Great appreciations to my family for their patience during this research. This research was possible due to the assistance and funding of Family Power Corporations. Correspondence concerning this article should be addressed to 750258, Department of the Psychology, Sam Houston State University, Huntsville, TX 77341, United States.
Facial Recognition, Bias in Race and Social Groups
Humans have a tendency to perceive facial recognition of cross-race individuals to look all alike. Unlike the perception in own-race, which has a higher facial recognition accuracy than the recognition of cross-race faces. Throughout the years, several replicated empirical researches have been made to comprehend the processing of facial recognition.
The superiority of facial recognition in own-race compared to the facial recognition of cross-race is known as the own-race bias. MacLin and Malpas stated that the first demonstration of the cross-race effect was done by Malpass and Kravitz, since then several meta-analyses have been made to explain this phenomenal ambiguity.
Since people have more familiarity with their own racial images, they become more acquainted with the proportions of features that are most essential when identifying faces of their own-race . In theory, if the same experiences were to be applied with cross-race individuals, facial recognition could improve. This psychological theory, known as contact hypothesis, suggest that bias between races can be minimized if cross-race individuals engage more frequently amongst one another. In a controlled experiment, the CRE was momentarily reduced when bias perception between same-race and cross-race were practiced. Thus, training in a long-term reverses CRE. Children who were adopted by a different race grew up to racially associate themselves more with their adoptive family. They saw t their adoptive race as own and their biological race as other. Increase exposure with cross-race faces can develop better CR recognition.
Despite these empirical findings, other studies have challenged the contact hypothesis behind the cross-race effect. Studies have shown that when individuals perceive other individuals that are considered an in-group, facial recognition was recognized better than individuals of an out-group, despite of race. We are shown that the accuracy of facial recognition is not about race but about the social groups people are affiliated with. The previous theory states that bias in facial recognition occur because of long-term exposure of same-race. Bernstein, Young and Hugenberg tested two hypotheses proving that individuals recognize in-group members better than out-group members without manipulating race. They tested 60 White Miami University students who would each do a facial recognition test on 80 White college-age males. The first 40 faces were labeled half “Miami University” in a red background and the other half, “Marshall University” in a green background.
The second set of 40 faces were labeled half red in a red background and the other half green in a green background. Their analysis of variance results revealed that the faces labeled Miami University were recognized better than the faces that were labeled Marshall University. These studies showed that the participants and the individuals in the photographs were the same race, there were no other physical distinction between them and yet affiliation alone was sufficient to elicit a face recognition trend similar to the CRE. Similar to cross-race faces, out-group members have poor recognition results.
Explaining this theory could be due to the social-cognitive model. It is the process individuals form to categorize other members of in-groups or out-groups. Race excluded, affiliations and other physical distinctions are considered in social categorization. During facial recognition in cross-race faces, people search for specific features to categorize. This leads to different forms in recognition aside from race. A study was conducted to support that social categorization affected facial recognition. In this one study, 61 White University of Delaware students were shown eight faces, in a random order. They consisted of four Black students, two from UD and two from James Madison University and four White students: two UD affiliations and two JMU affiliation. In the ANOVA results, the participants recognition for own-university faces had a better recall than those compared to the other-university faces. Race had no main effect on facial recall. Alternatively, the propensity to have a categorial perception of social groups contributes to the specifying features of social categorization.
In all, facial recognition works in with race and social class studies. Despite the empirical evidence of social categorization, own-race faces are better recognized then cross-race faces. The categorization of faces is flawed when studies have shown that Blacks are automatically racially categorized. Hehman, Mania and Gaertner had conducted a second study where instead of dividing the eight faces by affiliation they were divided by race, their affiliation was still labeled. Facial recognition had a higher accuracy for own-race faces than cross-race faces. Their affiliation was no longer a main effect; therefore, race is a significant factor in facial recognition.
Race does play a large role in recall, but social class has a main effect as well. Even though facial recognition of same-race faces is more accurate, mixing social context will reduce recognition. When White-middle class individuals were presented with White faces on impoverished backgrounds, recognition had a significant drop as if it were compared to same-race and cross-race recognition. Ultimately, my goal is to show that despite affiliations and constant contact between individuals race and social groups are a predominant factor in being one step closer in understanding the bias in facial recognition. Therefore, the current study connects race and social groups to have a higher accuracy in recognizing same-race and other-race.
Method
Participants
Eighty-two Sam Houston State University undergraduate students participated in a research experiment for their Research Methods course as partial fulfillment of a class requirement. A total of 66 females and 16 males. Twenty-seven were Black participants and 55 were White. The average age of the participants was 23.02 years (SD = 6.87). All participants were tested in multiple groups; they were given the same experiments with the same instructions.
Design
The current study utilized a 2 x 2 within-subjects design. The dependent variable was the proportion of faces recognized.
Materials
For the experiment, the encoding task presented 32 photographs of 16 White and 16 Black individuals. These individuals were randomly assigned “faculty” or “student”, with the faces ranging in age between 18 and 30 years. The photographs were a smiling headshot of the individuals. Each photograph was color coded by a blue background labeled “faculty” or an orange background labeled “student” and were numbered one through 32. Each photograph was presented at an eight second rate in a random order.
A rating sheet numbered one through 32 was used using a 5-point like scale. One being dissimilar and five being very similar. After the encoding task, there was a filler task of three sheets of hidden pictures. For the recognition task, 64 photographs of individual’s with neutral poses were presented. From the 64 photographs, 32 were from the encoding task and the other 32 were new. The photographs had been randomly arranged and labeled one through 64. The photographs were randomly shown for five seconds each and the social group labels were removed. There was no orange or blue background, only a white background. A recognition form numbered one through 64 was utilized with “yes” or “no” responses. All photographs were taken from Minear and Park.
Procedure
The day of the experiment, participants were given a consent form to participate in the experiment. They were instructed to read the form fully and sign if they wish to participate. Participants were also given a brief demographic questionnaire. After, each participant was assigned a number, that number was to be written down on top of three sheets of paper. None of the participants names were written on any page other than the consent form. The first page that was given out was a rating sheet. For the encoding task, the participants were informed that they will be seeing a series of photographs of individuals on the projector for eight seconds per photograph.
They were instructed to rate each individual on how similar the individuals were to their current social group. When the encoding task was completed, each participant received another assignment, this was served as their filler task. The participants were instructed to find as many items from the 3 hidden picture sheets provided in five minutes. The last step of the experiment was for the participants to take a recognition test. The participants were once again shown a series of photographs. The participants were instructed to circle yes or no on the recognition sheet if they recognized any of the individuals that were presented in the encoding task. Once the experiment was fully completed, all sheets were collected, and the participants were informed of the purpose of the experiment. In the end, the students were thanked for their participation.
References
- Bernstein, M. J., 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(8), 706–712. https://doi.org/10.1111/j.1467-9280.2007.01964.x
- Hehman, E., Mania, E. W., & Gaertner, S. L. (2010). Where the division lies: Common ingroup identity moderates the cross-race facial-recognition effect. Journal of Experimental Social Psychology, 46(2), 445–448. https://doi.org/10.1016/j.jesp.2009.11.008
- MacLin, O. H., & Malpass, R. S. (2001). Racial categorization of faces: The ambiguous race face effect. Psychology, Public Policy, and Law, 7(1), 98–118. https://doi.org/10.1037/1076-8971.7.1.98
- Minear, M., & Park, D. C. (2004). A lifespan database of adult facial stimuli. Behaviour Research Methods, Instruments, & Computers, 36, 630-633. http://dx.doi.org/10.3758/BF03206543
- Shriver, E. R., Young, S. G., Hugenberg, K., Bernstein, M. J., & Lanter, J. R. (2008). Class, race, and the face: Social context modulates the cross-race effect in face recognition. Society for Personality and Social Psychology, 34(2), 260-274. http://doi.org/10.1177/0146167207310455
- Young, S., & Hugenberg, K. (2012). Individuation motivation and face experience can operate jointly to produce the own race bias. Social Psychological and Personality Science, 3, 80-87. https://doi.org/10.1177/1948550611409759
