What's the most accurate poll concerning the November election ?

[steve2470]

Related question, do any of the pollsters call cell phones ? Thanks for your time.

[dimbear]

1. The most accurate indicator would be Intrade or some other

gambling den. At Intrade you see people actually wagering money on the outcome, they aren't kidding around. And as you can imagine they keep extemely well informed as to the odds. Plus as gamblers they don't (supposedly) let their emotions get involved. It's a rough parallel to the stock market.

Not a recommendation that anybody play.

[steve2470]

2. Here is what I found at Intrade:

Barack Obama to be re-elected President in 2012

53.7% CHANCE
Last prediction was: $5.37 / share
Today's Change: +$0.07 (+1.3%)
Contract Type: 0-100
Event: 2012 Presidential Election Winner (Individual)

http://www.intrade.com/v4/markets/contract/?contractId=743474

[dimbear]

4. It's a little hard to interpret, because if you add Obama's chances to Romney's chances, you don't

quite get to 100 per cent. I don't know who else they expect to drop down from a cloud.

[SoutherDem]

6. I think the missing votes are undecided.

Some polls list them as undecided others just don't talk about them.

[jenmito]

8. Here's Intrade's electoral college map:

http://electoralmap.net/2012/intrade.php The states in gray have "No Intrade Data," but you can pretty much figure which ones will go to Obama (like IL, for example). They also have a place you can click to create your own map. It's fun.

[Drunken Irishman]

3. An average of all the polls should be pretty accurate...

It helps weed out the outlier polls, whether positive/negative for Obama/Romney. I suspect the Bloomberg poll is probably an outlier.

RCP puts its average at 2.2% in favor of Obama. I think that's about right, maybe a bit more in favor of Obama. Unfortunately, as odd as it might sound, we've had a lack of polls the past month. Outside Bloomberg, the only polls released since the very first of June are the daily tracking polls, which have their own problems.

[SoutherDem]

7. I agree

RCP at least doesn't rely on one organization who may have a bias. They probably discard outliers which is common in statistics.

[demwing]

13. Before the Bloomberg poll, Obama was only up 0.6%

A month ago he was up 1.5%, but has been losing points each week. The 2.2% lead that shows currently is almost entirely dependent on the Bloomberg "outlier" lead.

[SoutherDem]

15. I was taught in both basic and advance statistics

to drop outliers, unless you had a reasonable purpose for the drastic difference.

But never the less I am not placing much trust in the Bloomberg poll, nor any national poll for that matter for two reasons.

It is too soon to start being concerned about polls except for talk on websites after the conventions maybe I normally don't start celebrating or sweating until late September or early October.

It is the electoral college is what that really matters I took the current average polls of each swing state along with the states RCP has deemed leaning or strong applied that to one of the interactive maps on the internet and Obama won with 303. I am not saying it is in the bag, but I am showing that a 2.2 up for Obama on the national poll still equates to a good win. All of these national polls don't matter.

[SoutherDem]

5. The one held on November 6th

But, realclearpolitics.com has a summary giving one number, today it is Obama +.2.2, they take several polls and combined them.

I am not sure if any one polling company or one poll should be given too much confidence or despair.

Also, who cares if Obama wins by 1% or 13%, it is the electoral college which will elect the president most news networks websites have a map with their current predictions. Most are rather close, the only difference is some count leaning states in their totals while some keep them as toss ups. You can also play around to see some what if scenarios.

Lastly, we have 4.5 months till the election. So much could change between now and then, so the best poll in the world today will be ancient history by then. I don't give much stock in any poll till late September or early October, certainly not before the conventions.

These are just my opinion others may and most likely will disagree, it is DU and we are Democrats.

[steve2470]

9. anyone know about the cell phones ? thanks nt

[jenmito]

10. Every poll I see DOES call cell phones. Gallup, for example, states:

Does Gallup call cell phones?

Gallup includes cell phones in each national Gallup poll. Gallup has been including cell phone-only households in all national telephone Gallup polls since January 2008. Further, cell phone-only households are now as likely to fall into national Gallup polls samples as those living in traditional landline households.

Gallup and the survey research industry as a whole are exceedingly aware that increasing numbers of Americans use cell phones in addition to traditional landline phones and that more than 20% of households have only cell phones. Gallup's methodologists and survey scientists have studied, researched, conducted test polling; analyzed existing data; and examined the implications of the inclusion or exclusion of cell phones and determined that it is necessary to include cell phone numbers in survey samples despite the increased costs of doing so to ensure that the obtained samples continue to be representative of all Americans. This is particularly important given recent studies that show that people in cell phone-only households tend to be younger, are more likely to be racial and ethnic minorities, and more likely to be transient.

Gallup methodologists continue to consider and study new approaches to cell phone sampling to refine and improve upon current techniques in use.

In addition, recent annual conferences of the American Association for Public Opinion Research have included multiple scientific sessions dealing specifically with the cell phone issue. The organization and Gallup methodologists continue to consider and study new approaches to cell phone sampling to refine and improve upon current techniques in use.

http://www.gallup.com/poll/110383/Does-Gallup-call-cell-phones.aspx

[steve2470]

11. ok, thanks, I'm very out of the loop lol nt

[jenmito]

12. No problemo! n/t

[hrmjustin]

14. i like ppp. n/t