COVID-19 Vaccination Coverage in the U.S.: Comparison Among National Probability-Based Surveys - National Immunization Survey Adult COVID Module, Household Pulse Survey, and the AmeriSpeak Panel

Thursday, May 12, 2022

Vicki J. Pineau

NORC at University of Chicago
Jason Fields

US Census Bureau
James Singleton

US Centers for Disease Control and Prevention
Elizabeth Allen

NORC at the University of Chicago
David Dutwin

NORC at the University of Chicago
Sarah Kornylo

NORC at the University of Chicago
Xiuli Tang

NORC at the University of Chicago
Michael Chen

US Centers for Disease Control and Prevention
Benjamin Fredua

NORC at the University of Chicago

This paper presents exploratory research into differences in survey estimates of the proportion of adults reporting receipt of =1 dose of a COVID-19 vaccine in the U.S. by three national probability-based surveys from early May to December 2021: National Immunization Survey Adult COVID Module (NIS-ACM), the Household Pulse Survey (HPS), and the AmeriSpeak Panel Omnibus survey. Though each survey is probability-based, each has unique features (e.g., sample frames) that are in some respects expected to produce somewhat different estimates of COVID-19 vaccination coverage; while other features (e.g. question wording) would be expected to produce similar estimates to each other and to official vaccination administration data available in the CDC COVID-19 Data Tracker. We explore estimated vaccination coverage for subpopulations across the three surveys to CDC's administrative data for receipt of =1 dose of a COVID-19 vaccine. Some findings as to the potential impact of differing sample designs, unit nonresponse, survey mode, and estimation methods on the differences between vaccination coverage estimates from the three surveys and COVID-19 vaccine administration data are discussed.