The Blind Spot of Metacognition — The Dunning-Kruger Effect and the Structure of Not Knowing What One Does Not Know

"To notice that one does not know what one does not know" — this is a cognitive task harder than we think. The tendency for those with lower capacity in a domain to estimate their capacity higher — commonly called the Dunning-Kruger effect — is the most cited phenomenon in overconfidence research and at the same time the most misunderstood finding. This volume traces the effect honestly, from the original 1999 paper through the 2020s statistical reassessment, and discerns what is genuinely established and what is exaggerated. Justin Kruger and David Dunning's classical experiment (1999), John Flavell's metacognition theory (1979), the contemporary statistical critiques of Nuhfer and Gignac & Zajenkowski, Zhuangzi's well-frog, Bertrand Russell's warning, and the analogy from the neuroscience of anosognosia (病態失認) — at the intersection of eight perspectives we observe the deepest root of overconfidence. The blind spot cannot be fully seen. But that the blind spot "exists" can be consciously observed. This is the central proposition of the volume.

01The Problem of Not Knowing What One Does Not Know

In 1989, in Pittsburgh, there was a bank robbery. A man named McArthur Wheeler openly robbed two banks in front of the security cameras. In the police interrogation after his arrest, he is recorded as saying, with astonishment, "But I had lemon juice on my face." From the fact that lemon juice can be used as an invisible ink, he had "convinced" himself that, with lemon juice on his face, the camera would not capture his face.

In 1999, the Cornell psychologist David Dunning (1950–) and his then-graduate student Justin Kruger (1969–) cited this episode as the starting point of overconfidence research. "Those with extremely low capacity often cannot notice their own low capacity" — this is the theme of the volume. Because one cannot notice, one cannot correct. Because one cannot correct, capacity remains low. And the person is full of confidence. This triple structure distorts self-evaluation.

What this volume asks is why this structure arises, and how far we all — experts included — are governed by the same structure. The previous volume's theme, "overconfidence as the most lethal bias," is dug into from the perspective of metacognition.

02Kruger & Dunning (1999) — The Discovery of the Classical Experiment

Kruger and Dunning's 1999 paper "Unskilled and Unaware of It" (Journal of Personality and Social Psychology) is one of the most cited papers in contemporary psychology. They had university students take tests of grammar, logical reasoning, and humour (whether a joke is funny), and compared the actual scores with the self-estimated scores.

"Our finding was striking. Subjects whose scores were in the bottom 25% believed that they had scored around the top 40%. Subjects in the top 25%, in contrast, tended to slightly underestimate their scores. The lower the capacity, the higher the self-evaluation relative to reality — and the higher the capacity, the less the person estimates themselves above others." — paraphrase of Kruger & Dunning, JPSP (1999)

The interpretations they offered are double, and remain at the core to this day. First, those with low capacity in a domain also lack the capacity required to evaluate their own performance in that domain. The person who makes grammar mistakes lacks the grammatical capacity to notice that the grammar is mistaken. Second, those with high capacity tend, conversely, to assume "others must understand as much as I do" (related to the false-consensus effect).

This finding had decisive influence on overconfidence research and educational psychology. "Awareness of ignorance presupposes a certain amount of knowledge" — this is the core proposition of the Dunning-Kruger effect.

03Metacognition as a Capacity — Flavell's Theory

The theoretical background of the Dunning-Kruger effect is the psychology of metacognition. The one who first systematised the concept is the Stanford developmental psychologist John Flavell (1928–). In his 1979 paper in American Psychologist, "Metacognition and cognitive monitoring," he defined:

"Metacognition is knowledge about one's own cognitive processes and the capacity to monitor them. What am I knowing or not knowing right now; do I hold the information needed to solve a given problem; can I trust my own answer — the capacity to judge these constitutes the core of human higher cognition." — paraphrase of John Flavell, American Psychologist (1979)

What Flavell showed is that knowledge itself (cognition) and knowledge about knowledge (metacognition) are different capacities. To hold much knowledge in a domain and to grasp accurately how far one understands in that domain are different capacities. When the latter does not function, the "Overprecision" of Moore & Healy treated in Vol. 1 rises. The precision of metacognition is the breakwater against overconfidence. This is the practical core of the series.

Developing Flavell's framework, Thomas Nelson and Louis Narens, in 1990, presented the two-layer model of monitoring and control of metamemory. The monitoring of "do I think I remember?" and the control of "do I take action to strengthen memory?" — the two are different capacities, both required. The practices for correcting overconfidence treated in this series are also based on this two-layer structure.

04Contemporary Reassessment of the Dunning-Kruger Effect — The Statistical Critique

In the twenty-first century, an important statistical critique of the Dunning-Kruger effect has appeared. The 2016 paper of Edward Nuhfer and colleagues (Numeracy), and the 2020 paper of Gilles Gignac and Marcin Zajenkowski (Intelligence), pointed out that the original figure of Kruger & Dunning may contain a statistical artefact.

"The original figure of Kruger & Dunning (overestimation by the lower-capacity group, underestimation by the upper) can also be reproduced by a combination of data containing random noise, the better-than-average effect, and regression to the mean. In independent verifications of capacity and self-evaluation, the effect size is smaller, and the overestimation by the lower group in particular has been disproportionately emphasised." — paraphrase of Gignac & Zajenkowski, Intelligence (2020)

To be honest as scientists, let us convey the implication of this critique candidly. The Dunning-Kruger effect does not entirely fail to exist, but it is not as dramatic as the early claim. The extreme picture in which the low-capacity person believes themselves at the top is not accurate in the form often related in the media. The actual effect is modest and depends on context.

* Dunning himself, since around 2018, has responded to parts of these critiques, maintaining the existence of the effect while admitting that extreme figures such as "the bottom 25% evaluate themselves as top 40%" are context-dependent. The position of this series is consistent with the contemporary standard evaluation — "restrain extreme interpretation of the effect; preserve the core implication."

05What Is the True Discovery? — Where the Core of DK Remains

So where does the core of the Dunning-Kruger effect remain after the statistical critique? Three implications continue to be confirmed by empirical research.

The most important point for the reader of this series is Dunning's own repeated warning that even after learning about the effect, applying it to oneself is not easy. One can observe others — "those foolish people have fallen into the DK effect" — but to observe oneself — "I, too, may have fallen into the DK effect" — is far harder. This volume takes the final step with the reader.

06Zhuangzi's Well-Frog and Russell's Warning — The Same Place in the Classics

The structure of not knowing what one does not know is not a discovery of contemporary psychology. In ancient China, Zhuangzi (c. 4th c. BCE), in the Qiushui (Autumn Floods) chapter of Zhuangzi, holds the famous metaphor of "the frog in the well".

"The frog in the well cannot speak of the sea, because it is bound to its hollow (its world). Cannot speak of ice to the summer insect — for the same reason. We are all bound to the hollow we inhabit." — paraphrase of Zhuangzi, Zhuangzi, "Autumn Floods"

The implication of Zhuangzi is deep. The frog in the well does not know the sea. Not only that — it does not know that it does not know. Because the very concept of "the sea" is outside the experience inside the well. This is the expression of the Dunning-Kruger effect 2,400 years earlier.

The twentieth-century British philosopher Bertrand Russell (1872–1970), in his 1933 essay "The Triumph of Stupidity," wrote:

"The fundamental cause of the trouble in the modern world is that the stupid are cocksure while the intelligent are full of doubt." — Bertrand Russell, "The Triumph of Stupidity" (1933)

The same place was pointed to by the Irish poet W. B. Yeats (1865–1939) in his 1919 poem "The Second Coming" — "The best lack all conviction, while the worst / Are full of passionate intensity." Zhuangzi, Russell, Yeats — three figures in different times and different cultures pointing to the same place as the Dunning-Kruger effect. This is not a discovery of contemporary psychology but a phenomenon woven into the very structure of human cognition.

07The Expert's Blind Spot — The Known Covers the Unknown

Vol. 1 cited Tetlock's research. The more confident the expert in their prediction, the more they tend to miss. This is overconfidence arising by a different route from the Dunning-Kruger effect, but the root is at the same place — the blind spot of metacognition.

The expert's blind spot takes two forms. First, "the known covers the unknown." In a domain known well, one feels as if "what one does not know" is also visible in one's own way. But what one truly does not know — questions not yet asked, new revisions, unexpected contexts — is by definition outside the frame of the known. Second, "the confusion of breadth and depth of knowledge." The broadly knowledgeable tend unconsciously to assume they have depth as well. In fact they are different capacities.

The implication for the reviewer is clear. It is in the domain of long-accumulated experience that the Dunning-Kruger trap operates quietly. The sense "I understand this domain" generates oversight. Checking new revisions; serious listening to beginners' questions; cross-checking with another reviewer — these are concrete means of observing the blind spot. This is the foreshadowing of "calibration" treated in Vol. 8 of this series.

08Anosognosia — The Neuroscience by Which One's Own Disorder Is Invisible

From neurology, the most extreme form of the blind spot. Anosognosia (病態失認), named in 1914 by the French neurologist Joseph Babinski (1857–1932), is the phenomenon in which a patient with damage to a specific brain region (mainly the right parietal lobe) cannot recognise the very existence of their own paralysis or disorder.

"When a patient with left-side paralysis from a right-hemisphere stroke is asked, 'Please move your left hand,' the patient answers, 'I just moved it.' In fact it has not moved. The patient is not lying; they truly recognise that they 'moved it.' Because the brain circuit that recognises the disorder itself is not functioning, the disorder is invisible." — neurological observation (clinical reports since Babinski 1914)

Anosognosia is pathological, not overconfidence itself. But there is no other phenomenon that so vividly shows that "the capacity to recognise one's own deficit" is an independent cognitive function. Even in a healthy brain this capacity is not complete. The Dunning-Kruger effect can be observed as a mild functional anosognosia — this is one way contemporary neuropsychologists understand it.

The implication is solemn. We can all enter an anosognosia-like state in some domain. The circuit recognising one's own blind spot does not fully operate in that domain. For just this reason, external feedback, the eyes of others, and objective cross-checking become structural means of compensating for the blind spot.

09In the Pharmaceutical Workplace — The Reviewer's Metacognition

Let us bring the abstract down to the field. Concrete scenes in the reviewer's daily work in which the blind spot of metacognition is tested.

10Five Practices for Observing the Blind Spot

To bring the same place to which the eight traditions pointed down into daily practice, five practices are set here.

The structure of not knowing what one does not know is a phenomenon woven deeply into human cognition. The dramatic figures of the Dunning-Kruger effect's original paper (1999) have been moderated by contemporary statistical critique, but the core proposition — the precision of metacognition is the breakwater that holds overconfidence in check — remains established. Flavell's theory of metacognition, Zhuangzi's well-frog, Russell's warning, the neuroscience of anosognosia — different traditions have pointed to the same place. The blind spot cannot be fully seen. But that the blind spot "exists" can be consciously observed. This is the central proposition of the volume.

The next volume of this series (Vol. 3) treats another layer of the unconscious that supports overconfidence — Freud's defence mechanisms. Where the Dunning-Kruger effect observes the "blind spot of cognition," the defence mechanisms treat the "motivated blind spot." The two arrive at the same place by different routes.