[nbouscal]

> religion makes you happier

In this case the distinction between correlation and causation is actually very important. The alternative causal chain, that happy people are more likely to become religious, is highly plausible. Further, there are myriad plausible confounding variables. In many cases correlation can be taken as evidence of causation, but in this particular case it really should not be.