ABSTRACT
This chapter reviews the so-called dual system estimator (DSE) and the related census coverage adjustment method. It presents some of the challenges of replacing the census enumeration with a register one are set out. The chapter proposes a trimmed dual system estimator (TDSE) that can be used to explore, reduce and potentially remove the bias of the standard DSE, in the presence of erroneous enumeration, such as is the case with the Patient Register. Capture-recapture models for population size estimation can be used to deal with the under-coverage errors that exist in multiple lists. A notable application is census under-enumeration adjustment using an independent U-sample coverage survey (UCS) to generate recapture data. For census over-coverage adjustment, therefore, the standard approach is to deploy a separate O-sample coverage survey (OCS), selected directly from the census reports. Nonparametric survey sampling theory can be used to yield an estimate of the number of population units in the census enumeration.
