Well if we want to be all logical and science-y about this, we can't assume humans are rational beings. People obviously make bad decisions all the time.

Just because lots of people have kids, for example, doesn't mean that having kids makes people happier, and in fact studies seem to show the opposite to be true on average. That said, I'm not trying to say that having kids is always an irrational choice, and certainly it makes some people happier (or it might make people "unhappier" but lend them a greater sense of meaning and purpose that balances things out). But clearly you can't just say "well, correlation must imply causation because otherwise why would people do X if it makes them less happy?"