There's a real, measurable cognitive trait called "intellectual humility" that turns out to predict accuracy on almost every kind of judgment task — and the people who score highest on it aren't the ones who know the most, they're the ones who hold their existing beliefs slightly more loosely than everyone around them

The standard cultural register, when it considers the question of who is good at making accurate judgments, tends to focus on intelligence as the primary variable. The smarter person is, in the standard framing, the more reliable judge. The framing is intuitive. The framing is also, on the available evidence, considerably less accurate than the wider culture has been treating it as.

The variable that turns out to predict accuracy on a wide range of judgment tasks, on the basis of research conducted across the last fifteen years, is a different one. The variable is called intellectual humility. The variable is structurally distinct from intelligence, and the variable is, by the available empirical measures, a better predictor of accuracy on certain kinds of judgments than intelligence is.

The people who score highest on intellectual humility are, on close examination, not the ones who know the most. They are, more accurately, the ones who hold their existing beliefs slightly more loosely than the people around them. The slight looseness is, in some real way, what most of the predictive power is being produced by.

What the trait actually is

It is worth being precise about what intellectual humility consists of, because the term has been used in a number of different ways across the various research programs that have studied it.

The wider research literature has converged on a working definition. The Open Encyclopedia of Cognitive Science defines intellectual humility as “an awareness of and appropriate attentiveness to one’s personal intellectual limitations.” The term started being used in this technical sense in the 2010s, followed by exponential growth in the research literature after multiple research groups developed measurement scales for it simultaneously. The trait is distinct from modesty, which involves the social presentation of oneself as not being overly impressive. Intellectual humility is, more specifically, a cognitive disposition involving the recognition that one’s own beliefs might be wrong and the corresponding willingness to revise them in light of evidence.

The research has consistently distinguished intellectual humility from intellectual diffidence, which involves excessive doubt about one’s beliefs regardless of evidence. Intellectual humility is calibrated. The intellectually humble person holds their beliefs with the appropriate level of confidence given the available evidence, neither over-confident in beliefs the evidence does not fully support nor under-confident in beliefs the evidence does support.

This is, on close examination, a structurally different cognitive style from what the wider cultural register typically associates with intelligence. Intelligence, in its standard measurement, is calibrated to the ability to produce correct answers on tests with known correct answers. Intellectual humility is calibrated to the ability to accurately assess one’s own confidence in beliefs across a wider range of situations, including the many situations in which the correct answer is not, at the time of the judgment, known.

What the trait actually predicts

The empirical research on intellectual humility has, in the last decade, produced a consistent set of findings about what the trait predicts. The trait predicts, on the available evidence, better performance on a wide range of judgment tasks that the standard intelligence measures do not particularly predict.

One of the more striking findings comes from research on misinformation discernment. A 2024 study by Prike and colleagues tested participants on their ability to distinguish true from false news headlines, while also measuring their intellectual humility on validated scales. The result was that individuals with higher intellectual humility were significantly better at distinguishing true from false headlines. The effect was specifically related to discernment rather than response bias. The intellectually humble participants were not, in other words, simply more skeptical of everything. They were, more specifically, better at distinguishing what was true from what was not. The strongest predictor in the study was a specific facet of intellectual humility called Actively Open-Minded Thinking about Evidence, which produced a correlation of 0.48 with misinformation discernment.

The effect extends beyond misinformation. Research on climate change interpretation has found that more intellectually humble participants were better at distinguishing correct from incorrect interpretations of evidence about a polarized scientific topic. The effect held across the political spectrum, meaning that the trait was producing better judgment regardless of which political conclusions the participant initially preferred. The effect was specifically associated with what the researchers called metacognitive ability, the capacity to introspect on one’s own performance and to assign appropriate confidence to varying accuracy of one’s own judgments.

The effect also extends to memory and learning. Research on recognition memory has found that intellectually humble participants considered sentences counter to their own opinions for longer, and also distinguished more successfully between sentences they had read previously and sentences they had not. The intellectually humble participants were, in some real way, processing new information more carefully than their less intellectually humble counterparts, and the more careful processing produced more accurate downstream judgments.

Why the trait works, mechanically

The mechanism by which intellectual humility produces better judgment is, on the available evidence, reasonably well understood. The mechanism operates through what the research literature calls metacognitive calibration. The intellectually humble person is, by structural design, better at assessing the accuracy of their own judgments. They know, more accurately than other people, when they are likely to be wrong. The knowing-when-wrong is what allows them to seek out additional information in the cases where it would be most useful, to defer to others on the questions where their own confidence does not warrant deference, and to revise their beliefs in response to new evidence at the rate the evidence actually warrants.

The wider population, by contrast, tends to display what the research literature calls overconfidence. Most people, when asked to rate the accuracy of their own beliefs, rate the accuracy higher than the beliefs actually warrant. The overconfidence is not, in most cases, conscious. The overconfidence is, more accurately, a structural feature of how the average cognitive apparatus operates. The intellectually humble person has, by some combination of temperament, training, and accumulated self-correction, developed an apparatus that operates with less of this structural overconfidence than the average.

The less-overconfident apparatus is what produces the better judgment. The apparatus is not, in any deep sense, smarter. The apparatus is, more accurately, more accurately calibrated to its own limitations. The accurate calibration is what allows it to operate more effectively in situations where the standard cognitive bias toward overconfidence would otherwise be producing systematic errors.

Why this is not what intelligence measures

The structural distinction between intellectual humility and intelligence is worth dwelling on, because the wider cultural register has tended to collapse the two.

Intelligence, as standardly measured, predicts performance on tasks with clear correct answers. The intellectually humble person is not, by their humility, necessarily better at these tasks. The intellectually humble person is, more specifically, better at the second-order task of knowing how confident to be about their own answers, which is a different cognitive operation.

This means that the two traits can vary independently. A highly intelligent person can be intellectually arrogant, in which case their intelligence produces correct answers in many cases but their arrogance produces systematic errors in the cases where their initial intuition is wrong. A less intelligent person can be intellectually humble, in which case their initial answers are correct less often but their willingness to revise produces better final judgments than the more intelligent but more arrogant alternative.

The empirical research has been finding, on close examination, that the second configuration often produces better real-world outcomes than the first. The configuration is not what the wider culture has been selecting for. The wider culture has been selecting, in most cases, for visible intelligence, which is correlated with confidence but not, on close examination, with the calibration of confidence to actual accuracy. The selecting has produced, in most organizational contexts, a structural over-representation of intellectually arrogant high performers and a structural under-representation of intellectually humble ones. The over-representation has, on the available evidence, been producing systematic costs in domains where accurate judgment under uncertainty matters more than the visible production of confident answers.

The acknowledgment this article wants to leave

The empirical research on intellectual humility has, over the last decade, established a reasonably clear set of findings. The trait is real. The trait is measurable. The trait predicts accuracy on a wide range of judgment tasks better than intelligence does in the same domains. The trait is, on the available evidence, calibrated to the second-order question of how confident to be about one’s own beliefs, which turns out to be the question that most affects real-world judgment accuracy in domains where the correct answer is not, at the time of judgment, available.

The people who score highest on the trait are not, on close examination, the people who know the most. They are, more accurately, the people who hold their existing beliefs slightly more loosely than the people around them. The slight looseness is what produces the better calibration. The better calibration is what produces the better judgment. The better judgment is what the research, across multiple independent labs and