Scott Rasmussen believes in the theory that people are more accepting of information that reinforces their established views. It’s called confirmation bias.

The founder of Rasmussen Reports, a national polling firm, sees it all the time.

“When you have a lot of different polls showing different results, almost everybody favors the one with the results they like,” Rasmussen said during a presentation to members of the Maine Credit Union League in Freeport last week.

That may explain why Republicans have taken a shine to Rasmussen’s polls, while Democrats view his work with suspicion, or outright contempt.

The divergent sentiment has been on display over the last few weeks of the presidential race. Most major national polls like CBS News/New York Times, Fox News and Gallup showed President Obama with a narrow lead over Republican hopeful Mitt Romney. But Rasmussen’s polls were an outlier, giving Romney a slight advantage.

His numbers reignited a debate about his methodology and his motives. The charismatic Rasmussen has taken it all in stride, defending his numbers as a guest on Fox News while the conservative aggregate blog The Drudge Report screamed the results in large, bold type.

During his visit to Maine last week, Rasmussen talked about the variance in polls, the impact on campaigns, and voting decisions in a “world where nobody believes what the other side is saying.”

Rasmussen knows that world all too well.

Recent claims that he cooked polls on the presidential race were not the first. He heard similar criticism in 2010, when his polls were the first to show Republican Scott Brown closing on Democrat Martha Coakley in the Massachusetts U.S. Senate race. Other polls taken at the time had Coakley leading.

Brown eventually won.

Rasmussen has had some good polling years and some bad ones. He was within a percentage point of nailing the 2008 presidential race, but 4.5 points off the mark in the 2000 presidential contest.

His methods are not without controversy. While some pollsters conduct surveys over several days, Rasmussen does automated polling in four hours. He is not an issue pollster, exploring and double-checking respondents’ answers. He is, unabashedly, a horse-race pollster.

Rasmussen weights his results by party. He only polls likely voters. The two methods have led some critics to claim that his polls have a Republican lean, or “house effect.”

He is also criticized for not fully disclosing how his surveys are conducted. The lack of disclosure is the reason Rasmussen polls are not endorsed by the National Council of Public Polls and the American Association for Public Opinion Research and Transparency Institute. The two groups have endorsed most of Rasmussen’s competitors.

For the above reasons, many national news organizations, like The Associated Press, The New York Times and The Washington Post, don’t publish Rasmussen’s results. Those that do often include some kind of disclaimer that Rasmussen’s polls skew Republican.

Rasmussen has heard that charge plenty of times. He heard it in Maine in 2010.

Polls are supposed to measure public sentiment. However, in some instances, they can change how people vote.

During the 2010 governor’s race, Rasmussen released a poll that some political observers believed changed the dynamic of the race. For months the race had been between Republican Paul LePage and Democrat Libby Mitchell. Independent Eliot Cutler was considered a formidable candidate, but most Mainers simply didn’t know him. Polls released in September showed that Cutler had been unable to get above 15 percent — the so-called viability threshold for third-party candidates.

But in October, Rasmussen published poll results that changed the complexion of the race. The poll gave Cutler 20 percent, equal to Mitchell.

Cutler’s sudden surge caused confusion among progressive voters who were opposed to LePage but unsure whether Cutler or Mitchell had the best chance to beat him.

The result was a splitting of the progressive and independent vote. LePage walked away with the narrow victory.

It was Republicans’ dream scenario — one they’re trying to replicate in this year’s U.S. Senate race.

At the time, observers struggled to explain Cutler’s surge. Some settled on conspiracy.

Rasmussen’s reputation as a “Republican pollster” surfaced again. Others noted that Rasmussen was scheduled to speak at a luncheon hosted by the Maine Heritage Policy Center, a conservative advocacy group, just days after the election.

Speaking in Freeport last week, Rasmussen said that the 2010 gubernatorial race was a rare occurrence, one that probably won’t be duplicated in the U.S. Senate race.

“It used to be that the third-party candidate faded away as Election Day gets closer,” he said. “That’s not going to happen this time. You have a former governor running.”

Rasmussen was one of several pollsters to recently survey the Senate race. There were no surprises this time. His poll gave former Gov. Angus King a 12-point lead over Republican Charlie Summers. It was a larger advantage than the single-digit lead reported by Public Policy Polling, a national firm that’s been pegged with a Democratic house effect.

Rasmussen suspects that Summers’ polling numbers won’t change much between now and Election Day.

“The variance will come between King and (Democratic candidate) Cynthia Dill,” he said. “That’s something to watch for.”

Rasmussen will be watching, too. He plans to poll the U.S. Senate race again.

If it produces a controversial result, it’s unlikely to change how he polls.

“It’s worthless to get into arguments about these things,” he said. “People are always going to pick the poll with the results they like.”

Steve Mistler can be contacted at 791-6345 or at: smistler@pressherald.com

Twitter: stevemistler