When Statistics Strike Back

The 2012 US Presidential Election is almost upon us, and just like last time I have been completely engrossed in it. I also admit to compulsively checking FiveThirtyEight to give me a daily (or hourly?) dose of sanity. Nate Silver’s blog has truly become a sensation this election season. The premise is really simple: take national and state polls, adjust for various statistical anomalies, factor in economic fundamentals, present all this in a clear and rational way, and voila.

The result so far? Since June when the campaigning went into full swing, the model has not given Mr Obama less than a 59% chance of retaining the presidency. As Mr Silver has patiently explained, this race is a close one, with Mr Obama having a slight edge over Mitt Romney. Despite the Democratic Convention boost in ratings, and despite the Denver first debate drop in ratings, the general theme has prevailed. Now as the race enter into the final stages, the model has ticked towards Mr Obama, since Mr Romney does not have much opportunity left to turn things around. The latest figures give Mr Obama a 83.7% chance.

It all sounds so simple and reassuring (if you’re an Obama supporter). And yet, there’s one group of people who simply cannot take the words of reason that are coming out of Mr Silver’s blog. Actually, make that two. First are those people cocooned in the conservative alternative reality. These people believe that Mr Obama is a Muslim Kenyan socialist seeking to sabotage the country. Defying logic and evidence, they predict that Mr Romney will either win comfortably or win by a landslide. They don’t like what Mr Silver is saying, not one bit. And so the specious reasoning and ad hominem attacks have been flying thick and fast in the last few weeks.

The other group is more fascinating. These are the political pundits, the media commentariats, the people who are paid to talk and to write and to get attention and ratings and internet traffic. Would it surprise you that it is in their interests to play up the drama as much as possible? To insist that the race just has to be a toss-up, a dead heat, a coin flip, neck-and-neck. Obama had the advantage, but look! Romney now has the momentum. It is therefore not surprising at all that a blog using statistics, incorporating facts and applying reason, showing a race that has remained steady for the most part with a small advantage to Obama, would get under the skin of the punditry.

I am loving it actually. It is so good to see someone who knows what he is talking about, has the facts to back it up, and has now become popular in mainstream media and the public consciousness. Could this be a turning point in the coverage of US politics? Pundits thrive on gut instinct, ‘feelings’ and emotions. As a student of psychology, I have learned that humans are especially prone to think and act in those ways, rather than a more effortful — and also more accurate — way of doing things. This is just a few examples of what political pundits are guilty of, now and in the past: narrative formation, rationalisations, cherry picking, counting the hits and forgetting the misses, and maths- and statistics-illiteracy. I’m not going to be completely cynical and say that they are only doing this for the  attention (ie, money). Rather, it is a very human shortcoming that takes deliberate effort to overcome.

For a particularly egregious example of media dramatisation, here is Chris Cillizza of the Washington Post explaining why he is calling the crucial state of Ohio a ‘tossup’ rather than ‘lean Democrat’:

After reviewing all of the available public polling data as well as talking to operatives in both parties about the private polls they are privy to, we are convinced that Ohio is a 1-3 point race in President Obama’s favor at the moment.

That — coupled with the state’s electoral history and the absolute necessity for Romney to win the state if he wants to be president — leads us to move it back to the “tossup” category.

Can you see what’s happening here? Yeah, yeah, I can see that all the polls indicate a 1-3 point advantage for Mr Obama. But… state electoral history! Ohio has always been a swing state that’s balanced on a knife edge. So it’s going to be a close one. And, Mr Romney simply has to win the state! Without it, his campaign is doomed. His overwhelming need to win Ohio makes me feel that it is a ‘tossup’ now. I am a political hack.

Read the comments at the bottom of the article for some brutal (and satisfying) takedowns.

I will end on the last (most recent?) words of Nate Silver, who puts it all into perspective and sticks his finger into the eyes of political pundits everywhere:

My argument, rather, is this: we’ve about reached the point where if Mr. Romney wins, it can only be because the polls have been biased against him. Almost all of the chance that Mr. Romney has in the FiveThirtyEight forecast, about 16 percent to win the Electoral College, reflects this possibility.

Yes, of course: most of the arguments that the polls are necessarily biased against Mr. Romney reflect little more than wishful thinking.

Nevertheless, these arguments are potentially more intellectually coherent than the ones that propose that the leader in the race is “too close to call.” It isn’t. If the state polls are right, then Mr. Obama will win the Electoral College. If you can’t acknowledge that after a day when Mr. Obama leads 19 out of 20 swing-state polls, then you should abandon the pretense that your goal is to inform rather than entertain the public. [Emphasis added].

Amen, Nate. Amen.

Advertisements