Saturday, October 11, 2008

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.


John Arkwright said...

Don't bugger the data, Dr. H. Torture it! Waterboard it! It will talk. It always does.

Hollander said...

I have to try waterboarding the data, but I'm not sure it's going to break. Seems awfully stubborn and unwilling to cooperate.