An Alternative to the Census Bureau’s Distributive Housing Unit Method for Estimating Population Size for Sub-County Areas

Warren A. Brown, Cornell University

The U.S. Census Bureau produces population estimates for sub-county areas using the distributive housing unit method. This method distributes county population estimates to sub-county areas based on estimates of housing units (HU), persons per household (PPH) and occupancy rates (OCCR). The Census Bureau holds PPH and OCCR from Census 2000 constant due to the absence of updates. When all sub-areas within a county are not changing to the same degree or direction the variability of error for the estimates increases substantially. We develop and test an alternative approach that uses sample survey data to update estimates of PPH. Census Block Groups are clustered on the basis of distributions of household characteristics that underlie PPH. Households in sample surveys are coded by Block Group Type yielding estimates of PPH that result in a reduction in the variability of error for the estimates as compared with the currently used distributive method.

  See extended abstract

Presented in Session 78: New Directions in Small-Area Population Estimation and Forecasting