Journal of Official Statistics, Vol.9, No.4, 1993. pp. 815–823
Transforming Hypotheses for Test of Homogeneity with Survey Data
Jeffrey R. Wilson and Mark Reiser
Abstract:Hypotheses tested using X2 (Pearson statistic) and G2 (likelihood ratio) in SAS, BMDP, SPSSX and similar statistical programs assume that the data have been obtained from a simple random sampling scheme. In the use of survey data such an assumption is hardly ever met and adjustments must be made to the test statistic for the possible presence of extravariation. Instead of making test statistic adjustments, this paper considers transformation of the original hypotheses and the construction of test statistics with familiar forms to investigate the transformed hypothesis. Thus the hypothesis is transformed when survey data are used and the form of the basic statistic is maintained.
Keywords:Design effects; complex sampling; extravariation.
Copyright © Statistics Sweden 1996-2018. Open AccessISSN 0282-423XCreated and Maintained by OKS Group