Total Survey Error in Practice

20Total Survey Error Assessment for Sociodemographic Subgroups in the 2012 U.S. National Immunization Survey

Kirk M. Wolter,¹ Vicki J. Pineau,¹ Benjamin Skalland,¹ Wei Zeng,¹ James A. Singleton,² Meena Khare,³ Zhen Zhao,² David Yankey,² and Philip J. Smith²

¹ NORC at the University of Chicago, Chicago, IL, USA

² National Center for Immunization and Respiratory Diseases, Centers for Disease Control and Prevention, Atlanta, GA, USA

³ National Center for Health Statistics, Centers for Disease Control and Prevention, Hyattsville, MD, USA

20.1 Introduction

Total survey error (TSE) is the difference between a survey estimate and the true value of the corresponding population parameter. TSE is the net effect of sampling error and all forms of nonsampling error, including sample‐frame coverage error, error due to nonresponse, and errors of measurement (such as reporting, coding, and other processing errors). In our work, TSE excludes conceptual errors committed in deciding what should be measured in the survey and judgmental errors committed in analyzing the data (see Chapter 22 of this volume) or in making policy or business decisions based on the data. An intrinsic feature of any high‐quality survey should be an attempt by the survey organization to assess TSE for some of the survey’s key statistics.

The long history of the study of component survey errors goes back decades. Sampling error was described in the 1930s by Neyman (1934). The U.S. Census Bureau’s model for measurement ...