Now that we have reformatted the data, we are already ready to run the second time-based model.
The CoxModel.2 variable is created in a similar fashion as CoxModel.1, however Xsatisfaction2 is substituted for Xsatisfaction to account for the change in satisfaction due to the follow-up study. The new time0 variable is also substituted for the original tenure variable, and Churn2 is substituted for the original churn response variable:
CoxModel.2 <- coxph(Surv(time0,Xtenure2, Churn2) ~ Xeducation + Xgender + Xsatisfaction2 + Xservice.calls + Xpurch.last.month + Xmonthly.charges, data=SURV2)
Avoid overwriting existing variable names. Either create new variables within an existing dataframe or preserve the old ...