Events per Person-Time: A Misleading Statistic?
Jun 05, 2008 (Thu) | 4:00 PM -5:00 PM
Hoover Pavilion, 211 Quarry Road / Farquhar Room, 4th floor
: Stanford, CA
When epidemiological studies are later contradicted, such studies are often based on the statistic: events per person-time. There are circumstances when this statistic is ideal for the purpose of comparing high versus low risk groups or treatment versus control, but such circumstances are relatively rare in clinical contexts. When it is used otherwise, it is uninformative and often very misleading. Survival methods are better choices for comparing groups in such studies.
Department: Stanford Prevention Research Center
Contact: Susan Ayres | 723-6145 | susan.ayres@stanford.edu
Presenter(s):
- Helena Chmura Kraemer Professor of Biostatistics in Psychiatry, (Emerita), Department of Psychiatry and Behavioral Sciences, Stanford University
Sign Up For Email 
Announcements