SC PrimR: ARG: Gingrich leads by double-digits

[Tender Branson]

New Poll: South Carolina President by ARG on 2011-11-28

Summary:

Gingrich:	33%
Romney:	22%
Cain:	10%
Paul:	8%
Perry:	8%
Other:	7%
Undecided:	12%

Poll Source URL: Full Poll Details

[Tender Branson]

Among those who say they are Definite Voters:

36% Gingrich
20% Romney
10% Cain

[CLARENCE 2015!]

This is actually not bad for Romney

[Swing low, sweet chariot. Comin' for to carry me home.]

This is actually not bad for Romney

it's a 3 man race at this point:

Gingrich: 33%
Romney: 22%
Paul: 8%
========
Total:  63%

considering where the remaining 37% is likely to land...plus the influence of the IA and NH results...Newt will likely beat Romney 2:1 in SC:

Newt     54%
Romney 27%
Paul       14%
Other       5%

[Swing low, sweet chariot. Comin' for to carry me home.]

it's a 3 man race at this point

actually, it's probably a 4 man race:  Newt/Romney/Paul/Huntsman 

[California8429]

They did have Huckabee up 7 in SC right before the primary...so ARG doesn't mean too much but is still somewhat in line with the Insider Advantage polls.

[Simfan34]

it's a 3 man race at this point

actually, it's probably a 4 man race:  Newt/Romney/Paul/Huntsman 

He has decent numbers here... above nat'l polling.

[© tweed]

so does ARG exist?

[Mr. Morden]

so does ARG exist?

My assumption is that they've been making up their polls since Super Tuesday 2008.  Their polling before that point was too awful to be fake.

[© tweed]

so does ARG exist?

My assumption is that they've been making up their polls since Super Tuesday 2008.  Their polling before that point was too awful to be fake.

has Nate Silver run an end-digit randomness regression?

[Mr. Morden]

so does ARG exist?

My assumption is that they've been making up their polls since Super Tuesday 2008.  Their polling before that point was too awful to be fake.

has Nate Silver run an end-digit randomness regression?

I don't think so, but one of us should email him and ask him to do it.

[© tweed]

so does ARG exist?

My assumption is that they've been making up their polls since Super Tuesday 2008.  Their polling before that point was too awful to be fake.

has Nate Silver run an end-digit randomness regression?

I don't think so, but one of us should email him and ask him to do it.

I'll do it.  why not?.  not only am I a paid member of Baseball Prospectus, of which Nate used to be the CEO, I am also a good writer.

[Mr. Morden]

I'll do it.  why not?.  not only am I a paid member of Baseball Prospectus, of which Nate used to be the CEO, I am also a good writer.

Go for it.  Feel free to quote from this post, to explain why ARG is so terrible:

Here are some of ARG's final pre-election polls from this year:

IA: Clinton up by 9
NH: Obama up by 9
MI: McCain up by 1
SC: Obama up by 3
SC: Huckabee up by 7
FL: Romney up by 2
CA: Romney up by 2

But those are just the final pre-election polls.  If you look back at polls they've conducted from from weeks or months in advance of the primaries, they are far worse.  I mean, there's no real way to test that, but they are usually wildly off from all the other pollsters.  For example:

Iowa:

They had Clinton leading by double digits in early 2007, while every other pollster had Edwards leading.

South Carolina:

They had several polls in the spring/summer of 2007 in which Edwards was in 2nd place in SC (behind Clinton), while every other pollster had Edwards in 3rd.

In the spring/summer of 2007, they had McCain still leading in the state, when everyone else had him trailing by double digits.  Then from about August-October, everyone had Thompson leading, but ARG had Giuliani or Romney leading.

California:

In mid-January, when every other pollster had McCain leading in CA and Giuliani in the low teens, ARG had Giuliani leading by double digits with 33% of the vote.

Those are just a few of their greatest hits.  Do any other pollsters have so many laughably bad results?

Basically, they were getting every contest backwards in the early 2008 primaries.  That is, they were on the wrong side of the polling average in nearly all the early states.  Then after Super Tuesday, they suddenly started offering more reasonable results.  So either they suddenly became competent pollsters, or they started fudging their results.