In ba.transportation Jack May wrote:
> "Merlin Dorfman" wrote in message
> news:d9i47p$hgm$5@blue.rahul.net...
>> In ba.transportation Jack May wrote:
>> There is no possible application of this theory that would
>> account for more than 2 to 3 IQ points in the average for different
>> races; even "The Bell Curve" does not claim any more. And there must
>> be a huge overlap, so that the probability of a person of one race
>> picked at random being more intelligent than one of another race
>> also picked at random is very, very close to 50%.
> Real world statistics are seldom Gaussian (Bell curve) Most real world
> statistics are power law with a distribution such as 1/X^C where C is a
> constant. They are called heavy tail distributions because very large
> variation have a far higher probability than Gaussian tails.
> Human capabilities in all areas of endeavor are power law. The curves show
> that the top 1% for example produce five times more than the average person.
> The bottom 20% produces almost nothing. That is why you see the extreme
> people for example make many times more or accomplishing many times more
> than the average person.
> The same distribution applies to lines of code written, touch downs made,
> enemy aircraft shot down, and innovations.
I've seen the "long tails" theory, but I would hardly state it
as fact at this point; especially if you claim the characteristics
of the tails are different for different racial groups. That's way
beyond the frontiers of anything that's been investigated.
> A Gaussian distribution is seen when many independent things are added
> together
This is called the Law of Large Numbers or the Central Limit
Theorem, and of course it does require that the samples are
independent.
> (or in simple homework problems) :-). When many things interact
> with each other, especially in a non-linear way, you get power law
> statistics. Fractals are power law distributions for example.
|