ABSTRACT
Up to now we have drawn and interpreted correspondence analysis (CA) maps in two possible ways. In the asymmetric map, for example in the row analysis, the χ2-distances between row profiles are displayed as accurately as possible, taking into account the masses of each profile, while the column vertices serve as references for the interpretation. In the symmetric map, the rows and the columns are both represented as profiles, thus the χ2-distances between row profiles and between column profiles are approximated. The biplot is an alternative way of interpreting a joint map of row and column points. This approach is based on the scalar products between row vectors and column vectors, which depend on the lengths of the vectors and the angles between them rather than their interpoint distances. In the biplot only one of the profile sets, either the rows or the columns, are represented in principal coordinates. In fact, asymmetric CA maps, with one set in principal coordinates and the other in standard coordinates, are biplots. But there are alternative choices of coordinates for the other set of points serving as the references for the interpretation.
