Contribución de modelos estadísticos a la equidad en la asignación del programa de alimentación escolar. (PAE), 1991

Abstract

Statistical models to classify the primary school in relation to the need of school lunch are presented. The datas come from a special survey done in all the first and fourth courses, at beginning of fate 1990 year. (Half a million of school age children in eight thousand schools). Instrument PAGE was elaborated by #te analysis of the datas by the software called Logistic regression (LOGIT), and the results are better than the wed early in the past four (PTIL) years elaborated by the software known as Vector Analysis, with the annual information of the census of the first course. One region appears to be postponed by the new model, but it is highly explained because of the under-registration in 15% of the urban schools and 19,5% to in the rural ones. The national average of under-registration was 6%, but it was concentrated in the named Region, and in this, only in some countries. The two models are compared in their eficacy in front of four dependant variables, given the advantage to PAJE over PTIL, mainly in rural schools.