Who Will Win

Predictions of who is going to win a political campaign are fascinating stuff. Many people project their own preferences onto an outcome. In the literature that's called wishful thinking. The origins of wishful thinking research is out of studies of sport fans who think their team is going to win even when all the evidence suggests otherwise. Then the notion got moved over into politics. A few studies have been done. Here is some my own data.

Respondents were asked who is going to win. I then combined that with who they supported. Looking at presidential elections from 1968 to 2004, you see how much preference guides prediction.


Republicans: Out of 6,960 surveyed, 81 percent predicted their preferred candidate would win.
Democrats: Out of 6,664 surveyed, 64 percent predicted their preferred candidate would win.


The amount of wishful thinking from year to year is fairly stable, but predictive accuracy (getting it right) ranges wildly from election to election. This is important. When elections are close (like 2000), personal preference and wishful thinking gets in the way and people are less accurate in predicting the outcome.


When it's a runaway election, like 1972, then accuracy improves. You'd think media exposure would improve accuracy, if for no other reason than reading or watching the news would expose people to public opinion polls and the viewpoints of others, but it does damn little to dampen the preference -> prediction bias.

Data Gets in the Way of Theory

First, the background. There's research that suggests being on the losing side in an election will cause one to be less trustful of government. Makes sense.

So I extended this idea to something I blogged about a day or so ago -- perception of opinion.

It's simple. In the ANES are these questions about who you think is going to win an election. I figure that people who think their candidate is going to win and end up on the losing side will be even less trustful of government and democracy than those who are on the losing side but guessed their guy was going to lose.

Dammit. Data gets in the way of good theory.

I looked at data over three decades, certain I'd stumbled on this brilliantly publishable idea. To add a dash of mass comm to the recipe, I figured news consumption would help dampen this effect. Media tend to include poll results, so even if you're on the losing side you'd be more likely to be exposed to polls that tell you this and be a little more accepting when you lose -- even if you thought you might win.

Again, dammit. Data. In the way. Good theory.

None of this really worked as expected. It should have. It should have been this groundbreaking piece of research destined to appear in Political Communication or some journal of that stature. Shoulda woulda coulda. Instead I have this great theoretical idea that simply doesn't pan out in the data. I will revisit the idea at some point, or maybe someone will take up the challenge and prove that I was an idiot and missed something obvious, making an academic name for themself. Maybe.

Bugger the data.

Addendum to Below

On my earlier post today I talked (at length, sorry) about perception of opinion and why it matters. In reporting Obama's lead over McCain, a new Fox poll tells us:

A 61 percent majority of voters believes Obama is going to win the election — more than three times as many as believe McCain will (18 percent). A month ago it was evenly divided: 41 percent Obama and 40 percent McCain (Sept. 8-9). This summer, voters were more likely to say Obama would win: 51 percent Obama and 27 percent McCain (July 22-23).

I wish I could remember where I once read that perception of who is going to win is a better predictor of an election outcome than merely seeing what a poll tells you on who people prefer. Let me say that another way. Asking who you are for ends with Obama getting 45 percent of the vote to McCain's 39 percent (according to this very poll). But asking who is going to win gets Obama with 61 percent.

I vaguely recall reading once, as a grad student perhaps, that asking who you think is going to win is a better predictor than just counting up those for or against the candidates. This question, neat as it is, suggests a sense of inevitability among even McCain supporters. That gets right at the power of perception.

Some other day I'll blog about a related matter -- wishful thinking. Again, this is about perception and I've done some work in the area, but basically we tend to wishfully think and see our own preferred candidate as likely to win. Fun stuff, as of course all mass comm research is -- if you're into the whole research thing.

Public Opinion

Public opinion is no more than this,
what people think other people think.

- Alfred Austin, 1891
in Prince Lucifer

I've always been partial to the above definition of public opinion -- and to Victorian-era plays named after the devil, but that's a different story. The play itself sucks. I read it in grad school when I came across this definition. Kept the definition, forgot the play.

Scholars in various fields have nibbled around the idea that the perception of opinion matters. Read spiral of silence, or false consensus, or pluralistic ignorance, or bandwagon effect, or third-person effects and you'll find the roots lie in what people think other people think. The consequences depend on your theoretical approach. In spiral of silence, for example, the perception that one is in a distinct minority on some issue creates a fear of isolation and results in people being a little less likely to speak out on their position. The "spiral" in this case is the slow disappearance of minority viewpoints.

Okay, we've had our dose of Theory for the Day. So what?

Perception matters. If people sense a shift in the opinion climate, some will follow and some will simply grow more quiet, some will jump on the bandwagon and some will root louder for the underdog. It's interesting that as the presidential campaign gets closer to ending (thank God!) and Obama eases away from McCain, we've seen some ugliness on both sides. The perception of opinion going against your way can lead to people staring at a TV screen in disbelief. The poster child for this is the guy at the McCain rally the other day who begged McCain to go negative, that he can't believe America is following Obama. It's fascinating stuff.

I should also mention polls here and the discrediting of polls when they go against our side. "They never asked me or anyone I know" is often heard from people who don't like what the polls tell us about the opinion climate. Since we tend to hang out with people like ourselves, we create these artificial bubbles of like-minded opinion. When polls say otherwise, they just don't make sense. My guess is people for McCain right now are scratching their heads, wondering how this is happening (it's the economy, stupid). Supporters of Kerry and Gore did the same thing in 2004 and 2000.

It's human nature. Our perception of opinion is biased by those we hang out with, those we talk to, those we see in our neighborhoods or at work. When a broader measure of opinion disagrees with that, we're stumped. And that screws with our perceptions of opinion. Some people correct their own estimations of opinion, some argue with the source of the disagreement.

Again, cool stuff.

What Happened in 1066?

Begin Gushing

The best book-length treatment of political knowledge is probably What Americans Know about Politics and Why It Matters by Delli Carpini & Keeter. The thing is full of information, statistics, and analysis of why political knowledge matters in a democracy. A brilliant book.

End Gushing

So I'm looking at a 2-page table (Table 2.7, for those who care) and it has all these questions and the percent who answered them correctly over many many years, over many many surveys. They have two categories on this table: geography and social & political history. The highest percent correct for geography was locating Texas on a map (91 percent). The lowest? Where is Belize (4 percent). On social and political history, the greatest accuracy was on identifying the first president of the U.S. (93 percent), and the lowest asked what happened in 1066 (10 percent). Oh, the answer to the blog title -- Battle of Hastings.

I think it's kinda neat that at least 10 percent got the last one right. Honestly, it could have been a lot lower.

For those of you fascinated by political knowledge and what people know, I cannot recommend this book high enough. It ranges from what we mean by political knowledge to why it matters. I'll blog about it some more between debates, polls, and other presidential stuff. Hell, you could blog the appendices for a week and not run out of material.

Debates and Drinking Games

A CNN poll has Barack Obama winning last night's debate. Fifty-four percent said Obama won, 33 percent went with John McCain. If my math is right, 13 percent called it a tie. Yada yada yada.

Let's get to what really matters -- drinking games.

For the VP debate it was "maverick" as the signal to down or chug or shoot or whatever the hell it is people do, depending on what they're drinking. Last night it hit me what the right word/phrase would have been -- "my friends."

By the fifth or so time McCain used it, I realized that was the key phrase of the night. Not as good as a wink, but still it came often enough that you'd get a serious buzz if you followed through with the downing or chugging or shooting.

What did people learn, though, from last night's festivities? Obama is smooth, McCain can have his moments but seems outclassed, or out-presidentialed. They disagree on some key fundamental issues. We got to hear the same talking points again, in case you missed them in the previous debate. And we learned that the town hall format can surprise you with the quality of the questions and how well the candidates respond.

Cool Site

Been meaning to put this up ever since Mark Johnson pointed it out to me. Had a minute, so here it is, a graphical display of the amount of coverage received by McCain and Obama. Very cool.

Be interesting to put this next to a graphical display of opinion polls to see if there is a correlation. If all publicity is good publicity, then this should reflect that. Sometimes, though, you'd rather not have some of the coverage you get. This is a blunt instrument to be sure, but it's still neat.