Hacker News new | past | comments | ask | show | jobs | submit login

1) is the source reputable? (esp. the journal, magazine, author)

2) is the study well-designed and big enough to detect the signal as being reproducible and significant?

3) did the study ask the question it now supposedly answers?

4) did the study employ methods that discriminate precisely and robustly, like use of random groups and correction for complicating variables? Or is it based mostly on observations? And did it overlook the a priori distributions of variables known to confound the outcome of interest?

If the article/study does not pass all these tests, it's unworthy of trust as an accurate test of its hypothesis or its claims of significance.




Consider applying for YC's Spring batch! Applications are open till Feb 11.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: