| Science Forum Index » Space - Consult Forum » effect coding of categorical data in Cox Regression?... |
|
Page 1 of 1 |
|
| Author |
Message |
| Lina... |
 Posted: Tue Oct 13, 2009 5:16 am |
|
|
|
Guest
|
Hello,
I have one question concerning coding of categorical data in Cox
Proportional Hazard Model for survival data and would be pleased about
any help.
Take the following example: I have survival data and want to analyze
the influence of a categorical covariate with 6 levels on the
survival. As there is no ordering within the categorical covariate, I
would like to avoid using a reference group. It would be better to get
hazard ratios for the comparison of each factor level with an average
effect. But how can I model such hazard ratios? Is there something
like "effect coding" for Cox regression (like for linear and logistic
regression)? Would it be possible in the presence of the "undefined"
baseline hazard and the lack of an intercept?
I am happy about any suggestions or references
Lina |
|
|
| Back to top |
|
|
|
|
