We investigate the presence of gender bias in student evaluation of teaching (SET) scores with a natural field experiment in India. We employ a novel experimental design, randomly assigning 504 economics students to an identical audiovisual class taught either in a female or male voice. We find...