The University of Arkansas’s Arkansas Poll was released early today and it showed broad Republican preferences among likely voters in major election contests, including Tom Cotton over Mark Pryor for Senate (49-36), Asa Hutchinson over Mike Ross (50-39) for governor and generic Republican preferences in congressional and legislative races.

On ballot issues, the poll also showed heavy opposition to the legislative ethics amendment and sale of alcohol, but strong support for an increase in the minimum wage.


The poll also found a pessimistic outlook on conditions, a low opinion of journalists and little changes in past preferences on such issues as abortion and same-sex marriage.

I think I’ll go drink hemlock.


The poll interviewed 747 people between Oct. 21 and 27 and has a 3.6% margin of error (but 4.1 percent on the 568 “very likely” voters who constituted the findings on the political races. The Arkansas Poll seemed to me to delve more deeply into political questions, but the boss, Janine Parry, says the questions follow the pattern of past years. It posts its results on past questions and the results are generally good. You can get complete results here.

The gaps, by the way, were narrower, but still 7 points, in the races for Senate and governor counting preferences of all respondents, not just very likely voters.


The margins reflected here are, in my recollection, outliers of outliers in dozens of polls on the U.S. Senate and gubernatorial races. David Ramsey wonders about the definition of “very likely” voter. Is that more likely than just a “likely” voter? Has the polling reached into base-level voting? I don’t have a clue. Turnout of the less likely voters still seems an important factor. Also: The sample is 9 percent black, against a 13 percent black percentage in the voting population, a percentage the Democratic Party turnout effort hopes to increase.

UPDATE: Parry responded to my question about her feelings on the gap in the Senate and governor races, wider than other polling:

The gap is wider than the latest RCP [Real Clear Politics] average for sure. Gave me pause. But the rest of the sample looks as it always does, and when the head-to-heads have performed well come election day. So, it could be that I drew the short sampling straw and respondents were not representative on this question (and yet appear to be on all the other items, which is puzzling), or it could be that that barrage of negatives against Pryor is widening the lead Cotton has maintained – except for two polls (out of 19) – since late May. 

As to the meaning of “very likely”:

Very likely for us has always meant, said they were registered and described themselves as very likely to vote in the next election. It’s about the same percentage of the total sample as in years past.

Back to the hemlock.


PS: I dug up two pieces of information that didn’t make me weep. Gun control: 43 percent said laws are strict enough and 32 percent said they should be stricter. Take that, gun nuts. Abortion: Only 40 percent think abortion should be even harder to obtain. 37 percent want no change, 11 percent said it should be easier. No movement on same-sex marriage. 42 percent, split evenly, favor gay marriage or civil union. 48 percent oppose both. There was a sharp drop in support for a question about a path to citizenship for immigrants (from 59 to 49) and a rise in support of immediate deportation (from 21 to 29).

Democrats lead Republicans 31-29 in respondent identification, but independents (33 percent of poll) split Republican leaning. I’d say so.

You won’t be surprised to learn President Obama’s favorable/unfavorable at 27/66. Gov. Mike Beebe, on the other hand, was 67-18. Too bad one has coattails and the other doesn’t.