3 ms·
Reproducibility: Expect Less of the Scientific Paper
- kkoncevicius 5y agoI think this article has one key insight: that generating new ideas and rigorously testing them are orthogonal activities. More than that - if people who generate ideas are also responsible for selling them they will likely overstate their evidence and hide inconvenient facts. But after this excellent point the rest of the article seemed like incomplete speculation: "Maybe there will be two kinds of scientists, maybe some scientists will have to take less autonomous roles, maybe we need to pick most promising high-impact ideas and assign them for testing". I wish authors spent more time distilling their proposal and making it concrete.
- geoalchimista 5y agoPretty sure few people would be enthusiastic about doing those "testing" jobs. A graduate student spending six years testing others' "high-impact ideas" would have no future on the cutthroat academic job market.