Total Survey Error in Practice

9A Total Survey Error Perspective on Surveys in Multinational, Multiregional, and Multicultural Contexts

Beth‐Ellen Pennell,¹ Kristen Cibelli Hibben,¹ Lars E. Lyberg,² Peter Ph. Mohler,³ and Gelaye Worku⁴

¹ Survey Research Center, University of Michigan, Ann Arbor, MI, USA

² Inizio, Stockholm, Sweden

³ University of Mannheim, Mannheim, Germany

⁴ Department of Statistics, Stockholm University, Stockholm, Sweden

9.1 Introduction

The focus of this chapter is on surveys that are comparative by design, that is, surveys that collect data from two or more populations and deliberately design their instruments and other aspects of the survey to compare these populations (Harkness et al., 2010a). Examples of such surveys are the European Social Survey (ESS) conducted in 20 or more countries, the World Values Survey (WVS) conducted in up to 75 countries, and the Gallup World Poll conducted in 160 countries. There are a myriad of other multinational, multiregional, and multicultural surveys in the academic and private sector. However, these surveys vary considerably in their organizational structure and level of coordination, and the extent to which aspects of the survey design are specified. These design decisions have important implications for each of the total survey error (TSE) sources as well as for overall comparison error.

In this chapter, we introduce a TSE typology that integrates error sources with methodological and operational challenges that are unique to or may be more ...