Tuesday, October 30, 2007

U of A Poll and Arkansas Republican Presidential Primary

The University of Arkansas conducted its annual poll on Arkansas values. I think they oversample towards the left since they say things like only "53% oppose homosexual marriage" when the actual number in the real marriage amendment election was more like 75%. They also threw in a so-called "equal rights amendment" question with no context in order to get uninformed folks to say it was a good idea. Still, these are quibbles. There was some good data in this poll.

The question I think many of us are interested in is the looming Arkansas Presidential Primary on February 5th. Not the Democratic side- Hillary! is a forgone conclusion in that one. Rather, the numbers on the Republican side are what people are interested in.

The poll basically lumped all respondents, Democrat, Republican, and Independent together. It did not directly measure strength against other Republicans. So in the poll Hillary! was listed with 35% of the vote, while the top three Republican vote getters were Guiliani and Huckabee with 8% each and Fred Thompson with 5%. That is NOT indicative of what their vote percentages will be in absolute terms, but it is relative to each other.

Here is my estimate based on their results as to how the figures translate into a GOP-Primary. I repeat, their poll did not have the data this way, I simply used their data and calculated what the totals would have been if the responses had been sorted by party....

(analysis continued on jump. hit TUESDAY below and scroll down)


Blogger Mark Moore (Moderator) said...

The poll had 49% Democratic leaners and 31% Republican leaners. That means .61 of respondents are more likely to vote in the Democratic Primary, and Hillary! has 35% which is 57% of that .61 percent. Hillary is ahead with 57% of the Democratic vote. All the other dem candidates and the dem. undecideds add up to only 43%.

Of course we are an open-primary state, so this is all back-of-the envelope stuff. Still, we can do the same thing on the Republican side.

39% of repondents leaned GOP and Guiliani and Huckabee tied with 8% among ALL repsondents Take that 8% and divide it by the .39 percent that are Republican-leaning and the Republican primary share of the vote for both Huckabee and Guiliani is 20.5%. Thompson is at about 13% by that reckoning.

I would point out that the respondent had to name the candidate they would vote for. This greatly biases the poll in favor of those with the highest name-ID. In Arkansas on the Republican side that would be Huckabee, Guiliani, and Thompson. Thus, these numbers might represent a high-point for these candidates given the methodology.

Eleven percent in the poll named some other candidate. My guess is that at least half of those other candidates were on the GOP side. Let's say 6% in the poll which converts to an estimated 6/.39 = 15% who named other candidates on the GOP side. The big winner is "don't know/refused". It got 28%. I think it is fair to say that about half of them, say 13%, were on the GOP side. That would be 33% of the GOP vote that could be described as undecided.

Due to rounding errors, the totals add up to a little more than 100%

The way the rules are written, the bulk of the Republican delegates will be allotted to the top three finishers if no one candidate gets over 50% of the vote. I am amazed that the race for the top spot is still so wide open, and the top three slots are even "opener". The three "leaders" still seem to have very low totals, and they are such known quantities that there is little room for them to go up, and in the case of two of them, lots of room to go down as their records are better known.

12:32 PM, October 30, 2007  
Blogger Mark Moore (Moderator) said...

Nobody wanted to comment on this all week? Did I mess up the numbers and ya'll are too kind to tell me?

I thought this would be big news.

7:09 PM, November 02, 2007  

Post a Comment

Links to this post:

Create a Link

<< Home