We use the National Longitudinal Study of the High School Class of 1972 and the Education Longitudinal Study of 2002 data sets to evaluate changes in the college matching process. Rising attendance rates at 4-year institutions have not decreased average preparedness of college goers or of college graduates, and further attendance gains are possible before diminishing returns set in. We use multinomial logic models to demonstrate that measures of likely success (grade point average) became more predictive of college attendance over time, while other student characteristics such as race and parents’ education became less predictive. Our evidence suggests that schools have become better at sorting while students have efficiently responded to changes in the return to higher education.