I have developed a case-control study, matched 1:2, and I have addressed the conditional logistic regression analysis with COXREG (SPSS), via recode. The model runs OK (confirmed with Stata).

My doubt is about the use of Cox residuals (e.g., Schoenfeld residuals for the proportional hazards) to validate a model which has been made not for a survival study.

For instance, must I only select Cox-Snell residuals to ascertain the goodness of fit? Must I reject Schoenfeld and deviance residuals because of its closely linkage to time?

Please, does anyone know the answer?

Thanks in advance.

Conditional logistic regression using COXREG:

http://www-01.ibm.com/support/docview.wss?uid=swg21477360