This paper develops a signalling model of HIV testing, and evaluates an information experiment designed to test the theory and reduce stigma. We find that stigma is in part explained by statistical discrimination between potential sexual partners, and that this form of stigma is a significant barrier to HIV testing. Discrimination is based on false beliefs about antiretroviral drugs, the medication used to treat HIV. In our setting, we find that providing precise, new information about the public benefit of antiretroviral drugs reduces stigma and increases HIV testing. The results demonstrate that social stigma can be due to rational behavior by a misinformed public, and that providing new information can be an effective way to mitigate its effects. * l.c.derksen@lse.ac.uk. Laura Derksen is grateful to her supervisor, Oriana Bandiera and her advisor, Greg Fischer, for their continued guidance and encouragement. This paper also benefited from discussions with Gharad Bryan,