ALSO VISIT

 

We are hearing constantly that the Republican National Convention has yielded a political windfall in public opinion for President Bush, with polls shifting from a Kerry lead of 3 to 5%, to an astonishing Bush lead of 11% in a post-convention poll. But this is not the whole story. According to CNN, as reported today, polling data put Bush ahead by 3% prior to the convention and by 5% afterward, a bounce of only 2%, not the apparent 11 to 16% hinted at by haphazard mass media parroting of single or conflicting polls. That 2% is the smallest increase in polling for an incumbent since before Richard Nixon's presidency, meaning that the news about Bush's bounce is actually less than fabulous.

What's more, the reporting of a lackluster Kerry bounce is based on Kerry's receiving a 2% bounce of his own. This implies that somehow expectations, or pop-culture pretense or else just widespread media bias account for the media's erroneous reporting of the RNC-related polling information. When more sophisticated polling data are analyzed, it is possible to ascertain the meaning of the data reported. That 2% jump came from the reaction of male voters, many of whom were electrified by the high-noon rhetoric of the RNC. Among males, Bush gained 6%, but among women, he lost 1%. This means that he acquired fewer new votes among men than he lost among women, judging by the 2% figure tied to his lead over Kerry.

When other polls are examined, it becomes less clear whether Mr. Bush actually has the lead certain convention-tied polls say he has. For instance, Rasmussen Reports finds that during and after the convention, polls showed Bush with 47.6% to Kerry's 46.4%, a margin of just 1.2%, shockingly divergent from the 11% touted by TIME magazine. As of today, Labor Day, Kerry had gained one-tenth of one percent, moving to 46.5%, just 1.1% behind Bush. This would be reported by mainstream media as a 48% to 47% lead for Bush, literally negligible for being so far below the margin of error of any serious poll. Another major problem with the TIME magazine poll is that is asked which candidate respondents were sure of voting for in November and did not account for whether those who did not respond would vote differently than those who did or for whether those who responded to the poll might actually do something other than what they reported.

TIME gave Kerry 41% support among respondents, who were asked "If the election were held today, who would you vote for?" Of course, the election is not being held today, and the campaigns are in a sense just revving up. The Republicans have sought to portray Kerry as having presented no plan to support his campaign promises, but in fact his plan has been completed and available for months: Kerry website. This fact automatically distorts the overall election picture, so that when the question is asked, placing the election suddenly today, it requires that respondents base much of their response on information either contained in media reports or in the poll itself.

When we examine the telling Rasmussen survey, we find that while Bush may have "gained more than five percentage points over John Kerry during the past three weeks," his overall lead has actually shrunk between 2 September (the day of his acceptance speech) and today, 6 September: from 2.8% to 1.1%. The other interesting fact that emerges is that the Rasmussen weekly tracking poll has shown no lead greater than 2.8% for either candidate going as far back as 29 January. On 12 August, Kerry led by 2.8%, and as of 2 September, Bush led by 2.8%. Rasmussen is more reliable in some ways because it polls 1,000 voters every day, but publishes polling data for 3 days' worth of surveys, meaning that a broader pool of research contributes to the published data, without being compromised by reaching beyond the statistically valid range.

What we come away with from even this minor scrutiny of the mainstream media's poll-data reporting is that it utterly fails to convey a meaningful picture of the race and seems entirely abstract in relation to the actual concerns and tendencies of voters. There is also a clear attempt to conflate polling data with prominent news items, without any scientific means of adjudicating the intellectual honesty of the analytic process. In the same way that stock analysts are constantly telling us the reasons for the actions of millions of investors, even as the trends are happening, the mainstream news media are telling us that there are clear links between polling numbers and news stories, then hyping those stories within the context of their now inflated importance. What we do not see is the citizen being asked to shape the debate; the citizen is asked yes or no or poor, fair or good, and then blunt figures are concocted out of that raw material.

The fine carving of the full picture is not conveyed by this daily ritual of repetition of data gleaned from the most shocking or zealous source of information. The TIME poll showed an 11% lead, but no matter how reputable any polling organization may be, it is only as accurate as fact, or as the overall trend in polling, and the fact remains that the TIME poll is the most extreme, making it an untrustworthy outlier, statistically. What is not included in the reporting of the 11% lead is that among respondents who vote on the issues, Kerry leads Bush (in the same poll) by 60% to 39%, a stark 21% margin. On "likeability", Bush has a lead of 60% to 37%. But the crucial question is which group would be more likely to vote. Issue voters tend to be more passionate and more resolute in their electoral leanings; they mobilize, and this year, there are reports predicting at least a 10% increase in turnout, mostly driven by passion related to issues which do not favor President Bush. The TIME poll does not account for these factors and does not work them into the questioning or the reporting.

Karl Rove has taken note of the trend and is desperately trying to mobilize a population he claims is out there waiting for his call, some 4 million evangelical Christians and rural voters. It has been frequently stated this week that the RNC was much too rough and angry to appear moderate to the independent voter, and that it appeared Karl Rove had followed through on his pledge to use every opportunity to "mobilize the base". Some analysts have suggested that the White House has given up hope on capturing the independent vote. Some reports have suggested that as much as 70% of undecided voters say they will not cast their ballot for Bush. Add these many factors to the polling data so prominently reported across our media, and we find that the small lead the President currently appears to hold, which does indicate a small increase in overall support, may actually show something like the upper limit of his support, whereas Kerry has a lot of persuading still to do.

Pollster John Zogby has already said this election is "Kerry's to lose", and indicates that many groups, including Arab-Americans, have pulled their support from the Republican party virtually nationwide due to the events of the last few years. We are consistently reminded by our media dons that polling is a great public service, allowing the public, and our leaders, access to all that we don't know about our fellow citizens, about what our nation is doing and thinking at a given moment. But are we getting that kind of mystical insight, or is our own insight being stripped from us in a haze of off-kilter numbers put artfully together to tell us something other than what we imagine?

In the end, issues will decide this race, along with the question of ballot integrity and the fairness and openness of the process. Hopefully, there will be no judicial intervention needed, and the results will be clear and irrefutable, but the election of Tuesday, 2 November, will almost certainly have its own numbers, its own logic, and its own meaning, not yet foretellable by telephone survey teams. [s]