Thursday, 4 November 2004

Another Triumph for the Glug Election Indicator

The Glug Election Indicator performed creditably in picking the United States election result. It had George W. Bush a 56.6% chance of remaining President which translates in to a prediction of a 50.3% national vote for Bush. That was lower than the 51.1% shown by the current actual figures but substantially closer than the 49.4%prediction of Messrs Zogby and co.
If you are wondering how a 56.6% chance of winning becomes a predicted vote of 50.3%, we assume a normal distribution with a standard deviation of 1.5 percentage points. That is the standard deviation which the University of Iowa found applied to its Iowa Electronic Markets on political elections. (See So Who is the Underdog) in our Election Diary archive.
The Glug indicator, incidentally, predicted that the Liberal-National Coalition in Australia was a 77.8% chance of winning. That assumed a two party preferred vote of 51.2%. The final figure was 52.7%.

Another Crushing Defeat for Journalists and the Opinion Poll Industry .

There have been three predictions of cliff hanger election results by political journalists and opinion pollsters in English speaking democracies this year. And three clear cut wins by incumbent governments have made the pollsters and the other pundits look less reliable than astrologers.
Consider this wonderful document posted on the web site of Zogby International – the much quoted American polling guru.
Released: November 02, 2004

Our Call
Zogby International's 2004 Predictions
(as of Nov. 2, 2004 5:00pm EST)

2004 Presidential Election

Electoral Votes





Too close to callNevada (5)
Too close to callColorado (5)
Zogby International Finds: Bush at 49.4%, Kerry at 49.1%
The nationwide telephone poll of 955 likely voters was conducted (November 1-2, 2004). The MOE is +/- 3.2
(as of Nov. 2, 2004 5:00pm EST)
901 Broad Street, Utica, New York 13501 USA
1600 K Street, Suite 600, Washington,DC, 20006 USA NY phone 315.624.0200
Toll Free in the U.S. and Canada 1-877-GO-2-POLL | fax 315.624.0210
Contact sales and marketing at
Contact our web manager with any comments regarding this web site.

Copyright by Zogby International.
How about that for a fearless prediction! No wonder the British journalist Christopher Hitchens commented on the ABC’s Lateline last night that after Bush’s clear win "above all else people will laugh when they hear the word Zogby." In my opinion we should laugh at every pollster in every country and positively heckle any journalist stupid enough to quote one.
Zogby’s prediction of a 49.4% vote for Bush and 49.1% for Kerry was well wide of the mark. At the time of writing this comment, the New York Times website showed the primary vote as 51.1% for Bush and 48.0% for Kerry. In terms of electoral college votes the Times had it 249 to Bush and 242 for Kerry with 47 undecided with the paper’s lead story saying that "President Bush today seemed headed toward winning enough Electoral College votes to assure his re-election."
Heading indeed! That very cautious approach smacks of a paper that had been taking notice of Zogby and wanted to preserve the fiction of the cliff hanger (and thus their own credibility) for as long as possible. Only the ardent supporters of a party dominated by trial lawyers like the Democrats, who are itching to show their cleverness with a series of court challenges, would be hesitating about declaring George Dubya the clear victor.
I far prefer the honesty of the Toronto Star which had a headline after the Canadian Government was comfortably returned back in June that should be indelibly imprinted on the mind of anyone with the slightest interest in elections:
Results Confound Pollsters
Numbers Meant Little in the End
Were People Lying to Them?
The Canadian equivalents of Zogby had predicted that the election would be neck and neck with the governing Liberals on 32% and the opposition Tories on 31%. The Liberals ended up with 37% of the vote and a comfortable victory. Hence those headlines.
In between the Canadian and United States elections was our own poll where the government of John Howard increased its vote in the face of the pollsters predicting a decline.
In the face of these three disastrous results for the pollsters I can but repeat the advice I gave in the first entry in my Australian election diary back on 1st July 2004: "we should studiously ignore the opinion polls."