> Square .14 and you get 0.0196. So age explains 2% of income. Not very
> much.
Well, explaining a little bit of such a vast variation as that of U.S. personal income is nothing I'd sneer at.
But the main reason why a measure of linear association (the Pearson correlation coefficient) comes up so small is the type of distribution of these variables. Take logs and the Pearson of the logs is .25 or so.
Actually, I have the PUMS right here. Let me run a decent regression and I'll report back to you.