Rigging the poll: NYT sample nearly triples real-population share of ultra-conservatives

After Washington Post/ABC News put out their poll drastically over-representing GOP and Republican leaning voters, New York Times and CBS News have come up with brand new blunders of their own on their poll released late afternoon yesterday, showing the president at an approval rating of "all time low" 41%. This poll has the most glaring, dumbfounding proof that its sample is rigged staring at you in the very front page of the poll's crosstabs.

In the polling sample, Republican primary voters are 29.8% of all those polled (301 of 1009 total polled). But luckily, we don't have to look at the Times or CBS News to know what the actual turnout of GOP primary voters are as a percentage of the voting age population is, what with actual primaries going on right now and stuff. The actual percentage of voters turning out in GOP primaries as a percentage of the voting age population? 11.5%. So the pollsters nearly triple the representation of GOP primary voters over what the actual voting numbers are.

Now think about this. Who votes in Republican primaries? Uber-right wing conservatives. These are the same group of gems that think that the president is both a Muslim and a member of a radical Christian church. You more than double - nearly triple their proportion in the sample from what their percentages actually are, just what exactly do you expect to happen to the president's numbers from that cooked, rigged sample?

This stunning and blatant rigging of the numbers comes on top of the biases also present in the ABC News/Washington Post poll: this poll too narrows the Democratic-Republican difference to the same level as the ABC news poll (D +4) as opposed to the actual voter registration data showing a much wider spread in favor of Democrats (D +12).

So yes, you can believe this poll if you suspend your disbelief and also believe in the following:
  • People who are voting in GOP primaries are nearly a third of the voting age population, even though empirical data shows that they are only slightly over one in 10.
  • The actual data showing a big Democratic voter registration edge is simply false and roughly the same number of people identify with each party.
  • A boadload of voters will suddenly turn to the beliefs and voting patterns of the far far Right (they will more than double in number, to be exact) in November even though they're not participating in picking the GOP nominee. Or, in the alternative, the voting age population will suddenly shrink but all the wingnuts will still be there.
  • Pigs fly. Actually, scratch this one. This one may be more believable than this poll.
This is the height of insanity. What the Times and CBS News are polling is a sample severely skewed towards extreme conservatives (people who vote in Republican primaries). Why? Because they really want to poll the Republican primaries. But you add the president's approval to this set of cooked sample, and you can create a nice stir in the media based on absolute garbage. And the rest of us are left no smarter for the clever tricks of pollsters vastly over-representing the far-right in their samples and passing it on as a legitimate national poll.

That's why it's not just the president's numbers that went down 8 percentage point (duh), but every other ultra right wing position fared better in the GOP poll, including support for requiring insurance plans covering women's contraceptives as preventive care going from a 61-31 in favor to 51-40 opposed - a full 41 point swing in favor of the wingnut position. In a month. Stop the presses!

I bet that had they added a question about whether Barack Obama should be lynched, a pretty sizable portion of this sample would answer in the affirmative. Oh, I dunno, about 25 percent or so.

Give me a freaking break. Pollsters, you need to have your heads slammed together and start doing your actual job and learn something about big, glaring, sampling errors. Pundits and hosts, please stop flying off the god-damned nest based on poll numbers and 'statistics' you know nothing about.

I understand that the media needs to build a narrative of a "president in crisis" for the election year. But you know what? There are real crises going on in our country and in our world right now. American children are going hungry to bed every night. The middle east is one misstep away from blowing up. Women's rights are under attack. Minority voting rights are under seize. We have real crises. It's time for the media and the pollsters to quit making up fake ones.