ABSTRACT
Previous studies exploring the effect of college quality on individual earnings (or hourly wage rate) use more or less the same method as in Equation 2.1: Individual i’s log earnings (1n(Yi)) is a function of quality of institution j he or she actually attended (Qij), demographic characteristics (Di), family background (Fi), academic background (Ai), job market conditions (J i), and an individual disturbance term (μ i). This model has been tested in several recent studies by Thomas (2000a, 2003) and Thomas and Zhang (2005). The same model is used in this study as the baseline model for several reasons. First, it is desirable to maintain consistency with previous research (provided it is good) so that results can be compared without confusion due to the method. Second, although it is interesting to employ statistically more advanced methods such as correction for selection and HLM, it is my intention to keep the technical aspect of this study as parsimonious as possible. Finally, I experiment with other methods, and results do not differ substantially from the baseline model.
