psymon100's forum posts
Oh good another angsty atheist trying to confirm his beliefs.:roll:LJS9502_basic
Great, now if only I could access the article.
The fact that the metric they report is correlation makes me doubt they did anything too legitimate. Psychology isn't known for being good at statistics.
I was unaware it was an issue, this should work: https://dl.dropboxusercontent.com/u/23555618/psymon100_Posts/waifu/Meta%20Analysis%20Religion%20and%20Intelligence.pdf
I've heard statisticians claim that no academics other than statisticians have a comprehensive grip on statistics. A typical strategy in our larger studies is to simply collaborate with a statistician and give them an author credit on the paper. In the acknowledgements they mention the stats was done by a Satoshi Kanazawa - however I'm not sure which Satoshi Kanazawa.
It's all there in the results section for you dude. Such colourful things as 'heterogeniety of chi squared' and what not. I've had a glance but need a more critical read through myself.
But all in all, I just want to say that yes - one certainly couldn't make any predictions from this data.
Of course, any actual interpretation of this number assumes no sample bias, no omitted variable biases, no endogenous relationships, etc., and honestly I'm highly skeptical that psychology studies actually did a decent job of addressing these.chessmaster1989
Their methods section explains how bias was assessed, exclusion criteria, etc. These Meta-Analysis can be tough to follow though...
Personally I wouldn't make any assumption about one academic discipline. I would give them all equal weight if they're rigorous with their science.
Ah, my apologies to you and Lostrib. It slipped my mind completely that it is R² which is used that way, not r.
Wait, didn't the abstract give the number .24 and .25 for the correlation which mean it is positively related? Or is it past my bed time?
I'm afraid I have made a mistake - my apologies. For some reason when copying and pasting - the negative signs disappeared. I have fixed the OP, the abstract should now be identical to the published paper.