A comprehensive review of the literature regarding the impact of college on students has indicated that the most striking principle in these studies is the tendency of initial differences in characteristics among groups entering diverse settings to increase over time. The purpose of the present study is to investigate this phenomenon of accentuation by analyzing the nature of its occurrence and the distinctive processes that account for its existence at both the group and individual level. The study is an exploratory one designed to test the adequacy of two alternative models of accentuation that change at the two levels. Using several data sets, this study assesses the extent to which accentuation does occur at each level and the nature of the relationship between the different types of accentuation across the two levels. The analyses in this study demonstrate that when accentuation is found, there is no consistent relationship between the pattern of selection and change at the individual level, indicating that neither of the models had any explanatory power for the processes at work in the data sets. (Author)