Friday, 28 March 2014

What IS somebody's IQ? Only an approximate measure - the example of Robert M Pirsig

*

IQ is not a precise measurement - especially not at the individual level, and especially not at the highest levels of intelligence when the whole concept of general intelligence breaks-down and there are increasing divergences between specific types of cognitive ability.

http://iqpersonalitygenius.blogspot.co.uk/2012/08/problems-with-measuring-very-high-iq.html

There is a tendency to focus upon a person's highest-ever IQ measure - for example in the (excellent!) philosophical novel Zen and the Art of Motorcycle Maintenance the author Robert Pirsig notes the startling fact (and it is a fact) that his (Stanford-Binet) IQ was measured at 170 at the age of nine - which is a level supposedly attained by one in fifty thousand (although such ratios are a result of extrapolation, not measurement).

*

But an IQ measure in childhood - even on a comprehensive test such as Stanford Binet, is not a measure of adult IQ - except approximately (presumably due to inter-individual differences in the rate of maturation towards mature adulthood).

A document on Pirsig's Wikipedia pages (Talk section) purports to be an official testimonial of Pirsig's IQ measurements from 1961 (when he was about 33 years old) and it reads:

*

UNIVERSITY OF MINNESOTA
COLLEGE OF EDUCATION
MINNEAPOLIS 14
 
 INSTITUTE OF CHILD DEVELOPMENT AND WELFARE
  
   June 14,1961
  
   To Whom it May Concern:
  
   Subject: Indices of the Intellectual Capacity of Robert M. Pirsig
 
Mr. Pirsig was a subject in one of the institute’s longitudinal research projects and was extensively evaluated as a preschool, elementary, secondary, college and adult on various measures of intellectual ability. A summary of these measures is presented below.
 
Childhood tests: Mr. Pirsig was administered seven individual intelligence tests between the ages of two and ten. He performed consistently at the 99 plus percentile during this period.
 
His IQ on the Stanford Binet Form M administered in 1938 when he was nine and a half years old was 170, a level reached by about 2 chilldren in 100,000 at that age level.

In 1949 he took the Miller's Analogy at the Univer. of Minn.. His raw score was 83 and his percentile standing for entering graduate students at the University of Minnesota was 96%tile.
 
In 1961 he was administered a series of adult tests as part of e follow up study of intelligence. The General Aptitude Test Battery of the United States Employment Service was administered with the following results:
  
   General Intelligence .......99 % ile
  
   Verbal Ability .............98 % ile
  
   Numerical Ability ..........96 % ile
  
   Spacial Ability ............99 % ile
  
  
   John G. Hurst, PhD   Assistant Professor

*

So, as well as the stratospheric IQ 170, there are other measures at more modest levels around 130 plus a bit (top 2 percent).

Of course there may be ceiling effects - some IQ measures don't try to go higher than the top centile.

But still, lacking that age nine test - and most nine year old's don't have a detailed  IQ personal evaluation - Pirsig's measured IQ would be quoted at about around one in fifty or one in a hundred - rather than 1: 50,000.

Ultra-high IQ measures must be taken with a pinch of salt; because 1. at the individual level IQ measures are not terribly reliable; 2. high levels of IQ do not reflect general intelligence, but more specialized cognitive ability; and 3. even when honest, the number we hear about may be a one-off, and the highest ever recorded from perhaps multiple attempts at many lengths and types of IQ test.

*

2 comments:

William Zeitler said...

Also, with ultra-high IQs, the test taker is likely smarter than the test writers. That does not seem to me to be a recipe for accuracy...

Bruce Charlton said...

Indeed.

In Anne Roe's book, The Making of a Scientist she could not find items that were difficult enough, and she had to discriminate between the high IQ scientists mostly by speed of completion (i.e. they got almost everything right, and the discrimination was in how many items were completed in a fixed time) - which means that factors such as differences in visual acuity, concentration, motor skills start to blur things.

But the problem is much deeper than that, because the tests are supposed to be measuring 'general' intelligence, and the fact seems to be that at very high levels of cognitive ability, intelligence becomes specific and is not general .