Total Survey Error Analysis: National Immunization Survey Adult COVID Module COVID-19 Vaccination Coverage Estimates

Thursday, May 12, 2022

A. Elizabeth Allen

NORC at the University of Chicago
Vicki J. Pineau

NORC at University of Chicago
Kirk Wolter

NORC at the University of Chicago
Jason Boim

NORC at the University of Chicago
James Singleton

US Centers for Disease Control and Prevention
Michael Chen

US Centers for Disease Control and Prevention
David Yankey

US Centers for Disease Control and Prevention
Jennifer Kriss

US Centers for Disease Control and Prevention
Yi Mu

US Centers for Disease Control and Prevention

The difference between an estimated survey outcome and the true population value can be impacted by a variety of sources. This overall difference is defined as the total survey error which arises from both sampling and nonsampling errors, such as nonresponse error, sample-frame coverage error, and measurement error. In this analysis, we will present results from a total survey error evaluation of national estimates of the proportion of adults reporting receipt of =1 dose of a COVID-19 vaccine in the U.S. as reported each week by the National Immunization Survey Adult COVID Module (NIS-ACM) since April 2021. We will examine the potential stages in which survey error can occur in the NIS-ACM process and seek to measure the error associated with each survey stage. External data sources, such as the American Community Survey, the National Health Interview Survey, and COVID-19 vaccine administration data from CDC will be used to assess and quantify the potential sources of error.