Media outlets duped by bogus polls
Published: March 19, 2010 at 7:37 am
Regardless of what you may have read elsewhere, no one has a decisive lead in the GOP primary race to replace Congressman John Shadegg, who is retiring at the end of the year. There have been three “polls” released by candidates in the race so far, and none of them are worth a lick.
Nor are they worth reporting on as news – much less fact – yet there has been a disturbing trend in the Arizona political media to parrot the results of these “polls,” even though nothing about them should pass the smell test.
Take the “poll” results released by Jim Waring Feb. 11 that showed 50 percent of 3rd Congressional District voters would choose him, dwarfing his nearest competitor, Pamela Gorman, who pulled in about 27 percent. Local media foolishly pounced on the story and reported the numbers that were in the press release. Dennis Welch of the Arizona Guardian wrote Waring had “a commanding lead,” James King at the Phoenix New Times unquestioningly wrote he “holds a double-digit lead” over his competition, and The Arizona Republic national political reporter Dan Nowicki regurgitated the results that showed Waring has “a big early lead” in the race. Even nationally respected political news outlets like Real Clear Politics helped push the Waring camp’s fairy tale that he was the early favorite.
Only my colleague Luige del Puerto pointed out the truth behind the poll: Its numbers were fantastically incomplete, as nearly three- quarters of the 300 people polled were undecided.
The political media here continued to irresponsibly report other bogus poll numbers in the race. A Feb. 25 “poll” of 1,000 voters released by Vernon Parker’s campaign showed he trailed Waring by a single point, 16 percent to 17 percent, with Ben Quayle close behind at 15 percent.
Unlike Waring’s poll, this one noted 41 percent of those surveyed were undecided, though that number is still alarmingly small for a poll conducted six months before a primary election in which none of the candidates are household names.
Once again, local media helped a candidate by reporting phony numbers.
Welch and King even noted how different the results of the Parker and Waring polls were, yet they didn’t even attempt to explain why. King went so far as to attribute the drop to Quayle entering the race.
On March 10, Waring’s campaign released another “poll” that should have been ignored. It wasn’t, as Welch again wrote a dispatch that gave the survey – like the first, conducted by GOP campaign firm Summit Consulting Group – credibility it doesn’t deserve. He noted that the report’s authors acknowledged there were “a high number of undecided” voters, but he didn’t report those numbers.
Arizona Capitol Times chose to cover none of those “polls,” outside of del Puerto writing about the flaws in the first one. Why? For starters, they wouldn’t release additional information about the polling, like the questions that were asked, demographic breakdowns of those surveyed, margin of error, and cross-tabs. Any valid poll will readily release that information, and an unwillingness to do so should set off warning sirens for journalists.
Those are important things for reporters to look for when evaluating polls, said Tom Jensen, the director of Public Policy Polling. His firm works across the country and is considered one of the more reputable pollsters.
“The No. 1 thing reporters should look for is transparency,” Jensen said.
More than anything else, the wording of the questions – all of them asked of respondents, not just the one or two cited in a press release – determines the validity of the polls. Did the poll include pass- throughs, where voters are told bits of information about each candidate to see how those surveyed react? Those are commonplace for internal campaign polls, as they can help shape strategy, but they don’t make for reliable results and should serve as a red flag.
The questions can also reveal if the poll is, in fact, a push-poll, which isn’t designed to produce valid numbers as much as it is aimed at poisoning the respondents’ views of a candidate by asking unflattering questions.
Finding out who paid for the poll also is critical, said pollster Bob Moore of Oregon-based Moore Information. If a candidate has paid for it, then the numbers can’t be taken at face value, he said, and reporters should do everything they can to speak to the pollster about the results, not someone working with the campaign.
“The pollster may get some numbers that the campaign doesn’t like, and won’t release,” Moore said.
Unfortunately, it’s likely we will continue to see spurious poll results being reported as the campaign season drags on. Jensen said it is “very frustrating” for reputable pollsters to see obviously disingenuous – if not fraudulent – polls being swallowed hook, line and sinker by reporters who accept the results at face value.
“As long as that happens, people will continue to do it,” he said.
- Jim Small is a reporter for the Arizona Capitol Times.