Could it be that there is some selection bias in this study? I don't mean to say the conclusions are wrong per se, but I would think that people working very few hours (unemployed?) or too many hours (factory workers?) are typically less educated than those in the middle, which could explain the IQ disparity.
Over the last few decades longer hours have become associated with higher earnings and higher educational backgrounds http://www.nber.org/digest/jul06/w11895.html (old data but a good start)
If you're allowed to, you go after every minute of that sweet, sweet, overtime. Time and a half (or double time, if you work holidays) is a beautiful thing.
That's over 40 vs white collar. This study was just comparing people who worked 40 vs 25.
There's a lot more studies showing that working 60 vs 40 is bad for you cognition, but this one is unique in showing the same effects at a much loader working load.
