Identifying Students in the Most Need in Low-Income Bolivian High Schools Through Surveys: A Data Mining Application

• In Event: 32.083 - Roundtable Session 12
  In Roundtable Session: 32.083-10 - Social Contexts Research in International Education Settings

Abstract

This paper session describes a series of data-mining techniques used to analyze survey responses for early identification of students who may benefit from reach-out initiatives. The methodology is applied in a low-income environment in a developing country, Bolivia, where the main factor for student abandonment is the need to contribute to their household budget by working, whether in the fields, at home, or in low-paying jobs. The methodology is being tested at schools run by Fe y Alegria:Bolivia, a network of schools in low-income communities, which is instituting a reach-out vocational process to help maximize student learning potential. We describe the methodology, present its application both as a descriptive and as a predictive tool, point out limitations, and invite reactions.

Authors

• Kathleen Campbell, Saint Joseph's University

• Miguel Angel Marca Barrientos, Fe y Alegria Bolivia

• Joao Neiva de Figueiredo, Saint Joseph's University

• Christine Wolf, Deloitte & Touche