Trait Ratings for the Radboud Faces Database

Abstract

Since its publication, the Radboud Faces Database (RaFD; Langner et al., 2010) has become one of the most widely used face databases. At the time of writing, it has been cited more than 1,400 times. The database includes validation data such as rated genuineness, clarity, and intensity of the displayed facial expression. Ratings of models’ attractiveness based on their neutral, frontal gaze image is also available and age data for most models can be found at http://gijsbijlstra.nl/rafd-ratings/. These ratings are useful for researchers who want to (a) select models who score particularly low or high on a certain characteristic, (b) ensure that different image sets do not significantly differ on certain characteristics, or (c) control for these characteristics in statistical analyses. Here, additional trait ratings for all 39 models of the RaFD’s Caucasian Adult Subset (neutral expression, frontal gaze) are presented. The models were rated on 19 trait dimensions: trustworthiness, dominance, attractiveness, competence, openness, conscientiousness, extraversion, agreeableness, emotional stability, aggressiveness, friendliness, health, intelligence, anger, disgust, fear, happiness, sadness, surprise.

Can I Trust This Person? Evaluations of Trustworthiness From Faces and Relevant Individual Variables

“…Some literature reports agreement on trustworthiness evaluations across observers and cultures (e.g., Sutherland et al, 2018 , 2020 ). An exploratory analysis revealed that our ratings agree strongly with those obtained in the Netherlands by Jaeger ( 2020 ) for the RaFD faces (ICC = .95 and .92, with our young-adults online and laboratory data, respectively). Good agreement was also achieved between our young-adult laboratory ratings and those collected in Spain by Aguado et al ( 2011 ) for the KDEF faces (ICC = .78), in spite of the fact that faces were presented in grayscale in that study.…”

Judgements of Social Dominance From Faces and Related Variables

“…Given that some studies support a relatively widespread cross-cultural agreement on several trait inferences from faces (e.g., Sutherland et al, 2018 ), we expect our data to be useful for researchers from other countries and cultures. Favoring this argument, an exploratory inspection of the agreement between our ratings and those obtained in the Netherlands by Jaeger ( 2020 ) for faces from the RaFD revealed an excellent level of agreement (Intra Class Correlation Coefficient [ICC] of 0.93 and 0.90, for young adults' data collected in the laboratory and online, respectively). Nonetheless, compared to other inferences, the evaluation of dominance seems to be more flexible and susceptible to contextual factors (e.g., Sutherland et al, 2020 ).…”

How facial masks alter the interaction of gaze direction, head orientation, and emotion recognition

“…Many factors could be responsible for such differences, for instance invariant features such as eye color and shape or eye and pupil distance, or variable features like mimic movements. Even though the models had been trained by FACS experts (Facial Action Coding System) and the displayed facial expressions had been validated (Langner et al, 2010), they still differ in their emotional expressions as well as in their neutral expression (Jaeger, 2020). Since the eye region is considered the most variable facial area (Itier and Batty, 2009), model differences could be particularly relevant when faces are covered by a mask.…”