# Florida polls: Now Bush is up by 1% in two different ones. **URL:** \ **Category:** Great Debates\ **Created:** [November 2, 2004, 12:07am UTC](https://boards.straightdope.com/t/florida-polls-now-bush-is-up-by-1-in-two-different-ones/272447 "2004-11-02T00:07:09Z")\ **Posts on this page:** 20\ **Page:** 1
**Author:** ![rfgdxm](https://avatars.discourse-cdn.com/v4/letter/r/c0e974/32.png) [@rfgdxm](https://boards.straightdope.com/u/rfgdxm)\ **Post date:** [November 2, 2004, 12:07am UTC](https://boards.straightdope.com/t/florida-polls-now-bush-is-up-by-1-in-two-different-ones/272447/1 "2004-11-02T00:07:09Z")
Both Zogby and Survey USA have Bush ahead of Kerry by 1%. [http://olympics.reuters.com/newsArticle.jhtml?type=electionsNews&storyID=6681154](http://olympics.reuters.com/newsArticle.jhtml?type=electionsNews&storyID=6681154) [http://www.local6.com/news/3881411/detail.html](http://www.local6.com/news/3881411/detail.html) Bad news for Kerry. However, still within the margin of error. And undecides usually break about 2-1 for the challenger over the incumbent. ---
**Author:** ![BrainGlutton](https://avatars.discourse-cdn.com/v4/letter/b/82dd89/32.png) [@BrainGlutton](https://boards.straightdope.com/u/BrainGlutton)\ **Post date:** [November 2, 2004, 12:11am UTC](https://boards.straightdope.com/t/florida-polls-now-bush-is-up-by-1-in-two-different-ones/272447/2 "2004-11-02T00:11:58Z")
The latest Zogby poll, 10/31/04, has Kerry leading Bush in Florida, 48% to 47%, according to [http://www.electoral-vote.com/](http://www.electoral-vote.com/). ---
**Author:** ![tnetennba](https://avatars.discourse-cdn.com/v4/letter/t/7993a0/32.png) [@tnetennba](https://boards.straightdope.com/u/tnetennba)\ **Post date:** [November 2, 2004, 12:13am UTC](https://boards.straightdope.com/t/florida-polls-now-bush-is-up-by-1-in-two-different-ones/272447/3 "2004-11-02T00:13:12Z")
I have a feeling the polls tomorrow will tell us the real outcome, one way or the other. ---
**Author:** ![rfgdxm](https://avatars.discourse-cdn.com/v4/letter/r/c0e974/32.png) [@rfgdxm](https://boards.straightdope.com/u/rfgdxm)\ **Post date:** [November 2, 2004, 12:17am UTC](https://boards.straightdope.com/t/florida-polls-now-bush-is-up-by-1-in-two-different-ones/272447/4 "2004-11-02T00:17:30Z")
> [@BrainGlutton](#): > > The latest Zogby poll, 10/31/04, has Kerry leading Bush in Florida, 48% to 47%, according to [http://www.electoral-vote.com/](http://www.electoral-vote.com/). No. The latest Zogby poll came out today. [http://www.electoral-vote.com](http://www.electoral-vote.com) had their last update before this latest Zogby poll. ---
**Author:** ![Squink](https://avatars.discourse-cdn.com/v4/letter/s/b5e925/32.png) [@Squink](https://boards.straightdope.com/u/Squink)\ **Post date:** [November 2, 2004, 12:25am UTC](https://boards.straightdope.com/t/florida-polls-now-bush-is-up-by-1-in-two-different-ones/272447/5 "2004-11-02T00:25:08Z")
> [@](#): > > In Florida, the latest Gallup poll showed that about a third of the polled voters had already cast ballots and that Kerry led 51 percent to 43 percent among them. [here](http://www.dailykos.com/) ---
**Author:** ![andros](https://avatars.discourse-cdn.com/v4/letter/a/e480ec/32.png) [@andros](https://boards.straightdope.com/u/andros)\ **Post date:** [November 2, 2004, 12:25am UTC](https://boards.straightdope.com/t/florida-polls-now-bush-is-up-by-1-in-two-different-ones/272447/6 "2004-11-02T00:25:10Z")
If it’s within the MOE, it doesn’t exist. ---
**Author:** ![Malodorous](https://avatars.discourse-cdn.com/v4/letter/m/e47774/32.png) [@Malodorous](https://boards.straightdope.com/u/Malodorous)\ **Post date:** [November 2, 2004, 12:36am UTC](https://boards.straightdope.com/t/florida-polls-now-bush-is-up-by-1-in-two-different-ones/272447/7 "2004-11-02T00:36:21Z")
> [@andros](#): > > If it’s within the MOE, it doesn’t exist. If someone is leading within MOE, then they are still the most probable winner. If your leading outside MOE, then the chances of actually being ahead are greater then some threshold, usually 90-95%. But even if your a fraction of a precent ahead in the polls, then you have a greater then 50% chance of being ahead in the actual election. Of course, all this is in theory. Given all the assumptions and flaws in political polls, they’re probably not a great example of statistics at its best. ---
**Author:** ![Malodorous](https://avatars.discourse-cdn.com/v4/letter/m/e47774/32.png) [@Malodorous](https://boards.straightdope.com/u/Malodorous)\ **Post date:** [November 2, 2004, 12:41am UTC](https://boards.straightdope.com/t/florida-polls-now-bush-is-up-by-1-in-two-different-ones/272447/8 "2004-11-02T00:41:40Z")
> [@Squink](#): > > In Florida, the latest Gallup poll showed that about a third of the polled voters had already cast ballots and that Kerry led 51 percent to 43 percent among them. This of course just means that more Kerry voters went to the polls early, not that he will have more voters total when the results are tallied. Given that the precentage of Dems that are worried about problems and fraud at the polls is far greater then the number of Repubs (can’t find the poll now for exact numbers, but it was a huge difference), I would expect that Dems would be showing up for early polling in far greater numbers. ---
**Author:** ![rfgdxm](https://avatars.discourse-cdn.com/v4/letter/r/c0e974/32.png) [@rfgdxm](https://boards.straightdope.com/u/rfgdxm)\ **Post date:** [November 2, 2004, 12:44am UTC](https://boards.straightdope.com/t/florida-polls-now-bush-is-up-by-1-in-two-different-ones/272447/9 "2004-11-02T00:44:18Z")
> [@Malodorous](#): > > If someone is leading within MOE, then they are still the most probable winner. If your leading outside MOE, then the chances of actually being ahead are greater then some threshold, usually 90-95%. But even if your a fraction of a precent ahead in the polls, then you have a greater then 50% chance of being ahead in the actual election. > > Of course, all this is in theory. Given all the assumptions and flaws in political polls, they’re probably not a great example of statistics at its best. Right. Also, IMO even if Bush wins FL, that alone will still leave him behind Kerry in the EC. It is possible that Kerry could still win this without FL. ---
**Author:** ![Knorf](https://avatars.discourse-cdn.com/v4/letter/k/58956e/32.png) [@Knorf](https://boards.straightdope.com/u/Knorf)\ **Post date:** [November 2, 2004, 12:45am UTC](https://boards.straightdope.com/t/florida-polls-now-bush-is-up-by-1-in-two-different-ones/272447/10 "2004-11-02T00:45:17Z")
Any difference with the margin of error can be neither a good nor a bad sign for either candidate, excepting the general observation that, historically, a tie in polls before an election is bad for the incumbent. ---
**Author:** ![Airblairxxx](https://avatars.discourse-cdn.com/v4/letter/a/f9ae1b/32.png) [@Airblairxxx](https://boards.straightdope.com/u/Airblairxxx)\ **Post date:** [November 2, 2004, 12:53am UTC](https://boards.straightdope.com/t/florida-polls-now-bush-is-up-by-1-in-two-different-ones/272447/11 "2004-11-02T00:53:48Z")
At this point, there isn’t much to divine from a poll showing someone leading by 1%. The real question is: Is the incumbent above 50%? Or is he hitting a ceiling somewhere in the neighborhood of 46-48%? If the latter, then the incumbent is in serious trouble. Presumably, voters would have had enough time to decide if the current guy deserves another four years. If they enter the polling booth and still haven’t been convinced, they’ll likely say, “Fuck it, let’s put someone else in there.” That’s why both sides are relying on massive Get Out The Vote campaigns. But if you want to put credence in Florida polls at this point, what about the one with [Kerry in the lead by five?](http://www.foxnews.com/story/0,2933,137279,00.html) (And for shits and giggles, look where that poll came from.) ---
**Author:** ![rfgdxm](https://avatars.discourse-cdn.com/v4/letter/r/c0e974/32.png) [@rfgdxm](https://boards.straightdope.com/u/rfgdxm)\ **Post date:** [November 2, 2004, 1:02am UTC](https://boards.straightdope.com/t/florida-polls-now-bush-is-up-by-1-in-two-different-ones/272447/12 "2004-11-02T01:02:50Z")
> [@Airblairxxx](#): > > But if you want to put credence in Florida polls at this point, what about the one with [Kerry in the lead by five?](http://www.foxnews.com/story/0,2933,137279,00.html) (And for shits and giggles, look where that poll came from.) And that Fox News poll was released just today. ---
**Author:** ![AZCowboy](https://avatars.discourse-cdn.com/v4/letter/a/97f17d/32.png) [@AZCowboy](https://boards.straightdope.com/u/AZCowboy)\ **Post date:** [November 2, 2004, 1:44am UTC](https://boards.straightdope.com/t/florida-polls-now-bush-is-up-by-1-in-two-different-ones/272447/13 "2004-11-02T01:44:39Z")
> [@Malodorous](#): > > If someone is leading within MOE, then they are still the most probable winner. Uh, no, statistically speaking, if the lead is within the margin of error, you cannot deduce anything about who is the most probable winner. That is why it is a margin of error, and not a level of confidence. ---
**Author:** ![AHunter3](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/ahunter3/32/368_2.png) [@AHunter3](https://boards.straightdope.com/u/AHunter3)\ **Post date:** [November 2, 2004, 3:06am UTC](https://boards.straightdope.com/t/florida-polls-now-bush-is-up-by-1-in-two-different-ones/272447/14 "2004-11-02T03:06:56Z")
No, he’s right. The conventional standard is 95%, or, to expand it a bit, “the distribution within the sample is not going to differ from the actual population by more than the stated margin of error (which is usually expressed as plus or minus some figure) in 95% of all cases, assuming a normal distribution within the population”. The 95% figure is just a convention. Nothing magic about it. No reason not to express the statistic with a wider margin of error and therefore be able to say it with a higher confidence, say, 98% of the time. So if a measurement is within a stated margin of error, that is for a given confidence only, and for some other confidence (less confidence) it could be said to be outside the margin of error. People for some reason expect stats to say something “is so” when statistics by their very nature never say that anything “is so”, but only that based on sample size and assuming a distribution shape you can say that X% of the time the real population will be within + or - Y points of finding Z which is what you measured in your sample. ---
**Author:** ![AZCowboy](https://avatars.discourse-cdn.com/v4/letter/a/97f17d/32.png) [@AZCowboy](https://boards.straightdope.com/u/AZCowboy)\ **Post date:** [November 2, 2004, 3:16am UTC](https://boards.straightdope.com/t/florida-polls-now-bush-is-up-by-1-in-two-different-ones/272447/15 "2004-11-02T03:16:39Z")
I agree with everything you wrote, but I’m not sure where you disagreed with what I wrote. Inside the margin of error, the test has no statistical validity. There is no value within the margin of error with higher or lower predictive value. You can simply assert, with 95% confidence, that the value is somewhere within the range. It doesn’t mean that any value within the range is more probable. ---
**Author:** ![Sam\_Stone](https://avatars.discourse-cdn.com/v4/letter/s/ecccb3/32.png) [@Sam\_Stone](https://boards.straightdope.com/u/Sam_Stone)\ **Post date:** [November 2, 2004, 3:32am UTC](https://boards.straightdope.com/t/florida-polls-now-bush-is-up-by-1-in-two-different-ones/272447/16 "2004-11-02T03:32:58Z")
There are two margins of error we have to consider here. The first is the mathematical margin of error of the poll given a certain sample size. The other, which is far more important, is the error built in to the assumptions of the pollster and the information they are missing completely. For example, should you measure likely voters, or registered voters? What about breakdown by party? If 35% of Republicans voted last time but only 30% of Democrats, should you make sure that your sample has 35% Republicans and 30% Democrats, or should it be weighted 50/50? How about those who are traditionally under-represented? To give you an example of how big this swing is, consider the difference between the Zogby poll in Wisconsin (Kerry +7) and Gallup (Bush +8) is about eight Sigmas, way, way way outside the margins of error for either poll. Clearly the error is coming from somewhere else. Also, notice that the polls we’ve been looking at the last few days don’t have undecideds, and yet anywhere from 3-8% of the people claim to be undecided. So where do their votes go? Some pollsters use the historical 3/4 break for the challenger among undecideds, which pushes up Kerry’s numbers. Others split them 50/50, which makes it a wash. Which is the ‘correct’ assumption? Who knows? This is a different election from most in the past. The ‘margin of error’ around this assumption alone may be as great as the sample error of the poll. Plus there’s the cell phone users, the Amish, the new Republican ‘ground campaign’ which they claim is going to push their numbers up way over the historical values, which would make the assumptions of these polls way out of line. How will it all break out? Hell if I know. There are more ‘unknowns’ in this election than in any other elections in my memory. I suspect we’re going to get a bit of a surprise tomorrow, one way or the other. ---
**Author:** ![BrainGlutton](https://avatars.discourse-cdn.com/v4/letter/b/82dd89/32.png) [@BrainGlutton](https://boards.straightdope.com/u/BrainGlutton)\ **Post date:** [November 2, 2004, 3:37am UTC](https://boards.straightdope.com/t/florida-polls-now-bush-is-up-by-1-in-two-different-ones/272447/17 "2004-11-02T03:37:02Z")
> [@Sam Stone](#): > > Plus there’s the cell phone users, the Amish . . . Do the Amish vote? ---
**Author:** ![rfgdxm](https://avatars.discourse-cdn.com/v4/letter/r/c0e974/32.png) [@rfgdxm](https://boards.straightdope.com/u/rfgdxm)\ **Post date:** [November 2, 2004, 3:44am UTC](https://boards.straightdope.com/t/florida-polls-now-bush-is-up-by-1-in-two-different-ones/272447/18 "2004-11-02T03:44:40Z")
> [@Sam Stone](#): > > There are two margins of error we have to consider here. The first is the mathematical margin of error of the poll given a certain sample size. The other, which is far more important, is the error built in to the assumptions of the pollster and the information they are missing completely. [http://www.electoral-vote.com/index.html](http://www.electoral-vote.com/index.html) Scroll down the page to the final 2000 national poll results. While Gore was just slightly higher in the final popular vote, only 2 polls had Gore ahead, with 2 tied. This suggests to me that however the pollsters were doing things in 2000, they had a sample that was skewed Republican. Perhaps younger voters hang up on pollsters more than older people? With the cell phone factor missing a lot more younger voters than in 2000, this may skew the poll sample even more Republican. ---
**Author:** ![rfgdxm](https://avatars.discourse-cdn.com/v4/letter/r/c0e974/32.png) [@rfgdxm](https://boards.straightdope.com/u/rfgdxm)\ **Post date:** [November 2, 2004, 3:47am UTC](https://boards.straightdope.com/t/florida-polls-now-bush-is-up-by-1-in-two-different-ones/272447/19 "2004-11-02T03:47:44Z")
> [@BrainGlutton](#): > > Do the Amish vote? Yes, some do. [http://www.google.com/search?sourceid=navclient&ie=UTF-8&q=amish+voting](http://www.google.com/search?sourceid=navclient&ie=UTF-8&q=amish+voting) ---
**Author:** ![rjung](https://avatars.discourse-cdn.com/v4/letter/r/45deac/32.png) [@rjung](https://boards.straightdope.com/u/rjung)\ **Post date:** [November 2, 2004, 7:38am UTC](https://boards.straightdope.com/t/florida-polls-now-bush-is-up-by-1-in-two-different-ones/272447/20 "2004-11-02T07:38:40Z")
> [@Airblairxxx](#): > > The real question is: Is the incumbent above 50%? Or is he hitting a ceiling somewhere in the neighborhood of 46-48%? If the latter, then the incumbent is in serious trouble. Bush hasn’t gotten higher than 48%-49% in the last few weeks, IIRC. He’s in **deep** doo-doo here, and the campaign folks know it. I could practically smell the desperation when I heard Bush refer to Kerry as a “flip-flopper” this morning… :rolleyes: [Next page](https://boards.straightdope.com/t/florida-polls-now-bush-is-up-by-1-in-two-different-ones/272447.md?page=2)