I have been following the math in this election a bit. It’s interesting. You have guys like Sam Wang – and to a slightly more compromised extent, Nate Silver – who are math nerds. Their algorithmic approach to polling creates a more neutral – I want to say “pure” but won’t – vision of what is happening on the campaign trail. Ideally, they are creating real-time snapshots of what is actually happening, as opposed to partisan interpretations of what is happening.
Those partisan interpretations are problematic in the sense that they are dishonest – or at least they are willing to be dishonest. Their objective is to drive the narrative in order to move the numbers. If your candidate is winning, you play that up in order to buttress their lead or even expand on it. If they are not, then you have to create a narrative – a story – in which they actually are, or are just about to because they have the momentum, and so forth.
Political journalism and punditry at all levels tends to treat campaigns as narratives that can be massaged and nurtured into preferred tales with defined endings. But there are consequences to that approach, not the least of which is it tends to deprive people of meaningful information that can help them formulate who to support and how.
Something has to give.
Math nerds are really my first experience with an approach to following politics that separate the math from the narrative. But it raises an interesting question. What is the proper relationship between narrative and math in politics?
Take, for example, Sam Wang’s Princeton Election Consortium. That site’s math – I am being inelegant here – gives President Obama a well over 85% chance of winning the election with only ten days to go. Compare that to Dan McLaughlin’s analysis over at Redstate, a well-known conservative site. McLaughlin – also relying on math – concludes that “Obama is toast.”
Both men are using math. How can we reconcile the difference in their conclusions?
One way to do it is to consider the role of narrative. Wang is not an obviously partisan player. While his site attracts a lot of Obama supporters, there are a fair number of Republicans as well. And the comment streams often feature some fairly spirited debate between people who think Wang is out to lunch and Romney is cruising to an outsized victory and those who are comfortable that Obama is going to win an election that he has led narrowly (in terms of popular vote) but consistently for months.
Redstate, on the other hand, is less about dialogue from a conservative standpoint than simply generating unqualified support for the Republican candidate. Commentators are often banned for “concern trolling” – expressing even the hint of an opinion that all is not proceeding apace for that outstanding Romney victory.j
Thus, one way to think about Redstate’s front page writers like McLaughlin is to realize that they are carefully generating pieces that push the favored narrative. In the immediate case, so far as I can tell, McLaughlin is cherry-picking polls – focusing almost exclusively on Rasmussen and Gallup – in order to say without blushing that “Obama is toast.”
His math is probably fine in the calculations, but one questions the premise. If there are dozens of polls out there and you only choose the ones whose conclusion best suits the narrative you want, then you are not really being honest. You are privileging story over math.
Of course, I can appreciat the impulse. Elections are essentially unfolding narratives – I suspect more voters are in tune with politics at that level than at the math level. Ideally you can point to sites like PEC or even 538 and use them to ground a convincing narrative about how voters are responding to this or that aspect of your candidate/campaign and so forth.
But if the aggregate numbers don’t work, then what happens?
On the one hand, you can be honest and say the math isn’t working so well for your candidate. But that’s hardly inspiring. If you’re in it to win, you can’t really dwell on how bad you’re doing. You’ve got to change the narrative.
And that happens in one of two ways. You take McLaughlin’s approach and restrict yourself to numbers that create the desired result. Then you tell a story based on those numbers that gets people fired up and excited.
Or you can simply ignore the math altogether, tell the story you want to tell, and hope nobody catches on.
If we are being coherent, then the narrative we construct is going to be consistent with the facts – whether they are mathematical or scientific or what have you. Indeed, “construct” may be too strong a word. We seek a narrative that naturally and honestly arises from what I am going to call the factual ground.
In politics, which is so often a bloodthirsty and zero sum game, that’s generally not acceptable. In fact, we take it as truth that politicians – and their stories – are not truthful. If a suitable narrative “arises,” then great. If not, who cares? We can always invent one – or switch the polls (or climate warming reports) or what have you around.
Stories are malleable; the truth is not. A coherent political system would recognize this and act accordingly. A coherent citizenry would demand nothing less.
We have a lot of work to do!