Coursework: ITP Seminar

Dr. Rob Olsen
George Washington Institute for Education Policy

Well-designed impact evaluations of educational interventions produce rigorous evidence of their causal effects. However, evaluations of educational interventions are often conducted in purposive or convenience samples of schools that are not designed to represent the populations that policymakers care about. This presentation will introduce participants to methods that can be used to assess the representativeness of an impact study sample to the population of interest. It will also introduce methods for improving the generalizability of the impact study findings, including methods for re-weighting the sample to more closely resemble the population. The presentation will focus on Recommendations 5 and 6 from a new guide to generalizability for education researchers that our presenter co-authored with Dr. Elizabeth Tipton from Northwestern University.

Dr. Fabian Pfeffer
University of Michigan

The Higher Education Act (HEA) is the central federal legislation regulating the financing of higher education and student financial aid. The HEA reauthorization of 1992 introduced pronounced shifts in student financial aid. We take advantage of this policy change to identify the heterogeneous impact of federal financial aid on not just the short-term college outcomes of students, such as college access and graduation,but also on longer-term and broader outcomes, including their later economic well-being and demographic transitions. First, we draw on the National Postsecondary Student Aid Study to trace changes in federal aid and federal loans entailed by the HEA-1992 reform. We find that after the policy change students tended to access more money in federal aid, in particular, in the form of federal loans. Second, we draw on the Panel Study of Income Dynamics, which provides rich measures of the socio-economic background of students during high school as well as their own educational outcomes(college access and college graduation), their adult family income,wealth, home ownership, debt, partnership status, and fertility during adulthood. Drawing on a novel difference-in-difference approach,our results suggest that the policy change did not entail direct effects on college access and graduation but that substantial impacts emerge during later stages of the life-course: Those students who have gained additional access to financial aid thanks to the policy reform,were able to achieve higher incomes, lower debt burdens, higher home ownership rates, and even increased their childbearing. As such, theHEA-1992 is one example of an educational policy that impacts non-educational outcomes more so than educational outcomes, once we can trace its longer-term impacts. 

Dr. Soobin Kim
Education Analytics
 
Studentbehavioral outcomes have been identified by recent research as importanteducational inputs that are positively related to achievement and long-termoutcomes. For example, higher attendance is associated with increases instudent achievement, high school graduation, and college enrollment. Similarly,suspensions contribute to achievement gaps and to the likelihood of involvementwith the criminal justice system. Recent studies identified teacher racialdiversity and teacher-student racial matching as factors that could improvebehavioral outcomes for students of color. One intervention that has not yetbeen seriously investigated as an approach to improving behavioral outcomes andteacher diversity involves non-traditional teacher preparation programs. Inthis presentation, I use data from all students attending New York City publicschools in grades three through twelve between 2014 and 2019 and examine theeffects of teachers who are prepared and developed by one such alternativepreparation program on student attendance, suspension, and teacher diversity.Results indicate that teachers trained by the non-traditional preparationprogram are more effective in improving student attendance and suspension. Theprogram effects are substantially larger among students of color and studentsfrom historically marginalized backgrounds. There was greater racial diversityamong first-year teachers from the non-traditional program relative to othernew teachers and thus Black students were more likely to be taught by Blackteachers if the teachers were trained by the non-traditional program. Finally,I provide suggestive evidence that same teacher-student race effects are largerfor non-traditional teachers. Together, these findings provide an importantexample of how non-traditional teacher preparation programs can play animportant role in improving students’ non-test outcomes and remedying largeracial mismatches between teachers and students, and ultimately improveoutcomes for historically marginalized students. 

MMSD Office of Institutional Research and Evaluation
Title: Collaborative Translational Research Part I: The Challenge and the Teams

ITP Fellows Only

No seminar meeting

Dr. Katie Jajtner
Scientist, UW-Madison

Abstract: Education can break the intergenerational cycle of disadvantage but, like socioeconomic status, education is also persistent across generations (Fletcher & Han, 2019; Hertz et al., 2007). This persistence can erode the power of education to fuel the mobility prospects of members of some groups relative to others. To describe variation in the mobility-enhancing potential of education, we examine racial/ethnic patterns in intergenerational education mobility – both upward and downward mobility. Preliminary results suggest persistence in educational attainment across generations may be lower for Hispanic, Asian, and American Indian/Alaska Native (AIAN) demographic groups relative to non-Hispanic White Americans. While Asian Americans appear to have greater upward mobility relative to non-Hispanic Whites, non-Hispanic Black, Hispanic, and AIAN Americans have greater downward mobility. Females across all races/ethnicities tend to experience greater educational mobility relative to their male counterparts.

Lois Miller
UW-Madison

Abstract: Many students attend colleges close to their homes. How important is the proximity of colleges in students’ enrollment choices? In this paper, I explore this question using administrative data on the universe of public high school students in Texas matched to all within-state college enrollment at public 2-year and 4-year colleges. I use an event study analysis around the openings of community colleges to estimate the causal effect of the availability of a 2-year college within 20 minutes driving time of students’ high schools. I find that local students do enroll in the new community colleges, but this enrollment response is primarily substitution away from other colleges rather than the effect of enrollments from new students who otherwise wouldn’t have enrolled in any college. I also find suggestive evidence that students who gain access to a local community college are more likely to complete 2-year degrees. I discuss potential mechanisms and policy implications from these results.

MMSD Institutional Research and Evaluation
Title: Collaborative Translational Research Part II: Presenting to Senior Leaders

ITP Fellows Only