Impact Study of a Predictive Approach to Student Achievement and Targeted Pedagogical Interventions

• In Event: Using Data to Predict Student Outcomes

Abstract

This paper presents an approach where predictive modeling is used at the student level to anticipate future achievement on large scale tests. For prevention purposes, the approach also provides guidance in reducing identified risk factors and targeting individual pedagogical interventions for students at risk. Students’ report cards are used as a source of independent variables. Models are adjusted for sex and for differences in grading severity between schools. It is considerably advantageous and economical to anticipate results on large scale tests from an already existing source of data like the report card. The approach has been used for a few years. We propose to present the results of an impact study.

Authors

• Serge Boule, Centre canadien de leadership en évaluation

• Mario Gagnon, Ontario Ministry of Education