> IMHO even a statement like '... has an IQ of ...' is grossly 
> misleading.
> There are tests that can be applied to a person that will produce a
> figure, but there is no guarantee that two such tests will agree. 
> The
> only scientific definition of IQ is 'the figure that results from 
> this
> test'.

Your point is well understood, but I believe you can get a bit better 
than that. The "meaning" of IQ is a position on a distribution 
relative to the population as a whole. The weightings will indeed vary 
with culture, education, context, phase of moon. last bank holiday, 
and more. But, if a person rates in eg the top quartile on a given 
test that is meant to be appropriate to general performance in a given 
society, it seems unlikely that they would rate in say the bottom 
quartile in another test with the same intended target group. One 
would hope (perhaps forlornly) that two tests which had the same 
general objectives and were based on substantial experience would rate 
the same person within say 5% or so on the same distribution. If they 
failed to do so and the difference was substantial then it may point 
to a specific factor in the individuals performance / mindset / 
education / cognitive ability etc that was explored substantially 
differently by the different tests. Such a difference could be 
extremely valuable in identifying certain aspects of a subjects 
capabilities.

I generally rate "well up the scale" on a range of IQ tests. A while 
ago I took one online in uncontrolled circumstances without an 
adequate understanding of certain conditions under which the test was 
run. It was a reputable enough test that the score should have been 
meaningful. It went very badly indeed and AFAIK I got a result in the 
105-110 range. I understand why and if I took a similar test again I 
would score differently BUT it is quite possibly valid to say that 
within the overall circumstances of that test my IQ was validly around 
the range assessed. ie I blew it because it found and exploited a 
weakness in my 'system'. The key was that the test was timed on a per 
question basis and you did each question in order and were not able to 
go back. Failure to complete any question in the relatively short 
allotted time presumably returned a zero score. I did not expect and 
was unprepared for this approach and by the time I had adjusted to it 
I had irrevocably dropped out of "the top of the tail". If I had been 
subjected unprepared to a similar situation in real life I may have 
failed as badly. This is exactly the sort of thing that I (and most of 
us) try to avoid in real life as life tends to be real-time, 
sequential and the time-arrow usually only points one way :-). As I 
rode a motorcycle on and off road for many years and am still alive 
(only one head-on collision with a car on road :-) ) it suggests that 
I am usually reasonably OK at this process :-).

This highlights another problem with many tests. As they are designed 
to sort IQs across a wide distribution and as the majority of the 
distribution has the majority of the data points they tend to "cram" 
the top end into the outcome of a small range of questions. (Badly 
put). eg in the Mensa workout I scored 25/30. Several questions that I 
got right were open to interpretation and I could as well have got 
them wrong (poor) and several I got right for valid (I thought) 
reasons which were not their reasons. Two I could well have got right 
but didn't and I judge it happenstance as to whether I did or didn't. 
It's quite possible that a Goethe or James [:-)] would have got them 
correct "as of right" and this would have reflected their high IQ but 
my result could presumably have been pushed quite substantially up or 
down the distribution by happenstance.

All good fun though :-)


        RM 

-- 
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist