Seriously, and I'm really, really serious here, have you ever taken a statistics course in your entire life? And have you really actually understood it?
Seriously --- and I'm really, really, serious here -- I've not only
taken statistics courses, but I've
taught them, and published peer-reviewed articles in top-flight journals on the subject.
In a universe of nations, the United States is an individual nation. And in many respects, an outlier -- it's more or less the
only prosperous first world nation with as high a "religiosity" score as it has; almost all of the other nation-states that test as high on religious belief as the USA are third-world nations.
If you want to see this kind of research, check
this report here (and
its rebuttal here). Note that Dr. Paul is attempting to establish exactly what ID would like to see -- that there is a negative correlation between a society's degree of religiosity and its health, although Paul is looking at other factors than poverty. And in the context of this study, the point on the various graphs labelled "U" is eactly that -- a single point, and in many/most graphs, an unrepresentative outlier. Paul's figure 7, in particular, shows this most clearly -- the United States is practically off the graph w.r.t. the other data points, which suggests that any "correlation" among this data has little to do with anything (and any reputable statistican would have treated a data set this noisy with much more respect than Paul has.)
Now, of course, I could easily "cook the books" and do another study, padding the list of countries with highly religious third world countries to make religion look bad. I could also cook the books in the other direction by including lots of former-Soviet bloc countries that look really bad in terms of poverty and societal health but that have low religiosity rates. Or I could do a
proper study and look at at several matched collections -- for example, do a Western-Europe substudy, a Soviet-bloc substudy, and a Subsaharan-Africa substudy -- to control for possible outliers.
But in no case would I be able to say -- "look, there's a single point on the graph -- therefore there is no correlation." Or even "look, there's a single point -- therefore there
is a correlation." Because statistics doesn't work like that.