The Changing Impact of School Suspensions on Student Outcomes: Evidence from North Carolina Public Schools
by Lewis Zhu
Abstract
Motivated by a historic decline in standardized test scores among US students, this paper investigates whether exclusionary discipline—specifically out-of-school suspensions (OSS)— contributes to changes in academic performance. Drawing on administrative data from the North Carolina Education Research Data Center, which span a period marked by substantial discipline policy reform, I assess whether OSS rates are associated with school-level achievement and estimate the effect of OSS on individual student outcomes. I find that these relationships vary over time. As statewide suspension rates have declined in recent years, the negative association between school-level OSS rates and achievement has weakened, while the effect of OSS on
individual student outcomes has grown more negative. One interpretation is that, as suspensions
become less common, being suspended is a stronger negative signal for the child, possibly
inducing stigma and differential treatment which worsen outcomes. Another possibility is that
suspensions have become more targeted, such that those who are still suspended may have
engaged in more serious misbehavior associated with worse outcomes. Meanwhile, school-level
estimates may appear less negative as suspensions now target a smaller group of (on average)
more disruptive students.
Professor Jason Baron, Faculty Advisor
Professor Duncan Thomas, Faculty Advisor
JEL Codes: H75; I21; I24; I28
Keywords: Suspension; Discipline policy reform; Test scores; Student achievement
Affirmative Action and Human Capital Accumulation: Evidence from Brazil Marcos Hirai Catao
by Marcos Hirai Catao
Abstract
In this study, I examine the effects of affirmative action (AA) policies on high school students’ incentives to invest in human capital, focusing on the Brazilian Quotas Law (QL). This law mandates that federal higher education institutions reserve half of their seats for students from public high schools. Utilizing administrative data on schooling, college enrollment, and performance on standardized tests, I observe an increase in test scores among private high school students who attend public colleges. This increase corresponds with the reduction in available non-reserved seats. Conversely, no significant change is observed in the performance of public school students, despite a substantial increase in reserved seats, indicating a potential behavioral response. To estimate the effects of the policy, I analyze variations in policy exposure across regions and cohorts using difference-in-differences methods, which predominantly yield precisely estimated null results. Finally, I discuss potential reconciliations of these, proposing avenues for further research to explain the discrepancies.
Professor Jason Baron, Faculty Advisor
Professor Duncan Thomas, Faculty Advisor
Professor Michelle Connolly, Faculty Advisor
JEL Codes: I2, I23, I24
Bailing on Justice: Plea Bargain Sentencing Outcomes
by Esmé Lise Mailloux Govan
Abstract
In 2020, over 630,000 American adults were detained in local jails each day, 74% of whom had not yet been convicted of a crime. These defendants were detained before going to court because they did not make bail. There is a large body of work documenting the negative impacts of pretrial detention on a variety of outcomes, both short term such as sentencing and long term such as labour market outcomes. However, most of these studies take place in the largest cities in the U.S., which is partly a result of data availability. Thus, it is unclear if these results replicate outside of these urban cores. This paper uses data from Berkshire, which is disproportionately rural and White, to test whether the negative effects of pretrial detention extend to these less studied areas. First, using Durham data, the negative effects of pretrial detention in urban areas that previous studies have reported are replicated. Then, using Berkshire data, the negative effects of pretrial detention are shown to not only extend to rural populations, but are in fact more severe.
Professor Bocar A. Ba, Faculty Advisor
Professor Jason E. Baron, Faculty Advisor
Professor Kent Kimbrough, Faculty Advisor
JEL Codes: K14; K41; K42
Questions?
Undergraduate Program Assistant
Matthew Eggleston
dus_asst@econ.duke.edu
Director of the Honors Program
Michelle P. Connolly
michelle.connolly@duke.edu
