Who's really ahead? Bush or Kerry?

[chroot]

I keep seeing poll reports with wildly differing statistics. I was pretty well convinced by the media that Kerry was clearly ahead until this morning, when the newly-released CNN/Gallup poll surprised me by saying that Bush is strongly ahead. Most of the statistics seem pretty useless, since their margin of error is larger than the difference between the candidates. Which candidate is really ahead? Is there any way at all to tell with any reliability in a sample size of only around a thousand people? Which polls have been historically accurate?

It seems certain that the majority of voters believe Kerry did better in the debates, and the (vast) majority of Americans disapprove of the job Bush is doing. How in the world is Bush supposedly holding a six-percent lead?

- Warren

 

[member 5645]

It seems certain that the majority of voters believe Kerry did better in the debates, and the (vast) majority of Americans disapprove of the job Bush is doing. How in the world is Bush supposedly holding a six-percent lead?

- Warren

Because
1>lesser of two evils for some
2>Kerry's campaign is pathetic at best.

 

[Aquamarine]

The market is very efficient, usually more so than polls:
http://128.255.244.60/graphs/graph_Pres04_WTA.cfm

 

[russ_watters]

I keep seeing poll reports with wildly differing statistics. I was pretty well convinced by the media that Kerry was clearly ahead until this morning, when the newly-released CNN/Gallup poll says Bush si strongly ahead.

HERE is USAToday's article. It says:

-The president leads Kerry 52%-44% among likely voters
-Kerry led by 1% a week ago
-Bush leads 49%-46% among registered voters
-Error margin 3-4%

That, to me, adds up to not a whole lot - that big of a swing that fast doesn't seem reasonable.

 

[NateTG]

The race is essentially tied

The most accurate polling is done by the campaigns, and it's too close to the election for them to tell the truth about their results.

Depending on who you talk to, various polling organizations have biases, or problems in methodology. Moreover, it's been suggested that as much as 28% of the electorate has cell phones, and is consequently not being polled at all.

In general, the democrats are more mobilized than they have been in a long time, and I expect that with Bushco running a get-out-the-base campaign both parties are going to see large turnout. That means that the 'likely voter' models may not be applicable.

It's also worth noting that Bush led Gore by several points in the polls coming into the last presidential election, and you already know that that one was very close.

National polls aren't really that useful for presidential elections anyway unless there is a large difference, since electoral votes are distributed by state - last time around, Gore won the popular vote, and lost the electoral vote. The electoal-vote sites that I'm familiar with:
http://www.race2004.net/
and
http://www.electoral-vote.com/
both have the race as too close to call, and have, more or less, been there for the last week with Kerry making very slow gains.

http://www.pollkatz.homestead.com/
Astutely observes that the race is almost certainly going to be decided in Ohio, Pennsylvania, and Florida, and that whomever wins two of the three is likely to win. (Three of three would be a landslide.)

 

[check]

I’m not going to trust any polls with predicting a winner simply because telephone polling excludes people who have all cellular telephone service to their household. Some 5% of the total US population has service like this with about 15% of 18-25 year olds get this kind of service. It's leaving out a big chunk of a demographic.

http://www.cnn.com/2004/ALLPOLITICS/10/15/polls.cellphones.ap/index.html

 

[chroot]

I feel dumb for never having considered that, check (and NateTG) -- I am one of those 18-25 year old people who have only cellular telephone service!

- Warren

 

[cragwolf]

The market is very efficient, usually more so than polls:
http://128.255.244.60/graphs/graph_Pres04_WTA.cfm

There are also the betting exchanges:

Go to http://www.betfair.com and on the sidebar click on "Special Bets", then click on "USA", then click on "Politics", then click on "US President 2004", then click on "Winning Party". Bush is trading around $1.75 and Kerry is trading around $2.35. So Bush is favoured to win. Chalk another one up for the good ol' boys! Yee ha!