Multilevel Modelling of Hierarchical Data in Developmental Studies
This report attempts to give nontechnical readers some insight into how a multilevel modelling framework can be used in longitudinal studies to assess contextual influences on child development when study samples arise from naturally formed groupings. We hope to achieve this objective by: (1) discussing the types of variables and research designs used for collecting developmental data; (2) presenting the methods and data requirements associated with two statistical approaches to developmental data—growth curve modelling and discrete-time survival analysis ; (3) describing the multilevel extensions of these approaches, which can be used when the study of development includes intact clusters or naturally formed groupings; (4) demonstrating the flexibility of these two approaches for addressing a variety of research questions; and (5) placing the multilevel framework developed in this report in the context of some important issues, alternative approaches, and recent developments. We hope that readers new to these methods are able to visualize the possibility of using them to advance their work.
No Supplementary Data
No Article Media
Document Type: Research Article
Affiliations: 1: McMaster University and Hamilton Health Sciences Corporation, Hamilton, Canada, 2: University of New Brunswick, Fredericton, Canada
Publication date: January 1, 2001