The aim of this study is to show the different advantages and drawbacks of the main quantitative indicators used in research on diachronic religious behavior. We will demonstrate that religious affiliation has to be used in long-term studies; although it is extremely imprecise, it is often available over long periods of time. Instead, the best solution for medium-term studies is to use frequency of attendance at religious services. This indicator is more accurate than religious affiliation, but is only widely available from the 1960s–1970s onwards. Finally, the most suitable indicator for short-term analysis is obtained from diary-based time-use studies. It is the most precise of the three indicators, but the source is less readily available as these surveys have only been conducted in most countries in the last 30 years. This study is based on data from Time Use Surveys conducted in the Netherlands every 5 years from 1975 to 2005.