I dialed into that Clinton conference call this afternoon and was left absolutely speechless by some of the things Mark Penn said. For example, he argued that we should disregard polls showing Obama doing better than Hillary against McCain. That's fine--I don't actually put much stock in polls about hypothetical matchups. But Penn was making a much more ambitious point--arguing, essentially, that polls are unreliable in general. He cited several data points: Polls showed John Kerry winning against George W. Bush in 2004; polls showed Hillary losing to Rick Lazio in the 2000 Senate race; polls showed Obama doing well in California and Massachusetts last Tuesday; early Super Tuesday exit polls showed Obama rolling to big victories across the country...

Wow, where to start? The Bush-Kerry polls showed a pretty close race in the final week or so of the 2004 campaign (which it basically was), and a lot of them got the result right. So I don't think you can summarily dimiss the 2004 polling. Likewise, most of the polling going into Hillary's 2000 election showed her up slightly (see this New York Times editorial from the day before). She won relatively comfortably, but it's not like the polls were highly misleading. As for February 5, the vast majority of Masachusetts polls showed Hillary winning comfortably (see here for the average of the last few polls there, which were Obama's best Massachusetts polls); it's only California that was really screwy--and that seems forgivable since the outcome hinged on a remarkable and unexpected surge in Latino turnout. 

And who said anything about exit polls? They've definitely been misleading at times. But most journalists know that, or have learned it by now. Anyway, it's not like anyone's been citing early exit polls showing Obama beating McCain. I'm not even sure what that would mean at this point.

(In fairness, the New Hampshire polls were pretty misleading, though I don't remember Penn mentioning them...)

More importantly, consider the logical fallacy in what Penn's saying, which is basically: X is sometimes misleading. X shows Y. Y must be wrong because X shows it. It's a bit like a woman refusing to believe she's pregnant because pregnancy tests are sometimes wrong. While it's true that you might not be pregnant if a pregnancy test says so, the occasional fallibility of pregnancy tests is hardly the strongest argument for why you're not pregnant. Call it the Juno fallacy. 

Finally, Penn's ostensibly a pollster. Have you ever heard of a pollster who questions the value of polling? Would you hire one who did? And what's the alternative to polls? Going with your gut? Taking the word of campaign flacks? I'm not sure what the broader take-away was supposed to be here.

Three more quick nuggets from the conference call:

1.) Howard Wolfson seemed very eager to remind voters of what Obama strategist David Axelrod said this morning about superdelegates, which is that their role is/should be to exercise their judgment about what's best for the party. The conventional wisdom is that Obama wants superdelegates to stand down and let the voters decide things, while Hillary wants to leave the door open to winning on the backs of superdelegates. This was either a slip on Axelrod's part or a reflection of a change in thinking within the Obama campaign. Either way, the Clintonites were eager to embrace it.

2.) Several journalists on the call asked why Wisconsin would favor Obama rather than Hillary. Wolfson seemed torn between arguing that Hillary could do well there and raising expectations for her. That's okay, though, I'm happy to field this one for him. The short answer is that Wisconsin's a lot like Iowa, except younger and with more black voters. The long answer is the subject of this blog item from last week. The only thing I'd add is that Wisconsin also has slightly fewer working class people than Iowa--49 percent of primary voters there made under $50,000 in 2004, versus 53 percent in Iowa in 2004. (True, Iowa holds a caucus and Wisconsin holds a primary. But Iowa had such high turnout this year that it effectively became a primary.)

3.) Wolfson was asked at one point why Hillary wasn't more competitive in Maine, where the demographics seemed to favor her. His short answer: It's a caucus. We do better in primaries, where the maximum number of voters can show up and express their preferences.

Update: Commenter psantillana suggests Axelrod wasn't necessarily reversing the campaign's position on superdelegates. That's certainly possible, particularly given the weird way in which the question was asked by Matt Lauer. But the implication is that he doesn't thinks superdelegates should be bound by the final pledged delegate tally. (Otherwise why make an electability argument?) It could be hard for him to walk back that implication down the road.

Here's the transcript of what he said:  

LAUER: Let me ask you about superdelegates. A lot of discussion about the role they're going to play in all this by the time it's over and the debate being whether they should vote out of loyalty to a candidate, particularly if they're elected officials, or whether they should vote the way their state or district votes.

So, when it comes to Senators John Kerry and Senator Ted Kennedy, should they vote for had Barack Obama because they've endorsed him or should they vote for Hillary Clinton because the state of Massachusetts voted for Hillary Clinton?

AXELROD: I think that the role of the superdelegate is to act as, sort of, a party elder. These are elected officials from across the country and they're supposed to exercise their judgment as to what would be best for the party. And as they look at this, they need to decide who would be the strongest candidate for the party…

LAUER: David, you're not answering. Should those two senators vote for Barack Obama or Hillary Clinton?

AXELROD: I think they and all the superdelegates should vote according to what they think is best for the party and the country. And I think that we need the strongest possible candidate against John McCain…

--Noam Scheiber