Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

I don’t think this is the right way of looking at the result. Imagine you have cardiovascular disease. If you take vitamin D, the study suggests that you will on average a 9% reduction in risk. There is a ~97% chance that it provides you some improvement, but there is a ~3% chance that your risk actually increases. This seems like an easy decision to me if you are a high risk individual.

Looked at another way, imagine the authors used a 94% confidence interval instead of 95%. Suddenly, the risk reduction from vitamin D becomes “statistically significant”. There is nothing magical about 95%, and use of p values to assess boolean effective/ineffective should have been done away with a long time ago.

Related:

The Earth is Round (p < 0.05) [1994]: https://www.sjsu.edu/faculty/gerstman/misc/Cohen1994.pdf

Sadly, the Earth is Still Round (p < 0.05) [2012]: https://cmapspublic2.ihmc.us/rid=1P501PBWQ-1S6L809-28MX/Zhu_...



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

Search: