Research Paper Sample on Other-Race Effect

📌Category: Articles, Race and Ethnicity, Sociology
📌Words: 788
📌Pages: 3
📌Published: 18 January 2022

Facial recognition is an important process that can be easily taken for granted. Successful social interactions require the ability to recognize and interpret individuals faces in order to understand identity and social cues. Issues with facial recognition can create an inability to properly connect with other people. A heavily researched issue with facial recognition is the other-race effect (ORE). The ORE is the phenomenon when one is able to recognize faces of their own race more accurately than faces belonging to people of other races. This may impact the way people of different races connect with each other. Recent studies have been attempting to determine how facial recognition can be improved, reducing the ORE. 

In their article, “Learning context and the other-race effect: Strategies for improving face recognition”, Cavasos et al set up two experiments to determine if learning context and image variability impact facial recognition accuracy for ones’ own race and other races. In their first experiment, they examined if learning context (contiguous or distributed) had an impact on facial recognition, hypothesizing that distributed learning would yield higher recognition accuracy than contiguous learning. 133 University of Texas Dallas (UTD) students were included in the study(74 Caucasian and 59 East Asian). Participants were randomly assigned to either the continuous or distributed learning context. They first completed a learning phase where 144 images were shown for two seconds each, including 4 images of each of 36 identities(18 Caucasian, 18 East Asian). In the contiguous condition, the 4 images of each identity were shown consecutively wheras in the distributed condition they were shown randomly. Then they entered the testing phase where they were shown 72 identities; 36 previously seen and 36 new (36 Caucasian, 36 East Asian). When an image appeared on the screen, participants pressed old(previously seen)or new on a keyboard. In their second experiment, they again examined if learning context impacts facial recognition accuracy of ones’ own race and other races now using single, repeated images of each individual, hypothesizing that they would find a strong ORE and greater accuracy in the distributed learning context. 140 UTD students were included in the study(75 Caucasian, 65 East Asian). The experimental design was identical to the first experiment, except participants were shown only one randomly assigned photo of each individual four times. For both experiments, the independent variable was learning context and the dependant variable was recognition accuracy.

Statistical analyses were done for accuracy and response bias in both experiments. No main effect of learning context was found on recognition accuracy in experiment 1, but participants in the distributed learning condition were found to have greater recognition accuracy than those in the contiguous condition in experiment 2. Criterion data were also analyzed to show how liberal(more likely to choose old) or conservative(more likely to choose new) participants were during the testing phase in order to measure response bias. In experiment 1 only, a significant effect of learning context was found showing that participants in the contiguous condition were more conservative than those in the distributive condition. Both found that there was a main effect of stimulus race such that participants were more conservative when recognizing Caucasian faces, and this effect was especially strong for Caucasian participants recognizing Caucasian faces. 

A cross-experiment analysis was then conducted in order to look for differences in recognition accuracy based on image variability and to analyze if learning context and image variability interact in a way that influences recognition accuracy. Greater  recognition accuracy was found for the multiple variable image condition than for the single repeated image condition. Participants responded more conservatively to the multiple variable image condition than to the single repeated image condition. They also responded more conservatively to Caucasian faces(Caucasican participants particularly), and more conservatively during contiguous learning. 

The authors main finding was that learning context affected recognition accuracy for faces of both own and other race. Distrubuted learning gave greater facial recogntion but only in the single repeated image condition, leading the authors to conclude that distributive learning is only of use when participants perceive that the distributed photos are of the same person. Their second finding was that muti-image learning gives better facial recognition accuracy for both own and other races, leading them to conclude that seeing multiple images of the same face allows for generalization to new images of the face. The authors note that this conclusion was determined using the cross-experimental analysis, and should be considered cafefully. 

Future research could solidify the conclusions that the authors came to. The authors propose a future experiment containing both learning context and multi-image learning in order to better understand how they interact and are best used for facial recognition training. Another mentioned future experiment would add more facial recognition training to ensure that participants understand that different photos are of the same person, allowing for better comparison of learning contexts. These future studies may discover an efficient way for people to learn facial recognition, reducing issues like the ORE, and allowing for better connections to be made between people. 

References

Cavazos JG, Noyes E, O'Toole AJ. Learning context and the other-race effect: Strategies for improving face recognition. Vision Res. 2019 Apr;157:169-183. doi: 10.1016/j.visres.2018.03.003. Epub 2018 Apr 6. PMID: 29604301.

+
x
Remember! This is just a sample.

You can order a custom paper by our expert writers

Order now
By clicking “Receive Essay”, you agree to our Terms of service and Privacy statement. We will occasionally send you account related emails.