By Simona Toti, Marco Di Zio, Alessandra Ronconi
In the 2018 Italian Population Census, counts are computed by using the Base Register of Individuals (BRI) corrected for over- and under-coverage.
The aim of this paper, which stems from the efforts of the Istat Advisory Committee on Statistical Methods, is to propose a method for the evaluation of the accuracy of those count estimates.
The correction coefficient is the ratio between the complement of the probabilities of under- and over-coverage. Those probabilities are estimated through logistic regressions with random effects applied to survey data. To compute the variance of the estimator, we study some procedures based on the bootstrap technique. The resampling procedure is applied considering two settings: in the first we suppose that BRI is affected by measurement errors, in the second BRI is free of measurement errors.
The paper reports the results of an experimental application to the BRI and sample survey data for the 2018 permanent census.