Since Dave nicely pointed out that /PATTERN gives me the option of changing the categories that /PLOT uses in coxreg, i had a couple cups of coffee and starting playing around with my syntax.
However, I want to confirm something about the default.
If i just do:
/PATTERN by race
/PLOT
The syntax guide says that it will plot "at the means of the set of contrasts generated to replace them".
Does this mean that the graph which shows the survival (or predicted survival??) for each race is computed when every other variable in the model is at the *reference* category? The term "mean" is throwing me
off since i have only categorical and bivariate variables in the model.
My other question is with regards to getting predicted probabability or hazard for various combinations of covariates.
Does:
/SAVE survival hazard
give these values? they appear to be what i'm looking for but i haven't been able to replicate samples in a spreadsheet - my gold standard of comprehension :-). So i'm either asking for the wrong values or i'm not using the correct equation
in my spreadsheet.
Finally, more of a non-SPSS question: Does anyone have an explanation as to why with logistic regression, there is an intercept in the model but not in cox regression?
Thanks much and happy weekend.
Carol
Locked
