Aaron Scfiff has a great data problem on his blog at 26econ.com. Does talking about a recession in blogs correlate with the probablility of a recession given in a prediction market? (see his blog for the caveats in the data).
The relationship is pretty intruguing. But I wonder whether he has captured the entire story. What about that elusive third factor that drives both? What about the amount of time 'recession' occurs in the old media, how many times by politicians. There appear to be two reversals around 9/14/07 and 10/26/07 -- are these correlated with major events or stories? I had just blogged about a great data problem for student econometricians and I think this one is even better. Again this one could be loosed on students for a PBL exercise and I just may.Aaron goes on to posit a relationship and I wonder if the relationship has trend or drift. Facinating stuff. Thanks Aaron.