This chapter intends to data analysis for the social sciences. In quantitative data analysis, the main goal is usually to describe the ­distribution on a single variable and the association between the units' values on two or more variables, and to draw inferences from a sample to the relevant population. In the framework oriented toward model testing, one specifies a model for the associations between a limited number of variables. Depending on the statistics of the model fit, the model is either accepted, revised or rejected. If several models fit the data, the most parsimonious model is usually the one that is selected. The affinity between Bourdieu's multidimensional, spatial approach to studies of power and inequality and the geometric approach to statistical analysis has been strong. This has particularly been the case in studies of the associations between social classes and cultural preferences and in analyses of elites.