Bad correlation coefficients still

Brad has us take out the odd ball points in these measurements, then look at the trend when they are removed.  The measurement error still sucks.  I haven't looked at income inequality, maybe I should spend time in that debate?