Conceived and designed the experiments: GP MD LC GG. Performed the experiments: GP MD. Analyzed the data: GP MD LC GG. Contributed reagents/materials/analysis tools: LC GG. Wrote the paper: GP MD LC GG.
The authors have declared that no competing interests exist.
Viewing a face with averted gaze results in a spatial shift of attention in the corresponding direction, a phenomenon defined as gaze-mediated orienting. In the present paper, we investigated whether this effect is influenced by social factors. Across three experiments, White and Black participants were presented with faces of White and Black individuals. A modified spatial cueing paradigm was used in which a peripheral target stimulus requiring a discrimination response was preceded by a noninformative gaze cue. Results showed that Black participants shifted attention to the averted gaze of both ingroup and outgroup faces, whereas White participants selectively shifted attention only in response to individuals of their same group. Interestingly, the modulatory effect of social factors was context-dependent and emerged only when group membership was situationally salient to participants. It was hypothesized that differences in the relative social status of the two groups might account for the observed asymmetry between White and Black participants. A final experiment ruled out an alternative explanation based on differences in perceptual familiarity with the face stimuli. Overall, these findings strengthen the idea that gaze-mediated orienting is a socially-connoted phenomenon.
People's gaze conveys several information to an observer. Humans and other animal species have developed the ability to rapidly orient attention towards the location where the gaze of another is directed in order to detect potentially relevant events in the environment
To our knowledge, no previous work has systematically addressed the impact of intergroup processes on gaze cueing. In the present research we explored this issue by focusing on racial group membership, comparing the responses of Black and White people when presented with Black and White faces. There is strong evidence that the interaction between the racial group membership of the perceiver and that of the person perceived may deeply shape human basic cognitive processes. For instance, it has been shown that empathic sensorimotor resonance when observing the pain of other individuals is present for racial ingroup but not outgroup members
In the following experiments we aimed at testing the two alternative scenarios described above. In particular, in Experiment 1, we presented White participants (i.e., the majority group) with a modified spatial cueing paradigm including pictures of White and Black individuals.
Thirty-seven White Italian students (25 females) from the University of Padova participated in the study. Their mean age was 24 years. All participants provided a written informed consent prior to taking part in the experiment. Twenty participants took part in the study on a voluntary basis whereas the other seventeen were paid 10 euros for their participation. The experiment was conducted in accordance with the guidelines laid down in the Declaration of Helsinki, and participants were fully informed that their data would be analyzed anonymously.
Upon arrival in the lab, participants were greeted by a White experimenter. Presentation of the stimuli and registration of the responses were controlled by E-prime 1.1. Stimuli were presented on a 17″ monitor with a resolution of 1024×768 connected to an IBM compatible Pentium IV computer. The participants sat 57 cm from the computer monitor.
Sixteen avatar 3-D full-color faces created with FaceGen 3.1 software (2006) were used (4 Black females, 4 Black males, 4 White females, and 4 White males). Independent observers showed perfect agreement in the categorization of stimuli as representing either Black or White faces. For each face, the same software was used for creating a first image with direct gaze, a second image with averted gaze to the right and a third image with averted gaze to the left. Faces did not display additional characteristics such as hair or clothes. Each face subtended a visual angle of 16.8° in height and 14.4° in width. Faces presented to paid and unpaid participants were the same, except for the fact that those presented to paid participants were matched for luminance (2.5 cd/m2).
Each trial began with a white fixation point which remained on the screen for 900 ms, then a face with direct gaze appeared remaining on the screen for another 900 ms. Next, the image of the same face with gaze averted leftwards or rightwards was superimposed, thus conveying the impression of the eyes looking left or right. A target letter (either L or T) then appeared to the left or to the right of the face after 200 ms (see
The Figure illustrates the experimental procedure within a single trial. Gaze direction was not informative as to target location. A cue-target spatially incongruent trial with a Black face is illustrated. Stimuli are not drawn to scale.
Gaze direction was not informative with respect to target location and trial order was randomized. The target letter appeared at approximately 11° from the center of the screen, aligned with the horizontal meridian. The background color of the monitor was set to black and the target letter was set to white in 24-point Arial bold font. The specific color of the target is likely to be irrelevant, in line with previous research showing that chromatic features do not affect gaze cueing
Thirty-two students (16 females) from the University of Padova who self-identified as Black individuals were paid 10 euros for their participation. They were all born in different countries of Sub-Saharan Africa. Their mean age was 26.5 years. The experiment was conducted in accordance with the guidelines laid down in the Declaration of Helsinki, and participants were fully informed that their data would be analyzed anonymously.
Apparatus, stimuli, and procedure were the same as those used in Experiment 1.
Seventy-two White Italian students (60 females) from the University of Padova participated in partial fulfillment of course requirements. Their mean age was 20 years. All participants provided a written informed consent prior to taking part in the experiment. The experiment was conducted in accordance with the guidelines laid down in the Declaration of Helsinki, and participants were fully informed that their data would be analyzed anonymously.
The apparatus, stimuli, and procedure were the same as in the previous experiments, with only one exception, namely the manipulation of an additional between-participants factor. Indeed, participants were randomly assigned to either a Mixed condition (N = 36) or a Blocked condition (N = 36). In the Mixed condition, we aimed at replicating the results obtained in Experiment 1. The procedure was exactly the same as in Experiment 1 (see
Seventeen White Italian students (14 females) from the University of Padova took part in the experiment and were paid 10 euros for their participation. Their mean age was 23 years. All participants provided a written informed consent prior to taking part in the experiment. The experiment was conducted in accordance with the guidelines laid down in the Declaration of Helsinki, and participants were fully informed that their data would be analyzed anonymously.
Apparatus and procedure were the same as in Experiments 1 and 2 (see
A 2 (cue-target spatial congruency: congruent vs. incongruent)×2 (racial group membership: White vs. Black)×2 (payment: paid vs. unpaid) ANOVA was performed on mean reaction times for correct responses. Participants were faster to identify the target when it appeared in the gazed-at location (i.e., spatially congruent trials,
Panel A shows participants' mean reaction times in Experiment 1 as a function of Congruency and Racial group membership. Panel B shows participants' mean reaction times in Experiment 2 as a function of Congruency and Racial group membership. The black line illustrates spatially congruent trials, the dashed line illustrates spatially incongruent trials. Bars represent SEM. *p<.05.
An identical ANOVA was conducted on the percentage of errors (3.8%). No significant effect emerged, thus making the occurrence of any speed-accuracy tradeoff unlikely.
The present results show that racial group membership plays an important role in gaze cueing.
This pattern is even more remarkable given that it was obtained with a sample drawn from a population of undergraduate students who typically hold rather liberal social attitudes
The observed pattern cannot distinguish between the two scenarios illustrated in the
Data from one participant were excluded from analyses because more than 10% of his reaction times were above 1000 ms thus leaving thirty-one participants for the analyses.
A 2 (cue-target spatial congruency: congruent vs. incongruent)×2 (racial group membership: White vs. Black) repeated measures ANOVA was performed on mean reaction times for correct responses. Participants were faster to identify the target when it appeared in the gazed-at location (i.e., spatially congruent trials,
A further 2 (cue-target spatial congruency: congruent vs. incongruent)×2 (racial group membership: White vs. Black)×2 (Experiment: 1 vs. 2) mixed ANOVA was conducted on correct reaction times in order to compare the behavior of White and Black participants. The three way interaction was significant,
Overall, the two experiments seem to indicate that social factors impact onto the cueing effect. In order to further support this interpretation, in Experiment 3, we employed a specific experimental manipulation that is known to shape the salience of social factors such as group membership. Indeed, research from the person perception domain has shown that the automaticity of category activation is modulated by the task environment, so that categorical knowledge is activated when participants are presented with exemplars belonging to two different categories (i.e., in mixed order), but not when stimuli are blocked according to their category
For completeness, a 2 (cue-target spatial congruency: congruent vs. incongruent)×2 (racial group membership: White vs. Black) repeated measure ANOVA with Condition (mixed vs. blocked) as a between participant factor was performed on mean reaction times for correct responses. The three-way interaction was not significant,
Panel A shows mean reaction times for the Mixed Condition as a function of Congruency and Racial group membership. Panel B shows mean reaction times for the Blocked Condition as a function of Congruency and Racial group membership. The black line illustrates spatially congruent trials, the dashed line illustrates spatially incongruent trials. Bars represent SEM. *p<.05.
These results show that, when White and Black faces were presented in a mixed order, White participants exhibited a gaze-cueing effect only in response to White faces. In contrast, when participants faced two different blocks in which race was invariant in a block, they shifted attention to every available cue and a cueing effect emerged in response to both White and Black faces. This pattern is taken as additional evidence that the modulation of gaze cueing observed in the previous experiments is related to social rather than low-level perceptual properties of the stimuli. In this regard, one may have argued that the absence of cueing effect exhibited by White participants for Black faces in Experiment 1 was simply due to the different physical properties (e.g., contrast between the skin and the sclera) of the faces, with faces of Black individuals eliciting a weaker perception of averted gaze than faces of White individuals. The observation that Black faces were able to trigger a cueing effect only under specific contextual circumstances allows us to rule out this alternative account and highlights that the broader experimental context can have a fundamental role in the emerging of the cueing effect, thus challenging the unconditioned automaticity of gaze-mediated orienting of attention
So far, we have discussed our findings as reflecting socially relevant differences associated to racial group membership. However, one may argue that the effects obtained in the previous experiments were instead due to differences in perceptual familiarity. Indeed, it could be hypothesized that White participants have a perceptual advantage in processing White faces because of the familiarity they have with people of their racial ingroup. On the contrary, Black participants might be expected to be equally good at processing faces of both White individuals, who represent the majority of the population in their living environment, and Black individuals, who represent their direct relatives.
In order to investigate whether perceptual familiarity of the face stimuli could play a role in driving our results, we performed another experiment in which we presented participants with perceptually familiar and unfamiliar faces as stimuli in a gaze-cueing paradigm.
Data from two participants were excluded from analyses because more than 10% of their reaction times were above 1000 ms thus leaving fifteen participants for the analyses.
A 2 (cue-target spatial congruency: congruent vs. incongruent)×2 (familiarity: familiar vs. unfamiliar) repeated measures ANOVA was conducted on mean reaction times for correct responses. A significant effect of Congruency emerged,
The results of Experiment 4 demonstrate that a difference in the perceptual familiarity of the face stimuli is not sufficient to modulate the gaze-cueing effect. Indeed, even though it could be reasonably expected that White participants had a better expertise in processing White faces over green multi-racial faces, the cueing effect was not modulated by the nature (familiar vs. unfamiliar) of the cueing faces. This suggests that the results obtained in the previous experiments, in which Black and White faces were used, are unlikely to be driven by any eventual difference in perceptual familiarity associated to these two classes of stimuli.
In recent years, research has focused on the social side of gaze-mediated orienting of attention by investigating the interplay between this phenomenon and social information
Although perceptual familiarity represented a potential alternative explanation for these findings, the results of Experiment 4 suggest that the different degree of perceptual familiarity with the face stimuli is not sufficient to modulate gaze-mediated orienting. Indeed, the cueing effect exhibited by participants in Experiment 4 was not modulated by the presentation of either a White Caucasian face or a green multi-racial face. Had perceptual familiarity played a relevant role, we should have observed a significantly reduced cueing effect for the unfamiliar green multi-racial faces, which clearly was not the case.
In light of the arguments discussed above, we feel confident that the effects emerged in Experiments 1 and 2 genuinely reflect the impact of race membership. In this regard, it is worth noting that, in the present experiments, we focused on overall intergroup differences, comparing responses towards White and Black individuals in the lack of any more specific information about the faces used as stimuli. However, when specific characteristics of a single individual are inconsistent with the social information associated to his/her group (e.g., the face of a well-known high-status Black person), it might be predicted that individualized information can override the effects of group membership, even for White respondents. In addition, it was hypothesized that group memberships have to be salient in the specific social context in order to exert a modulation. Experiment 3 was specifically designed to test whether the modulation of gaze cueing observed in Experiment 1 for White participants was sensitive to the context. The results showed that when the experimental context allowed for a comparison between Black and White faces (i.e., mixed condition), White faces were prioritized over Black faces. On the contrary, when Black and White faces were kept constant within two distinct blocks (i.e., blocked condition), they both drove gaze-mediated orienting. This result is important in that it shows that the modulation of gaze cueing as a function of race membership critically depends upon whether or not the context favored the activation of different social information related to different racial groups.
Overall, the results of the present set of experiments demonstrate that social information associated to racial group membership can affect orienting of attention. It is likely that the specific kind of modulation observed here (i.e., only White participants differentiate between White and Black faces) is not universal but it occurs each time the two groups are not associated to equally positive representations in terms of attributes and social status. In support of this view, a questionnaire administered to a sample of White and Black individuals from the same population as participants in the present experiments confirmed that Black people were considered as the minority group with respect to White people, independent of the race of the respondent (see
It could be argued that the modulation emerged in the present research reflects the impact of different factors related to the perception of a particular social group. Indeed, each social group is associated to different affective responses, evaluations, stereotypes and perceptions of social status, and it is difficult to identify which specific factors are involved in the observed modulation. In addition, these affective and cognitive aspects of attitudes towards the outgroup are so strictly interconnected that the identification of the unique impact of each of them is problematic. For instance, prominent models about intergroup perception
In sum, differences in perceived social status are a likely candidate factor for explaining the current findings. In this domain, a recent study on non-human primates investigated the role of status in modulating gaze-mediated orienting of attention
To conclude, we found that racial group membership modulates gaze-mediated orienting of attention in humans. All together, results obtained in the current set of experiments show that gaze-mediated orienting of attention is not an ubiquitous phenomenon which takes place every time an averted gaze is seen, but it depends on the kind of stimuli presented and on the social relationship between the observer and the person perceived.
(DOC)
(DOC)