Let us suppose that we wish to build the confidence interval for an indicator (e.g. the ratio of the production of a crop per hectare or industrial production in two different situations) when sample design is stratified sampling. The Combined Ratio estimate will be considered. The problem of interval estimator for a ratio has been studied in simple random sampling in Galeone-Pollastri in 2012 and in Pollastri-Tulli in 2011 it is considered the situation when the indicator must be positive. The problem of building the confidence interval for the indicator is very similar, mutatis mutandis, to the one studied for the simple sample survey because the asymptotic distribution of the estimator is based on the ratio of two correlated normal random variables. The correlation between the numerator and the denominator may be estimated. It is also possible to compare the size of the confidence intervals for the indicator using the simple random sampling and the Combined Ratio in stratified sampling.
Pollastri, A., Maffenini, W. (2013). interval estimation of indicators in stratified sampling. In BOOK OF ABSTRACT Italian Conference of SUrvey Methodology (pp.131-132). Padova : Libreria Universitaria Webster srl.
interval estimation of indicators in stratified sampling
POLLASTRI, ANGIOLA;MAFFENINI, WALTER
2013
Abstract
Let us suppose that we wish to build the confidence interval for an indicator (e.g. the ratio of the production of a crop per hectare or industrial production in two different situations) when sample design is stratified sampling. The Combined Ratio estimate will be considered. The problem of interval estimator for a ratio has been studied in simple random sampling in Galeone-Pollastri in 2012 and in Pollastri-Tulli in 2011 it is considered the situation when the indicator must be positive. The problem of building the confidence interval for the indicator is very similar, mutatis mutandis, to the one studied for the simple sample survey because the asymptotic distribution of the estimator is based on the ratio of two correlated normal random variables. The correlation between the numerator and the denominator may be estimated. It is also possible to compare the size of the confidence intervals for the indicator using the simple random sampling and the Combined Ratio in stratified sampling.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.