Categorization Errors and Differences in the Quality of Questions in Comparative Surveys

Differentials in measurement error and in question quality can bias findings in comparative survey research. This chapter examines methods to identify and to remedy such bias. The first section of the chapter discusses models used to estimate the measurement error coefficients of survey questions, starting with a basic response model. Then, the chapter presents questions and data from the European Social Survey (ESS). It presents estimates from prior research of the authors, and briefly reviews possible explanations for large differences found in these estimates across countries. The next section presents the model to account for categorization errors. The chapter illustrates what is meant by such errors, and compares the results obtained from categorical versus continuous models. Next, it introduces the method of estimation and discusses the results. Finally, the chapter presents a meta‐analysis of the many findings and discusses general conclusions on the basis of this.