Ottawa Mike
2012-07-03 17:57:54 UTC
From the abstract: "We present computer simulations and a pair of actual experiments that demonstrate how unacceptably easy it is to accumulate (and report) statistically significant evidence for a false hypothesis."
What peaked my interest this week is a report about one of the authors of that study, Uri Simonsohn. He is apparently an expert in this area and has developed a statistical method to detect data fraud. He has used this to expose several papers in psychology where data was used to create false positives to support a hypothesis: http://news.sciencemag.org/scienceinsider/2012/07/fraud-detection-tool-could-shake.html?ref=hp
The interesting part is that Simonsohn intends to publish his statistical method for detecting this type of data misconduct. It may very well apply to all science which uses data and statistics to support scientific hypothesis through published and peer-reviewed studies. From sciencemag article: "If it proves valid, Simonsohn’s technique might find other possible cases of misconduct lurking in the vast body of scientific literature."
How might climate science be affected by this?