Live by the Poll, Die by the Poll - The American Spectator | USA News and Politics
Live by the Poll, Die by the Poll

According to a poll conducted June 18-19, Barack Obama had a15-point lead over John McCain. A poll conducted July 25-27, however, showed McCain with a 4-point lead over the Democrat.

Given this 19-point swing in favor of the Republican candidate over a period of 38 days, if the trend continues at the current rate, on what date will Obama’s support reach zero? (Extra credit: If the Democrat gets no votes, will McCain get 100 percent? If not, please estimate the percentage of the total popular vote that will be won by Bob Barr, Cynthia McKinney and Ralph Nader.)

While you run through those equations with your calculator, let me remind you that the USA Today poll that made headlines Monday at the Drudge Report — showing McCain ahead 49-45 percent among likely voters — was the first poll in nearly three months to show the Republican leading Obama.

The USA Today poll was conducted by Gallup during the same three-day period when Gallup’s own daily tracking showed Obama with an 8-point lead. When two polls by the same firm at the same time disagree by a margin of 12 percentage points, what’s a political junkie to do?

First, relax — it’s still July, the election’s more than three months away, and most of the independent voters whose preferences will decide the election on Nov. 4 aren’t really paying attention yet.

Second, realize that the methodology of polling is not an exact science. The variability of results is caused by a lot of things, including the fact that nobody knows who will actually turn out to vote on Election Day.

The most recent Gallup tracking poll, for example, reported that 88 percent of registered voters indicated a preference — 47 percent for Obama, 41 percent for McCain. But in 2004, only 85 percent of registered voters actually cast a ballot for president. The 12 percent of the Gallup’s undecided voters is smaller than the percent of registered voters who didn’t vote four years ago.

THIS IS WHERE the issue of “likely voters” comes in. A series of questions (including about past voting habits) is used to identify those survey respondents who are most likely to vote in November. Republicans have traditionally done better among “likely voters,” and this was the case with the USA Today poll released Monday that showed McCain leading Obama for the first time in nearly three months.

Perhaps the thorniest methodological puzzle of polling, especially in polls taken months before an election, is whether a random sample will match the actual electorate that shows up to cast their votes on Election Day.

Consider that mid-June poll by Newsweek that thrilled Democrats by showing Obama with a whopping 15-point lead. Newsweek‘s sample of registered voters showed only 23 percent identified themselves as Republican compared to 38 percent who said they were Democrat, a skew unmatched by other polls.

Bandwagon psychology — the tendency of people to want to be part of the winning team — can sometimes turn polls into self-fulfilling prophecies. This became a danger to the McCain campaign as the summer dragged on with poll after poll showing Obama ahead.

Psychology, however, cuts both ways. In early July, while John McCain was reshuffling his campaign team, bringing in Karl Rove protege Steve Schmidt, Team Obama was showing signs of overconfidence and planning a nine-day trip overseas.

While the Democrat basked in media adulation abroad, the McCain campaign hit hard on the home front, rolling out a TV ad that blamed high gas prices on Obama’s opposition to offshore oil drilling. McCain repeatedly slammed Obama for opposing the “surge” in Iraq. And when Obama canceled a visit with wounded troops at a military hospital in Germany, the McCain campaign quickly rolled out a TV ad criticizing his decision.

Poll-watchers waited to see the impact. A Fox News poll last week showed Obama down to a one-point lead, but that was taken too early to gauge the impact of the Democrat’s foreign trip or of McCain’s newfound aggression at home.

THEN CAME A WEEKEND of worry for Republicans. The Rasmussen tracking poll, which had shown the race tied on June 19, on Saturday reported Obama ahead by 6 points. The next day, Gallup reported Obama ahead by 9 points.

Had McCain’s attacks backfired? Or had the wall-to-wall media coverage of Obama’s trip simply buried the Republican message?

Monday afternoon’s release of the USA Today poll, with McCain leading by four points among likely voters, answered “no” to those questions, and was echoed yesterday when Rasmussen’s tracking numbers showed Obama’s lead shrunk to 47-46. Meanwhile, the Democrat’s margin in yesterday’s Gallup tracking poll was down to 6 points — the same 47-41 edge he had on July 21.

It is still only July and, as pollsters routinely caution, their results are only “snapshots” of current opinion, not predictions of Election Day results. The dog days of August and national conventions loom ahead, and the outcome will ultimately depend on the events of the next three months.

For the past three months, however, all those polls showing Obama in the lead fueled a sense of inevitability that cast shadows of doom among Republicans.

What happens if the polls should shift the other way? What fate befalls the frontrunner who falls behind? Just ask that formerly “inevitable” nominee, Hillary Clinton.

Robert Stacy McCain