On the effects of sampling errors on inverse projection results: an empirical study Odoardo Bussini - Giorgio E. Montanari A new statistical method has been proposed in previous papers to estimate historical population aggregated moviment data over a region using a sample of elementary areal unit observations. Such a methodology employs data already available, it does not require further collections of supplementary sample observations and consists in specifying and identifying a superpopulation model capable of describing the probabilistic structure of survey variables. The proportional to the unit population model, applied to a real population of the seventeenth century, allowed the estimation of the aggregated time series of births, marriages, and deaths using data from a stratified sample of parishes, and allowed the estimation of their precision. The present paper is aimed at verifying the effect of the level of sampling errors on the method subsequent demographic analysis, with particular attention to the population recostruction called inverse projection, proposed by Lee. Reducing the sample base, we compare the recostruction results achieved by summing elementary data with those obtained with the predictive approach based on superpopulation models. Such a comparison will enable to standardize and simplify the aggregation process required by population recostruction.
L'effetto dell'errore di campionamento sui risultati dell'inverse projection: uno studio empirico
BUSSINI, Odoardo;MONTANARI, Giorgio Eduardo
1996
Abstract
On the effects of sampling errors on inverse projection results: an empirical study Odoardo Bussini - Giorgio E. Montanari A new statistical method has been proposed in previous papers to estimate historical population aggregated moviment data over a region using a sample of elementary areal unit observations. Such a methodology employs data already available, it does not require further collections of supplementary sample observations and consists in specifying and identifying a superpopulation model capable of describing the probabilistic structure of survey variables. The proportional to the unit population model, applied to a real population of the seventeenth century, allowed the estimation of the aggregated time series of births, marriages, and deaths using data from a stratified sample of parishes, and allowed the estimation of their precision. The present paper is aimed at verifying the effect of the level of sampling errors on the method subsequent demographic analysis, with particular attention to the population recostruction called inverse projection, proposed by Lee. Reducing the sample base, we compare the recostruction results achieved by summing elementary data with those obtained with the predictive approach based on superpopulation models. Such a comparison will enable to standardize and simplify the aggregation process required by population recostruction.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
