Barrett campaign: Latest guv poll is "a stretch"

Unsurprisingly, U.S. Rep. Gresham Barrett's gubernatorial campaign takes a pretty dim view of the reliability of the poll I wrote about in my column today.

They have a point. Little has happened from the beginning of December, when Barrett was tied for first place with McMaster in two polls, to Dec. 16, when the most recent poll showed weaker numbers for him than McMaster, Bauer or even Nikki Haley, a state representative. 

B.J. Boling, the affable Barrett spokesman whom overattentive readers will remember from John McCain's (successful) S.C. primary campaign, sent a note today explaining the flaws the Barrett camp sees with the Insider Advantage methodology. Definitely worth a read - and remember my column's caveat, that we're so far out from the primary that all this is just idle speculation anyway. The governor's race won't begin in earnest until after the New Year (though I bet it gets real good, real quick).

B.J.'s letter in its entirety appears below the jump.

---------------------------------------------------------------------------------------------------------------

Hey Rob –

Hope all is well. Read the polling piece from this morning’s paper and wanted to share a couple of thoughts…

1.       This poll was conducted using robo calls. That’s a process where a machine makes a call to a phone number and proceeds to poll whoever answers. So, my eight year old nephew could have responded to this poll and been counted the same as an actual voter. To say its findings are scientific would be quite a stretch.

2.     Insider Advantage did not consider geography in their poll. Since McMaster and Bauer are statewide elected officials, and Barrett’s vote is concentrated in one congressional district at this point, getting the regional balance right is absolutely critical to getting the numbers right.  But Insider Advantage failed to take that into consideration.  Consequently IA shows an extraordinarily popular congressman (Barrett) trailing a state house member (Haley) on the Republican gubernatorial ballot.   That doesn’t make any sense to me, and it conflicts with other polling.  A Republican primary survey we conducted last February showed the top three candidates tightly bunched together, with Bauer at 25 percent, McMaster at 22 percent, and Barrett at 21 percent, with 32 percent undecided.  Nothing has happened since to change the overall picture that this is a three-person, wide-open race.

I also asked our pollster, Whit Ayres, to put together a few thoughts on the Insider Advantage poll. Please see below:

 

MEMORANDUM

 

TO:           Interested Parties

FROM:     Whit Ayres

DATE:      December 23, 2009

RE:           Insider Advantage SC Republican Primary Survey

Insider Advantage, an Atlanta firm, published a South Carolina survey this week that included gubernatorial results from the respondents who said they intend to vote in the Republican primary.  The poll showed Henry McMaster and Andre Bauer tied at 22 percent, with Nikki Haley in third place with 13 percent, Gresham Barrett in fourth place with 9 percent, Larry Grooms in fifth place with 6 percent, and 28 percent undecided. 

In contrast, a Republican primary survey our firm conducted last February, before Haley and Grooms entered the race, showed the top three candidates tightly bunched together, with Bauer at 25 percent, McMaster at 22 percent, and Barrett at 21 percent, with 32 percent undecided.

The Insider Advantage results look curious to me.  I find it hard to believe that Nikki Haley, a state representative, has almost fifty percent higher standing on the ballot than Gresham Barrett, a very popular U.S. congressman who enjoyed a 75 to 9 percent favorable-to-unfavorable rating in his congressional district in our February poll.  I also find it hard to believe that Haley is tied with McMaster among Independents who intend to vote in the Republican primary.

I have asked Insider Advantage for more details about their sample, and I have not yet heard back from them.  But I strongly suspect that their results were driven by two factors:

1.      Under-representation of 3rd Congressional District Republican primary voters.

Insider Advantage uses automated phone calls (press 1 if McMaster, 2 if Bauer, etc.) and, at least in past South Carolina surveys, they did not pick up county or zip code.  If that is the case here, they have no idea how many calls come from the 3rd District versus the rest of the state.  If you do not know the county from which the calls come, you can not possibly know whether or not you have a sample that reflects likely Republican primary turnout.  Congressman Barrett's support, at this point, is heavily concentrated in his congressional district, so primary results are strongly related to region.  The crosstabs Insider Advantage presented with their survey did not include any regional or geographic breakouts, which suggests they did not collect that data.  Their primary results probably over-represent the midlands and under-represent the 3rd district.

2.      Under-representation of religious conservative voters.

Data for this survey was collected in only one night, Wednesday, December 16, which is not good polling practice in any event.  Where are religious conservative voters on a Wednesday evening during Advent?  More likely at church than at home to receive a survey call.

Other pollsters whose expertise and judgment I respect have raised serious questions about Insider Advantage's polling.  One example is available here:

http://www.pollster.com/blogs/the_insider_advantage_crosstab.php.

The bottom line is that the Republican nomination for governor in South Carolina is wide open, and no candidate has a significant lead.  The contest will go to the candidate with the best message, the best fund-raising, and the best campaign. 

---

I hope you would take these thoughts into consideration should you write anymore on this particular poll or any others conducted by Insider Advantage.

Thanks,

B.J. Boling

Communications Director

Barrett for Governor

---------------------------------------------------------------------------------------------------------------

Hey Rob –

Hope all is well. Read the polling piece from this morning’s paper and wanted to share a couple of thoughts…

1.       This poll was conducted using robo calls. That’s a process where a machine makes a call to a phone number and proceeds to poll whoever answers. So, my eight year old nephew could have responded to this poll and been counted the same as an actual voter. To say its findings are scientific would be quite a stretch.

2.     Insider Advantage did not consider geography in their poll. Since McMaster and Bauer are statewide elected officials, and Barrett’s vote is concentrated in one congressional district at this point, getting the regional balance right is absolutely critical to getting the numbers right.  But Insider Advantage failed to take that into consideration.  Consequently IA shows an extraordinarily popular congressman (Barrett) trailing a state house member (Haley) on the Republican gubernatorial ballot.   That doesn’t make any sense to me, and it conflicts with other polling.  A Republican primary survey we conducted last February showed the top three candidates tightly bunched together, with Bauer at 25 percent, McMaster at 22 percent, and Barrett at 21 percent, with 32 percent undecided.  Nothing has happened since to change the overall picture that this is a three-person, wide-open race.

I also asked our pollster, Whit Ayres, to put together a few thoughts on the Insider Advantage poll. Please see below:

 

MEMORANDUM

 

TO:           Interested Parties

FROM:     Whit Ayres

DATE:      December 23, 2009

RE:           Insider Advantage SC Republican Primary Survey

Insider Advantage, an Atlanta firm, published a South Carolina survey this week that included gubernatorial results from the respondents who said they intend to vote in the Republican primary.  The poll showed Henry McMaster and Andre Bauer tied at 22 percent, with Nikki Haley in third place with 13 percent, Gresham Barrett in fourth place with 9 percent, Larry Grooms in fifth place with 6 percent, and 28 percent undecided. 

In contrast, a Republican primary survey our firm conducted last February, before Haley and Grooms entered the race, showed the top three candidates tightly bunched together, with Bauer at 25 percent, McMaster at 22 percent, and Barrett at 21 percent, with 32 percent undecided.

The Insider Advantage results look curious to me.  I find it hard to believe that Nikki Haley, a state representative, has almost fifty percent higher standing on the ballot than Gresham Barrett, a very popular U.S. congressman who enjoyed a 75 to 9 percent favorable-to-unfavorable rating in his congressional district in our February poll.  I also find it hard to believe that Haley is tied with McMaster among Independents who intend to vote in the Republican primary.

I have asked Insider Advantage for more details about their sample, and I have not yet heard back from them.  But I strongly suspect that their results were driven by two factors:

1.      Under-representation of 3rd Congressional District Republican primary voters.

Insider Advantage uses automated phone calls (press 1 if McMaster, 2 if Bauer, etc.) and, at least in past South Carolina surveys, they did not pick up county or zip code.  If that is the case here, they have no idea how many calls come from the 3rd District versus the rest of the state.  If you do not know the county from which the calls come, you can not possibly know whether or not you have a sample that reflects likely Republican primary turnout.  Congressman Barrett's support, at this point, is heavily concentrated in his congressional district, so primary results are strongly related to region.  The crosstabs Insider Advantage presented with their survey did not include any regional or geographic breakouts, which suggests they did not collect that data.  Their primary results probably over-represent the midlands and under-represent the 3rd district.

2.      Under-representation of religious conservative voters.

Data for this survey was collected in only one night, Wednesday, December 16, which is not good polling practice in any event.  Where are religious conservative voters on a Wednesday evening during Advent?  More likely at church than at home to receive a survey call.

Other pollsters whose expertise and judgment I respect have raised serious questions about Insider Advantage's polling.  One example is available here:

http://www.pollster.com/blogs/the_insider_advantage_crosstab.php.

The bottom line is that the Republican nomination for governor in South Carolina is wide open, and no candidate has a significant lead.  The contest will go to the candidate with the best message, the best fund-raising, and the best campaign. 

---

I hope you would take these thoughts into consideration should you write anymore on this particular poll or any others conducted by Insider Advantage.

Thanks,

B.J. Boling

Communications Director

Barrett for Governor