About that Gallup abortion poll...

Last week, Gallup announced “More Americans 'Pro-Life' Than 'Pro-Choice' for First Time,” a finding that got a great deal of attention in light of the manufactured “controversy” surrounding President Obama's speech at Notre Dame.

That Gallup result seems a little fishy, though, for reasons I touched on in my latest column:

Gallup says the large swing from a year ago is attributable entirely to a 10-percentage-point increase in Republicans and Republican-leaning independents who call themselves pro-life. But that 10-point increase can only result in the overall swing Gallup claims has occurred if more people are Republican or lean Republican today than a year ago. That's possible, but is inconsistent with other polling that shows fewer Americans than ever consider themselves Republicans.

On Saturday, Pollster.com's Charles Franklin confirmed that the Gallup poll “has party identification tied at 32-32,” which he explains is an “outlier.”

But don't take Franklin's word for it. Take a look at some other recent Gallup releases:

April 30, 2009
Democrats Maintain Seven-Point Advantage in Party ID
Have held significant edge since 2006

May 18, 2009
GOP Losses Span Nearly All Demographic Groups
Only frequent churchgoers show no decline in support since 2001

So, according to Gallup, Democrats have held a “significant edge” in Party ID since 2006, and continue to hold one today, as Republicans have lost support among “nearly all demographic groups” since 2001.

And yet Gallup hypes a poll that finds a majority of Americans self-identify as “pro-life” for the first time ever, even though that finding is based on the implausible premise that the two parties are tied in Party ID -- a premise that Gallup itself contradicts elsewhere. And, get this, Gallup didn't mention the tie in Party ID in its release touting the abortion findings. Had it done so, the findings would have (appropriately) been greated with much more skepticism.

That's pretty dishonest -- Gallup withheld information about its own poll that undermined the sensational claim it was making about that poll's findings. And it's a useful reminder that broad announcements like “More Americans 'Pro-Life' Than 'Pro-Choice' for First Time” shouldn't be taken particularly seriously unless they are accompanied by the complete poll.