Issues in weighting for longitudinal surveys

The production and application of survey weights is rather more complex in the context of a longitudinal survey than that of a cross-sectional survey. And it can be particularly complex for some types of longitudinal surveys, due to certain common features of design, such as sample units and survey instruments that are hierarchically related or cross-classified (individuals within households; pupils within both schools and households) and dynamic sampling. We note that general issues in survey weighting - that apply to any survey - apply also to longitudinal surveys. We do not discuss those issues here but instead refer the interested reader to the excellent texts by Bethlehem and Keller (1987), Biemer and Christ (2008), Brick (2013), Elliot (1992), Kalton and Flores-Cervantes (2003), Little and Vartivarian (2005), Särndal (2011) and Valliant et al (2013). In this chapter we outline some of the aspects of weighting that are unique to the longitudinal context and discuss some of the advantages and disadvantages of different possible solutions to some of the issues that are encountered.