| This study represents a first attempt to explore other sophisticated statistical techniques. Its results are not conclusive claims. However, its methodology and statistical techniques might contribute, in part, to the enhancement of the SIMCE assessment in Chile. In the early '80s, the Chilean government implemented an educational reform based on decentralization and privatization policies. Chilean schools were divided into three school administrations: municipal, private subsidized, and private schools. Although several educational measures have been implemented, Chilean schools still have significant differences in student achievement. Since 1988, the Ministry of Education has used the SIMCE assessment in order to examine the quality of Chilean education. However, there are several concerns about the ways in which the SIMCE results are analyzed and reported. Usually, the traditional statistical analyses do not provide a full explanation of Chilean education. This study examined the application of Hierarchical Linear Models (HLMs) in order to reanalyze 4th grade students' math achievement by using 1996 SIMCE data. Several types of Hierarchical Linear Models were taken into account. For the effects of schools on math achievement, two models were developed: the One-Way ANOVA Model and the Random-Intercept Model. For the student gender differences in math achievement, three models were developed: the One-Way ANOVA Model, the Random-Coefficient Model, and the Intercept and Slope as Outcomes Model. The primary results revealed that municipal, private subsidized and private schools differed significantly in terms of their math achievement. However, these differences in math achievement tended to be much smaller, after controlling for the effect of socio-geographic variables. Furthermore, the results suggested that student gender, as a predictor, did not have much effect on 4 th grade math outcomes. In the majority of school administrations, 4th grade female students tended to have very similar math achievement to their male schoolmates. Note that I examined mathematics as an entire subject area. In order to explore other possible differences in Mathematics, we need to include more students, classroom and school variables. Also, we need to organize the SIMCE math test in a way in which each sub-topic can be analyzed separately. |
