This looks like an analysis going the wrong way. Klout probably uses retweet counts as a signal in their score so it's not surprising to see Klout score being positively correlated with retweet counts at all. It does show that "Klout does work" but in a very simplistic way.