Al Quds university
"Analysis of survey data under informative sampling "
The material in this talk serves as a basis for the analysis of survey data under informative sampling design. Section 1 contains an introduction. Sections 2 and 3 define the methods of probability weighting and pseudo likelihood estimation. Section 4 defines the sample distribution and its relationship to the population distribution. In Sections 5 and 6 we treat the sample likelihood function and the sample distribution is derived under different models for sample selection probabilities. Section 7 gives results on the relationships between moments under the sample and the population distributions. Section 8 outlines estimation methods based on the sample likelihood. Section 9 focuses on multivariate normal distribution. In Section 10 we propose the Kullback- Leibler information as a distance measure between the sample and the population distributions. Finally in Section 11 we discuss briefly the relationship between the response-biased and informative sampling designs.