Unknown eligibility and quality measures of longitudinal surveys: critical review

Abstract

A challenge for longitudinal surveys is to identify and update eligibility status over time for all sample units, including those who cease participating in the survey. As they are not part of the population of interest, ineligible units should be identified and excluded from the sample. This paper reviews the cause, impact and treatment methods of the unknown eligibility problem. The emphasis is theoretical rather than practical. We explain the effect of unknown eligibility on the calculation of survey quality measures (e.g. response rate) and summarise the existing methods of estimating the proportion of eligible sample units amongst units of unknown eligibility. We suggest and critique an alternative strategy for dealing with the problem.