Public opinion polls: take with grain of salt by Marjorie Arons-Barron

The entry below is being cross posted from Marjorie Arons-Barron’s own blog. Follow this link to find it.

Last week I was surveyed by the Gallup organization, confirming two things: first, that the venerable public opinion survey actually talks to real people, and second, that the results of questions they ask can’t possibly be a reliable reflection of, er, public opinion.

One problem is that the questions are designed to force choices,  elicit simple answers which often  don’t reflect either the complexity or the nuance of public policy issues.  I was asked whether I approve or disapprove of President Obama’s handling of the economy.  My position is that, while he initially failed to grasp the magnitude of the problem and was ill-served by some of his appointees, I now generally approve of his approach to the economy, but he has been unable to implement his ideas because of the recalcitrance of the Congress.    Give him a B+ on concept and a C-minus on implementation? So, do I approve or disapprove?

Other questions were like that as well.  Did I think that the No Child Left Behind Act was a good law or a bad law?  Are they referring to the goal of the law or to the unintended consequences of having teachers “teaching to the test?”  Some parts of the law are good; others are not.

Other problems with polls are laid out in David W. Moore’s 2008 book “The Opinion Makers: An Insider Exposes the Truth Behind the Polls.”  Moore, a former executive at the Gallup organization, notes that polls don’t measure the level of information a respondent has on a given subject.  The questionnaire forces the respondent to choose a position (approve, strongly or somewhat, and disapprove, strongly or somewhat)  and doesn’t offer a “don’t know” category.  Beyond that, such polls don’t have what I would describe as a “don’t care” category.

This is especially problematic when it asks people to identify whom they would vote for ”if the election were held today.”  Many respondents would probably say “don’t know” or  ”none of the above” if they could, and the either-or format tells nothing of the intensity of their selection….or the lack thereof.  The result, says Moore, is the “systematic distortion of public opinion.”

Why would they do this?  One reason is that the polling organizations partner with media organizations (for example, Gallup with USA Today; CNN with ORC International, formerly Opinion Research Corporation), who want the juiced up headlines to create a news story.  They certainly can’t run with a story that says that two thirds of the people sampled didn’t know anything about the topic or that three quarters of them had opinions but couldn’t care less if the outcome of the policy question went the other way.

They could do more sophisticated polling and reporting, but it costs time and money. The  polls as currently  designed  and used  are little more than a form of political entertainment.

What’s particularly toxic about pollsters and the reporting of polls is how they distort our political process. They give an artificial sense of what would- be voters really think (or don’t think) about candidates, often conflating name recognition with support. When, at the early stages of a campaign, candidates register in low single digits of “support,” the media will tune out their candidacies, depriving them of exposure they might deserve, and limiting their ability to raise money, which has become a marker of a serious candidate. It all becomes a self-fulfilling process.

Now we regularly see national polls, measuring Romney versus Obama, but,  under our electoral college system, as Moore reminds us, there is no such thing as a national electorate. The numbers should be published with warnings, like those on cigarette packages. Poll Reports indicating that there are few “undecideds” left may be artificially distorting how weak support may be for one candidate or the other and how easily some small item might cause them to change sides or stay home.

Moore’s belief is that “media polls give us distorted readings of the electoral climate, manufacture a false public consenses on policy issues, and in the process undermine American democracy.”  It’s hard to argue with his analysis.

I welcome your comments in the section below.