On August 12, CNN's Anderson Cooper program had a brief report on the Garthwaite and Moore paper on Oprah's support for Obama. The transcript:

JASON CARROLL, CNN CORRESPONDENT (voice-over): She endorses a book, it becomes a best seller. She mentions one of her favorite products, stores soon sell out. Oprah, it seems, has the Midas touch. But does it apply to voters? After Oprah endorsed Sen. Barack Obama, two grad students decided to find out.

TIM MOORE, ECONOMIST, UNIVERSITY OF MARYLAND: Statistically we found an effect that it`s very unlikely to be random.

CARROLL: Tim Moore and Craig Garthway (ph) of the University of Maryland used a formula that looked at county by county subscriptions to Oprah`s magazine "O" and Oprah`s book club. They compared the data with votes cast for Obama. A lot of equations later, their math showed Oprah was responsible for 1,015,559 votes for Obama, though they also admit it could be much more or much less.

BETH FRERKING, POLITICO.COM: We`ve never had an academic study that looked at the effect and the impact of celebrity endorsement.

CARROLL: Others are less impressed.

BILL SCHNEIDER, CNN POLITICAL ANALYST: 1,015,559 votes she brought to Barack Obama. Well, that`s a little absurd. I mean, how can you get specific to that degree?

What's really absurd is that Scheider has a Ph.D. from Harvard but he doesn't seem to know that the 1,015,559 is a point estimate and that the authors aren't claiming that to be the precise effect Oprah had on the election. In fact, the cutting room floor probably has a snip of one of the authors telling CNN something like "based on our estimates there's a 95% chance that the true effect of Oprah lies between 750,000 and 1,300,000 million additional votes for Obama."