Evidence Check
“IQ pilled” and “intelligence pilled” get thrown around online to describe someone convinced that IQ quietly runs everything — careers, money, relationships, decisions. Some of that claim is backed by real, well-replicated research. A lot of it is overstated. This page walks through what the actual evidence says, domain by domain, with sources you can check yourself.
Short version: IQ is a genuinely useful predictor in a few specific areas, a weak one in others, and nowhere close to deterministic in any of them.
The “-pilled” construction traces back to the red pill / blue pill scene in The Matrix, later adopted across internet culture to mean “converted to a particular belief.” It gets attached to almost anything — gym-pilled, finance-pilled, book-pilled — usually lightly. Applied specifically to IQ, the term more often shows up in heavier discourse: tech and rationalist-adjacent corners of the internet arguing that raw intelligence is an underrated, dominant predictor of who succeeds at what.
That claim deserves an actual answer instead of vibes in either direction. The research below is real, peer-reviewed, and decades deep — and it points to a more specific, more interesting picture than either “IQ is everything” or “IQ is meaningless.”
Five commonly claimed domains, ranked by the actual strength of the research behind each one — not by how confidently people talk about them online. Expand any row for the specific study.
Where IQ Actually Predicts Real-World Outcomes
Ranked by strength of evidence, not assumption. Expand any row for the specific research behind it.
One of the best single predictors of performance across nearly all occupation types.
A landmark 1998 meta-analysis by Frank Schmidt and John Hunter, published in Psychological Bulletin, found general mental ability has a validity coefficient of roughly 0.5 for predicting overall job performance — rising further in more complex roles. Among more than 20 selection methods studied (interviews, personality tests, reference checks), it was one of the strongest single predictors identified.
Source: Schmidt & Hunter (1998), Psychological Bulletin
Strongly associated with years of education and academic achievement.
IQ-type measures correlate with academic achievement and years of completed education at roughly 0.5–0.6 in most large studies — among the more robust associations in the social sciences. The relationship runs in both directions, though: schooling itself measurably raises scores over time, which is part of why the correlation is so strong.
Source: Widely replicated finding in educational psychology research
A real but modest relationship — other factors matter more.
Economist Jay Zagorsky's 2007 study in the journal Intelligence, using long-running U.S. survey data, found IQ correlates with income and wealth in roughly the 0.2–0.4 range — real, but leaving most of the variation explained by other factors: family background, education access, career choice, and financial behavior.
Source: Zagorsky (2007), Intelligence
Childhood IQ measurably predicts later-life health outcomes.
Psychologist Ian Deary's research using the Scottish Mental Surveys of 1932 and 1947 — which tested nearly every child born in Scotland those years — found childhood IQ scores predict adult health outcomes and mortality risk decades later, a field now known as cognitive epidemiology. The effect is real and replicated, though modest relative to factors like smoking, socioeconomic conditions, and healthcare access.
Source: Deary et al., Scottish Mental Survey cognitive epidemiology research
The most commonly overestimated correlation on this list.
Psychologist Keith Stanovich's research distinguishes IQ from rational thinking — the ability to make good decisions under uncertainty. His work found people with high IQ scores are just as prone to common reasoning errors (overconfidence, sunk-cost fallacy, myside bias) as anyone else. Intelligence and rationality are correlated, but far less tightly than most people assume.
Source: Stanovich, What Intelligence Tests Miss (2009)
Strength labels reflect the general weight of evidence across the cited research, not a single precise statistic — different studies use different methods and populations. None of these relationships are deterministic: within any domain, individual outcomes vary widely regardless of score.
The strongest correction to the “IQ pilled” worldview isn't that IQ doesn't matter — it's that people consistently overestimate what it predicts and underestimate everything else in the equation.
This is the single most counterintuitive finding in the research above. Keith Stanovich's work on “dysrationalia” found that smart people fall for overconfidence, sunk-cost reasoning, and motivated thinking at rates not meaningfully different from everyone else. Intelligence gives you more raw processing power; it doesn't automatically point that power in a rational direction. Rational thinking is trainable as its own skill, separate from raw reasoning speed.
Researcher James Flynn documented average IQ scores rising across most of the 20th century — a trend, now called the Flynn Effect, driven by environmental factors like education access, nutrition, and exposure to abstract problem-solving, not genetic change across generations. At the individual level, testing conditions and format familiarity produce real score movement between attempts too — see the IQ maxxing guide for what specifically moves the number.
Even in the strongest domain above — job performance — a correlation of 0.5 leaves most of the variation in outcomes explained by other things: conscientiousness, interpersonal skill, domain-specific experience, and plain opportunity. A correlation, however strong, describes a population trend. It does not predict any individual person's outcome with confidence.
~0.5
Job performance validity
Schmidt & Hunter's 1998 meta-analysis found this among the strongest of over 20 selection methods studied.
1936
First Scottish Mental Survey
Nearly every child in Scotland tested that year, later enabling decades of cognitive epidemiology research.
5
Life domains examined
From strong (job performance, education) to weak-to-moderate (decision-making quality).
It's internet shorthand from the "-pilled" slang family (originating from the red pill/blue pill scene in The Matrix) applied to intelligence. In practice it gets used two ways: casually, to describe someone deep into intelligence-related topics, and more seriously, to describe someone who has become convinced IQ is a dominant, largely fixed force in life outcomes. This page focuses on the second use, and checks it against actual research rather than online rhetoric.
Yes, they're used interchangeably online — "intelligence pilled" is just the less-abbreviated version of the same idea.
No. IQ is one of the stronger predictors of job performance and educational attainment in the psychological research literature, but it explains a meaningful minority of the variation in most life outcomes, not most of it. Conscientiousness, opportunity, family background, health, and plain circumstance all do substantial work too. The honest summary is "IQ matters, and it is nowhere near the whole story."
Less than most people assume. Psychologist Keith Stanovich's research found that high-IQ individuals are just as susceptible to common reasoning errors — overconfidence, sunk-cost thinking, motivated reasoning — as anyone else. Intelligence and rational decision-making are related but meaningfully distinct skills, and the second one is trainable independent of the first.
Individual scores are relatively stable in adulthood under consistent testing conditions, but population-level scores are not fixed at all — the Flynn effect documents measured average scores rising across much of the 20th century due to environmental factors like education access and nutrition, not genetic change. Testing conditions, format familiarity, and practice also produce real score movement between individual attempts. "Fixed" is an overstatement in either direction.
What actually moves your score, and what doesn't — the practical counterpart to this page.
The specific cognitive construct most IQ-style tests, including this one, are built to measure.
What standard error means and how much weight to put on any single score.
Free to start, built on a fixed Raven-style item set with a consistent scoring model — a directional baseline, not a verdict on anything the research above covers.