In the closed CMR sample simulation, try repeating the simulations but with N=50 and p=0.2 and 0.5, and then with N=200 and p=0.2 and 0.5. For each of these, find levels for k that achieve CV=0.1 or less
Modify the "robustness" program to evaluate how well the Mt (time variation in p, p=c) model performs under individual heterogeneity
Hint: a simple way to do this is the replace the m0 model with an mt model in all the places where m0 appears
The attached data set contains CMR data for males and females over 10 occasions. However, there is unidentified heterogeneity in the data within sex, with sub-groups in each sex having very different survival.
Test goodness of fit with a het.otstrap test using these data and the global Phi(group*time)p(group*time) model
Hint: you can easily modify the Dipper bootstrap gof code to do this by
Loading the "sim.het.data" object and replacing the dipper data object with sim.het.data (i.e., instead of data(dipper), dipper=sim.het.data
Proceed as before except be sure n.occasions =7 not 10