The Subtle Art of Lying with Statistics
summary
Shows concrete ways numbers mislead: arithmetic vs harmonic mean gives 1.56 versus 0.53 for the same dataset; geometric mean 1.01 and mode 1. Demonstrates range matters (Death Valley mean 16.1°C hides −5°C to 40°C) and identical means can have SD 3.3 or 44.8. Highlights sampling and selection tricks (tiny samples, 4/5 doctor stunt), graph manipulations like cutouts and log scale distortion, misleading percentages and data dredging (cherry‑picked 1997–2012 window) to create false claims.
AI-written summary · about
chapters (19)
- 0:00Introduction and welcome
- 1:03Statistics can easily mislead
- 3:34Choosing the right mean
- 5:23Range and deviation matter
- 7:12Sample size and aggregation
- 8:17Bias in surveys
- 10:07Common types of bias
- 13:59Graphs can deceive
- 16:54Misleading visuals
- 18:16Percentage comparisons
- 19:43Pie charts and correlation
- 21:10Correlation vs causation
- 22:57Confounding variables
- 24:00Data dredging and p-hacking
- 26:29Political misrepresentation
- 28:22Question framing matters
- 29:05Unconscious bias
- 30:13Numbers and pretty pictures
- 30:56Q&A and closing
related talks
27mThe Subtle Art of Lying with Statistics
43mThe Subtle Science of Misleading with Statistics
40mHow to Lie With Stats and Charts
32mIntroducing Reality into Data Metrics
38mKnow your data: The stats behind your alerts
37mHow to Start Critical Thinking in Data Science
05mPerception and bias and metrics, oh my!
32mLearning Statistics Will Save Your Life