Model reduction justified?

irma

New Member
#1
Hi!
I'm working with cox proportional hazard models in R and I'm trying to find the minimum model. I have a very simple question but I cannot find a direct answer:). These are the models and the result of the anova:

anova(model1,model2)
Analysis of Deviance Table
Cox model: response is Surv(tstart, tstop, tcensors)
Model 1: ~ tint + factor(tstage) + frailty(tdish) + factor(tfem) + factor(tcoll) + factor(ttemp)
Model 2: ~ tint + factor(tstage) + frailty(tdish) + factor(tfem) + factor(ttemp)

loglik Chisq Df P(>|Chi|)
1 -16521
2 -16502 37.797 11 8.464e-05 ***
---
Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1



Is it justified the model reduction in this case? I guess, yes, because the loglik is lower in the shortest model and so the AIC, but the fact that the anova is significant confuses me...
thanks for the help!!