’science’s dirtiest secret: The “scientific method” of testing hypotheses by statistical analysis stands on a flimsy foundation.’ Anthony Watts starts learning about statistics! Will he be shocked to learn that failing the 95% confidence interval doesn’t mean that something is false? (See the whole “no warming since 1995” tomfoolery.)

[Note that an excellent analysis of this can also be found at Open Mind. Tamino makes the point that the author’s main concern is with medical studies and with understanding the implications of the null hypothesis.]

P value = probability of an observed result arising only from chance.

Anthony has copy-and-pasted science journalist Tom Siegfried’s ScienceNews article “Odds Are, It’s Wrong“. This rather loose discussion of the flaws of statistical analysis serves as a jumping-off point for more of Anthony’s misplaced accusations. Short version: poor statistical analysis can obscure correct scientific interpretations. Here’s the “dirty secret” part:

Even when performed correctly, statistical tests are widely misunderstood and frequently misinterpreted.

Ouch! I think he’s talking to you Anthony! Here’s another good quote from the article (italics mine), especially in light of Anthony endless insinuations about the “Mann’s tree ring proxy hockey stick” (he can’t even get six words in a row right):

Replicating a result helps establish its validity more securely, but the common tactic of combining numerous studies into one analysis, while sound in principle, is seldom conducted properly in practice.

Guess what Mann’s climate analysis was? A combination of numerous studies. Not just tree ring proxies. This is why is stands up so well, and why the denialists are so desperate to tear it down.