> On Oct 31, 2007, at 12:58 AM, Miles Jackson wrote:
>
>> IQ performance is increasing, but this is obscured by the fact that IQ
>> tests are standardized so that the average score remains 100,
>> regardless of raw score. Thus IQ test scores cannot be meaningfully
>> compared over decades, because the meaning of 100 is always relative
>> to what the average performance is at a given time.
>
> So how do you/we/they know that IQs are rising?
IQ scores can be meaningfully compared over decades, but only if you are looking at results from the same test (i.e., based on the same standardization sample). In other words, you can compare how people in the 1970's did on the WAIS-R with how people in the 90's did on the WAIS-R. (I.e., you can compare the mean score of a sample in 1976 with the mean score on the same test of a sample in 1996.) In fact, that is exactly how Flynn observed the effect of rising IQ scores over time.
Some IQ tests stay around for decades, their administrators ever oblivious to their distorting effects over time. Before Flynn actually published his observation, that was understandable, as I don't think anybody assumed that IQ scores would rise over time. But old IQ tests with outdated norms do continue to be administered even today. (Respectable tests like the Wechsler will undergo periodic re-standardizations every 15 years or so, which is why we're up to the WAIS-III and Stanford-Binet IV.)