What’s Wrong With Betting?

The forecasting model of Nate Silver, perhaps the best known pollster of the 2012 election cycle, predicts an Obama victory, and conservatives have been predictably upset by this fact. For example:

On MSNBC’s “Morning Joe” today, Joe Scarborough took a more direct shot, effectively calling Silver an ideologue and “a joke.”

Scarborough rejected the validity of Silver’s model. Silver’s response to Scarborough was, essentially, put your money where your mouth is. He challenged Scarborough to a bet whereby the loser would have to donate $1000 to charity, if their predicted candidate did not win the election. If Obama won, Scarborough would pay, and Silver would pay in the event of a Romney victory.

Alex Tabarrok defends the bet:

Overall, I am for betting because I am against bullshit. Bullshit is polluting our discourse and drowning the facts. A bet costs the bullshitter more than the non-bullshitter so the willingness to bet signals honest belief. A bet is a tax on bullshit; and it is a just tax, tribute paid by the bullshitters to those with genuine knowledge.

However, the Public Editor of the New York Times, where Silver works, criticized Silver. Her statement is bewildering. Challenging people to a bet is “a bad idea – giving ammunition to the critics who want to paint Mr. Silver as a partisan who is trying to sway the outcome.” Why? She doesn’t elaborate.

Perhaps the intuition is that calling people on bullshit is generally rude? And Silver’s rudeness to a Republican partisan demonstrated that he was a Democratic partisan? Perhaps calling out a friend or acquaintance on bullshit is rude, but the same consideration should not apply between pundits, particularly when one pundit has essentially called the other a partisan hack. A bet would be a strong signal that Silver wasn’t trying to sway the election, that he actually had independent faith in his model’s predictive strength. That’s an entirely appropriate signal for a pundit to give, particularly when his objectivity has been called into question.

Caplan, Online Education, Signalling

Caplan defends hedging his prediction of online education’s dismal prospects:

My recent post on online education specifies:

When I talk about “online education,” I don’t just mean students at existing brick-and-mortar colleges taking some classes from their dorm rooms.  I mean students enrolling in virtual colleges instead of physical colleges.

Tyler objects:

I would say he is defining away the most likely model, namely a hybrid model which has a significant on-line component.

I think my definition is much closer to standard usage than Tyler’s.  In any case, though, there’s a simple rationale for my usage: If online education in my sense takes off, education will become drastically cheaper, and most existing schools will crumble into bankruptcy.  In contrast, if online education in Tyler’s hybrid sense takes off, education will be at most marginally cheaper, and most existing schools will stay in business.

As Caplan kind of admits here, I think his previous posts on the subject substitute the question of whether the signalling model of education is correct for the more difficult question of whether online education will succeed.

The analogy that I keep coming back to is Latin. Once upon a time, arcane subjects like Latin and religion were considered rather important parts of the curriculum. Plugged into Caplan’s previous signaling story, students who stopped taking these classes would fail to participate in the signalling and suffer accordingly.

The problem with this narrative is that it provides no reason to think that “Latin and religion” were actually part of the signal, rather than, say, going to the right university. More likely, the university is a large part of the signal itself, because getting in to a particular university can be quite hard. Once inside, the effort required to maintain good grades completes the signal. The subject matter taken hardly matters, so the grades sort students according to a relevant measure of their aptitude.

Likewise, there doesn’t seem to be any important reason to think that the physical or online nature of an institution should play a large role in the success of educational signalling. A less “pure” signalling model of education might provide reasons to doubt the prospects of online education, but signalling alone only seems to demonstrate that a valuable online education will still be expensive.