Political polling is both a science and an art; and it’s come a long way since it wrongly led the Chicago Tribune to proclaim that Thomas Dewey defeated Harry Truman in the 1948 presidential election. That doesn’t mean that many of today’s polls aren’t also misleading, especially if a wild card is in play.

Political polls were in their infancy in the 1940s. Few people conducted them, and they were only done periodically. They failed to predict President Truman’s win because polling stopped several days before the election; and they missed Truman’s last-minute surge.

Today, a plethora of polls are conducted every day using sophisticated methodologies. Now, however, polls are taken also to form opinion not just to gauge it. Campaigns and their supportive media understand that polls that show their candidate with a decisive lead, whether that lead exists in reality or not, can dampen the enthusiasm of their opponent’s supporters and sway on-the-fence Independents.

Presidential tracking polls, those that ask “if the election were held today,” rely on assumptions about the relative turnout of Republicans and Democrats on Election Day to predict the likely outcome. All that’s necessary to skew these polls is to alter those assumptions.

Most Americans that follow elections and campaigns closely understand this. They tend to discount polls sponsored by media outlets sympathetic to one political party or the other or that oversample one political party, usually Democrats. They know these polls are less likely to predict election outcomes.

There is another phenomenon that accounts for discrepancies between opinion polls and election outcomes. It's called the Bradley Effect.

In 1982, Democrat Tom Bradley, the long-time African-American mayor of Los Angeles, ran for Governor of California against Republican George Deukmejian. Post-election research indicated that a smaller percentage of white voters actually voted for Bradley than polls had predicted, costing him the election. It concluded that many white voters that told pollsters they would vote for Bradley didn’t because of race.

Although some pundits predicted we would see a repeat of the Bradley Effect in the 2008 presidential election, it failed to materialize.

A similar pattern, however, emerged in the recent recall election of Wisconsin Gov. Scott Walker. Polls predicted that the election would be close, and that Walker might win by two or three points at best. He won by seven points, a larger margin of victory over the same opponent than he won in the 2010 election. Most likely it was union members that didn’t tell pollsters the truth, fearing reprisals if Democratic activists or union leaders knew how they voted.

What these two elections, and a few others like them, have in common is that, from time to time, a wild-card factor such as race or union activism gives a significant portion of those sampled reason to withhold their true opinion from pollsters. When you answer a pollster’s phone call you have no idea who you are really speaking with.

Will a wild card like the Bradley or Walker effects play a role in the 2012 election, and if so how will it manifest itself?

One possibility is that many Independents and Republicans that voted for Barack Obama in 2008 won’t vote for him again, but they won’t admit that to a pollster. They may have voted for Barack Obama because they thought he would govern from the center and bring an end to racial divisiveness, but are disappointed that he hasn’t. Democrats over the past three-and-a-half years have often accused people who disagree with President Obama of racism, and these voters don’t want to invite perceptions that they harbor racial prejudice.

Another possibility is that, as in Wisconsin, members of various Democratic constituency groups—minorities, union members, etc.—won’t vote for President Obama, but they want to avoid explaining why. There are a certain number of voters like this in every election, but this year the number could be statistically significant.

If one or both of the above possibilities are true, some conservative pundits argue, a tidal wave of anti-Obama votes could strike American on November 6 denying the President reelection.

Even so, we're not likely to see the receding support for President Obama and the wave of support for Gov. Romney reflected in most polls until after the presidential debates just before the election or perhaps not until people vote on Election Day.

Obama supporters will say this is wishful thinking on the part of Republicans; and, indeed, wishful thinking is common phenomenon in elections that leads supporters of the candidate behind in the polls to disbelieve those they don’t like.

In addition to all the public polls, however, the candidates have their private pollsters; and the Obama campaign has paid millions on daily private polling. Those polls likely are the reason Democrats appear to be in panic mode despite all the public polls favorable to President Obama.

The public polls most likely to mirror private polls are those that sample likely voters. Pollster Scott Rasmussen and others that poll likely voters say the race is a dead heat. Some have Gov. Romney with as much as a five-point lead.

Is it any wonder then that Democrats have begun hurling every accusation they can think of at Romney hoping one will stick? The war-on-women and Bain attacks have failed; now Democrats are attempting to paint a false picture of Gov. Romney as a felon that doesn’t pay taxes and keeps all his money overseas bank accounts.

As in 2008, a wild card again may fail to materialize. In nearly every election since the special elections in Massachusetts and New Jersey in mid-2010, however, a wave of anti-Obama, anti-Democratic votes have materialized; and that suggests they are likely to again.