Dear all,
I am using a dataset in the counting process style to perform a Cox Proportional Hazard model, meaning I have multiple observations per id at each quarter.
There are no recurrent events, once the firm failed the are removed from the dataset.
However, I am unclear whether I would still need to add tvc to my model, as I have read that this data formatting already accounts for the time-varying nature of the variables.
As a note I'm looking at team composition diversity which can differ between different quarters.
Should I still look at the PH assumption if this is the case? What if it is violated in this case?
Thanks in advance!
Best regards,
Laura
I am using a dataset in the counting process style to perform a Cox Proportional Hazard model, meaning I have multiple observations per id at each quarter.
There are no recurrent events, once the firm failed the are removed from the dataset.
However, I am unclear whether I would still need to add tvc to my model, as I have read that this data formatting already accounts for the time-varying nature of the variables.
As a note I'm looking at team composition diversity which can differ between different quarters.
Should I still look at the PH assumption if this is the case? What if it is violated in this case?
Thanks in advance!
Best regards,
Laura
