In this paper we have defined “representativeness” in terms of the ability to support overall population inferences and subpopulation inferences, associations and causal relationships at the same time. We have note some challenges that need to be balanced to achieve this, for example between population coverage and precision and between cross-sectional and longitudinal populations of interest. We next offer some comments on how representativeness should be assessed. Subsequent sections review existing and new evidence on the representativeness of Understanding Society in these two dimensions. We then conclude by discussing current and planned initiatives to enhance the representativeness of Understanding Society.
Michaela Benzeval, Christopher R. Bollinger, Jonathan Burton, Thomas F. Crossley and Peter Lynn
(2020) The representativeness of Understanding Society, Understanding Society Working Paper 2020-08, Colchester: University of Essex