Archbald, Douglas A.
School Choice, Magnet Schools, and the Liberation Model: An Empirical Study
University of Delaware
This study examined whether school choice that is implemented through magnet schools affects the segregation of low-income students in school systems.
Journal Name or Institutional Affiliation:
Sociology of Education
Vol. 77, No. 4, pp. 283-310
- A typical low-income student is in a school with about 62 percent non-low-income students in the average "no magnet school choice" (NoCHC) district, but in a school with about 45 percent non-low-income students in the average district with "much magnet school choice" (MuchCHC).
- In the nonmagnet districts, 16.6 percent of children aged 5-1 7 are classified as living in poverty, compared to about 23 percent in the magnet districts.
- The size of a district and the demographic characteristics of the community show a clear and consistent pattern of associations with economic segregation among schools.
- Private school enrollment has small effects in two of the models.
- Whether the district is a nonmagnet or a magnet district does not make an additional difference after these other variables are accounted for.
- Neighborhood economic stratification is reproduced in the schools.
- After other variables are controlled, more children in poverty in the community means more children in poverty in the community's school district and less exposure of poor children to children who are not poor.
Journal Article Historical Analysis
Choice, Magnet Schools, Neighborhood, Poverty, Private Schools, SES Composition, Student Assignment Policies
Interviews, Secondary Data
Method of Analysis:
Unit of Analysis:
- Compared magnet-choice districts to central-assignment districts using statistical indices of economic segregation and statistical controls for city and district demographics.
- Uses data collected in 1991-92, in which the U.S. Department of Education conducted telephone interviews with officials of 600 school districts to collect information on magnet schools and other types of policies for school choice and desegregation.
- Interviews were conducted with individuals in the school district offices-usually at the program coordinator or director level-who were knowledgeable about student- assignment policies, magnet programs, and school desegregation.
- This study excluded districts with 5 or fewer schools.
- Final sample consisted of 355 districts.
- Data from 1990 US Census used for control variables.
- Poverty measured as eligibility for free lunch.
- DV: Dissimilarity Index (percentage of low-income children who would have to be redistributed to have the same percentage of low-income children in each school as in the district as a whole), Exposure Index (percentage of non-low-income students in the typical low-income child's school), Standard Deviation (reflects average deviation from mean of the percentage of low-income scores of the schools in a district).
- IV: Student assignment policies such as presence of magnet schools (measured continuously and as "much" /"some"/"no" choice), types of magnet schools,
- Control Variables: Variation in household income, proportion of children living in poverty, mean household income, number of schools, percent Black, percent enrolled in private school.