SAIPE County Poverty Models Using Data from the American Community Survey

摘要

The U.S. Census Bureau's Small Area Income and Poverty Estimates (SAIPE) program produces model-based estimates of income and poverty using data from Census 2000, the American Community Survey (ACS), administrative records, and intercensal population estimates. This work assesses SAIPE county poverty models under log rate and log count data transformations. Equivalent rate and count models are described as the basis for comparison, which differ only in their corresponding dependent variables. Scale invariance and homoskedasticity are assessed, and goodness of fit comparisons are made in terms of root mean square error (RMSE). The estimation results support use of the log count model.