Back to Annual Meeting Page		American Public Health Association 133rd Annual Meeting & Exposition December 10-14, 2005 Philadelphia, PA

Abstract #104442

Jichuan Wang, PhD¹, Michael Kepler, MS², Russel Falck, MA¹, and Robert G. Carlson, PhD¹. (1) Community Health, Wright State University, 3640 Colonel Glenn Hwy., Dayton, OH 45435, 937 775-2084, jichuan.wang@wright.edu, (2) Computing & Telecommunications Services, Wright State University, 3640 Colonel Glenn Hwy, Dayton, OR 45435

During the past decade, interest in the use of multilevel analysis to investigate public health problems has grown. The problems often occurr in the practice of multilevel modeling involve: 1) non-normality of the outcome measures under study; and 2) a small number of level-2 units in the hierarchical data (i.e., less than the required by the “30/30” rule). An approach to address these issues in multilevel modeling is bootstrap. This study demonstrates how to apply the parametric and non-parametric residual bootstraps to conduct robust multilevel modeling with only 20 level-2 units, using frequency of injection drug use data from the National Institute on Drug Abuse's Cooperative Agreement Project on HIV/AIDS. The original model results and the results of models with different bootstraps are compared. Details about conducting the parametric and non-parametric residual bootstrap for robust multilevel modeling are illustrated, using SAS Proc Mixed procedure.

Learning Objectives: At the conclusion of the presentation, participants (learner) would be able to know

• why bootstrap is necessary for multilevel modeling when the normality assumption is violated and/or the number of level-2 units is small.
• how to generate level-1 and level-2 residuals parametrically and non-parametrically for residual bootstrap, using SAS Proc IML.
• how to conduct parametric and non-parametric bootstrap multilevel modeling, using SAS Proc Mixed.
• how non-normality of the outcome measures and a small number of level-2 units in a read data set would impact the results of multilevel modeling.

Presenting author's disclosure statement:

I wish to disclose that I have NO financial interests or other relationship with the manufactures of commercial products, suppliers of commercial services or commercial supporters.