000 02002cam  2200325za 4500
0019.841695
003CaOODSP
00520221107152142
007cr |||||||||||
008170815s2004    onc    |o    f|0| 0 eng d
040 |aCaOODSP|beng
041 |aeng|bfre
043 |an-cn---
0861 |aCS11-619/2004-3E-PDF
1001 |aWong, Paddison.
24512|aA note on variance estimation in multi-stage sampling |h[electronic resource] / |cPaddison Wong.
260 |a[Ottawa] : |bStatistics Canada, |c[2004].
300 |a14 p.
4901 |aWorking paper ; |v2004-3
500 |aDigitized edition from print [produced by Statistics Canada].
500 |a"HSMD-2004-003E."
504 |aIncludes bibliographic references.
5203 |a"In multi-stage sampling surveys, the variance of an estimator has contributions from all stages of the survey design. The expressions of the total variance and its estimator are usually complicated. However, with certain choice of sampling scheme, they can be simplified. In the case of sample survey with r stages, if the first stage is with replacement, the total variance of an estimator has a simpler expression because the combined variance of the last r-1 stages has been integrated into the other terms in the equation. This is one advantage of sampling with replacement in the first stage over sampling without replacement. However, it should be noted that the variance of an estimator for sampling with replacement in the first stage is always greater than that for sampling without replacement in the first stage"--Abstract.
546 |aAbstract also in French.
69207|2gccst|aMethodology
69207|2gccst|aStatistical analysis
7101 |aCanada. |bStatistics Canada. |bMethodology Branch.
830#0|aWorking paper (Statistics Canada. Methodology Branch)|v2004-3|w(CaOODSP)9.834763
85640|qPDF|s1.79 MB|uhttps://publications.gc.ca/collections/collection_2017/statcan/11-613/CS11-619-2004-3-eng.pdf