Despite decades of research and policy efforts, socioeconomic disparities in educational outcomes persist in the United States. In this dissertation, I examine these disparities, focusing on access to higher education during the later years and on academic achievement during elementary school. I explore the importance of family, neighborhood, and school contexts and introduce new approaches for conceptualizing, measuring, and analyzing educational disparities. My findings reveal significant neighborhood disparities in early academic achievement and widening gaps in college application and enrollment, underscoring the need to identify viable solutions. The dissertation begins with an overview of the research questions and theoretical framework in Chapter 1, and concludes with a discussion of implications and future research directions in Chapter 5.

In Chapters 2 and 3, I explore disparities in higher education application and enrollment, focusing on the family context. In Chapter 2, I expand upon prior research by examining trends in socioeconomic gaps in not just whether students enroll, but also where they apply and enroll. I find that, despite decades of efforts to broaden access, socioeconomic gaps have actually become more pronounced over time, even as aspirations for educational attainment increased across all family backgrounds. Chapter 3 explores the role of financial information to socioeconomic gaps in college application, and finds that equalizing information for both students and their parents reduces differences in the perceived affordability of 4-year colleges by about one-fourth and in college application by nearly one-fifth, though it actually increases socioeconomic differences in application to highly selective colleges.

In Chapter 4, coauthored with Geoffrey T. Wodtke, I focus on neighborhood disparities in elementary school achievement. We assess whether neighborhood differences in school contexts contribute to neighborhood test score gaps, challenging prior studies that draw on limited measures of school quality. Using 171 school context measures across five dimensions – composition, resources, instructional practices, climate, and effectiveness – and applying machine learning methods for high-dimensional data, we find that only school composition and climate vary meaningfully between high- and low-poverty neighborhoods, and that differences in school contexts account for just 4% to 8% of the neighborhood poverty gap in student test scores.