That's not evidence, that's Krugman patting himself on the back. For evidence I would like to see a testable prediction made before the fact.

In this particular case, Krugman's original position was that the stimulus could be useful but was not sufficient. There is enough wiggle room for two elephants there -- if the stimulus failed, Krugman would have pointed at himself saying it was insufficient and if it obviously worked, he would have pointed at himself saying it would be useful.

In general, I find Krugman to be an interesting example of a very smart guy who either became mindkilled or deliberately decided to do propaganda "for the greater good". His columns are full of classic motivated reasoning.

As to Keynesianism, see this keeping in mind Yvain's recent Beware the Man of One Study.