Why no mention of penalized regression techniques in Applied Logistic Regression, 3rd edition, by Hosmer, Lemeshow, and Sturdivant? Just ordered this textbook, and Wow, the complete omission of this subject from an otherwise excellent reference on logistic regression is a bit surprising. The 2nd edition was published in 2000 - there's been a lot of research into penalized regression techniques since then: for example, the lasso and elastic net.
I know this text is aimed at epidemiologists and public health specialists. Have researchers in these fields ignored any developments that smack of data mining?
 A: It is possible that there is a prejudice against anything that comes close to data mining, but there can also be good reasons for omitting this area. Without further information I would start with the assumption that there was a good reason. If you assume a good reason and there was actually a prejudice the debate starts as much more friendly and is likely to become much more productive than when you assume prejudice and there was actually a good reason. 
Any textbook needs to make a selection on what it covers and what it does not cover. The fact that something has developed recently is on its own not a sufficient reason to include it. I use logistic regression a lot in my own research and I have never been in a situation where I needed the lasso or elastic nets. So if I were writing such a textbook, I would probably ignore that part too. This is not to say there is anything wrong with these techniques; they just weren't useful for the problems I wanted to solve. 
Since different (sub-)disciplines tend to focus on different types of questions, it is not surprising that some techniques are less useful for that (sub-)discipline than others. This should translate to which topics are covered in textbooks aimed at different (sub-)disciplines.
