In a discussion with a knowledgeable person they said that IQ's are based on 1/2 the population is above 100 and 1/2 are below. I found this questionable but have not been able to secure a source to prove the point either way. Please advise.

-- mike erlenbach (mrerlenbach@aol.com), June 11, 2002

Answers

The IQ was originally the person's "mental age" (the age at which half of people are as good as this person at solving the problems of the test) divided by the person's chronological age. That's the "quotient." This number is then multiplied by 100 for convenience. So yes, the average IQ is *defined* as being 100. Since the original tests, however, the scores have been normed on various populations and then "frozen" (until the next effort to set norms comes along). Last I heard, the average IQ score was about 105. Indeed, there is a literature on the "Flynn Effect" which notes that IQs seems to be getting higher, on average, by a certain amount (I forget the exact figure) each decade.

-- Christopher Green (christo@yorku.ca), June 11, 2002.

You can find information on the Flynn Effect at http://www.indiana.edu/~intell/flynneffect.html

-- Christopher Green (christo@yorku.ca), June 14, 2002.

As noted by Christopher Green, the original 100 was derived from the MA/CA ratio where a ratio of 1 was designated to be an IQ of 100. When L. M. Terman et al. at Stanford University revised and "Americanized" Binet's test to form the Stanford-Binet, the raw score mean of the combined subtests was set to 100 to continue the tradition of expressing the mean IQ as 100. For some reason unclear to me, 16 was the standard score chosen to represent the raw score standard deviation. When David Wechsler developed the Wechsler-Bellvue and later the Weschsler Adult Intelligence Scales for Adults (WAIS) and Children (WICS), he set the rawm score mean at 100 but the standard deviation at 15, which mean, for example, that a Stanford-Binet IQ of 132 was statistically equal to a Wechsler IQ of 130, a minor, perhaps, but annoying inconsistency between the two tests. One can read about standard scores and how they can be established in books such as L. J. Cronbach's, Essentials of Psychological Testing.

-- Roger K. Thomas (rkthomas@uga.edu), June 14, 2002.