McDaniel Leads DeLay in Attorney General Race
Little Rock's KTHV has commissioned a poll in the Attorney General's race that shows Democrat Dustin McDaniel ahead of Republican Gunner DeLay 50-35% with 14% undecided.
Is the poll accurate, and how firm are those numbers? The answers may surprise.....
(CONTINUED, click "Wednesday" below and scroll down for rest of article, or if sent straight here just scroll down.)
posted by Mark Moore (Moderator) at Wednesday, July 26, 2006
15 Comments:
With my science background, I am not one to discount polls. However, any research is only as solid as the methods that were used to collect data. This is the second Survey USA poll this week, and a reasonable person could suspect taht the poll methodology is tilted in favor of the Democrats. This is not simple partisan grousing. There are at least three solid reasons I come to this conclusion.
1) Survey USA yesterday released a poll on the Governor's race that had the Democrat up by 10, 48-38. Yet two more polls on that same race had been taken at roughly the same time, and they had the gap at only 4 and 7 points respectively. When one poll has the gap at about twice the size of the average of the other two (5.5%), it is a signal something may be wrong.
2) The Republican Base for a credible candidate is larger than the numbers both Hutchinson and DeLay are showing- and they are both credible. I realize that there are undecideds out there, and some of them will gravitate to the Republicans, but the "credible candidate" base vote for the GOP is now widely believed to be 40-42%. Those are the people that tend to decide early and reflexively to vote Republican (the Democratic base in Arkansas is a few percent larger, leaving only a small slice of true swing voters).
3) The last Republican to run for statewide office (Holt in 04) got 44% of the vote against an incumbant despite being outspent 40-1 in relative terms and outspent by 6.6 million dollars in absolute terms.
Neither DeLay nor Hutchinson is running against an incumbant. They are both running against weaker candidates than Lincoln, McDaniels is much weaker. Not even Bebee has that kind of money to throw at the race. The numbers don't add up. I suspect their methodology is flawed.
The numbers broke down that DeLay is behind 46%-44% among men, and an astonishing 54-27 gap among women voters. Hutchinson had a deficit with women voters that was almost as sharp.
This gives us a clue as to what might be wrong. Happily married women with children have tended to vote GOP recently. Single women tend to vote Democrat. Is the poll taking this into account or are they counting a "woman is a woman" and somehow polling a disproportionate share of single women?
Maybe the mom's are all out at ball games and dance classes in the evenings and don't get polled as much. I am just guessing here, but something is fishy.
As far as age group goes, DeLay is winning among 18-34 year olds and losing among the rest. That may be a good sign for the GOP's future, but it does not look good for the 04 AG race.
If the numbers are accurate, are they firm? No way. There is no way a fellow with McDaniel's thin resume, indeterminate charisma, and record of a family that sued gun and ammo manufacturers can hold a lead with male voters. Not in Arkansas.
DeLay is going to have to go negative to win, but he needs to do it in a way that does not turn away female voters. I suggest using the GOP's data machine to find households where guns is an issue and send a mailer that has information about the lawsuit on the outside of the envelope. Even if they throw it away, they see it. And the households that don't care about the issue much never see the attack. They can see commercials of Gunner holding a baby.
Now DeLay is right when he points out that McDaniel just spent $700,000 in a primary while he has not yet spent anything, but it's not like he has his own $700,000 lying around unused. Plus, McDaniel can cut into the GOP base some with his background as a policeman- a subculture that tends to stick together.
In conclusion, there is good reason to suspect that the polling methods for the Survey USA poll are flawed, but flawed in a way that is not detectable in a party primary (since their primary polls were accurate). Failure to realize that the "woman vote" is two groups and making sure each is polled proportionally would produce just the sort of error we are predicting. In other words, it would not be a factor in primary voting, since each woman voter would be sorted by what primary they tended to vote in, but would produce innaccurate results in a general election poll.
Also, McDaniel's numbers are not firm, but DeLay is going to have to mount an effective targeted negative campaign to win this race.
This is the same group that predicted Formicola to win the primary. We saw how badly off they were.
It has also been suggested to me that Survey USA may have over-polled the 2nd district in order to get enough voters to have a good sample on the 2nd district race, whose results will be announced tonight. If so, that under-represents the most populous 3rd district, from where both Asa and Gunner, (and Jim Holt) hail.
We will see when they release their poll tonight- all these polls were done at the same time. How many voters did they include in the 2nd district poll? If is is more than 509 (their statewide total) divided by 4 then it is more likely that they over-polled the liberal 2nd. It means they wanted a good statistical sample in that district, so their top priority was getting enough polling data from there, even if it slanted the state-wide races. I am repeating a theory that was advanced to me.
Excellent Excellent point.
Or it could just be that your boy is 15 points down. I hate polling as much as the next guy but even if this one is off a bit, gunner still has a long way to go. If it were the other way around, you guys would be calling for Dustin to concede.
Robbie you have a point there, but what about the fact that the same poll was out of line with the other two polls done about the same time as the Governor's race?
Do you disagree with the idea that married women with kids tend to vote GOP while single women overwhelmingly vote Democrat? There are REASONS why we think the poll is off. It is not just grousing. You either think this poll is off, or the other two are.
Can you take issue with any of those reasons?
A true 4 district sample would be needed to be correct. A republican can win by simply running close in the 1st and winning hugh in the 3rd
If a poll was taken only in the 2nd and 4th we all know the ole yellow dog wins
They used 473 voters on the second district poll and only 500 and something for the whole state poll. If they did them on the same calls it is easy to think that they over-represented the 2nd.
As you've said before, KTHV is usually right on.
6:38, is that the only thing you heard after all that analysis? "the poll is usually right on"?
Or, Its just not a good year to be a Republican. This whole thing goes in cycles, and being an (R) is just not as good as it was 2 years ago.
Im saying its gonna be a Democratic sweep statewide. Although, the R's will probably pick up a few house seats.
Or, Its just not a good year to be a Republican. This whole thing goes in cycles, and being an (R) is just not as good as it was 2 years ago.
Im saying its gonna be a Democratic sweep statewide. Although, the R's will probably pick up a few house seats.
I think we need to do a thread next week on House Races.
This site is one of the best I have ever seen, wish I had one like this.
»
I like it! Keep up the good work. Thanks for sharing this wonderful site with us.
»
Post a Comment
<< Home