Sunday, February 19, 2006

OT: Electoral Sabermetrics

I grew up as a big fan of Bill James and his analysis of baseball; it's now pretty much mainstream, but "sabermetrics" -- the systematic analysis of baseball statistics -- was absolutely revolutionary at the time. There was this idea that baseball analysis could only be done by insiders, people who really knew how the game worked. James came along and proved that an "outsider" could look at the numbers, analyze them creatively, and come up with analysis that was more accurate than the insiders', not less.

Since political blogging took off, I've been waiting for something like that to spring up, a political equivalent to sabermetrics -- a way for outsiders to look at political polling numbers and come to a better understanding of politics than the "insiders," with their obsession with marginal stuff (does anyone really, truly vote on most of the issues that politicians and consultants are obsessed with?). One thing I've found recently may be the closest thing to a Bill James approach to politics that I've seen. It's in this post as well as this earlier post from the blogger Chris Bowers at the group blog (one of the better political blogs out there).

Bowers was looking at old poll data, and he came across data that showed him two interesting things about U.S. Congressional elections. One, if a sizeable majority of respondents thinks a particular party controls Congress, that party will lose seats in the Congressional elections. And two, there are some years when the vast majority of respondents are flatly wrong about who controls Congress.

Specifically, he discovered that there were three years in which most people identified the wrong party as controlling the U.S. House of Representatives, 1982, 1986, and 2002. In 1982, Democrats controlled the House, yet "68% of the electorate believed that Republicans controlled the House of Representatives" -- and that November, the Republicans lost a bunch of seats in the House. In 2002, the Republicans controlled the House, yet "72% of the electorate believed that Democrats controlled the House of Representatives" -- and that November, the Democrats lost seats in the House.

Bowers realized that what these three "wrong identification" years had in common was that they were years when different parties controlled the House and Senate. In 1982, the Republicans had the Senate but not the House -- but people thought they had the House too, and so they lost seats in the House. In 2002, the Democrats had the Senate but not the House -- but people thought they had the House too, and so they lost seats in the House.

Bowers goes into more about this in his posts, but what I find fascinating is that with a little examination of data, he's found a better predictor of the results of U.S. Congressional elections than anything I've seen before, particularly mid-term Congressional elections. Simply put, whoever controls the Senate will lose seats in the House in the mid-terms, because the voters focus their attention on that party, assume that that party controls Congress, and set out to punish that party. If there's a lot of consensus as to who runs Congress (Bowers goes into more detail about what it takes to create such a consensus), then there will be a swing in seats.

I like stuff like this because, as Bowers himself says, a lot of the big questions about what swings an election -- what is the party's message, what are their policy positions -- may not be all that relevant in the long run. They're relevant in individual races, but if you're talking about the national results of a whole bunch of little elections, you need to look at the bigger trends, and the bigger trend is just that voters tend to want to punish the party that controls Congress. There were all these thousands of articles written analyzing why the Democrats lost seats in 2002; but it all seems to boil down to the fact that voters mistakenly thought that they controlled the House, since that exactly parallels what happened to the Republicans in 1982 and 1986. This makes sense to me, and is true to the way a lot of voters vote.

Therefore, Bowers recommends that the message of the opposition party should simply be to repeat over and over that the other party controls Congress: increase the knowledge of who controls Congress, and you increase your chances of winning. And the best part of it is, this is a winning message that is impossible to contradict, because it's literally true, no ambiguity, no "other side."

If the Democrats make gains in Congress this November (and it doesn't seem far-fetched to think that they will: the split-power factor that worked against them in 2002 doesn't apply this time, and voters will know the Republicans control Congress because of the news reporting on Tom DeLay and so forth) then Bowers' analysis will really seem valid and useful for future horseraces. Just like Bill James's sabermetric tricks are more useful than baseball's conventional wisdom, these political sabermetric tricks seem to describe the real world of voter patterns better than any political pundit.

Update: I see some people are linking here from political blogs (like the good old Daou Report. For anyone linking here from a political blog, I should note that this is not a political blog (hence the "OT," for "off topic"), and is mostly about film and TV and cartooons and stuff.

Admittedly I've gotten more politicized and anti-war in the last few years, like most previously non-political people have. The reason for this can probably be summed up as follows:

TIME: September 12, 2001

SCENE: A hallway

CHARACTERS: Me and my crazy lefty friend

CRAZY LEFTY FRIEND: Do you support the invasion of Afghanistan?

ME: Of course. It's a rational response. Don't you support it, Crazy Lefty?

CRAZY LEFTY FRIEND: No. If the Bush Administration invades Afghanistan, the next thing you know they'll use 9/11 as an excuse to invade Iraq for no reason.

ME: Don't be silly. Iraq has nothing to do with this. The Bush Administration would never be so stupid as to invade Iraq.

CRAZY LEFTY FRIEND: I'm telling you, if you don't oppose the Bush Administration now, they'll exploit 9/11 for political purposes and use it as an excuse for invading Iraq.

ME: Oh, Lefty. You're such a crazy lefty. Only a crazy lefty would say something so crazy -- Lefty.

Flash-Forward to 2003.

ME (chastened): You were right all along, Lefty. They did invade Iraq for no reason. Where can I sign up for the "These People Are Completely Crazy Book of the Month Club?"

And there you have it. However, I comfort myself with the notion that the world is no longer quite as batshit crazy as it was in 2002-4, and otherwise grit my teeth. And so I don't do political posts very often, which is just as well.

Instead, go to the main page if you like posts about King of the Hill and obscure Otto Preminger movies.

No comments: