Whilst looking, I hit this great site of some talks on social policy and the psychology of happiness. The difficulty of separating measurements of happiness from the impact of income level is discussed at length. Also, the observation that happy people may not be the best for some jobs, like power plant safety monitors. They're more likely to be feeling positive and not expecting the worst: "It was alright yesterday, why should it be any worse today?"
Look for the link at the bottom to the panel talk transcript: Informing Policy Choices Using the Economics of Happiness