Multilevel data are pervasive in research in education, social science, and public health. The data arise from randomized trials in school, community, and hospital settings, longitudinal studies of student learning, and large-scale surveys of individuals within social contexts. The problem of missing data is equally pervasive. Until recently, efficient methods for analyzing incomplete multilevel data have not been available. Standard methods of multilevel analysis are not flexible in handling missing values, and efficient methods for missing data have assumed a single-level structure and will produce bias if applied to multilevel data.
This workshop will present new methods for the analysis of incomplete multilevel data. Special HLM software designed for this purpose will be distributed, its use illustrated, and participants will actively participate in exploring how it may be employed in their own research. Participants will be expected to provide their own laptops using a Windows-based operating system.
Stephen Raudenbush, Ph.D., Department of Sociology, Chair Committee on Education, University of Chicago
Yongyun Shin, Ph.D, Department of Biostatistics, Virginia Commonwealth University