Association for Public Policy Analysis & Management

The use of virtual courses has been rapidly expanding in the K-12 sector over the last several decades. One reason for the burgeoning popularity of these courses is that they offer students the possibility of allowing atypical course progression; for instance, they allow students to remediate courses to catch up with their peers on their own schedule. Survey research suggests that the desire to facilitate off-time course progressions are a factor that motivates administrators to offer them (Picciano et al., 2012). Popular media reports suggest that these considerations may be impacting summer school offerings: For instance, a 2012 National Public Radio story[1]reported that many districts in Florida have substantially scaled back summer school offerings, instead encouraging students to enroll in online courses to remediate courses failed during the school-year.

However, little is known about how student course progression differs following online remedial courses compared to face-to-face alternatives. We exploit variation in face-to-face summer offerings to explore whether virtual or face-to-face alternatives better promote student academic outcomes for 9^th and 10^thgrade students in Florida. Specifically, we use an instrumental variables strategy that exploits variation over time in whether a school offers a particular course in a face-to-face summer school session to predict the likelihood that students who fail the course during the academic year take online (rather than face-to-face) remedial courses. Florida is an ideal setting for this study in that it 1)hosts the largest K-12 virtual schooling sector in the country (Watson, 2015), 2)offers highly detailed data that allow us to observe virtual enrollment at the course level, and 3)offers variation over time in the extensiveness of summer course offerings.

Our primary outcomes of interest will focus on course progression. Specifically, we look at whether online remediation affects both the likelihood that students are observed in a future, follow-on course, as well as performance in that follow-on course. For instance, for English I, English II is examined as the logical follow-on course. Focusing on next-course enrollment and performance avoids the concern that contemporaneous course performance outcomes (such as remedial course grade) may be subject to different standards in online versus face-to-face institutions. As a final outcome, we additionally explore whether students who take online courses are differentially likely to persist in school (i.e., to continue to appear in our data as of the year they would be expected to enter 12^thgrade).

These results will have important implications for districts as they consider how heavily to rely on virtual courses to meet the remediation needs of students.