Based on four years of student-level achievement and demographic data provided by the Hartford Public Schools (HPS), our quantitative analysis sought to answer two questions: (1) Continuity: Who stays and leaves the HPS dataset, and are these behaviors associated with student characteristics, school composition, or neighborhood demographics? (2) Clustering: Are high-achieving students widely distributed across the district, or are they more likely to be clustered with peers who have similar characteristics, or attend similar schools, or reside in similar neighborhoods? By analyzing statistically significant patterns among over 33,000 Hartford-resident HPS students in grades 3 to 8 from 2008-09 to 2011-12, we found that the proportion of high-achieving students who left the HPS dataset is not significantly different from the proportion who stayed (around 15 to 18 percent) over time, but there are significant differences in school zone, magnet school status, and other variables.
Zannoni, Diane, Jack Dougherty, Ben Rudy, and Evan Sternberg. Student Continuity and Achievement Clustering in Hartford Public Schools, 2008-2012: A Preliminary Data Report. Cities, Suburbs, and Schools Project. Hartford, CT: Trinity College, January 18, 2013. http://digitalrepository.trincoll.edu/cssp_papers/.