Andreas Johansson wrote:
> Quoting Stone Gordonssen > > I remember from my university classes in
psyche and biology, XYYs
> > were
> > faddishly called _supermales_, and it was asserted that they were given
> > to
> > unusually high rates of violence.
> If they had tested a reasonably large number of convicts of violent
crimes, and
> a reasonably sized test group without such convictions, they could, even
if
> these samples represented a very tiny proportion of the total population,
> probably offer a very satisfying statistical proof.
IIRC, that is exactly how the research was carried out; I don't remember
exactly when, but quite some time ago (60s maybe?). Again IIRC, it created
quite a stir at the time (Oh wow, we're really onto something here....!) but
one doesn't here much about it anymore.
My psych prof's favorite examples of a spurious statistical correlation:
the softness of asphalt in NYC streets / infant and elderly death rate.