Wednesday, August 1, 2012

Liberal media bias chronicles: Biased polling

We were overseas for the 2008 Presidential election, but even from abroad the thumb-on-the-scale coverage the US media gave Obama was eminently visible. Whether it was the double standard regarding experience ("no executive experience for Obama" = no big deal, but "limited executive experience for Palin" = instantly disqualified from being President), the refusal to seriously dig into his background, or the Pavlovian cries of Racism! whenever anyone began to seriously criticize him, the media at all levels protected, fawned over, and virtually fellated Barack Obama.

And the scary thing is that they're looking to outdo themselves in 2012. The Dear Leader is in deep trouble this time around. The media are frantically searching for something ... anything ... to save their beloved.

One of their favorite tricks is conducting biased polls. There are several ways to bias polls. Asking biased questions of the "Do you still beat your wife?" variety is generally considered an amateurish way to do it. Far better to either sample "all adults" or "registered voters" instead of sampling "likely voters." "Likely voters" are, after all, the ones whose opinions we want, because they're the ones who are the most likely to actually vote. The "likely voter" sample is rarely Liberal enough for the media, though, so most of their polls generally tend to be "all adults" or "registered voters" in order to get a sufficiently Liberal result. Another tactic, much more subtle, is to skew the partisan makeup of the poll. This allows pollsters to present their results as if they represent America (or a given state) at large, and it's harder to check, which dissuades casually informed observers from looking deeper for the bias.

The upshot of biased polls is that they can create a sense of inevitable victory for one side where there's actually a close race, and they can demoralize opponents who think, "The guy I don't want to win is winning and there's nothing I can do about it."

The latest swing state poll provides a clear example of biased polling. In a year that is shaping up to be much closer to 2010 (when Republicans and Conservatives were fired up and Democrats were demoralized) than 2008 (when it was reversed), CBS and the New York Times put out a poll assuming that partisan breakdowns will be even more favorable to the Democrats in 2012 in PA, VA, and OH than they were in 2008. That is, in an election where the economy is the weakest we've seen in a generation, the incumbent is weighed down by his own wildly unpopular policies, and the opposition consistently shows itself to be much more enthusiastic than the part in power, the CBS/NY Times poll assumes that voters will vote even more Democratic than they did in 2008.

This kind of nonsense would be shot down in a research paper from a high school student. Either the people putting it out there are less competent than a high school student or they're too frightened of what the results of an actual poll would be to conduct one.

Here is what a less biased poll looks like: Obama is actually underwater in almost 75% of the country. Gallup had more integrity in conducting the poll, but they couldn't help themselves with the headline: Thirteen States and DC Give Obama Majority Approval.

No comments: