In this paper, we propose a method that can be used in intercensal years to calculate the coefficient of variation (CV) for the Minimum Change method estimates of the number of children in poverty for school districts used in the Small Area Income and Poverty Estimates (SAIPE) program at the U.S. Census Bureau. The Small Area Income and Poverty Estimates program provides estimates for selected income and poverty statistics for states, counties and school districts. The Minimum Change methodology, outlined in Maples and Bell (2007), incorporates current IRS income tax data about sub-county-level poverty for school-age children. These estimates are used in the administration of federal programs and the allocation of federal funds to local areas. Additionally, we will attempt to empirically quantify the possible improvement in CV that might be made by improving the geocoding process (assigning the address of an income tax return to a census block) to reduce the percentage of nongeocoded exemptions. Comparisons of CVs using appropriate year IRS income tax data for school district poverty will be made against CVs using only the Census longform CVs from 2000 and 1990. The Minimum Change method will use Census 2000 as the “previous census” when estimating poverty in 1990.

School district estimates for the number of poor children are the sum of their school district piece estimates. School districts that cross county lines are split into pieces that correspond to the intersection of county and school district. Making estimates at the level of a school district piece rather than as whole school district allows for a simpler method of controling the number of poor school-age children to the county level estimates to maintain consistency between different geographical levels.