In Question 8.5(iv) of Chapter 8 Practice Questions, how is it that the female smokers have a higher risk of heart attack than female non-smokers , shouldn't it be other way around, as z*beta value is negative for the former and Zero for the latter
Because the "Sex x smoking" covariate is in addition to the sex covariate and the smoking covariate, so for a 50 year smoking female, the value of each of these three covariates is equal to one. But for all males and non-smoker female, the "Sex x smoking" covariate is equal to zero. Hence it's baseline hazard multiplied by exp(-0.15) for female smoker and baseline hazard multiplied by exp(-0.4) for female non-smoker (and note that exponential of negative number is smaller than one).