Total Survey Error in Practice

13Total Survey Error for Longitudinal Surveys

Peter Lynn¹ and Peter J. Lugtig^1,2

¹ Institute for Social and Economic Research, University of Essex, Colchester, UK

² Department of Methodology and Statistics, Utrecht University, Utrecht, The Netherlands

13.1 Introduction

In this chapter, we describe the application of the total survey error (TSE) paradigm to longitudinal surveys, by which we mean surveys that collect data on multiple occasions from the same sample elements. Longitudinal surveys are of great importance and considerable investment has been made in them by government agencies and academic funding bodies in recent years (Berthoud and Burton, 2008; Lynn, 2009; Tourangeau, 2003). However, there are several aspects of survey error, and of the interactions between different types of errors, that are distinct in the longitudinal survey context. Furthermore, error tradeoff decisions in survey design and implementation are subject to some unique considerations.

Previous literature on TSE mostly deals with survey design and implementation issues pertinent to cross‐sectional surveys and does not explicitly consider uniquely longitudinal issues (Biemer, 2010; Groves and Lyberg, 2010). Weisberg (2009) describes several relevant features of longitudinal surveys but does not address their error structures. Smith (2011) draws attention to this limitation and points out that the relationship between error structures at different waves may be at least as important as the nature ...