When Texas Governor George W. Bush declared victory on Super Tuesday, he talked issues and outlined several of his top priorities. While some commentators, such as Brit Hume on Fox News Channel, seemed more interested in whether Laura Bush had changed her outfit before walking on stage, the likely GOP candidate for president spelled out ideas. Bush said he wants to reform Social Security, rebuild the military and cut taxes.
On this last point, Bush charged that
After eight years of Clinton-Gore, we have the highest tax burden since World War II. And yet we are told that taxes are not an issue . . . The polls say cutting taxes is not popular. I'm not proposing tax relief because it's the popular thing to do; I'm proposing it because it's the right thing to do.
And, thus, we see another example of the decline of public deliberation and leadership due to the irresponsible use of polling. As Bush says, political courage has come to be synonymous with opposition to the simple-minded solutions and expectations embodied in polls.
Bush's tax cutting proposal isn't doomed — though you would think so if you listen to the press and the wrongway Corrigans of the punditocracy. For months now, the glittering phalanx for the status quo made up of producers and reporters has sought to upend Bush's tax cutting plan as unpopular. When Bush lost New Hampshire, pundits gleefully attributed the loss to his plan to cut taxes.
Polling is nowhere near the exact science portrayed by the press. Results are often soft or "mushy." And even when hard majorities are found, the media don't show how conflicted, how inconsistent or just how plain confused voters are. Support for tax cuts is a great way to show just how detrimental polling can be to public debate when the media has its own agenda.
Wording plays the decisive role in all polling questions because Americans are generally ignorant about matters political. What the media doesn't want you to know is that most voters are making up their answers on the spot.
The public does or does not support tax cuts depending on how you frame the question.
Compare the wording and results of two Pew Research Center polls that followed Clinton's last State of the Union address. The pollsters set up the question in the news context:
resident Clinton has proposed setting aside approximately two thirds of an expected budget surplus to fix the Social Security system. What do you think the leaders in Washington should do with the remainder of the surplus?â€¦
Under the first variation:
â€¦ Should the money be used for a tax cut, or should it be used to fund new government programs?
Phrased this way, four out of five (60%) of those polled supported a tax cut. Only 25% backed spending on new programs. Another 11% wanted the money spent on "other purposes," and 4% didn't know. What seems like an overwhelming majority disappears when the wording reorients the average American's mind to think of other issues.
In the second variation, Pew Research Center asked the same set-up question but then provided the following options for spending:
â€¦Should the money be used for a tax cut, or, should it be spent on programs for education, the environment, health care, crime-fighting, and military defense?
Framed this way, only 22% of persons polled favored a tax cut. Nearly seven in ten (69%) liked the wish list of spending presented by the pollster. Finally, six percent thought the money should be used for "other purposes" and three percent didn't know.
Both these wording variations found a strong majority. Pew Research Center for the People and the Press conducted a scientific sampling with a sound methodology, but the results could not be more different. How could this be?
Results of this kind are not unknown to pollsters. But they go unreported by reporters anxious to use polls for their stories. Pollster Daniel Yankelovich, writing in 1991, observed that, "people automatically oppose government spending in the abstract but simultaneously endorse programs that involve government spending for causes they support, such as the war on drugs."
Polling on taxes or any other policy issue is limited in value because the media often frames the issue with a liberal bias. For example, it is unquestionably liberal bias to frame the tax cutting question the way a NBC News/Wall Street Journal poll of 2,014 registered voters did in December. It asked voters to choose between Candidate A who "wants to use the rest of the federal budget surplus to strengthen Social Security and Medicare" or Candidate B who "wants to use the rest of the Federal budget surplus to provide across-the-board tax cuts for families and business." Sixty-three percent chose Candidate A and 35% chose Candidate B. For most journalists, this is a perfect way to start handicapping the November elections. Forget any discussion of the issues. For the media, these polls show who's up and who's down.
As Richard Noyes of the Media Research Center found in one of his studies, the media rarely note the arguments in favor of tax cuts. They tend to ignore the statistics that illustrate the burden of taxation while emphasizing the "risky" nature of tax cuts. For example, reporters will often ask congressional leaders how they intend to "pay for" tax cuts with subtle hints that pit tax reform against, say, the future of Social Security. Voters are left without a foundation of balanced knowledge when they hear politicians talk about tax cuts.
How a poll is worded, then, is crucial because the press does such a poor job reporting issues substantively. In an poll of 880 likely voters, Zogby International asked the tax question this way:
The U.S. government will take in $3 trillion over the next 15 years that it currently has no plans to spend. Do you want to keep some of the money in the form of a tax cut or have the government make plans to spend it on federal programs?
Zogby found much more support for tax cuts when the debate is framed in this way. A strong majority of those polled — 56.3% — said they preferred a tax cut, while 31% wanted more government spending. Presented this way, Zogby found that support for tax cuts also went up among Democrats, with 41.6% saying they'd prefer tax cuts under this scenario. Less than a majority — 44.7% — said they preferred more government spending. Key Democratic constituencies also favored tax cuts, using the Zogby wording. More than half of union members, or 50.4%, who responded said they preferred a tax cut, while only 40.8% said they preferred government spending. And in the suburbs where both parties battle for the vote, 54.5% said they preferred a tax cut, while a relatively scant 30.4% preferred government spending.
"If every pollster asked the question using Zogby's words, the wave of polls showing support for tax cuts would lead to a very different conventional wisdom than we have now, it would seem," said Noyes.
The wording and background reportage of polling questions aren't the only factors that affect perceived support for tax cuts. The polling on tax issues are also a valuable reminder of the limitation of polls to explain the complex and nuanced public debate about taxes. While many journalists enjoyed reporting a diminution in support for tax cuts based on polls, none crafts their explanations based on polling questions that specifically test whether tax cuts "don't resonate with the public." It would not be unfair to say they just make it up as they go along.
The media usually portray the lack of support for tax cuts as a sign that the public positively supports the economic status quo. But at least one candidate, who spoke on the issue and actually interacted with citizens from around the nation, offers another explanation. Steve Forbes, publisher and former GOP presidential candidate, says that a lack of credibility has undermined the political support for tax cuts. "People tend to regard the tax issue the way they do the weather, because they've been burned so often," he says. This, too, goes unexplored by a press corps that prefers to use the so-called hard numbers of polls to ratify its own beliefs and prejudices in the horserace between candidates.