Tuesday, January 15, 2019

The Flynn Effect (Rising IQ Scores Over Time) Reverses

The "Flynn effect" refers to a pattern observed by James Flynn, a professor at the University of Otago in New Zealand. It points out that for most of the 20th century, scores on IQ tests have been rising.  The reasons behind this pattern have been a subject of controversy.  For example, are the rising IQ scores a result of some factor like improved nutrition, both prenatally and for young children? Are they a result of improved schooling? Or a  job environment that puts greater emphasis on cognitive skills? Is there something about the design of IQ tests, perhaps combined with  that has made scores go up even if underlying intelligence hasn't moved?

It's clear that intelligence has a genetic aspect: studies that looked at twins who were separated at birth and raised in different environments show a high correlation of their IQ scores. But rising average IQ scores over a couple of generations is clearly not the result of a sharp genetic shift; instead, the Flynn effect strongly suggests that differences in measured intelligence are not purely genetic, but also have a strong environmental component.  However, the Flynn effect now seems to be moving in reverse.

As a starting point, Flynn offers a readable overview of his perspectives on IQ research  in "Reflections about Intelligence over 40 Years" (Intelligence, September-October 2018, 70: pp. 73-83). A few snippets (citations omitted): 
A few years later, I documented what became called the `Flynn effect'. The 20th century had been dominated by massive IQ gains from one generation to another. Americans had gained 14 IQ points on the standard IQ tests (Stanford-Binet, Wechsler) between 1932 and 1976; and 14 nations had made massive gains on a whole range of IQ tests, the largest on Raven's Progressive Matrices. ... This phenomenon now covers at least 34 nations and is accepted by all scholars. The 21st century may well be different, with gains tailing off or reversing in some nations beginning in 1995, although
not in the US. ... 
After a few years of inactivity, my Catholic Youth Organization (CYO) basketball team came back to play the current team. They killed us. They had all sorts of skills we lacked, they could shoot with either hand, could pass with either hand, do fade-away jump shots. I doubt that any of them had superior genes for basketball. Rather it was the passage of time that had given them a basketball environment a world away from our own. I take it that it is easy to apply this to the realm of cognition. Within  a cohort, genetic quality tends to dictate how you respond at school, how hard you work, whether you join a book club, how you will do in high school, what university you attend – your genetic quality will eventually tend toward a matching quality of environment for cognition. Between cohorts spaced over time, different forces operate.
Since the industrial revolution began, social change has caused new cognitive exercise ... more schooling, more cognitively demanding work, and more cognitively demanding leisure. These environmental factors initially triggered a mild rise in average performance, but this rise was greatly magnified by feedback mechanisms and over a century average IQ escalated. As the average years of schooling rose, the rising mean itself became a powerful engine in its own right as people chased it to keep up. ...

I want to make it clear that although enriched environment dominated the 20th century, IQ gains are not destined to persist like the law of gravity. Factors that were immediate triggers of IQ gains included more adults per child in the home, more and better schooling, more people at university, more cognitively demanding jobs, and better  health and conditions of the aged. There are signs that these are beginning to show diminishing returns.
What are some of the "signs" which Flynn is referring? For an overview of the recent findings about recent reversals in the Flynn effect, Flynn and Michael Shayer wrote "IQ decline and Piaget: Does the rot start at the top? (Intelligence, January–February 2018, 66: 112-121). They write:
The IQ gains of the 20th century have faltered. Losses in Nordic nations after 1995 average at 6.85 IQ points when projected over thirty years. On Piagetian tests, Britain shows decimation among high scorers on three tests and overall losses on one. The US sustained its historic gain (0.3 points per year) through 2014. The Netherlands shows no change in preschoolers, mild losses at high school, and possible gains by adults. Australia and France offer weak evidence of losses at school and by adults respectively. German speakers show verbal gains and spatial losses among adults. ...

After our analysis, we will suggest two tentative hypotheses. First, trends on conventional tests show those at most risk of IQ decline are high school students aged 14 to 18. However, Piagetian results in Britain imply losses at earlier ages. Second, Piagetian tests signal something extra: conflicting trends between top scorers (those at the highest or formal level of cognitive development) and those in the early stages of the next level (concrete generalization). Large losses at the formal level may be accompanied by gains at the concrete level.
A recent study from Norway struck me as especially interesting, because in Norway many men 18-19 years of age were given standardized IQ tests as part of assessment for compulsory national service. With data from 1970-2009, one can look both at trends for 18-19 year-olds, but also look link together the scores of fathers and sons--and include other variables about a given family.  Using this data, Bernt Bratsberg and Ole Rogeberga argue that the "Flynn effect and its reversal are both environmentally caused" (PNAS,  June 26, 2018, 115, #26).
The results show that large positive and negative trends in cohort IQ operate within as well as across families. This implies that the trends are not due to a changing composition of families, and that there is at most a minor role for explanations involving genes (e.g., immigration and dysgenic fertility) and environmental factors largely fixed within families (e.g., parental education, socialization effects of low-ability parents, and family size). While such factors may be present, their influence is negligible
compared with other environmental factors.
The questions of how intelligence translates into economic outcomes and into broader human well-being are big ones, and I won't make even a feeble gesture at tackling them here. But intelligence is a real thing, albeit hard to measure, and it shapes lives and society. What looks like a reversal of the Flynn effect, apparently for broad-based reasons environmental reasons that reach across families, is worth some thought.