How to Assemble Science
2 days ago
- Science news often contradicts itself because studies are reported in isolation; real understanding requires weighing all evidence together.
- The FDA's warning on paracetamol during pregnancy was based on cherry-picked studies, later contradicted by a comprehensive review.
- The massive volume of research (19,000+ papers daily) creates confusion, with many unconfirmed results and flawed studies.
- Scientists and journalists focus on new studies, neglecting the synthesis of existing knowledge—a 'scandalous failure' noted by Iain Chalmers.
- Gene Glass invented meta-analysis in 1976 to combine results from multiple studies, leading to systematic reviews.
- The Cochrane Collaboration, founded by Chalmers, produces rigorous systematic reviews that underpin modern medical guidelines.
- Synthesizing evidence earlier could have prevented thousands of SIDS deaths by linking back-sleeping to risk decades earlier.
- Evidence synthesis now extends to education (e.g., metacognition boosts learning), climate policy, and conservation.
- Critics note systematic reviews often ignore qualitative and Indigenous knowledge, but new methods like mixed-methods syntheses address this.
- AI may accelerate evidence synthesis, but risks include poor-quality reviews and reliance on hallucinating tools like ChatGPT.