This dissertation consists of three essays in applied econometrics. The first two explore how students and parents respond to changes in educational policies. The last one implements bias corrections for nonlinear panel data models with fixed effects.
The first essay addresses the effect of decreasing the age at which compulsory schooling begins in Mexico. A policy change in 2002 required all children between ages 3 and 5 to attend preschool before entering first grade. School entry laws create a discontinuity that induces a sharp increase in preschool enrollment at the expected birthday cutoff. Using this sharp increase, I find that requiring children to attend preschool increases their achievement in elementary and middle school. There is also an increase in parental investments during elementary school. Only parents who would not send their kids to preschool influence the results, as access to preschool was free and universal prior to this policy change.
The second essay uses an experiment in Mexican high schools to analyze how estimates of peer effects change under different types of monetary incentives aimed at increasing math achievement. I find that the estimated peer effect differs between incentivized and non-incentivized groups, and across incentivized groups. I argue that changes in the impact of classroom peers' ability (measured by their individual fixed effects) resulting from the experimental intervention should be interpreted as evidence of peer effects.
The third essay (with Ivan Fernandez-Val and Martin Weidner) develops two Stata commands that provide bias-corrected estimates of panel probit and logit models with two-way fixed effects. Nonlinear panel data models with individual fixed effects can be severely biased because of the incidental parameter problem, and yet these models are used in many empirical applications. The commands developed in this essay correct for both the bias arising from the inclusion of individual fixed effects and the additional bias arising from the inclusion of time fixed effects. They also provide corrected estimates of the average partial effects.
| Identifer | oai:union.ndltd.org:bu.edu/oai:open.bu.edu:2144/33083 |
| Date | 07 November 2018 |
| Creators | Cruz Gonzalez, Mario Alberto |
| Contributors | Lang, Kevin |
| Source Sets | Boston University |
| Language | en_US |
| Detected Language | English |
| Type | Thesis/Dissertation |
Page generated in 0.0033 seconds