We are 99.99999% sure that broken models produce stupid climate statistics

Amazing what they can discover with data from just the last 130 years. Hey but it must be right. It’s bootstrapped!

Finally, the study you’ve been waiting for. Now we can be absolutely certain — it’s practically proven beyond all doubt — your SUV changes the climate.

We just need to assume the climate models understand the climate and that there are no longer natural cycles at work AND that there are no effects from the sun from the solar wind, solar magnetic fields, or spectral changes. Easy. (Don’t look at the evidence, the pause or all the model failures.)

The new headline:

99.999% certainty humans are driving global warming: new study

It’s at The Conversation – -the government funded site where government funded scientists discuss their bestest ideas.

WARNING – Modern global temperatures (red) were homogenized with imaginary Vostok data for entertainment purposes only. See footnote*

Obviously being 95% certain is not enough. Desperate believers are upping the ante. I guess all those people who were not convinced by 95% certainty will now switch over, blown away by the last 4.999% certainty that was missing before. It’s the third decimal place that does it.

PS: There’s […]

Abusing statistics in the name of global warming

UPDATED: Lovejoy has responded (in PDF and in comments. See below.)

I could tell from the headline below this was going to be a candidate for the Top-Ten most vacuous papers. It lived up to expectations, and then some.

“Odds that global warming is due to natural factors: Slim to none”

Is there anyone with the lights on at McGill University or “Climate Dynamics“? Surely ScienceDaily ought to have laughed at the press release and sent it back?

Seriously, people wield the magic wand of “statistical significance” without realizing that a/it isn’t magic, and b/ tiny p values can still mean nothing. (Depends on the hypothesis and assumptions underneath, hmm?). LoveJoy looked at 4500 years of a very squiggly line (the last 5% of this graph) and pronounced his magic tool could tell whether the last wiggle was …. ahem, unnatural. If that looks like tea-leaf reading to you, join the club.

Modern climate science can predict virtually none of the spikes and wiggles on this graph. Note the graph doesn’t include the last 100 years which adds about 1C to the rise.

Don’t look now, but it’s another Nail In the Deniers Coffin.



The future of climate alarmism is bogus statistics

Dr David Evans and Joanne Nova

The temptation is all too strong. How many bureaucrats would work just as hard to show that their department was less important, less necessary, and less deserving of funding? It’s the fatal trap of socialist management. The incentives are wrong.

When governments are faced with poor reports, but they write their own report cards, they have many options to upgrade their “score”. It’s insane to think that people might not take every opportunity they can to improve their mark. They are human.

Big problems like inflation, unemployment, national growth, or global temperatures can be “improved” two ways –one way takes tough decisions and years of work, and the other way takes a quiet statistical summit, a white paper and an in-house training weekend. It’s easier to “solve” big problems by changing the way you measure them. By changing definitions, methods of interpreting the data, or through sheer statistical chicanery it’s possible to issue press releases with the words “improvement”, “better than expected” or at least “figures have plateaued”.

10 out of 10 based on 4 ratings […]