We conduct an empirical test of the relationship between religious pluralism and religious participation in U.S. counties using a fixed-effects panel estimation technique. The empirical technique allows us to control for unobserved heterogeneity across counties resulting from various cultural and historical factors. Contrary to prior cross-sectional research on the 1980 and 1990 Glenmary U.S. counties data, we find a significantly positive relationship between pluralism and participation from panel estimation on the same data. However, we also explain how changes between 1980 and 1990 in the composition of denominations in the Glenmary samples can generate a false positive relationship with the panel estimator. The results show the importance for future research on pluralism and participation of data that have a consistent denominational composition across time.