Programming the statistical procedures from SAS

DF=0 in PROC PHREG output

Accepted Solution Solved
Reply
New Contributor
Posts: 2
Accepted Solution

DF=0 in PROC PHREG output

I have a binary covariate using PROC PHREG that states it has 0 DF in the output. I have checked and there are definitely two categories in the data for this covariate. I am not sur ehow to proceed.


Accepted Solutions
Solution
Sunday
Super Contributor
Posts: 271

Re: DF=0 in PROC PHREG output

It is not enough you have more than one value of the covariate in your data. There should also be events in all levels ( I mean, not all persons with same level should be censored).

It is even more complicated: At the time a person has events, there should also be persons with different values of covariates at risk at that time, if that event should contribute anything. If all events fail this, then it is possible to have events at several levels and still no estimate (and df=0).

View solution in original post


All Replies
Grand Advisor
Posts: 16,838

Re: DF=0 in PROC PHREG output

Can you post some of the output that illustrates your issue?

Solution
Sunday
Super Contributor
Posts: 271

Re: DF=0 in PROC PHREG output

It is not enough you have more than one value of the covariate in your data. There should also be events in all levels ( I mean, not all persons with same level should be censored).

It is even more complicated: At the time a person has events, there should also be persons with different values of covariates at risk at that time, if that event should contribute anything. If all events fail this, then it is possible to have events at several levels and still no estimate (and df=0).

New Contributor
Posts: 2

Re: DF=0 in PROC PHREG output

Thank you this is very helpful and I can see my error because to model mortality and whether there is an extreme weather event but of course each date has the same value hence my error

☑ This topic is SOLVED.

Need further help from the community? Please ask a new question.

Discussion stats
  • 3 replies
  • 79 views
  • 0 likes
  • 3 in conversation