Measuring Religiosity across countries and across time

JOSÉ RAMÓN MONTERO; MARÍA CELESTE RATTO; GUILLERMO CORDERO

Córdoba, España

Congreso; V Congreso de Metodología de Encuestas IESA (España); 2008

IESA

Comparative cross-national datasets have become a crucial tool in comparative politics. But their indicators have to a large extent being used regardless of their quality, i.e., of their reliability and face validity. The paper we are proposing deals with one of the most relevant indicators for comparative survey research, namely religiosity. This paper is a preliminary attempt to analyze some methodological dimensions of religiosity measurement. Therefore, we will assess the contribution of several cross-national datasets to understand the religious phenomena. The argument will proceed in three stages. In the first one, we will outline the existing way for studying religiosity using cross-national data sets. In this sense, we are building a dataset containing most of the huge number of existing religious indicators for assessing their reliability and validity. It contains so far 143 surveys, and shows the existing state of the art of religiosity measurements. The most relevant surveys that will be analyzed are European Social Survey, Eurobarometer, Latinobarometer; Asianbarometer; Afrobarometer; European Values Surveys and World Values Survey; International Social Surveys Program; Comparative Study of Electoral Systems; and  New Europe Barometer. We will analyze the items used and its implications for studying religiosity. Secondly, for checking the quality of religiosity indicators we will use Multitrait-Multimethod Matrix (MTMM). It works calculating errors, convergence, and discrimination capacity. The Survey Quality Prediction (SQP) program allows us to check the quality of different measures for religiosity. In the third stage, we will draw on these results to discuss the contribution of each cross-national dataset to the analysis of empirical data and we will propose some tentative ways for improving them. In the appendix we have detail same datasets whose indicators on religiosity have been interpreted in our own dataset.