
When the Robots Take Over... - Cautionary Questions
Cautionary Tales with Tim Harford
00:00
Beware the Correlation Fallacy
Statistical correlations can be misleading, as exemplified by the false relationship between the number of stalks and babies in various countries, where large populations support both due to sheer space rather than causation. This misuse of correlation is further illustrated by Darrell Huff's testimony on smoking and lung cancer, where he dismissed evidence of causation despite the significant association. Such skepticism towards statistical evidence can lead to denialism, obscuring the importance of understanding context and causality in data interpretation.
Transcript
Play full episode