A major theme of this blog has been investigating some mysterious changes to human intelligence that have occurred since the 19th century. Since intelligence is highly genetic, and genetic change tends to be extremely slow, it was shocking to find evidence that intelligence significantly changed since the 19th century, and it’s especially shocking to find evidence of it going dramatically up and down at the same time. For example, on measures of “fluid IQ”, intelligence has increased by over 30 points since the 19th century. However measures of general intelligence indirectly estimated from simple reaction time (a measure of physiological brain speed) suggest that real IQ has actually dropped 15 points.

How do we resolve this fascinating paradox? By looking for other ways to measure intelligence. In a previous post I looked at the SAT but confined myself to only trends among the gifted. In this post I will look at the average person, because that’s how long term changes in intelligence are usually measured.

Of course the average American doesn’t take the SAT. Usually only the top third most academically ambitious take the test, however on page 422 of the book The Bell Curve by Herrnstein & Murray, we learn about rare studies where an abbreviated SAT was given to nationally representative samples of ALL high school juniors from 1955 to 1983. Juniors are more nationally representative than seniors because they include students who drop out before their senior year, and the data was statistically adjusted to reflect how the juniors would have scored had they been tested at the normal SAT age of 17.

The results were as follows. In 1955 and 1983, the average verbal score was 348 and 376 respectively. Converting to the IQ scale, that’s a gain of 1.29 IQ points a decade.

Meanwhile the average math score was 417 in 1955 and 411 in 1983. Converting to the IQ scale, that’s a loss of 0.25 IQ points a decade.

In my previous analysis of the highest SAT scores, I found that verbal scores went down while math scores went up, yet when looking at average people, we find the reverse trend. Maybe it’s just statistical noise that shouldn’t be over-interpreted and the overall score is the more reliable indicator. If we add the verbal and math means, we find that the average American teen could get a combined SAT score of 765 in 1955 and 787 in 1983. Converting to the IQ scale, that’s a gain of 0.5 IQ points a decade.

It’s interesting that on self-described IQ tests, scores have increased by 3 points a decade, but when intelligence gains are estimated from physiological reaction time tests, IQ has DECREASED by 1.2 IQ points a decade. Meanwhile changes in SAT scores are intermediate.

How do we explain this? Well, self-described IQ tests include a lot of spatial ability and according to scholar Richard Lynn, spatial ability is more sensitive to prenatal nutrition which was impaired in your parents’ generation who were born with smaller brains. Secondly, self-described IQ tests are given to people with a wide range of education levels, while the SAT study controlled for education since all the test takers were high school juniors. This is relevant because education has increased a lot over the 20th century, and many mental tests are biased against the less schooled.

So what this may tell us is that largely controlling for both nutrition and education, IQ scores have been increasing by 0.5 points a decade, implying a 6 IQ point gain since the Victorian era.

Controlling for 20th century socio-economic gains

The SAT study controls well for nutrition because the SAT measures little nutrition loaded spatial ability, and controls well for education because only high school students are tested, however it does not control for the huge 20th century rise in socio-economic status. Scholastic achievement tests might be much more sensitive to socioeconomic status than self-described IQ tests are because according to page 475 of The g Factor by Arthur Jensen, African Americans teenagers raised by professional white parents scored 10 IQ points higher than black Americans raised by their biological parents when scholastic scores were converted to IQ. This suggests that SAT scores can be propped up by having educated rich parents. Well modern teens have far richer and more educated parents and far more access to technology and media than teens in the 19th century. Indeed considering the huge 20th century gains in wealth, education, media and technology, the cultural advantage of teens today over teens in the 19th century is probably double the cultural advantage of white professional homes over average black homes living contemporaneously, so if the latter caused a 10 IQ point boost, then the former should cause a 20 IQ point boost. So if we subtract those 20 spurious culturally caused IQ points from modern SAT performance, then instead of IQ going up by 6 points since the Victorian era, it’s actually gone down by 14 IQ points, which is exactly what the studies of physiological reaction time suggest.

Why are people getting so unintelligent so quickly?

Scholars Michael A. Woodley, Bruce Charlton and their colleagues suggest it’s because advances in health care have caused genetic mutations to accumulate at an alarming rate, meanwhile the earliest IQ experts noted that IQ is highly genetic and low IQ people have more kids.

If we’re getting so dumb, so quickly, why hasn’t society collapsed?

Evidence suggests innovation rates have collapsed but Bruce Charlton believes we were so innovative in the 19th century that we are still benefiting from that momentum. Also, the wealth society has inherited has caused a huge rise in education and nutrition. The latter boosts brain size and probably brain development which I believe has boosted spatial ability, but it hasn’t increased physiological brain speed because that’s so essential to our intelligence that evolution likely made it largely insensitive to nutrition. When nutrition is scarce, our biology probably evolved to send all the nutrients to the most fundamental and general parts of intelligence, and only when nutrients are abundant do the spatial abilities needed to build our high tech era get sufficiently fed. Spatial ability would be a lower priority than verbal-numerical ability because humans live in groups so we access centuries of accumulated cultural knowledge through language and symbols and seldom have to figure out to build or create objects through our own spatial reasoning. And both verbal and spatial would be lower priorities than general intelligence so our biology gives it all the nutrients first which is why it’s so insensitive to changes in nutrition.