Frank Bosse is a guy who gets around. Retired Senior Scientist of Molecular Neurobiology at Heinrich-Heine Unversity in Duesseldorf, he’s also a man who knows his statistics.
Now, anybody with a bit of stats background might look at the graph above, note the r-statistic along with the limited data stet and scatter, and have more than a few, uhm, err, ahh, “concerns” regarding conclusions.
Any conclusions beyond saying “seriously?”
Frank took issue with this graph and its related paper in Nature by Li & Chakraborty (L&C 2020, hereinafter). Among other things, L&C 2000 reports a statistically significant increase of the decay time when a North Atlantic hurricane makes a landfall, all due to warmer SST in a warming environment.
Small datasets are bad enough. Data pre-conditioning by L&C is borderline criminal. Or fraudulent. Do you prefer “incompetent?”
Certainly, “fake news,” as they say.
Here is what Frank comes up with when he backs out time series data smoothing.
Which, of course, has a trivial slope with a high error term. And some wild behavior in the tails.
Frank has some great lines regarding the sophomoric (literally) stats analysis:
“The peer review process of “Nature” for L&C 2020 lasted more than 8 months, it makes wonder if there was no reviewer with some fundamental skills in statistics involved.
“However, this must be the case unfortunately: In the ‘methods-statistical significance’ section the authors mention a test for autocorrelation and there is written: ‘(which we test using the Dublin–Watson test)”’ This must be a typo, the name of the test is ‘Durbin– Watson‘.
“One should hope that the peer review process of ‘Nature’ would be improved soon to avoid overconfident, obviously flawed papers like L&C 2020.”
Don’t hold you breath.