Jensen Redux

From Boing Boing -

The brain and intelligence

You know what they say about people with big brains ... Or, actually, maybe you don't.

Despite being a major concept underlying of the neurobiology of intelligence for the last 150 years or so, the connection between brain size and smarts isn't well-understood by Joe and Jane Average. Does it mean smaller people—including women—are less intelligent? What about animals, like elephants, that have much larger brains than ours? Are our academic destinies really written in our hat size?

It's complicated. We know that brain size and intelligence are correlated, but that simple fact is only a starting point for a much more intricate story—one that science is only beginning to understand.

First off, yes, bigger brains really do seem to be smarter brains. That correlation has been pretty solidly proven, experts say, and the connection gets stronger when you calculate total brain volume via MRI technology or post-mortem analysis, rather than simply running a tape measure around somebody's head. Basically, the more accurate and precise the brain measurement, the more size and smarts are connected.

How connected varies a bit, depending on the methodology, but an analysis of previous research, published in 2005 in the journal Intelligence, found a .33 correlation at the population level. Which means, if you look at humans as a whole, a little more than 10% of the difference in intelligence from person to person can be accounted for by brain size.

That's statistically significant. But it also means overall brain size isn't the only thing affecting intelligence. Case in point: Gender.
By Maggie Koerth-Baker at 7:35 AM February 26, 2010

If anybody has actual library access to the "analysis of previous research, published in 2005 in the journal Intelligence" could they check a couple of points?

1) Controls.
Large and tall people are more likely to be well-fed, and therefore on balance more likely to be rich. Unless size variation is tracked within comparable income groups, that's an obvious possible source of 'bias' (used in its non-technical sense).

2) 'an analysis of previous research"...
when I last looked at this issue (http://home.vicnet.net.au/~borth/JENSEN.HTM) some twenty years ago the then popular brainsize/intelligence review (Van Valen) went back to include data from Pearson in 1906, at a time when methodologies were, to say the least, less developed. I commented then that "One of the besetting sins of psychometrics is that it continues to hoard its references long after they have gone thoroughly rotten, as if the transformation into number raised experiments into a sphere where their methodology could not date." Could somebody check this out?

In my experience, people with a commitment to the measurement of intelligence are prepared to be satisfied with evidence and arguments that are extraordinarily flimsy, and it is necessary to check their arguments, their references, and their bona fides carefully.