Which US pollster is best for you?
September 26th, 2004
-
Is the Bush-Cheney ticket a certainty?
Since the Republican Convention in New York at the end of August the betting markets have followed the polls and Bush is a red hot favourite. It’s hard to call anything other than Bush.
But it’s worth looking back at what happened last time to check which pollsters were most accurate. These were the final polls.
The final result was Gore 500,000 votes ahead. But after the pro-longed legal battle Bush won Florida, came top in the “electoral college” and took the White House. Only Zogby and CBS got the vote shares right.
A useful tool is this site which has an excellent interactive map showing the latest polls in each state and a running total. What is interesting is that you can select your pollster and, beating in mind Zogby’s success last time, the picture for the Democratic ticket does not look as bad as some are painting it.
It’s a brave punter, though, who goes with Kerry even at the current odds but it might start to move and it does look less certain for the President if you just follow the Zogby national and state polls.
If you think that this election is over there’s now a market on who will by the Republican candidate in 2008. This is a long time to lock up your money, many things could happen in the meantime, and the prices don’t look very generous.
MessageSpace Advertising
GO to “David Leip’s Election Atlas Site” and visit the forum there… a pollster called Vorlon will give you a very comprehensive assessment of the various polling firms and CBS and Zogby are not rated as good firms, Zogby is generally all over the map “dead on ½ the time and way off the rest” while CBS only got close to the real result in their polls in 2000 and before and since have been way out of step with other polling firms and actual results…
Interesting Fox News was closest in 2000 and still shows a close race with a modest Bush lead. The Polls for Newsweek and Time are both utter bull£$(* and should be disregarded they gave Kerry double digit leads not long ago and that was also unreliable in the extreme… of all the networks Fox has pretty much the strongest polls and CBS has the weakest and is closely followed by CNN and MSNBC who both have very bad records as far as polls go… Mason Dixon are great at poll Florida (Where they are based) while James Carville’s “Democracy Corps” firm are also good so long as you shave around 1 piont of the Dem total… but if you invest your faith in Zogby and CBS you will lose a great deal of money.
Out of interest, any chance of betting on the Ausie Election? although I doubt there would be any good prospects as almost all elections over there are necessarily close
blue sq bet
Coalition 4/11
Labor 2/1
Ben - for Aussie elections see - this
Rasmussen now has Kerry 1 point up in Florida - it still looks close, though this Thursday\’s debate could trigger a significant shift either way. If you can get 2/1 againt Kerry it\’s tempting.
BV - I agree with you assessment. I\’m waiting till I get some money back from Hartlepool before going in again.
The security of a man with a flat position!
If nevertheless you stay up for the Hartlepool result on Thursday night, you may as well watch the first Bush-Kerry debate on http://www.c-span.org at 2am UK time.
Interesting quote on the Gallup method from Electoral-Vote.com:
\”It is becoming increasingly clear that the pollsters are producing the results that the people paying the bills want to hear. Even pollsters who were once thought to be above suspicion are now suspicious. Gallup, for example, is now normalizing its samples to include 40% Republicans, even though the 2000 exit polls showed the partisan distribution to be 39% Democratic, 35% Republican. There is scant evidence that the underlying partisan distribution has changed much since then. Other pollsters also normalize their data, but most don\’t say how. Normalizing the sample to ensure the proper number of women, elderly voters, etc. is legitimate provided that the pollster publicly states what has been done.\”
But are we sure Gallup are _normalising_ to 40%? (As opposed to normalising based on demographics and letting party ID be as it is measured). When we looked into this last week it sounded as if they were treating party ID as a measurable variable rather than assuming it had to come out at fixed proportions.
Electoral-Vote are completely wrong. Gallup don\’t weight by party ID at all, they definitely treat it as a variable. Zogby do weight by party ID, and weight to something like 39% Dem, 34% Rep.
Party iD isn\’t weighting to the fixed stats on the number of people who are registered as Rep/Dem/Ind, because lots of states don\’t have registration - it is merely how people answer the question \”which party do you identify with\”. The dispute is whether or not party ID changes quickly over time - some believe it changes very slowly over time, others believe it is affected by people\’s voting intentions and hence changes from day to day (i.e. if someone decided they are going to vote for Bush they are more likely to say they identify with the Republican Party.) If the former is correct, then Zogby is right to weight by Party ID, if the latter is correct Gallup are right not to.
Either way - it is simply untrue to say that Gallup are weighting (or \”normalising\”) on party ID - read it straight from the horses mouth here.
It seems like a linguistic nuance, really. The hypothesis that party ID is essentially fixed would predict voters to say, for example, \”I\’m a Democrat but I\’m voting for Bush\”. Now certainly in Britain you wouldn\’t hear that much. The closest you would get is \”I\’ve always been Labour but this time I\’m voting Lib Dem\” (vary parties to suit taste). So in the context of the way we speak about politics here, \”party identification\” is no less fluid than voting patterns.
To at least some Americans, it _does_ seem to be more fixed a concept. However, it seems so subtle a matter of language that you would expect wide variations between demographics (and probably between different poll questions) in how tightly a voter\’s declared party ID is tied to his/her declared voting intention.
And on next week\’s edition of Wittgensteinian Psephology…