A Stanford psychologist spent 35 years trying to prove that high IQ produced genius. He selected 1,528 of the smartest children in California and tracked them for the rest of their lives. Not one of them won a Nobel Prize. Two of the boys he had rejected from the study won the Nobel Prize in Physics. The trait he had built his entire career on did not predict the thing he thought it predicted. His name was Lewis Terman. The study is one of the most honest accidents in modern psychology. In 1921, Terman was the most famous psychologist in America. He had translated and adapted the original French intelligence test into the version that would dominate American schools for the next 50 years. He called it the 'Stanford-Binet'. He believed, with the certainty of a man who had built a career on a single idea, that intelligence was the master variable behind every form of human achievement. The doctors, the inventors, the senators, the artists, the great writers and great scientists. All of them, in his model, were sitting at the top end of the same bell curve. If you could find the children with the highest scores, you could predict the future leaders of the country. So he set out to prove it. He sent his research team into California schools and screened roughly 168,000 children. He had teachers nominate their brightest pupils. He gave the nominees the Stanford-Binet. He kept the ones who scored 135 or higher, which placed them in roughly the top one percent of the population. The final sample was 1,528 children, average age 11. They had a name in his lab notebooks within a year. Termites. He planned to follow them for the rest of their lives. He died in 1956 having tracked them for 35 years. Stanford kept the study going. The last surviving Termites were tracked until the 2000s. The data set is one of the longest continuous psychological studies in human history. Here is what the data showed. The Termites did well. They went to college at higher rates than their peers. They earned more money. They became professors and engineers and lawyers and physicians at higher rates than the general population. Terman was not entirely wrong. High IQ is correlated with conventional success. The correlation is real and the effect size is meaningful. But that was not what he had set out to prove. He had set out to prove that high IQ produces genius. The kind of genius that wins Nobel Prizes, writes great novels, founds new fields, and reshapes the technological direction of the world. And on that specific question, the dataset turned on him. None of the 1,528 Termites won a Nobel Prize. None of them won a Pulitzer. None of them became world-class musicians. None of them produced a single piece of work that historians of science or art still talk about. They were accomplished. They were comfortable. They were not, in any sense Terman would have recognized in his original ambition, geniuses. The detail that haunts the study is what happened to the children he rejected. In the screening phase, his team had tested two boys named William Shockley and Luis Alvarez. Both scored below the cutoff. Both were sent home. Shockley went on to co-invent the transistor and win the 1956 Nobel Prize in Physics, the same year Terman died. He founded the company that seeded the entire ecosystem we now call Silicon Valley. Alvarez won the 1968 Nobel Prize in Physics for his work on subatomic particles, and later proposed the asteroid impact theory of dinosaur extinction that turned out to be correct, too. Two of the most consequential American physicists of the 20th century had been measured by Terman's own instrument and judged not gifted enough to be worth tracking. There is an important caveat here that the more honest critics have raised in recent years. A 2020 simulation study from researchers at Utah Valley University showed that even with a perfect IQ test, the base rate of Nobel Prizes is so vanishingly low that Terman would have been statistically unlikely to catch a future laureate in any sample of his size, no matter where he set the cutoff. The Shockley and Alvarez story is dramatic but it does not, on its own, prove that IQ does not matter. It proves that rare outcomes are hard to predict from any single variable, including a very good one. That caveat is real. It is also not the most important thing the study showed. The most important thing the study showed is what Terman himself eventually admitted, late in his career, in a quieter voice than he had used for the previous three decades. He wrote that the relationship between intelligence and achievement was, in his words, far from perfect. Within the Termite sample itself, the highest-IQ children did not become the most accomplished adults. The variation in outcomes inside the group of geniuses was enormous, and IQ explained almost none of it. Some of the Termites had unremarkable careers. Some of the Termites had remarkable ones. The thing that distinguished the two groups was not the score he had used to select them. What distinguished them, when researchers eventually analyzed the data more carefully, was a cluster of traits Terman had not been measuring. Persistence. Curiosity. Health. Stable family circumstances. The willingness to keep going when a project stopped being interesting and started being hard. Most of the Termites who went on to do meaningful work were not the ones with the highest scores. They were the ones who had spent decades grinding on a single problem. The lesson is the part that should change how anyone reading this thinks about talent. The trait you select for is the trait you optimize for. If you measure children on a test of pattern recognition and verbal recall, you will find children who are good at pattern recognition and verbal recall. You will not find the children who will spend 30 years thinking about a single equation. You will not find the children who will quietly read the same difficult book six times. You will not find the children whose curiosity is wider than their working memory. Those traits do not show up on the test you are running, which means they do not show up in the dataset you build. Terman spent his life trying to find genius and ended up proving that he had been measuring the wrong thing all along. The kids he rejected were not stupider than the kids he kept. They were running a different program underneath, and his instrument could not see it. The trait you can measure is almost never the trait that actually matters. Most people building careers, hiring teams, and raising children are still selecting for the version of the trait that fits on a test.
