Docstoc

Slide 1

Document Sample
Slide 1 Powered By Docstoc
					Development of the Own-Race Bias in Face Processing David J. Kelly

Why Face Processing?
• Faces are important for many reasons:
– Allow identification at the individual level – Convey information, such as emotion

• Necessary for many species that live in social groups
– Humans – Apes – Monkeys

Early face processing system
• Preference for face schema at birth (Fantz, 1963; Johnson & Morton, 1991; Valenza et al., 1996).

Early face processing system
• Nelson (2001) suggested that like language, face processing develops from a broad nonspecific system to a human tuned face processor at the end of the first year of life • The environment is ‘sculpting’ the face system

Multidimensional Face Space (Valentine, 1991)
• Valentine proposes a model in which faces are encoded as vectors according to their deviation from a prototypical average • At birth, the dimensions of the prototype are likely to be broad with the development of one’s prototype dependent on facial input • The resulting dimensions will differ according to the input received with certain salient, individuating dimensions carrying more “weight” than others

The Own-Race Bias
• People are typically better at discriminating between faces from their own-ethnic group • Adults are 2.23 times more likely to correctly identify own-race faces as opposed to other-race faces (Meissner & Brigham, 2001) • Available data suggests children also demonstrate the ORB (e.g. Pedzek et al., 2003; Sangrigoli & de Schonen, 2004a) • One study has found evidence of the ORB in 3month-old infants (Sangrigoli & de Schonen, 2004b)

Experience and face processing
• We tested newborns and 3-month-olds infants for sensitivity to ethnicity • Only Caucasian neonates & infants were tested

Kelly et al. (2005). Three-month-olds, but not newborns, prefer own-race faces. Developmental Science, 8, F31-F36.

Experiment 1
• 64 Caucasian newborn infants (36 female) viewed faces from a range of ethnic groups using a Visual Preference (VP) task • VP task used to assess spontaneous preferences for stimuli
– – – – African vs. Caucasian (n=16) Chinese vs. Caucasian (n=16) Asian vs. Caucasian (n=16) Caucasian vs. Caucasian (n=16)

Example Stimuli

Experiment 1

Experiment 1

Experiment 1 - Predictions
• Due to a lack of exposure to faces in general, newborns may not display a preference for faces from any ethnic group. • Alternatively, newborns may have already encoded skin colour information from their mother’s face and subsequently will prefer faces which most closely match that colour

Results
• A paired samples 2-tailed t test conducted on overall looking time at Caucasian vs. otherrace faces yielded a non-significant result (t = .036, df = 47, p= ns) • Observation of mean looking time confirms newborn infants spent equal time looking at both own (49.73%) and other-race faces (50.27%)

Results
Newborns' Looking Time to Own- and Other-Race Faces
L o o kin g T im e (Secs) 6000 5000 4000 3000 2000 1000 0
es e ian n n n n as ia uc Ca Ca uc ric an as ia as ia as ia As as ia in n

Ch

Af

uc

uc

Ca

Ca

Condition

Ca

uc

Discussion
• Newborns show no preference for faces from own- or other-ethnic groups during the first days of life • They can perceive differences, but no group of faces elicits a greater attraction

Experiment 2
• 64 Caucasian 3-month-old infants (42 female) viewed faces from a range of ethnic groups using a Visual Preference (VP) task • Identical to Experiment 1

Predictions
• It was predicted that the infants would demonstrate a familiarity preference for Caucasian faces over other-race faces on the basis of greater experience with same-race faces. • Infants should demonstrate a null preference in the Caucasian / Caucasian condition.

Results
• A paired samples two-tailed t test conducted on overall looking time at Caucasian vs. other-race faces yielded a significant result (t = 4.179, df = 47, p < .0001) • Observation of mean looking time confirms 3-month-old infants spent more time looking at own- (58.80%) than other-race faces (41.20%)

Results
3-Month-Olds' Looking Time to Own- and Other-Race Faces
Looking Time (secs) 7000 6000 5000 4000 3000 2000 1000 0

*

*

*

C hi ne se C au ca si an

A fri ca C n au ca si an

A si an C au ca si an

Condition

C au ca si an C au ca si an

Chinese Infants
• We replicated the VP study in China with 3– month-old Chinese infants • Design was identical to experiment conducted with Caucasian infants

Kelly et al. (2006). Cross-race Preferences for Same-race Faces Extend Beyond the African versus Caucasian Contrast in 3-month-old Infants. Infancy, In Press.

Results
Chinese 3-Month-Olds' Looking Time to Own- and Other-Race Faces
Looking Time (Secs) 7000 6000 5000 4000 3000 2000 1000 0

*

*

*

C au ca si an C hi ne se

A si an C hi ne se

Ethnic Condition

C hi ne se C hi ne se

A fri ca n C hi ne se

Discussion
• 3-month-old infants are sensitive to ethnic differences • Results from primary exposure to own-race faces during the first months of life • Origin of sensitivity to ethnic differences and the own-race bias? • Preference not recognition

VPC Task Experiment
• Testing recognition of own- and other-race faces in 3-, 6- & 9-month-old infants using the Visual Paired-Comparison task • 48 infants to be tested in each age group split between 3 conditions: Caucasian, Chinese, African • Conducted in England & China

VPC Study

VPC Study

VPC Study

3-Month-Olds
3-Month-Olds' VPC Data
80 70 60 50 40 30 20 10 0

Looking Time (%)

*

*
Novel Familiar

African (9)

Chinese (9) Face Category

Caucasian (9)

6-Month-Olds
6-Month-Olds' VPC Data
70 Looking Time (%) 60 50 40 30 20 10 0 African (13) Chinese (13) Face Category Caucasian (14) Novel Familiar

*

*

9-Month-Olds
9-Month-Olds' VPC Data
70 Looking Time (%) 60 50 40 30 20 10 0 African (12) Chinese (11) Face Category Caucasian (12) Novel Familiar

*

Chinese 3-Month-Olds
3-Month-Olds' VPC Data
70 Looking Time (%) 60 50 40 30 20 10 0 African (10) Caucasian (10) Face Category Chinese (11) Novel Familiar

*

*

Chinese 6-Month-Olds
6-Month-Olds' VPC Data
70 Looking Time (%) 60 50 40 30 20 10 0 African (10) Caucasian (10) Face Category Chinese (11) Novel Familiar

*

Discussion
• Results are analogous to Pascalis et al. (2002) • They found that 6- and 9-month-old infants & adults could discriminate human faces, but only 6-month-olds could discriminate monkey (Rhesus Macaque) faces • Attributed to ‘environmental tuning’

Discussion
• Different finding to Sangrigoli & de Schonen (2004). Why?
– Methodological difference
• Test trials

– Change of point of view – Colour images – Only different to experiment 1

Conclusions
• Sensitivity to ethnic, morphological differences emerges early in life • Result of faces seen within the visual environment • Possible precursor of the own-race bias • Evidence for the onset of the own-race bias between 6 & 9 months of age

Thanks to…
• • • • • Olivier Pascalis – University of Sheffield Paul C. Quinn – University of Delaware Kang Lee – University of Toronto Alan M. Slater – University of Exeter Leizhong Ge - Zheijiang Sci-Tech University
– Shaoying Liu – Qinyao Liu


				
DOCUMENT INFO
Shared By:
Tags: Slide
Stats:
views:339
posted:1/22/2010
language:English
pages:36
Description: Slide 1