Polls, Surveys & Studies

other topics under JUDGMENT: Avoiding Bias, Be Enterprising, Dealing with Profanities, Public Meetings

Polls, surveys and studies have increasingly become a newscast staple. Nearly every day we seem to be bombarded by a poll, survey or study indicating, or seeming to indicate, how everything from a disease to a politician (if the two are distinct) might be faring among the public. It's always a good idea to figure out beforehand the types of poll and study results you plan to tell your listeners and why.

Get an expert

Polling data are among the most misunderstood and misused information with which we deal. It helps to have an expert available to explain to our listeners...AND TO US...what a poll's results can mean. Usually there's an expert on polling at a nearby college or university. Get in contact with that person and try to build a working relationship. Both you and your listeners will be grateful.

Polls are conducted by various groups -- polling organizations, news operations, colleges and universities, businesses and interest groups, as well as political consulting firms. These groups will generally provide some information on how any given survey was taken, such as how many people were interviewed, how those people were chosen, and how they were interviewed. There are advantages and disadvantages to various polling techniques, and your expert should enlighten you on how trustworthy the results may be.

The numbers game

Care must be given when providing poll results to our listeners. On the one hand, a simplifying vagueness about the results can be deceptive. To say

SENATOR ALISON WILLIAMS IS LEADING IN HER RE-ELECTION BID OVER CHALLENGER MARC ANTONY IN THE LATEST MIDDLEVILLE TIMES/CHANNEL 8 POLL.

and to leave it like that gives listeners no impression on how close the race may be. Something must be said about the size of the lead, keeping in mind how the the method of sampling used affects the margin of error and other aspects of result reliability.

Another presentation problem involves tossing ALL of the numbers from a poll to our listeners:

SENATOR ALISON WILLIAMS HAS A 9-POINT LEAD...42 TO 33...IN HER RE-ELECTION BID OVER CHALLENGER MARC ANTONY IN THE LATEST MIDDLEVILLE TIMES/CHANNEL 8 POLL. 25 PERCENT OF THE 607 PEOPLE SURVEYED STILL HAVEN'T MADE UP THEIR MINDS, AND THE POLL HAS A MARGIN OF ERROR OF PLUS OR MINUS 4 POINTS, SO WILLIAMS COULD BE AHEAD AS MUCH AS 46 TO 29 OR ONLY 38 TO 37.

Aside from not being a particularly helpful interpretation of the polling data, this story throws so many numbers that listeners might think they're hearing sports scores or the lottery results. Numbers are, of course, necessary to this story, but they should be included judiciously, and hypothetical results created by manipulating the margin of error should not be included.

Whose study is it?

Reporters who take great care to reveal the source and accuracy of political polling data may be cavalier about surveys and studies that do not directly address political issues. This carelessness is especially apparent in the reporting of lifestyle surveys, medical studies, environmental-impact studies, and surveys of economic activity. As an example, the American Automobile Association released a study that said most pollution in the air came from factory smokestacks rather than from automobile exhausts. One major-market, clear-channel station told the story this way. First the anchor read the sentence,

ARE CARS THE BIGGEST POLLUTERS OF OUR AIR?

Then followed a buzzing noise such as is heard when a wrong answer is given on "Family Feud" or some other TV gameshow. The anchor continued:

WRONG! FACTORIES ARE THE NUMBER-ONE POLLUTERS ACCORDING TO TRIPLE-A....

Aside from the gimmicky (and credibility-destroying) presentation, this story gives undue credence to Triple-A's study. Triple-A is an interest-group for drivers of passenger vehicles, and the link between Triple-A and the study must be prominently reported. Stating the report's conclusions as fact at the beginning of the story provides too much bias and may prevent listeners from acquiring a critical understanding of the report.

It ain't over 'til it's over

Political polling provides a measure of how well a campaign is doing at a particular point in time. Surveys are useful in reporting on the progress (or lack thereof) of a campaign and on any changes that may take place in the campaign as a result of this feedback. But, as we all know, they are not official election returns, and while they may reflect upon the substantive issues in a campaign, poll results do not themselves represent a substantive issue worthy of extensive coverage. In other words, don't spend your time covering the polls instead of covering the election.

The contents of the Newswriting for Radio website are under copyright and may not be printed or reproduced in any form without permission.