How Did "IQ" Become The Standard For Measuring Intelligence?

It's Wednesday, August nineteenth. The show is called HigherIQ, and it occurred to me that I've never actually done an episode about what IQ means. I've used the name for months without fully understanding the thing I named the show after, so I figured it was time. And the first thing I learned is that the guy who invented the IQ test explicitly said it should not be used to measure intelligence. He created it for a completely different purpose, warned against using it to rank people, called the idea of reducing intelligence to a single number "brutal pessimism," and then watched the world do exactly what he said not to do. That same test became the foundation for the SAT and ACT, which determine which college you get into, which shapes the trajectory of your career, which means a test that was never designed to measure intelligence is quietly influencing millions of people's lives every year. In 1904, the French government asked a psychologist named Alfred Binet to develop a way to identify schoolchildren who were struggling and needed extra help. That's all it was, a tool for teachers to figure out which students needed more support, not a way to rank kids or measure innate ability. Binet built a series of tasks that tested memory, attention, and problem-solving, and compared each child's performance to what was typical for their age. He called it "mental age," so a ten-year-old performing like a typical twelve-year-old had a mental age of twelve.

Binet was clear about the limits of what he built. He said intelligence was too complex to capture in a single number. He said the score could be improved with practice and better education. And he specifically warned that using the test to label children as permanently smart or not smart would be, in his words, "a brutal pessimism" that would damage their futures. He died in 1911, and within five years an American psychologist named Lewis Terman at Stanford took Binet's test, renamed it the Stanford-Binet Intelligence Scale, introduced the IQ score formula, and did everything Binet warned against. Terman believed intelligence was heritable and fixed, and he used IQ scores to sort people into categories that determined what kind of education and career they deserved.

Through the 1920s and 30s, IQ tests were used to justify the Immigration Act of 1924, which restricted entry to the US from Southern and Eastern Europe based partly on the claim that those populations scored lower on intelligence tests. They were also used to support forced sterilization programs in over 30 states, where people deemed "feebleminded" based on their test scores were sterilized without meaningful consent. The test that was built to help struggling French schoolchildren became a tool for deciding who was allowed into the country and who was allowed to have children. So if IQ tests aren't a great measure of intelligence, what does the science say about why some people seem to process things faster or solve problems more easily than others?

There is real research on this and some of the findings are surprising. One of the most consistent results in neuroscience is something called neural efficiency: people who score higher on cognitive tests actually use LESS brain energy when solving problems, not more. Their neural pathways are more streamlined, so the brain does the same work with fewer resources. It's not that their brains are working harder. They're working more efficiently, like a newer computer that runs cooler while processing faster.

Brain size matters less than most people assume. What matters more is white matter connectivity, which is the wiring between different brain regions. How quickly and efficiently different parts of your brain communicate with each other correlates more strongly with processing speed and reasoning ability than the size of any individual region. Two people can have similar-sized brains and very different cognitive profiles because the connections between the regions are structured differently.

And the genetic component is real. Twin studies have consistently shown that genetics account for roughly 50 to 80 percent of the variation in IQ scores. Identical twins raised apart, in completely different environments with different families and different schools, end up with remarkably similar IQ scores. So there IS a biological foundation that some people are born with, and pretending otherwise isn't honest. But the other 20 to 50 percent is environment, education, nutrition, and exposure to complex thinking, which is a massive variable that the IQ number alone can't separate from the genetic piece.

This is where the Flynn Effect comes in. Researcher James Flynn discovered that IQ scores have been rising by about 3 points per decade throughout the twentieth century, consistently, across every country where data exists. The average American today would score in the top percentiles on a test from the 1930s. Human brains didn't evolve in 100 years. The conditions people grew up in got better: better nutrition, more education, more exposure to abstract reasoning. The scores went up because the environment improved, which is exactly what Binet said in 1905 when he warned that treating the number as fixed was wrong. The SAT was originally derived from an Army IQ test used to sort military recruits during World War I. The connection between IQ tests and college entrance exams isn't a loose analogy, it's a direct lineage. And the College Board's own data shows that students from families earning over $200,000 per year score an average of 388 points higher on the SAT than students from families earning under $20,000. The test measures something, but separating raw ability from the environment that shaped it is a lot harder than a single score suggests.

Angela Duckworth at the University of Pennsylvania found that among students at an elite university, those with higher SAT scores actually reported slightly less grit, while grittier students earned higher GPAs even after controlling for test scores. In a separate study of eighth graders, self-control predicted grades twice as strongly as IQ. The willingness to keep working at something difficult, to persist through frustration and maintain effort over long periods, was a better predictor of academic success than raw cognitive ability, and it's a trait that no standardized test measures. Carol Dweck's research at Stanford found something similar: students who were praised for being "smart" after a test actually performed worse on the next one because they avoided harder challenges to protect the label, while students praised for "working hard" sought out harder problems because they valued the effort over the identity. How you think about your own intelligence changes how much of it you develop.

And for everyone wondering whether any of this translates to money: there is a correlation between IQ and income, but it's a lot weaker than most people assume. IQ accounts for roughly 10 to 15 percent of the variation in how much someone earns. Family wealth, personality traits like conscientiousness, and social connections all predict income independently of IQ and in some studies predict it more strongly. And above an IQ of about 120, additional points don't really correlate with additional earnings at all. It's a threshold, not a ladder, meaning you need to be smart enough to do the work, but beyond that, the things that determine how far you go have almost nothing to do with a test score.

And this is where I think the whole framework needs to be reconsidered, because ChatGPT scores above 120 on most IQ tests. A machine with no consciousness, no lived experience, and no common sense can outscore most humans on the test that's supposed to measure intelligence. If AI can beat the test, then whatever the test measures, pattern recognition, logical reasoning, data analysis, isn't uniquely human intelligence. It's the stuff machines can replicate. What AI can't do is be creative in the way humans are: asking a question nobody has thought to ask, connecting ideas that don't obviously belong together, feeling something and turning that feeling into a solution. If that's what actually separates human intelligence from machine intelligence, then IQ tests have been measuring the wrong thing since 1905, and the fact that a chatbot can outscore most of us on it kind of proves the point. The show is called HigherIQ. I'm going to keep the name.

Stay informed, stay curious, and we'll see you tomorrow.