Total Survey Error in Practice

23Mixed‐Mode Research: Issues in Design and Analysis

Joop Hox,¹ Edith de Leeuw,¹ and Thomas Klausch²

¹ Department of Methodology and Statistics, Utrecht University, Utrecht, The Netherlands

² Department for Epidemiology and Biostatistics, VU University Medical Center, Amsterdam, The Netherlands

23.1 Introduction

Surveys increasingly use mixed‐mode data collection (e.g., combining face‐to‐face and web) to control costs and maintain good coverage and response rates. Common types of designs are concurrent mixed‐mode designs, offering respondents a choice of mode, and sequential mixed‐mode designs, for example starting with the least expensive mode (e.g., web) and follow‐up with more expensive (interview) modes (de Leeuw, 2005; Chapter 6). Given the extremely high costs of panel surveys, an increased interest in applying a mixed‐mode approach in longitudinal studies can be observed (Dex and Gumy, 2011), where the first data collection occasion uses a mode that is known to have high response rates (often face‐to‐face), and subsequent data collections are carried out using less expensive modes (typically web).

The strength of mixed‐mode surveys is their potential to reduce coverage and nonresponse problems, while at the same time attract as many respondents as possible for an affordable budget (de Leeuw, 2005). When mixed‐mode data collection designs are used, there is the potential for systematic differences in the data collected due to the different modes used. As Tourangeau et ...