What is it with wingnuts and pollsters?
A post last year about polling generated a lot of discussion. Since then we’ve come across plenty of conservatives complaining about this or that poll — as though polling is little more than the devil’s work because after all, Satan loves his representative samples displaying no signs of heteroscedacity when running OLS regressions. The latest should be in a category all its own however. Writing about the results of yesterday’s elections in Germany, Davids “Brent Bozell” Medienkritik writes:
One of the biggest losers in this election will be the pollsters as they clearly and dramatically failed to accurately predict the correct outcome. [Emphasis in the original.]
If you want to predict the future, you call Mrs. Cleo’s Psychic Hotline. Polls can no more predict the future than you can predict next week’s weather by looking out your window today. Properly run, polls are a wonderful thing. The latest polls provided estimates that, given their margin of error, were within spitting distance of the actual results. So why is it that some people are always yapping on about pollsters? Is it because they tell us a certain failure is extremely unpopular, unless the US is being attacked or is attacking? Is it something else? Maybe our conservative reader (Dr. BLT) can tell us.
Bonus points: As long as we’re writing (kind of) about the German election — the Wall Street Journal’s John Fund writes:
Most voters and nearly all of the business community wanted a decisive result. Instead the peculiarities of Germany’s parliamentary system delivered a complete mess… [Emphasis added]
For those not familiar with Germany, the peculiarities of its parliamentary system are usually, and normally, called a proportional representation system.
Satan loves his representative samples displaying no signs of heteroscedacity when running OLS regressions
I hate the fact that I know what this means.
Except the part about Satan, that is.
spencer, clue the rest of us in.
Yes, down with the evil overlord pollster class, the world is rife with their virulence!
All that fancy-schmancy statistical talk is just obfuscation of the plain facts, which are clear to anyone with an ounce of common sense, and don’t need to be asked of your “representative sample.” W polls at 90+% approval? Plain as the nose on your face, some ejucated librull will try to blame it on the fact that the sample was crazy Aunt Adolph and Kaye Grogan. So to hell with Mr. Bartlett AND his sphericity test, as well as Kurt Osis, whoever that is.
Dear lord, I’m drunk already.
Proportional representation is a moronic system. Where is room for creativity, like artistically gerrymandered districts?
Actually, German system does have elements of first-past-the-post and it is probably too complicated for the simple folks at WSJ.
Seriously, Ho hard would it be to do a poll with a one percent (+/- 0.5) MOE?
statisticians?
this 4 percent is beginning to feel like stroking the wounded ego because he has ‘the football’.
Seriously, How hard would it be to do a poll with a one percent (+/- 0.5) MOE?
statisticians?
this 4 percent is beginning to feel like stroking the wounded ego because he has ‘the football’.
There’s PR, and then there’s PR. Germany’s PR system does have its peculiarities, you know. For instance, there will be a special election in one district of Dresden in October. (An obscure neonazi candidate had the good form to die shortly before the voting, hence needed to be repalced. German law requires a certain minimum time before voting on any given panel of candidates, hence the small delay.) In that election,if the CDU do really, really well… their number of MPs in the Bundestag will decrease by one.
But the peculiarities of German electoral law, such as they are, were not what delivered Sunday evening’s result. That the German did not elect the government the WSJ wished for is down to this: the party the WSJ backed ran an abortion of a campaign, squandering what had been, not very long ago, a 20% lead. In the event they polled less than 1% better than the incumbent SPD, and have but three more seats in the Bundestag (assuming they don’t do excessively well in Dresden). Indeed, another perverse Dresden outcome would see the CDU and SPD evenly matched for seats. That’s pretty much a purely mathematical possibility, though — virtually impossible in real-world terms.
(Click through on the URL for more on the German elections than you could possibly want to read.)
Catastrophic success?
Sadly, No! takes the Wall Street Journal to task for saying that the German elections failed to produce a clear result because Germany’s parliamentary system is ‘peculiar’. As S,N!