Standardization and decomposition are widely used analytic techniques in population studies for adjusting the impact of compositional factors on rates. This study demonstrates the application of these methods to behavior and health studies. Bootstrapping is used to estimate standard errors of the component effects and to conduct significance tests for them. The authors have developed a Windows-based computer program that is demonstrated in the study for standardization and decomposition analysis by using empirical data on HIV seropositivity rates in two injection-drug-using populations in the northeastern United States.

JOURNAL ARTICLE *Data Interpretation, Statistical Female Human HIV Seropositivity Male *Models, Statistical *Reference Standards Sex Distribution Software Support, U.S. Gov't, P.H.S. United States/EPIDEMIOLOGY

www.aegis.org