Non replicable research

2012 April 6
by Daniel Lakeland

An article in Nature (paywalled, or read a Reuter's summary) suggests that many scientific "discoveries" are nothing of the kind, being the product of most likely too little skepticism on the part of the researchers that did the work. It's easy to fool yourself into thinking you've made a discovery if you don't take into account the inevitably large amount of filtering that is done during the course of discovery.

Getting a result with a small p value should be the first step followed by a rigorous test of alternative causal hypotheses and replication in alternative hands. But people who are that careful will never get funded compared to those who throw up a flashy paper claiming a novel and truly groundbreaking discovery.

No comments yet

Leave a Reply

Note: You can use basic XHTML in your comments. Your email address will never be published.

Subscribe to this comment feed via RSS