(1) Statistical significance is not a measure of importance. It is a general measure of how well the sample population represents the true population.

(2) The standard 0.05 significance level is just a convention. There is no mathematical basis for it.

(3) Since statistical significance is a function of sample depth (see #1 above), determinations based on the period 1998-2012 or whatever are just as likely Bullsht as not.

(4) Who cares? The only temperature signal that matters in terms of AGW is the 20-30 year average. That is the reason the trend for the last 18 years is not statistically significantly different from the trend for the preceding 18-year period. It is the reason that both the raw and adjusted temperature data show the warming trend (see graph below). The data adjustments improve relative accuracy and make for more reliable calculations in models and reconstructions, but they do not change the overall end result because of the robustness of the empirical data (we know this because of the Law(s) of Large Numbers and the Central Limit Theorem).

The fact that the authors have made it clear that the significance level is 0.10 shows that they have reservations about the accuracy of their hypothesis. If you’re able to access the supplementary notes that accompany the full article the SL is explained in detail.

In science something with an SL of 0.10 would never be regarded as being proven, for that you need a SL of 0.00006 (5 σ or 99.999994% certainty).

All the paper does is to put forward a possible explanation for the warming pause. In the climate science community it has been noted as a point of interest but little credence has been given to it; further research is needed together with a comprehensive revocation of other explanations for the pause.

.10 WHAT? is that 1/10, 0.1 percent or what?

FYI: a 0.1 percent chance of an error is acceptable in this type of research. 10 percent is okay as a preliminary probability.

I believe that Karl et al actually calculates the figure. The IPCC just guesses one. So this is quite a step forward.

I think that must make it acceptable.

Generally in statistics, we use 0.05 as the break point. However, 0,10 gives you an error rate of only one in ten. It is certainly, "indicative" and suggests the need for refinement.

I'm afraid the 'pause' is alive and well. Rumors of it's demise have been greatly exaggerated.

https://wattsupwiththat.files.wordpress.com/2015/0...

about as acceptable as the non significance of the so called pause

trevor is right. maxx is wrong