40 likes | 383 Views
Survival Analysis in Stata. First, declare your survival-time variables to Stata using stset For example, suppose your duration variable is called timevar and it ranges from 0 to 15 (where 15 marks the end of the sample)
E N D
Survival Analysis in Stata • First, declare your survival-time variables to Stata using stset • For example, suppose your duration variable is called timevar and it ranges from 0 to 15 (where 15 marks the end of the sample) • You observe failure for all durations less than 15 but when timevar = 15 the data is censored • You need a variable that indicates for which durations you observe failure; call this variable failureind • Failureind must take the value 1 when duration is less than 15 and 0 otherwise • Given timevar and failureind, you tell stata you have duration data with the command: Stset timevar, failure(failureind)
The Non-Parametric Hazard Function • Sts can be used to generate the non-parametric survivor function or the non-parametric hazard function • To graph the hazard function, use: Sts graph, hazard • To list the hazard function, use: Sts list, hazard
Survival Models • Estimate survival models using streg • Specify which distribution you want using the distribution() option • You can use the exponential, Weibull, log-normal, log-logistic, Gompertz, and gamma distributions Streg x, distribution(exponential) • You may need to use the nohr option to get Stata to report coefficient estimates rather than exponentiated coefficients when you are working with the exponential, Weibull, and Gompertz distributions: Streg x, distribution(exponential) nohr