# Junkscience.com endorses Climate Audit to win

LAST CHANCE TO VOTE – POLL CLOSES NOV 8th @ 5PM EST

junkscience .com endorses Climate Audit to win – see their main page

Vote for best science weblog here; and if so inclined, click on Climate Audit. Voting is permitted once every 24 hours. Vote early, vote often.

Why Climate Audit? – see the reasons.

Article Rating
Inline Feedbacks
Steven Mosher
November 7, 2007 10:04 am

Call Hannity.

Larry Grimm
November 7, 2007 3:50 pm

Mosher: you cracked me up! Thanks.
I hope CA wins. Thanks to Steve and Anthony. They’ve made it fun to see science in action in real time. Thanks guys! What’s really cool is I can watch the science play out while having a scotch – I couldn’t do that in the lab.

Editor
November 7, 2007 10:25 pm

St. Mac made a provocative observation last June. He didn’t go there. Perhaps wise, ’cause it’s can ‘o worms city.
But *I* am going there.
I am going to have a little tromp on the old Excel. but I’ll be back–unless my head explodes first!

Editor
November 8, 2007 12:00 am

So here goes, Rev.
Writing last June, St. Mac makes the following observation.
“But the oddest pattern is surely the degree to which red and blue states on these maps match their political counterparts. There are a few exceptions – Arizona, Montana, Utah [Red States, NB.–EJ], but it looks to me like voting patterns would be a better proxy for the existence of a 20th century temperature trend (by state) than tree rings. ”
Now, I thought it looked closer to 5, so I calculate on that basis.
What are those odds? Mac just passed lightly over it. He never ran the numbers (or if he did, he gave no indication).
The odds against this are:
One million three hundred thousand to one.
Calculated thus:
=1/(((1)+(48)+(48*47)+(48*47*46)+(48*47*46*45)+(48*47*46*45*44))/2^48)
If this calc. is correct and I am not failing to account for major factors, I can no longer believe in the integrity of the historical data collection from 1980 to present. And I’m not talking microsite violations here, either.
Poor Mac! No wonder he steered clear of this one! Yes, it seems to me as if some massive “informal data adjustment” has been going on here!
What say the stats authorities around here to this? Can 1,300,000 flies be wrong?

Editor
November 8, 2007 12:33 am

I can’t get away from the “Texas Sharpshooter” fallacy (aka the “Cancer Cluster” fallacy): The observation was not predetermined. That that can’t be helped. But there is a direct correlation, and it is not cherry-picked so far as I can see.
The rural states are the ones with the most station creep. Yet they are the ones which tend to show a decline.
And the only hard evidence that surface temps are actually rising are–the measurements, themselves. So our “expectation” of rising temps is largely based on the results in question.
And there HAS been a population shift in the US. But from Blue to Red! Blue states have been hemorrhaging electoral votes for the last three decades. Yet it’s the Blue states showing the increase.
I’m not laying all the blame on the Blue states, either. The trend goes both ways, so I don’t trust anyone involved.
How many full houses in a row do we have to see before we get suspicious? Sounds like Mark Twain’s immortal “Science Vs Luck” to me!
No, it’s not positive proof. But what are the odds that 1,300,000 flies are wrong?

Freddy
November 8, 2007 8:59 am

Is ClimateAudit down at the moment ?

Steven Mosher
November 8, 2007 9:04 am

Other gems of mine you might enjoy, Larry.
Free the code.
Treemometer.
Glad I could make you chuckle. Hope you didnt have a mouth full
of coke. Nose enemas are brutal when the liquid has dissolved C02

Craig W
November 9, 2007 7:40 am

On tuesday I tried to vote, but, the page wouldn’t load, I refreshed and when the page loaded it read that I’d already voted. Anyway, it looks like a statistical tie.

doug
November 9, 2007 8:40 am

=1/(((1)+(48)+(48*47)+(48*47*46)+(48*47*46*45)+(48*47*46*45*44))/2^48)
“If this calc. is correct and I am not failing to account for major factors, I can no longer believe in the integrity of the historical data collection from 1980 to present. And I’m not talking microsite violations here, either.”
I think the calculation is flawed because the states are not broken into red and blue randomly. All you need is a physical condition and political condition to coincide, to lump the NE and NW together, to lump the south and middle together, and you will get a corrrelation which exceeds the expected match of two random sets.
I’m not sure how the fact that the NE is made of many small states affects it, but I imagine it does.

Editor
November 9, 2007 2:20 pm

“I think the calculation is flawed because the states are not broken into red and blue randomly.”
Yes, I think you are right, here.
And the fact that it was not a predetermined measurement and dataset makes the sample less reliable.
Even so, assuming the calculation was off by a factor of a thousand one would STILL get odds of > 1300-to-1.
Would you say the odds flawed by that margin?
Perhaps we can break the country up into larger, more regionaal chunks and recalculate.