The paper presents a value-added approach in the context of the literature of school effectiveness. Relevant aspects such as the use of the value-added indicator, the choice of the statistical model, and the scientific properties of consistency and stability in Brazilian education are discussed. The empirical evidence presented is based on the variance component models applied to GERES 2005 data collected in the municipality of Campinas. The results show that the indicator obtained by the contextualized results model is rather different from the one obtained by a value-added model, supporting the need of conducting longitudinal data studies. The findings about consistency and stability suggest the value-added usefulness to timely diagnose schools that systematically perform as effective schools or as ineffective schools.
School performance; Evaluation; Value added; Multilevel model; GERES 2005
