101 – Statistical Inference with Clustered Data in Survey Sampling
Discussion: Statistical Inference with Clustered Data in Survey Sampling
Michael P. Cohen
American Institutes for Research
Inference with clustered data is an important practical problem in statistics. When the sample is obtained from complex sampling design including unequal probability sampling and cluster sampling, the inference becomes more complicated because the sampling design is often informative. We have three excellent papers on this topic, two papers from frequentists' approach and one paper from Bayesian approach. Two papers in frequentists' approach are quite different because one is based on classical survey sampling approach and the other is based on the so-called h-likelihood approach. The three papers will be discussed by an experienced practitioner in this area. Speakers: 1. Youngjo Lee Paper title: h-likelihood method for analyzing clustered survey data Author contact information: Department of Statistics, Seoul National University, Seoul, Korea, E-mail: youngjomail@gmail.com 2. Jae-kwang Kim Paper title: Inference with cluster data under informative sampling Author contact information: Department of Statistics, Iowa State University, Ames, IA 50014. Phone: 515-294-7068 E-mail: jkim@iastate.edu 3. Yajuan Si Paper title: Bayesian survey inference with cluster sampling design Author contact information: Department of Biostatistics, University of Wisconsin, Madison, WI E-mail: ysi@biostat.wisc.edu