• Corpus ID: 145099711

Supplemental Educational Services and Student Test Score Gains: Evidence from a Large, Urban School District

@article{Springer2009SupplementalES,
  title={Supplemental Educational Services and Student Test Score Gains: Evidence from a Large, Urban School District},
  author={Matthew Springer and Matthew J. Pepper and Bonnie Ghosh-Dastidar},
  journal={Journal of Education Finance},
  year={2009},
  volume={39},
  pages={370 - 403}
}

• Matthew Springer, Matthew J. Pepper, B. Ghosh-Dastidar
• Published 12 August 2009
• Education
• Journal of Education Finance

This study examines the effect of supplemental education services (SES) on student test score gains and whether particular subgroups of students benefit more from NCLB tutoring services. Our sample includes information on students enrolled in third through eighth grades nested in 121 elementary and middle schools over a five-year period comprising the 2003–04 to 2007–08 school years. A total of 17 elementary and middle schools were required to offer SES at some point during the period under study, and 9,861 student-year pairings in the sample were eligible to receive SES. Although our preferred analytic approach takes advantage of panel data to implement an analysis strategy with student fixed effect regression methods, we also test the robustness of these estimates to a number of alternative approaches, including a comparison of student test score gains between current and future SES participants. We find consistently significant and positive average effects of SES on test score gains in mathematics. Results in reading tend to be insignificant. SES tutoring does not appear to disproportionately benefit a particular racial/ethnic group or ability level. Female students and students with disabilities appear to benefit more from participating in SES. SES has a significant, cumulative effect on students in both mathematics and reading. We also demonstrate that not accounting for content area of tutoring can cause downward bias in estimates of the SES treatment effect. Our findings are qualified on a couple of dimensions.