Race Biases in Face Recognition - Research Paper Example

Several studies on race biases in face recognition have discovered a high-frequency of the phenomenon of own-race bias in the experiments conducted. The own-race bias phenomenon is a state wherein the memory for own-race faces are better retained in the memory than for another race faces. In lieu of this, this report will provide an empirical explanation to further support the own-race bias phenomenon through conducting an experiment that will investigate on the respondents’ ability to match faces. The experiment is composed of three phases, namely, the study phase where the respondents will be shown with various target faces to determine their capability to remember faces; the activity task where the respondents will answer a crossword problem within an allotted time; and the test phase where the respondents will be shown with faces that are distorted and they will be asked to discern if the altered faces were displayed to them in the study phase. The findings of the experiment reinforced the own-race bias phenomenon because the resulting data showed greater recognition accuracy for aligned faces from own race. Introduction This report has explored a number of theories and assumptions in understanding the high tendency for own-race bias in face recognition experiments. Basically, three of the models in the study of race biases in face recognition, namely, the contact hypothesis, prejudiced attitudes and physiognomic differences between races are used as guiding principles the data interpretation and analysis. The contact hypothesis postulates that the own-race bias in face recognition could be greatly affected by the extent of the contact with several other-race faces. Pieces of evidence available regarding the accuracy of the contact hypothesis are mixed. Some studies show that individuals living in an isolated neighborhood are more prone to exhibit own-race bias in face recognition than those who reside in an integrated neighborhood; nevertheless, Blacks people tend to display the opposite. On the other hand, the prejudiced attitude principle maintains that racial intolerance might affect the intensity of own-race bias in face recognition. The assumption of the prejudiced hypothesis is that racially discriminating attitudes consequently produce reduced ability to recognize other-race faces (Chiroro and Valentine 1995: p. 880). Closely related to the prejudiced attitude hypothesis is the physiognomic difference between races or the assumption that an individual’s outer appearance, particularly the facial features, may provide an idea about the individual’s character or personality. Another theory that is thoroughly connected with the own-race bias in face recognition is the configural coding in face recognition. Diamond and Carey (1986 as cited in Rhodes et al. 1989: p. 313) have stated that “expertise in face recognition depends on the ability to code configural properties in addition to isolated features”. They have verified their hypothesis through conducting two dissimilar experiments in order to evaluate the impact of distortion on identification of “own race” (high expertise) and “other race” (low expertise) faces (ibid). Moreover, Diamond and Carey (1986) proposed that a significance element of expertise with face recognition is the aptitude to predetermine slight deformation of the permanent facial configuration. Therefore, the featural processing or face recognition abilities for beginners are strongly determined by discrete or isolated characteristics or those that are recognizable without orientation to the other parts of the face (ibid: p. 314). Furthermore, studies that are indispensably connected with the expertise and configural coding of face recognition are the inversion effect, thatcherizing and composite faces. Yin (1969 as cited in Rhodes et al 1989: p. 315) discovered that for upright faces respondents tried to obtain a broad sense of the entire image but for the inverted faces respondents tend to look for some unique features to get a general impression of the image shown. On the other hand, Thompson (1980 as cited in ibid) offered an explanation regarding the disorientation effects inherent to inversion which is termed as the Thatcher illusion or the loss of the entire face when inverted with distorted facial features. Also, Young, Hellawell and Hay’s (1988 as cited in ibid) study of composite faces showed that face recognition decreases once the top halves of famed faces are combined with the bottom halves of different faces but the effect is reversed once the faces become inverted. Using these theoretical backgrounds, this report will attempt to convey an experiment finding that supports the own-race bias in face recognition and the effects of inversion in the accuracy of face recognition for both the “own race” (high expertise) and “other-race” (low expertise). The research questions to be answered in the proceeding sections of the paper are: 1.) Can an experiment using composite faces show the extent of accuracy in face recognition for both the “own race” and “other-race”; 2.) Are familiar misaligned faces easier to recognize than aligned faces; 3.) Is configural processing difficult to inhibit? The presumptions that could be obtained from these research questions are: 1.) the experiment using the composite faces can produce accurate results on the problem of race-bias in face recognition; 2.) Familiar misaligned faces are easier to recognize than aligned faces; 3.) and configural processing is difficult to inhibit. Methodology The experiment has three phases, namely, the study phase, the activity task and the test phase. These three phases aim to find out the ability of the respondents to recognize, remember and match faces from their “own race” and from the “other race”. a. Participants The sample for the experiment was randomly chosen from a roster of heterogeneous population. The researcher randomly selected 20 respondents. b. Materials/ Apparatus The study was conducted using these following materials: computer units for the presentation of pictures in the study phase and the crossword puzzles for the transitional phase of the experiment. c. Procedure The experimenter asked the 20 respondents to occupy their designated seats and computer unit before the briefing or the handing out of instructions. Afterwards, the respondents underwent the three phases of the experiments. For the first phase, the computer monitor displayed a series of pictures and the respondents are instructed to remember the faces flashed in the screen as much as they could. After the study phase, the respondents are then asked to answer a crossword puzzle which would be an activity task while waiting for the final phase of the experiment. After the allotted 3 minutes for the completion of the crossword puzzles, the respondents were then instructed to look carefully at their computer monitor as the series of altered or composite faces; these faces were of two types, the aligned and the misaligned. For each altered face, the respondents were asked to press the key “Y” if they recall it and press the key “N” if they do not. After the final phase, the experimenter expressed his/her deepest gratitude for the participation of the respondents and debriefed them regarding the experiment afterwards. Results The experiment using composite faces of “own race” and “other-race” showed that altering the fixed facial features could produce result that would testify to the phenomenon of own-race bias. There is greater accuracy for own race faces. For the own race faces, as the composite faces become misaligned, the accuracy to recognize it goes down whereas for the other race faces there is no much difference between the result of the aligned and the misaligned. Moreover, the results show that familiar aligned faces are easier recognized than misaligned faces which are shown by the statistics in the table:   OWN RACE   OTHER RACE   Subject No Aligned Misaligned Aligned Misaligned 1 853 608 989 901 2 1184 807 1102 956 3 1209 688 888 724 4 754 867 1044 888 5 966 537 871 748 6 944 917 1167 1084 7 1062 877 1009 954 8 1381 752 1099 933 9 905 1088 1438 1272 10 784 752 754 775 11 936 931 1079 1000 12 1135 674 991 871 13 887 848 1267 1042 14 1173 701 1146 957 15 994 569 947 808 16 1037 602 1031 843 17 971 631 966 866 18 769 775 1123 949 19 1094 637 1132 842 20 1299 843 1310 1002 This confirms that the featural processing of the novice is predetermined by the distinctiveness of the basic or fixed characteristics of the “own race” and the “other race”. Discussions The result of the experiment shows that the respondents tend to recognize more the “other race faces” than the “own race faces” which is illustrated in the graph above. The configural coding in face recognition of the respondents manifested that they have an expertise in isolating the distinctive characteristics of each of the “own race face” and the “other race face” to come up with an accurate impression of two aligned pairs of different faces. On the other hand, for the “other race faces” the respondents showed little disparity in recognizing both the aligned and misaligned composite faces. This means that the respondents cannot discriminate the dissimilarities between the aligned and misaligned composite faces of the “other race”. Using the contact hypothesis, it may be argued that the respondents’ insufficient interaction with the other race affected their ability to discriminate the differences among the facial features of the other race. On the other hand, excessive contact could downsize the ability of an individual to recognize or remember some of the distinctive facial features of their own race. The prejudiced attitude could also be attributed to the results of the experiments since having stereotypes create memory retention for characteristics that are assumed to be inferior or not part of dominant traits. The prejudice attitude could arise from the physiognomic aspect of race bias in face recognition. The respondents could have ascribed a unique personality and characteristic to the other race faces which enabled them to remember the composite faces of the other race. This experiment, in conclusion, supports the assumption of the own race bias. Works Cited Carey, S. and Diamond, R. (1977). From piecemeal to configural representation of faces. Science, 195, 312-314. Rhodes, G., Brake, S., Taylor, K. and Tan, S (1989). Expertise and configural coding in face recognition. British Journal of Psychology, 80, 313-331. Thompson, P. (1980). Margaret Thatcher: a new illusion. Perception, 9,483-484. Yin, R. K. (1968). Looking at upside-down faces. Journal of Experimental Psychology, 81, 141-145. Young, A., W., Hellawell, D. and Hay, D. C. (1987). Configural information in face perception. Perception, 16, 747-759. Read More

Another theory that is thoroughly connected with the own-race bias in face recognition is the configural coding in face recognition. Diamond and Carey (1986 as cited in Rhodes et al. 1989: p. 313) have stated that “expertise in face recognition depends on the ability to code configural properties in addition to isolated features”. They have verified their hypothesis through conducting two dissimilar experiments in order to evaluate the impact of distortion on identification of “own race” (high expertise) and “other race” (low expertise) faces (ibid).

Moreover, Diamond and Carey (1986) proposed that a significance element of expertise with face recognition is the aptitude to predetermine slight deformation of the permanent facial configuration. Therefore, the featural processing or face recognition abilities for beginners are strongly determined by discrete or isolated characteristics or those that are recognizable without orientation to the other parts of the face (ibid: p. 314). Furthermore, studies that are indispensably connected with the expertise and configural coding of face recognition are the inversion effect, thatcherizing and composite faces.

Yin (1969 as cited in Rhodes et al 1989: p. 315) discovered that for upright faces respondents tried to obtain a broad sense of the entire image but for the inverted faces respondents tend to look for some unique features to get a general impression of the image shown. On the other hand, Thompson (1980 as cited in ibid) offered an explanation regarding the disorientation effects inherent to inversion which is termed as the Thatcher illusion or the loss of the entire face when inverted with distorted facial features.

Also, Young, Hellawell and Hay’s (1988 as cited in ibid) study of composite faces showed that face recognition decreases once the top halves of famed faces are combined with the bottom halves of different faces but the effect is reversed once the faces become inverted. Using these theoretical backgrounds, this report will attempt to convey an experiment finding that supports the own-race bias in face recognition and the effects of inversion in the accuracy of face recognition for both the “own race” (high expertise) and “other-race” (low expertise).

The research questions to be answered in the proceeding sections of the paper are: 1.) Can an experiment using composite faces show the extent of accuracy in face recognition for both the “own race” and “other-race”; 2.) Are familiar misaligned faces easier to recognize than aligned faces; 3.) Is configural processing difficult to inhibit? The presumptions that could be obtained from these research questions are: 1.) the experiment using the composite faces can produce accurate results on the problem of race-bias in face recognition; 2.) Familiar misaligned faces are easier to recognize than aligned faces; 3.) and configural processing is difficult to inhibit.

Methodology The experiment has three phases, namely, the study phase, the activity task and the test phase. These three phases aim to find out the ability of the respondents to recognize, remember and match faces from their “own race” and from the “other race”. a. Participants The sample for the experiment was randomly chosen from a roster of heterogeneous population. The researcher randomly selected 20 respondents. b. Materials/ Apparatus The study was conducted using these following materials: computer units for the presentation of pictures in the study phase and the crossword puzzles for the transitional phase of the experiment. c. Procedure The experimenter asked the 20 respondents to occupy their designated seats and computer unit before the briefing or the handing out of instructions.

Afterwards, the respondents underwent the three phases of the experiments. For the first phase, the computer monitor displayed a series of pictures and the respondents are instructed to remember the faces flashed in the screen as much as they could.

Read More