In this paper, we report the results of two experiments that randomise the share of individuals who are taking an action in subjects’ immediate environment. Despite the differences between our two settings (face masks and online camera use), we uncover some empirical results that are common to both. First, we find that the share of individuals taking the relevant action is increasing in the share of others who take the action (although the relationship need not be linear). Second, and despite this, we find that many individuals nonetheless defy social pressure. Our results point both to the importance of social pressure as well as its very real limits in our settings.