https://www.youngju.dev/blog 천천히 올바르게. AI Researcher & DevOps Engineer Youngju's blog. GPU/CUDA, LLM, MLOps, Kubernetes AI workloads, and data engineering — plus mindset essays on confidence, routines, health, and sport psychology. ko fjvbn2003@gmail.com (Youngju Kim) fjvbn2003@gmail.com (Youngju Kim) Sun, 26 Jul 2026 00:00:00 GMT https://www.youngju.dev/blog/psych-papers/dunning-kruger-revisited.en https://www.youngju.dev/blog/psych-papers/dunning-kruger-revisited.en The Peak of Mount Stupid graph that circulates online appears nowhere in the 1999 paper by Dunning and Kruger. The original was four studies that gave Cornell undergraduates tasks in humor, logic and grammar and compared self-assessments across performance quartiles, and the real result shows the top quartile underestimating itself about as much as the bottom quartile overestimates. The sharper objection cuts deeper still: feed in pure random numbers with no relationship between self-rating and score, and the very same graph appears. Regression to the mean and the illusion it manufactures, the 2020 reanalysis of 929 people, and what is still worth keeping about metacognition training. Sun, 26 Jul 2026 00:00:00 GMT fjvbn2003@gmail.com (Youngju Kim) psychologypaper-reviewmetacognitionself-assessmentstatistics https://www.youngju.dev/blog/psych-papers/dunning-kruger-revisited.ja https://www.youngju.dev/blog/psych-papers/dunning-kruger-revisited.ja ネットで出回っている「愚かさの山」のグラフは、ダニングとクルーガーの1999年論文のどこにも存在しません。元論文はコーネル大学の学部生にユーモア、論理、文法の課題を課し、成績四分位ごとに自己評価を比べた4つの研究であり、実際の結果は下位層の過大評価と同じくらい上位層の過小評価を示しています。さらに痛い指摘は別にあります。自己評価と実際の成績にまったく関係のない乱数を入れても、まったく同じ形のグラフが描けてしまうのです。平均への回帰がつくる錯覚、929人を再分析した2020年の論文、それでも捨てなくてよいメタ認知トレーニングの価値までを整理します。 Sun, 26 Jul 2026 00:00:00 GMT fjvbn2003@gmail.com (Youngju Kim) psychologypaper-reviewmetacognitionself-assessmentstatistics https://www.youngju.dev/blog/psych-papers/dunning-kruger-revisited https://www.youngju.dev/blog/psych-papers/dunning-kruger-revisited 인터넷에 떠도는 우매함의 봉우리 그래프는 던닝과 크루거의 1999년 논문에 존재하지 않습니다. 원 논문은 코넬 학부생에게 유머, 논리, 문법 과제를 시키고 성적 사분위별로 자기평가를 비교한 네 개의 연구였고, 실제 결과는 하위권의 과대평가만큼이나 상위권의 과소평가를 보여 줍니다. 더 아픈 지적은 따로 있습니다. 자기평가와 실제 성적이 아무 관계가 없는 난수만 넣어도 똑같은 모양의 그래프가 그려진다는 것입니다. 평균으로의 회귀가 만드는 착시, 929명을 다시 분석한 2020년 논문, 그리고 그럼에도 버리지 않아도 되는 메타인지 훈련의 가치까지 정리합니다. Sun, 26 Jul 2026 00:00:00 GMT fjvbn2003@gmail.com (Youngju Kim) psychologypaper-reviewmetacognitionself-assessmentstatistics https://www.youngju.dev/blog/psych-papers/dunning-kruger-revisited.zh https://www.youngju.dev/blog/psych-papers/dunning-kruger-revisited.zh 网上四处流传的“愚昧之巅”曲线图，在邓宁与克鲁格1999年的论文里根本不存在。原论文是四项研究：让康奈尔大学的本科生完成幽默、逻辑与语法任务，再按成绩四分位比较自我评价；真实结果显示，高分组低估自己的幅度，和低分组高估自己的幅度不相上下。更扎心的批评还在后面——把自我评价与实际成绩换成毫无关系的随机数，照样能画出一模一样的图。向平均回归造成的错觉、重新分析929人的2020年论文，以及即便如此仍然值得保留的元认知训练价值，本文一并梳理。 Sun, 26 Jul 2026 00:00:00 GMT fjvbn2003@gmail.com (Youngju Kim) psychologypaper-reviewmetacognitionself-assessmentstatistics