Cross-national comparisons frequently involve working with rates (aggregate data, namely collected from Eurostat statistics) to identify countries profiles. This statistical design deals with quantitative indicators i.e. rates, and qualitative indicators, in this case the country. Quantitative and qualitative variables can be mixed by applying Categorical Principal Component Analysis [1, 3, 5], thus enabling us to identify the structure defined by the quantitative variables and to overlap the countries on that structure. In a second step, the detailed description of the profile types was achieved from a Multiple Correspondence Analysis following the categorization of quantitative variables in a suitable form for MCA. Different approaches to transforming quantitative variables into a set of categorical ones are found in the literature [1, 2, 4]. However, in light of knowledge of the data structure from CatPCA results, we explored an alternative means that simultaneously identifies the best cut-points for defining the new classes and the best number of classes in order to minimize the loss of information due to categorization. The illustration of the proposal linkage between CatPCA and MCA was supported by reported crime rates in European Union countries.

--

Categorical Principal Component Analysis (CatPCA),Multiple Correspondence Analysis (MCA),categorizing quantitative data,reported crime rates