Secularization Index Modeling

An age effect means individuals change religiosity as they grow older; a period effect means an event or era shifts everyone at once; a cohort effect means each birth generation enters adulthood at a different baseline that it largely keeps for life. Secularization modeling separates these because they imply very different futures: an age effect could reverse as a population ages, a period effect can rebound, but cohort-driven decline is largely irreversible because it proceeds through generational replacement. Voas found European decline to be overwhelmingly cohort-based.

Fuzzy fidelity is Voas's term for the large group that is neither actively religious nor firmly secular - people who believe something vaguely, identify nominally, or attend occasionally. It matters because it is a transitional state, not a destination: it expands as societies move away from committed religion and then contracts as later cohorts become more definitely secular. Recognizing it prevents two errors - mistaking nominal affiliation for genuine religiosity, and mistaking a society still full of fuzzy believers for one that has finished secularizing.

Because cohort equals period minus age, the three are perfectly linearly dependent, so a regression cannot identify all three sets of effects simultaneously without additional constraints. Analysts impose identifying assumptions - fixing or constraining one component, using nonlinear specifications, or bringing in external information. These choices can materially affect the results, which is why responsible secularization modeling is explicit about its identification strategy and checks the robustness of the cohort-versus-period conclusion.