PDA

View Full Version : American Research Group - low polling #s for Ron Paul




Nathan Hale
12-10-2007, 07:17 PM
I check www.realclearpolitics.com every day to see the new polls as they come in. I've been following it daily for about a month now, and I've noticed a pattern. While Paul's numbers jump up and down with a bit of randomness (there is a 4% average margin of error in all polls), one company in particular always shows Paul polling WAY below his average result. That company is the American Research Group.

Nationally, ARG's last poll put him at 4%, which isn't excessively low considering that Fox put him at 3% soon after. But let's drill down to the early primary states.

In Iowa, ARG put him at 3% on 11/29. What were the other Paul pollings from that week? A slew of 5% results and one 7% from the Des Moines register.

In NH, ARG has him at 2% as of their last poll on 11/29, which is the lowest he's been since the ARG's LAST poll where he came in at 1%. The other Paul poll results surrounding ARG's 2%? A slew of 8%, a 9%, a 4% (Fox), and a 6%.

In Nevada, ARG has Paul at 3% on 12/6. The other recent polls? 5%, 7%, 8%.

In SC, ARG had him at 3% when everybody else was reporting 6%.

What's the deal with ARG?

Nathan Hale
12-11-2007, 08:56 PM
Anybody else notice this? Any information out there? I know this isn't as sexy as FEMA conspirators and the evil MSM, but I figured at least somebody would have some info.

DRV45N05
12-11-2007, 08:58 PM
ARG has sucked for a long time at polling.

dspectre
12-11-2007, 09:01 PM
Forget the old media's polls.

saahmed
12-11-2007, 09:01 PM
I have noticed that as well. They must be polling differently. Perhaps leaving his name off the polls or only giving a few names and then asking you to press a number to get the other names.

Nathan Hale
12-12-2007, 09:44 PM
Forget the old media's polls.

Not a wise idea, friend.

10thAmendmentMan
12-12-2007, 09:48 PM
My guess would be their polling practices are biased against Paul supporters -- whether it be that they take samples from older, less internet-savvy respondents or target dyed-in-the-wool Bush Republicans.

hawkeyenick
12-12-2007, 09:50 PM
Also, where the hell did this "mason-dixon" poll come from?

They randomly add and remove polls from the average whenever they like.

Ron LOL
12-12-2007, 09:51 PM
ARG isn't just their name. It's the sound I exclaim whenever they release a new poll :)

The Only Woj
12-12-2007, 09:52 PM
the only poll that matters is the one each state will have to nominate a candidate ... and it won't be some bullshit phone poll, but a show up and vote poll.

Nathan Hale
12-13-2007, 10:00 AM
My guess would be their polling practices are biased against Paul supporters -- whether it be that they take samples from older, less internet-savvy respondents or target dyed-in-the-wool Bush Republicans.

I disagree. There's no evidence that this is a bias situation. Perhaps there is an unintentional slant, but there's no bias. It's simply easier to do a random sample of, as an example, landline phones, because there is no directory of cellular phones to work with.

Nathan Hale
12-13-2007, 10:01 AM
Also, where the hell did this "mason-dixon" poll come from?

They randomly add and remove polls from the average whenever they like.

I think that the realclearpolitics average includes the shaded polls at the top of the list. IT's a good system if thats the case, as it keeps enough data points in play to compute a genuine average, while simultaneously filtering out old data for the average is up to date.

Nathan Hale
12-13-2007, 10:03 AM
the only poll that matters is the one each state will have to nominate a candidate ... and it won't be some bullshit phone poll, but a show up and vote poll.

Yes, at the end of the day that's what matters.

However, these polls are useful. They do filter out people unlikely to vote, so they are a fairly accurate indicator of how people who vote think.

werdd
12-13-2007, 10:04 AM
The polls are just subliminal brainwash much like the magical knob debate sentiment mind change graphs VIA FRANK luntz.

FUCK YOU FRANK.

These are the devices frank uses.

http://www.estarland.com/images/products/62/22962/39133.jpg

10thAmendmentMan
12-13-2007, 10:07 AM
I disagree. There's no evidence that this is a bias situation. Perhaps there is an unintentional slant, but there's no bias. It's simply easier to do a random sample of, as an example, landline phones, because there is no directory of cellular phones to work with.

You just described a sampling bias. A statistical bias can be unintentional and I used the term in the same way you used "slant." I wasn't suggesting that they're trying to cook their numbers. :)

Bigboyen
12-13-2007, 10:10 AM
I read how they did it in Iowa and I guess they do the same in other states:

The call randomly (not after registration voter lists).

From pollster.com:

"Asks four screen questions:

They ask whether respondents are registered to vote, and whether they are registered as Democrats or Republicans. Non-registrants are terminated and not interviewed.

They ask registrants how likely they are to participate in the Caucus "a 1-to-10 scale with 1 meaning definitely not participating and 10 meaning definitely participating." Those who answer 1 through 6 are terminated and not interviewed.

They ask unaffiliated registrants ("independents" registered as neither Democrats nor Republicans) whether they plan to participate in the Democratic or Republican caucus. Registered Democrats and independents who plan to caucus with the Democrats get the Democratic vote question; Registered Republicans and independents who plan to caucus with the Republicans answer the Republican question.

After asking vote question, they asks the question that appears on the web site: "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?" Only the definite are included in the final sample of likely caucus voters."

So I can't see anything in this metode that should give Paul lower scores then in other polls.

Elwar
12-13-2007, 10:15 AM
Not a wise idea, friend.

John McCain? Is that you?

Bigboyen
12-13-2007, 10:21 AM
They didn't disclose to polling.com on how many first time caucus goers they had (but for democrats they had 43%) and they didn't disclose the breakdown into age groups. So it can be that they are to heavy on the older voters compared to other polls.

werdd
12-13-2007, 10:32 AM
http://img91.imageshack.us/img91/3008/werdqh8.jpg

Nathan Hale
12-14-2007, 06:33 PM
You just described a sampling bias. A statistical bias can be unintentional and I used the term in the same way you used "slant." I wasn't suggesting that they're trying to cook their numbers. :)

You're right, I assumed that your use of the word "bias" implied intent. Miscommunication.

Nathan Hale
12-14-2007, 06:34 PM
John McCain? Is that you?

???