Sunday, November 6, 2011

Not toast

Has Nate Silver jumped the shark? I hesitate to criticize, because Silver has such a great track record when it comes to analyzing polls, but today's New York Times Magazine cover article seems to take Nate into dangerous and uncharted territory.

But before I even get to that, I have a beef with the headline. For an article that concludes with the overall assessment that President Obama has a 50/50 chance of being re-elected, it seems a bit misleading to title it "Is Obama Toast?" It also seems a bit misleading to take the most unfavorable possible scenario (zero economic growth next year, and the strongest possible Republican candidate) and blast those figures on the magazine's cover. This is the New York Times, after all, not the New York Post. So you wouldn't think that giant letters trumpeting a "17% chance of an Obama victory," which turns out to be only one possible scenario (and a pretty unlikely one at that, unless Europe's economy sinks into the toilet), would be appropriate for the cover of the usually-respectable New York Times Magazine. I guess this is par for the course for the media, however, which continually likes to feed into the narrative of portraying Obama as struggling or unpopular, regardless of what the facts might support.

As to the analysis, it seems a bit more dangerous than Nate Silver's usually astute compilations and weighing of poll results. Maybe because this forecast seems to mix in some apples and oranges. It somehow combines two variables--the ideological position of the Republican nominee, and the growth rate of next year's economy--then throws in President Obama's current approval rating (43%), and through some complicated unspecified formula, comes up with a probability rating for Obama winning the popular vote next year under all possible scenarios. The weight given to all these variables seems to be based on an historical analysis, but even Silver admits that there are exceptions to all of his assumptions about the weight each factor should be given. For example, the elder George Bush failed to win re-election even though he had a high approval rating the prior year; Carter lost re-election even though his opponent (Reagan) was the furthest to the right of all possible Republican candidates; Eisenhower and Reagan both won re-election despite weak economic performance in their re-election years. And so on. But if all Silver is saying is that Obama is more likely to be re-elected if the economy keeps growing at a decent rate, and if his opponent is perceived as too far to the right of the mainstream, whereas he is less likely to be re-elected against a moderate opponent during another economic downturn, that seems obvious enough.  How he can put these kinds of percentage figures on the various scenarios, or why the chosen variables are the most relevant ones, is never very well explained, however.

If the only conclusion to be fairly drawn from this article is that President Obama should be viewed as a slight underdog next year, I'm fine with that. I'm still annoyed with the false media narrative that Obama is struggling or unpopular, but I'm ok with portraying the 2012 election as an uphill battle. That will generate interest in the president's re-election. And Obama performs well as an underdog. And Americans like comeback stories.

(photo from Cracked)

1 comment:

  1. I agree, he seems to be looking for attention.

    I had no idea that he has been so accurate in the past:

    For other readers; he correctly predicted 49 of 50 states in the November 2008 presidential election the winner of all 35 Senate races that year.