But...
If you compare those numbers to the variance in the chart Tom posted, they don't seem to indicate anything out of the ordinary. It smacks a bit of cherry-picking to try to find data that supports something one might want supported, while ignoring data that's less convenient.
