[net.misc] IQ Statistics, Anyone?

cbd@iham1.UUCP (Carl Deitrick) (09/19/84)

I know what an Intelligence Quotient (IQ) is, but I don't know the distribution
of IQs through the general population (besides, of course, one per person :-)).
I assume the mean IQ is 100, but what is the standard deviation?  Do IQs even
follow a normal distribution? (You'll pardon me if my statistics's terms are
used improperly, but I've not studied statistics in more than 10 years.)  I'd
appreciate it if someone out there could anwer my question or point me to a
reference that would have the answer.  I don't expect this to be of general
interest, so mail replies to me.  Thanks.
					Carl Deitrick
					iham1!cbd

renner@uiucdcs.UUCP (09/24/84)

>   I know what an Intelligence Quotient (IQ) is, but I don't know the
>   distribution of IQs through the general population (besides, of course,
>   one per person :-)).  I assume the mean IQ is 100, but what is the
>   standard deviation?
>   				-- Carl Deitrick (iham1!cbd)

IQ scores are normally distributed.  The mean score is 100.  The standard
distribution is right around 17.

Scott Renner
...ihnp4!uiucdcs!renner

ken@ihuxq.UUCP (ken perlow) (09/26/84)

--
>> IQ scores are normally distributed.  The mean score is 100.
>> The standard distribution is right around 17.

>> Scott Renner

1 out of 2.  The mean is 100.  The other statistic is the
standard deviation, and it's 15.  The index "IQ" was invented
and scaled to meet exactly those two criteria.  Whether this
index is correlated with any real phenomenon is, of course,
another matter.

For those of you unfamiliar with statistics, mean=100 and sd=15
imply that approximately 95% of the population is expected to score
between 70 and 130.  A sobering thought.
-- 
                    *** ***
JE MAINTIENDRAI   ***** *****
                 ****** ******  25 Sep 84 [4 Vendemiaire An CXCIII]
ken perlow       *****   *****
(312)979-7188     ** ** ** **
..ihnp4!ihuxq!ken   *** ***

billp@azure.UUCP (Bill Pfeifer) (09/27/84)

--------
>   I know what an Intelligence Quotient (IQ) is, but I don't know the
>   distribution of IQs through the general population (besides, of course,
>   one per person :-)).  I assume the mean IQ is 100, but what is the
>   standard deviation?
>   				-- Carl Deitrick (iham1!cbd)

That depends on the test which is used to measure the IQ.
Although the tests are centered at 100, the standard deviation varies.
MENSA obviously has done a lot of research on IQs, and here are some of
their data:

   Test                           Std. Dev.  Admission Score for MENSA
Revised Binet                       15.8            133
Cattell                             23.7            148
Wechsler Adult Intelligence Scale   15.0            130
California Test of Mental Maturity  16.0            132

	Bill Pfeifer
{ucbvax,decvax,ihnp4,allegra,uw-beaver,hplabs} !tektronix!tekmdp!billp

gjphw@iham1.UUCP (09/27/84)

   This is just a brief contribution to the topic of the distribution of IQs
 based upon an article disussing this in the present Encyclopaedia Britannica
 (Brintannica III).

   While the Stanford-Binet IQ test has been standardized to yield a normal or
 bell-shaped distribution with mean of 100 and standard deviation of 15 or 16,
 the normal distribution fails beyond the 95% level.  For IQs above 125, about
 twice as many people exist in the general population as would be predicted
 from a normal distribution.  Below about 85, there are three times as many
 people as indicated from a normal distribution.

   The high IQ population is explained as due to the high correlation that
 exists between marriage partners and their offspring.  Statistically, high IQ
 people marry other high IQ people and raise high IQ children.  A normal
 distribution assumes that all measured quantities are uncorrelated, so having
 correlated marriage partners violates this assumption and reveals itself where
 the population become small (IQs above 125).

   The low IQ population is described as due to a variety of causes.  Any
 injury to the brain, due to illness, accident, or prenatal development, almost
 always lowers the measurable IQ.  Since all kinds of problems that can happen
 to people virtually never acts to raise IQs, the ills that may befall anyone
 provide the source to increase the low IQ population.

   IQs are supposed to be reasonably well correlated (better than 0.5) with
 students' grade point averages through elementary and secondary schools (for
 IQs between approximately 90 and 125).  Both IQs and GPAs are poorly
 correlated with success in adult life.

   Excuse me while I return to playing with my mental blocks...

-- 

                                    Patrick Wyant
                                    AT&T Bell Laboratories (Naperville, IL)
                                    *!iham1!gjphw

gino@voder.UUCP (Gino Bloch) (09/28/84)

>> >> IQ scores are normally distributed.  The mean score is 100.
>> >> The standard distribution is right around 17.
>> >> Scott Renner

>> 1 out of 2.  The mean is 100.  The other statistic is the
>> standard deviation, and it's 15.
>> ken perlow

Both right/both wrong.  Various tests have different standard
deviations, usually around 15 to 17, but there are some funny
ones too - I mean ones with SD's like 12 (sorry, I don't remember
specifics; there is a problem with my IQ).
-- 
Gene E. Bloch (...!nsc!voder!gino)