We deduce hypotheses from theories on religious change to explain changes in church attendance rates. Using a new dataset with 51 countries across a long period we apply panel regression models, which enable us to test well-known theories in a more strict and dynamic fashion than do cross-sectional studies.Our results provide new evidence for a few old ideas, but also show striking lack of evidence for ideas that appear well-accepted. Tertiary education proved to be a strong predictor of changes in church attendance. Theories about individualization were also supported. The evidence of existential insecurity as a cause of change was ambiguous: economic development and life expectancy showed significant effects but income inequality did not. We found no support for theories on social globalization and social benefit policy. Finally, we found that income inequality and urbanization were driving forces of change during the 70s and 80s, but not since 1990.