Week 9 RF-The other race effect Flashcards
What was found in Meissner & Brigham’s (2001) meta-analysis?
- People are more likely to correctly recognise a face of their own race (greater
number of true positive identifications) and incorrectly recognise a face of another race (greater number of false positive identifications). - People are more cautious to say that they have seen own race faces (more conservative criterion) than other race faces (more liberal criterion i.e., much more likely to say you have seen this person before=greater false positives).
- Shorter exposures and longer retention intervals increase the other race effect
(own>other) suggesting a link between weak encoding and face effect. - Research has a clear and important applied focus in relation to eyewitness
testimony. - The ORE can be found across racial groups.
- Caucasians show good discrimination for Caucasian faces, Black for Black,
Chinese for Chinese but a relative impairment for others - Some early evidence that white participants show greater other race effect than (at least) black participants.
- Amount of interracial contact is a significant but weak predictor of the ORE
- Racial attitudes do not appear relevant (but difficult to measure reliably and
meaningfully)
Does other race contact matter?
■ Meissner & Brigham’s (2001) meta-analysis found significant but weak correlations (r= -.13) with amount of contact.
■ A more recent meta-analysis (Singh et al. 2021) replicated the earlier findings (r= -.15) and identified additional moderators:
– If we experimentally manipulate contact (i.e. train participants), effects are higher than if we simply measure individual differences in other race contact.
– Contact early in life (infancy and childhood) with other race groups reduces the ORE more than later contact (adolescence and adulthood) suggesting the need for hard-wired changes in the brain.
■ BUT, quantity and quality of ORE contact is difficult to measure accurately and studies vary widely in the instruments they use making any firm conclusions difficult. (e.g., the use of Likert scales will not capture the meaningfulness of relationships and contact with other race groups)
What is the Developmental evidence for the ORE? (Kelly et al., 2007)
- Infant studies rely on indirect measures of memory (because we can’t just ask them). Memory is inferred by measuring how long infants spent looking at novel faces in pairs consisting of a novel and a studied face.
- Infants show habituation to studied faces. These faces have become familiar and therefore there is no need to explore
them as much as novel faces. - Looking to the novel face for longer suggests memory for the studied face in the pair. No difference in looking times suggests lack of memory.
Findings:
* Early and gradual development of memory effect.
- No effect at 3 months (i.e., no preference in race) but it appears fully developed by 9 months.
- There was also some evidence that at 6 months, infants were more likely to look at novel Chinese and Caucasian faces than
expected by chance-see dotted baseline at 50%. - Although the mean looking times for African and Middle-Eastern faces at 6 months were similar to Chinese and
Caucasian faces, infant responses were more varied.
Developmental Evidence: Although other race effects can appear quite early, do they remain fixed or can they be reversed? (Sangrigoli et al., 2005)
- Korean adults who between the ages of 3-9 years were adopted and raised by Caucasian families in Europe showed a memory effect for Caucasian and not Asian faces.
- It appears that the system or processes responsible for other race effects are plastic (i.e. subject to change). Theories of the effects should include mechanisms to account for such plasticity.
- Recent evidence suggests that this “other race face plasticity” ends by the age of 12 (McKone et al ., 2019).
What are Face processing resources?
- If processing own race faces is more resource demanding, then reducing our available resources should eliminate the other race effect.
- Reduced other race effect
- If processing of other race faces is more laborious (resource intensive), then reducing available resources should increase the other race effect.
- Greater other race effect
What is The role of attention on the other race effect? (Zhou et al., 2015)
- Zhou et al. manipulated attention to examine the effects of resource availability and allocation in the ORE.
- One group studied own and other race faces under full attention conditions. A second group had their attention divided between face processing and another task (tone detection).
- The results showed that under full attention there was a clear other race
effect (own>other). With divided attention the effect disappeared (own=other). - Own race processing is more attention and resource intensive. We engage
same processes more or use additional processes when dealing with own race faces. - In agreement with findings of slower race classification of own compared
with other race faces which suggests that own race faces receive additional
processing.
What are Expertise-based theories of the other race effect?
- Expertise theories: Lack of contact between races (racial segregation) responsible for developing greater expertise in processing and/or remembering members of the same race better than those of other races.
- Developmental evidence discussed earlier is in line with the predictions of expertise
theories. - No initial differentiation but effect develops quickly.
- Effect can be reversed at least at a young age.
- Two main expertise-based theories: differential processing mechanisms theory;
representational theory.
What is Configural processing?
- Faces are processed configurally: Configural processing involves the extraction of the relationships between permanent facial features (e.g. eyes, nose,
mouth) and enables the processing of the face as a whole rather than a collection of individual features. - Face-specific processing detected early in brain potentials (differences in ERPs visible by 200ms).
- Evidence of configural processing can be seen in studies that present inverted
(upside down) faces. - Inversion disturbs the relationship between features (e.g. eyes no longer above
nose and nose no longer above mouth). Recognition accuracy drops significantly for all inverted faces.
What is the link between Configural processing and the other race effect? (Rhodes et al., 1989)
- Configural processing is related to expertise. We are more likely to process something as a whole if we have a lot of experience with the particular item.
- Reduced contact with other race faces will make them less likely to be processed
configurally. - If same race faces are processed configurally but other race faces are not, then face inversion will affect same races faces more than other race faces.
- Rhodes et al. (1989) found exactly this. Viewed inverted, other race faces were recognised equally well to same race faces.
Representational models: What is The face space model?
- Representational models explain the face race effect with regards to memory
representations. - No difference in how faces are processed initially
- The face-space model (Valentine, 1991) assumes that each facial feature and inter
feature relationships are normally distributed in a population and represented as dimensions of the face’s memory representation. - Multiple dimensions define face space and this changes based off the number of faces we see
- Each face is encoded on each dimension and represented in memory space by its
coordinates. Most faces will be close to the average with few outliers - Typical and atypical faces
- The number of dimensions is defined by and evolves with the number of exemplars. Same race faces are more frequent and will be responsible for shaping the dimensions. The more dimensions, the greater the spread of exemplars in face-space (more distinctive).
- Other race faces may not conform fully to the established dimensions and will be
seen as atypical and will be represented by fewer dimensions. As a result they will
be represented in a smaller, denser face-space. - It will be more difficult to distinguish between other race faces than for same race faces and recognition will suffer.
- More false alarms for other race faces
What are some Critiques of the experience-based models?
- Meissner & Brigham’s (2001) meta-analysis
- There is only weak evidence that amount of contact with other race faces is responsible for the other race effect in adults.
- But that may be due to the fact that measures of contact rely on self-report and are not particularly sensitive.
- But quality of interracial contact may be a better measure
- Will consider individuation studies later.
- Not all studies have found evidence that disruption of configural processes eliminates the other race effect.
- In the face-space model, the actual dimensions are not specified. Training to improve discrimination has short-term effects. We look into some findings that experience-based models cannot easily explain.
What happens with angry faces?
Angry facial expressions capture our attention (more resources allocated)
because:
* Signify threatening intent and potential danger. Should be detected!
- It is also important to encode individuating characteristics of angry persons to make sure you remember them even when angry expression disappears (as the threatening intent may still be present).
- Anger expressed by out groups may be perceived more threatening than in groups.
What happens with angry faces? (Ackerman et al., 2006)
- An other race effect was only seen with
neutral faces. - Angry faces were recognised equally well
across races (i.e., ORE eliminated) so perhaps we can eliminate it in an adult population. - This finding is difficult for processing and
representational theories to explain. - It is not clear how people overcome
fundamental processing and representational shortcomings without
extensive training.
What is The ambiguous-race face effect? (MacLin & Malpass, 2001; 2003)
- What happens when the faces are racially
ambiguous? - Faces (on the right) are identical apart from a single, race specifying characteristic (hair).
- According to experience-based models there should be no memory differences between the Hispanic and Black looking faces.
- Yet, Hispanic participants were more likely to remember the Hispanic looking faces than the Black looking faces.
- “…other race faces are encoded categorically and …categorization drives the perceptual process” (MacLin & Malpass, 2001, p.112)
-There is an element of social categorisation and that we apply a different type of processing which is socially derived rather than expert derived
What evidence shows that Race is not the only issue? (Wiese et al., 2013)
- Several studies (e.g. Rhodes & Anastasi, 2012; Wiese et al. 2013) show other age effects where young and old participants tend to remember faces of their respective age groups better.
- Others report differences in face recognition based on sex (Herlitz & Loven, 2013) with women more likely to remember female than male faces. No sex bias is seen
in men. - expert accounts fails to account for these face biases.