Re: Race, intelligence, and anti-racist prejudice (Was: Genetic Evolution)
ron house (house@helios.usq.EDU.AU)
Thu, 2 Feb 1995 02:05:16 GMT
st26h@rosie.uh.edu (JAMES BENTHALL) writes:
>In article <3glfl1$b5m@vixen.cso.uiuc.edu>, pcollins@prairienet.org (Patrick R. Collins) writes...
>>Frank Fujita (ffujita@sun1.sun1.iusb.indiana.edu) wrote:
>>: JAMES BENTHALL (st26h@elroy.uh.edu) wrote:
>>: : All I.Q. tests fail the t-test. Variation within a population is always
>>: : greater than between populations.
>>
>>: I'm unfamiliar with your usage of "t-test." Could you elaborate?
>>: Frank Fujita
>>
>>A t-test is a statistical test. It lets you know if the average of a sample
>>is different from a known population when the standard deviation is not
>>known.
I don't know about the specific question at issue here, but James
Banthall's statement is dubious from a mathematical viewpoint.
Variations between two groups can indeed be much less than the variations
within a group and still be statistcally significant. With billions
of 'samples' involved (as is the case with questions about various
populations of people), even a minute average difference can be
significant.
Note: 'significant', statistically, means that the probability that
the difference is random is less than a certain percentage: often
set at 5% or 1%. The smaller the percentage at which a difference
is significant, the less likely it is that the difference is due
to chance alone.
--
Ron House. USQ | A nonviolent diet is the
(house@usq.edu.au) Toowoomba, Australia. | foundation for a nonviolent world.
|