Estimating the Undercoverage of a Sampling Frame Due to Reporting Delays
Dan Hedlin, Trevor Fenton, John W. McDonald, Mark Pont and Suojin Wang
One of the imperfections of a sampling frame is miscoverage caused by delays in recording real-life events that change the eligibility of population units. For example, new units generally appear on the frame some time after they came into existence and units that have ceased to exist are not removed from the frame immediately. We provide methodology for predicting the undercoverage due to delays in reporting new units. The approach presented here is novel in a business survey context, and is equally applicable to overcoverage due to delays in reporting the closure of units. As a special case, we also predict the number of new-born units per month. The methodology is applied to the principal business register in the UK, maintained by the Office for National Statistics.
Frame quality, births and deaths, birth lags, right-truncated data