Spring Break
UA political scientists conduct unique opinion poll
Jack Willems
Issue date: 9/29/08 Section: News
The Blair Center of Southern Politics and Society will conduct the tenth annual Arkansas Poll this October, and there will be new improvements in the poll this year, said professor Janine Parry, poll director and UA associated professor of political science
"It will be so much larger than before," Parry said.
The Arkansas Poll is the only public-opinion poll in the state, Parry said. It began in 1998 and has been funded by the Blair Center for eight years. The poll surveys public attitudes in Arkansas on public officials, party identification, ideologies, ballot measures and issues.
Arkansas is one of 35 states with a statewide poll, Parry said. The poll will begin on Oct.1, and the results should be released Oct. 23.
Normally, the response goal is to get 800 respondents, but this year the makers of the poll hope to get 1,600 respondents, Parry said. Pollsters are hoping to get 400 responses from all four of Arkansas's congressional districts so they can determine public opinion for Arkansas's representatives.
It will also allow them to compare public opinion of one region of the state to another, she said.
"We can now compare Northwest Arkansas to southeast Arkansas or central Arkansas to the delta," Parry said.
The Arkansas Poll has been more accurate than commercial polls in predicting election results in the state, Parry said. Of the predictions the poll has made, 16 of 18 have been correct. In 2004, the poll called the state for President George W. Bush by nine points, and it also predicted Mike Beebe would beat Asa Hutchinson by 15 points two years later, she said.
"There is no more scientific or objective way to measure public opinion in the state," Parry said.
There are several methods the Arkansas Poll uses to get better results, said Molly Longstreth, director of the survey research center, which conducts the poll.
First, households are randomly sampled using a computer that ensures that the sample is proportionally represented by area code, Longstreth said. Within households, the survey will randomly ask for different adults of voting age, not just survey the first adult to answer the phone, she said. The order of the questions asked and the answers listed for those questions are changed to avoid order bias, Longstreth said.
"It will be so much larger than before," Parry said.
The Arkansas Poll is the only public-opinion poll in the state, Parry said. It began in 1998 and has been funded by the Blair Center for eight years. The poll surveys public attitudes in Arkansas on public officials, party identification, ideologies, ballot measures and issues.
Arkansas is one of 35 states with a statewide poll, Parry said. The poll will begin on Oct.1, and the results should be released Oct. 23.
Normally, the response goal is to get 800 respondents, but this year the makers of the poll hope to get 1,600 respondents, Parry said. Pollsters are hoping to get 400 responses from all four of Arkansas's congressional districts so they can determine public opinion for Arkansas's representatives.
It will also allow them to compare public opinion of one region of the state to another, she said.
"We can now compare Northwest Arkansas to southeast Arkansas or central Arkansas to the delta," Parry said.
The Arkansas Poll has been more accurate than commercial polls in predicting election results in the state, Parry said. Of the predictions the poll has made, 16 of 18 have been correct. In 2004, the poll called the state for President George W. Bush by nine points, and it also predicted Mike Beebe would beat Asa Hutchinson by 15 points two years later, she said.
"There is no more scientific or objective way to measure public opinion in the state," Parry said.
There are several methods the Arkansas Poll uses to get better results, said Molly Longstreth, director of the survey research center, which conducts the poll.
First, households are randomly sampled using a computer that ensures that the sample is proportionally represented by area code, Longstreth said. Within households, the survey will randomly ask for different adults of voting age, not just survey the first adult to answer the phone, she said. The order of the questions asked and the answers listed for those questions are changed to avoid order bias, Longstreth said.
Be the first to comment on this story