¿Cuánto oro hay entre la arena? Minería de datos con los resultados de España en PISA 2015

Since the start of the PISA evaluations there have been numerous studies that have metaphorically tried «to separate the gold from the sand», in other words, to derive useful knowledge to guide educational practice and policy from the vast amount of data collected. However, research that uses data mining techniques to extract knowledge from the databases provided by the OECD has been less common. This paper analyses the context questionnaires from a metric perspective using a methodology based on data mining with «regression trees». Its main goal is to discover how much value (how much «gold») is in the items that compose these questionnaires, considering their use as predictors of the performance of Spanish students. The results provide a list of the items selected in the six questionnaires and their predictive value. It also provides a methodological approach to help improve the productivity of educational research derived from PISA.