There are a lot of nonfiction books where I expect the data would show me as having abandoned it, but really I think the books are just too long. Popularizations of scientific research fit this category very well -- halfway into most of them I'm pretty convinced I've gotten everything out of them already.