By Josh Goodman, CQ Guest Columnist

As voters in Maine cast votes today in a referendum on gay marriage, most polls are giving supporters of same-sex nuptials reason to cheer. Surveys from three of the four pollsters that have looked at the vote indicate that the effort to block gay marriage in the state will narrowly fail.

Before California’s vote on gay marriage last fall, however, the polls showed the exact same thing. Those forecasts were wrong and California banned gay marriage in its constitution. The key question in Maine is whether the polls will be wrong again.

That, of course, is a difficult question to answer. One thing I’ve learned watching politics is that while it’s completely predictable that polls will sometimes be wrong, it’s not at all predictable precisely when they will be wrong. If it was obvious that a poll was wrong, the pollster would have come up with some different numbers.

That said, there is one explanation as to why the polls were wrong in California that might be at least somewhat instructive for the Maine vote: a “Bradley Effect” for gay marriage.

The theory here is that some voters might not tell strangers over the phone their true feelings about gay marriage for the same reason they might not tell strangers on the phone their true feelings about a black candidate: They fear that an honest answer wouldn’t be a socially acceptable answer — that it might even make them look like bigots in the eyes of some.

Never mind that there is quite a bit of debate about whether a Bradley Effect ever existed for black candidates in the first place. When it comes to gay marriage, the theory is plausible enough that it deserves to be entertained. And the important thing is that there’s no obvious reason why such an effect would be less potent in Maine than in California. If respondents are fibbing to pollsters, Maine’s gay marriage law probably will be repealed.

In fact, there’s at least a small bit of evidence that voters are fibbing to pollsters.

The only poll that has shown Maine voters overturning their gay marriage ban came from Public Policy Polling, which also is the only firm that conducted a poll in the state using Interactive Voice Response (IVR) technology — essentially an automated message. Tom Jensen of Public Policy Polling offered an explanation on his blog: Opposition to gay marriage was stronger because voters are more willing to tell the truth to a machine than a real person.

That’s a plausible explanation, but it’s not the only one. It’s worth noting that, for polling nerds, one of the storylines of this year’s elections have been the conflicting results of IVR and traditional live-interviewer polls. In Virginia, New Jersey and New York’s 23rd Congressional District, IVR polls all show Republicans running stronger than live-interviewer polls do. So, it could be that the divide in Maine has nothing to do with a Bradley Effect.

Plus, there are other reasons the polls might have failed in California that aren’t relevant to Maine. Perhaps in California pollsters were underestimating the turnout of blacks and Hispanics (groups that happened to be more likely to oppose gay marriage) who were motivated by President Obama’s campaign. Maine doesn’t have a lot of blacks and Hispanics.

Due to this uncertainty, the vote in Maine won’t just be a test of how voters feel about gay marriage. It also will be a test of whether, when it comes to gay rights issues, a Bradley Effect exists.

Josh Goodman is a staff writer for Governing Magazine and its blog about politics, Ballot Box.