Are we Getting Smarter?

For the last 75 years the average IQ (intelligence quotient) test score has been increasing in every industrialised country in the world. Some suggest that this increase could be as much as 25 points. So are we really more intelligent than our grandparents? The first inkling of this effect dates from the early part of last century. In 1949 the Scottish Council for Research in Education looked at scores from two different generations of people of the same age. Comparing the IQ test scores of 11 year olds in 1933 and 1947 suggested that the average IQ score had increased by 2-3 points. But the effect wasn't solely confined to Scotland. James Cattell carried out a similar study on schoolchildren in the English city of Leicester between 1936 and 1949, which showed an increase of 1.28 IQ points, and in 1948, using databases of US Army recruits from the first and second world wars, Reid Tuddenham also found that IQ scores had increased significantly over this time. Subsequent studies carried out on the same groups of people by researcher James Flynn, using two styles of intelligence tests, have also shown that the average IQ in America appears to have risen by 15 points between 1930 and 1980.

But why should this rise in IQ have come about ?

Could this be a result of genetic changes? The process of assortative mating occurs when there is selective mating between individuals with similar characteristics, and there is evidence this occurs in the context of intelligence. In other words, individuals of similar intelligence are more likely to mate with each other. But, rather than increase the intelligence of the population as a whole, this should only increase the variance, or range, of IQ scores because a child is the result of genes from both parents, half coming from the mother and half from the father. If we think of intelligence as partly being a product of various genes, then the offspring of two intelligent people, who are likely to have even more 'good genes' for intelligence than their parents (as they will receive these good genes from both their mother and their father), are likely to be even more intelligent than their parents. So, over time, one would expect the higher and lower ends of the IQ scale to be stretched as those humans with the highest (and lowest) IQ scores mate and reproduce.

A further, more damning piece of evidence against the genetic argument is that if the effect were down to genes, any change they produced should be in the opposite direction, that is, making the population appear less intelligent. This is because the association between fertility and IQ shows a negative correlation. In other words, if IQ score were just down to genes, then the fact that individuals with lower IQs are having more children than people with higher IQs ought to decrease the average IQ of a population. It seems, then, that the answer to the riddle of the rising IQ score does not lie in genetics.

But before we probe more deeply for the reason, let's first consider what is 'intelligence'? Is it the ability to understand and profit from experience? The faculty of thought and reason? Should short term memory, which is 'measured' in IQ tests, be part of our definition of 'intelligence'? The technology boom that we have observed since the industrial revolution may have made people more adept at many of the skills which IQ tests measure, but intelligence is surely more than simply the ability to perform well on a test. Perhaps the increase in IQ merely reflects an improved exam technique ?

Whatever intelligence actually is, if nature (genes) isn't responsible for boosting our brain power, then it must be down to nurture - the environment of our upbringing - and one hotbed of contention is that of education. Since 1920 the number of years that people spend at school has increased significantly, as too has the proportion of students going on to higher education. But significant increases in IQ test scores were seen in children below the school leaving age as early as 1920, and both six year olds in Japan and nine year olds in Canada have been found to have IQ scores higher than their representatives in earlier generations. Perhaps the way that education has changed in style in the last 60 years has acted to increase the IQ of children receiving formal schooling? It is easy to see that the change in emphasis from 'parrot-fashion' learning to learning by discovery and conceptual understanding may make children better at solving problems (such as the Ravens Matrices subtest, which seeks to measure an individual's ability to reason clearly. The test is normally presented in a symbolic format, where the individual is required to perceive the pattern or link between the symbols, and predict the appropriate continuation of the sequence). But what does the increase we are looking at actually represent? Does this change in teaching style actually increase the intelligence of children, or just their performance in IQ tests? Of course, IQ tests strive to measure intelligence itself, but such a complex and multi-factorial entity is extremely difficult to quantify in a meaningful way.

For instance, Ulric Neisser has highlighted the example of Chinese schoolchildren, pointing out that they improved by 22 IQ points on Ravens Matrices between 1936 and 1986, a period when literacy in China was increasing dramatically following urbanisation. Learning to read Chinese characters involves memorising complex symbols and combining them to alter their meaning and to signal pronunciation. Of course this may help performance on tests such as Raven's Matrices, but should we view this as this a gain in intelligence or just an increase in IQ score performance?

Researchers now think that changes in socio-economic status have may have played a significant part in the rising IQ phenomenon. In the last 60 years there has been an increase in the proportion of the population who are in middle-class professions. Since we know that IQ correlates very strongly with social class, many have suggested that this increase in class has caused the increase in IQ test scores. Indeed, when we study the correlations between IQ tests and socio-economic status, we find that verbal abilities play the leading role, and as socio-economic status increases, so we see an accompanying increase in verbal IQ test scores. (A verbal IQ score is taken from the performance on language related tests, such as vocabulary, comprehension of written material and general knowledge. The non-verbal IQ score relates to tests of spatial ability, pattern recognition and other subtests that do not encompass any use of language.) But if we study the raw data for IQ tests we find that the rise is much more pronounced in non-verbal tests, and a very small increase, if any, is observed for verbal IQ tests. So, if socio-economic status is most strongly linked to verbal IQ, and the most significant rises in IQ score are for non-verbal IQ, then the change we have seen in socio-economic status cannot underlie the rising IQ scores. Moreover, this rising IQ seems to be occurring across all social classes, rather than merely a group phenomenon, and if it were due to a change in social status we ought to find different IQ rises in different social groups.

A more universal environmental change that we have experienced since the First World War is an improvement in public health and nutrition. This is especially relevant in the aftermath of the Second World War when rations were removed and more food was available. Indeed, in this time Japan saw an average height increase in the population of 2-4cm, and the latter half of the century saw an increase in both head and brain size, indicating that this nutritional improvement had a real physiological effect. In 1970, Michael Rutter and his colleagues found a significant correlation between birth weight and later IQ, so we can imagine that better nutrition for pregnant mothers may have resulted in more intelligent offspring. We also see a very significant effect on children of very low birth weight, who are likely to produce relatively low IQ scores. However, it is difficult to know whether the effect of nutrition on IQ scores is very significant. A study of 1-year-old infants showed a positive correlation with maternal nutrition during pregnancy, but one of 18-20 year old men showed no correlation at all, so it is possible that any effects of maternal nutrition on IQ are only transient. Studies on the relationship between post-natal nutrition and IQ scores yield a variety of different results, making it difficult to draw a reliable conclusion. However, studies of the severely malnourished show a correlation with very low IQ scores, implying that there may be a threshold below which nutrition is very important and above which it makes little difference.

James Flynn, after whom the rise in IQ scores is named 'The Flynn Effect', suggests the 'multiplier effect' as the cause of rising IQ scores. He proposes that as the average IQ rises, parents create an environment for their children which is increasingly constructive for developing their mind from a very early age. He claims that a person's IQ develops partly in relation to the IQs of the people he or she comes into contact with. As IQ rises in this way, so will this effect be multiplied in the following generation, thus causing a fairly rapid and ongoing increase. This is a plausible hypothesis and the effect may well contribute to the significant IQ score rise.

The implications of these increases in IQ scores may be one of two things: either the populations of industrialised countries are becoming more intelligent, or IQ scores are not in fact measuring intelligence and we are merely adapting to score highly on these tests. The fact that the most significant increases in test score are seen for non-verbal test components, and that these gains do not appear to have been accompanied by an escalation in other cognitive skills, might suggest that we are not measuring 'real world' intelligence. IQ tests certainly measure something similar to intelligence, and they do measure some of the manifestations of it. Yet it is possible that they measure factors that we would not define as intelligence, whilst equally missing out some of its fundamental areas. Tests have become more sophisticated over time and we believe that they are close to measuring intelligence, but disputes about what they measure and indeed the nature of intelligence continue.

So maybe we're not becoming more 'intelligent' after all, despite what the rise in scores of IQ tests would initially seem to imply!