Fig 1.
Example stimuli representing positions 1, 4, 7, and 10 on each dimension.
Rows represent shifts in physiognomy; columns represent shifts in skin tone.
Fig 2.
Summary of omissions in each stimulus set.
Rows represent levels of physiognomy, columns represent levels of skin color / hue. Red fill indicates a missing cell; yellow fill indicates a stimulus that was duplicated.
Table 1.
95% confidence intervals describing estimate of each parameter of interest in adults as well as the mean value of the maximally prototypical Asian and White faces, separately for each of three stimulus sets as well as the primary model collapsing across stimulus set and a subsequent replication set.
All estimates are in units on the categorization scale, ranging from 1 to 100.
Fig 3.
Main effects of (mean-centered) skin color and facial features when the other factor is set to its mean level, adult participants only.
Dashed red line represents overall model predictions and light grey lines represent the predictions for each participant, estimated from the random effects component of the models.
Fig 4.
Interaction between facial features and skin color for adults.
Lighter and darker faces are depicted at the 3rd and 8th level of skin color, respectively. Light lines indicate uncertainty in the fitted regression.
Fig 5.
Main effects of (mean-centered) skin color and facial features when the other factor is set to its mean level, child participants only.
Dashed red line represents overall model predictions and light grey lines represent the predictions for each participant, estimated from the random effects component of the models.
Fig 6.
Interaction between facial features and skin color for children.
Lighter and darker faces are depicted at the 3rd and 8th level of skin color, respectively. Light lines indicate uncertainty in the fitted regression.