Selected ongoing projects
Classmates: College Context and Socioeconomic Disparities in College Outcomes
(Presented at the 2019 PAA Annual Meeting)
This paper extends previous research by proposing, and evaluating a meso-level explanation for socioeconomic inequality college outcomes: college social context. Colleges are bounded social environments that facilitate students’ educational experience and social interactions, and can influence student goals, perspectives and behaviors in college and can be consequential for student outcomes. I evaluate this argument with information from a nationally representative cohort of high school students in the 2000s (ELS 2002), along with detailed and external information on the socio-economic context of students’ campuses obtained from the College Scorecard data. Results show that (1) low-SES students are substantially more likely to enter lower SES campuses, even net of a comprehensive set of social, demographic, academic, and organizational characteristics. (2) Net of these factors, campus SES context is positively associated with students’ likelihood to obtain a BA, and negatively associated with their likelihood to dropout, or to obtain an associate’s degree or certificate. (3) This relationship holds even net of college characteristics, including college selectivity, sector, revenues, student-faculty ratio, and size, and does not vary by student background. Results from a weighted regression technique suggest there is little heterogeneity in the treatment effect of college SES context, and it is not the result of negative selection. (4) Decomposition analyses suggest that campus context account for a larger share of the net socioeconomic gap in college outcomes than college selectivity, resources and size together. Taken together, college socioeconomic status emerges as an important, yet unexplored, mechanism that inhibits the intergenerational mobility of low-SES students.
"Pipeline Dreams? Gender Differences in Occupational Plans and STEM Major among US College Entrants” (with Kim Weeden and Stephen L. Morgan)
We use data from the Educational Longitudinal Survey (2002-2012) to document and examine the sources of the gender gaps in the probability of earning a BA in either a STEM or biomedical field or a health field. Even among this recent cohort, gender differences in the completion of these two types of science-related degrees are substantial, and they are strongly predicted by gender differences in the content and stability of occupational plans, but not by gender differences in prior academic achievement and math and science preparation, family-work orientation, or self-assessed math ability. Attrition from STEM/Biomed majors is also gendered: women who had declared STEM/Biomed majors as sophomores were more likely to leave without a degree than they were to graduate with a degree in STEM, the reverse pattern as men. Among the subset of college sophomores who declared STEM/Biomed degrees, gender differences in academic achievement attenuate the gender gap in STEM persistence (because of STEM women’s higher GPAs), self-assessed math ability and family-work orientation have no effect on the gender gap, and gender differences in occupational plans account for between 10 and 20% of the gender gap in STEM persistence. (Invitated to Revise and Resubmit at Sociology of Education)
How Organizational Cultures in US Higher Education (may) Promote Inequality: Analyzing College Mission Statements with Networks-Based Methods for Text Analysis. (With Eric Gladstone)
This project evaluates the impact of educational organizational culture in generating stratification in student outcomes in US higher education. We posit that organizational cultures, defined here as shared understanding about the preferred modus operandi of postsecondary institutions, both differentiate between organizations, and provide guiding principles that shape the social and academic environments students encounter in higher education, rendering them potentially consequential for social stratification. We evaluate this argument by analyzing the mission statements of allaccredited postsecondary institutions in the US higher education in 2018. Mission statements are codified, often digitally represented, strategically written statements aimed at communicating the core values, missions, and purpose of the organization to external and internal audiences. They are practically ubiquitous in higher education. We argue that mission statements are cultural products, reflecting shared understanding about the desired modus operandi of higher education institutions, regardless of whether they reflect desired or existing culture in the organization. We analyze the entire corpus of mission statements in higher education using automated network-based text analysis methods that construct semantic networks that depict cultural themes in higher education, and their relations. From there, we are able to place the network of cultural concepts in an n-dimensional meta-social space where we can both identify salient organizational characteristics that define cultural associations, and view organizations as organisms who search out and exploit different types of cultural resources necessary for survival. We then evaluate the impact of the identified variations in organizational culture on student outcomes.
"Unintended Consequences: Educational Expansion and Rising Income Segregation in American Higher Education" (Paper was presented at the 2017 PAA Annual Meeting).
Higher education has responded to the growing demand for postsecondary education over the past few decades with a dramatic expansion of open admission institutions. Changes in the supply of slots at open admission colleges can alter the characteristics of the population of students entering higher education, and extend opportunities to disadvantaged students. But unequal expansion of open admission colleges can also impact the distribution of students across college types, and, consequently, the social environment they encounter while in college. Using data from two nationally representative cohorts of high school students in the 1990s and the 2000s (NELS and ELS), and information on college campuses compiled by the US Department of Education, this study sets out to examine whether and how changes in the availability of slots in higher education in general, and open admission colleges in particular, impact the social environment students encounter while in college, measured here by the average income of students on college campus. Results show that low-income students in the later cohort were more likely to access higher education, especially two-year colleges. However, they were also more likely than their counterparts in the 1990s to enroll in colleges characterized by significantly lower average family income. Thus, students in younger cohorts may have fewer opportunities to engage with students from different social backgrounds while in college. These results, which hold even when social and compositional changes are accounted for, underscore the importance of macro-level organizational processes in higher education to micro-level behaviors and circumstances.