This article examines the classic question of how religious diversity in a community affects church membership in a period of high growth and social change. Using panel data on local U.S. communities from 1890 to 1926, the authors estimate models specified to overcome likely artifactual problems, deal with unobserved community‐specific heterogeneity, and model state dependence. In general, the findings support the plausibility of mechanisms based on pluralistic deobjectivation and identity activation; they do not support predictions from mechanisms based on organization‐environment matching and interdenominational competition. The findings also show that the overall effect of urbanization on church participation was positive in all but the most religiously diverse communities.