Association for Public Policy Analysis & Management

Millions of students arrive at higher education institutions, especially community colleges, only to find out that they could not take college-level classes that count toward their degrees right away. About 60% of entering community college students are referred to remedial education (Complete College America, 2016). Only 16% of these students complete a college-level math course within three years (Bailey, Jaggars, & Jenkins, 2015), and 34% of them obtain a credential within six years (National Center for Education Statistics, 2012). There are mounting empirical evidence suggest that the traditional remediation delays and diverges students from taking college-level courses; it is often not effective in improving academic outcomes for underprepared students (e.g., Boatman & Long, 2018; Martorell & McFarlin, 2011; Scott-Clayton & Rodriguez, 2012).

In light of these pieces of evidence, college systems start to experiment with reforms that compress developmental course sequence and accelerate students into college-level courses. Among these efforts, several colleges and state systems begin to implement the co-requisite remediation: all entering students could take college gateway courses, while colleges provide those who need additional help a concurrent course or lab that offers just-in-time academic support. Early findings on the effects of co-requisite models are promising: a randomized control trial in City University of New York suggests that students who were randomly assigned to co-requisite statistic courses are 16 percentage points more likely to pass gateway math courses, comparing to students in traditional developmental algebra courses (Logue, Watanabe-Rose, & Douglas, 2016).

However, more questions on the effects of co-requisite need to be answered, before one makes any sweeping conclusion about the reform. Is co-requisite remediation effective for both math and English? Does co-requisite remediation improve academic outcomes beyond gateway course completion rate?

In this paper, we intend to address these research questions using administrative data from the 13 community colleges in the Tennessee Board of Regents (TBR) system. TBR is among the first in the country to implement a system-wide co-requisite reform in math and English since 2015. The data we obtain include student demographic and academic attributes, placement test scores, college transcripts, enrollment records, credential attainment, as well as transfer records from National Student Clearinghouse (NSC), from academic year 2010-11 to 2017-18.

We use a difference-in-regression discontinuity (DiRD) design to estimate the causal impact of co-requisite remediation, exploiting the variation of co-requisite implementation timeline across colleges. We identify the effects of placing into remediation within a bandwidth of college-ready cutoff scores both before and after the reform, and then to compare the two. We find that co-requisite is significantly more effective than pre-requisite remediation: students placed into co-requisite learning support are more likely to pass gateway math and English by 15 percentage points and 12 percentage points by year one, comparing to otherwise similar students placed into traditional pre-requisite remediation. We also explored the impacts of co-requisite on longer-term outcomes, including enrollment persistence, transfer to four-year colleges, and early credential completion. But we do not find any significant effects on these longer-term outcomes.