Multicollinearity and ecological regression

The use of the linear models in aggregate voting has now received widespread attention in political science. The aim of ecological regression technique, the use of aggregate data to describe behaviour at the individual ? level. In this thesis Goodman?s classical model of ecological regression technique is developed to estimate voters transitions probabilities. In doing this, ridge regression technique further developed steps and have restricted modified generalized ridge estimator. To illustrate proposed estimation technique, results of two consecutive elections in various electrol districts used for Swedish, Czech Republic and Turkey.