Bifulco, Robert, Cobb, Casey D., & Bell, Courtney
Can Interdistrict Choice Boost Student Achievement? The Case of Connecticut’s Interdistrict Magnet School Program
Presents evidence that interdistrict magnet schools have provided students from Connecticut's central cities access to less racially and economically isolated educational environments; estimates impact of attending magnet school on achievement.
Journal Name or Institutional Affiliation:
Educational Evaluation and Policy Analysis
Vol. 31, No. 4, pp. 323-345
- Experimental results (lottery based):
- Estimates of the treatment-on-treated effect from the models that include covariates indicate that the reading test scores of students in these magnet schools are nearly 0.28 standard deviations higher
- Math test scores of students in magnet schools are nearly 0.14 standard deviations higher than what they would be if those students had attended other schools
- Nonexperimental results:
- The eighth grade results indicate that magnet middle schools have had similar effects on the mathematics achievement of suburban and central city students; effects are positive but statistically insignificant for suburban students and only marginally significant for city students
- The positive effects of magnet middle schools on reading scores are larger for students from the suburbs, but they are statistically significant for students from the central cities as well
- Estimates imply that 3 years of exposure to a magnet school in the middle school years increases reading achievement between 0.093 and 0.152 standard deviation for city students and between 0.219 and 0.265 standard deviations for suburban students.
- High school results indicate that, on average, interdistrict magnet schools have had positive and statistically significant effects on the 10th-grade mathematics and reading achievement of central= city students.
Journal Article Historical Analysis
Academic Achievement, Choice, Desegregation, Magnet Schools, Math, Reading, Segregation
Experiment, Secondary Data
Method of Analysis:
5th grade magnet school applicants
Unit of Analysis:
- Sampling frame includes all 5th grade students applying for lottery-based admission to two Connecticut magnet schools (one serving grades 6-8, the other serving grades 6-12) in the Hartford area in 2003 and 2004
- Admissions data was matched to test score records from 2001-2002 to 2006-2007
- Test score data provided two pre-treatment (i.e., pre-magnet school enrollment) measures from the fall of fourth grade and the fall of sixth grade and one post-treatment measure from the spring of eighth grade
- Final sample included 553 students drawn from 11 lotteries
- Treatment group consists of students selected by random lottery to attend magnet schools, control group consists of students not selected by lottery to attend magnet schools
- Additionally non-experimental data set was constructed to determine whether experimental and non-experimental results would vary.
- Used nonexperimental results to estimate average achievement effects for larger sets of interdistrict magnet schools.
- Sample included all students attending one of the two magnet schools or attending a high school in a district that sends at least 10 students to one of the magnet schools.
- 2005-2006 and 2006-2007 8th and 10th grade Connecticut Academic Performance Test records
- 10th grade records matched to earlier 8th, 6th, and 4th grade records; 8th grade records matched to earlier 6th and 4th grade records
- Final sample of 10th graders included 1,073 magnet school students and 6,676 nonmagnet students
- Final sample of 8th graders included 849 magnet school students and 7,045 nonmagnet students
- DV: Mathematics achievement, reading achievement
- IV: Magnet school enrollment, race, age, free lunch eligibility, special education, suburban vs. urban residence