Bram Hogendoorn
I have a couple of questions regarding your suggestions.
1. Time fixed effect with Cox regression?
- It seems quite controversial to include time-fixed effect in the Cox regression since time (measured in years) and survival time at risk are correlated. (posted here: https://www.statalist.org/forums/for...ysis-e-g-stcox)
2. Poisson vs Cox?
- Poisson uses dependent variable as (_d, which is binary choice) and Cox has duration of time as dependent variable. Both of the regression adopts the exponential function, however it would be questionable whether it is a valid comparison with one another since we can't use the same dependent variables.
I have a couple of questions regarding your suggestions.
1. Time fixed effect with Cox regression?
- It seems quite controversial to include time-fixed effect in the Cox regression since time (measured in years) and survival time at risk are correlated. (posted here: https://www.statalist.org/forums/for...ysis-e-g-stcox)
2. Poisson vs Cox?
- Poisson uses dependent variable as (_d, which is binary choice) and Cox has duration of time as dependent variable. Both of the regression adopts the exponential function, however it would be questionable whether it is a valid comparison with one another since we can't use the same dependent variables.
Comment