UC San Diego

In the first chapter of this thesis, the 1995 TIMSS eighth -grade mathematics score is used to proxy for home country education quality for U.S. immigrants. On average, a one standard deviation increase in TIMSS magnifies the marginal returns to post-migrational education by 0.83 percentage points. This pre-migrational education quality effect remains positive and significant for individuals at the 25th percentile of the conditional wage distribution. In addition, diminishing returns to post-migrational years of schooling is observed at all wage quantiles, but evidence is mixed in regards to pre-migrational years of education. Using the 2000 Census, the second paper finds that, compared to another immigrant holding a job that requires less human-interaction, an immigrant worker who possesses knowledge in speaking a non-English language and who works in a human-interaction-intensive occupation would enjoy an average wage benefit of 4.47%. For an immigrant, other immigrants from a different home country are perceived as complements, while those from the same country of origin would be substitutes. Moreover, a one standard deviation increase in bilateral trade volume between the United States and the immigrant's country of origin is predicted to enhance the immigrant's returns to working in the Wholesale Trade industry by 3.36% on average, a pattern that is very different for immigrants whose country of origin uses English as an official language. A positive relationship between parental involvement in reading-related activities before the student began schooling and the student's 2001 PIRLS test score is found in the third chapter. On average, having a parent who played alphabet toys, played word games, and read signs and labels out loud during the student's preschool years is predicted to carry an effect size of 0.2, holding other attributes constant. However, the effect of watching reading programs on television on this test score seems negative. Under a quantile regression framework, the effect of these parental inputs continues to be observed for students belonging to the 25th quantile of the conditional score distribution. Lastly, these academic variables are predicted to not affect an immigrant student's PIRLS score, although small sample size may be an issue