The xerox copier company is tracking how long the toner wheel on their copiers lasts until it fails. they have a data set of 563 observations from 269 copy machines. some machines have been measured more than once because they had the toner wheel replaced after it failed and then the lifetime of this new toner wheel was measured. machines that were taken out of commission for other reasons were marked as censored.

(a) we have a few covariates that we might want to include in the model.

• year the copier was built • model type of the copier • the number of maintenance calls made to the copier before the toner wheel failed • the episode number of the toner wheel (1= original part, 2= first replacement part, 3 = second replacement wheel)

which of these covariates should we not include in our cox proportional hazards model? why?