Abstract
Despite the long-lasting interest in religious change, debates on the topic have been heated and are still far from being settled. In order to provide a reliable data source through which to study these dynamics, the CARPE project harmonizes well-known international surveys containing items concerning religiosity (the ESS, Eurobarometer, EVS, ISSP and WVS). This makes it possible to broaden the available observation window, both across countries and over time. Moreover, the opportunities this provides for comparing different survey programmes also enable researchers to analyse the consistency of the results, minimizing the impact of random fluctuations and providing useful information with respect to the degree of confidence which can be placed on the relevant estimates. The main focus of this cumulative approach is the variable regarding church attendance, which has been harmonized in various ways. All in all, the CARPE dataset contains figures of religious practice for 45 countries spanning the period of 1970–2016 and derived from 1665 national surveys. This results in a sample of approximately 1.8 million individual observations. The aim of this contribution is to present the dataset’s composition, the harmonization procedure adopted, the strategy used to combine the single datasets and the reliability tests which have been performed. Finally, some possible applications of the CARPE dataset will be introduced.