Hacker News Reader

No, really, it's not.

There are very similar problems in most of the social AND natural sciences. In 2012, Amgen reported that they could only replicate 6/53 "landmark" cancer studies. Bayer had a similarly hard time replicating papers suggesting new drug targets in a smattering of fields including cardiology and oncology (14/67 successfully replicated).

View on HN

You are both right. The softer the science, the worse the effect. I trust physics more than biology, and I trust biology more than sociology. Well, that implies I trust sociology... I really don't. Anything softer than physiological psychology is basically in the 'don't trust this until you have verify it, and then still mostly don't trust it'.

Absolutely Matt. It is also shocking that most people believe that "peer reviewed" is synonymous with "reproduced study". It is not. One is editorial the other requires longer and heavier work.

Reply on news.ycombinator.com