|
|
| |
| Sat, Jul. 5, 2008 |  |
 |
|
||
| | The pollster and methodology
Sunday, Oct 10, 2004 Arkansas News Bureau LITTLE ROCK - The Arkansas News Bureau/Stephens Media Group political poll on the 2004 election was conducted by Opinion Research Associates Inc., of Little Rock, which polled with accuracy for the bureau two years ago. The firm also polled voters on education reform issues for the bureau while the Legislature was in special session early this year. Opinion Research was the only public polling firm in 2002 to consistently show now-Sen. Mark Pryor defeating then-Sen. Tim Hutchinson by a several-point margin. All of the results of the company's final poll before the election were on target. The firm is a full-service marketing research company that has conducted a variety of surveys for private commercial, governmental, political and media clients. It was founded in 1983 by the husband-and-wife team of Ernest and Zoe Oakleaf. It is wholly independent, owned and operated solely by the Oakleaf family. No other research firm has conducted as many media-sponsored polls in Arkansas. The firm also has done work for politicians, mostly Democrats, but also has worked for Republicans, including the late Frank White, a former governor, and Jim Keet, a former state legislator. The most recent poll, conducted Monday through Wednesday, included interviews with 502 registered Arkansas voters - either 125 or 126 from each of the four congressional districts - who indicated they were likely to vote in the Nov. 2 general election. The poll has a margin of error of plus or minus 4.5 percentage points at a 95 percent level of confidence. That means that 95 percent of samples of this size could be reasonably expected to arrive at results varying no more than 4.5 percentage points from the true population statistics. Percentages don't always add up to 100 percent, owing to rounding error. The secretary of state's office says this is how statewide voter registration breaks out among the four congressional districts: 1st, 23 percent; 2nd, 26 percent; 3rd, 27 percent; 4th, 24 percent. Ernest Oakleaf says that adjusting the sample on the basis of those variables would be statistically insignificant. The random sample was 53 percent female and 47 percent male, and 91 percent white and 8 percent African-American. Twenty-nine percent of the interviewees were age 65 or older, and 60 percent had a household income of $50,000 or less. Fifty percent identified themselves as Democrats or independent leaning Democrat; 37 percent said they were Republican or independent leaning Republican. |
