Sunday, January 29, 2006

Intelligence in UK declining?   posted by DavidB @ 1/29/2006 03:33:00 AM

Today's London Sunday Times (January 29) has an article in the Education section on new research which claims that British children's 'intelligence' has declined dramatically in the last 30 years. If the link works, the article is here.

The research is by Profs. Adey and Shayer of King's College London. Adey claims, based on a sample of 25,000 children, that 'the intelligence of 11-year-olds has fallen by three years' worth in the past two decades'.

Naturally this is of interest in the context of the Flynn Effect - the long term trend of rising IQ scores. Several recent reports suggest that the Flynn Effect has halted or gone into reverse.

I haven't been able to find any further details of the research than those in the ST article, and I suggest a need for caution. The tests used do not appear to be standard IQ tests but rather tests of 'scientific reasoning', which combine general intelligence (g) and more specific mathematical and physical concepts. In IQ terms, a fall of 3 years in average mental age at chronological age 11 would be massive: if we suppose the baseline 30 years ago is IQ = MA/CA = 11/11 x 100 = 100, the new IQ would be MA/CA = 8/11 x 100 = approx. 73. I don't think mean IQ can possibly have fallen by 27 points in 30 years! The school at King's College is also known for unorthodox views on the nature of intelligence, including the belief that 'thinking skills' can be radically improved by fairly small amounts of direct 'thinking' teaching.

I also note that there is no mention of the ethnic composition of the samples, which must certainly have changed in the last 30 years. However, in IQ terms the fall is far too large to be explained by compositional changes of this kind.

[Added: The last point should be sufficiently self-evident, but let me expand on it for the benefit of the innumerate. In 1975 the proportion of non-whites at age 11 in Britain was around 5%. In 2005 it was around 15%. (These are very rough figures, but good enough for the present purpose.) Let us suppose that in 1975 the mean 'intelligence' of white 11-year-olds, by standard IQ tests or any other valid instrument, was 100, while that of non-whites was 85. This is about the size of the black-white differential in the US, or the difference between whites and the offspring of recent non-white immigrants in European countries. It probably overstates the differential between whites and non-whites in Britain, since non-whites in Britain include large numbers of Indians and other high-achieving groups. Assume that white and non-white IQ is unchanged between 1975 and 2005. These assumptions gives us mean population IQ of 99.25 in 1975 and 97.75 in 2005 - a fall of less than 2 percentage points. This is far too small to account for the kind of decline reported by Adey and Shayer. To explain such a large decline by changes in the composition of the population, either the magnitude of the compositional change, or the differential between the different components, or both, must be much larger than is at all credible.]

Despite these reservations, this is clearly interesting research, and I will try to follow it up.

Added: I found a more informative account of the research in the Guardian here. The full report will be published in the British Journal of Educational Psychology next year.