Last-Minute Polls All Over the Place, But Obama Leads in Virtually All of Them

The final round of presidential election polls offer nothing certain about Tuesday's results -- except that Obama leads in all of them. Despite Obama's consistent lead, the polls show a lot of variance, and the "fine print" often raises more questions.

Variation

Although the presidential election is really a series of state-to-state contests, the national popular vote polls can provide some indication of the overall sentiment of voters. The nationwide polls, however, lack uniformity: they either show that the race is very close or tied or that Obama will win in a landslide.

The most recent CBS poll, for example, has Obama up by 13 points (54-41), but Obama leads by only 2 points (47-45) in the final IBD/TIPP poll. Last week, Pew reported that Obama had a 12-point lead; today, the pollster gave Obama a comfortable, but much smaller, lead of 6 points. Overall, the Pollster.Com average of all the major polls has Obama up by 6 points nationally.

Fine Print

Reading the details in some of the polls only makes things even more confusing -- or interesting, depending upon your perspective. In the Pew poll, for example, people who have "already voted" favor Obama 52-39. The remaining 9% are in a category "undecided/other." I assume that people who have already voted are not undecided about how they voted. On the other hand, I cannot imagine what "other" candidate could garner 9% of the national vote among early voters. Furthermore, despite Obama's lead among those who have already voted, the Pew poll actually shows McCain winning among election-day voters 46-45. A significant amount of likely voters remain undecided, which I find baffling.

In an earlier post, I analyzed the CBS poll that gives Obama a 13-point lead. In that poll, Obama leads among early voters, but 50 percent of early voters in that survey are Democrats, 60% are women, and 16% are black -- which explains Obama's substantial lead.

The TIPP poll, co-sponsored by IBD (or Investor's Business Daily) does not provide the racial demographics of the individuals it polled (which is a startling omission). In that poll, a large degree of voters also remain undecided. For example, 11% of "northeastern" and 12% of "midwestern" voters have not picked a candidate yet, according to the poll. Interestingly, the pollster has a "hot topic" poll which seeks to determine whether Americans are "ready for socialism." The results will shock you (not really).

Pretty soon, the only relevant polls -- the election-day results -- will give us an actual winner, and pundits will drown us with repetitive and unlearned commentary about the returns. Until then, the poll-obsessed readers can only speculate about the accuracy of the pre-election data. For a great resource in this process, visit the presidential polling site on Pollster.Com.