AN ANALYSIS OF GRADE RETENTION AND DROPOUT IN THE HIGH SCHOOLS OF SANTA CATARINA USING ADMINISTRATIVE MICRODATA

Abstract

Based on an unexplored microdata database of secondary students from the public schools of Santa Catarina, the article analyzes three points that are rarely addressed in the literature of grade retention and dropout in Brazil. The first is that, in administrative data, school dropout is partially observed through of the attrition, and it is discussed how to solve this with a systematic tabulation of information. Second, in the case of Santa Catarina it is possible to observe family income, which does not happen in the School Census, so this research may be the first to model grade retention and school dropout with Brazilian administrative data controlling this covariate. Third, an econometric approach is taken assuming that unobservable factors can affect both retention and dropout. Thus, the effect of the ultimate retention on the probability of dropout is estimated, controlling the characteristics of the student and the school, it would be close to 35 p.p. However, this effect would be closer to 20 p.p. for students who are not in age-grade distortion, and it could reach 45 p.p. for students who was retained at least once.