153 British Journal of Psychology (2014), 105, 153 161 2013 The British Psychological Society www.wileyonlinelibrary.com Increased facial width-to-height ratio and perceived dominance in the faces of the UK s leading business leaders Shuaa Alrajih 1 and Jamie Ward 1,2 * 1 School of Psychology, University of Sussex, Brighton, UK 2 Sackler Centre for Consciousness Science, University of Sussex, Brighton, UK The relative proportion of the internal features of a face (the facial width-to-height ratio, FWH) has been shown to be related to individual differences in behaviour in males, specifically competitiveness and aggressiveness. In this study, we show that the Chief Executive Officers (CEOs) of the leading UK businesses have greater FWHs than age- and sex-matched controls. We demonstrate that perceivers, naive as to the nature of the stimuli, rate the faces of CEOs as higher in dominance or success, and that ratings of dominance or success are themselves correlated with the FWH ratio. We find no association with other inferred traits such as trustworthiness, attraction or aggression. The latter is surprising given previous research demonstrating a link between FWH and ratings of aggression. We speculate that the core association may be between FWH and drive for dominance or power, but this can be interpreted as aggression only in particular circumstances (e.g., when the stimuli are comprised of faces of young, as opposed to middle-aged, men). When encountering an unfamiliar face in a neutral pose, perceivers make use of certain structural characteristics of the face to make inferences about their personality and, hence, their likely behaviour (for a review see Todorov, Said, & Verosky, 2011). One facial characteristic that has attracted considerable attention recently is the facial widthto-height ratio (FWH) measured as the maximum horizontal distance from the left facial boundary to the right facial boundary (width) divided by the distance from the top of the lip to the highest point of the eyelids (height). In males, variability in the FWH has been linked to aggression (Carre & McCormick, 2008) and deception (Haselhuhn & Wong, 2012). In this study, we provide evidence that it is linked to business leadership. Specifically, we show that the Chief Executive Officers (CEOs) of the UK leading businesses have a higher FWH than age- and sex-matched controls and, moreover, that variability in the FWH predicts naive participants ratings of dominance and success based on the face alone. Carre and McCormick (2008) reported the first experimental evidence that FWH predicts aggressive behaviour in men (but not women). In one study, participants played a game against another player in which one button earns points, another button steals points from other players, and a third button protects one s own points. Aggression was operationally defined as the number of button presses that steal other players points and *Correspondence should be addressed to Dr Jamie Ward, School of Psychology, University of Sussex, Falmer, Brighton BN1 9QH, UK (email : jamiew@sussex.ac.uk). DOI:10.1111/bjop.12035
154 Shuaa Alrajih and Jamie Ward this was found to correlate with FWH of the male players. In a different paradigm in which players could cheat to increase their chances of winning in a lottery, FWH was found to be related to deceptive behaviour (Haselhuhn & Wong, 2012). Other research using multiplayer games suggests that, in men, higher FWH is related to a tendency to exploit someone else s trust (Stirrat & Perrett, 2010) but, also, increased cooperation to other ingroup members during an intergroup competition (Stirrat & Perrett, 2012). As such, it may be linked to a desire to succeed rather than strictly selfish or anti-social behaviour. Finally, there is also a link between FWH and subjective ratings of aggressiveness based on the face alone (Carre, Morrissey, Mondloch, & McCormick, 2010; Geniole, Keyes, Mondloch, Carre, & McCormick, 2012). Of particular relevance to this study is intriguing evidence for a relationship between FWH and both leadership success and leadership style. Lewis, Lefevre, and Bates (2012) examined the relationship between FWH and rated personality traits in 29 former US Presidents. The factor of achievement drive (tendency to persist and not quit) was found to be related to FWH, but not other traits relating to aggression and dominance (such as forcefulness, pacifism). Among a student sample, Haselhuhn and Wong (2012) found a correlation between FWH and self-reported power (the latter determined via questionnaire items such as I can get people to listen to what I say ). A number of studies have examined the relationship between success, measured objectively, and ratings of power or dominance based on ratings of faces alone. Rule and Ambady (2011) found that uninformed participants ratings of power of faces of partners of leading law firms correlated with a measure of their financial success. A similar finding is found between ratings of power or dominance and company profits of CEOs of US-based Fortune-500 companies (Rule & Ambady, 2008). However, neither of these studies attempted to correlate particular facial characteristics (such as FWH) with their trait ratings. The original theoretical motivation for postulating a link between FWH and aggression or dominance behaviour derives from the suggestion that testosterone is responsible for the development of male secondary sex traits such as facial masculinity during puberty (Verdonck, Gaethofs, Carels, & de Zegher, 1999). High testosterone levels are associated with aggressive behaviour and pursuit of dominance status in men (Mazur & Booth, 1998). Moreover, there is a link between salivary testosterone and ratings of facial masculinity at baseline (Penton-Voak & Chen, 2004) or in response to winning (Pound, Penton-Voak, & Surridge, 2009). Initial findings suggested that the FWH was sexually dimorphic (males having greater FWHs than females) consistent with the idea that testosterone may be the mediating factor between facial structure and behaviour (Weston, Friday, & Lio, 2007). However, further studies with larger samples have shown that FWH is not sexually dimorphic (Lefevre et al., 2012; Ozener, 2012). The findings that FWH is not sexually dimorphic do not logically invalidate a potential role of testosterone in influencing FWH within men. For instance, it may be that the effects of testosterone on FWH act within sex (or possibly only within men). Whatever the causal mediator between FWH and behaviour, the evidence for a relationship between them is compelling. The aim of this research was to establish, for the first time, whether leading CEOs have a higher FWH than sex- and age-matched controls. A secondary aim was to determine whether, for this same sample of faces, there is a relationship between FWH and ratings of dominance and success when the faces are shown to naive participants who are asked to evaluate them. Our stimuli are based on images of CEOs of the UK FTSE-100 Index, the most widely used stock market indicator in the United Kingdom.
Faces of business leaders 155 Method Participants A total sample of 20 (12 women, 8 men; age range = 18 35 years) students from the University of Sussex participated in the ratings of personality and an additional set of 10 students (7 women, 3 men; age range = 20 27 years) were involved in rating the images to ensure that they were matched. Two further participants (the first author and an independent rater) measured the FWH ratios. The independent rater was blinded to the hypotheses and the nature of the stimuli. The study was approved by the School of Psychology Research Ethics Committee at the University of Sussex. Materials and stimuli Stimuli consisted of two sets of facial photographs. The first group featured the faces of CEOs of UK FTSE-100 companies (as listed in the review of the last quarter of 2011). The photographs of CEOs were obtained from company websites and Google Images. Only male CEOs were included (there were six female CEOs at that time) and it was not possible to obtain a clear image for one CEO, so the final set contained 93 faces. The mean age of the CEOs was 52.5 years (SD = 6.37). The second group, which served as the control, consisted of an initial set of 100 photographs (reduced to a final set of 93) taken from the internet including university websites (but not the University of Sussex). The controls were all male and were matched to the CEOs in terms of the presence or absence of facial hair, ethnicity, wearing of glasses, and approximate age (given that exact age was unknown for the controls). Each photograph was cropped, so that the background appeared white and clothing was not visible. Only the head and neck were visible. All photographs were in colour, facing forward, and had a neutral expression. However, given that we were unable to manipulate the tilt of the head or the expression, the images were rated along these dimensions to ensure reasonable matching. Procedure: facial measures Following the methodology of Carre and McCormick (2008), the FWH ratio was measured with ImageJ (National Institutes of Health open-source software) for each photograph. Specifically, bizygomatic width was measured, which is as the maximum horizontal distance between the right and left facial boundary; face height was measured as the vertical distance between the highest point of the upper lip and the highest point of the eyelids (see Figure 1). The FWH ratio was calculated as width divided by height. Procedure: ratings of personality and image characteristics Participants were seated in front of a 14-inch LCD monitor at a comfortable viewing distance. They were instructed that they would be shown a series of male faces and that their task was to judge the personality of these people based on their faces alone. They were reassured that there is no right or wrong answer and that they should go with their intuition. They were told that most of the faces would probably be unknown to them, but they should indicate if they did recognize anyone. All faces were presented to each participant in a randomized order. However, to prevent the experiment being too long, each participant provided ratings for only three of the possible six dimensions (in addition to a familiarity judgement). The selection of traits
156 Shuaa Alrajih and Jamie Ward Figure 1. An example illustrating how the facial width-to-height ratio (FWH ratio) was calculated for our stimuli. In the rating experiment, single images were presented to participants with the rectangle removed. The faces have FWH ratios of 1.81 (left) and 2.20 (right). to rate was based on a Latin square design. The procedure within each trial was as follows. A fixation point (+) was presented for 1000 ms at the centre of the screen. This was replaced by a photograph (4 9 6 cm) with the following text beneath it Is this person familiar? Y/N. Participants were required to respond in their own time. Following this, the first question appeared (again requiring a keyboard response), followed by the second and third questions (each requiring a keyboard response). The face remained on the screen during all three trait judgements. After all three questions had been asked, the fixation cross reappeared followed by the next face and the familiarity question. The six possible questions were as follows, with all but the last having a 0 7 scale: 1. How dominant? (0 = extremely non-dominant, 7=extremely dominant). 2. How trustworthy? (0 = extremely untrustworthy,7=extremely trustworthy). 3. How aggressive? (0 = extremely unaggressive, 7=extremely aggressive). 4. How attractive? (0 = extremely unattractive, 7=extremely attractive). 5. How successful? (0 = extremely unsuccessful, 7=extremely successful). 6. What age? (in years). Age was included in addition to personality traits because the exact age of the controls was not known and nor was it known when the photos of the CEOs had been taken. Participants were offered a break at the half-way point and the experiment itself lasted no more than 1 hr. A second set of participants were asked to rate the images themselves using the procedure described above, but replacing the personality questions with the following three questions: How expressive? (0 = completely neutral, 7 = extremely expressive) ; How tilted (up/down)? (0 = completely forward facing, 7 = extremely tilted) ; and How tilted (left/right)? (0 = completely forward facing, 7 = extremely tilted). The order of questions was rotated across participants. All of these factors may distort the
Faces of business leaders 157 measured FWH ratio and could possibly affect the CEO group more (given that appropriate facial images were harder to obtain for this group). Results Image ratings The ratings of facial expressiveness in the CEO group and the control group were low (means [SD] of 0.92 [0.32] and 0.92 [0.36] on a 0 7 scale) and did not differ significantly (t(9) = 0.01, p =.991). Similarly, the ratings of left right tiltedness in the CEO group and the control group were low (means [SD] of 0.46 [0.29] and 0.50 [0.26] on a 0 7 scale) and did not differ significantly (t(9) =.45, p =.66). Although the ratings of up down tiltedness in the CEO group and the control group were low (means [SD] of 0.36 [0.18] and 0.53 [0.29] on a 0 7 scale), they did differ significantly with the control images rated as more tilted (t(9) = 2.94, p =.017). Removing the seven most extreme control images (thus, leaving 93 images in each group) resulted in satisfactory matching along this dimension too (t(9) = 1.59, p =.147). This matched set of images was carried forward to the subsequent analyses. Facial width-to-height Inter-rater agreement across the two researchers was high for both CEO faces (Cronbach s a =.99) and control faces (a =.99). As such, the mean FWH value was used as the measure for all further analyses. The results show that CEO faces (M = 2.04, SD =.16) had a higher FWH ratio than the control group faces (M = 1.87, SD =.15). This difference was statistically significant (t(184) = 7.55, p <.001). Personality measures Six participants were familiar with a total of 33 photographs (14 were CEOs), and these items were excluded from the analysis. The results for the five personality traits are summarized in Figure 2. The CEO group was rated as more dominant than the faces of the control group (t(9) = 2.42, p=.04) and as more successful (t(8) = 3.53, p =.008). However, they were not rated as differing on perceived trustworthiness (t(10) =.13, Figure 2. Mean rating for the five personality traits between the two groups. Error bars represent one standard deviation. *p <.05; **p <.01.
158 Shuaa Alrajih and Jamie Ward Table 1. Correlations between facial width-to-height ratio (FWH) and mean personality trait ratings for each face FWH Dominant Trustworthy Aggressive Attractive Success Dominant.28** Trustworthy.08.01 Aggressive.06.20**.14 Attractive.10.16*.02.33** Successful.30**.52**.01.41**.38** Age.12.45**.04.32**.11.40** Note. *p <.05; **p <.01. p =.90), attractiveness (t(10) = 1.48, p =.17), or aggression (t(9) = 2.10, p =.07). The latter showed a trend in the reverse direction. The average perceived age (in years) of the CEOs was 51.6 years (SD = 4.73) and for the controls was 47.7 years (SD = 4.43). Although numerically similar, this difference was in fact significant (t(8) = 13.15, p =.001). The analyses below, however, suggest that age is not the mediator variable for the effects we observe. Relationship between FWH and inferred personality The correlation among FWH, age, and the five personality traits is shown in Table 1, considering the entire pool of faces. Only two scales correlated significantly with FWH: these were dominance and success. To ensure that these correlations were not mediated by age, partial correlations were run controlling for age. The correlations between FWH and dominance and success remained significant (r =.258,.278, respectively, p <.001). As such, we can conclude that CEOs and controls differ in their FWH ratio and that people (unfamiliar with the faces) use this dimension to infer dominance or success. Trustworthiness did not correlate with any of the other rating scales. Attractiveness was significantly negatively correlated with aggressiveness and age, but positively correlated with perceived dominance or success. Discussion In this study, we show that the CEOs of the leading UK businesses have greater FWHs than age- and sex-matched controls. We demonstrate that perceivers rate the faces of CEOs as higher in dominance or success (but do not differ on other trait measures), and that ratings of dominance or success are themselves correlated with the FWH ratio. 1 There are two possible interpretations of this finding (and they are not mutually exclusive): namely that FWH serves as a reliable marker of some internal disposition relating to dominance or, alternatively, that people select their leaders according to this facial characteristic. On balance, the evidence is more consistent with the former. In experimental situations, men with higher FWH behave in a more competitive way (Stirrat & Perrett, 2012) suggesting that the difference lies within them rather than others. Moreover, Haselhuhn and Wong (2012) showed that although FWH was correlated with deceptive behaviour, this is mediated by individual differences, an internal disposition 1 Given that ratings of dominance and success are themselves highly correlated (and conceptually related), we consider this as a single finding rather than two separate ones.
Faces of business leaders 159 relating to social dominance. If social dominance is the underlying psychological trait, then the underlying biological mechanism remains to be determined. This study compared CEOs against controls. However, another recent study examined differences within a group of CEOs themselves. Wong, Ormiston, and Haselhuhn (2011) found a correlation between FWH of company CEOs (from the USbased Fortune 500) and the financial performance of their firms (return on assets), after taking into account other variables (such as size of the firm). Our study replicates previous findings that FWH is unrelated to perceived trust or attractiveness (Stirrat & Perrett, 2010), but we fail to replicate the relationship between FWH and perceived aggression (Geniole et al., 2012; Short et al., 2012). Aggression and dominance are conceptually related, but are not the same. Although aggression is defined by its intent to cause harm (Anderson & Bushman, 2002), dominance need not entail this. Social dominance can be achieved either by coercive or pro-social tactics and dominant individuals typically adopt a mixed strategy (Hawley, Little, & Rodkin, 2007) rather than resorting to bullying and intimidation alone. The discrepancy between our findings and previous ones, with respect to aggression, may be related to differences in the stimuli. Our faces were of middle-aged and elderly men, but previous research has used young men. Our explanation is that high FWH is perceived as aggressive in young men, but dominant, non-aggressive in older men. Consistent with this, we found a significant negative correlation between ratings of age and aggressiveness, but a significant positive correlation between age and dominance or success. It is important to consider, and discount, various alternative account of this data. Firstly, one might wonder whether there is something unusual about the faces of our control group (as opposed to the CEO group). A comparison with other studies suggests that the FWH of our control group (= 1.87) is typical. The average adult male FWH reported by other groups selected by opportunistic sampling are 1.78 (Haselhuhn & Wong, 2012), 1.84 2.12 (in three separate Caucasian samples; Lefevre et al., 2012), and 1.89 ( Ozener, 2012). Instead, the FWH of 2.04 from our CEOs is similar to the mean FWH of 1.99 reported for US Presidents (Lewis et al., 2012) and 1.96 for the US-based CEOs reported by Wong et al. (2011). We also do not mean to imply that our controls are unsuccessful. Some may well be at the top of the hierarchy in their niche, but many will not be. It would be important to address this systematically in the future. Secondly, one could entertain the possibility that the CEOs and controls are not equally unfamiliar (given that the former are likely to receive some media coverage), despite our best efforts to control for familiarity. Of course, this is possible, but it remains unclear why it would have produced this particular pattern. Based on the existing literature, mere exposure would be expected to increase ratings of traits such as attractiveness (Rhodes, Halberstadt, & Brajkovich, 2001), but not necessarily traits such as dominance or success. Finally, one might wonder whether the CEO faces and controls differed in other ways not considered here (e.g., image clarity, hairstyle). Although this remains possible, the presence of an additional variable would not negate the current findings unless it could be demonstrated that such a factor also accounted for differences in FWH, differences in dominance or success, and the relationship between them. In summary, we have provided evidence for a link between FWH and business leadership and we have established that social dominance may be the mediating psychological trait. Acknowledgements S. A. was supported by a scholarship from the King Saud University, Saudi Arabia.
160 Shuaa Alrajih and Jamie Ward References Anderson, C. A., & Bushman, B. J. (2002). Human aggression. Annual Review of Psychology, 53, 27 51. doi:10.1146/annurev.psych.53.100901.135231 Carre, J. M., & McCormick, C. M. (2008). In your face: Facial metrics predict aggressive behaviour in the laboratory and in varsity and professional hockey players. Proceedings of the Royal Society B-Biological Sciences, 275, 2651 2656. doi:10.1098/rspb.2008.0873 Carre, J. M., Morrissey, M. D., Mondloch, C. J., & McCormick, C. M. (2010). Estimating aggression from emotionally neutral faces: Which facial cues are diagnostic? Perception, 39, 356 377. doi:10.1068/p6543 Geniole, S. N., Keyes, A. E., Mondloch, C. J., Carre, J. M., & McCormick, C. M. (2012). Facing aggression: Cues differ for female versus male faces. PLoS ONE, 7, e30366. doi:10.1371/journal. pone.0030366 Haselhuhn, M. P., & Wong, E. M. (2012). Bad to the bone: Facial structure predicts unethical behaviour. Proceedings of the Royal Society B: Biological Sciences, 279, 571 576. doi:10.1098/ rspb.2011.1193 Hawley, P. H., Little, T. D., & Rodkin, P. C. (2007). Aggression and Adaptation: The Bright Side to Bad Behavior. Hillsdale, NJ: Lawrence Erlbaum Associates. Lefevre, C. E., Lewis, G. J., Bates, T. C., Dzhelyovaa, M., Coetzeed, V., Deary, I. J., & Perrett, D. I. (2012). No evidence for sexual dimorphism of facial width-to-height ratio in four large adult samples. Evolution and Human Behavior, 33, 623 627. doi:10.1016/j.evolhumbehav. 2012.03.002 Lewis, G., Lefevre, C., & Bates, T. (2012). Facial width-to-height ratio predicts achievement drive in US presidents. Personality and Individual Differences, 52, 855 857. doi:10.1016/j.paid.2011. 12.030 Mazur, A., & Booth, A. (1998). Testosterone and dominance in men. Behavioral and Brain Sciences, 21(3), 353 397. doi:10.1017/s0140525x98001228 Ozener, B. (2012). Facial width-to-height ratio in a Turkish population is not sexually dimorphic and is unrelated to aggressive behavior. Evolution and Human Behavior, 33, 169 173. doi:10. 1016/j.evolhumbehav.2011.08.001 Penton-Voak, I. S., & Chen, J. Y. (2004). High salivary testosterone is linked to masculine male facial appearance in humans. Evolution and Human Behavior, 25, 229 241. doi:10.1016/j. evolhumbehav.2004.04.003 Pound, N., Penton-Voak, I. S., & Surridge, A. K. (2009). Testosterone responses to competition in men are related to facial masculinity. Proceedings of the Royal Society B: Biological Sciences, 276, 153 159. doi:10.1098/rspb.2008.0990 Rhodes, G., Halberstadt, J., & Brajkovich, G. (2001). Generalization of mere exposure effects to averaged composite faces. Social Cognition, 19, 57 70. doi:10.1521/soco.19.1.57.18961 Rule, N. O., & Ambady, N. (2008). The face of success: inferences from chief executive officers appearance predict company profits. Psychological Science, 19, 109 111. doi:10.1111/j.1467-9280.2008.02054 Rule, N. O., & Ambady, N. (2011). Face and fortune: Inferences of personality from managing partners faces predict their law firms financial success. The Leadership Quarterly, 22, 690 696. doi:10.1016/j.leaqua.2011.05.009 Short, L. A., Mondloch, C. J., McCormick, C. M., Carre, J. M., Ma, R., Fu, G., & Lee, K. (2012). Detection of propensity for aggression based on facial structure irrespective of face race. Evolution and Human Behavior, 33, 121 129. doi:10.1016/j.evolhumbehav.2011.07.002 Stirrat, M., & Perrett, D. I. (2010). Valid facial cues to cooperation and trust: male facial width and trustworthiness. Psychological Science, 21, 349 354. doi:10.1177/0956797610362647 Stirrat, M., & Perrett, D. (2012). Face structure predicts cooperation: men with wider faces are more generous to their in-group when out-group competition is salient. Psychological science, 23, 718 722. doi:10.1177/0956797611435133
Todorov, A., Said, C. P., & Verosky, S. C. (2011). Personality impressions from facial appearance. In A. Calder, J. V. Haxby, M. Johnson & G. Rhodes (Eds.), Handbook of Face Perception (pp. 631 652). Oxford: Oxford University Press. Verdonck, A., Gaethofs, M., Carels, C., & de Zegher, F. (1999). Effect of low-dose testosterone treatment on craniofacial growth in boys with delayed puberty. European Journal of Orthodontics, 21, 137 143. doi:10.1093/ejo/21.2.137 Weston, E. M., Friday, A. E., & Lio, P. (2007). Biometric evidence that sexual selection has shaped the hominin face. PLoS ONE, 2, e710. doi:10.1371/journal.pone.0000710 Wong, E. M., Ormiston, M. E., & Haselhuhn, M. P. (2011). A face only an investor could love: CEOs facial structure predicts their firms financial performance. Psychological Science, 22, 1478 1483. doi:10.1177/0956797611418838 Received 14 November 2012; revised version received 5 March 2013 Faces of business leaders 161