Help to de-convert!

[bennyboy]

You seem to think this is a one-off. It isn't. You'll find countless examples in the PubMed database. Here's a  systematic review (basically the highest level of evidence within peer-review). Depending on what is being studied (in that case dementia) religion either comes out well in front, or doesn't make a difference. You can pretend that it's not the case, or do your own literature search and see for yourself.

That's not a systematic review.  That's an abstract ABOUT a systematic review.  Based on what you link, there's no evidence that having a religion benefits you-- there's evidence that religious people tend to do better, but they haven't demonstrated a causal link.

There's a whole world of difference there-- because there are many things that are different about religious people.  For example (and I'm just assuming) it may be that religious people use less drugs, drink less, stay up late partying less, etc.  They may have more aid when they're sick-- church members willing to take care of them, etc.  It may simply be that in the case of age and dementia, getting out of the house and having people to talk to will help one stay mentally engaged longer in life.

Herein lies the rub, though.  In ALL those things I just mentioned, someone could find a secular equivalent, and both maintain their mental health AND avoid religious bullshit.

And back to the value of correlations. It often happens that biased reviewers will collect maybe dozens of studies, and treat that data like a truth score (9/10 researchers agree that X). But it takes only one really good research result to make a proof. If you need dozens, it's likely that there ISN'T proof, and that we're back to juggling numbers.