La educación secundaria y sus dos dimensiones. Efectos del barrio y del colegio sobre los resultados saber 11

This paper studies the mathematics and language results of 32,000 students on the Saber 11 test for 2008 in the city of Bogota D.C. This analysis considers that individuals are contained in neighborhoods and schools, but not all individuals from the same neighborhood attend the same school, and vice versa. With the purpose of creating a proper model for this data structure various econometric models were used, including a crossed random effect multilevel hierarchical regression. The central objective is to identify the extent to which neighborhood and schooling conditions are correlated with the educational results of the objective population, and which neighborhood and school features are more strongly associated to this test’s results. We used data from the Saber 11 test, the C-600 school census, the 2005 population census and the Bogota D.C. metropolitan police department. Our estimations show that both neighborhoods and schools correlate with this test’s results; but the school seems to be a much stronger factor than the neighborhood. School features that have the strongest correlation with these test’s results are the teacher’s education, the school day’s schedule, schooling expenses and the school’s socioeconomic context. Neighborhoods features that are mostly associated with these test’s results are the presence of university students within the upz, a cluster of higher educational levels as well as the crime rate within the neighborhood, which correlates negatively. Previous results were found taking in account family and personal controls.