ABSTRACT
In CA the way distance is measured between profiles is a bit more complicated than the one that was used implicitly when we drew and interpreted the profile plots in Chapters 2 and 3. Distance in CA is measured using the so-called chisquare distance and this distance is the key to the many favourable properties of CA. There are several ways to justify the chi-square distance: some are more technical and beyond the scope of this book, while other explanations are more intuitive. In this chapter we choose the latter approach, starting with a geometric explanation of the well-known chi-square statistic computed on a contingency table. All the ideas embodied in the chi-square statistic carry over to the chi-square distance in CA as well as to the related concept of inertia, which is the way CA measures variation in a data table.
