Het toepassen van statistiek voor enkelvoudige aselecte steekproeven terwijl de steekproef niet enkelvoudig aselect is

Content Provider	Semantic Scholar
Author	Knol, Hans Roelof
Copyright Year	1976
Abstract	For financial reasons, the samples derived from the Dutch population are usually not simple aselect. However, the statistical methods used with those samples are those developed for simple aselect samples. The question is whether the use of these statistical methods results in a lesser reliability of the estimated population parameter. This difference in reliability is computed for a sample of 1,622 Dutchmen 16 years & older, from 1970. Following one of the usual methods of Dutch surveys, the sample has 2 levels, one composed of 7 sublevels (large municipalities) & the other composed of 67 clusters (from the smaller municipalities), weighted according to size. In the English language literature this is usually referred to as 'two-stage sample'. It results in estimated proportions being rarely underevaluated by the use of statistical methods designed for simple aselect samples. However, correlations of 'product moment' type were found to be slightly underevaluated. 5 Tables. HA Tr & Modified by A. Orianne.
Starting Page	179
Ending Page	200
Page Count	22
File Format	PDF HTM / HTML
Volume Number	51
Alternate Webpage(s)	https://rjh.ub.rug.nl/MenM/article/download/13266/10765
Language	English
Access Restriction	Open
Content Type	Text
Resource Type	Article