He has evidence. He just hasn't looked to see if it's right or wrong.
Seriously - that was his excuse for the most recent blunder on using the wrong month's data from a Russian dataset. Data from a summer month was used in a fall month - surprise! it must have been a hot year! - and when a complete amateur pointed out that it couldn't possibly be that two months had identical readings everywhere, Hanson admitted they had never looked at the data at all.
The "hot year" anomaly went away when the right data was used.
_________________________
"If they can't picture me with a knife, forcing them to strip in an alley, I don't want any part of it. It's humiliating." - windsock