El uso de las Técnicas de Segmentación en la evaluación del rendimiento en lenguas. Un estudio en la comunidad autónoma vasca

Abstract

This paper is an illustration of the use of segmentation statistical techniques in the data analysis of an evaluative research. The CART method has been applied to a set of data on academic achievement in Basque and Spanish language. The data set was obtained from a sample of Secondary Education students in the Basque Autonomous Community. The aims were to devise and purge a predictive model of the achievement in these areas, and to assess the possibilities that the use of this kind of technique offers both in the data analysis phase of evaluative research and in that of dissemination of results. The results obtained through segmentation have been triangulated by means of multiple regression, factor analysis and multiple correspondence analysis. The results basically agree, but segmentation has the advantage that one can operate simultaneously with both quantitative and qualitative variables. Furthermore, the graphical results offered (decision trees) are very easy to interpret.