A 2013 oct 19 ECONOMIST article reviews problems with scientific research - i've summarized and reordered them:
1. most scientists
are not statisticians
2. statistical mistakes are common
3. pressure to publish incomplete results is strong
4. a lot of research is poorly designed and executed
5. many models are ‘overfitted’
6. bias to publish positive results leads to bias in overall knowledge
7. peer review is unreliable
8. amounts of data are often overwhelming
9. intellectual property rights restrict the use of code or methods
10. replication is often difficult and unrewarding
11. fraud is more common than realized
2. statistical mistakes are common
3. pressure to publish incomplete results is strong
4. a lot of research is poorly designed and executed
5. many models are ‘overfitted’
6. bias to publish positive results leads to bias in overall knowledge
7. peer review is unreliable
8. amounts of data are often overwhelming
9. intellectual property rights restrict the use of code or methods
10. replication is often difficult and unrewarding
11. fraud is more common than realized
