Abstract
This study analyzes the importance of the quality of higher education in the educational input-output process based on two different samples. The sample of fresh college graduates in 1991 is used to measure the impact of school quality on earnings for entry level workers. The sample of college graduate in 1992 with 2-3 years of work experiences is used to measure the impact of school quality on earnings for experienced workers. Six input indices such as “the teacher-student ratio”, “library collections” and “the ratio of teachers with doctor’s degrees” are employed to represent school quality. The potential problem of sample selection bias is taken into account by applying the Heckman’s two-stage procedure and the Bivariate Probit Selection model. The results suggest that in the early 90’s of Taiwan, the input of education does not necessarily have positive effect on the output no matter what kind of sample is adopted, and the school quality effect vanishes with the extension of a graduate’s career. We also find that the impact caused by the input indices varies for college graduates with different fields of major. Commonly, the quality indices show stronger influence on the graduates majoring in natural science than those who major in social science.