Score contribution per author:
α: calibrated so average coauthorship-adjusted count equals average raw count
Using appearance scores created by facial-recognition and machine-learning programs that incorporate tens of thousands of individuals’ appearance preferences, we find in China’s migrant schools that students’ appearance has a statistically significant and positive effect on their teachers’ evaluation of their exam performance, even after netting out the influences of important confounders such as physical growth, cognitive ability, mental health status, family background, and school quality.