Intelligence, Aging, and Societal Change
Issues surrounding human intelligence are complex and controversial: What is intelligence? What is the best way to measure it? Is there one type of intelligence or are there many? Are there gender and ethnic differences in intelligence? How much of our intelligence is shaped by ‘nature’ i.e., genetically determined, and how much by ‘nurture’, that is, by socioeconomic and cultural variables?
These are difficult and important questions, as is the one addressed here: how do our intellectual abilities change as we grow from early to late adulthood?
The ‘Classic’ Aging Pattern
The scientific research on the age-intelligence relationship in the adult population is based to a significant extent upon the extensive use of the Wechsler Adult Intelligence Scale (WAIS), a psychometric test developed in the 1950s in the US which has undergone several revisions since then.
The WAIS consists of two parts: a verbal scale, so called because it requires verbal responses, and a performance scale, which does not; each scale consists of subtests which tap a specific skill.
The verbal subtests measure our store of world knowledge and our ability to apply it. They assess a person’s knowledge of appropriate behaviors in everyday situations; of biological, historical and literary facts, of math and of the meaning of words.
By contrast, most of the performance subtests require the testee to manipulate blocks, interpret pictures, solve puzzles, etc. The object of these tests is to measure on the spot intellectual ability: to find out how, and how quickly, a person can approach and solve an unfamiliar problem, the solution of which is not importantly helped by his or her level of knowledge, as is the case with the verbal tests. A (raw) score is computed for each subtest; for the overall verbal and performance scales separately; and for the whole test. These latter three scores are then converted into IQ scores, which index how well a person functions intellectually relative to their age group.
The large scale administration of the WAIS over the decades has produced results consistent enough to be referred to as the ‘classic aging pattern’. It shows that, over the approximate age period of 20-70 years, the aspect of intelligence measured by the verbal scale shows a very modest or even no decline (and can even improve on some subtests). By converse, the performance scale of the WAIS shows a steadily increasing decline from the early twenties onward.
Explanations of the Classic Aging Pattern
Several explanations of this pattern have been offered. One favored by many is the so-called two-factor theory of intelligence (which was based on a variety of tests aside from the WAIS). Its main tenet is that there are different types of intelligence, of which two are the most important. Crystallized intelligence is involved in acquiring, remembering, and applying information. This type of intelligence is primarily engaged by the tests which make up the verbal scale of the WAIS. Because of its nature, it depends upon our lifetime experiences and learning; therefore, it should remain stable as long as the rate at which we learn new information balances out the rate at which we forget. This is the case for a majority of people until their reach their seventies. This type of intelligence is expected to fall in extreme old age because the cumulative effects of losses: of jobs, health, relationships, etc. lead a person to become increasingly disengaged from society and culture and thus less able to acquire new information.
By converse, fluid intelligence, which is tapped by problems the solution of which does not depend appreciably upon our level of knowledge (think for example of solving Rubik’s cube versus a crossword puzzle), is more affected than crystallized intelligence by genetic and physiological factors. Because of this, fluid intelligence, as is true of abilities most directly dependent upon the efficiency of the central nervous system, is expected to peak in early adulthood, and then to begin to increasingly decline.
Are Age Related Changes in Intellectual Performance Caused by Factors Other Than Aging Per Se?
Can we then regard the matter as basically settled? Not quite.
Here is a radically different way of looking at the above data. Could it be that the obtained differences between young and old are due, not to aging per se, but to the fact that successive generations tend to get 'smarter', that is, to perform at higher levels of functioning, at least in some cognitive tests?
An example may illustrate this idea. Think of a computer that was built in 1980. Imagine that it was so well preserved that it works as well as new thirty years later, in 2010. This being the case, its performance is unaffected by age. Compare now that computer with a 2010 model. The more recent model will vastly outperform its1980 counterpart. The latter will underperform, not because it is no longer functioning properly due to its age, but because it was built earlier, and hence it reflects a less competent technology.
Could something like this be happening to people? The hypothesis here is that older people intellectually underperform younger ones, not because they are older, but because even when young their level of intellectual ability (as measured by these tests) was inferior to that of today's youth. It is not aging per se that matters, but simply the fact that society as a whole gets smarter with time, and the young are the primary beneficiaries of this trend.
One way to assess this hypothesis is by comparing the intellectual abilities of people of identical age growing up in different historical periods.
Some of the best data originate from European countries where intelligence tests have been administered to all males registering for the military draft upon reaching 18 years of age. In many of these countries, data are available from over 90% of the 18-year-old males across a period of almost 50 years.
Analyses of these data reveal a consistent pattern of substantially higher intelligence (as measured by the same standardized tests) for recent generations relative to earlier ones.
In fact, the size of this generational difference over a period of 30 years is nearly as large as the average IQ difference between 20-and 50-year-olds in cross-sectional comparisons.
Similar improvements are found in US data. For example, the average score on the Alpha Army test of US soldiers in World War II matched the score of the top 15% WWI soldiers.
The Flynn Effect
Other data from a variety of standardized intelligence tests from the early 20th century onward show that IQ scores have been rising at an average rate of around 3 to 6 IQ points per decade. This finding was named the ‘Flynn’ effect after its discoverer.
One of the most recent studies I am aware of was conducted in 2010. It compared the IQ scores of today's 20 years old with today's 70 years old. The researchers then adjusted the scores of today's 70 year old for the Flynn effect, that is by taking into account the growth in IQ as a result of the effect. They found that between 85 and 100 per cent of the overall difference between these two age groups was in fact attributable to the Flynn effect alone.
Importantly, these generational improvements are not restricted to crystallized or knowledge-based types of intelligence. In fact, the generational improvements are larger on measures of fluid than of crystallized intelligence. Thus it is not simply the case that successive generations are acquiring more knowledge than earlier generations, since fluid intelligence measures reflect ‘pure’ abstract or novel problem-solving ability.
At present, the reasons for these improvements are not well understood. One factor that can almost certainly be ruled out is genetic improvement in the species, because evolutionary changes require much longer than one or two generations to occur.
More likely factors include educational changes, historical changes in prenatal care, public health practices, nutritional habits, and increased cultural stimulation. It is also likely that increased familiarity with tests of all kinds helps the younger groups get better results.
As noted at the beginning, the age-intelligence relationship is indeed extremely complex, and the debates surrounding it are likely to be with us for years to come.
© 2014 John Paul Quester