Strata variables can be selected to perform a stratified analysis when the trait is binary or survival. In a stratified analysis for a binary trait, separate intercepts are fit for each stratum, but a single estimate is obtained for each model effect conditional on the intercepts. Stratified survival analysis involves fitting a separate baseline hazard function for each stratum but again calculating a single set of effect estimates.
A Censor Variable must be selected when performing analysis of a Survival trait. This variable indicates which observations have been censored - often, this is an indicator of whether a patient is living at the end of a study.
All of the variables in the specified data set are displayed in the Available Variables field.
|
|
|
Highlight the desired variable(s). To highlight multiple variables, hold down while left-clicking.
|
|
|
Click
 to add the highlighted variable(s) to the Strata Variables field. |