The political world was more or less stunned Friday by the revelation that, according to Research 2000, less than half of Republicans think President Obama was born in the U.S. In fact, 28 percent of the poll's GOP respondents said he wasn't.

After initial questions of whether this can be, it's clear that the polling methodology looks pretty good: Research 2000 polled a decent sample of 2,400 adults nationwide, and the regions are split evenly.

Polling on Republican opinion is undergoing dual phenomena right now. With fewer and fewer people counting themselves as Republicans, sample sizes of Republicans are smaller, meaning results could be less reliable. The size of today's GOP sample was 527, compared to 601 independents and 743 Democrats surveyed--not a huge difference, but a difference.

The other phenomenon is the legitimate polarization of the shrinking Republican base, as die-hard conservatives remain and moderates leave the party.

In 2006, Pew clocked GOP party identification at 28 percent of likely American voters; in April of this year, it was 22 percent. If we assume that all those Republicans who left became independents, we can guess at how today's poll would have turned out in 2006, since we know that 83 percent of independents today think Obama was born here*.

(This means running with a flawed premise: that the Republicans who left their party and became independents are equally likely to question Obama's citizenship as all other independents. It also ignores the notion that party affiliations are more fluid than 21 percent of Republicans simply bolting for independence, as a chunk--without some becoming Democrats, and without the GOP picking up some new members as those left.)

According to a mathematical expansion on today's pool of GOP respondents, the 2006 results among Republicans on the question "Do you believe Barack Obama was born in the United States of America or not?" would have gone as follows:

              R2000 results     educated guess at 2006 results
Yes                42%                  51%     
No                  28%                  24%
Not sure        30%                  25%

More than half would have said Obama was born in the U.S...but the coalition of doubters would still be strong, even with the old members of the GOP base counted in.

(Maybe we should assume that former Republicans are more likely to question Obama's birthplace than the rest of the independent field, and intentionally skew the independent effect here; maybe not.)

Since today's poll made such a splash, it'll be interesting to see if anyone else polls on the same question, both to see if Research 2000's results were a fluke, and to see how the birther "dialogue" is going--whether suspicions of foreignness have subsided. Major polling firms could, conceivably, include birtherism with other current-events questions in weekly surveys.

The established media is wholly against the essence of these citizenship doubts, and respected journalists freely refer to birtherism as racist. But, at the same time, everyone's rubber-necking. 42 percent of Republicans seems so low--can it be right?

Maybe a poll on this issue only fosters birtherism, in the same fashion that Conor Clarke has suggested polls can do with all public matters.

Maybe we can track birtheristic sentiment along with Obama's favorability and approval as his presidency goes on...Do you like President Obama? Do you approve of his job as president? Do you think he was born here?...at that point, things will have truly gone weird. Whether more news happens around birtherism will depend on how weird we're all willing to get.

*Given the 28 percent ID in 2006, and the 22 percent ID today, we can conclude that a proportioned GOP sample size would be 670, not 527, by multiplying the 527 sample size by 28/22 (the percent change in the size of the GOP's base). 143 of which would be independents. From those numbers, we can derive weights for the GOP and independent percentage results in today's poll--weighting the independents at .2134 and the current-day Republicans at .7866--and can produce percentages for a purported 2006 survey.

We want to hear what you think about this article. Submit a letter to the editor or write to letters@theatlantic.com.