["Journal for the Scientific Study of Religion, Volume 60, Issue 2, Page 424-441, June 2021. ", "\nAbstract\nThe study of religion typically assumes that individuals can only identify with one religion, which is inconsistent with lived religious experience. Surveys that only allow respondents to select one choice may underrepresent religious traditions in which multiple religious identities are more likely. We argue that religious “nones” and those who identify with Asian religions will be more likely to identify multiple (non)religious identities and that this will lead to these identities being undercounted in surveys allowing individuals to select only one identity. We use a sample of Mechanical Turk participants half of whom received the standard one choice (non)religious identity option and the other half could select as many (non)religious identities as they wanted. This allows us to show which (non)religious identities are undercounted when only one selection is permitted and to identify what factors are associated with reporting more than one (non)religious identity. We find support for our hypotheses.\n"]