Reporting on Polls: Facts and Value

Reporting on Polls: Facts and Value

Last week, I wrote about the overabundance of polls tracking the progress of various candidates in various races for public office here in Florida. I acknowledged my own frustration as a homeowner (even though I am a former pollster) over the constant barrage of phone calls, some soliciting support for campaigns, some seeking to assess that support.

It’s enough to make one refuse to answer the phone.

But beyond the irritation lies a more important question. Why so many polls?

Part of the answer, as I observed, is perfectly reasonable. Campaigns need polls to assess their progress and make tactical adjustments. The more competitive races, the more campaigns tracking voters, the more polls.

But other polls are conducted and released for press consumption, not for the team in the war room. The ink and air time spilled on polls often tops every other category of campaign coverage, for a very simple reason:

Poll results make a seemingly unambiguous claim to being “true.”

  • Generally speaking, poll results are released when an actual poll was conducted; rarely do we encounter a report of an alleged poll that was not conducted.
  • While various aspects of survey methodology can be the subject of controversy, reporting that the poll found that candidate X is leading candidate Y by Z percent is, quite simply, factual reporting (however important, or unimportant, these facts may be).
  • Because poll results are considered to be plausible public indicators of campaign progress (often the only plausible public indicator, other than periodic reports on contributions, which are available too infrequently to sustain campaign news coverage), campaigns care about poll results and are almost certain to be willing to comment on the latest poll, providing journalists with ready-to-use quotations to give life to the story.
  • Because poll results almost invariably affect the rival campaigns differently, one is guaranteed of contrasting viewpoints, another journalistic requirement.
  • Finally, because the basics of poll interpretation are straightforward and numbers-oriented, journalists can feel pretty well-insulated against accusations of bias in reporting polls, as long as they give each campaign a chance to offer their own interpretation of the poll’s results and significance.

There are a number of problems with this analysis, but one in particular is of greatest concern to me.

How do poll results assist voters in deciding for whom to vote?

Granted, in a contested primary with multiple candidates, or a nonpartisan race (again, with multiple candidates), or a race (like this year’s governor’s race) with so-called third-party candidates attempting to be taken seriously, poll results might prompt some voters to vote strategically, casting their ballot for a candidate who is not their first choice, but is their most preferred candidate with an apparent potential to win. Alternatively, improving poll numbers for a most preferred but third-party or long-shot candidate might inspire more voters to vote for that candidate.

But outside of these situations, polls tell us nothing that will help us choose our representative or executive.

What would happen if the coverage of polls was reduced by 80 percent, and those column inches/telecast minutes were devoted to other news about the election? What would journalists write about?

We could get more scandal-mongering, to be sure.

But we also might get stories about candidates’ careers (beyond the obligatory “this is candidate X” stories we encounter early in the campaign cycle). We might get candidates and others to engage the critical issues of the day, giving us better information about the state of the economy, say, or education, and the likely effects of alternative policies proposed by the candidates. Indeed, we might actually get alternative policy proposals with enough substance to discuss and critique.

We might . . . as far-fetched as this might seem . . . we might just find more voters knowing a bit more about the candidates and their policy plans than we do today. We might find voters choosing candidates based on that additional substantive and relevant information.

In other words, we might just improve the electoral process, holding candidates to a higher standard of knowledge and issue engagement, and allowing voters to hold themselves to a higher standard of informed decision making.

Voters do learn from the communications they receive during the campaign. Most of that learning, for better or (likely) worse, results from the ads and mailers the campaigns fire off.

From the press, alas, they mostly learn about upticks and downturns in popular support. There’s plenty of that information to be had . . . but none of it, for a functioning democracy, has any real value.

2 Responses to Reporting on Polls: Facts and Value

  • jimfrishe

    So true, sad to say. The press is not interested in the differences of candidates, beyond the bumper sticker brevity of a list of check marks. Although it may have started earlier, I think the change came after they brought down the Gary Hart campaign. After that, the press was only interested in the “gotcha” moment. Hanging a pelt on the wall was the goal of every reporter.

    Length of article may have also been a factor. Editors did not want long in-depth essays that got to the nuanced differences; they needed to sell papers. Every candidate had to fit a pre-determined mold/category. Every candidate I know well, loves to sit down with 10-20 people and spend an hour or two discussing issues with voters. That is real democracy at work. There is, unfortunately, not enough time to do much of that when you represent 200k – 600k people. The media could help bridge that time issue, but they cannot afford the column inches it would take to do that. They are, after all, a business first. So they look for the gotcha moment to define a candidate. Perhaps the internet will help. But then, a quote lifted out of context may get us right back to square one.