— We see something like this on almost a daily basis: Candidate A is leading Candidate B by   X-number of percentage points.

But really, how much faith should we put in political polls? Or should we put any faith in them?

At this point, it’s hard to tell what to believe. Take, for instance, the race for the 12th district seat in the U.S. House of Representatives. Incumbent Mark Critz, a Democrat from Johnstown, and Keith Rothfus, a Republican from suburban Pittsburgh, each crowed last week about their strong polling numbers. The thing is, they were talking about polls that reach wildly different conclusions.

Rothfus’ campaign was touting the survey by McLaughlin & Associates that had him in a dead heat – with each candidate taking 38 percent of the vote – with six weeks to go before the election.

Not surprisingly, the Critz camp was quick to dismiss the poll, which showed a remarkable   24 percent undecided. A Democratic poll taken by Anzalone Liszt Research revealed numbers that the local candidate found much more to his liking: Critz was pulling in 52 percent of the vote, while Rothfus had 41 percent, with 7 percent undecided.

It’s surprising to see such a wide disparity between polls, but that doesn’t mean it’s uncommon.

Late last month, the Republican Party of Pennsylvania touted a poll conducted by Susquehanna Polling & Research for the Tribune-Review that showed Barack Obama holding a slight 2-point lead over Mitt Romney in the presidential race in our state.

That stood in stark contrast to other polls released in the same time period.

A Philadelphia Inquirer poll had Obama leading by 11 points in Pennsylvania while Muhlenberg College’s survey showed the president with a 9-point edge.

So which polls should we believe?

Perhaps the more important question is this: Are we putting too much stock in them?

A poll may give us an idea of what might happen on Election Day, but it’s by no means a perfect indicator. Just ask former congressman Jason Altmire, who was favored to beat Critz in this spring’s primary election. Union members, who largely backed Critz, came out in extraordinary numbers to turn the tide in his favor.

And too often, it seems, the polls conveniently fit the organizations paying for them. Whether it’s tweaking the questions to get the response you’re looking for or overpolling a supportive demographic, it’s easy to manipulate the numbers.

We understand that and, while we by no means condone the practices, we realize it’s part of today’s political world.

The part that’s truly frightening is that skewed numbers can have a real-life impact. Too many people look at the polls and say “My candidate’s down by 8 percentage points. There’s no way my vote will make a difference.”

That kind of attitude, when shared by thousands or even millions of others, can swing the results of a race at any level.

We urge all voters to take the polls being published with a grain of salt and don’t let what could end up being faulty numbers steal your voice on Nov. 6.

Click here to subscribe to The Tribune-Democrat print edition.

Click here to subscribe to The Tribune-Democrat e-edition.