Steven Pedlow
Senior Survey Statistician, NORC at the University of Chicago, Chicago, IL
Steven Pedlow (M.S. in Statistics from the University of Chicago) is a Senior Statistician for NORC at the University of Chicago. He is NORC's imputation expert, specializing in hot-deck imputation and multiple imputation methods, but also has expertise in sample design, weighting, and variance estimation, especially for large face-to-face area probability surveys.![IconGems-Print](images/IconGems-Print.png)
Bias Reduction Through Rural Coverage for the Amerispeak Panel
Steven Pedlow
Senior Survey Statistician, NORC at the University of Chicago, Chicago, IL
Jie Zhao
NORC at the University of Chicago
The AmeriSpeak Panel is a household, multi-client panel to support NORC's mission to deliver reliable data to guide critical programmatic, business, and policy decisions. AmeriSpeak uses the NORC 2010 National Sampling Frame to draw an address-based nationally representative sample of all Americans with full urban and rural coverage. The rural coverage allows the AmeriSpeak Panel to have the most coverage of the U.S. population among multi-client panels in the U.S. The U.S. postal address database used for address-based sampling (ABS) provides city-style addresses for 92 percent of the U.S. population. In areas not well covered by the postal address database, the ABS frame is enhanced with in-person listing, which increases the coverage by an estimated five percentage points. Another benefit is a possible reduction in bias by including these households in rural areas that would be missed by ABS studies. Our focus in this paper is a comparison of AmeriSpeak estimates with and without this rural component against national benchmarks. Our main question is, "For which variables does the rural coverage improve the representativeness of AmeriSpeak?"