Dems go after pollsters?

Unbelievable. From Politico

Democrats are turning their fire on Scott Rasmussen, the prolific independent pollster whose surveys on elections, President Obama’s popularity and a host of other issues are surfacing in the media with increasing frequency.

The pointed attacks reflect a hardening conventional wisdom among prominent liberal bloggers and many Democrats that Rasmussen Reports polls are, at best, the result of a flawed polling model and, at worst, designed to undermine Democratic politicians and the party’s national agenda.

On progressive-oriented websites, anti-Rasmussen sentiment is an article of faith. “Rasmussen Caught With Their Thumb on the Scale,” blared the Daily Kos this summer. “Rasmussen Reports, You Decide,” the blog Swing State Project recently headlined in a play on the Fox News motto.

“I don’t think there are Republican polling firms that get as good a result as Rasmussen does,” said Eric Boehlert, a senior fellow with Media Matters, a progressive research center. “His data looks like it all comes out of the RNC [Republican National Committee].”

“Whether intended or not, Rasmussen polls have been used by conservative voices as talking points, and when that happens on one side it inevitably produces a reaction from the other,” explained Mark Blumenthal, a polling analyst and the editor and publisher of Pollster.com. “Rasmussen produces a lot of data that appear to produce narratives conservatives are promoting, and that causes a reaction.”

While Scott Rasmussen, the firm’s president, contends that he has no ax to grind — his bio notes that he has been “an independent pollster for more than a decade” and “has never been a campaign pollster or consultant for candidates seeking office” — his opponents on the left insist he is the hand that feeds conservative talkers a daily trove of negative numbers that provides grist for attacks on Obama and the Democratic Party.

Leave it to the hacks like David Brock's (the hack) "Media matters" to be idiots. If the methodology was challenged, that's one thing. This is just bitching about the numbers.

Now I don't always like the polls, but always remember one thing with a polls. At its best, it's a snapshot in time of the current political situation. It also does not measure intensity of the views. People's opinions can change.

As far as pollsters goes, It's always good to know the bias behind them and factor that into the race. That's why I always look at multiple polls. If four pollsters say the same thing, it's probably close. There's the DailyKos, Penn, Field, and NY Times poll on the left. There's Market Research and ARG on the right. There's survey usa, Gallup, Rasmussen, USAToday, Fox, and CNN in between. Mitchell and EPIC/MRA are right and left, but sometimes actually have the most unfavorable numbers for their side.

Real Clear Politics shows the most recent polls regarding job approval: First numbers are dates, then polled, then whether adults, likely voters, or registered voters are polled. One thing I noticed different with Rasmussen is that he polled only likely voters. Only Battleground (Bipartisan with one republican and one democrat) also did that.

Gallup	12/28 - 12/30	1547 A	51	42	+9
Rasmussen Reports	12/28 - 12/30	1500 LV	46	53	-7
Quinnipiac	12/15 - 12/20	1616 RV	46	43	+3
CNN/Opinion Research	12/16 - 12/20	1160 A	54	44	+10
NBC News/Wall St. Jrnl	12/11 - 12/14	1008 A	47	46	+1
Associated Press/GfK	12/10 - 12/14	1001 A	56	42	+14
USA Today/Gallup	12/11 - 12/13	1025 A	49	46	+3
ABC News/Wash Post	12/10 - 12/13	1003 A	50	46	+4
Pew Research	12/9 - 12/13	1504 A	49	40	+9
Battleground	12/6 - 12/10	1000 LV	50	45	+5
FOX News	12/8 - 12/9	900 RV	50	44	+6

Now is Rasmussen accurate? At times he's been inaccurate (2000), but I know he changed his style after that. In 2008, these were the final polls.

Marist	11/03 - 11/03	804 LV	4.0	52	43	Obama +9
Battleground (Lake)*	11/02 - 11/03	800 LV	3.5	52	47	Obama +5
Battleground (Tarrance)*	11/02 - 11/03	800 LV	3.5	50	48	Obama +2
Rasmussen Reports	11/01 - 11/03	3000 LV	2.0	52	46	Obama +6
Reuters/C-SPAN/Zogby	11/01 - 11/03	1201 LV	2.9	54	43	Obama +11
IBD/TIPP	11/01 - 11/03	981 LV	3.2	52	44	Obama +8
FOX News	11/01 - 11/02	971 LV	3.0	50	43	Obama +7
NBC News/Wall St. Jrnl	11/01 - 11/02	1011 LV	3.1	51	43	Obama +8
Gallup	10/31 - 11/02	2472 LV	2.0	55	44	Obama +11
Diageo/Hotline	10/31 - 11/02	887 LV	3.3	50	45	Obama +5
CBS News	10/31 - 11/02	714 LV	--	51	42	Obama +9
ABC News/Wash Post	10/30 - 11/02	2470 LV	2.5	53	44	Obama +9
Ipsos/McClatchy	10/30 - 11/02	760 LV	3.6	53	46	Obama +7
CNN/Opinion Research	10/30 - 11/01	714 LV	3.5	53	46	Obama +7
Pew Research	10/29 - 11/01	2587 LV	2.0	52	46	Obama +6

The final number was Obama +7. Not dead on, but he was close. 1 pt off is real good and within the margin of error. If you go with the runup before the election, there were several 9-10-11% leads for Obama, and it wasn't that far apart.

2004 Rasmussen was very close.

GW/Battleground (1000 LV)	10/31 - 11/1	50%	46%	Bush +4
Rasmussen (3,000 LV)	10/30 - 11/1	50.2%	48.5%	Bush +1.7
TIPP (936 LV)	10/30 - 11/1	46.9%	44.3%	Bush +2.6
FOX News (1200 LV)	10/30 - 10/31	46%	48%	Kerry +2
CNN/USAT/Gallup (1573 LV)	10/29 - 10/31	49%	47%	Bush +2
CBS/NY Times (643 LV)	10/28 - 10/30	50%	47%	Bush +3
ARG (1258 LV)	10/28 - 10/30	48%	49%	Kerry +1
Newsweek (882 LV)	10/27 - 10/29	51%	45%	Bush +6

TIPP was the champ by margin, but Rasmussen was only 1/2 a percent off of Bush's numbers and .3% off on Kerry's numbers. That's good work for a pollster. One thing good about the Real Clear Politics site is that they show a large number of the polls and average them, and also compare them.

State polls are tougher, but here's the polls here. The final result was McDonnell by 17 and Christie by 4 with the independent taking almost 6%

SurveyUSA	10/30 - 11/1	574 LV	58	40	McDonnell +18
Times-Dispatch/MD	10/28 - 10/29	625 LV	53	41	McDonnell +12
Rasmussen Reports	10/27 - 10/27	1000 LV	54	41	McDonnell +13
Daily Kos/R2000	10/26 - 10/28	600 LV	54	44	McDonnell +10
Suffolk University	10/26 - 10/28	400 LV	54	40	McDonnell +14

Monmouth/Gannett	10/31 - 11/1	722 LV	41	43	8	Corzine +2
SurveyUSA	10/30 - 11/1	582 LV	45	42	10	Christie +3
Quinnipiac	10/27 - 11/1	1533 LV	42	40	12	Christie +2
Rasmussen Reports	10/29 - 10/29	1000 LV	46	43	8	Christie +3
Stockton/Zogby	10/27 - 10/29	1093 LV	39	40	14	Corzine +1

He was off in Virginia (towards the dem), but close in Jersey. One thing about pollsters, you can always judge them by their final results.

Politico also said this.

Last year, the progressive website FiveThirtyEight.com’s pollster ratings, based on the 2008 presidential primaries, awarded Rasmussen the third-highest mark for its accuracy in predicting the outcome of the contests. And Rasmussen’s final poll of the 2008 general election — showing Obama defeating Arizona Sen. John McCain 52 percent to 46 percent — closely mirrored the election’s outcome. 

Rasmussen, for his part, explained that his numbers are trending Republican simply because he is screening for only those voters most likely to head to the polls — a pool of respondents, he argues, that just so happens to bend more conservative this election cycle.

Polling all adults — a method used by Gallup, another polling firm that conducts a daily tracking poll of Obama — Rasmussen acknowledged, is “always going to yield a better result for Democrats.”

 

Now the left isn't the only one to complain about the numbers. 

Last month, conservative radio host Rush Limbaugh accused the Gallup polling organization of “doing everything they can — they're upping the sample to black Americans — to keep” Obama’s approval at 50 percent.

And PPP, which has also increased its profile in the past two election cycles, has drawn criticism from Republicans for repeatedly showing North Carolina GOP Sen. Richard Burr with low ratings heading into his 2010 reelection.

PPP IS democrat and that should be kept in mind, and Gallup's numbers are more friendly to the dems. Is it biasness looking for a result, bad methodology, or just educated guess on the samples? 

Gallup and Rasmussen are the most opposite of recent polls. Gallup has Obama +9 and Rasmussen -7. 

Rasmussen - Likely Voters - 1500 

Gallup - Adults  - 1547

Gallup by demographics there - I don't see the weights to the responses however

Crosstabs are for premium members on Rasmussen's site.

The main thing I try to find out with polls, is the weight (Party ID). That usually determines how far left/right it leans. If the methodology is wrong, that's one thing. If the weight is wrong, that's one thing. If the numbers are unfriendly, that's just sour grapes. 

Most importantly though is that in the end one poll matters, and outside of vote fraud, it's the most accurate one. Election Day.