October 2, 2012 by Julia
Here’s the deal. I hate math–no, really, that’s why I went to law school. So, this post will not be about math. Instead, let’s take more of a political science angle (emphasis on the political, because I laugh whenever it’s called a “science.” The only thing “scientific” about it is statistics, in the same way that all social “sciences” use and manipulate statistics–i.e. not scientifically at all).
To start out, check out Charlie Martin’s math-based discussion of how polls work and why oversampling renders poll “results” completely useless. Then consider that even for all that oversampling, only 9% of sampled households even give answers to polls. Think of everyone that excludes, including people with no landline, people who work during the times that the polling companies call, people who simply refuse to answer, people who don’t trust the media, etc. It’s unlikely that the 9% of people who take the time to answer polls (I am one of them, by the way, but only when I am home to answer the phone, which is rarely) are actually representative of the population at large.
So then, here’s the situation: we’ve got polls that oversample people who aren’t representative of the country as a whole, and the people who may disagree with the oversampled people are more likely to not respond in the first place. The inescapable conclusion? Polls have no value whatsoever as anything other than propaganda tools. The public agrees, according to a poll by Daily Kos and SEIU (imagine what the results were before they tampered with them).
That only holds, of course, for publicly disseminated polls. Internal polling, which I would imagine they spend more time getting right, is probably a different story–and in this election in particular, probably yields completely different results. We know Obama’s “lead” over Romney in any recent national poll has been within the margin of error, and that the polls have oversampled Democrats. If the Obama campaign has done a better job of finding a representative sample in the 9% (or even got a higher response rate than 9%), it’s a good bet that the internal numbers show the president in serious trouble.
If not, if the numbers showed Obama with a healthy lead, would there be a reason to bribe Ohioans with federal money? Offer federal money to bribe defense contractors who break the law? Make creepy calls to Catholics that lie about the LDS church, Obama’s abortion record (I mean, really, Obama would probably be offended that people are running around saying he is pro-life), and Planned Parenthood‘s services? Make an ad about how Mitt Romney doesn’t hug his garbage man? (yes, that is real)
Well, okay, the Left may do those things anyway, because bribery and lies are just part of the game plan. But doesn’t it feel just a tad desperate? I mean, really, they blamed Romney for a woman’s death from cancer. Their internal poll numbers must be terrible.