Stumbling on Happiness (Daniel Gilbert)

A fun book, recommended by lightreads, about why we are so bad (except when we're good) at predicting the future. It's not a bad pop pysch book, though I had my usual pop-social-science problem: he describes an experiment, what it's supposed to convey, and I go, "Yes, but..." Which maybe I wouldn't do if I saw the actual paper, but I don't find the level of detail in pop-social-science convincing.

For example, he has this thesis that people deriving satisfaction from having children is a myth propagated by human society, and isn't actually true, which he "proves" by showing a graph of how marital satisfaction diminishes once people have children, and showing that women find looking after kids less satisfying than doing other tasks. Yes, but... but I find working less satisfying than reading books, but I can tell you right now that my overall satisfaction would diminish if I were reading books all the time instead of working, which lets me exercise my technical skills, spend time with interesting geeky people, gives me satisfaction that I'm contributing to society (maybe that is the "myth" part -- but if it really does make me happy because I believe that it should, is it a myth?), and not least, makes me money to buy the books with... And there are lots of things -- he wouldn't even disagree with this -- that can honestly make you happy once you've convinced yourself they ought to make you happy.

ETA: Not that I disagree with his hypotheses (certainly, for his hypothesis on child-rearing, it is supported by kids being a Big Pain to raise) but it irks me that his proof standard is as bad as it is, even for my pleasure reading.