Originally posted by Hydro:
The average IQ of the population as a whole is, by definition, 100. IQs range from 0 to above 200, and among children, to above 250. However, about 50% of the population have IQs between 89 and 111, and about 80% of the population have IQs ranging between 80 and 120, with 10% lying below 80, and 10% falling above 120.
For IQs below 120, IQ is the best predictor of socioeconomic status of any psychometric measurement. In more complex jobs, IQ is better than even education or experience at predicting job performance. In her article "The General Intelligence Factor", Scientific American Presents "Exploring Intelligence", pg. 24, 1999, Linda Gottfredson states,
"Adults in the bottom 5% of the IQ distribution (below 75) are very difficult to train and are not competitive for any occupation on the basis of ability. Serious problems in training low-IQ military recruits during World War II led Congress to ban enlistment from the lowest 10% (below 80) of the population, and no civilian occupation in modern economies routinely recruits its workers from that below-80 range. Current military enlistment standards exclude any individual whose IQ is below about 85."
"Persons of average IQ (between 90 and 100) are not competitive for most professional and executive-level work but are easily trained for the bulk of jobs in the American economy. By contrast, individuals in the top 5 percent of the adult population can essentially train themselves, and few occupations are beyond their reach mentally."
"People with IQs between 75 and 90 are 88 times more likely to drop out of high school, seven times more likely to be jailed, and five times more likely as adults to live in poverty than people with IQs between 110 and 125. The 75-to-90 IQ woman is eight times more likely to become a chronic welfare recipient, and four times as likely to bear an illegitimate child than the 110-to-125-IQ woman."
Nice bit of stats there, but in general, IQ scales still have a mean of 100 and a standard deviation of 15. Without going and measuring the Bell curve, the corresponding figures work out to produce a similar net effect of the 80-120 calculation.
What IS important about an SD of 15 lies in the definition of a SD. Put simply, if you have two people who are one SD apart from each other, you really don't need any measurement instrument to notice the difference.
The key aspect of IQ has also been missed out here, namely what it means. The formula for calculating IQ is (Mental Age/Chronological Age) X 100, which works out effectively into a percentage, e.g. an IQ of 125 states that you have the intellectualy ability expected of an average person who's 125% of your age.
Also, there's little point in talking about IQs of over 160. The sample of the population represented is so small that no score above that can be considered statistically sound.
While the passages you cited are all right as a general reference, I would take them with a pinch of salt. IQ tests, unfortunately, are subject to bias, be it in the form of gender, language or education. For example, you could have someone who's pretty intelligent from say, the Australian outback, but who would not be able to perform on any tests of common Western knowledge. Research indicates that men have the advantage over women in terms of visuospatial abilities, so tests involving things like rotating figures would be biased in favour of men. So on and so forth. Fact of the matter is that the only thing that the result tells you is how well you did on the test. Theoretically, the ideal measure of intelligence would be Spearman's
g, but in a century nobody's been able to define it properly in an operational sense, much less test it. At the end of the day, you could have someone with an IQ of 145 and he could still be dumb as dogsh*t in a practical sense.
When it comes to measurement of intelligence (or to be politically correct, intellectual ability), the instruments currently being used are far from perfect. Over and above the bias issues described above, you've also got a problem with assessing the difficulty of the test. Traditional test theory is based on a vicious cycle: the worse people do on a test, the harder it is, but the less able the people are, the worse they do. The question this system fails to answer is: are the people less able or is the test harder? Recent developments in Item Response Theory have started to address this issue, but it'll be a while yet before we get a system that comprehensively replaces the current system.
The last time I head, Mensa requires a verified IQ score of 145 for membership (I believe it was raised from 135 about a decade ago).
As to what's considered a high IQ, it's a subjective matter. let's face it, if you've got an IQ of 101, you're smarter than the average bear already. As a rule of thumb, if you're over 120, you're doing all right.
From a clinical perspective, I'd choose the Weschler Adult Intelligence Scale (WAIS) as a measure of intellectual ability to produce an IQ score. However, if I did want to work out how quick someone is intellectually, I'd use Raven's Standard and Advanced Pattern Matrices.