Abandon significance test



You may be surprised that people like me support the idea of abandoning the test of significance. However, I do support the idea/motion. Although null hypothesis significance testing (NHST) has been a ‘standard’ practice in scientific research for approximately 100 years, it has done damage to science. And, it is a high time to retire the practice.


A group of experimental and epidemiologic scientists led by Valentin Amrhein (Zoological Institute, University of Basel, Basel, Switzerland), Sander Greenland

(University of California, Los Angeles), and Blakeley McShane (Kellogg School of Management, Northwestern University) have drafted a commentary for Nature to call for the abandonment of NHST. In the commentary, among others, they argue that:


(a) researchers should not treat their result as black-and-white distinction (eg “significant” vs “non-significant”) based on some arbitrary P-value cut-off such as 0.05. This is a thoughtless interpretation. We should embrace uncertainty in every finding!


(b) researchers should not categorize continuous data into distinct groups – it is not a good idea at all. Many many studies have shown that such a practice results in the loss of information and reduction of statistical power;


I agree with all the points raised in the commentary.


They are calling for support from scientists worldwide to sign the motion. If you agree with the idea and support the motion, please email to retire.significance2019@gmail.com (until March 7). Alternatively, you can add our name here: https://bit.ly/2tVjv1X.



Featured Posts
Recent Posts
Archive
Search By Tags
Follow Us
  • Facebook Basic Square
  • Twitter Basic Square
  • Google+ Basic Square

Opinions expressed are my own and not the views of my employers

© 2018 Tuan V Nguyen

  • RG
  • Twitter Social Icon
  • LinkedIn Social Icon