People I (Mostly) Admire 163. The Data Sleuth Taking on Shoddy Science
107 snips
Aug 2, 2025 
Uri Simonsohn, a behavioral science professor at Esade Business School, investigates fraudulent academic research. He shares insightful methods for detecting academic fraud, emphasizing the difference between mere 'red flags' and definitive evidence. Simonsohn discusses notorious cases like the misleading study on name similarity in marriages and the implications of statistical manipulations in research. He also navigates the challenges of maintaining integrity in academic settings and highlights the need for strict standards and community support to combat fraud.
AI Snips
Chapters
Transcript
Episode notes
Name Initials and Marriage Flaw
- Uri Simonsohn discovered a key flaw in a study about name initials and marriage.
- He identified that effects were driven by people marrying those with the exact same last name, likely due to remarriage to the same partner.
Selective Reporting Inflates False Positives
- Running many analyses but reporting only the successful ones inflates false positives.
- This practice misleads readers into believing false hypotheses are true.
Stopping Data Collection Is Bias
- Adding more data until desired results appear biases experiments.
- Stopping when winning but continuing when losing inflates false positive rates.