People’s group memberships bias the way we perceive, remember, and interact with social surroundings. One of such biases is the so-called own-race bias in face recognition (also known as the cross-race effect). The own-race bias (ORB) is the tendency for perceivers to have better recognition for faces of their racial ingroup than for faces of their racial outgroup (Meissner & Brigham, 2001). The ORB is a statistically robust phenomenon with important social implications. For example, research on the causes of erroneous guilty verdicts (in USA, Canada, or the UK), found that the majority of false convictions involved other-race eyewitness misidentifications (Scheck, et al., 2000; Smith, et al., 2004).

The vast majority of experimental studies on this phenomenon uses paradigms where participants view photographs of unfamiliar individuals with the goal of memorizing them for a future recognition test. These photographs are usually presented in a passport-like format without any kind of contextual information. The manipulations employed typically involve asking participants to individuate the target-faces (versus memorizing; e.g., Hugenberg et al., 2007), offering rewards as a function of recognition performance (DeLozier & Rhodes, 2015), or presenting faces with different expressions (Young & Hugenberg, 2012).

In our view, such experimental paradigms greatly constrain the theoretical conclusions one can extract from existing findings as they do not reflect the dynamic nature of everyday-life face perception and memory. In everyday life, we are continuously extracting information to guide person perception. During such process, our prior knowledge about people, our stereotypic beliefs, and other top-down factors come into play and help shaping the way we perceive and remember other people (Macrae & Bodenhausen, 2000; Roese & Sherman, 2007). For example, it has been shown that stereotypes provide perceivers with expectancies that guide information processing such that any information (e.g., a behavior) about someone that fits the stereotype expectancy is understood much faster than information that contradicts stereotype expectancy (e.g., Bargh & Thein, 1985; Stern, et al., 1984).

Inspired by these findings, the present research investigated whether stereotypic information about the target individuals moderates the ORB. To the extent that the ORB is driven by the poor individuated processing of other-race faces, presenting these faces together with information that either fits or contradicts the race stereotype should influence the perceptual encoding of these faces. Specifically, while presenting other-race faces with stereotype-consistent information should lead to an encoding based more on category than on individuating facial features, presenting these faces with stereotype-inconsistent information should lead to more attention being allocated to individuating information.

Study 1 – Generation and Pre-testing of Stereotypical Behaviors

We first selected 16 personality traits from Petsko and Bodenhausen (2019): 4 consistent (2 positive and 2 negative) and 4 inconsistent (2 positive and 2 negative) traits with the stereotype of blacks; 4 consistent (2 positive and 2 negative) and 4 inconsistent (2 positive and 2 negative) traits with the stereotype of whites. Then, for each trait, we generated a set of behavior descriptions that seemed illustrative of the respective trait. This effort resulted in a total of 154 behavior descriptions. Finally, we asked a group of 30 participants to rate each behavior in two dimensions: in the first block, they rated how stereotypic each of these behaviors seemed of blacks and whites; in the second block, they rated each behavior on valence.

Study 2 – The effect of Behavioral Information in the ORB I

Study 2 (N = 85) consisted of a study phase, a distractor phase, and a test phase. In the study phase, white participants were presented with faces paired with stereotype consistent, inconsistent, and neutral behavioral descriptions. Half of the faces were white males (own-race faces) and half black males (other-race faces; Chicago Face Database, Ma et al., 2015). The computer randomly paired behaviors and faces. Participants’ task will be to establish a mental connection between faces and behaviors. After the study phase, participants completed a distracter task followed by a recognition test. In this test, they saw the old faces (i.e., faces presented in the study phase) intermixed with the same number of new faces and had to indicate whether they saw each face before using a 6-point confidence scale ranging from 1 - “Definitely no” to 6 - “Definitely yes.” Upon completing the study, when preparing the data for analysis, we noted that some faces and behaviors appeared more than once in the study phase. We then realized we had made a mistake in the javascript programming that critically compromised our data.

Study 3 – The effect of Behavioral Information in the ORB II

Study 3 (N = 85) was a close replication of Study 2. Results showed that the predicted interaction between behaviors and face race did not reach conventional significance levels (p < 0.05), even though the pattern of means was in the predicted direction: the difference between own- and other-race faces was numerically smaller when these faces had been presented with stereotype inconsistent behaviors. Following a common practice in the recognition memory literature, we dichotomized the 6-point scale in order to calculate the signal detection index d' (Macmillan & Creelman, 1991). On repeating the same analysis on d', the predicted interaction reached statistical significance.

Study 4 – The effect of Behavioral Information in the ORB III

Given the results obtained in the previous study when doing the analysis on signal detection index d', in Study 4 (N = 90), we replicated Study 3 with only two modifications: instead of using a 6-point confidence scale in the recognition test, in this study we used a standard yes/no scale; after each "yes" response, participants made remember-know-guess judgment (for details, see Yonelinas, 2002). If stereotype-inconsistent behaviors lead to a more individuated processing of other-race faces, then we should obtain more remember responses for these faces than for other-race faces presented with stereotype-consistent behaviors. Contrary to our expectations, the analysis of d' scores yielded a non-significant interaction between behaviors and race. Regarding remember-know-guess judgments, we only obtained a main effect of race on remember judgments such that participants gave more remember judgments for own- than other-race faces, which is consistent with previous findings (Horry et al., 2010).

Study 5 – The effect of Behavioral Information in the ORB IV

In the previous studies, we only used positive behaviors. In Study 5 (N = 96), we used both positive and negative behaviors. Thus, half of the stereotype consistent behaviors were positive and half negative, and the same happened for stereotype inconsistent behaviors. We changed the paradigm such that behaviors were presented before faces, on a separate screen. Participants were instructed to read each behavior and try to imagine the person who did it.

Moreover, after the study phase, participants completed the Racial Contact Questionnaire (RCQ; Hancock & Rhodes, 2008) and the Internal and External Motivation to Respond Without Prejudice Scale (Plant & Devine, 1998), and responded to a single item measuring their political orientation (1 – "Extremely liberal", 9 – "Extremely conservative"). Results showed no significant interactions between behaviors and race nor between behaviors, race, and valence of behaviors. Interestingly, we obtained a strong significant interaction between race, valence, and political orientation, showing that more conservative participants had better recognition accuracy for other-race faces presented with negative than positive behaviors, while more liberal participants showed no such difference; for own-race faces, we did not observe any effect of valence and political orientation on recognition accuracy. 

Study 6 – The effect of Behavioral Information in the ORB V

Study 6 (N = 96) was designed to explore further the effect of political interaction obtained in the previous study. Instead of simply measuring political orientation, in this study, we manipulated this variable by selecting a sample constituted by 50% liberals and 50% conservatives. Thus, this study used a 2 (political orientation: liberals vs. conservatives) X 2 race (own-race vs. other-race) X 2 (stereotypic behavior: consistent vs. inconsistent) X (behavior valence: positive vs. negative). The procedure was similar to that of the previous experiment. Contrary to our predictions, the political orientation x race x behavior valence was not significant.

Concluding Remarks

In the present studies, we hypothesized that stereotype-inconsistent behaviors should lead to an encoding of faces based on individuating information, particularly of other-race faces as own-race faces are by default individuated. Taken together, the results obtained in the five studies did not support this hypothesis. Thus, although these findings may suggest that the ORB is immune to the violation of stereotypical expectancies, future research should further investigate this question using different methods.

This research was conducted in collaboration with Joana Quarenta and Ana Sofia Santos from the University of Lisbon and Joshua Correll from the University of Colorado, Boulder, USA. A manuscript is currently in preparation. We are thankful for the opportunity afforded by the EASP Seedcorn Grant, allowing us to conduct the above-presented studies.

References

• Bargh, J. A., & Thein, R. D. (1985). Individual construct accessibility, person memory, and the recall-judgment link: The case of information overload. Journal of Personality and Social Psychology, 49, 1129 –1146.
• Bernstein, M. Young, S. G., & Hugenberg, K. (2007). The cross-category effect: Mere social categorization is sufficient to elicit an own-group bias in face recognition. Psychological Science, 18, 706–712.
• Correll, J., Hudson, S. M., Guillermo, S., & Earls, H. A. (2017). Of kith and kin: Perceptual enrichment, expectancy, and reciprocity in face perception. Personality and Social Psychology Review, 21, 336–360.
• DeLozier, S., & Rhodes, M. G. (2015). The impact of value-directed remembering on the own-race bias. Acta Psychologica, 154, 62–68.
• Hancock, K. J., & Rhodes, G. (2008). Contact, configural coding and the other‐race effect in face recognition. British Journal of Psychology, 99(1), 45-56.
• Horry, R., Wright, D. B., & Tredoux, C. G. (2010). Recognition and context memory for faces from own and other ethnic groups: A remember-know investigation. Memory & Cognition, 38(2), 134-141.
• Hugenberg, K., Miller, J., & Claypool, H. M. (2007). Categorization and individuation in the cross-race recognition deficit: Toward a solution to an insidious problem. Journal of Experimental Social Psychology, 43, 334–340.
• Macmillan, N. A. & Creelman, C. D. (1991). Detection theory: A user’s guide. Cambridge, England: Cambridge University Press.
• Ma, D. S., Correll, J., & Wittenbrink, B. (2015). The Chicago face database: A free stimulus set of faces and norming data. Behavior Research Methods, 47, 1122-1135.
• Macrae, C. N., & Bodenhausen, G. V. (2000). Social cognition: Thinking categorically about others. Annual Review of Psychology, 51, 93–120.
• Meissner, C. A., & Brigham, J. C. (2001). Thirty years of investigating the own-race bias in memory for faces: A meta-analytic review. Psychology, Public Policy, and Law, 7, 3–35.
• Plant, E. A., & Devine, P. G. (1998). Internal and external motivation to respond without prejudice. Journal of Personality and Social Psychology, 75(3), 811-832.
• Roese, N. J., & Sherman, J. W. (2007). Expectancy. Social Psychology: Handbook of basic principles, 2.
  Scheck, B., Neufeld, P., & Dwyer, J. (2000). Actual innocence: Five days to execution and other dispatches from the wrongly convicted. New York: Doubleday.
• Smith, S. M., Stinson, V., & Prosser, M. A. (2004). Do they all look alike? An exploration of decision-making strategies in cross-race facial identifications. Canadian Journal of Behavioural Science, 36, 146–154.
• Stern, L. D., Marrs, S., Millar, M. G., & Cole, E. (1984). Processing time and the recall of inconsistent and consistent behaviors of individuals and groups. Journal of Personality and Social Psychology, 47, 253–262.
• Yonelinas, A. P. (2002). The nature of recollection and familiarity: A review of 30 years of research. Journal of Memory and Language, 46(3), 441-517.
• Young, S. G., Hugenberg, K., Bernstein, M. J., & Sacco, D. F. (2012). Perception and motivation in face recognition: A critical review of theories of the cross-race effect. Personality and Social Psychology Review, 16, 116–142.