In search of good teachers: Patterns of teacher quality in two Mexican states

Mexico's national teacher incentive program awards significant salary bonuses to teachers based on attributes such as educational attainment, experience, and scores on tests and training evaluations. Yet there is little consistent evidence that these attributes are positively associated with student outcomes. This study uses longitudinal data on teachers in the Mexican states of Aguascalientes and Sonora to examine whether the teacher attributes rewarded by the national teacher incentive program are positively associated with students' gains in achievement. The study also explores whether all children have equal access to teachers with important attributes. To identify important teacher attributes, I use a value-added multiple regression approach that relates changes in student achievement scores over time to changes in teacher characteristics. To control for the non-random assignment of teachers into schools, I use school fixed effects, thereby isolating variation to changes within schools over time. After identifying teacher attributes that are consistently and positively related to student achievement gains, I describe how teachers with important attributes are distributed according to average school achievement, municipal poverty, and across urban and rural schools.; I find that teachers' scores on tests or training evaluations are consistently and positively related to student achievement gains in both states, while teachers' education levels and experience are not. I also find that teachers with higher scores are more concentrated in schools with higher average achievement, lower municipal poverty, and urban schools, patterns which are likely to have perverse effects on educational equity in both states. One exception is the case of schools with high municipal poverty in the state of Sonora. Teachers' scores in these schools are not significantly different from teachers' scores in schools with very low poverty, but they are significantly higher than teachers' scores in schools with medium poverty. These results may indicate an equity-enhancing effect of incentives to teach in high-poverty schools. The results of this study can help to inform decisions regarding teacher incentive, hiring, and assignment policies in Mexico. They also contribute to the broader debate over the components and distribution of teacher quality.