I am tooling up for my debate on the economics of religion with Lawrence Iannaccone. Studying data from the General Social Survey, it's clear that people who attend church more are a bit happier. On a three-step scale (very happy/pretty happy/not too happy) people who attend church several times per week are about a quarter of a point happier than people who never attend. (And that's quite a bit smaller than the effect of income, which most happiness researchers think is over-rated).
However, it's striking that controlling for attendance, religious beliefs do little or nothing for happiness. In particular, it doesn't matter how firmly you believe in God - or if you believe; and it doesn't matter if you believe there's an afterlife.
This strongly suggests that the psychological benefits people get from religion stem from the social aspect, and not the doctrine. So while people often use findings like this to make a pragmatist's case for religion, a better interpretation is that people would be happier if they joined a group of some sort with regular meetings. It doesn't have to be a church; it could just as well be a regular gaming group. Back in high school, I did both, and guess which one gave me more happiness!