Lies, Damned Lies, and Medical Science
theatlantic.com
theatlantic.com
I'd really like to see:
1. Any study with public funding is published free of charge.
2. A registry of proposed studies underway prior to results publication. Addressing the issue of companies cherry picking the 1 in 5 study that shows X is effective in treating Y.
3. Some form of standardized results reporting: Drug X did/did not perform better than a placebo.
4. Some form of confidence rating based upon the methodology the study was undertaken with.
Seeing as the problems extent beyond medicine, certainly to psychology but also physics, maybe the problem lies with the very foundations of science.