Stories like this pop up frequently on HN. But I'm intrigued by your post post script. What do you think is the reason for the proliferation of these studies, and why do they so often ignore the correlation-does-not-prove-causation fallacy?
In this case, I see two possibilities. The Bureau of Labor Statistics author is too dumb to understand the fallacy. The second is that she does understand the fallacy, but thinks that at least some Americans won't. If the latter, what purpose would that serve?
I think it's because not because these numbers are imperfect, or because many things can be read into them that shouldn't, that they are completely useless. In the end you need some data that is intuitive to understand, even if there are many ifs and buts involved.
In this case, I see two possibilities. The Bureau of Labor Statistics author is too dumb to understand the fallacy. The second is that she does understand the fallacy, but thinks that at least some Americans won't. If the latter, what purpose would that serve?
Interesting.