1997 Raising Darwin's Consciousness: Female Sexuality and the Prehominid Origins of Patriarchy. Human Nature 8:1-49. Kalma, A. 1991 Hierarchisation and Dominance Assessment at First Glance. European Journal of Social Psychology 21:165-181. Kenny, D. A., and L. La Voie 1984 The Social Relations Model. In Advances in Experimental Social Psychology, L. Berkowitz, ed. Pp. 141-182. Orlando: Academic Press. Leibel, R. L., N. K. Edens, and S. K. Fried 1989 Physiological Basis for the Control of Body Fat Distribution in Humans. Annual Review of Nutrition 9:417-443.

WHR and Social Dominance

361

Manning, J. T., and P. E. Bundred 2000 The Ratio of 2nd to 4 th Digit Length: A New Predictor of Disease Predisposition? Medical Hypotheses 54:855-857. Manning, J. T., D. Scutt, J. Wilson, and D. I. Lewis-Jones 1998 The Ratio of 2nd to 4 th Digit Length: A Predictor of Sperm Numbers and Concentrations of Testosterone, Luteinizing Hormone and Estrogen. Human Reproduction 13:3000-3004. Marti, B., J. Tuomilehto, V. Saloman, L. Kartovaara, H. J. Korhonen, and P. Pietinen 1991 Body Fat Distribution in the Finnish Population: Environmental Determinants and Predictive Power for Cardiovascular Risk Factor Levels. Journal of Epidemiology and Community Health 45:131-137. Maslow, A. H. 1937 Dominance-Feeling, Behavior, and Status. Psychological Review 44:404429. Miller, G. F.

2000 The Mating Mind: How Sexual Choice Shaped the Evolution of Human Nature. New York: Doubleday. Penton-Voak, I. S., D. I. Perrett, D. L. Castles, T. Kobayashi, D. M. Burr, L. K. Murray, and R. Minamisawa 1999 Menstrual Cycle Alters Face Preference. Nature 399:741-742. P~russe, D. 1993 Cultural and Reproductive Success in Industrial Societies: Testing the Relationship at Proximate and Ultimate Levels. Behavioral and Brain Sciences 16: 267-322. Rebuffe-Scrive, M. 1987a Regional Adipose Tissue Metabolism in Men and in Women during Menstrual Cycle, Pregnancy, Lactation and Menopause. International Journal of Obesity 11:347-355. 1987b Regional Adipose Tissue Metabolism in Women during and after Reproductive Life and in Men. In Recent Advances in Obesity Research, E. M. Berry, S. H. Blondheim, H. E. Elihau, and E. Schafrier, eds. Pp. 82-91. London: John Libbey. 1988 Metabolic Differences in Fat Deposits. In Fat Distribution during Growth and Later Health Outcomes, C. Bouchard and F. E. Johnston, eds. Pp. 163-173. New York: Alan R. Liss. 1991 Neuroregulation of Adipose "Iissue: Molecular and Hormonal Mechanisms. International Journal of Obesity 15:83-86. Rebuffe-Scrive, M., P. Marin, and R Bjorntorp 1991 Effect of Testosterone on Abdominal Adipose "Iissue in Men. International Journal of Obesity 15:791-795. Rosenberg, M. 1965 Society and the Adolescent Self-image. Princeton: Princeton University Press. Sadalla, E. K., D. T. Kenrick, and B. Vershure 1987 Dominance and Heterosexual Attraction. Journal of Personality and Social Psychology 52:730-738.

362

Human Nature, Vol. 13, No. 3, 2002

Schmitt, D. P., and D. M. Buss 1996 Strategic Self-promotion and Competitor Derogation: Sex and Context Effects on the Perceived Effectiveness of Mate Attraction Tactics. Journal of Personality and Social Psychology 70:1185-1204. Simpson, J. A., S. W. Gangestad, P. N. Christensen, and K. Leck 1999 Fluctuating Asymrnetry, Sociosexuality, and Intrasexual Competitive Tactics. Journal of Personality and Social Psychology 76:159-172. Singh, D. 1993 Adaptive Significance of Female Physical Attractiveness: Role of Waistto-hip Ratio. Journal of Personality and Social Psychology 65:293-307. 1995 Female Judgment of Male Attractiveness and Desirability for Relationships: Role of Waist-to-hip Ratio and Financial Status. Journal of Personality and Social Psychology 69:1089-1101. Symons, D. 1979 The Evolution of Hurnan Sexuality. New York: Oxford University Press. Thornhill, R., and S. W. Gangestad 1993 Human Facial Beauty: Averageness, Symmetry, and Parasite Resistance. Human Nature 4:237-269. Thornton, G. C., and J. N. Cleveland 1990 Developing Managerial Talent through Simulation. American Psychologist 45:190-199. Tooby, J., and L. Cosmides 1992 The Psychological Foundations of Culture. In The Adapted Mind: Evolutionary Psychology and the Generation of Culture, J. Barkow, L. Cosmides, and J. Tooby eds. Pp. 19-136. New York: Oxford University Press. Trivers, R. 1972 Parental Investment and Sexual Selection. In Sexual Selection and the Descent of Man, B. Campbell, ed. Pp. 136-179. New York: Aldine de Gruyter. Walters, S., and C. B. Crawford 1994 The Importance of Mate Attraction for lntrasexual Competition in Men and Women. Ethology and Sociobiology 15:5-30. Zahavi, A. 1975 Mate Selection--A Selection for Handicap. Journal of Theoretical Biology 53: 205-214.