Reporting a presidential election is a big challenge for television networks. They have have huge interest in a news event - but not much information to share.

Enter, the exit polls: Early surveys administered at about 1,000 polling locations across the country. At their best, exit polls give election junkies an early sense of how the American electorate is leaning. At their worst, their data can be incomplete and misleading. Early exit polls don't always capture the full picture of who is voting; supposed "leaks" are often inaccurate. 

"It's an interesting contest of peek-a-boo," says Frank Newport, editor in chief of the Gallup Poll. "The only people who have access are the paying clients. Everybody else can only see what those clients leak out or report. It's a bit of a game in that sense, depending on what comes out, like reading tea leaves when the Vatican is choosing a new pope."

It's hard to divine when polls will prove accurate - and when they'll lead election watchers astray. Pollsters do, however, have a few tips on how to make the most of election polls. 

1. The first rule of exit polls is this: Before 5 p.m., there are no exit polls. That's when the "quarantine" lifts on the exit poll data, collected by interviewers stationed at polling places across the country, that major television networks use to project outcomes.

"Given the precautions we take, the chances are infinitesimal that you'll see correct information before 5 p.m.," says Larry Rosen, president of Edison Research, which runs the exit poll that the major networks and newspapers use (The Washington Post is among the newspapers that has paid for access). 

It's not like the floodgates open at 5 p.m., either: Networks are prohibited from releasing information that could be used to project the race until after polls have closed. It won't be until 7 p.m. that projections get the green light on the East Coast. 

"You will often see, early on, talking about the issues that people listed as important," Rosen says. "You will not have someone saying that one candidate is leading among both men and women, because that would pretty much give it away."

2. Take early exit polling results with a "giant grain of salt." That quote comes from Tom Jensen, director of Public Policy Polling, which runs polls for many Democratic clients.

"We saw in 2004 that early results were wrong and saw that even as recently as the Wisconsin recall," he says. "I know its hard for people not to just want to eat every piece of information. A lot of time though, watching the first wave you're going to get burned."

Rosen adds that there "are definitely patterns of whose supporters will vote at different times of day," which they've observed during their decade running national exit polls. That means "You're seeing partial data if you see data that you suspect was leaked during the day," he adds. 

Perhaps the most famous incident came from the 2004 election, where exit polls initially showed Sen. John Kerry (D) leading President George W. Bush by 3 points. A later report suggested a simple culprit: Many Republicans had refused post-voting interviews. Quirks like these explain why even some even polling junkies tend to tune out some of the early action, favoring information that comes out later in the evening. 

"Frankly, I'm not that interested in some early guesses," says Frank Newport, editor in chief of the Gallup poll. "It's kind of fun, but it's mostly people on television with time to fill."

3. Early exit polls may favor Obama more than the overall vote. That has to do with the Obama voters tending to be more likely to vote early (the exit polls that Edison Research runs includes early voting in 14 states where it is prevalent.

"What we have found in most of the important swing states, Obama has run up the score with early voting," says Jensen. "That means that earlier voting might be a little too optimistic, because they're including Obama's early voters but not the Election Day vote where Romney would do better."

David Flaherty, CEO of the right-leaning Magellan Strategies, agrees. "The Obama strategy has been to really go after the early voters," he says. "The Republicans are putting together efforts to get people to vote early, but the whole argument is we will win the vote on Election Day."

4. Keep an eye on the demographics. Jensen and Flaherty both say this is going to be one of the crucial things they'll watch tomorrow night. 

"Here in Colorado, for example, Latinos were 13 percent of the vote in 2008," says Flaherty. "If it comes back at 5 pm, that they're 18 or 20 percent of the vote, that would be a key observation. Then the folks in Chicago are going to be extremely pleased. If we see it at 9 percent, it might cause the Romney folks to start thinking that group isn't as enthusiastic."

Here's Jensen's take: "What I'm going to be looking out for are the demographics of who voted along racial and age lines. We Democratic pollsters think that the African American and younger voters will have as high turnout as they did in 2008. Republican pollsters are projecting enthusiasm won't be there." 

More from The Washington Post:

Early exit polls show electorate leans more Republican

The seven most important counties in Election 2012

Wonkblog footnotes the election