Most public secondary and postsecondary systems of education are badly misaligned. Standards for academic success vary widely between high schools and colleges. This variation poses a significant challenge to students and policy makers, the consequence of which is a great deal of confusion and even ignorance among students about the academic demands of college.

Our study focuses on California’s efforts to improve college readiness through the Early Assessment Program (EAP).  The EAP is an intervention designed to increase the quality of information about academic preparedness available to high school students.  We aim to understand how an increase in information about college readiness affects students’ need for remediation once enrolled. In this paper we extend previous analyses to explore heterogeneous treatment effects of the program across important subgroups. Our aim is to determine whether California’s Early Assessment Program is equally effective for all types of students (e.g. by race/ethnicity, sex, and prior academic achievement).  Access to statewide data now enables us to replicate our earlier study and extend it in important ways.

This study takes place in California; California serves students from a tremendous range of ethnic and socioeconomic origins. Our paper relies on longitudinal student-level data beginning when California public high school students are in 11^th grade and following these same individuals as they enter and proceed through the 23 campuses of the California State University system.  The data span the eight academic years between 2001/2002 and 2008/2009, a time period that includes several years prior to and six years following the statewide implementation of EAP (approximately 640,000 students). Our large samples allow us to investigate whether the effect of EAP varies systematically by student characteristics (e.g. by race/ethnicity, gender, achievement levels), or high school context.

We employ quantitative data and analytic methods to answer our research questions. The quasi-experimental nature of the data will enables us to employ a treatment-control research design to evaluate the effect of EAP on the probability of needing remediation in college. We estimate two types of treatment effects. First, by taking advantage of the temporal disjuncture in the availability of the EAP along with measures of other covariates, we estimate the intent to treat effect by comparing remediation rates for students eligible to participate in EAP by virtue of the year they entered11^th grade (between 2003/2004 and 2008/2009) to the rates of those students ineligible to participate because the program was not yet available (in 2001/2002 and 2002/2003). Second, we estimate the effect of the treatment on the treated by comparing remediation rates for those who do and do not complete the EAP among all those who were eligible for the assessment (cohorts between 2003/2004 and 2008/2009). To maximize the validity of the treatment on the treated estimates we will capitalize on the statistical power afforded us by having student-level data for the population of public school students who complete high school in California between 2000/2001 and 2008/2009. We match treated to untreated students using a variety of approaches, including propensity scores, and difference-in-difference.