Search Results

Title: Diploma Effect
Resulting in 1 citation.
1. Frazis, Harley Jay
Diploma Effect
Ph.D. Dissertation, University of Wisconsin - Madison, 1988. DAI-A 49/12, p. 3825, Jun 1989
Cohort(s): Young Men
Publisher: UMI - University Microfilms, Bell and Howell Information and Learning
Keyword(s): Bayesian; College Education; College Graduates; Earnings; Educational Attainment; Educational Returns; Panel Study of Income Dynamics (PSID)

This dissertation deals with the college diploma effect on earnings. A "diploma effect"--a particularly large return to completing the final year of college compared to earlier years of college--has been found in some studies but not in others. The major part of the dissertation examines the existence of the diploma effect. Most of the analysis is performed using the Panel Study of Income Dynamics (PSID) and the NLS of Young Men. OLS analysis of earnings reveals that neither the addition of IQ, allowance for different time paths of earnings for different educational levels, or interaction of schooling with other characteristics accounts for the diploma effect. To correct for selection bias, a model of choice of level of educational attainment and earnings is developed. The results of correcting for selection bias indicate that selection bias does not account for the diploma effect in either data set. A sensitivity analysis (performed using a Bayesian technique developed by Leamer) shows that the results are not sensitive to the exclusion of family background variables from the earnings equation. Estimated diploma effects are not consistently statistically significant in the NLS or in the PSID. Increasing the sample size by combining the NLS or the PSID with the 1970 Census--treating IQ and selection bias correction terms as missing observations in the Census--the estimated diploma effects greatly increase in magnitude, and the effects are consistently statistically significant. The final part of the thesis examines explanations of the diploma effect. A version of the Spence screening model where family background variables affecting the cost of schooling are observable to both the employer and the analyst implies that selection bias correction should account for the diploma effect. A version where family background variables observable to the analyst are not observable to the employer is not supported by the data. A model is developed where workers signal that they know the degree is expected by obtaining the degree. If such knowledge is correlated with productivity, and under certain conditions, an earnings differential for the diploma above that reflecting the acquisition of human capital can be sustained as one of multiple equilibria. [UMI ADG89-03016]
Bibliography Citation
Frazis, Harley Jay. Diploma Effect. Ph.D. Dissertation, University of Wisconsin - Madison, 1988. DAI-A 49/12, p. 3825, Jun 1989.