Lies, damned lies, and Ramesh Ponnuru's use of polling

In a “Right Matters” entry on the Washington Post web site, Ramesh Ponnuru argues:

President Obama's speech at Notre Dame yesterday is another sign that pro-lifers are slowly winning the political battles over abortion. It was not the speech of a man who is confident that his position is right and popular.

You know what else is not the approach of a “man who is confident that his position is right and popular” ? Using bogus polling data to back up a point.

And yet, here's Ramesh Ponnuru, just a few sentences later: “The latest Gallup poll even has a majority of Americans calling themselves pro-life."

That Gallup poll is, however, so deeply flawed that it simply cannot be taken seriously. The flaw is obvious from simply reading Gallup's write-up of the poll, and had been pointed out publicly days before Ponnuru's post. Yet he cites it as though it is meaningful.

Now, here's a Ponnuru post on National Review's blog, writing about Barack Obama speaking at Notre Dame: “the same polls that showed that most Catholics approved of the invitation showed that a plurality of weekly churchgoers do not.”

No, they didn't.

Ponnuru is presumably referring to this Pew poll. But that Pew poll didn't say a plurality of weekly churchgoers disapproved of Notre Dame's invitation. It said a plurality of white, non-Hispanic Catholics who attend church weekly disapproved. “White, non-Hispanic Catholics” is not a synonym for “Catholics,” and it is certainly not a synonym for “churchgoers.”

UPDATE: Ponnuru offers a partial correction:

On Monday I wrote that a poll showed that a plurality of Catholics who attend Mass weekly disapproved of Notre Dame's invitation to President Obama. Via Jamison Foser, I see that I got it wrong: The poll only gave the numbers for those weekly Mass attendees who are white and non-Hispanic—which seems to me a little strange of the pollsters but there you go. (Needless to say, I don't find the rest of Foser's post persuasive.)

As for the rest of my post, the part he doesn't find persuasive, that was about Gallup's claim that a majority of Americans now self-identify as “pro-life.” The problem with that poll is that it had an equal percentage of Americans self-identifying as Democrats and Republicans -- something way out of line with other polling, including Gallup's polls. In order to believe Gallup's finding that a majority of Americans self-ID as “pro-life,” you have to believe as many Americans self-ID as Republican as describe themselves as Democrats.

So, a simple question for Ponnuru: Do you believe that, right now, as many Americans self-ID Republican as describe themselves as Democrats? If so, how do you square that with the vast polling data to the contrary?