A Framework for Identifying and Addressing the Risks of Exclusion from Social Surveys

This paper discusses the issue of population subgroups that may be under-represented in, or missing entirely from, UK population surveys. I do not attempt to quantify the extent of under-representation or to evaluate the impact that this may have on survey estimates. Instead, I attempt to provide a framework for considering the problem: for identifying the groups of interest, for understanding the mechanisms that lead to them being under-represented or excluded, for identifying methods to reduce the under-coverage in future surveys, and for evaluating the merits of implementing such methods on national surveys. The discussion is informed both by a statistical accuracy perspective and an ethical and moral perspective.