Top Ecological Correlations With IQ: -Blacks, +Nordics, -Binge Drinking

The top State-wide ecological correlations* with IQ are rather fascinating**. Make of them what you will but here are the thoughts that occur to me.

“United States or American” and “Unclassified or not reported” as US Census “nation of origin” categories are pretty much the same category. People who don’t classify themselves could choose either of these as their “nation of origin”. This may mean we’re looking at a proxy for southern blacks so it isn’t suprising that this is the most negative ecological correlation with IQ.
“Swedish” and “Finnish” are Nordic “nations of origin” so their position as the highest positive ecological correlation with IQ matches the known (although slight***) correlation between latitude and average intelligence for Europe (Source: Greek IQ by Dienekes Pontikos) and confirms my suspicion that a big reason the motion picture industry casts blond males in negative roles is a resource competition from Jews—specifically for high IQ niches.
The negative association between alcohol binging and IQ is interesting but the explanations are not quite obvious. Are we dealing with fetal alcohol syndrome? Or is it the fact that among the highest ecological correlates of binge drinking is the presence of blacks in the State and blacks generally test at lower IQs?

*Standard ecological correlation disclaimer: Ecological correlation is a correlation where the cases are, themselves, populations. They are useful for low-cost, preliminary investigations of hypotheses. If an hypothesis predicts an ecological correlation, and that correlation exists, one may not conclude that the hypothesis is correct but one may justifiably further test the hypothesis on the heightened chance that it may be correct.

Concluding that an hypothesis is correct on the basis of an ecological correlation is a common statistical error known as “the ecological fallacy”. There is an opposing fallacy which is the failure to follow up on further study of an hypothesis that predicted an ecological correlation simply because it would be premature to conclude that the hypothesis is correct. Both fallacies are damaging.

Read a good high-level Power Point presentation of some of the issues surrounding ecological correlation as a tool in epidemiology at Epidemiology Kept Simple. **Standard data dredging disclaimer: Looking at rank-ordered correlations with no preconceived ideas can be misleading. You can’t use the normal statistical significance tables that tell you the odds of the correlation being due to chance unless you have a preconcieved idea, also known as an hypothesis, that would tend to predict the correlation you observe. For example, the top correlation with UFO reports is Amerindians.

If you have no preconceived ideas about UFO reports relating to Amerindians, then the value of the finding that Amerindians per capita is the highest correlation loses much of its significance due to “data dredging”. Even if it is the highest correlation out of hundreds, if you don’t have any preconceived ideas about UFOs that would tend to predict that they would correlate with Amerindians, then you have to start looking at the degree of correlation and compare it to the degree of correlation you would expect from a rank ordering of hundreds of variables all of which were entirely random, correlated with UFO reports. You would expect that even if they were entirely randomly correlated that, out of hundreds of such correlations, some distribution of them would turn out, by chance, to have very high correlations even if there were many data points to line up. ***I should caution that the European correlation is not statistically significant by itself (r=.15 df=19). Significance may be imputed, however, by looking at the US data, which does show this correlation at a highly significant level, and the fact that any correlation among nations may appear lower due to institutional differences in measurement/sampling between countries. The positive correlation with latitude may be due, as Dienekes points out regarding the higher variation in IQ with lower latitude, to a more heterogenous population present in more southerly climates.

Posted by James Bowery on Wednesday, November 23, 2005 at 06:16 PM in IQ and Heredity
Comments (6) | Tell a friend

Comments:

Posted by friedrich braun on November 23, 2005, 06:42 PM | #

Dienekes is a Greek propagandist; anything he says about Greeks should be independently verified. Nonetheless, that doesn’t mean that his blog is without interest.

Posted by JRM on November 23, 2005, 06:42 PM | #

I appreciate your posts.

Is the IQ data from the year 1965? Does it make sense to run a correlation from 1965 vs. data from 1990?

Posted by James Bowery on November 23, 2005, 07:02 PM | #

Except for illegals, the variables we’re looking at haven’t changed all that much since the “mid 1960s” (which is when this IQ study, supplied by Harpending to Sailer, was conducted). Perhaps I should strike illegals from the hypothesis for #1.

Posted by James Bowery on November 23, 2005, 08:22 PM | #

Corrected with an additional caveat regarding the low significance of the European correlation between IQ and latitude.

Posted by John J Ray on November 24, 2005, 04:30 AM | #

James
I think you should mention the usual cautions about ecological correlations

Posted by James Bowery on November 24, 2005, 05:04 AM | #

Standard disclaimer for ecological correlation added. Also added is the standard disclaimer for data dredging.

Next entry: The descent of Geoff Beck

Previous entry: California’s near-term future

Existential Issues

White Genocide Project

Of note

General News

Science News

Science Categories

All Categories

The Writers

Each author's name links to a list of all articles posted by the writer; the hashes link to authors' homepages.

Links

Endorsement not implied.

Controlled Opposition

Crime

General

Immigration

Islam

Jews

Nationalist Political Parties

Science

Whites in Africa