Hypothesis, then data, then inference, THEN conclusion

George Will, Stanley Fish, and Terrence Jeffrey have been showing their quantitative asses with some weird rap about President Obama referring to himself (I, me, my, myself, etc.) excessively.   Mark Liberman over at Language Log has blown the whistle on this nonsense, twice, showing it’s unfounded in the data.  Liberman, whose only obvious fault is having a name I can’t spell, suspects this is conservative confirmation bias.  I suspect he’s right, but we’ll need a bit more data to tell.