Charting fields and spaces quantitatively: from multiple correspondence analysis to categorical principal components analysis

Multiple correspondence analysis (MCA) has started to gain popularity within sociology as a method of mapping ‘fields’ and ‘social spaces’ in the style of Pierre Bourdieu, its capacity to document multidimensional geometric relationships within data being a snug fit for the relational mode of thought he championed. There is a risk, however, of over-relying on MCA when the data suggest alternative methods and, as a result, drawing unsound conclusions. As a case in point, I take a recent analysis of political attitudes in the UK using MCA that drew bold inferences about the relationship with social class and reanalyse the same data with categorical principal components analysis (CatPCA). The results suggest the opposite conclusion to what was originally argued. I thus urge greater methodological flexibility and openness among those wishing to chart fields and social spaces and, more specifically, I make a case for CatPCA as a tool of geometric data analysis.