Um estudo sobre o plano amostral do Saeb

Abstract

The case studied discuss aspects of Saeb’s data collection process, specifically about sampling and estimation, to answer the question: In view of the various changes in the elements that guide the data collection process of the Saeb since its creation, is the adopted model still adequate? This thesis aims to analyze the Saeb’s sampling design and how to estimate its results, looking for opportunities to improve the methodologies. To achieve this goal, this thesis presents the history of educational reforms around the world and how it influenced the emergence of largescale assessment in the world, in Brazil and the creation of Saeb. The thesis also brings the evolution history of Saeb and its sampling design to understand the calculation of results using the current sampling design, trying to find limitations in the process. After that, it was possible to suggest strategies to reduce the effects of these limitations, making Saeb more adequate to large-scale assessment in Brazil. The theoretical reference tells about the origin of large-scale educational assessment, how this movement influenced the Saeb’s creation and the evolution of its sampling design through the years. As hypothesis, it was assumed that after many changes the Saeb's sampling design must be studied and analyzed to understand how it can be updated. The employed methodology was bibliographic and document revision, followed by analysis collected data from Saeb 2017, using simulation to evaluate the efficiency of current sampling design, comparing estimated results from Sampling and results from Census. The comparison between different methodologies shows that the current methodology is efficient and adequate, but there are opportunities to improve it. It was also assumed that as a large-scale educational assessment, Saeb has characteristics and limitations natural from this assessment model. The knowledge of this context leads the opportunities of improvement for the assessment model and create an action plan that can make changes and that have a capacity to evolve Saeb. Therefore, the action plan presents some suggestions to improve the sampling design, to enlarge the data collection cycle, with the goal to obtained deepen information through the data collected and given context inputs for the assessment results.