We contribute to an understanding of the determinants of voter support for US prohibition policies in the early 1900s, by separating substantive preferences for wet versus dry policies from preferences for centralized versus decentralized control. Prior studies of prohibition referenda have generated various conclusions about which groups supported and opposed prohibition, whether regarding the role of religion, urban/rural residence, immigrant status, gender, or class. But none of these studies has considered the impact of preferences regarding decentralization on voter support for prohibition measures. We exploit a combination of referenda unique to the 1933 Ohio ballot, where voters considered prohibition-repeal measures alongside a county home-rule amendment. By viewing support for home rule as a proxy for decentralization preferences we clarify and explain anomalies in prior studies regarding determinants of support for prohibition and its repeal, especially regarding urban counties and some evangelical denominations, which are shown to have been guided by a preference for local control of alcohol policy, and counties with larger proportions of women, which are associated with greater support for more centralized and uniform alcohol policy.