Unfortunately, looks matter. Everyone should start life on a level playing field, but they don’t. The egalitarianism embedded in modern social science abhors "built in" disadvantages, whether in social structure or in personal capital. But while explicating the consequences of structural inequalities, social science ignores equally important differences in how we look, in the effects on our fortunes of our faces and physiques. The late sociologist Louis Dotson used to remind his colleagues that "we have bodies," but few heard his message in the epoch after World War II, when the echoes of biological racism were still loud. Today the issue is ripe for attention, if approached with a proper respect for past abuses. We cringe at reviving the term "physiognomy," which carries so much baggage, but we do indeed address here the behavioral importance of people’s faces.
Effects of Facial Dominance
When raising a controversial topic, a good place to start is Square 1: Good looking people have different life experiences than unattractive people. This fact was ignored, minimized, even occasionally denied by social scientists during the postwar period. Since 1974, when Berscheid and Walster’s review in Advances in Experimental Social Psychology reinstated facial beauty as a proper topic for research, there has been a flow of work showing that attractive people are perceived more positively than the unattractive on all sorts of dimensions that seem unrelated to physical appearance. It has become clear that an attractiveness stereotype pervades society: what is beautiful is good. Beautiful and homely people apply this stereotype to themselves. We respond to beautiful people -- especially beautiful females -- in more positive ways than we respond to ordinary-looking ones, and they expect that response (Patzer 1985; Hatfield and Sprecher 1986). Thus, attractiveness may be the most invidious basis for approved discrimination in our society.
Obviously faces convey many impressions beside degree of attractiveness. Judges consensually (if not correctly) rate faces on such traits as intelligence, honesty, sympathy, or athleticism. From portraits, judges accurately distinguish Americans of Mediterranean descent from "WASPs", although they cannot generally distinguish a Jew from an Italian (Mazur 1973). Judges accurately judge a woman’s social class from her bridal portrait (Mazur 1993). Facial displays of political leaders affect emotional responses among their supporters (Sullivan and Masters 1988).
Our focus here is on facial dominance, the degree to which a person is judged from his or her facial appearance to be dominant, assertive and a leader, as opposed to someone who is subordinate, submissive, a follower. The face is one among several status signs which each individual displays, suggesting this his status is, or ought to be, high or low. Other status signs appear in our speech, posture, dress, and mannerisms. It is useful to divide facial signs into two categories: those that are fairly constant such as shape of the chin or degree of baldness, and those that are temporary or controllable such as smiling, gaze direction, or eyebrow movement. While individuals can do little about their constant attributes, they may manipulate their controllable signs to signal a dominant or deferent demeanor (Mazur 1985; Weisfeld and Linkey 1985; Kalma 1991).
Our research on human facial dominance began with controllable signals. Eyebrow position was known to play a role in the dominance displays of nonhuman primates, and it seemed plausible that the brows might signal status in humans too. American subjects, shown portraits of male and female models and asked to judge their dominance, gave significantly higher dominance ratings when models were posed with lowered brows than when they were posed with raised brows (Keating, et al. 1977). Inquiring if this were a universal feature of human status signaling, Caroline Keating, in a doctoral dissertation supervised by Marshall Segall, collected comparable data from 11 industrial and agrarian cultures in nations around the world. She found the brow effect only in Westernized cultures, concluding that it was not a universal signal. (In all cultures, nonsmiling faces were judged more dominant than smiling faces; Keating, et al. 1981.)
More important for our story was a serendipitous finding: Some stimulus faces were consistently rated dominant, independent of their expression, while other faces were consistently rated submissive. Thus, facial dominance was judged reliably across cultures. Some people always looked dominant, others always submissive (Keating, Mazur and Segall 1981).
An obvious question arose. Does facial dominance, like facial beauty, affect one’s life experiences? More specifically, does a facially dominant person have an advantage in moving up a status hierarchy? The United States Military Academy at West Point provided an ideal setting to answer this question. Graduates who make their careers in the military are continually evaluated for promotion, beginning in their junior and senior years at West Point, so their progress up the hierarchy is well measured. West Point’s yearbook, like all college yearbooks, shows portraits of the men in the graduating class, which can be rated for facial dominance.
What do dominant faces look like? Everyone knows because anyone can sort portraits on this basis, but facial dominance seems to be a gestalt concept, difficult to describe in simple terms. Faces identified as dominant are more likely to be handsome -- with striking exceptions, to be muscular, to have prominent as opposed to weak chins, and to have heavy brow ridges with deep set eyes. Submissive faces are often round or narrow, with ears "sticking out," while dominant faces are oval or rectangular with close-set ears (Mazur, et al. 1984). The descriptor "babyfaced" (Zebrowitz, et al. 1993) presumably refers to a submissive appearance.
Data were collected for West Point’s Class of 1950, a cohort which had nearly finished its military service by the time of study. For these men, facial dominance was strongly related to promotions in junior and senior years at West Point. However, facial appearance did not predict rank attainment after graduation, in early or middle career (Mazur, et al. 1984). Collins and Zebrowitz (1995) similarly found that possessing a "babyface" had no special effect on military promotion.
Our initial West Point data base was considerably strengthened in 1989 when questionnaires were sent to the men of the Class of 1950 for a follow-up study. By this time the highest ranking generals from the class had finished their service. With these changes and improvements, there was still no indication that facial dominance had aided promotions at mid-career. However, to our surprise, facial dominance -- measured from cadet portraits taken 20+ years earlier -- significantly predicted promotion to the highest ranks -- to the various levels of general officer. Having a dominant face was an advantage in reaching the top (Mazur and Mueller 1996; Mueller and Mazur 1996).
Apparently, facial appearance plays a role when the promotion board knows the candidates personally, as is true in the cadet years at West Point, and again near the end of the career line, when a small, highly select group of men compete for the highest ranks of general. At mid-career, promotions are made impersonally by boards who generally have no direct contact with the candidates; in this situation, dominant appearance is no help.
Many theorists suggest that dominant males leave more offspring than submissive males. While this is true for humans in premodern societies, the data are ambiguous for industrialized societies. Part of the problem is that class membership -- determined in modern societies by formal education and wealth -- and individual status within one’s class/occupational group are confounded when male dominance is measured by socioeconomic status. There are some studies that report a positive relation between individual professional success and family size (Haggod 1948; Freedman 1963; also see Wrong 1980 for citations of several earlier studies). Among our West Point graduates, generals had about 20% more children than other officers. Since dominant looking men, if they met average requirements for competence and expertise, found it easier than their classmates to become generals, a direct relation between facial dominance and family size was no surprise. This relation remained, though less strong, after the effect of career was controlled. Thus, there may be an effect of facial dominance per se on fitness.
High status human males have more copulatory opportunity (Perusse 1994). Portraits of boys in a study of teenage sexual behavior were rated for facial dominance, for handsomeness, and for stage of pubertal development. The handsome boys had more coital activity than their unattractive peers, an obvious finding. However, dominant looks were a better predictor than handsomeness of coital activity, and the effect of dominant looks on coital activity was significant even after controlling for handsomeness and pubertal development (Mazur, et al. 1994).
Why does facial dominance aid someone in obtaining valued prerogatives like promotion and sexual access? The standard "constructionist" explanation in modern social science is that facial dominance is a purely conventional signal, which carries no real information about intrinsic professional capabilities. It is a sign that certain corporate groups -- here the military -- use to distinguish themselves from others. Dominant looks, in this view, are an expression of adherence to the values of the group. Insofar as certain traits of a face are universally judged as dominant -- like strong jaws, broad cheek bones, or a prominent forehead, which Thornhill and his colleagues conjecture to be testosterone dependent features (Grammer and Thornhill 1994; Thornhill and Gangestad in press) -- the standard explanation would be that these facial features might have indicated actual dominance in the social and physical context of early human evolution, but do not carry such information anymore. In other words, looks are no longer truly indicative of behavior, but dominant looking men do well simply because the grantors of prerogatives are fulfilling expectations set up by a substanceless facade.
Problems with the standard constructionist explanation are suggested by game theoretical investigations into the evolutionary stability of signals that reveal one’s internal qualities, i.e., qualities not directly observable to others. The reasoning goes as follows. Facial dominance obviously matters today: it makes others give in to, or take orders from, dominant looking individuals, thus providing the latter better access to the good things in life. Why, then, don’t subordinate individuals cheat, trying to look more dominant than they are? Mathematical modeling of "evolutionarily stable strategies" shows that signals like facial dominance may honestly indicate an internal quality -- here, actual dominance -- provided only that (1) signaling carries some costs, and (2) marginal costs of signaling are higher for "low quality" individuals (Zahavi 1975, 1977; Enquist 1985; Grafen 1990). While physiological costs of signaling dominance are difficult to measure, the social costs of falsely signaling dominance have been demonstrated in birds. For example, brown chest spots are a sign of dominance in house sparrows. A subdominant sparrow with an artificially enlarged brown spot is attacked more fiercely by dominant birds than subdominant birds with appropriately smaller spots. If birds with artificially enlarged brown spots are given testosterone supplements, making them more aggressive, then after some fights they are accepted as having high status (Moller 1987; also see Rohwer 1977; Roskaft and Rohwer 1987). Among humans, baby faced individuals who perceive themselves, and are perceived by others, as warm and unaggressive, receive relatively harsh punishment for wrongdoing. Also, among our officers, for those who were below average in intellectual and social skills, dominant looks possibly were a handicap in their careers rather than an asset. Therefore, contrary to the conventional constructionist view, facial dominance may be an honest signal of capabilities that actually make a man dominant in his group. This evolutionary speculation seems as compelling as the constructionist version. In either case, dominant looking actors are more likely to get (or take) what they want, whether because of their intrinsic qualities or because people simply expect them to dominate.
In most studies described above, facial dominance was measured from black-and-white graduation portraits found in yearbooks typical of American high schools and universities. Portraits are copied on slides. In order to control the size of the portrait, all slides are made in vertical format with the head nearly filling the frame.
Slides are projected in front of 20-40 judges (usually undergraduate classes) who view each for about 10 seconds and independently rate faces on a seven-point scale of dominance-submissiveness (1 = very submissive, 4 = neutral or undecided, 7 = very dominant). Judges are instructed that a dominant person tells other people what to do, is respected, influential, and often a leader; submissive or subordinate people are not influential or assertive and are usually directed by others. The judges are then shown several slides before the ranking task begins in order to accustom them to the range of variation. To avoid fatigue, a maximum of 24 slides is shown in a single series, though judges sometimes rate more than one series separated by a break of several minutes. Judges’ ratings are not affected by the order of slide presentation, by whether the model is facing left or right, or by sex of the judge.
The median score for each slide is taken as the value of facial dominance. Among 416 West Point cadets, these medians ranged from 2 (moderately submissive) to 7 (very dominant), with a mode of 5 (slightly dominant). On 85% of cadet slides, at least halfthe judges’ choices fell within two adjacent scale points, indicating more clustering than would be expected if choices were uniformly distributed across the seven-point scale. Slides that do not meet this cluster requirement are dropped as unreliable.
There are pitfalls in this method. While we want a dominance rating of each face, the rating actually obtained may also reflect the subject’s expression and pose, or features of the portraiture (e.g., removal of facial blemishes, lighting). For example, faces are known to be judged less dominant when they are smiling than when the same faces are unsmiling (Keating, et al. 1981). To evaluate this effect, cadet portraits were sorted into categories of no smile, broad smile with teeth showing, and slight smile (no teeth). The no smiles had a mean dominance score .3 higher (more dominant) than the slight smiles (ns, t-test). The broad smiles had a mean score .7 less than all other portraits (p < .001). Also, cadets who posed gazing at the camera scored .2 higher than those gazing away (ns). Corrections for these various factors have little effect on analytic results, so our practice is to use the scores without correction.
Faces change with age (Zebrowitz, et al. 1993). Since dominance scores measured from cadet portraits predicted promotions two decades later, we wondered if portraits taken in middle age would have done even better. We asked a subset of men from the Class of 1950 if they would send us a formally posed portrait taken around 1970. We received 30 usable portraits (24% response) taken between 1960 and 1982 (average year = 1972). The final military ranks of men who sent portraits (33% generals, 43% colonels, 23% lieutenant colonels) are high compared to final ranks for all men who remained in the military for 20 years or more (12% generals, 50% colonels, 38% lieutenant colonels), however their distribution of facial dominance scores, measured as cadets, is similar to that of the whole class.
These later pictures, more diverse than the cadet portraits, were brought to uniform size and format during conversion to slides; and badges of military rank were obscured. While most men were photographed in uniform, some wore civilian clothes (including three in casual dress). Seven portraits were in color. Whereas cadet portraits were touched up to remove facial blemishes, and no cadet wore glasses, none of the later portraits seem to be touched up, and glasses are worn in four.
Slides were rated for dominance by 35 undergraduate judges, following the procedure used for cadet portraits. (Cadet portraits were rated in 1982, later portraits in 1992.) On 84% of the slides, at least half the judges’ scores fell within two adjacent scale points, virtually the same reliability obtained for cadet portraits. As usual, the median score for each slide was taken as its measure of facial dominance.
The 30 later portraits received somewhat lower dominance scores (range 2-6, mode 4) than the same 30 men evaluated as cadets (range 2-6, mode 5), but not significantly so.
Does the relative dominance ranking of men persist over two decades? The Pearson correlation for these men between their cadet score and their later score is r= .36 (p= .05). (Using ordinal statistics, gamma = .44; p= .04.) The second row of Figure 1 shows middle-age portraits of the men pictured as cadets in the first row, allowing a comparison of their changes with age.
To test if ratings were affected by arbitrary features of a pose or portrait, dominance scores for the 30 later portraits were simultaneously regressed against dummy variables representing clothing (uniform or tie-and-jacket versus casual dress), whether the portrait was in color versus black-and-white, whether glasses were worn, the presence of a full smile (teeth showing), whether the subject was looking toward the camera versus away, and the year of the portrait. Men in casual dress had dominance scores 1.9 points lower than men in uniform or in a tie and jacket (p< .001), men with glasses scored 1.1 points lower than other men (p = .02), and those who looked away from the camera scored a half point lower than those looking at it (p= .07). Year of the portrait, and whether it was in color, have little effect on dominance score. (Dominance scores adjusted for casual dress and glasses were no more strongly related to cadet dominance than the raw scores.)
Unlike cadet portraits, where men with broad smiles were scored less dominant than those with slight or no smiles, broad smiles did not affect ratings of the later portraits. Remarkably, three of the four men who smiled broadly as cadets were among the five men who smiled broadly in middle age portraits. Apparently, smiling for the camera is a persistent habit.
Three judges independently rated the 30 men, as both cadets and in middle age, on a seven-point scale of attractiveness, and their median score was assigned to each portrait. Attractiveness was fairly stable across the twenty years between portraits (r= .48, p= .01), moreso than dominance. Handsomeness is significantly related to dominance among cadets (r= .40, p= .03) but not among the middle-age men.
If dominant facial appearance improves one's chance of
promotion during the later part of a career, we would expect final military
rank to correlate more strongly with dominance ratings from middle-age
portraits than from cadet pictures. Unfortunately we have only 30 middle-age
portraits, too few for a powerful test of this hypothesis, but the results
in Figure 2 suggest it is correct. There, mean facial dominance is plotted
against highest military rank achieved, using both middle-age portraits
and cadet portraits (for the same 30 men). Facial dominance is seen to
increase more smoothly with military rank when middle-age portraits are
used than when cadet portraits are the basis for ratings.
Military rank at midcareer, which was uncorrelated with facial dominance as measured from cadet pictures, remains uncorrelated to dominance as measured by later portraits.
Facial dominance, measured from portraits, has emerged as an important variable in predicting rank attainment in the military, and coital opportunity for teenage boys. It ought to become one of the variables routinely considered in explaining behaviors related to power, prestige, leadership, influence, status, and the attainment of valued prerogatives.
Facial dominance is sufficiently persistent over time that measurements taken from West Point graduation portraits correlate moderately well with portraits of the same men taken roughly two decades later. Dominance measurements from cadet portraits are sufficiently valid to predict promotions 20+ years later, although the use of middle-age portraits strengthens these predictions slightly.
Results to date suggest that facial dominance is a robust
variable, which can be measured from a variety of photographs, even when
the portraiture is not uniform in style or dating. Interviewers might easily
be trained to take portrait photos during standard face-to-face interviews,
thus incorporating facial dominance as a variable in survey data sets.
Portraits in historical archives, and taken from diverse print and electronic
media, should be adequate for facial ratings, although additional validation
studies are advisable before setting out on this approach.
Berscheid, E. and E. Walster. 1974. "Physical Attractiveness." In Advances in Experiment Social Psychology, ed. L. Berkowitz, 7: 157-215. New York: Academic Press.
Collins, M. and L. Zebrowitz. 1995. "The Contributions of Appearance to Occupational Outcomes in Civilian and Military Settings." Journal of Applied Social Psychology 25: 29-163.
Enquist, M. 1985. "Communications during Aggressive Interactions with Particular Reference to Variations in Choice Behavior." Animal Behaviour 33: 1152-1161.
Freedman, D. 1963. "The Relation of Economic Status and Fertility." American Economic Review 53: 414-426.
Grafen, A. 1990. "Biological Signals as Handicaps." Journal
of Theoretical Biology 144: 517-546.
Grammer, K. and R. Thornhill. 1994. "Human Facial Attractiveness and Sexual Selection: The Roles of Averageness and Symmetry." Journal of Comparative Psychology 108: 233-242.
Hagood, M. J. 1948. "Changing Fertility Differentials among Farm Operator Families in Relation to Economic Size of Farm." Rural Sociology 13: 314-328.
Hatfield, Elaine and Susan Sprecher. 1986. Mirror, Mirror... The Importance of Looks in Everyday Life. Albany: State University of New York Press.
Kalma, Akko. 1991. "Hierarchisation and Dominance Assessment at First Glance." European Journal of Social Psychology 21: 165-181.
Keating, C., A. Mazur, and M. Segall. 1977. "Facial Gestures Which Influence the Perception of Status." Sociometry 40: 374-378.
Keating, C., A. Mazur, and M. Segall. 1981. "A Cross-cultural Exploration of Physiognomic Traits of Dominance and Happiness." Ethology and Sociobiology 2: 41-48.
Keating, C., A. Mazur, M. Segall, P. Cysneiros, W. Divale, J. Kilbride, S. Komin, P. Leahy, B. Thurman, and R. Wirsing. 1981. "Culture and the Perception of Social Dominance from Facial Expression." Journal of Personality and Social Psychology 40: 615-626.
Mazur, A. 1973. "The Role of Predisposition in Identifying Jews." Jewish Social Studies 35: 290-291.
Mazur, A. 1985. "A Biosocial Model of Status in Face-to-Face Primate Groups." Social Forces 64: 377-402.
Mazur, A. 1993. "Signs of Status in Bridal Portraits," Sociological Forum 8: 273-283.
Mazur, A. and U. Mueller. 1996. "Channel Modeling: From West Point Cadet to General." Public Administration Review.
Mazur, A., C. Halpern, and J. Udry. 1994. "Dominant Looking Male Teenagers Copulate Earlier." Ethology and Sociobiology 15: 87-94.
Mazur, A., J. Mazur, and C. Keating. 1984. "Military Rank Attainment of a West Point Class: Effects of Cadets’ Physicals Features." American Journal of Sociology 90: 125-150
Mueller, U. 1992. "Is there a Stabilizing Evolution around Average Fecundity in Modern Human Populations? Findings from Two Samples of U.S. Military Men. Abstracts of the Fourth International Behavior Ecology Congress. Princeton.
Mueller, U. and A. Mazur. 1996. "Facial Dominance In West Point Cadets Predicts Military Rank 20+ Years Later." Social Forces.
Patzer, Gordon. 1985. The Physical Attractiveness Phenomenon. New York: Plenum.
Perusse, D. 1993. "Cultural and Reproductive Success in Industrial Societies: Testing the Relationship at the Proximate and Ultimate Levels." Behavioral and Brain Sciences 16: 267-322.
Rohwer, S. 1977. "Status Signaling in Harris Sparrows: Some Experiments in Deception." Behaviour 61: 107-128.
Roskaft, E. and S. Rohwer. 1987. "An Experimental Study of the Function of the Red Epaulettes and the Black Body Colour of Male Red-winged Blackbirds." Animal Behaviour 35: 1070-1077.
Sullivan, D. and R. Masters. 1988. "Happy Warriors: Leaders’ Facial Displays, Viewers’ Emotions, and Political Support." American Journal of Political Science 32: 345-368.
Thornhill, R. and S. W. Gangestad. In press. "The Evolution of Human Sexuality." Trends in Ecology and Evolution.
Weisfeld, G. and J. Linkey. 1985. "Dominance Displays as Indicators of a Social Success Motive." In Power, Dominance, and Nonverbal Behavior, J. Dovidio and S. Ellyson (eds.). New York: Springer-Verlag: 109-128.
Wrong, D. 1980. Class Fertility Trends in Western Nations. New York: Arno Press.
Zahavi, A. 1975. "Mate Selection - A Selection for A Handicap." Journal of Theoretical Biology 53: 205-214.
Zahavi, A. 1977. "The Cost of Honesty." Journal of Theoretical Biology 67: 603-605.
Zebrowitz, L., K. Kendall-Tackett, and J. Fafel. 1991. "The Influence of Children’s Facial Maturity on Parental Expectations and Punishments." Journal of Experimental Child Psychology 52: 221-238.
Zebrowitz, L., K. Olson and K. Hoffman. 1993. "Stability of Babyfaceness and Attractiveness Across the Life Span." Journal of Personality and Social Psychology 64: 453-466.