Schools, Neighborhoods and Selection: Outcomes Across Metropolitan Los Angeles

What is the relationship between school segregation and neighborhood segregation across school districts in Los Angeles County, and are school district outcomes on reading and mathematics scores related to levels of school segregation across these districts? We compute segregation scores using US tract level data for 2000 and use reading and mathematics scores from California State tests. Data from the Los Angeles family and neighborhood survey are used to track individual residential changes and differences in the associated Woodcock Johnson scores. We show that there is a close link between levels of school segregation and neighborhood segregation and that many suburban school districts are relatively integrated across both neighborhoods and schools. When we examine average school district outcomes on reading and mathematics scores we do not find any relationship with levels of segregation. At the same time there is clear evidence of spatial sorting with poorer and lower scoring children moving into school districts (or zones with poorer achieving schools). The multi-level models of segregation and achievement show that the variance in achievement levels across districts is significantly greater than across schools within districts.