Typology of Orthodox Russians: the problem of constructing a generalized religiosity indicator

There are many approaches to religiosity measurement in sociology of religion, but most of the methodological literature is devoted only to the first step of operationalization of this concept — defining the aspects of religiosity and choosing the indicators for each of these aspects. At the same time almost no attention is paid to the inverse operation — the construction of the generalized measure of religiosity based on the operationalization developed. We consider three basic approaches to this problem and discuss their strengths and weaknesses. The first approach is using individual indicators. It allows minimizing the loss of information about the respondent’s answers, but it lacks generalizability and doesn’t allow for comparison of religiosity effects if religiosity is measured with different indicators. The second approach is construction of a generalized measure of religiosity on the basis of several indicators using summation, averaging, the principles of the strongest or the weakest response, and Factor Analysis. An advantage of this approach is that it produces a religiosity scale of at least ordinal measurement level. But one of the problems that emerge when we try to apply Factor Analysis is that it implies a measurement model with reflective (effect) indicators while the theory rather favors the formative measurement model. The principle of the strongest response overestimates the number of highly religious people while the opposite — the principle of weakest response — underestimates it. The third approach is the construction of a typology on the basis of several indicators with relatively homogeneous religiosity groups using Cluster Analysis or Latent Class Analysis. The advantage of this approach is that it doesn’t imply unidimensionality of religiosity phenomena and tries to find types that actually exist in the data. We come to the conclusion that in current situation, characterized by the plurality of social forms of religiosity, it would be most productive to use the latter approach. Further, we develop a typology of Russian Orthodox Christians based on the data of the Orthodox Monitor survey (2011) using Hierarchical Cluster Analysis. The typology is based on four indicators: believing in God, frequency of Church attendance, frequency of attendance at religious services, and frequency of Confession and Communion.