AI Snips
Chapters
Transcript
Episode notes
Data Colada's P-Hacking Wake-Up Call
- Hosts described Data Colada's origins and their influential 2012 p-hacking paper.
- That work showed common reporting practices could manufacture statistical significance.
Sleuthing Revealed Widespread Data Issues
- Data Colada found statistical and internal inconsistencies suggesting data tampering in high-profile papers.
- Harvard's 1200-page investigation corroborated those concerns and moved toward retractions.
No Formal Police For Research Fraud
- There is no formal policing body for research fraud, so peers must volunteer to audit data.
- That creates reliance on a few motivated sleuths and uneven oversight across the field.