Thursday, February 11, 2010

In Which I Try To Play Nate Silver

Meandering over to Wonkette yesterday I ran across another amusing tale of Fox News much vaunted propagandizing balance falling flat on face. It comes in the form of a poll published in the Fox Forum that gives some remarkably non-foxy results. I'll go ahead and reproduce the poll results as of 7:00 this morning (click image to enlarge):

Now, it's certainly amusing to see Fox publish a poll in which a full 75% of respondents characterize the Tea Party movement as a "fruitless mix of racism, conspiracy theories." It was, no doubt not the results that the creator of the poll was hoping for. Most likely, once the poll began yielding unexpected results it got picked up by the liberal blogs who then sent it waves of traffic, further skewing the results in a direction the poll creators were neither expecting nor hoping for. These sorts of things have a way of snowballing once they reach a certain critical mass, and currently there are over 155,000 responses to the poll --by my count more than three times the number of responses for the previous two polls. It bears mentioning, furthermore, that this is an unscientific, self-selecting, internet poll. So the results would have been dubious at best no matter how they came up.

That said, Fox News does have a storied history of FAIL when it comes to charts and basic math. The most notorious example of which I have taken the liberty to reproduce below:

If you successfully completed 1st grade, you've probably realized that the percentages in the above poll add up to 193%.If you've gotten as far as 2nd grade then you know they should add up to 100% But if you were home schooled by parents seeking to shield you from Satanism in the form of the Theory of Evolution (as the Fox News intern who produced the above chart likely was) then you're probably pretty comfortable with the results.

But the tragedy of Fox News' epic stupid doesn't end there. As I mulled over the amusing results of the first poll, I realized that the creators of the poll, in their propagandistic zeal and lack of basic math competence had drafted a poll that actually favored the sort of result they were hoping to avoid. How so, you ask? Let me explain:

One of the surest ways to guarantee a position a strong showing in a multiple-choice poll is to provide several subtly different alternatives for opponents of your position to chose from, while providing only one choice for proponents to pick. So, for instance, if I wanted to prove the superiority of hamburgers over pasta in a population of evenly divided hamburger lovers and pasta lovers, I would produce a poll such as this:

Now, assuming these pasta lovers all love different kinds of pasta, I can expect a poll result like this:

Though poll responses are evenly divided between pasta lovers and hamburger lovers, the way this poll was constructed, and especially the bar graph format gives the false impression that the overwhelming majority of respondents are hamburger lovers. And this is exactly what Fox has done, crafting a poll that gives the president's opponents three options to chose from, but supporters only one (we can pretty much disregard the meaningless 5th "other" option).

So why did editors who produced this poll for Fox News make such a basic mistake? I have a theory that goes beyond their appallingly obvious lack of basic math competence. My suspicion is that the producers of the graph are so blindly driven in their ideological partisanship that they can think of all sorts of reasons to hate President Obama and the Democrats and have a hard time narrowing it down. But when it comes to providing the opposite choice, they find the idea so distasteful that they can bring themselves to provide only one pro-Obama option, and that only to avoid producing a poll so ridiculously partisan in its framing that it makes a complete mockery of Fox's disingenuous "fair and balanced" mantra. You can't defend a poll that doesn't give the opposition at least one voting option.

And what happens then, is that the few liberals who do hang around at Fox can more easily tilt the poll in their favor. Once this starts to happen the unexpected results get linked to by some high-volume liberal blogs and before you know it 75% of respondents think the movement you're trying to promote is made up of racists and crazy conspiracy theorists.

Nice going Fox.


MeadHallPoet said...

I would say two of the above.

Partly outrage at politicians (which everyone shares to some extent) and partly a toxic mix of racism and conspiracy theories.

Katie Cooks said...

Interesting thouughts