A Classification Approach to Mapping Cultural Differences: An Illustration Using Survey Data from 60 Russian Regions

We argue that a classification-based approach to measuring cultural differences across countries or subnational regions is a promising complement, and sometimes an alternative, to the widely used dimensional method in cross-cultural research. The latter summarises cultural variation using continuous dimensions, for example, Hofstede’s famous individualism-collectivism dimension. However, this approach relies on strong parametric assumptions, which are often violated in aggregate cross-cultural samples – particularly when distributions of individual indicators are multimodal or asymmetric, dimensions are highly correlated, or outlying cultural units are present. Using survey data from 60 Russian regions (N = 18,768), we implement the classification approach through latent profile analysis (LPA) to group these regions into discrete types within a multidimensional cultural space. Comparison with conceptually similar dimensional indices shows that the classification-based measure often demonstrates stronger associations with objective geographical, historical, economic, and social characteristics of Russian regions. Furthermore, it reveals additional details when relationships between culture and its potential antecedents or effects are complex and non-linear. Our take-home message is that this approach is a powerful exploratory tool that can uncover patterns in cultural differences and their correlates that might be overlooked by the dimensional approach alone.