Polls: Obama or McCain Is Winning

Two months ago I argued in these pages that Barack Obama will not win this year's presidential election. My prediction was based on essential factors about Obama himself -- his inexperience, his tax-and-spend liberalism, his history of race consciousness -- that make him an unacceptable choice to lead the nation in this time of difficult foreign and domestic challenges. I remain convinced that, on Election Day, most American voters will refuse to pull the lever for a candidate who falls so far outside the political and cultural mainstream of the country.

Now to the polls. There are three basic reasons to be skeptical about the validity and accuracy of polls: First, there is the well-known problem of bias that results from how polls are worded. Second, the raw data for the polls almost always is "adjusted" by the pollsters to give more weight to the Democratic responses. And third, the results of the polls almost always are within the reported "margin of error." The first two issues would require a detailed analysis that is beyond the scope of this article. But the third issue clearly proves my point that Obama is not "winning" the race.


As illustrations, let's look at three national polls, taken from the Real Clear Politics website: the NBC News/Wall Street Journal poll from October 4-5; the Reuters/C-Span/Zogby poll from October 7; and the Rasmussen poll from October 8.


According to the NBC/WSJ poll, Obama is leading McCain among registered voters nationwide by a margin of 49% to 43% (these figures combine those who said they supported and were leaning towards the respective candidates). So Obama is winning, right? Not so fast. The reported margin of error for the poll is +/- 3.8 points. The "margin of error" is simply the amount of potential error in a poll's results. This means that the results of the poll actually could be McCain 46.8% and Obama 45.2%. Very close, but with McCain ahead. The poll also includes 8% of respondents who said they were undecided. These undecided voters might decide to vote for McCain. Certainly there is no reason to assume they will vote for Obama. In short, this poll does not show that Obama is winning. At most it shows that Obama might be winning. On the other hand, it also shows that McCain might be winning.


Similarly, the Reuters/C-Span/Zogby poll shows Obama leading McCain among likely voters nationwide by a margin on 47.7% to 45.3%, with 7% undecided. The margin of error for this poll is +/- 2.8 points. Based on the data from this poll, therefore, McCain could be leading Obama by a margin of 48.1% to 44.9%. If the undecided voters decide to support McCain, the eventual margin of victory could be even larger. Again, this poll does not necessarily show that Obama is winning. Indeed, the pollsters correctly noted that "[t]he race for President of the United States remains far too close to call."


Lastly, the Rasmussen poll shows that Obama is leading McCain among likely voters nationwide by a margin of 51% to 45%, with 4% undecided. The margin of error for this poll is +/- 2 points. A-ha! Obama definitely is winning according to this poll (that is, assuming the undecided voters do not "break" for McCain). But there is something curious about this poll, which relates to the second issue I raised above about polls. This poll "weighted" its responses to reflect a predicted voting population of 39.3% Democrats, 33.3% Republicans, and 27.4% unaffiliated. However, in the last presidential election, just four years ago, the party affiliation of actual voters (according to the CNN exit poll) was 37% Democrat and 37% Republican. Perhaps Democrats are going to "out vote" Republicans by 6 points this year, but I think this is a very dubious assumption. In any event, there is no reason to accept the results of the Rasmussen poll over those of the other two polls (or over the results of many other polls I could have examined).


Before concluding, let's look at a typical poll of the races in two important "battleground" states. This poll was done by CNN/Time on October 7 (also taken from RCP). According to this poll, Obama is leading McCain in Ohio, 50% to 47%, and in Wisconsin, 51% to 46%. The headline for the poll: "New Obama gains in battleground states." Yet the margin of error for the poll was +/- 3.5 points. This means that, based on the results of the CNN/Time poll itself, McCain could be leading in Ohio by a margin of 50.5% to 46.5% and in Wisconsin by a margin of 49.5% to 47.5%.