Identifying influential observations through the intraclass correlation coefficient

Angel de Jesus Davalos, University of Texas at El Paso

Abstract

In this thesis we analyze the performance of adapting the DFBETA statistic for identifying influential observations on the intraclass correlation coefficient under the assumptions of the one-way random effects model. Additionally, we introduce an approach to transforming negative intraclass correlation coefficient estimation values using the method of moments estimator. We apply this method on a data set of repeated blood pressure measurements, after which we will investigate implications of identifying influential observations.^

Subject Area

Statistics|Health Sciences, Public Health

Recommended Citation

Davalos, Angel de Jesus, "Identifying influential observations through the intraclass correlation coefficient" (2010). ETD Collection for University of Texas, El Paso. AAI1479551.
http://digitalcommons.utep.edu/dissertations/AAI1479551

Share

COinS