Wednesday, October 12, 2005


* There's been some discussion about the recent article by Gregory Paul arguing that religiosity is correlated with societal dysfunction. You can find it online. It is a very weird paper, and noticeably so even to a non-statistician's eye. While the study is supposed to be on the noncorrelation of nonreligiosity with societal dysfunction (i.e., with providing a statistical refutation of the claim that religiosity is required for a health society), it spends most of its space talking about creationism and evolution. It makes many claims that are not clearly connected to its statistical argument. It claims not to give causal links but makes claims that require them (e.g., paragraph [17]). It doesn't seem to examine very carefully the fact that the U.S., as usual, is out in the middle of nowhere as an extreme outlier, whereas most statistics papers of this sort that I have seen would at least have a short explicit argument about whether the outlier was distorting the results (or was a symptom of some deeper distortion). For what is purportedly a statistical study it is highly polemical in tone. It mischaracterizes Catholic doctrine. It is vague in its use of terms, and its statistics seem to consist entirely of little graphs that are not very informative.

I would be the first to admit that statistics is an immensely difficult discipline, involving many things that are not entirely intuitive, but if this were standard statistical practice I would begin to be very skeptical of statistics. Fortunately, it doesn't seem to be anything remotely near standard. The best discussion of the article that I have seen is at Magic Statistics, the weblog of Scott Gilbreath. He gives his reasons for doubting the statistical value of the article. Alas, the atheistic part of the blogosphere has been spreading the study. Even if the conclusion of the paper is right, it doesn't seem to provide any clear statistical evidence for it; and if we're stuck with regarding it as a philosophical argument based on scattered evidences, it's not a very good one of those, either (and even if it were a well-formulated philosophical argument it wouldn't be contributing anything new, anyway).

* The next poetry carnival is going to be at Talking to Myself. Entries need to be in by the end of October 14. The host is asking for a theme:

NEW THIS MONTH: I would like to ask for a theme for all entries and this month's theme is short poetry -- limericks, haiku, tanka, renku, cinquains, you name it -- anything is welcome as long as it is 6 lines or less.

No comments:

Post a Comment

Please understand that this weblog runs on a third-party comment system, not on Blogger's comment system. If you have come by way of a mobile device and can see this message, you may have landed on the Blogger comment page, or the third party commenting system has not yet completely loaded; your comments will only be shown on this page and not on the page most people will see, and it is much more likely that your comment will be missed.