Pa Annoyed has what I think is a very important explanation of the problems with doing statistics on the kinds of signals found in climate science. It’s important because it makes available to the layman a level of understanding about the nature of the science. And understanding helps us come up with good arguments.
The article starts off about an email from dendrochronologist Ed Cook’s in which he complains that he has to review a paper that contains too much maths. Pa Annoyed points out that it’s exactly the kind of maths he should be intimately familiar with.
Pa Annoyed then proceeds to explain the kind of maths we are talking about. It is about how to tell the difference between signal and noise. The problem is there are different kinds of forms that the signal and the noise can take, and climatologists seem very keen to assume that the signal is a straight line and the wobbles up and down are noise. The straight line goes steadily up and is global warming; the noise wobbles up and down and is weather.
Pa Annoyed explains that there is no physical reason for that to be the signal to be a straight line, and every reason for it to be what the graphs actually look like: a random walk. In other words, there can be other hypotheses that fit the data.
Read the whole thing. Really. Read it now. You’ll learn something.