Gas chromatography retention data for 52 solutes on 7 stationary phases were examined by principal component analysis. Without relying on a priori knowledge of solvation parameters and regression analysis, we are able to study the applicability of Kamlet-Taft linear solvation energy relationships and provide independent insight into the most appropriate models. The abstract factors calculated by principal component analysis are correlated with known solute parameters. We compare regression equations, focusing particularly on the need for some type of 'polarizability correction factor', such as Abraham's R2 or Karolet's δ2. The results show that some type of 'polarizability correction' is required and that the continuous R2 parameter is more applicable than δ2.