The data set is from an US federally sponsored program for providing remedial assistance to disadvantaged primary school students. The firms are 70 school sites, and data are from entire sites. The variables consists of results from three different kind of tests, a reading score, y1, a math score, y2, and a self–esteem score, y3, which are considered outputs in the model, and five different variables considered to be inputs, the education level of the mother, x1, the highest occupation of a family member, x2, parental visits to school, x3, time spent with children in school-related topics, x4, and the number of teachers at the site, x5.