Categorical study outcomes. One sample, two sample and correlated data problems
Gui-shuang Ying, PhDSchool of MedicineUniversity of Pennsylvaniahttp://www.ncbi.nlm.nih.gov/pubmed/?term=Ying+GS
Description
"Session 4: Categorical study outcomes. One sample, two sample and correlated data problems, by Dr. Gui-shuang Ying"
This session will cover the sample size calculation based on the precision of proportion (with 95% confidence interval), the comparison of two proportions, the odds ratio or risk ratio for association between two categorical variables, under both independent setting and dependent setting (due to matching or correlation from two eyes of a subject). Real life examples will be used to demonstrate sample size/power calculation and how one may leverage two eyes per subject to optimize either sample size or power.