In the context of a voluntary remedial math course for university students, we conduct randomized field experiments to study whether social information can help counteract low participation rates. Incoming students receive postal invitation and reminder letters that inform them about past...