ABSTRACT

Underlying all international survey programs is the conviction that cultures can be compared. In the empirical analyses, comparability problems are either ignored or analyzed in terms of measurement equivalence. Criteria of metric and scale equivalence are so strict, however, that they often lead to a double reduction – a reduction of the number of indicators and a reduction of the number of countries. As a consequence of these reductions, the meaning of the broad theoretical concepts is often narrowed.

This empirical analysis on religiosity attempts to overcome these limitations, first by defining softer criteria of comparability. We consider a theoretical concept as fruitful if its relationships to the multiple indicators as well as to other independent and dependent variables are sufficiently similar to each other in all countries under investigation. In the early stage of comparative research only a moderate degree of similarity should be required. In applying this standard, it turns out that the various aspects of religiosity have to be specified by different concepts. Religiosity therefore rather designates a conceptual framework than a single concept. The concepts we use display an astonishing degree of similarity. Not only can they be reliably measured but most of them are also similarly related to gender, age, education, happiness, sexual morality and volunteering in almost all ISSP countries. This is particularly true for a two-item index of religious involvement.