Here is a unified, readable introduction to multipredictor regression methods in biostatistics, including linear models for continuous outcomes, logistic models for binary outcomes, the Cox model for right-censored survival times, and generalized linear models for counts and other outcomes. The authors describe shared elements in methods for selecting, estimating, checking, and interpreting each model, and show that these regression methods deal with confounding, mediation, and interaction of causal effects in essentially the same way.
I read it for my biostatistics course. Not wildly impressed. Not a fan of Stata on which most of the coding information is based. Did present concepts in a fairly simple fashion, but not the friendliest text for non-stats majors. Maybe I'm asking for too much here. I mean, it's a textbook.