A question about PPP Polling.

[slick8790]

What's the deal with these guys? How reliable are they? In the past week or so they've come out with two polls that show one candidate or the other WAY out in front.

Their PA poll had Hillary +26.
Their NC poll had Obama +21.

Those are way out of line with a bunch of other polls.

So what's the deal? How reliable have they been this year? Anyone got that table with the average deviation from the real results for all the pollsters?

Any info is much appreciated.

[slick8790]

1. Kick.

Anything?

[jenmito]

2. All I know is the same poll had Obama just +1 in NC a week or more ago when the Wright junk was in

full force. :shrug:

[slick8790]

3. Yeah, it just seems like they swing so wildly around.

Makes me wonder if they're any good at sampling.

[jenmito]

4. Sorry, but I have no idea. I want to believe them with one poll but not the other.

:)

[Teaser]

14. They changed their likely voter model

They were using a low turnout model earlier, but decided that it is more likely that this year is a high-turnout election.

The LV voter models cause the wide swings.

[spinbaby]

5. How does polling work anyway?

Do they call the same people every time? I've gotten three calls from the same polling outfit in Nevada, but no one else I know has had any calls.

[grantcart]

7. different methods some use people some have computers calling

asking respondents to push buttons - then you have the question of modeling - how many young people how many independents etc

[grantcart]

6. all i could find

http://www.politico.com/blogs/bensmith/1207/Two_polls_and_a_poll_of_pollsters.html





but this is Iowa only and that is a caucus situation so I wouldn't give it much credence

http://www.publicpolicypolling.com/pdf/PPP_Release_032508.pdf



Here you could check their polls against results

http://www.publicpolicypolling.com/surveys.asp?@spdT=ECF791E062CD404AAB75

[FlyingSquirrel]

11. Highest percentage (64%) didn't know PPP and skipped it.

Personally I just use the average of all polls and it works pretty well for me.

[Arkansas Granny]

8. I use the polls at Real Clear Politics. They give an average of national polls

and it's easy to click through to the various websites for detailed information about the polls. Even then, there is a lot of variation between the polls they list. I also like to see if they have polled likely voters (LV), registered voters (RV) or just spoke to an adult (A) who responded to the questions.

http://www.realclearpolitics.com/

[slick8790]

9. Yeah, but still.

If you take out the PPP polls out of the averages, it drops everybody pretty significantly, 3-4 points.

[Arkansas Granny]

10. I'm not sure what the PPP polls are. Do you have a link?

[slick8790]

12. Sure do.

NC - Obama +21 http://www.publicpolicypolling.com/pdf/PPP_Release_032508.pdf
PA - Hillary +26 http://www.publicpolicypolling.com/pdf/PPP_Penn_Release_031708.pdf

[Arkansas Granny]

13. Thanks for the info.

[Adelante]

15. They've made a sampling change, accounting for some of the wide lead

See here

http://www.democraticunderground.com/discuss/duboard.php?az=show_mesg&forum=132&topic_id=5253286&mesg_id=5259078

[Upton]

16. They were close in Texas and Ohio

http://publicpolicypolling.blogspot.com/2008/03/ppp-best-combined-democratic-numbers-in.html

[MadBadger]

17. They were on target in Wisconsin, Ohio, and Texas, the last three races they polled.

[Bullet1987]

18. Go to pollster.com and look at PPP recent history in NC

It's typically been very close like last weeks poll. So this recently released NC showing Obama way ahead is actually not normal for them.