Date of Degree
cross-race effect, eyewitness identification, contact theory, signal detection
Studies have shown that eyewitnesses are less accurate when the accused and foils (i.e., non-suspects in a lineup) are of a different race from the eyewitness than when of the same race. This has been termed the Cross-Race Effect (CRE). To study the CRE in eyewitness identifications, this experiment used a 4x2x2 between-subjects design with 492 participants. Participants watched a film act in which an African-American (AA) or Caucasian (C) either stole or recovered a cell phone. Afterwards they had to identify the person as well as answer a series of questions. There were 4 perpetrators for each race (4 C & 4 AA; 1 perp per video), 2 races of the witnesses (C & AA) by 2 framings of the film (moral transgression/stole or prosocial moral act/found). The purpose of the framings was to see if the moral relevance of the action, stealing versus a mildly prosocial act (recovering a client’s cell phone) affects the criterion applied by the witness and whether this interacts with the race of the eyewitness and those in the films. Signal detection analysis (SDA) was used to calculate identification accuracy. For the measure of decisional criteria, a significant main effect for race of participant was obtained. Caucasian participants committed fewer false positives than African-American participants. A significant interaction was also found between race of participant and race of actor. When the actor was Caucasian, African-American participants committed more false positives than did Caucasian participants. A three-way ANOVA on sensitivity/accuracy (d’) with race of the perpetrator, race of participant and framing of the act as factors, showed a significant main effect for race of actor. There was increased discriminability between perpetrator and foil when the actor was African-American than when the actor was Caucasian. There was also an interaction between race of participant and race of actor. When Caucasian participants viewed the film with an African-American actor they were better able to discriminate between perpetrator and foil than when the actor was Caucasian. However, African-American participants performed the same regardless of whether the actor was Caucasian or African-American. In addition, an interaction between race of actor and framing of the film. When the framing of the film was “found”, there was a significant difference seen, when the actor was African-American, the “eyewitnesses” discriminated better between perpetrator and foil than when the actor was Caucasian. However, when the framing of the film was “stole” there was no difference between the African-American actor and the Caucasian actor. A partial data set (n = 378) which excluded the least similar Caucasian and African actor was also analyzed. In this analysis, a significant main effect for race of participant was obtained. The Caucasian participants committed fewer false positives than did the African-American participants. When looking at sensitivity, a significant main effect was found for race of actor. There was greater discriminability between perpetrator and foils when the actor was African-American than when the actor was Caucasian. In addition, a significant interaction was found between race of actor and framing of the act. When the actor was African-American, participants were able to better discriminate between perpetrator and foil than when the actor was Caucasian. However, when the framing of the film was “stole”, there was no difference between the African-American actor and the Caucasian actor. Thus, the difference ‘disappears’ when the act was described as a transgression. No other main effect or interaction was found. Studying the Cross-Race Effect under different framings of the act should help us tease out what may be attributed to different degrees of contact and the use of different criteria for guilt.
Vidal, Bianca, "Cross-Race Effect of African-Americans and Caucasians" (2018). CUNY Academic Works.
This work is embargoed and will be available for download on Friday, March 15, 2019
Graduate Center users:
To read this work, log in to your GC ILL account and place a thesis request.
See the GC’s lending policies to learn more.