School matters: Perspectives on differences in student achievement in Mexico

The quality of schooling in Mexico is considered to be poor, and is characterized by vast differences across states. In the Program for International Student Assessment (PISA) 2003, Mexico was 38th out of the 41 countries that participated in the survey, at about the same level as other Latin American countries but far below the results for developed countries. Mexico's Mathematics scores were about one standard deviation below the OECD mean. However, within Mexico, at the highest end, Colima, Distrito Federal and Aguascalientes are three-tenths of a standard deviation below the OECD mean. Conversely, Oaxaca and Tabasco are at the bottom of the distribution, almost two standard deviations below (Vidal, 2004). Heterogeneous conditions prevail throughout the country, affecting the educational arena. The purpose of my dissertation is to uncover factors that are directly influencing academic achievement in Mexico.;This dissertation analyzes students' performance at the national level to ascertain the determinants of individual and school level variables. Second, it presents a simulation analysis to estimate the scores of the worst performing states (Campeche, Coahuila, Guerrero, Tabasco) with the parameters of Distrito Federal. The third part of the study is an exercise to estimate the average treatment effect of private schooling in urban areas, in an effort to tackle the problem of selection bias. The next section addresses the issue of the effect of students enrolled in a telesecundaria, distance education schools, versus other regular rural schools. The analysis for the third and fourth part is based on propensity score matching. This allows me to compare students that are very similar but that enrolled in different types of schools. Finally, the dissertation focuses on estimating the value of an additional year of schooling, which is calculated using a discontinuity regression approach. This analysis is based on the enrollment cutoff date for primary schooling, trying to show the correlation between an additionally year of schooling and academic achievement as measured by the PISA 2003.;The main findings of this dissertation show that grade level has a significant influence in mathematical achievement---approximately two-tenths of a standard deviation. Also, socioeconomic status at the individual level and the aggregate SES measure at the school level correlate highly with academic achievement. These results suggest that schools have a greater responsibility of providing support and educational resources to underprivileged students that face a lack of assistance from their families and the lack of access to educational materials and opportunities.;This dissertation provides evidence to sustain that the differences among public and private Mexican schools are closely related to differences in the student bodies in these types of schools. This suggests that if private schools are more effective than public, their effectiveness---for example, perhaps to cover more material in the curriculum---is associated with the greater amount of resources their students bring to school and not solely to possible differences in the organization of the type of institution. The estimates show that once individual and school socioeconomic differences are taken into account there is no difference in academic achievement among students from public and private schools.;My work engages in the analysis of telesecundarias, demonstrating that there may be no negative impact on student achievement related to this type of school. Even though there is a gap in achievement between students from telesecundarias and students enrolled in regular rural schools, once the comparison is made among similar students the gap disappears.;Mexico should focus on retaining students in school so that they finish more years of schooling, on closing the gender gap in mathematics and science, and on reducing achievement differences between students attending rural and urban schools.