The Undercoverage-Nonresponse Tradeoff

Total Survey Error in Practice
,
pp. 95-113
,
Hoboken, NJ
,
Wiley
,
2017,

Paul P. Biemer, Edith de Leeuw, Brad Edwards, Lars E. Lyberg, N. Clyde Tucker, Brady T. West,
Eckman, Stephanie, Frauke Kreuter
ISBN: 978-1-119-04167-2 (print); 978-1-119-04170-2 (online)
DOI: 
10.1002/9781119041702.ch5
Abstract

This chapter begins with documented examples of the tradeoff to study the undercoverage-nonresponse tradeoff and then presents a simple example of how the tradeoff might arise in the context of a landline and mobile phone survey. It further introduces the concept of response propensity (RP), which may be familiar to some readers, and the related idea of coverage propensity (CP): the likelihood of a given case to be included in a survey's frame. A simulation study reveals the effects of CPs and RPs on coverage and response rates and on undercoverage and nonresponse bias. Building on these concepts, the chapter presents a simulation model and explores how the relationship between coverage and response propensities affects response and coverage rates and undercoverage and nonresponse bias. It gives the simulation results back to the real world and discusses the implications of the simulation for survey practice and survey costs.