Experimenting with School Attendance Area Free Lunch Estimates: A Census 2000 Special Tabulation Case Study

Douglas Geverdt, U.S. Census Bureau

Free and Reduced-Price Lunch counts from the National School Lunch Program (NSLP) have become the de facto standard for identifying school-level poverty. Although these data have significant and well-known shortcomings, including age bias and administrative incentives to over-report disadvantaged students, no other systematic source of school-level poverty data currently exists. This lack of an independent, comparable school-level poverty indicator hampers attempts to evaluate NSLP data quality and program efficacy. This case study analysis used a GIS to define and construct boundary data for 233 schools in a large urban school system, and then applied Census 2000 sample data to create school-area estimates of Free Lunch and Reduced-Price Lunch eligible students. These school-area Census estimates were then compared to school-reported NSLP counts. The findings show a number of systematic differences between the reported and estimated values, even after accounting for differences in school demographics and NSLP data collection methods.