Zebrowitz Bronstad Lee 2007

Social Cognition, Vol. 25, No. 2,2007, pp.
306-338
THE CONTRIBUTION OF FACE FAMILIARITY TO INGROUP FAVORITISM AND STEREOTYPING

Leslie A. Zebrowitz and P. Matthew Bronstad Brandeis University Hoon Koo Lee Yonsei University
The familiar face overgeneralization hypothesis holds that an own-race positivity bias is, in part, a perceptual by-product of reactions to familiar people versus unfamiliar-looking strangers. Because prototypical facial structure varies across racial groups and communities are often racially segregated, strangers from one's own racial group should appear more familiar than strangers from a different racial group, contributing to ingroup favoritism and negative outgroup stereotypes. As predicted, the lower familiarity of own- than other-race faces mediated Koreans' and White Americans' ingroup favoritism in Study 1 and Black and White Americans' ingroup favoritism in Study 2. Lower familiarity of other-race faces also mediated negative stereotypes of other-race faces and partially suppressed positive ones, with familiarity effects confined to affectively valenced stereotypes. The results suggest that the unfamiliarity of other-race faces contributes not only to ingroup favoritism but also to a dual-process stereotyping in which both cultural beliefs and negative affective reactions to unfamiliarity make a contribution.
The authors thank Irene Blair for sharing her images of Black male faces with us, Judy Hall for providing assistance and facilities for recruiting and running Black participants, Ryeo-Jin Ko for collecting data in Korea, Jeong-Sook Lee and So Young Goode for translating the experimental materials into Korean, Aviva Androphy, Shayna Skelley, Naomi Skop, Rebecca Wienerman, and Glenn Landauer for their help finding, scanning, and standardizing facial images and collecting data in the United States, and Masako Kikuchi for coordinating the foregoing tasks. This work was supported by NIH Grants MH066836 and K02MH72603 to the first author. Requests for reprints should be sent to Professor Leslie A. Zebrowitz, Department of Psychology, MS 062, Brandeis University, Waltham, MA 02454, e-mail: zebrowitz@brandeis.edu. Correspondence should be addressed to Leslie A. Zebrowitz, Brandeis University, Department of Psychology, MS 062, Waltham, MA 02454. Email: zebrowitz@brandeis.edu. 306
FACE FAMILIARITY, INGROUP FAVORITISM, STEREOTYPING
307
The variable of physical appearance has remained at the periphery of psychology, with the notable exceptions of research on attractiveness (cf, Hatfield & Sprecher, 1986; Eagly, Ashmore, Makhijard, & Longo, 1991; Feingold, 1992; Langlois, Kalakanis, Rubenstein, Larson, Hallam, & Smoot, 2000; Rhodes & Zebrowitz, 2002) and babyfaceness (cf, Montepare & Zebrowitz, 1998; Zebrowitz, 1997), The scant attention to appearance qualities is particularly surprising in the domains of prejudice and stereotyping, since many stereotyped groups can be differentiated by their appearance. Ethnic and racial groups look different from one another, as do the elderly and the young, the disabled and the able-bodied. Some groups are stereotyped solely on the basis of their appearanceattractive people, obese people, short people, redheads. In the dominant model of stereotyping, appearance indirectly leads to stereotypical attributions, Perceivers use appearance features to make a racial or other categorization of an individual; once the classification is made, stereotype-relevant knowledge is activated and applied to the individual. In contrast, the face overgeneralization hypotheses proposed by Zebrowitz (1996,1997,2004) hold that appearance can have a direct effect on prejudice and stereotyping. Grounded in an ecological approach to perception (Gibson, 1966,1979; McArthur & Baron, 1983), these hypotheses hold that socially or evolutionarily adaptive responses to certain facial attributes are overgeneralized to individuals whose facial qualities merely resemble those with functional significance. The purpose of the present research was to investigate the contribution of familiar face overgeneralization to racial prejudice and stereotyping. The familiar face overgeneralization (FFO) hypothesis holds that the utility of differentiating known individuals from strangers has produced a tendency for responses to strangers to vary as a function of their resemblance to known individuals (Zebrowitz, 1996,1997, 2004), Such an effect is consistent with the exemplar-based model of social judgment, which holds that representations of specific individuals influence judgments about similar persons and groups (cf, Andersen & Berk, 1998; Smith & Zarate, 1992), Whereas that model considers many determinants of similarity, FFO focuses on perceptual determinants. Consistent with this hypothesis, reactions to people do depend on their facial resemblance to known others. To give just one example, peers appeared more trustworthy when their photographs were morphed so they resembled the perceiver (DeBruine, 2002; see also Hill, Lewicki, Czyzewska, & Schuller, 1990; Lewicki, 1985), Such effects exemplify what has been called episodic familiarity (Peskin & Newell, 2004), In addition to such idiosyncratic effects of FFO, a broader consequence is that prejudiced responses to strangers of another race may vary with the familiarity of their appearance. Such effects exemplify what has been called general familiarity (Peskin & Newell, 2004), This type of familiarity is illustrated by the finding that faces closer to the population average are rated as more familiar even when they have never been seen before (Langlois, Roggman, & Musselman, 1994), The present study focused on general familiarity by
308
ZEBROWITZ, BRONSTAD, AND LEE
examining reactions to faces as a function of mean ratings of their familiarity across multiple perceivers of the same race. There are several reasons to expect FFO to contribute to racial prejudice and stereotyping. First, prototypical facial structure varies across racial groups,^ Second, communities are often racially segregated, Consequently, strangers from one's own racial group should appear more familiar than strangers from a different racial group because they are more similar to a known facial prototype. Consistent with this argument, 6-month-old infants attended to a stranger's face that was an averaged prototype of known faces as if it were familiar (Rubenstein, Kalakanis, & Langlois, 1999), and, as previously mentioned, faces closer to the population average are rated as more familiar even when they have never been seen before (Langlois et al,, 1994), Finally, since it is well established that people prefer faces that have become familiar through repeated exposure (e,g,, Bornstein, 1989, 1993; Hamm, Baum, & Nikels, 1975; Peskin & Newell, 2004; Rhodes, Halberstadt, & Brajkovich, 2001; Zajonc, 1968), the greater familiarity of faces of strangers from one's own racial group should yield a preference for those faces. Indeed, the enhanced liking of known faces that results from exposure can generalize to similar-looking strangers (Rhodes et al,, 2001), and 3-month-old infants show a visual preference for own-race faces only if they are living in a racially segregated environment(Bar-Haim, Ziv, Lamy, & Hodes, 2006), Although ingroup preferences also have been well documented in adults (cf, Allen, 1996; Brewer, 1979; Cottrell & Neuberg, 2005; Dovidio & Gaertner, 1993; Fiske, 1998), they have not been linked empirically to the familiarity of facial appearance. Recent research examining White judges' responses to faces varying in the degree to which they resemble a prototypical Black face is consistent with the hypothesis that familiarity may contribute to own-race bias, Livingston and Brewer (2002) found that highly prototypical Black faces presented to White perceivers primed faster reaction times to stereotype-irrelevant negative nouns (e,g,, poison, despair) than did low prototypical Black faces or White faces, and they concluded that associations to facial primes reflect affective responses to the racial prototypicality of the faces' perceptual qualities rather than a conceptual evaluation of the faces' racial categories. Similarly, Blair, Judd, Sadler, and Jenkins (2002) found that White judges attributed more negative traits to Black or White faces rated higher in "Afrocentricism," Eberhardt, Goff, Purdie, and Davies (2004) further found
1, It should be acknowledged that anthropologists and biologists question the validity of race as a scientific concepf(e,g,, Lewontin, 2001), Nevertheless, race is a widely accepted concept in folk psychology (Zuckerman, 1990). While recognizing that clear decisions on category membership are problematic, we nevertheless use the "fuzzy" category system of racial groups and the terms White, Korean, and Black to denote the physical appearance of target faces and the group identification of perceivers in this research.
FACE FAMILIARITY, INCROUP FAVORITISM, STEREOTYPING
309
that Black faces rated high in stereotypicality of appearance were more likely to be perceived as criminal by police officers than Black faces rated low in stereotypicality. And Blair, Judd, and Chapleau (2004) demonstrated that inmates with more Afrocentric features received harsher sentences than those with less Afrocentric features, independent of their actual race. The results of all of these studies as well as evidence for perceivers' implicit sensitivity to different degrees of Asian/Caucasian appearance in morphed faces (Locke, Macrae, & Eaton, 2005) indicate that responses to people do not have sharp racial category boundaries, but instead are inherently graded, depending upon race-related perceptual cues. There is an explanation besides familiarity for the finding that Afrocentric features contributed to White judges' negative responses to faces in the foregoing studies. In particular, these negative reactions could reflect a more general prejudice against certain appearance qualities. The student raters and court judges who contributed data for the research in the foregoing studies were almost exclusively White. Thus, we do not know whether Black Americans would have reacted similarly to faces with more typically African American features, thereby disconfirming a familiarity effect. Indeed, there is evidence to indicate that negative affective associations to dark skin color are shared by people of varied ethnic and racial backgrounds (cf. Gergenl967; Hill, 2002; Maddox, 2004; Maddox & Gray, 2002; Uhlmann, Dasgupta, Elgueta, Greenwald, & Swanson, 2002; van den Berghe & Frost, 1986), and there are automatic associations between object darkness and negative valence (Meier, Robinson, & Clore, 2004). To differentiate main effects of facial qualities from effects that vary with their familiarity to perceivers, the present research examined reactions to faces by own-race and other-race judges. Study 1 investigated reactions to White and Korean faces by White Americans and Koreans. An advantage of studying these groups is that there is likely to be more symmetry in the unfamiliarity of other-race faces for Koreans and Whites compared to Blacks and Whites. Black judges are a racial minority in the United States, and they report more familiarity with outgroup members than do White judges (e.g., Judd, Park, Ryan, Brauer, & Kraus, 1995). Another advantage to studying Korean and White judges is that there are positive stereotypes of Asians among American students. Study 2 investigated reactions to White and Black faces by White Americans and Black Americans.
STUDY 1
Study 1 tested the FFO hypothesis that variations in the familiarity of appearance of Korean and White faces to Korean and White judges contribute to ingroup favoritism and race-related stereotypes. Given the previously cited evidence for ingroup preferences, we predicted that own-race faces would be judged more likeable than other-race faces by both Korean and
310
White raters. We further expected that ingroup favoritism for both groups would be mediated by the lower familiarity of other-race faces. Our predictions regarding stereotypes take into account the fact that cultural beliefs can make a significant contribution independently of any effects of familiarity. In U.S. culture, Asians are stereotyped by White Americans as law abiding, competent, and socially reserved, with the former two stereotypes indicative of what has been dubbed the model minority stereotype (Ho & Jackson, 2001; Kawai, 2005; Lin, Kwan, Cheung, & Fiske, 2005; Niemann, Jennings, Rozelle, Baxter, & Sullivan, 1994). Thus, independent of reactions to an unfamiliar appearance. White American raters may perceive Asian faces as less dangerous, more competent, and more reserved than White faces. Nevertheless, consistent with the FFO hypothesis, controlling familiarity should strengthen White Americans' positive stereotypes of Korean faces as less dangerous and more competent by removing the negative influence of an unfamiliar appearance. On the other hand, controlling familiarity should weaken negative stereotypes of Korean faces as more reserved. Although we are not aware of any research that has investigated race stereotypes held by Koreans, we expected their ratings to parallel those of White Americans either because they share the cultural stereotypes or because of an ingroup positivity biasnot only are high competence and low danger likely to be viewed positively by Koreans, but so is social reserve (Kim, Atkinson, & Yang, 1999). According to the FFO hypothesis, controlling familiarity should weaken Koreans' negative stereotypes of other-race White faces by removing the negative effects on impressions of an unfamiliar appearance. METHOD
Raters
Forty White American and 40 Korean college undergraduates participated. All White raters received course credit; all Korean raters received the equivalent of $10 in their local currency. Equal numbers of male and female raters from each racial group rated either Korean and White men's or women's traits and general appearance qualities. An additional 39 White American undergraduates and 48 Koreans rated the emotional expression of the faces.^
2. These participants also made several other ratings of the faces for a separate experiment including how African, White, and Asian they looked, their skin tone, and their femininity or masculinity.
311
Facial Stimuli There were 60 White face images and 60 Korean face images. White facial images were selected from four different databases. University of Stirling PICS database (http://pics.psych.stir.ac.uk/), the AR face database (Martinez & Benavente, 1998), NimStim,'' and two yearbooks (one high school and one university yearbook). Korean facial images were selected randomly from a Korean university yearbook. We used four criteria for image selection: neutral expression, no head tilt, no eyeglasses, and no beards. Images were edited in Photoshop to reraove jewelry and blemishes. For some images, minor changes were made to the naso-labial folds and the corners of the mouth to make the facial expression appear more neutral. All faces were shown in color against an identical beige background. Face Ratings Likeability and Stereotypes. Faces were rated on 7-point scales designed to assess ingroup favoritismLtfeaJ^/e (not at all likeable-very likeable)as well as race stereotypes: Intelligent (unintelligent-intelligent). Hardworking (not very hardworking-very hardworking). Hostile (not at all hostile-very hostile). Trustworthy (not at all trustworthy-very trustworthy), and Reserved (reserved / quiet-expressive /loud), Appearance. Faces were rated on 7-point scales designed to assess appearance qualities that may contribute to likeability and stereotypes: Familiar (not at all familiar-very familiar), Babyfaced (babyfaced-maturefaced). Attractive (unattractive-attractive). The faces also were rated on two emotion expression scales. First, participants judged whether each face had a neutral expression. If they answered "no," they then specified which emotion that face reflected, choosing among happy, sad, angry, fearful, disgust, and surprise. Procedure White and Korean raters viewed images and input responses on Pentium 4 personal computers with Windows XP and 19" CRT displays. Raters sat within 36" of the monitors, MediaLab 2004,2,10arvis, 2004) was used to display images and collect ratings. Faces were displayed for 5 seconds, or until a rating was made, and a random order of all Korean and White faces was rated on one scale before raters proceeded to the next scale in order to minimize carryover effects in ratings from one scale to another. A different ran3, Development of the MacBrain Face Stimulus Set was overseen by Nim Tottenham and supported by the John D. and Catherine T. MacArthur Foundation Research Network on Early Experience and Brain Development. Please contact Nim Tottenham at tott0006@tc.umn.edu for more information concerning the stimulus set.
312
dom order of faces was presented for each rating scale. All raters rated familiarity first followed by likeability, with attractiveness and babyfaceness rated after the traits. The order of trait ratings was Intelligent, Reserved, Hostile, Trustworthy, Hardworking for one group of raters and the reverse order for a second group. Familiarity was always rated first because we were interested in the relative familiarity of own- and other-race strangers who had never before been seen. Raters received the following additional instructions for familiarity ratings: "Sometimes a face looks familiar to us even though we've never seen the person before. Although you haven't seen any of the following faces before, please rate how familiar each one looks to you." Korean Translation All rating scales were translated into Korean by a native Korean speaker, A second native Korean speaker translated the Korean back into English, and these results were compared to the original English-language scales. For any discrepancies, the native Korean speakers were consulted to retranslate the scales so that the meaning in Korean was as close as possible to the meaning in English, RESULTS Reliability of Measures and Trait Composites Cronbach alphas for ratings of familiar and likeable were a=,94 and a = ,93 for Korean raters and a = ,95 and a = ,90 for White raters, Cronbach alphas for attractiveness and babyfaceness were a = ,90 and a = .93 for Korean raters and a = .87 and a = ,90 for White raters. We created two a priori trait composite variables related to the U,S, stereotype of Asians as a law abiding and competent "model minority," Ratings of hostile and trustworthy (reverse scored) were summed to create a danger composite, which had alpha coefficients of .90 and ,92 for Korean and White raters, respectively. Ratings of hardworking and intelligent were summed to create a competence composite, which had alpha coefficients of .93 and ,91 for Korean and White raters, respectively. Ratings of the individual trait reserved had alpha coefficients of ,81 and .87 for Korean and White raters, respectively. Confirmatory principal component factor analyses performed on the trait ratings with varimax rotation and results constrained to 3 factors supported the a priori trait composites. The results for four separate analyses performed for Korean and White raters within each race of face yielded a danger factor comprised of hostile and trustworthy, a competence factor comprised of hardworking and intelligent, and a separate reserved factor.
313
The high inter-rater reliabilities for the appearance ratings and trait composites justified data analyses utilizing mean ratings for each face across White judges and across Korean judges as the unit of analysis. These analyses tested the hypothesis that some faces look more familiar than other faces across all perceivers of a given race and that these variations account for part of the effect of face race on likeability and stereotypes. Emotional Appearance The majority of faces were perceived as showing no emotion by all participants (68% of White faces and 66% of Korean faces). Happiness was the most common emotion detected, identified, on average, in 11% of the White faces and 16% of the Korean faces. We therefore created a smile score for each face by dividing the number of times it had been identified as happy by the total number of participants. We did this for participants of each race, thus generating two smile scores per face. We did not control for other emotions, since they were even more rarely detected, identified on average in only 4% of the faces, Ingroup Favoritism and Stereotypes We performed separate 2 (face race) x 2 (rater race) ANOVAs on mean ratings of the faces on likeability and the trait scales. Means and f values are shown in Table 1, A tendency for Korean faces to be more likeable was qualified by the predicted face race x rater race interaction effect, indicating that own-race faces were judged as more likeable. Planned comparisons revealed that the simple effect for ingroup favoritism was significant for Korean raters, f(118) = 5,07, p < ,001, but not for White raters, t < 1. Consistent with the prediction of shared stereotypes, both White and Korean raters perceived Korean faces to be lower than White ones on the danger composite, and higher on the competence composite and the trait reserved. Significant face race x rater race interaction effects for the competence composite and reserved ratings revealed stronger effects for White than for Korean raters, although the effects were significant for both groups, respective fs(118) = 11,03 and 3,52 for competence and 9.37 and 2,60, for reserved, all ps < ,01. The face race x rater race interaction effect was not significant for the danger composite. Although we had no predictions for main effects of rater race. White raters gave higher ratings on likeability and competence and lower ratings on danger.
314
3
00 1-^ -* eN IN
Ov CN
ON T H
^ in
(N
(N CN
z
2 S
00 o eN o en en NO
en o
CO
o o LT)
(N
CO
CO
CO
eN
CO
2 M>
2 >^ fi "ra 'i < CQ cS
315
Does Face Familiarity Mediate Ingroup Favoritism and Stereotypes? As a first step in examining face familiarity as a mediator of ingroup favoritism and stereotypes, we performed separate 2 (face race) x 2 (rater race) ANOVAs on mean ratings of the faces on familiarity, as well as the other appearance qualities of attractiveness, smiling, and babyfaceness that also could affect likeability and trait ratings. Means and F values are shown in Table 1, Consistent with the hypothesis that face familiarity mediates ingroup favoritism, familiarity ratings paralleled the likeability ratings, A tendency for Korean faces to be rated more familiar was qualified by a highly significant rater race x face race interaction effect indicating that faces of own-race strangers were judged to look more familiar than those of other-race strangers. Planned comparisons revealed that this effect was significant for both Korean and White raters, respective ts (118) = 11,92 and 9,16, bothps <.OO1, In addition. White faces were less familiar to Korean than White raters, t(l 18) = 6,51, p < ,001, and Korean faces were less familiar to White than Korean raters, f(118) = 15,99, p < ,001, revealing the expected symmetry in other-race familiarity for Koreans and White Americans, Own-race faces were not perceived as more attractive, more babyfaced, or more smiKng, There was also no significant main effect of face race on attractiveness or smile ratings. On the other hand, a significant main effect for face race on babyfaceness revealed that Korean faces were perceived by both Korean and White raters as more babyfaced. Although we had no predictions for main effects of rater race. White raters gave higher ratings on familiarity and lower ratings on babyfaceness. Multiple regression analyses were performed within rater race to determine whether face familiarity mediated ingroup favoritism and stereotypes independently of other appearance qualities. Face race was entered into the first step of the regression analyses along with face sex, smile, babyfaceness, and attractiveness, each of which is likely to influence judges' ratings. Controlling attractiveness was particularly important, as faces of more attractive strangers are more likely to be perceived as familiar (Corneille, Monin, & Pleyers, 2005), Familiarity was entered at the second step, and we followed the procedure outlined by Baron and Kenny (1986) to determine whether it mediated the likeability or stereotype effects. Table 2 and Figure 1 provide the results of the regression analyses. Likeable. Faces higher in smiling or attractiveness were judged more likeable by both Korean and White raters. In addition. White raters gave higher likeability ratings to faces higher in babyfaceness,* With these variables controlled, own-race faces also were rated as more likeable, not only by Korean
4, There were few effects of face sex on ratings, probably because sex of face was varied between subjects in Study 1 and blocked in Study 2 so that faces were judged in comparison with others of the same sex (cf, Biernat & Manis, 1994),
316
ZEBROWiTZ, BRONSTAD, AND LEE
rH
r-<
S g
u o
CN CO
01
60
tu
en N
n (N -^ tN in T-H
en
O (N
en
T-H CO
en
O CO
00 CN
(N
^D t<
CN
in
p cD
V .D
CO CO 60
> 2
c
C O
p g
C N
o O
re
in
00
iJO
re
C
OJ
S
y, 'o
^^
^ O
CN
T I rH
CO
in in
CO O
( I O
I
ro
u C O
&
u _^ C O
Fa:
liai
FACE FAMiLiARITY, iNGROUP FAVORiTiSM, STEREOTYPiNG
317
(Koreans Higher)
1
n Face Race, Stepi
0.8 0.6 0.4 0.2
Face Race, Step2, Controlling Familiarity
m
-o
Si
-0.6 (Whites Higher)
XI
CO
m c C D Q
O)
"c
a> a> m ro a)
o O
O) ro Q
CD
S S.
E o O
CD
cn a>
Korean Raters
mediation by famiiiarity,*p<.05,+ p<.10
White Raters
FIGXJRE 1. Race of face predicting Korean and White trait ratings. Smiling, sex, attractiveness, and babyface controlled at Step 1; familiarity added as a control variable at Step 2. Positive standardized betas indicate higher ratings given to Korean faces.
raters but also by White raters for whom the likeability effect had not reached sigruficance in the ANOVA where the other appearance variables were not controlled. As predicted, these ingroup favoritism effects lost significance for both White and Korean raters when familiarity was controlled in Step 2 of the regressions. Indeed, controlling familiarity produced a nonsignificant reversal of the ingroup favoritism shown in likeable ratings by White judges. Sobel tests demonstrated that both of these mediation effects were significant, zs = 4.82 and 5.68, both ps <.OO1, for White and Korean raters, respectively. Danger Composite. Faces higher in attractiveness or babyfaceness were rated lower on the danger composite by both Korean and White raters. In addition, Korean raters gave lower danger ratings to faces higher in smiling. With these variables controlled, face race also predicted danger stereotypes.
318
with Korean faces receiving lower scores on the danger composite from both Korean and White raters. For Korean raters, this stereotype lost significance when familiarity was controlled at Step 2 of the regression. Familiarity met the criteria for mediation, and the Sobel test indicated that familiarity was a significant mediator of Koreans' stereotypes of Korean faces as less dangerous, z = 4.21, p <001. Familiarity also met the criteria for mediation of danger stereotypes by White raters. In this case, controlling familiarity increased their tendency to view Korean faces as less dangerous, as predicted for a positive other-race stereotype. The Sobel test revealed that this change was significant, z = 2.25, p = .02, indicating that the tendency for White raters to stereotype Korean faces as less dangerous than White ones was partially suppressed by Korean faces' lower familiarity. Competence Composite. Faces higher in attractiveness were rated higher on the competence composite by both Korean and White raters, although the effect was only marginal for White raters. With other appearance variables controlled, face race also predicted competence stereotypes, with Korean faces receiving higher scores on the competence composite from both Korean and White raters, as predicted. For Korean raters, this stereotype lost significance when familiarity was controlled at Step 2 of the regression. Familiarity met the criteria for mediation, and the Sobel test indicated that familiarity was a marginally significant mediator of Koreans' tendency to stereotype Korean faces as higher in competence, z = 1.78, p = .08. Familiarity also met the criteria for mediation of competence stereotypes by White raters. As for danger stereotypes, controlling familiarity increased their tendency to stereotype Korean faces as higher in competence, as predicted. The Sobel test revealed that this change was significant, z= 3.70, p < .001, indicating that the tendency for White raters to stereotype Korean faces higher in competence was partially suppressed by their lower familiarity. Reserved. Faces rated higher in smiling were rated as less reserved by both Korean and White raters. In addition. White raters gave higher reserve ratings to faces lower in attractiveness or higher in babyfaceness. With other appearance variables controlled, face race also predicted both Korean and White raters' reserved stereotypes, with Korean faces receiving higher scores from both groups. When familiarity was entered at Step 2, the effect of race on the reserved stereotype lost significance for Korean raters. Familiarity met the criteria for mediation, and the Sobel test was marginally significant, z = 1.74, p = .08. On the other hand, familiarity did not meet the criteria for mediation of the race effect on White judges' reserved stereotype. DISCUSSION The results of Study 1 are consistent with previous research demonstrating ingroup favoritism and race stereotypes. Both Korean and White American judges found faces of their own race more likeable when attractiveness.
FACE FAMiLiARiTY, iNGROUP FAVORiTiSM, STEREOTYPiNG
319
TABLE 3. Partial Correlations Between Likeability and Traits Korean Raters White Faces Danger Competent ^31^ .27^ Korean Faces ^ .23^ White Raters White Faces ^ .63^ Korean Faces ^ .32^
R e s e r v e d^^ ^ 3 ^ ^^ 5 - ^ 0 ^ 6 - 5 5
Nofe. Partial correlations control face sex, smile, attractiveness, and babyfaceness. p<.10; p<.05; p< .001.
babyfaceness, and smiling were controlled. Both groups also shared similar stereotypes, rating Koreans as more reserved than Whites, higher on a competence trait composite that included hardworking and intelligent, and lower on a danger trait composite that included hostile and vmtrustworthy. These stereotypes are consistent with other evidence that Asians are stereotyped, at least by Americans, as the model minoritycompetent and law abidingas well as socially reserved (Kawai, 2005; Lin et al., 2005). The high reliabilities and own-race bias in familiarity ratings by both Korean and White Americans indicated that these judgments reflected at least in part an implicit comparison between each face and a shared facial prototype rather than merely reflecting a rater's idiosyncratic perceptionsakin to "general" rather than "episodic" familiarity in Peskin and Newell's (2004) terms. As predicted, variations in the average familiarity of faces across all raters of a given race made a sigruficant contribution to ingroup favoritism and race stereotypes. When face familiarity was statistically controlled, Koreans' ingroup favoritism in likeability ratings and their negative stereotypes of White faces as higher in danger and lower in competence and reserve all lost significance. A negative impact of an urvfamiliar appearance was similarly shown by White Americans. Their ingroup favoritism lost significance when face familiarity was statistically controlled. Whereas Koreans' race stereotypes could be completely explained by the lesser familiarity of White faces, the results for White raters revealed effects of both the model minority stereotype and familiarity. White Americans' positive stereotypes of Korean faces as lower in danger and higher in competence were significantly increased when face familiarity was controlled, indicating that the greater imfamiliarity of Korean faces acted to suppress the positive "model minority" stereotype. The finding that perceived familiarity did not influence White raters' stereotypes of Koreans as more reserved cannot be accounted for by the FFO hypothesis, which predicts that any negative outgroup stereotype should be exacerbated by negative reactions to outgroup members' unfamiliar appearance. The findings of Lin et al. (2005) indicate that White Americans' stereotype of Asians as socially reserved is indeed a negative one, inasmuch
320
as the perceived low sociability of Asians predicted White Americans' negative reactions to Asian American individuals. However, the Lin et al. item "tends to be shy and quiet," which is comparable to our measure of reserved/quiet did not load as highly on their "unsociable" composite as some of their other items, such as the importance of and time spent socializing. It may be that our measure is more evaluatively neutral than those items. Since face familiarity should have stronger effects on social judgments that are more affectively loaded, this could explain its failure to influence White raters' stereotype of Asians as more socially reserved. The finding that face familiarity did mediate Koreans' stereotypes of Whites as less reserved suggests that positive affective reactions to social reserve in Asian culture (Kim, Atkinson, & Yang, 1999) is stronger than the negative reactions in American culture. As a proxy for assessing affective reactions to social reserve, we correlated it with likeability ratings, partialling out the control variables entered at Step 1 of the regression (face sex, attractiveness, babyfaceness, smiling). As shown in Table 3, there was no significant relationship between social reserve and likeability for White raters' judging White faces or Korean faces. On the other hand, there was a marginally significant positive relationship for Korean raters judging White faces, albeit not Korean faces. In contrast, the relationships between likeability and the danger and competence composites were significant for both White and Korean raters judging faces of either race, although the relationship between likeability and competence was only marginal for Koreans judging Korean faces. These results indicate that whereas competence and danger are affectively valenced for both Korean and White raters, reserve is affectively valenced only for Korean raters. Since face unfamiliarity is hypothesized to create negative affective reactions, it makes sense that statistically partialling out these effects would influence only those judgments that are also affectively loaded. The effects of familiarity held true over and above effects of smiling, babyfaceness, and attractiveness, each of which also had a predictable influence on various impressions, including similar effects of babyfaceness and attractiveness across Korean and White raters and faces (Zebrowitz, Montepare, & Lee, 1993). The tendency for faces higher in smiling to be rated as more likeable, less dangerous, and less reserved is consistent with the association of smiling with high affiliation (Knutson, 1996; Montepare & Dobish, 2003). The tendency for faces higher in babyfaceness to be rated as more likeable and less dangerous is consistent with the well-documented babyface overgeneralization effecf(Montepare & Zebrowitz, 1998; Zebrowitz, 1997). The tendency for faces higher in attractiveness to be rated as more likeable and competent as well as less dangerous and reserved is consistent with the well-documented attractiveness halo effecf(e.g., Eagly et al., 1991; Feingold, 1992; Langlois et al., 2000). The dissociation between U.S. raters' prejudice and stereotypes is also noteworthy. Despite the fact that they showed ingroup favoritism in
321
likeability judgments once attractiveness, babyfaceness, and smiling were controlled, U.S. raters nevertheless had positive stereotypes of Koreans, as revealed in impressions of their lower hostility and their higher trustworthiness, intelligence, and industry. One might suggest that the likeability judgments reflected higher valuation of other traits that Asians are perceived to lack, such as indices of sociability not assessed in our study (cf. Lin et al., 2005). Another explanation is that the affective processes that guide judgments of likeability are separable from the more cognitive processes that guide assessments of personality traits. This interpretation is consistent with Livingston & Brewer's (2002) finding that priming White perceivers with highly prototypical Black faces facilitated negative affective responses (automatic prejudice) but not semantic responses (automatic stereotyping), as well as with other evidence for the differentiation of affective and cognitive responses in reactions toward outgroups as well as in other domains (e.g., Eagly, Mladinic, & Otto, 1994; Jussim, Nelson, Manis, & Soffin, 1995; Stangor, SulUvan, & Ford, 1991; Tropp & Pettigrew, 2005; Zajonc, 2000). Although stereotypes may often be heavily influenced by cognitive processes, affective processes also have an impact as evidenced by our finding that negative reactions to unfamiliarity suppressed White raters' positive outgroup stereotypes and mediated Korean raters' negative ones.
STUDY 2
Study 2 tested the FFO hypothesis that variations in the familiarity of appearance of Black and White faces to Black and White American judges contribute to ingroup favoritism and race-related stereotypes. Whereas White Americans stereotype Asians as law-abiding, competent, and socially reserved, their stereotypes of Black Americans are precisely the opposite. For example, Devine & Elliot(1995) found that the traits "criminal," "low intelligence," and "loud' were endorsed as describing Blacks by at least a third of White participants. Similarly, Niemann, Jennings, Rozelle, Baxter, & Sullivan (1994) foimd that "antagorustic," "ambitionless," and "speak loudly" were among the top-ranked traits given for African Americans by White American college students. There is less research examining race stereotypes by Black Americans, and the existing findings are inconsistent. Early studies found that Blacks and Whites tended to have similar stereotypes (e.g., Bayton, 1941). More recent research suggests that Blacks, Uke Whites, give more positive ratings to ingroup members (e.g., Allen, 1996; Judd et al., 1995). (See Schneider, 2004 for a comprehensive review.) Thus, it may be that both White and Black Americans will judge own-race faces as higher in likeability and competence and lower in danger. As in Study 1, we predicted that ingroup favoritism in likeability judgments would be mediated by the familiarity of the faces. According to the FFO hypothesis, stereotyped impressions also should be at least partially mediated by face familiarity. More specifically, statistically controlling the lower familiarity
322
of other-race faces should weaken negative other-race stereotypes and strengthen positive ones. METHOD The method employed in Study 2 was identical to Study 1 except as indicated.
Raters
Forty White American and 13 Black American college undergraduates rated the faces on trait and general appearance scales. White raters received course credit for rating either Black and White male faces or Black and White female faces. Black raters received $25 for rating both male and female faces in two counterbalanced sessions. Facial Stimuli There were 60 White face images and 60 Black face images. White facial images were identical to those used in Study 1. The facial images of Black males were selected from a set of 84 faces that had been used in previous research by Blair, Judd, Sadler, & Jenkins (2002). These faces had been selected by the authors from various high school yearbooks, with the constraint that they all have neutral facial expressions and no visible accessories. The images of Black females were selected from the web site http://americansingles.com in a search for Black females aged 18-25. The criteria for image selection were the same as in Study 1. Dependent Measures The dependent measures were identical to Study 1 with two exceptions. First, we added the trait scale Athletic (not at all athletic-very athletic) because it is a dimension on which there are strong cultural stereotypes regarding Blacks and Whites (e.g., Devine & Elliot, 1995). Second, we used White participants' smile ratings in all analyses because these ratings were collected in another experiment for which we did not have Black participants. Since smile ratings by Korean and White raters in Study 1 were correlated .90 for White faces and .85 for Korean faces, it seems that smile ratings by White raters are a reasonable proxy for those that would be provided by Black raters.
FACE FAMILIARITY, INCROUP FAVORITISM, STEREOTYPINC
323
Procedure Black raters viewed images and input responses on a Dell Latitude D800 computer with a 15,4" screen. White participants rated images of either men or women, while Black participants rated images of both men and women in two separate sessions with the order of face sex counterbalanced across sessions, RESULTS Reliability of Measures and Trait Composites Cronbach alphas for ratings of familiar and likeable were a = ,72 and a = ,80 for Black raters and a = ,73 and a = ,84 for White raters, Cronbach alphas for attractiveness and babyfaceness were a = ,85 and a = ,88 for Black raters, and a = ,92 and a = ,88 for White raters. As in Study 1, two a priori trait composite variables were created. The danger composite had alpha coefficients of ,74 and ,85 for Black and White raters, respectively. The competence composite had alpha coefficients of ,58 and ,87 for Black and White raters, respectively. Although a higher alpha for Black raters could be obtained by creating a single positive trait composite, summing ratings of hardworking, intelligent, trustworthy, and hostile (reversed), we retained the separate competence and danger composites for consistency with Study 1 and for the interest value of examining stereotypes on these two dimensions. Ratings of reserved and athletic were analyzed separately. As in Study 1, the data analyses utilized mean ratings for each face across judges of each race as the unit of analysis. Emotional Appearance As in Study 1, we created a smile index for each face by dividing the number of times a face had been identified as happy by the number of participants. Like the results for White and Korean faces reported in Study 1, the majority of Black faces (63%) were perceived as showing no emotion by all participants. Happiness was the most common emotion detected, identified, on average, in 18% of the Black faces, Ingroup Favoritism and Stereotypes We performed separate 2 (face race) x 2 (rater race) ANOVAs on mean ratings of the faces on each dependent measure. Means and F values are shown in Table 4, Black faces were rated as significantly more likeable, but this main effect was qualified by the predicted face race x rater race interaction effect.
324
12
^'6N
CO ON CM

73
CN ON
T1
CO
rH
00 00 NO
o
00
to
CN
00
on
CO CO
69
CO
03
o
102
CO
"'co
13
I"
T-H
00
S g
57
07
T1
97
ro
18
IM NO
O CN CN
CO CO
CO
CO
CO
CN
in
00 CO CO
(N
t1
CO
o\ CN
NO ON
rx
ON
IN
rx m CO CN
CM
s
CN in CO
75
61
59
rH
IX
I1
S
C O ra
46
CO CO
37
38
o
CN
CO
CO
CO
5
1^
VO NO
T-H
in
CN NO
ON
in
NO ( "1
00
in
c^
20
85
30
28
in
tN.
30
-a
o
CO
ON
CO
CO
'^
CO CO
CM NO
(N
IN
ON
CN
rNi
tx
NO CM
rH
UH
C ra
O CO
in in CO
<N (N
on
in
CO CO CO
NO
in
ro CN CM
C Q
ikeable anger ompetent eserved thletic amiliar ttractive abyface U U

< ^
0 1
t i n <^
CQ
325
revealing greater likeability of own-race faces. The simple effect for this ingroup favoritism was significant for Black raters, t(118) =7.01, p < .001, whereas there was a nonsignificant trend in the predicted direction for White raters, f(118) = 1.19,p = .24. A face race x rater race interaction effect for ratings on the danger composite also revealed ingroup bias, and this effect was significant for both White raters, t(118) =2.22, p < .03, and Black raters, t(118) = 3.10, p < .01. White faces were rated marginally higher on the competence composite, but this main effect was qualified by a significant face race x rater race interaction, revealing higher competence ratings for own-race faces. However, the simple effect was significant for White raters, t(118) =3.13, p < .01 but not for Black raters, t<l. Consistent with cultural stereotypes, both White and Black raters perceived Black faces to be higher than White ones on the trait "athletic" and lower on the trait "reserved." Although we had no predictions for main effects of rater race. White raters gave higher likeable ratings and marginally lower reserve ratings. Does Face Familiarity Mediate Ingroup Favoritism and Stereotypes? As a first step in examining face familiarity as a mediator of ingroup favoritism and stereotypes, we performed separate 2 (face race) x 2 (rater race) ANOVAs on mean ratings of the faces on familiarity, as well as the other appearance qualities of attractiveness, smiling, and babyfaceness that also could affect likeability and trait ratings. Means and F values are shown in Table 4. Consistent with the hypothesis that face familiarity mediates ingroup favoritism, familiarity ratings paralleled the likeability ratings. A tendency for White faces to be rated more familiar was qualified by a highly sigruficant rater race by face race interaction effect indicating that faces of own-race strangers were judged to look more familiar than those of other-race strangers. Planned comparisons revealed that this effect was significant for both Black and White raters, respective ts (118) = 7.71 and 4.22, ps <.OO1. In addition, as expected from the minority status of Blacks in the U.S., White faces were equally familiar to Black and White raters, t < 1, whereas Black faces were more familiar to Black than White raters, t(118) = 10.69, p < .001. Black faces were rated as more attractive overall, but this main effect was qualified by a significant face race x rater race interaction, revealing that the higher attractiveness of Black faces was significant for Black raters, t(118) = 4.66, p < .001, but not for White raters, t < 1. Black faces were also rated as higher in smiling by all raters, but no different in babyfaceness. Finally, White raters gave higher attractiveness ratings, particularly to White faces, and lower famiUar ratings, particularly to Black faces. Multiple regression analyses were performed within rater race paralleling those in Study 1 to determine whether face familiarity mediated ingroup fa-
326
ZEBROWITZ, BRONSTAD, A N D LEE
(Blaci<s Higher)
u.o D Face Race, Step1

OR\J.\J
Face Race, Step2, Controlling
-amiliarity
(D
0.4 *
-1
0 ra 0.2 -
'S m
~]
0-0.2 -
u
^ _
O)
L|'
^. .2
o O
1'
thieti
o
I *
-U.4 (Whites Higher)

(0 (U .^
1
c
(0
c Q
(0
S.
E
(0
<
C (0
.2
(U
(/}
I " O
o
(D
<
Black Raters
mediation by famiiiarity,*p<.05,+ p<.10
White Raters
FIGURE 2. Race of face predicting Black and White trait ratings. Smiling, sex, attractiveness, and babyface controlled at Step 1; familiarity added as a control variable at Step 2. Positive standardized betas indicate higher ratings given to Black faces.
voritism and race stereotypes independently of other appearance qualities (see Table 5 and Figure 2). Likeable. Faces higher in attractiveness or smiling were rated as more likeable by both Black and White raters. In addition. White raters also judged faces higher in babyfaceness as more likeable. With these appearance variables controlled, own-race faces also were rated as more likeable not only by Black raters, but also by White raters for whom the ingroup favoritism effect was not significant in the ANOVA where the other appearance variables were not controlled. Moreover, as predicted, the ingroup favoritism effect was weakened for both Black and White raters when familiarity was controlled in Step 2 of the regressions. Sobel tests demonstrated significant partial mediation effects for both groups, zs = 3.33 and 2.44, ps < .01 for Black and White raters, respectively.
thieti
erve
1 ^ .
327
(U
p
B o U
e
a.
^ P CO Tf CO CO ^O (N CO CO CO CM * CO
r
C O LO ro
LD CD
00 o
r r
C<^
5.
C O
^
rt
T4
CN
CN
ON T-H
T-H \o
\D CO
O ^
fO (N ro ro 6s. ON "^ ^ rj T ( ^ en
(N '-'
' r
CD o
<* T I
CN o
tI o
CN CN rn (N CN| Ln
g o U <u so
Ln O
fO C O C O
s O
CM IN fS
o O
Csl 00 CO
'
OH
<
C D
ii
328
Danger Composite. Faces higher in attractiveness, babyfaceness, or smiling were rated as less dangerous by both Black and White raters. With these appearance variables controlled, face race also predicted danger stereotypes for White raters, with Black faces receiving higher scores, t = 4.80, p < .001. For Black raters there was a nonsignificant tendency to rate White faces as more dangerous, contrasting with the significant effect in the ANOVA when the greater attractiveness and smiling of Black faces were not controlled. The race stereotype for White raters was weakened when familiarity was controlled at Step 2 of the regression, although it remained significant. Familiarity met the criteria for mediation, and the Sobel test indicated that familiarity was a significant partial mediator of Whites' stereotypes of Black faces as more dangerous, z = 2.38, p < .02. Competence Composite. Faces higher in attractiveness were rated as more competent by both Black and White raters. In addition. White raters gave higher competence ratings to faces higher in smiling or babyfaceness. With these appearance variables controlled, face race also predicted White raters' competence stereotypes, with Black faces receiving lower scores on the competence composite. This race stereotype was weakened when familiarity was controlled at Step 2 of the regression. Familiarity met the criteria for mediation, and the Sobel test indicated that it was a significant partial mediator of Whites' tendency to stereotype Black faces as lower in competence, z =3,28, p < ,01, Face race did not predict Black raters' competence stereotypes in Step 1 of the regression. However, when the lesser familiarity of White faces was statistically controlled at Step 2, Black raters judged them significantly higher in competence than Black faces. The Sobel test revealed that this change was significant, z = 3.00, p < .01, indicating that a tendency for Black raters to stereotype White faces as higher in competence was partially suppressed by their lower familiarity. Reserved. Faces higher in attractiveness or smiling were rated as less reserved by both Black and White raters, while those higher in babyfaceness were rated as more reserved. With these appearance variables controlled, face race also predicted White raters' reserved ratings, with Black faces receiving lower scores. There was a nonsignificant trend in the same direction for Black raters, contrasting with the significant effect in the ANOVA when the greater attractiveness and smiling of Black faces were not controlled. Familiarity did not meet the criteria for mediation of the race effect on White judges' reserved stereotype. Athletic. Faces higher in attractiveness or lower in babyfaceness were rated as more athletic by White raters. With other appearance variables controlled, face race also predicted both Black and White raters' athletic ratings, with Black faces receiving higher scores. Familiarity did not meet the criteria for mediation of the race effect for either group.
329
DISCUSSION Like Study 1, the results of Study 2 are consistent with other evidence for ingroup favoritism and race stereotypes. Both Black and White judges found faces of their own race more likeable when smiling, attractiveness, and babyfaceness were controlled. White judges also revealed predictable stereotypes of Blacks, rating them lower on the competence composite, higher on the danger composite, less reserved, and more athletic, effects that held up when smiling, attractiveness, and babyfaceness were controlled (Devine & Elliot, 1995; Niemann et al,, 1994), Although Black judges perceived Black faces as less dangerous, less reserved, and more athletic than White ones, only the athletic stereotype remained significant when the greater attractiveness and smiling of Black faces were controlled. The failure of Black raters to judge own-race faces as more competent and less dangerous when controlling race differences in attractiveness and smiling appears to contradict other evidence for positive ingroup stereotypes by Black Americans (Allen, 1996; Judd et al,, 1995), The divergent results may be attributable to divergent methodologies, Livingston (2002) found that African Americans' explicit positive evaluations of their own group are not reliably shown in their implicit judgments. Similarly, "political correctness" may have less influence on Black students' ingroup assessments when they respond to specific individuals who vary in race, as in the present study, than when they respond to category labels, as in previous research that showed positive ingroup stereotypes. On the other hand, it is significant that Black raters did find Black faces more attractive, which White raters did not, and with this positive reaction to own-race faces uncontrolled. Black raters also found Black faces less dangerous. As in Study 1, the high reliabilities and own-race bias in familiarity ratings by both Black and White Americans indicated that these judgments reflected at least in part an implicit comparison between each face and a shared facial prototype rather than merely reflecting a rater's idiosyncratic perceptions. Again, variations in the average familiarity of faces across all raters of a given race made a significant contribution to ingroup favoritism and race stereotypes, Ingroup favoritism by both Black and White judges was partially mediated by the greater perceived familiarity of own than other race faces. Consistent with the FFO hypothesis, face familiarity also partially mediated White perceivers' tendency to stereotype Black faces as lower in competence and higher in danger, and it partially suppressed Black raters' tendency to stereotype White faces as higher in competence. Familiarity failed to mediate White raters' stereotype of Black faces as less reserved than White faces and Black and White raters' stereotypes of Black faces as more athletic than White faces. Consistent with the failure to find effects of familiarity on White raters' reserve stereotypes in Study 1, these results may reflect a weaker affective valence for the traits of reserve and athleticism than for danger or competence. Table 6 shows correlations of
330
TABLE 6, Partial Correlations Between Likeability And Traits

Black Raters White Faces Black Faces -,453 ,29^ -.30^ .03 White Raters White Faces -.70' ,533 ,03 .07 Black Faces -,653 ,653 -,01 -,10
Danger Competent Reserved Athletic
-,603 ,48^ -,06 ,11
Note. Partial correlations control face sex, smile, attractiveness and babyfaceness. p < .05; p < .001,
likeability ratings with each of the traits, partialling out the control variables entered at Step 1 of the regression (face sex, attractiveness, babyfaceness, smiling). There was no significant relationship between ratings of likeability and athletic for White or Black raters judging faces of either race or between ratings of likeability and social reserve for White raters' judging faces of either race. In contrast, the relationships between likeability and the danger and competence composites were significant for both White and Black raters judging faces of either race. These results suggest that the failure of face familiarity to influence race stereotypes on the dimensions of social reserve and athletic may be due to the fact that these traits are less affectively valenced than danger and competence,^ Effects of face familiarity again held true over and above effects of smiUng, babyfaceness, and attractiveness, each of which also had a predictable influence on various impressions, including similar effects of babyfaceness and attractiveness across Black and White raters and faces (Zebrowitz, Montepare, & Lee, 1993), Faces higher in smiling were perceived as more likeable, less dangerous, and less reserved. Consistent with the babyface overgeneralization effect, faces higher in babyfaceness were rated as more likeable and reserved, as well as less dangerous and athletic (Montepare & Zebrowitz, 1998; Zebrowitz, 1997). Consistent with attractiveness halo effect, more attractive faces were rated as more likeable, competent, and athletic as well as less dangerous and reserved (cf., Eagly et al., 1991; Feingold, 1992; Langlois et al., 2000),
GENERAL DISCUSSION
The support for familiar face overgeneralization shown in the ingroup favoritism and race stereotypes observed in Studies 1 and 2 indicates that peo5, It should be noted that Black raters did show negative affective reactions to social reserve, evidenced in a significant negative correlation with likeability for Black faces. However, this effect is not pertinent to explaining the failure of face familiarity to mediate race stereotypes on the dimension of social reserve, since Black raters showed no race stereotype on this trait.
331
pie respond with a general negativity to the more unfamiliar appearance of strangers of another race. This effect is consistent with evidence that other race faces elicit cardiovascular, neural, and conditioned responses paralleling responses to threateriing stimuli (Blascovich, Mendes, Hunter, Lickel, & Kowai-Bell, 2001; Hart et al,, 2000; Ohman, 2005; Olsson, Ebert, Banaji, & Phelps, 2005; Phelps et al,, 2000), The finding that negative reactions to the unfamiliar appearance of other race faces contributes to ingroup favoritism and negative other-race stereotypes while suppressing positive stereotypes provides a novel perspective on racial prejudice and stereotyping. At the same time, the results also support the well-established view that culturally transmitted stereotypes are attached to people based on their social category membership. The relative unfamiliarity of Korean faces merely suppressed White perceivers' strong Asian stereotypes, which have origins that cannot be accounted for by reactions to vmfamiliarity. Similarly, the relative unfamiliarity of Black faces only partially mediated White perceivers' negative stereotypes, which cannot be accoimted for solely by reactions to unfamiliarity. On the other hand, Korean perceivers' negative stereotypes of White faces were eliminated when the greater unfamiliarity of White faces was controlled, suggesting that Koreans may not have culturally transmitted stereotypes about the intelligence, industry, hostility, or trustworthiness of Whites that are independent of their reactions to unfamiliarity. Although face familiarity had parallel effects in both studies, some differences are noteworthy. First, controlling face familiarity eliminated the ingroup favoritism of White Americans and Koreans in Study 1, but only partially mediated the ingroup favoritism of White and Black Americans in Study 2, Second, controlling face familiarity eliminated Koreans' negative stereotypes of White faces, but only partially mediated White Americans' negative stereotypes of Black faces. Given the history of Black-White race relations in the United States, it is not surprising that factors other than an unfamiliar appearance made a more significant contribution to ingroup favoritism and negative stereotypes for Blacks and Whites, Finally, although Koreans' stereotypes of White faces as less quiet and reserved were mediated by face familiarity. Whites' identical stereotypes of Black faces were not. This may reflect a positive affective valence attached to social reserve in Asian culture in contrast to a more neutral reaction to variations in social reserve in U,S, culture. Similarly, the failure of face familiarity to mediate stereotypes of Black faces as more athletic may reflect the lack of strong affective reactions to this trait. The contribution of an unfamiliar appearance to racial prejudice and stereotypes documented in these studies is consistent with recent evidence that more prototypical Black faces or Afrocentric-looking faces of either race elicit more negative reactions from White perceivers (Blair et al,, 2002; Blair et al,, 2004; Eberhardt et al,, 2004; Livingston & Brewer, 2002; Maddox, 2004), Our findings cannot be explained by the hypothesis that group-related physical features activate group-related stereotypes over and above those
332
resulting from categorization (Blair et al,, 2002), If unfamiliarity were a proxy for features that activate stereotypes, then one would expect controlling familiarity to weaken all group-related stereotypes. However, as noted, controlling familiarity strengthened White raters' positive stereotypes of Koreans, and it failed to influence strong race stereotypes that were not affectively valenced. Our findings also cannot be explained by shared negative affective associations to darker skin color (cf, Maddox, 2004), since reactions to the darker Korean and Black faces race varied across rater race. Our findings are most consistent with the hypothesis that perceptual qualities of other-race faces elicit negative affective responses (Livingston & Brewer, 2002), More specifically, our results suggest that negative affective responses contribute not only to prejudice but also to a dual-process stereotyping, in which stereotypes are influenced by cultural beliefs about the outgroup, if any, as well as by negative affective reactions to the unfamiUar appearance of the outgroup, if applicable. Our instructions that "some faces look more familiar than others even when they have never been seen before" were sufficient to elicit judgments with high inter-rater agreement. This indicates that our measure did not assess idiosyncratic variations in the familiarity of the faces looking like my cousin Joethat exemplify what has been called episodic familiarity (Peskin & Newell, 2004), Rather, the familiarity ratings captured the familiarity of each face as compared with some abstract prototype shared by all raters, exemplifying what has been called general familiarity (Peskin & Newell, 2004), Since own-race faces looked more familiar than other-race faces, it appears that our measure of face familiarity captured resemblance to prototypical faces of the raters' own race. This measure was not isomorphic with face race, since familiarity predicted impressions with race of face controlled. Thus, some faces looked more familiar than others across all raters of a given race, independent of face race as well as attractiveness, babyfaceness, or smiling, and the more familiar-looking faces were perceived as more likeable, less dangerous, and more competent by raters of all races, as well as more reserved by Korean raters. Although our analyses examined effects of variations in the average familiarity of faces across all raters, it would also be valuable to investigate effects of idiosyncratic variations in face familiarity. For example, it may be that some perceivers Hnd faces of another race more familiar than do other perceivers and these differences among individuals may mediate differences in their ingroup favoritism and stereotypes. In considering the generalizabiUty of our results, it must be acknowledged that the faces employed were drawn from a variety of samples that could have introduced between race variability in the mean level of appearance qualities that can influence impressions. For example, Korean faces were judged as more babyfaced than White faces by raters of both races, a difference that would not necessarily generalize to other samples of Korean and White faces. The same can be said for the finding that Black faces were
333
judged more attractive than White faces by Black raters. However, the effects of face race on ingroup favoritism and stereotypes were shown with babyfaceness, attractiveness, and smiling controlled. Although it remains a possibility that something about the particular faces employed in our studies other than their race or these control variables contributed to the stereotyped trait impressions, our results replicate well-documented racial stereotypes using other methodologies. Moreover, neither the mediating effects of familiarity nor the differences between the two groups of raters, as in likeability ratings, can be explained by intrinsic differences between the faces of each race. In addition to supporting the FFO hypothesis, the present findings have two more general implications. First, they indicate that conceptualizing stereotypes along the two dimensions of competence and warmth (Fiske, Cuddy, Glick, & Xu, 2002) does not capture the nuances of other-race stereotypes that were revealed in our studies. Although impressions of intelligence and industry map onto the dimension of competence, impressions of trustworthiness, hostility, and social reserve are not adequately captured by the dimension of warmth. White perceivers stereotyped Koreans as low in danger (high trustworthy, low hostility), which would be construed as high warmth. However, they also stereotyped Koreans as high in reserve, which would be construed as low warmth. Similarly, White perceivers stereotyped Blacks as high in danger (low warmth) but low in reserve (high warmth). Another general implication of the present findings is to underscore the utility of the face rating method for assessing ingroup favoritism and stereotypes. Many stereotype researchers have resorted to methods that either conceal their purpose, such as the Modern Racism scale (e,g,, McConahay, 1986), or are relatively invulnerable to social desirability biases, such as implicit measures (e,g,, Olson & Fazio, 2003) or matching race-stereotyped biographical descriptions to faces (e,g,, Blair et al,, 2002), However, the present findings show that the simple procedure of asking participants to rate individual faces of varied races elicits ingroup favoritism and stereotypes. It appears that the social desirability biases that militate against negative judgments about an outgroup category are undercut when outgroup members are judged on an individuated, person-by-person basis. With this method, Korean raters voiced negative stereotypes of Whites and White raters voiced negative stereotypes of Blacks and all raters voiced preferences for their own group once variations in smiling, babyfaceness, and attractiveness were controlled. The finding that face familiarity mediated ingroup favoritism and race stereotypes extends previous evidence that the manipulated familiarity of faces or other stimuli through mere exposure has a positive effect on likeability (e,g,, Bornstein, 1989,1993; Hamm, Baum, & Nikels, 1975; Peskin & Newell, 2004; Rhodes et al,, 2001; Zajonc, 1968), In the context of this evidence, the present findings suggest that interventions designed to increase people's familiarity with other race faces may ameliorate reactions to indi-
334
viduals whose race differs from their own. Indeed, mere increases in face familiarity may contribute to the recently documented robust ameliorative effects of intergroup contact even when the ideal conditions specified by Allport(1954) are not mef(Pettigrew & Tropp, 2006),
REFERENCES Allen, B,P, (1996), African Americans' and European Americans' mutual attributions: Adjective generation technique (AGT) stereotyping. Journal of Applied Social Psychology, 26, 884-912, AUport, G,W, (1954), The nature of prejudice. Reading, MA: Addison-Wesley, Andersen, S, M., & Berk, M,S, (1998), The social-cognitive model of transference: Experiencing past relationships in the present. Psychological Science, 7,109-115, Bar-Haim, Y,, Ziv, T., Lamy, D,, & Hodes, R, M, (2006), Nature and nurture in own-race face processing. Psychological Science, 17,159-163, Baron, R, M,, & Kenny, D, A, (1986), The moderator-mediator variable distinction in social psychological research: Conceptual, strategic and statistical considerations. Journal of Personality and Social Psychology, 51,1173-1182, Bayton, J, A, (1941), The racial stereotypes of Negro college students. Journal ofAbnormal and Social Psychology, 36, 97-102, Biernat, M,, & Manis, M, (1994), Shifting standards and stereotype-based judgments. Journal of Personality and Social Psychology, 66, 5-20, Blair, I,, Judd, CM,, & Chapleau, K,M. (2004), The influence of afrocentric facial features in criminal sentencing. Psychological Science, 15, 674-679, Blair, L, Judd, CM,, Sadler, M,S,, & Jenkins, C (2002), The role of afrocentric features in person perception: Judging by features and categories. Journal of Personality and Social Psychology, 83, 5-25, Blascovich, J,, Mendes, W,B,, Hunter, S,B., Lickel, B,, & Kowai-Bell, N, (2001), Perceiver threat in social interactions with stigmatized others. Journal ofPersonality and Social Psychology, 80, 253-267, Bornstein, R,F, (1989), Exposure and affect: Overview and meta-analysis of research, 1968-1987, Psychological Bulletin, 106, 265-289, Bornstein, R,F, (1993), Mere exposure effects with outgroup stimuli. In D,M, Mackie & D,L, Hamilton (Eds,) Affect, cognition, and stereotyping: Interactive processes in group perception (pp, 195-211), San Diego, CA: Academic Press, Brewer, M,B, (1979), Ingroup bias in the minimal intergroup situation: A cognitive-motivational analysis. Psychological Bulletin, 86, 307-324, Corneille, O,, Monin, B,, & Pleyers, G, (2005), Is positivity a cue or a response option? Warm glow vs, evaluative matching in the familiarity for attractive and not-so-attractive faces. Journal of Experimental Social Psychology, 41,431-437, Cottrell, C A,, & Neuberg, S, L, (2005), Different emotional reactions to different groups: A sociofunctional threat-based approach to "prejudice," Journal of Personality and Social Psychology, 88, 770-789, DeBruine, L, (2002), Facial resemblance enhances trust. Proceedings ofthe Royal Society of London, B, 269,1307-1312, Devine, P, G,, & Elliot, A,J, (1995), Are racial stereotypes really fading?: The Princeton trilogy revisited. Personality and Social Psychology Bulletin, 21,1139-1150,
335
Dovidio, J,F,, & Gaertner, S,L, (1993), Stereotypes and evaluative intergroup bias. In D,M, Mackie, & D,L, Hamilton (Eds,), Affect, cognition, and stereotyping. San Diego, CA: Academic Press, Eagly, A, H,, Ashmore, R, D,, Makhijani, M, G,, & Longo, L, C (1991), What is beautiful is good: A meta-analytic review of research on the physical attractiveness stereotype. Psychological Bulletin, 110,109-128, Eagly, A, H,, Mladiruc, A,, & Otto, S, (1994), Cognitive and affective bases of attitudes toward social groups and social policies. Journal of Experimental Social Psychology, 30,113-137, Eberhardt, J, L,, Goff, P, A,, Purdie, V, J,, & Davies, P,G, (2004), Seeing black: Race, crime, and visual processing. Journal of Personality and Social Psychology, 87, 876-963, Feingold, A, (1992), Good-looking people are not what we think. Psychological Bulletin, HI, 304r-34:l. Fiske, S, T, (1998), Stereotyping, prejudice, and discrimination. In D, T, Gilbert, S, T, Fiske, & G, Lindzey (Eds,), Handbook of social psychology (4th ed,, Vol, 2, pp, 357-^11), New York: McGraw-Hill, Fiske, S, T,, Cuddy, A, J,, Glick, P,, & Xu, J, (2002), A model of (often mixed) stereotype content: Competence and warmth respectively follow from perceived status and competition. Journal of Personality and Social Psychology, 82, 878-902, Gergen, K, J, (1967), The significance of skin color in human relations, Daedalus, 96, 391-406, Gibson, J, J, (1966), The senses considered as perceptual systems. Boston, MA: Houghton Mifflin, Gibson, J,J, (1979), The ecological approach to visual perception. Boston, MA: Houghton Mifflin, Hamm, N,H,, Baum, M.R,, & Nikels, K,W, (1975), Effects of race and exposure on judgments of interpersonal favorability. Journal of Experimental Social Psychology, 11,14-24, Hart, A, J,, Whalen, P, J,, Shin, L, M,, Mclnemey, S, C , Fischer, H,, & Rauch, S,L, (2000), Differential response in the human amygdala to racial outgroup vs, ingroup face stimuli, Neuroreport, 11, 2351-2355, Hatfield, E,, & Sprecher, S, (1986), Mirror, mirror: The importance of looks in everyday life. Albany: State University of New York Press, Hill, M,E, (2002), Skin color and the perception of attractiveness among African Americans: Does gender make a difference? Social Psychology Quarterly, 65, 77-91, Hill, T,, Lewicki, P,, Czyzewska, M,, & Schuller, G, (1990), The role of learned inferential encoding rules in the perception of faces: Effects of nonconscious self-perpetuation of a bias. Journal of Experimental Social Psychology, 26, 350-371, Ho, C , & Jackson, J,W, (2001), Attitudes toward Asian Americans: Theory and measurement. Journal of Applied Social Psychology, 31,1553-1581, Jarvis, B,G, (2004), Medialab (Version 2004, 3,24) [Computer software]. New York, NY: Empirisoft Corporation, Judd, CM,, Park, B,, Ryan, CS,, Brauer, M,, & Kraus, S, (1995), Stereotypes and ethnocentrism: Diverging interethnic perceptions of African American and White American youth. Journal ofPersonality and Social Psychology, 69,460^81, Jussim, L,, Nelson, T,E,, Manis, M,, & Soffin, S, (1995), Prejudice, stereotypes, and la-
336
beling effects: Sources of bias in person perception. Journal ofPersonality and Social Psychology, 68, 228-246, Kawai, Y, (2005), Stereotyping Asian Americans: The dialectic of the model minority and the yellow peril. Howard Journal of Communications, 16,109-130, Kim, B, S, K,, Atkinson, D, R,, & Yang, P. H, (1999), The Asian Values Scale: Development, factor analysis, validation, and reliability. Journal of Counseling Psychology, 46, 342-352, Knutson, B, (1996), Facial expressions of emotion influence interpersonal trait inferences. Journal of Nonverbal Behavior, 20,165-182, Langlois, J,H,, Kalakanis, L,E,, Rubenstein, A,J,, Larson, A,D,, Hallam, M.J,, & Smoot, M,T, (2000), Maxims and myths of beauty: A meta-analytic and theoretical review. Psychological Bulletin, 126,390-423, Langlois, J,H,, Roggman, L,A,, & Musselman, L, (1994), What is average and what is not average about attractive faces? Psychological Science, 5, 214220, Lewicki, P, (1985), Nonconscious biasing effects of single instances on subsequent judgments. Journal of Personality & Social Psychology, 48, 563-574, Lewontin, R. (2001), Is race a sensible genetic concept? Invited talk in lecture series on Biology, race, and genetics: Old problems and new findings. Brandeis University, Lin, M, H,, Kwan, V,S,Y,, Cheung, A,, & Fiske, S, (2005), Stereotype content model explains prejudice for an envied outgroup: Scale of Anti-Asian American stereotypes. Personality & Social Psychology Bulletin, 31, 34-47, Livingston, R,W, (2002), The role of perceived negativity in the moderation of African Americans' implicit and explicit racial attitudes. Journal of Experimental Social Psychology, 38, 405-413, Livingston, R,W,, & Brewer, M,B, (2002), What are we really priming? Cue-based versus category-based processing of facial stimuli. Journal of Personality and Social Psychology, 82, 5-18, Locke, V., Macrae, CN,, & Eaton, J.L, (2005), Is person categorization modulated by exemplar typicality? Social Cognition, 23, 417-428, Maddox, K,B, (2004), Perspectives on racial phenotypicality bias. Personality & Social Psychology Review, 8,383-401, Maddox, K,B,, & Gray, S,A, (2002), Cognitive representations of Black Americans: Re-exploring the role of skin tone. Personality and Social Psychology Bulletin, 28, 250-259, Martinez, A,M,, & Benavente, R, (1998), TheAR Face Database. CVC Technical Report #24, June, McArthur, L,Z,, & Baron, R, M, (1983), Toward an ecological theory of social perception. Psychological Review, 90,215-238, McConahay, J,B, (1986), Modem racism, ambivalence, and the modem racism scale. In J,F, Dovidio & S,L, Gaertner (Eds,), Prejudice, discrimination, and racism (pp, 91-125). San Diego, CA: Academic Press, Meier, B, P,, Robinson, M,D,, & Clore, G,L, (2004), Why good guys wear white: Automatic inferences about stimulus valence based on brightness. Psychological Science, 15, 82-87, Montepare, J.M,, & Dobish, H, (2003), The contribution of emotion perceptions and their overgeneralizations to trait impressions. Journal of Nonverbal Behavior, 27, 237-254, Montepare, J,M,, & Zebrowitz, L, A, (1998), Person perception comes of age: The sa-
337
lience and significance of age in social judgments. Advances in Experimental Social Psychology, 30,93-161, Niemann, Y,F,, Jennings, L,, Rozelle, R,M,, Baxter, J,C,, & Sullivan, E, (1994), Use of free responses and cluster analysis to determine stereotypes of eight groups. Personality and Social Psychology Bulletin, 20, 379-390, Ohman, A, (2005), Conditioned fear of a face: A prelude to ethnic enmity? Science, 309, 711-713. Olson, M, A,, & Fazio, R, H, (2003), Relations between implicit measures of prejudice: What are we measuring? Psychological Science, 14, 636-639, Olsson, A,, Ebert, J,P,, Banaji, M,R,, & Phelps, F, A, (2005), The role of social groups in the persistence of leamed fear. Science, 309, 785-787, Peskin, M,, & Newell, F,N, (2004), Familiarity breeds attraction: Effect of exposure on the attractiveness of typical and distinctive faces. Perception, 33,147-157, Pettigrew, T,F,, & Tropp, L,R, (2006), A meta-analytic test of intergroup contact theory. Journal of Personality and Social Psychology, 90, 751-783, Phelps, E,A,, O'Connor, K,J,, Cunningham, W,A,, Funayama, E,S,, Gatenby, J,C,, Gore, J,C,, & Banaji, M,R, (2000), Performance on indirect measures of race evaluation predicts amygdala activity. Journal of Cognitive Neuroscience, 12, 1-10, Rhodes, G,, & Zebrowitz, L,A, (2002), Facial attractiveness: Evolutionary, cognitive, and social perspectives. Westport, CT: Ablex, Rhodes, G,, Halberstadt, J., & Brajkovich, G, (2001), Generalization of mere exposure effects to averaged composite faces. Social Cognition, 19,57-70, Rubenstein, A.J,, Kalakanis, L,, & Langlois, J,H, (1999), Infant preferences for attractive faces: A cognitive explanation. Developmental Psychology, 35, 848-855, Schneider, D,J, (2004), The psychology of stereotyping. New York: Guilford Press, Smith, E,, & Zarate, M,A, (1992), Exemplar-based model of social judgment. Psychological Review, 99, 3-21, Stangor, C , Sullivan, L,A,, & Ford, T,E, (1991), Affective and cognitive determinants of prejudice. Social Cognition, 9, 359-380, Tropp, L,R,, & Pettigrew, T,F. (2005), Relationships between intergroup contact and prejudice among minority and majority status groups. Psychological Science, 16, 951-957, Uhlmann, E,, Dasgupta, N,, Elgueta, A,, Greenwald, A,G,, & Swanson, J, (2002), Subgroup prejudice based on skin color among Hispanics in the United States and Latin America, Social Cognition, 20,198-225, Van den Berghe, F,P,L,, & Frost, P, (1986), Skin color preference, sexual dimorphism, and sexual selection: A case of gene culture co-evolution? Ethnic and Racial Studies, 9, 87-113, Zajonc, R, B,(1968), Attitudinal effects of mere exposure. Journal of Personality and Social Psychology, 9, Monograph Supplement, No, 2, part 2, Zajonc, R,B, (2000), Feeling and thinking: Closing the debate over the independence of affect. In J,P, Forgas (Ed,), Studies in emotion and social interaction, second series feeling and thinking: The role of affect in social cognition (pp, 31-58), New York: Cambridge University Press, Zebrowitz, L, A, (1996), Physical appearance as a basis of stereotyping. In N, MacRae, M, Hewstone, & C, Stangor (Eds,), Foundations of stereotypes and stereotyping (pp, 79-120), New York: Guilford Press,
338
Zebrowitz, L,A, (1997), Reading faces: Window to the soul? Boulder, CO: Westview Press, Zebrowitz, L. A, (2004), The origins of first impressions. Journal of Cultural and Evolutionary Psychology, 2,93-108. Zebrowitz, L, A,, Montepare, J,M., & Lee, H.K, (1993), They don't all look alike: Differentiating same versus other race individuals. Journal of Personality and Social Psychology, 65,85-101, Zuckerman, M, (1990), Some dubious premises in research and theory on racial differences, American Psychologist, 45,1297-1303,

Zebrowitz Bronstad Lee 2007

Caricato da

Informazioni sul documento

Copyright

Formati disponibili

Condividi questo documento

Condividi o incorpora il documento

Opzioni di condivisione

Hai trovato utile questo documento?

Questo contenuto è inappropriato?

Copyright:

Formati disponibili

Zebrowitz Bronstad Lee 2007

Caricato da

Copyright:

Formati disponibili

Social Cognition, Vol. 25, No. 2,2007, pp.

THE CONTRIBUTION OF FACE FAMILIARITY TO INGROUP FAVORITISM AND STEREOTYPING

FACE FAMILIARITY, INGROUP FAVORITISM, STEREOTYPING

ZEBROWITZ, BRONSTAD, AND LEE

FACE FAMILIARITY, INCROUP FAVORITISM, STEREOTYPING

ZEBROWITZ, BRONSTAD, AND LEE

FACE FAMILIARITY, INGROUP FAVORITISM, STEREOTYPING

ZEBROWITZ, BRONSTAD, AND LEE

FACE FAMILIARITY, INGROUP FAVORITISM, STEREOTYPING

ZEBROWITZ, BRONSTAD, AND LEE

2 >^ fi "ra 'i < CQ cS

FACE FAMILIARITY, INCROUP FAVORITISM, STEREOTYPING

ZEBROWiTZ, BRONSTAD, AND LEE

FACE FAMiLiARITY, iNGROUP FAVORiTiSM, STEREOTYPiNG

0.8 0.6 0.4 0.2

Face Race, Step2, Controlling Familiarity

-0.6 (Whites Higher)

ZEBROWiTZ, BRONSTAD, AND LEE

FACE FAMiLiARiTY, iNGROUP FAVORiTiSM, STEREOTYPiNG

ZEBROWiTZ, BRONSTAD, AND LEE

FACE FAMiLiARiTY, iNGROUP FAVORiTiSM, STEREOTYPiNG

ZEBROWiTZ, BRONSTAD, AND LEE

FACE FAMILIARITY, INCROUP FAVORITISM, STEREOTYPINC

ZEBROWITZ, BRONSTAD, AND LEE

ikeable anger ompetent eserved thletic amiliar ttractive abyface U U

FACE FAMiLiARiTY, iNGROUP FAVORiTiSM, STEREOTYPiNG

ZEBROWITZ, BRONSTAD, A N D LEE

u.o D Face Race, Step1

Face Race, Step2, Controlling

-U.4 (Whites Higher)

FACE FAMILIARITY, INGROUP FAVORITISM, STEREOTYPING

ZEBROWITZ, BRONSTAD, AND LEE

FACE FAMILIARITY, INGROUP FAVORITISM, STEREOTYPING

ZEBROWITZ, BRONSTAD, AND LEE

TABLE 6, Partial Correlations Between Likeability And Traits

Danger Competent Reserved Athletic

-,603 ,48^ -,06 ,11

FACE FAMILIARITY, INCROUP FAVORITISM, STEREOTYPING

ZEBROWITZ, BRONSTAD, AND LEE

FACE FAMILIARITY, INCROUP FAVORITISM, STEREOTYPINC

ZEBROWITZ, BRONSTAD, AND LEE

FACE FAMILIARITY, INCROUP FAVORITISM, STEREOTYPINC

ZEBROWITZ, BRONSTAD, AND LEE

FACE FAMILIARITY, INCROUP FAVORITISM, STEREOTYPING

ZEBROWITZ, BRONSTAD, AND LEE

Potrebbero piacerti anche