The Standard has been covering some admirable work on recording and presenting this year's political opinion polling on Wikipedia. Lynn Prentice noted the trend lines and an analysis of potential bias -- in any direction.
The developing Wikipedia article is partially the work of Denmark-based New Zealander Mark Payne, who emailed me last night to explain the project:
It's a "poll of polls" that I and some others have been doing via Wikipedia. It seems to suggest that the race is not over by any means, but:
1. the systematic difference between the individual polls is far greater than any statistical noise due to sampling and,
2. that the result is more dependent on the outcomes of certain key electorates than the party vote (eg will Winston make it?)
I've also looked for systematic biases in the polling firms. There certainly appear to be some (at least relative to the ensemble of polls):
Colmar-Brunton and Nielsen show significant bias towards National, TNS towards labour. The digipoll doesn't appear as bad as everyone makes it out to be, which is interesting.
So, what do the statistics tell us about the final outcome? No bloody clue. Ask me again November 9th, when I've finished the new and improved model (also known as watching the count on election night!).
Back at The Standard, Steve Pierson has his graph on again, with a "horse-race" analysis that quite vividly depicts the Parliamentary implications of several recent polls -- based on his assumption that the Maori Party will choose Labour if it is tasked with anointing a new government. Labour forms a government in two polls, National in one, and the race is tied under Colmar Brunton.
It ought to be required reading for the author or authors of the New Zealand Herald editorials of Friday and Saturday.
Friday's, as I noted, declared victory for National, and Saturday's effectively states that if would not be fair or legitimate if Labour were to form a coalition representing a majority in the House without itself being the single highest-polling party.
Such a majority government, it insists:
… is clearly not what most voters want or believe should happen. Around 80 per cent of them vote National or Labour and when they go to the polling booth they believe they are choosing a Government. If their party is beaten at the ballot box they accept it is fair and square. Parties trifle with that result at their peril.
As an argument, it's either feeble-minded or disingenuous. The fact that "around 80 per cent … vote National or Labour" is neither here nor there. As a rule, people vote for a party because they wish to see it in government. They might place conditions on that wish, but a Green voter is hardly likely to reject a coalition with Labour because it's somehow unfair to National.
The Herald is free to make an argument that a National-led coalition of three parties would be a sounder government than a Labour-led coalition of four parties, but it should do so without making presumptions on the public will. If a majority coalition can be formed without betraying undertakings made before the election, then by definition it represents the will of the majority of voters.
"It's like smoking," Shayne Carter observed to me recently. "It's bad for your health and it's addictive, but it's fun."
Shayne was referring to his practice of following the US presidential election via Fox News. Personally, don't have the constitution for that -- but I do love me some wingnut bloggery.
You'd turn crazy if you had to find this stuff yourself, but happily, there are people who'll do it for you. I'm a big fan of Village Voice columnist Roy Edroso, who rounds up the Joe the Plumber madness and details the latest crazy Michelle Obama story to be clutched to wingnut breasts.
The crazy went off the scale late last week after a troubled McCain campaign worker, 20 year-old Ashley Todd, reported that she had been viciously assaulted by a "6' 4" black man", who, upon seeing her McCain bumper sticker, became enraged and carved a "B" into her face to teach her a lesson. But said "B" was reversed -- much as it might look as if a person had scratched it into her own face in a mirror.
The incendiary allegation was soon debunked as a sad hoax -- but not before a McCain campaign aide had shopped the story to the media, and was indeed the source of the claim that the "B" stood for "Barack".
Edroso recaps the affair in his latest Voice column, noting the weird contortions of the wingnuts who had been screaming race war (Jonah Goldberg's dive for cover is particular cause for mirth). There's lots more in his Alicublog.
Blue Texan's always-amusing Instaputz focuses on Glenn "Instapundit" Reynolds and his chums, and he finds Reynolds in classic passive-aggressive intellectually dishonest form over the Todd business.
And self-described "reasonable conservative" Jon Swift surveys great moments in election-year conserva-blogging.
You may have other favourites you'd like to share.
And this just in: US authorities have released details of the apprehension of two racist skinheads who planned to kill Barack Obama as the culmination of a massacre of innocent black people.
PS: This week'sMedia7 programme looks at the unrelentingly bad economic news, and explores whether, at least in so far as it guides consumer behaviour, it could be self-fufilling. The panel is Bryan Gould, the former Waikato University vice-chancellor and British Labour MP; Briar Harland, executive director of Colmar Brunton in Auckland; and psychological therapist Jock Matthews. It's an afternoon record, so if you'd like to join us from 2pm at The Classic in Queen Street, hit "reply" and let me know asap.