Respondent-targeted survey design features
Methodological research into design features that aim to improve survey participation often finds that effects are heterogeneous. In other words, what works best for one sample sub-group may not work best for another. A corollary of this is that the most effective survey design is unlikely to involve standardised treatment for all sample members. Rather, different design features should be utilised in different ways for different groups of respondents. Longitudinal surveys are in a strong position to be able to identify subgroups at which various design features should be targeted, given the wealth of information they have available about preferences, circumstances, behaviours and past participation histories.
Such targeting of design features has the potential to both reduce survey costs (as costly features need not be applied to all sample members, but only to those for whom they are particularly effective) and improve survey participation. These outcomes are particularly pertinent to Understanding Society as budgets are being reduced while stakeholders are understandably concerned to maintain high participation rates. A design feature currently at the forefront of our search for greater cost-efficiency is the use of alternative survey modes. We are seeking the best ways of targeting the use of web data collection. Other design features suitable for targeting may be able to improve participation with little or no cost implication. These include the design of a range of respondent communications, including advance letters and between-wave mailings.
This research programme aims to identify ways in which targeted procedures can improve the trade-off between participation rates and survey costs. The prime focus is on Understanding Society, though many of the findings are expected to have broader applicability.