Search

Individuals can form consensual first impressions of others based on brief exposures to their faces. However, whether first impressions are accurate and the underlying mechanism remained unclear. Here, we tried to address these questions by asking subjects to evaluate employees’ professional capabilities and personal traits based on employees’ facial images. Results showed that individuals could form consensual subjective impressions of employees’ professional capabilities, which can be further predicted by employees’ traits evaluated by another independent group of subjects. However, there was no significant correlation between subjective judgments of professional capability and objective indicators of employees’ job performance. Also, with facial features extracted by a deep neural network (VGG-Face), we can build a model to predict the subjective evaluation of professional capability. In contrast, models built on the same facial features failed to predict objective job performance indicators. These results indicated that different individuals consensually judged others’ professional capabilities based on image-computable facial features, but their evaluations did not reflect the objective value of job performance.