|
Abstract:
|
Oftentimes surveys use weighting adjustments and post-stratification weights to produce reliable estimates that reduce bias and variance. The most recent Residential Energy Consumption Survey (RECS) used weighting adjustments for insufficient addresses, unknown vacancy and primary housing unit status, and nonresponse. After these adjustments, the weights were then post-stratified to several control totals from the American Community Survey (ACS). Post-stratification is a complex process, with the choice of control totals possibly having unknown effects on survey estimates. These processes take a significant amount of time, money, and effort, and there is interest in finding a faster, simpler, yet effective weighting adjustment method for the RECS. This research will consider the impact to survey estimates by eliminating the additional weighting adjustments as well as calculating alternative post-stratification weights.
|
