The Flynn effect is the trend where raw performance on IQ tests has been increasing in many countries at a rate of about 1 Standard Deviation (SD) per half-century.  In 1990, scholar Richard Lynn published a brilliant paper claiming that the 1 SD per half-century gains in IQ scores were perfectly paralleled by 1 SD gains in height and head size (and by inference brain size) over the same period.  Since the 1 SD gain in head size and height are thought to be entirely caused by nutrition (including disease reduction which affects the body’s ability to use nutrients), Lynn reasoned quite logically that the 1 SD gain in IQ was also entirely caused by nutrition.

I call this the parallel effects model, because it implies nutrition’s effect on IQ will be paralleled by its effect on brain size which will be paralleled by its effect on head size which will be paralleled by its effect on height.  So if you see a population that has increased by 1 SD in height and head size via nutrition, it’s reasonable to conclude that they have also increased in brain size and IQ by 1 SD via nutrition since all four of these variables are just indirect proxies for nutrition and all reflect it to an equal degree.

Parallel effects model:


In 1998, Arthur Jensen revised Lynn’s theory (The g Factor, page 326).  He agreed that nutrition had caused a 1 SD increase in height and head size and further agreed that this implied a parallel 1 SD increase in brain size, but he stopped short of believing that nutrition had also caused a parallel 1 SD gain in IQ.  Instead he felt that nutrition’s impact on IQ was only a byproduct of its effect on brain size, and so the effect of nutrition on IQ was limited to only the effect of brain size on IQ.  I call this the secondary effect model.

Secondary effect model:


Jensen probably preferred the secondary effect model because it saved him from believing that real biological intelligence had increased as much head size and height, which seemed counter-intuitive to him given that society did not seem that much smarter than it was 50 years earlier.  By limiting nutrition’s effect on IQ to its effect on brain size, Jensen could cite the 0.4 correlation between IQ and brain size (probably an overestimate) to argue that nutrition could only increase IQ by 40% as much as it increased brain size as opposed to 100% in the parallel effects model.  This put a ceiling on how much of the Flynn effect could be biological and thus “real”.  Any observed Flynn effect exceeding 40% of brain size gains could be dismissed as fake gains caused by culture (i.e. schooling, media).

However with the release of a new paper by scholar Michael Woodley et al, showing brain size gains have only been a small fraction of an SD per half-century (much smaller than head size gains and height gains), it might be time to propose a third theory: The tertiary effect model.  In this model, not only would rising IQ just be a byproduct of brain size, but rising brain size itself would just be a byproduct of rising height.

The tertiary effect model:


As a biological determinist, I personally prefer the parallel effects model and dislike the tertiary effect model because it reduces the biological component of the Flynn effect to just a byproduct of a byproduct, but I must admit it’s the model that seems to best fit the totality of the evidence.   For example, Lynn 1990 claimed height was increasing by 1 SD per half-century (probably an overestimate) and the recent Woodley paper seemed to imply brain size has been increasing by only 0.21 SD per half-century (probably an underestimate).  Nonetheless, a 0.21 SD increase in brain size is exactly what you’d expect if it were caused by a 1 SD increase in height, given that body size and brain size in adult humans correlate between +0.20 and +0.25 (Jensen, 1998; The g Factor, page 147) implying brain size gains should be only 20% to 25% as large as body size gains.

A tertiary effect model would imply that not only are biological IQ gains just a fraction of brain size gains, but brain size gains are just a fraction of height gains.  In other words, very little of the Flynn effect is biological, and thus, almost all must be cultural.  However this is hard to square with the fact that the Flynn effect has been sizeable on some tests that seem relatively culture fair such as Block Design, and for this reason, I have preferred the parallel effects model from the outset.

Future research must determine which of the above three models (if any) best illustrate the relationship between nutrition and the Flynn effect.