So if we ignore selection bias, restraint bias, small sample size, and the general lack of any sort of scientific rigor — I think we may have something here: charts are pretty.

/via dwineman