Some of you may have heard of a pack of paid shills called The Heartland Institute.
These are the people whom Walmart paid to publish articles supporting Walmart's treatment of workers, and whom Philip Morris paid to bring into question any link between second hand smoking and health risks.
They are also being paid by Exxon Mobile, so it is not a surprise that one of their number, Roy Spencer, has just published a paper:
Spencer, R.W.; Braswell, W.D. On the Misdiagnosis of Surface Temperature Feedbacks from Variations in Earth’s Radiant Energy Balance. Remote Sens. 2011, 3, 1603-1613 (link to full PDF)
which he is heralding in press releases as disproving Global Climate Change.
Those of you who remember polynomial graphs from school will know that the more variables you're allowed to change in an equation, the easier it is to make your equation fit closely to a set of data points. In fact, if you have the same number of variables as datapoints, you can always get an exact fit. This is why legitimate scientists, if they use regression analysis upon a bunch of data to come up with a curve that fits it, then carry out a confirmatory analysis using different data, to double check that the fit wasn't just coincidence.
Spencer, apparently, is fond of skipping this step. He specialises in taking the inputs he wants to be the cause of temperature records (for instance: cloud cover), then using lots and lots of free variables to combine these inputs to produce a a result that is something similar to what's been observed.
To quote one reviewer:
”Well, give me more than 30 parameters, and I can fit a trans-dimensional lizard-goat and make rainbow monkeys shoot out its rear end.”
For those who enjoy reading incredulous scientists writing in full spate, I heartily recommend reading any of the following destructions of Roy's work: