Analyzing Asymmetry of 2-wave 2-variable Panel Data with Generalized Correspondence-analysis and Loglinear Models

In this paper it is out objective to give a graphical display of the asymmetry of a square contingency table formed by two variables measured at two time points. This is done by making combined use of generalized correspondence analysis and loglinear symmetry models. The residuals of loglinear symmetry models have various forms of asymmetry in it. These residuals are submitted to a skewsymmetric form of generealized correspondence analysis. In this way a graphical display of asymmetry is obtained.