Romney vs. Obama 2012: Are Presidential Polls Accurate?

With election season in full swing, polls are constantly being thrown at the public, showing how many of a certain population favors each candidate.

Digtriad.com says polls are useful because they allow politicians the resources to decide which states the need to target, but it's very hard to tell how accurate they are because they represent a small portion of the population.

The article gives an example of the Elon poll in North Carolina, with the top issues being economy, health care and the deficit. The poll showed that if the election were held that day, Mitt Romney would win by a 4 percent lead over President Barack Obama. It showed voters favored Romney 52 percent to Obama's 39 percent. It gave the example of another poll done by Public Policy Polling, which showed 50 percent of voters disapproved of Obama's job performance, while 48 percent had an unfavorable opinion of Romney. Based on that poll, the candidates would tie.

Here, the article says, is where it gets complicated. These polls interview 1 to 2 thousand people max, and in North Carolina, there are more than 6 million registered voters. Is this a fair representation?

Each year, Gallup surveys voters, asking who they would choose as the next president.

History has shown that Gallup's polls don't deviate much.

In 2008, Gallup showed that 55.0 percent of people voted for Obama, while 44.0 percent voted for John McCain. The election result: 53.0 votes for Obama and 46.0 votes for McCain. That's a 2 percent deviation both ways.

In 2004, the Gallup poll had 49.0 percent voting for George W. Bush and 49.0 voting for John Kerry. The election results were 50.7 percent for Bush and 48.3 percent with Kerry--a -1.7 percent and 0.7 percent deviation, respectively.

The biggest deviation the poll showed was the first year Gallup polled-1936. It had both a negative and positive 6.8 deviation for President Franklin D. Roosevelt and his opponent Alf Landon.