Using Entropy Balancing to Reduce the Effects of Selection Bias in Afterschool Studies: An Example in Studying the Relationship between Intensity of Afterschool Program Participation and Academic Achievement

Abstract

Every since the enactment of the No Child Left Behind Act (2001) in the United States, achievement gains resulting from afterschool participation have been of particular interest. However, findings have been inconsistent. The challenge for researchers is partly due to the wide variation of program goals, difficulty in obtaining valid control groups, difficulty in obtaining clean records of data, the high transience rates of the students, and in particular, the failure to differentiate among the dosage students receive and the inherent potential of selection bias in the afterschool population. This study draw on a large dataset and allows for the analysis of effects over the course of several years. Using LA's BEST afterschool program as an example, this study employed advanced methodology to reduce selection bias in examining the relations between afterschool program participation and academic achievement.