The Problem With IQ Tests

IQ tests are widely treated as a clean, objective measure of “intelligence,” but the underlying science is messier: IQ is strongly linked to real-world outcomes, yet it is also shaped by culture, motivation, test-taking strategy, and historical test design choices. That combination helps explain why IQ scores can predict school performance, job success, and even longevity—while also fueling misuse, controversy, and claims that don’t hold up.

The modern IQ concept traces to early 1900s work on correlations between school subjects. In 1904, psychologist Charles Spearman found that students who did well in one subject tended to do well in others, with a correlation of 0.64 between math and English. He proposed a general intelligence factor, “g,” plus smaller subject-specific influences. Around the same time, Alfred Binet and Theodore Simon built the Binet-Simon test to identify children who needed extra help. Their approach used “mental age” relative to actual age, producing the original idea of an intelligence quotient. In the U.S., Lewis Terman standardized and modified the test into the Stanford-Binet, and later IQ batteries expanded to multiple abilities—memory, verbal, spatial, and numerical—then normalized scores so the population mean sits at 100 with a standard deviation of 15.

Those scores do relate to life outcomes. Higher IQ correlates with larger brain size (a 2005 meta-analysis reported 0.33), and with school achievement. A major study of 13,000 Scottish children measured IQ at age 11 and found correlations around 0.8 with later GCSE marks—suggesting a large share of variation in exam performance can be predicted from earlier IQ. IQ also tracks educational attainment and is often comparable to standardized tests like the SAT, ACT, and GRE (correlations around 0.8). Outside school, IQ shows moderate links to occupational success (often 0.2 to 0.6), with the strongest effects in complex roles; the U.S. military historically used IQ thresholds and found that lowering them increased failure rates and remedial training needs.

Yet IQ is not a pure readout of fixed ability. The transcript highlights the “Flynn Effect,” where average IQ scores rise over decades even though genetics don’t change quickly. James Flynn’s work suggests re-normalizations would otherwise reveal a roughly 30-point increase over the last century, likely driven by improved nutrition and health, better education, and shifts toward more abstract work. Even within IQ testing, performance can move with incentives: paying test-takers can raise scores, sometimes by up to 20 points, and coaching can boost results by several points. Time pressure, anxiety, and motivation all matter.

The controversy is also historical and political. The transcript links IQ’s U.S. adoption to eugenics: Henry Goddard’s interpretation of inherited, unchangeable intelligence helped justify forced sterilization laws, upheld by the Supreme Court in 1927, with tens of thousands sterilized and later influence claimed by Nazi Germany. Modern researchers emphasize that IQ is partly heritable but also environment-dependent, with estimates often around a 50/50 split in twin studies.

In the end, IQ is best treated as a useful predictor—not a verdict on worth. It can help identify strengths and support decisions (including in education and clinical settings), but it also measures more than “g,” and it can be distorted by culture and incentives. The transcript’s central message is that IQ scores can be informative while still being incomplete—and that how society uses them matters as much as what they measure.