Polls: Hillary Ahead In Key Swing States; Huckabee Rising

Comments (15)

mopper wrote on August 30, 2007 10:44 AM:

Given recent trends in SC, Obama being in the low 20's strains credulity. But I guess that's why you take averages/use regression analysis. Especially given ARG's consistent tilt towards Clinton throughout the primary season thus far.

bob wrote on August 30, 2007 10:53 AM:

Jeez, hard to square with this blurb from the morning roundup:

Poll: Edwards Continues To Lead In Iowa
John Edwards' campaign is betting heavily on a strong Iowa victory in order to catapult the candidate over his opponents. So far, so good, according to a new poll from Time, showing John Edwards with 29% of likely Iowa caucus goers, 24% for Clinton, and 22% for Obama.

bob wrote on August 30, 2007 10:58 AM:

Pollster.com has SC at 38 Clinton, 31 Obama, 14 Edwards. They also have closer races in NH and especially in Iowa.

American1989 wrote on August 30, 2007 11:02 AM:

I don't think that ARG has been favorable to Clinton entirely, they had some polls some weeks ago that showed Obama tying her in New Hampshire and Obama only 2-3 points behind in South Carolina. ARG is an unreliable poll with its results being really skewed.

IloveDemsNReps wrote on August 30, 2007 11:05 AM:

The American Research Group is really looking bad. No one believes Hillary Clinton is up in Iowa, must less South Carolina. I know she is not winning New Hampshire either.

mopper wrote on August 30, 2007 12:16 PM:

Well, here's my data to back up my initial take on SC. According to pollster.com, since May, including this poll, there have been 12 relevant SC polls taken.

Of those 12, 4 have Obama at/below 25 points in SC. 3 of those 4 are from ARG.

Indeed, the breakdown goes (from oldest to most-recent):

Obama Support in SC
ARG
18
21
33
21

Non-ARG
21
34
34
25
28
33
30

So, like I said: ARG showing Obama at 21% strains credulity. Interestingly enough, ARG is also the only firm since May to show Edwards over 20, and they do so 3 out of 4 times they poll the state. Nobody else ever gets him registering over 16%, with pollster's regression analysis putting him at 13.6%

Something is off. Either they're using a vastly different likely-voter model, or they're using a vastly-different expected demographic make-up of primary voters (and thus weighting responses differently). But there does seem to be a pro-Edwards, anti-Obama bias in their polling in the state. Whether that reflects a more accurate voter-modeling or not is not really my call to make.

magster wrote on August 30, 2007 12:19 PM:

Huckabee is the most electable in a general election. He came accross as a likeable guy at the cancer forum. Go Mitt!

Zhonni wrote on August 30, 2007 12:20 PM:

I thought I was the only one that noticed that ARG is a phony.

One can look back and say, how are they biased if they had Obama tied with Clinton earlier in the year. This is to give the appearance that he peaked too sooon and is on his way down.

They have consistently had Obama in the mid to high teens when an overwhelming number has him in the mid to high 20s.

magster wrote on August 30, 2007 12:22 PM:

Also, what are we: Fox News?

Fmr. Gov. Mike Huckabee, D-AR?

IloveDemsNReps wrote on August 30, 2007 1:24 PM:

ARG loves Hillary Clinton. First, ABC/Washington Post has a poll saying 27% Obama, 26% Edwards, 26% Clinton. Then the another polling company has a three way tie. Then Time Magazine came out with a three way tie with Edwards maybe a point or two ahead. Now ARG has Clinton ahead.

Same with N. H., and S.C.,

This is ARG thinking. They want Clinton winning in Iowa because everyone knows that Edwards is probably ahead there by 1 or 2 points. Everyone knows that N.H., is and Indie State were Obama is doing well. I'm sure he is ahead there. S.C. is Barack Country, he is leading in that State.

ARG puts Clinton ahead in Iowa with Obama 2nd and Edwards 3rd. They know if Clinton wins Iowa, Edwards is gone.

They put Clinton ahead in N.H., because they know Obama wins N.H., and Edwards Iowa, Clinton is in trouble.

They put Clinton ahead in S.C., and Edwards' 2nd because they know Obama is ahead in S.C., and they want it to look like Clinton and Edwards has the black vote.

ARG is apart of the Clinton machine. Simple

stlounick wrote on August 30, 2007 1:50 PM:

Eric, would you fix the D-AR for Huckabee? It should be R-AR. If you can't, then dump the picture. Thanks.

IloveDemsNReps wrote on August 30, 2007 2:42 PM:

Even politicos in S.C., know these poll is bogus. And when did the Clinton campaign stop worry about polls.

http://campaignsandelections.com/SC/articles/?ID=590

pjv wrote on August 30, 2007 3:16 PM:

"D-Arkansas"

Why is that on his name?

lol - fox news eat your heart out. TPM - go TPM

mopper wrote on August 30, 2007 4:43 PM:

Mark Blumenthal on ARG and Time in Iowa, and ARG in general:

What is the American Research Group Methodology? All they tell us on the website is that they completed 600 interviews and that respondents were asked:

Would you say that you definitely plan to participate in the 2008 Democratic presidential caucus, that you might participate in the 2008 Democratic presidential caucus, or that you will probably not participate in the 2008 Democratic presidential caucus?

Blogger speculation alert: If this was the only question used to screen, it is likely that ARG's incidence of eligible adults was much higher. Such a difference likely explains why they show Clinton doing consistently better in Iowa than other pollsters, but that is just an educated guess.

The resolution of that argument is neither simple nor obvious, but seems to have a profound impact on the results. Surveys that appear to include more past caucus goers (Time, Des Moines Register and One Campaign survey -- see our Iowa compilation) tend to favor John Edwards, while Hillary Clinton does better on surveys that define the likely caucus-goer universe more broadly.

It seems as though ARG's sample size might be historically large. If so, they are either expecting an enormous (tripling? quadrupling?) turn-out in primaries, or are over-sampling. The difference between a primary-goer and a non-primary goer is, most likely, relevant to the point that we should expect a change in preferences. As such, I don't think its unreasonable, based on what little knowledge we have, to think that ARG is sampling from non-representative samples.

Again, as Blumenthal notes:

Is one result more trustworthy than the other? That is always a tough question to answer, but one of these polls is considerably more transparent about its methods. And that should tell us something.

pollster blog

JCT wrote on August 30, 2007 8:48 PM:

I believe this is called the "Colbert Surge" as a result of Governor Huckabee appearing on the Colbert Report on Comedy Central.

Post a comment

Name:
Email Address:
URL:
	Remember personal info?
Comments: (you may use HTML tags for style)