CrowdScience
CrowdScience

How should we measure cleverness?

The team at CrowdScience have spent years answering all sorts of listener questions, which must make them pretty smart, right? IN this week’s episode, that assumption is rigorously tested as Marnie Chesterton and the team pit their wits against a multitude of mindbending puzzles from an old TV games

Featured Speakers

BBC World Service Host

Topics Discussed

Episode Summary

Executive Summary: The episode asks how cleverness should be measured, using puzzle challenges and a panel of psychologists to examine what IQ tests capture and what they miss. The guests argue intelligence is broad mental adaptability, strongly correlated across tasks, partly heritable, somewhat trainable in test-taking, but not the same as emotional intelligence, creativity, or overall life success.

Main Topics: What intelligence means (Priority: 5/5): The panel defines intelligence as reasoning, problem-solving, and the brain’s ability to adapt to context, while noting that tests only approximate this broad capacity. What IQ tests actually measure (Priority: 5/5): Discussion focused on positive correlations between different cognitive tests, the general intelligence factor, and specific abilities like working memory, processing speed, and novel problem solving. Limits of tests and missing skills (Priority: 4/5): The guests noted that IQ tests may undercount creativity, street smarts, personality, and emotional skills, and that these traits matter in real-world outcomes even if they are not captured well by standard tests. Trainability, practice, and the Flynn effect (Priority: 4/5): The episode explored how test performance improves with familiarity, schooling, and abstract reasoning habits over time, as reflected in rising IQ scores across decades. Genetics and brain research (Priority: 4/5): The panel explained that intelligence is substantially heritable in twin studies, but DNA-based prediction explains only part of that variance, with gene-environment interactions still poorly understood. Sex differences and variability (Priority: 3/5): Experts said average intelligence appears similar for men and women, though there may be small differences in some subskills and in variability at the extremes; they stressed the evidence is mixed and not sufficient to explain social outcomes.

Key Arguments: Intelligence is best understood as general mental adaptability: the capacity to reason, solve problems, and figure out what a situation requires. Standard intelligence tests are useful because scores correlate strongly across diverse tasks and predict academic, job, health, and longevity outcomes. IQ tests do not directly measure all valuable abilities; creativity, personality, emotional intelligence, and practical judgment are partially separate dimensions. Emotional intelligence may reflect a mix of personality and general cognitive ability, so it is not a clean replacement for IQ. Practice can improve test performance, especially through learned strategies, vocabulary, and familiarity with test formats, but fluid reasoning is harder to train directly. The Flynn effect suggests average scores have risen over time, likely due to education, diet, and more abstract modes of thinking, though some countries may be leveling off. Genes matter for individual differences in intelligence, but current molecular genetics captures only a fraction of the heritability seen in twin studies. Average intelligence appears similar by sex, but there may be differences in specific subskills and in score variability; the social significance of those differences remains unproven. IQ is strongly linked to educational and occupational pathways, but only weakly to happiness, relationships, and broader life quality. No single test should be treated as the final word on human potential; psychologists are trying to make assessments fairer and broader.

Data Points: Heritability of intelligence: 50% to 80% of variance - Alex and Sophie described intelligence as highly heritable in twin-study research. Mood-cognition study participants: 25,000 people - Sophie described the MooQ app study testing mood and cognitive function. Countries in MooQ study: 150 countries - The app reached participants around the world. Flynn effect increase: about 3 IQ points per decade - Stuart summarized the average rise in IQ scores across the 20th century. IQ average scale: 100 is average - Used when explaining the Flynn effect and standard IQ scaling. Meta-analysis outcome for emotional intelligence: no added predictive value beyond IQ + personality - Stuart said emotional intelligence did not improve prediction of job performance once IQ and personality were included. Britain gender pay reference: 90p per £1 - Sophie used this as an example while discussing whether IQ differences explain social inequality.

Pivotal Quotes: "intelligence is what the tests test" — Stuart Ritchie: He cited Edwin Boring to explain that test scores are the operational measure, even if they are imperfect. "the adaptability of the brain" — Sophie von Stumm: Her concise definition of intelligence emphasized contextual problem-solving. "we had put on scientific spectacles" — James Flynn: Flynn’s explanation of why modern people may perform better on abstract reasoning tasks.

Implications: IQ remains a useful but incomplete tool: it predicts important outcomes, yet educators and employers should also value creativity, personality, and practical skills. Future testing will likely become broader, fairer, and more biologically informed.

🔓 Sign Up for Unlimited Episode Search

About CrowdScience

We take your questions about life, Earth and the universe to researchers hunting for answers at the frontiers of knowledge.</p>]]></description><itunes:summary><![CDATA[<p>We take your questions about life, Earth and the universe to researchers hunting for answers at the frontiers of knowledge.

View all episodes from CrowdScience