The Analysis of Longitudinal Survey Data under Complex Sampling – A Review
Gad Nathan
Department of Statistics, Hebrew University, Jerusalem, Israel

The predominant method of analysis for longitudinal data, primarily used for experimental data, is based on the fitting of generalized linear models and the use of generalized estimating equations, for estimating the model parameters. This is not always suitable for survey data obtained under complex design. We examine the effects of informative sampling designs and propose an extension to longitudinal surveys of a general method of inference on the population distribution under informative sampling, developed for cross-sectional samples. Other methods of analysis, suitable for survey data, such as multilevel modelling can be modified to take complex survey design into account. In particular they can be used for the treatment of the effects of nonresponse in longitudinal surveys. Thus, the combined use of time series structures with hierarchical modelling based on an augmented regression method or on a state-space model can be used to treat informative non-response. The use of calibration techniques, using known external flow data, is examined, as a means to enhance the quality of flow data.

Keywords: Generalized linear models; Informative sampling; Multilevel modelling; Calibration

Biography: Gad Nathan is Professor Emeritus of Statistics at the Hebrew University of Jerusalem and has long been associated with the Israel Central Bureau of Statistics, most recently as Chief Scientist. He received his Ph.D. from Case Institute of Technology, Cleveland OH and has published numerous papers in leading statistical journals, including Journal of the Statistical Association, Journal of the Royal Statistical Society, Survey Methodology, Journal of Official Statistics and Sankhya. His main research areas are sampling methodology, inference from complex samples, computer assisted interviewing and telesurveys. He has held visiting and consulting positions at several academic institutions and statistical agencies in North America and in Europe and has served as Vice-President of the International Statistical Institute and of the International Association of Survey Statisticians, as well as President of the Israel Statistical Association and Chairman of the Israel Public Council of Statistics.