|Implementing Public School Choice in Charlotte, NC: Impacts on Student Outcomes, Competitiveness and Racial Segregation
|National Bureau of Economic Research (NBER)
|Improving Education Systems [Program Details]
|Efficacy and Replication
Purpose: In this study, researchers examined the impact of implementing lottery-based public-school choice, a systemic school reform strategy focused on allowing parents the right to choose which public school their child may attend within their home school district in Charlotte, North Carolina. Over the duration of this project, the researchers examined the impact of this program on several outcomes, including performance on standardized tests, student absences, disciplinary problems, parental satisfaction, participation in after-school activities, school demand, and racial re-segregation.
THE FOLLOWING CONTENT DESCRIBES THE PROJECT AT THE TIME OF FUNDING
Setting: The Charlotte-Mecklenburg School (CMS) District in North Carolina.
Sample: Student and parent samples will be drawn from the Charlotte-Mecklenburg public schools which serves a diverse population of 110,000 students and is 8 percent Hispanic, 43 percent African American, and 40 percent eligible for free or reduced-price lunch. Researchers will also utilize a dataset of all children who attended (or applied to attend) school in the district from 2001 to 2004. The parents of approximately 105,000 CMS students form the population for the parental choice measures. Approximately 13,000 youth from CMS will be used to estimate the impact of winning the lottery.
Intervention: The researchers will utilize a natural experiment created in 2001 when the Charlotte-Mecklenberg Board of Education chose to replace its three-decades old school desegregation plan for the public school system with a lottery-based school choice system. Under the new plan, the school district was divided into four "choice zones" containing a range of elementary, middle and high school choices, including magnet programs. During spring 2002, parents were asked to submit their top three choices of schools for the 2002-2003 school year. While parents were not restricted to choosing a school from the "choice zone" they were residents of, they would be required to provide their own transportation for their children to attend schools outside of their choice zone.
During spring 2002 parents submitted their top three choices of schools for the 2002-2003 school year, resulting in applications for approximately 95 percent of the district's 110,000 students. Each student was guaranteed admission to a default "home school" within their neighborhood if they were not admitted to any of their parents' top three choices. Additionally, students were guaranteed admission to continue in magnet programs in which they were enrolled in spring 2002.
Approximately 60 percent of the choice applications requested that students be assigned to their home school or a continuation magnet school as a first choice, while approximately 40 percent identified a first choice school to which students had no guarantee of admission. This resulted in an oversubscription of approximately one third of the schools in the district. Using a lottery-based system to determine enrollment in oversubscribed schools, students who were not guaranteed admission to their first-choice school were assigned to priority groups by school and grade. Groups were prioritized in the following order: students living in the home school zone, students who had attended the school in the prior year, free-lunch eligible students (in schools where less than half the students were free-lunch eligible), and students applying to schools within their choice zone. Within each priority group, admission was determined by lottery number; for each school slots were assigned in order of priority group and random number and proceeded until the school's capacity was reached. If a school was not filled by those students whose parents listed it as a first choice, this lottery process would repeat with students whose parents listed it as a second choice. Approximately 19 percent of students winning the lottery to attend the first-choice school selected by their parents subsequently attended a different school. Ultimately, for the fall 2002 semester, the school choice process resulted in a re-sorting of students across schools in the district. Approximately one-third of continuing elementary school students and two-fifths of continuing middle school students attended a different school than they had attended the previous year.
Research Design and Methods: The research design has four major components. The first will use actual parental choices to determine how parents in Charlotte-Mecklenburg traded-off considerations of school characteristics, such as school mean test scores, geographic proximity, racial composition, and the availability of magnet programs, when they prioritized which school they wished their child to attend. The second will use the randomized lottery results to estimate the impact of students being offered their parents' first-choice school on a range of outcomes including: student performance on standardized tests, student absences and disciplinary problems, parental satisfaction and participation in after-school activities. The third component will focus on testing whether actual impacts on academic achievement are consistent with parental preferences. Specifically, it will examine whether achievement impacts were any larger for the subgroup of students whose parents placed the highest value on school mean test scores in making their choices. The final research component will investigate whether factors related to the school choice policy contributed to racial re-segregation in Charlotte-Mecklenburg School District, will investigate how the public-school choice system combined with parental preferences to create a competitive environment for individual schools, and will advise the Charlotte Mecklenburg School District on ways to improve that competitive environment.
In phase 1, researchers will assemble student data and merge it with the characteristics of each parental school choice, the lottery results, the geographic boundary files, and responses to parental satisfaction surveys. Researchers will also develop a statistical package and an estimation theory for determining whether the impacts of school choice on student achievement are consistent with parental preferences. In phase 2, they will estimate the models of parental school preferences needed to determine the trade-offs parents made when choosing the schools they wished their students would attend with the Charlotte-Mecklenburg school system. They will also estimate the impacts of students being offered their parents' first choice school on various student outcomes. Phase 3 will involve merging previous findings to identify the subgroups of youth for whom researchers would expect the largest academic achievement impacts. The final tasks in phase 4 will include studying the causes of racial re-segregation and the competitive environment facing individual schools, and will also involve making recommendations to the school district.
Key Measures: Researchers have access to student level data from the Charlotte-Mecklenburg school system for the two years before and two years after the implementation of the school choice program that include demographic characteristics, residential location, parental school choice decision, lottery results, academic achievement, attendance, and self-reported time spent on studying. Additional school and neighborhood level data are also available.
Data Analytic Strategy: Researchers will use a mixed logit discrete choice model to estimate parental valuations of school characteristics such as travel distance, test scores and racial composition. OLS regressions will be used to estimate the impact of winning the lottery, the impact of attending one's parent's first choice school, and the impact of students' peers on various educational outcomes. Researchers will then use results from the choice model to identify the subsets of students for whom the researchers would expect large test score impacts, based on the parents revealed preference for academic achievement in their choice of schools. Lastly, researchers will use the choice model estimations to explain racial re-segregation, competition between schools for students, and whether parents are reporting their true preferences.
ERIC Citations: Find available citations in ERIC for this award here.
Hastings, J.S., Kane, T.J., and Staiger, D.O. (2006). Gender and Performance: Evidence From School Assignment by Randomized Lottery. American Economic Review, 96 (2): 232–236.
Hastings, J.S., Kane, T.J., Staiger, D.O., and Weinstein, J.M. (2007). The Effects of Randomized School Admissions on Voter Participation. Journal of Public Economics, 91 (5): 915–937.
Weinstein, J.M. (2016). The Impact of School Racial Compositions on Neighborhood Racial Compositions: Evidence from School Redistricting. Economic Inquiry, 54 (3): 1365–1382.
** This project was submitted to and funded under Education Policy, Finance, and Systems in FY 2005.