Skip Navigation
Evaluation of the DC Opportunity Scholarship Program:

NCEE 2009-4050
March 2009

Participation in the OSP

In interpreting the impacts of the OSP, it is useful to examine the characteristics of the private schools that participate in the Program and the extent to which students offered scholarships (the treatment group) moved into and out of them during the first 3 years.

School Participation

The private schools participating in the OSP represent the choice set available to parents whose children received scholarships. That group of schools had mostly stabilized by the 2005-06 school year. The schools that offered the most slots to OSP students, and in which OSP students and the impact sample's treatment group were clustered, have characteristics that differed somewhat from the average participating OSP school. Although 56 percent of all participating schools were faith-based (39 percent were part of the Catholic Archdiocese of Washington), 82 percent of the treatment group attended a faithbased school, with 59 percent of them attending the 22 participating Catholic parochial schools (table 2). Twenty-two percent of treatment group students were attending a school that charged tuition above the statutory cap of $7,500 during their third year in the Program (table 2) even though 38 percent and 46 percent of participating schools charged tuitions above that cap in 2006-07 and 2007-08, respectively.

While the characteristics of the participating private schools are important considerations for parents, in many respects it is how the schools differ from the public school options available to them that matters most. In the third year after applying to the OSP, students in the treatment and control groups did not differ significantly regarding the proportion attending schools that offered a separate library (88 vs. 91 percent), gyms (71 and 72 percent), and art programs (89 and 87 percent). There were the following statistically significant differences (at the .01 level):

  • Students in the treatment group were more likely than those in the control group to attend schools with a computer lab (96 vs. 87 percent), with special programs for advanced learners (48 vs. 32 percent), and that offered a music program (89 vs. 82 percent).
  • Students in the treatment group were less likely than the control group to attend a school with a cafeteria facility (79 vs. 88 percent) or a nurse's office (30 vs. 81 percent).
  • Students in the treatment group were also less likely than those in the control group to attend a school that offered special programs for non-English speakers (26 vs. 57 percent), special programs for students with learning problems (71 vs. 88 percent), counselors (69 vs. 82 percent), tutors (50 vs. 67 percent), and after-school programs (86 vs. 92 percent).

Student Participation

As has been true in similar programs, not all students offered an OSP scholarship actually used it to enroll in a private school. For students assigned to the treatment group, during the first 3 years of the Program (figure 1):

  • 25 percent (346 out of 1,387 students) of those offered an OSP scholarship never used it;
  • 34 percent (473 students) used their scholarship during some but not all of the first 3 years after the award; and
  • The remaining 41 percent (568 students) used their scholarship consistently for the entire 3 years after the lottery.

The reasons for not using the scholarship—either initially or consistently —varied. The most common reasons cited by parents whose child never used their scholarship at anytime in year 3 and who completed surveys were (figure 2):

  • Lack of available space in the private school they wanted their child to attend (22 percent of these parents);
  • Child moved out of DC (21 percent of these parents);
  • Child was accepted into a public charter school (19 percent of these parents); and
  • Participating schools did not offer services for their child's learning or physical disability or other special needs (16 percent of these parents).

The most common responses given by parents whose child initially used a scholarship in year 3 but dropped out of the OSP include:

  • Lack of academic support that the child needed (39 percent of these parents);
  • "Child did not like the private school" (25 percent);
  • There was another private school the child liked better (13 percent);
  • Work at the private school was too hard (11 percent);
  • It was too difficult to get the child to the private school each day (11 percent); and
  • The discipline or rules at the private school were too strict (7 percent).

Students who were partial users were more likely to have special needs and those entering the higher grades averaged lower baseline test scores than students who participated consistently across the 3 years.5

Students who never used the OSP scholarship offered to them, or who did not use the scholarship consistently, could have found their way into other (non-OSP-participating) private schools, public charter schools, or traditional DC public schools. The same alternatives were available to students who applied to the OSP but were never offered a scholarship (the impact sample's control group). Both the treatment and control groups moved between public (both traditional and charter) and private schools or between SINI and non-SINI schools. As a result, over the 3 years after they applied to the OSP:

  • Among the treatment group, 3 percent remained in the same school they were in when they applied to the Program; 46 percent switched schools once; 40 percent switched schools twice; and 11 percent switched three times.
  • Among the control group, 15 percent remained in the same school they were in when they applied to the Program; 40 percent switched schools once; 37 percent switched schools twice; and 8 percent switched three times.

These patterns of student mobility are important because previous studies suggest that switching schools has an initial short-term negative effect on student achievement (Hanushek, Kain, and Rivkin 2004).

Top

5 At baseline, partial users in grades 9-12 were lower performing in reading (27 National Percentile Ranks (NPRs) vs. 40 NPRs for full users, statistically significant at the .05 level) and in math (29 NPRs vs. 49 NPRs for full users, statistically significant at the .01 level); partial users in grades 6-8 were lower performing in math (34 NPRs vs. 41 NPRs for full users, statistically significant at the .01 level); and partial users were more likely to have special needs (5 percent vs. 10 percent for full users, statistically significant at the .05 level).