Take the example.data we used in the previous time- and individual specific analyses.
Create at least 2 additional models that contain the time-specific covariate as both linear and quadratic terms.
RMark- you will need to re-create the ddl and augment covariate data frame. with the quadratic term . These commands should works:
cap.ddl=make.design.data(cap.processed)
example.data$cov$cov2=example.data$cov$cov^2
cap.ddl$Phi=merge_design.covariates(cap.ddl$Phi,example.data$cov)
Run the new model set, compute AIC, perform model averaging, and plot model averaged estimates and CIs of survival versus the covariate value
Conduct the same analysis in MARK. Note: you will have to manually enter values for the quadratic term, although these can be copied and pasted from a spreadsheet
In RMark move to an individual covariate approach for the same analyses. Now produce plots of model-averaged estimates and CIs of survival at evenly spaced values over range of the covariate value (0 - 1).
CP&PPS*
*Conroy's proposed and probably partial solution