OT: a concrete example of scientific fraud

First, a bit of actual guitar content: I'll be
one of three guitarists performing at the Beacon
Hill Art Walk this afternoon. Should any of you
happen to be there, please say hello to the classical
archtop that Al Carruth built.

The rest of this is a follow-up to Miguel de Maria's
post citing and questioning an article by La Griffe
du Lion [1].

I had never before heard of La Griffe du Lion. His
article was so full of egregiously bad methodology
and other mistakes that I at first assumed it was a
satire of racist pseudoscience, the sort of thing
that might be written for the Annals of Improbable
Research or (if you're Alan Sokal) Social Text.

To mention just one example, the article's thesis is
that a nation's "smart fraction" of the population
predicts per capita GDP better than average IQ, but
then uses (highly questionable) "data" on average IQ
to pretend to estimate the smart fraction by assuming
a Gaussian distribution. As any competent scientist
would notice, however, the assumption of Gaussian
distributions can't be used to estimate the "smart
fraction" without the further assumption of a specific
variance or standard distribution, for which there are
no data available. In short, the article claims that
data on average IQ are inadequate to predict per capita
GDP, but then pretends to obtain a better prediction of
per capita GDP from precisely that inadequate data by
making a sequence of unsupported assumptions. We're
used to seeing that kind of argument in rmcg, but
scientists recognize it as a form of scientific fraud
known as drylabbing [2].

When I saw that the article had cited an "average
verbal IQ" of 59 for equitorial Guinea, I assumed
it was a joke (in extremely poor taste). According
to Wikipedia, however, that number really was given
by La Griffe du Lion's primary source, Richard Lynn
[3]. On the other hand, Wikipedia also cites Hunt
and Wittman's finding that a simple check of Lynn's
citation shows that 59 was the "mean IQ of a group
of Spanish children in a home for the developmentally
disabled in Spain." That summarizes the reliability
of Lynn's data, and all of La Griffe du Lion's IQ data
come from Lynn.

I don't know how many people have been taken in by
La Griffe du Lion, but one of them is Charles Murray,
the political scientist who co-authored The Bell Curve:
Intelligence and Class Structure in American Life.
In an editorial published by the Wall Street Journal
on 25 July 2006, Murray wrote [4]:

If you want to get deeper into the math, you
may visit a quirky and provocative Web site,
www.lagriffedulion.f2s.com, run by someone who
calls himself La Griffe du Lion. I surmise that
he is an established scholar--a quantitative
discipline seems likely--who once published on
the fraught topic of group differences, learned
how unpleasant and even professionally perilous
that can be, and decided to remain anonymous
henceforth. In any case, his technical skills
are first rate.

In any case, La Griffe du Lion's technical skills are
good enough to fool Charles Murray.

Will

[1] La Griffe du Lion. Smart Fraction Theory II:
Why Asians Lag. Self-published fraud, online at
http://www.lagriffedulion.f2s.com/sft2.htm

[2] http://en.wikipedia.org/wiki/Drylabbing

[3] http://en.wikipedia.org/wiki/Race_Differences_in_Intelligence

[4] http://www.opinionjournal.com/editorial/feature.html?id=110008701
.