Going into Tuesday, a little bit about those political polls

The Associated Press reported Oct. 28 that plenty of people think public opinion poll reporting affects voting behavior.

The people who think this go beyond you and me. They include the statisticians and political professionals – Republicans and Democrats, alike – who make a living off of the number filling your heads this fall.

The troublesome thing, pollsters and even some political workers interviewed for the story said, is that the polls end up affecting elections nowadays instead of predicting them. Actually, that's one of many troublesome things cited by those interviewed, but it's a big one if you consider that you can be swayed by a poll instead of the substance of what a candidate seeking office says.

Poll leaders use numbers as bandwagon fuel while those who trail lament that they aren't given a fair shake. But we report the poll numbers because this is a democracy. Insiders involved in political attempts to lead your government use polls to shape messages that reach voters and make other strategic decisions. You ought to have access to that information, too.

Even so, some of what you hear can be misleading. Some surveys have inadequate samples; I'm talking about online polls that invite you to register your opinion, as often as you'd like. Others have leading questions in order to get an intended answer. Some polls are based on faulty demographic data while others do not include cell phones.

Take cell phones out of the polls and you take a good share of voters out. A report this month by the Pew Research Center for the People & The Press went so far as to suggest that leaving cell phones out of polls skewered results toward Republicans, apparently more apt to have land line phones than Democrats, by 4 to 6 percentage points in three of four election polls it conducted since spring.

Plenty of reputable polls that include cell phones show Republicans making gains this year. They include a Pew center poll of 2,053 likely voters who have either land line or cell phones. That poll, taken Aug. 25-Sept. 6, showed 50 percent of likely voters opting for Republicans and 43 percent favoring Democrats in congressional races. That poll has a margin of error of plus or minus 3 percentage points.

The poll shows 53 percent supporting Republicans and 41 percent Democrats among likely voters with land line phones, only, with a margin of error of plus or minus 3.5 percentage points.

The margin of error in the Pew center's other polls ranges from 2.5 to 4 percentage points, depending upon the survey.

Here's a link to a poll that must be driving Democrats nuts. It was forwarded to me on Oct. 31 by Linn County (Iowa) Supervisor Linda Langston, a Democrat, after I wrote about polling for my Sunday column in The Gazette.

The Associated Press recommends that its editors make seven checks when considering the validity of polls and surveys:

• Who did the poll and who paid for it?
• How many people were interviewed?
• Who was interviewed?
• How was the poll was conducted?
• When was the poll taken?
• What is the poll's margin of error?
• What questions were asked, and in what order?

If you're tired of the polling, take heart. The finish line is near. The real winners, barring any unforeseen calamity, will be announced Tuesday night, Nov. 2. For Eastern Iowsa results go to one of two sites:

TheGazette.com or

kcrg.com