Advances in Longitudinal Survey Methodology

10Estimating the Measurement Effects of Mixed Modes in Longitudinal Studies: Current Practice and Issues

Alexandru Cernat¹ and Joseph W. Sakshaug^2,3,4

¹Department of Social Statistics, University of Manchester, Manchester, UK

²Statistical Methods Research Department, Institute for Employment Research, Nürnberg, Germany

³Department of Statistics, Ludwig Maximilian University of Munich, Munich, Germany

⁴School of Social Studies, University of Mannheim, Mannheim, Germany

10.1 Introduction

A mixed‐mode survey is a survey in which two or more modes of data collection are used to collect the same data from different respondents. This is in contrast to a multi‐mode survey, which uses multiple modes to collect different data from the same respondents. Currently, the most popular data collection modes are computer‐assisted personal interviewing (CAPI), computer‐assisted telephone interviewing (CATI), paper self‐administered questionnaires (SAQ), and web surveys (web), and combinations of these modes are frequently implemented within mixed‐ and multi‐mode surveys (De Leeuw 2005).

There are several advantages of mixing modes. First, it may reduce coverage error because not all population members have a telephone or access to the internet and thus offering an alternative mode may increase the likelihood of establishing contact, especially with hard‐to‐reach subgroups. Second, mixing modes may reduce non‐response error because some sample members may have a strong preference for responding ...