The closer the distributions for the individuals in the subsample are to those of the individuals in the global sample across the active set of variables. The more similar the results from the CSA will be to the results from the Multiple correspondence analysis (MCA), both in terms of dimensionality and in terms of the interpretation of the individual axes. To do a CSA is similar to running a non-normed principal component analysis (PCA) on a subset of individuals on the axes in the MCA. CSA is a newly developed technique that makes it possible to analyze the internal oppositions in a sub-group with reference to dominant oppositions in the global space. In a sociological analysis, a CSA also makes it possible to analyze a field within a field. CSA is therefore a very promising addition to geometric data analysis, and has the potential to open up a whole new avenue of possibilities.