Over the last 30 years, policy makers have framed the call for increased proficiency in mathematics among American students as a national imperative (Gardner, 1983; Tate, 1997). As a result, efforts have been made to increase the amount and rigor of mathematics instruction in schools. A body of descriptive research suggests that algebra, in particular, is a key determinant of future academic success and serves as a gatekeeper to higher-level mathematics (Adelman, 2006; College Board, 2000; Ladson-Billings, 1997; Stein, Kaufman, Sherman, & Hillen, 2011). Efforts to improve mathematics proficiency have focused on school algebra, theorizing that participating in algebra earlier will enhance students’ opportunity to access higher-level mathematics (Allensworth, Nomi, & Montgomery, 2009; Clotfelter, Ladd, & Vigdor, 2012; Cortes, Goodman, & Nomi, 2012; Gamoran & Hannigan, 2000). This has led to a proliferation of universal algebra policies, where all students in 8^th or 9^th grade are mandated to take Algebra I (Stein et al, 2011) as well as concurrent debates that these policies may be detrimental to some students—particularly those who lack appropriate preparation for the course (Loveless, 2008). In this paper, we investigate whether and when these underprepared students take Algebra I, and more advanced courses, under a targeted enrollment policy.

Empirical evidence of the success of universal algebra policies is mixed. Some studies show negative results for many students (see, for example, Clotfelter, Ladd, & Vigdor, 2013; Nomi, 2010). Responding to a desire to improve equity in and access to Algebra I in 8^th grade, but cognizant of the concerns around pushing underprepared students into the course, the Wake County Public School System (WCPSS) developed a targeted enrollment policy designed to increase the number of students taking Algebra I in the 8^th grade. WCPSS contracted with SAS to construct a predictive model indicating the probability that a student will score proficient or better on the end of algebra course exam. Students with a predicted probability of 0.70 or higher are advanced in the mathematics sequence in grades 6 through 8 (taking Algebra I in the 8^th grade). Students with lower than a 0.70 probability are assigned to the regular math sequence.

Using longitudinal student-level administrative data from WCPSS, we utilize discrete time survival analysis with an accelerated cohort design to investigate whether and when students who are not eligible to be advanced in the mathematics sequence enroll in Algebra I and courses beyond Algebra I.  We further investigate whether there are differences in timing of Algebra I enrollment (and courses beyond Algebra I) that are correlated with student demographic characteristics such as race, socio-economic status, and disability status. The WCPSS policy aims to restrict enrollment to those students who are most likely to succeed. Little empirical work has examined the trajectories of those students not accelerated by such policies. We anticipate that students who are assigned to the regular math sequence are less likely to enroll in both Algebra I and more advanced classes, and to do so later in their high school trajectories.