Reverse Bradley Effect? Barack Obama continues to out perform polls

[grantcart]

A closer examination of the polls taken immediately before the elections and the actual election results shows that 90% of the time Sen. Obama significantly out performs the polls taken right before the election, by an average of 6.8% of the time

Nineteen states were chosen based on a)primaries (not caucuses) b)having significant polling done before the election. The two states where Obama under performed were CA and NH where polls were very volatile. If these 2 polls are taken out the 17 states where he out performed average 6.8% higher than the polling. (With all 19 states + 6.2%)

The reasons for that include the following:

1) Obama does very well with younger supporters who may no longer use landline telephones that are a staple of pollsters
2) Obama's ground support is effective in getting a high percentage of followers to participate in elections
3) Obama does very well with late deciding voters - a recent TX poll showed 60% of people deciding in the few days before the election broke for Obama
4) Obama gains in polls right as his campaiging presses closer to election day. It could be that polls are simply 3-4 days behind momentum swings

Here are the states

State . . .Polling . . .Actual . . . Difference

AL . . . . .47% . . . . .56% . . . . .+9%

AZ . . . . . 40% . . . . .42% . . . . .+2%

CA. . . . . .44% . . . . .43% . . . . .-1.5%

CT. . . . . .48% . . . . .51% . . . . .+3%

DE. . . . . .42% . . . . .53% . . . . .+11%

DC. . . . . .63% . . . . .75% . . . . .+12%

GA. . . . . .50% . . . . .66% . . . . .+16%

IL. . . . . .58% . . . . .64% . . . . .+6%

ME. . . . . .54% . . . . .60% . . . . .+6%

MA. . . . . .39% . . . . .41% . . . . .+2%

MD. . . . . .44% . . . . .49% . . . . .+5%

NH. . . . . .39% . . . . .37% . . . . .-2%

NJ. . . . . .42% . . . . .44% . . . . .+2%

NY. . . . . .36% . . . . .40% . . . . .+4%

SC. . . . . .41% . . . . .55% . . . . .14%

TN. . . . . .36% . . . . .41% . . . . .+5%

VA. . . . . .56% . . . . .64% . . . . .+8%

WA. . . . . .52% . . . . .57% . . . . .+5%

WI. . . . . .49% . . . . .58% . . . . .+9%

[C_U_L8R]

1. it has nothing to do with Bradley

and everything to do with ... US

[Drunken Irishman]

2. I think a lot of it has to do with the fact there are so many new voters.

And since they're new, they're rarely ever contacted for polls.

[NJSecularist]

3. I'm not so sure.. some of those states were underpolled

Some of those states had very little data available to conduct polling averages.

[grantcart]

6. states that didn't have several recent polls were not included

I used pollster.com and tried to include an average of the polls one week before election

http://www.pollster.com/08-VT-Dem-Pres-Primary.php

[MindMatter]

16. Or were they over-hoped? :)

[Barack_America]

4. Will we soon talk about the "Hillary Effect"?

The phenomenon of actually LOSING support once you start campaigning.

IMO, it has been fascinating to see this play out time and time again.

[Liberal Dose]

30. I can't believe I was writing to Hillary 5 years ago, begging her to run. IMO she has changed so

much since then. Edited to add: So has Bill, for that matter.

[MadBadger]

5. I think its all those new voters Obama brings in that wouldnt be considered "likely" voters

[WHAT]

19. my theory is...

the youthful Obamas are "old School" in the best sense of the word. For example, being able to answers math questions before computers. You had to understand the relationships instead of just plug in the formula...if the variables and their relationships change, the formula doesn't work. I actually think they may have used common sense to derive a formula that works...

Anyway, not schooled in this but find it very fascinating.

a Thankyou to op and also prosense (?Ithink?) who had a very good discussion on the polling, etc

I think this has become a hobby...following the twists and turns and new insights...

[tritsofme]

7. Not really an assumption you can make when there are substantial undecideds

included in the poll.

They have to break in some way, and if the poll is in any way valid, both candidates would tend to over-perform the poll.

[grantcart]

14. that would only be true if all undecideds voted. All appearences point

to Obama holding a substantial margin in late breaking deciders.

[OzarkDem]

8. He's still losing in Ohio

[jefferson_dem]

9. Hi grantcart. I just posted a thread on this same topic.

:hi:

---> http://www.democraticunderground.com/discuss/duboard.php?az=show_mesg&forum=132&topic_id=4850577&mesg_id=4850577

[grantcart]

12. lol I didn't see it

great minds think . . well no that could not be it. . .

[theboss]

10. Cell phones have destroyed pre-election polling

Exit polling still seems to be reasonably accurate but pre-election polls have been a mess for most of this century.

[WillyT]

11. K & R !!!

:kick:

[tekisui]

13. NH and CA are the only negative anomalies.

NH because a late boost from Hillary didn't have time to show up.

CA because early voting.

Go, Obama!

[Tom Joad]

15. think it has something to do with people changing their mind and going with the momentum...

[grantcart]

26. also people who are not really committed find an excuse not to vote

[BearSquirrel2]

17. I suspect weightings ...

Polls are weighted according to demographics. So you get an idea of how many hispanics support a candidate and then you weight the outcome based on historical turnout.

I think the reason the polls are coming out low is that so many people are being energized and showing up in greater numbers.

[grantcart]

20. i agree

[VolcanoJen]

18. It's the turnout model thing

Pollsters still haven't figured out how to predict turnout accurately in this historic election.

Underestimated Latino turnout burned Zogby in California. Underestimated AA and under-30 turnout burned most other pollsters in most every other Obama state. :-)

For the record, what I'm seeing on the ground in Ohio isn't being caught by any polls.

[grantcart]

23. Also how do you factor 1,000,000 phone calls or 1,000,000

door knock. They can't factor in the enthusiasm factor.

[saltpoint]

21. Very useful post and good insight.

[grantcart]

24. tks it will be very interesting to see tomorrow if it holds true again

[Scurrilous]

22. K & R

:thumbsup:

[grantcart]

25. kick

[democrattotheend]

27. I have a bad feeling the opposite may happen tonight

Given the late-breaking trends in the polls.

Does anyone have exits yet?

[grantcart]

29. remember exits wont have early voters

[awaysidetraveler]

28. Nice post! I've noticed the same thing.

And I ran a post to this effect a few weeks ago, with a similar 6% bias against Obama.

If it's the case, then we may have a candidate by the end of the night.