John McLaughlin, who would be the worst GOP pollsters if any of the rest were any better.
The hits on Republican pollster John McLaughlin keep on coming, and a week later, they are
still hilarious.
MCLAUGHLIN, ROUND 3: Yesterday, we presented GOP pollster John McLaughlin’s rejoinder to criticism of his 2012 polling -- in light of his big miss in House Majority Leader Eric Cantor’s race. McLaughlin said that polls in early- and mid-October showing Indiana Senate candidate Richard Mourdock and Virginia Senate candidate George Allen ahead were conducted too early to be considered predictive, and both races moved away from Republicans in the final weeks.
But POLITICO yesterday obtained brushfire polling memos from McLaughlin dated about a week before the election: In a memo dated October 30, both Allen and Mitt Romney led in Virginia, 50-44. In an October 31 memo from Indiana, Mourdock led, 47-45.
McLaughlin’s response: “As for Virginia I said: ‘Our last internal poll was over a week out and the leads were lost in that week.’ This seems to be factually correct. Wasn't October 30 over a week out from the November 6 election? Hurricane Sandy hit Virginia and we didn't track there after that. I did hear reports that other Republican organizations did poll later right before election day and it wasn't good.”
“Your source appears to be proving my statements to you to be not only accurate, but also honest,” continued McLaughlin. “If your source thinks we should have won these races in a year when Republicans lost the Presidency, lost seats in the Senate and the House, the source should go on record and make their case. We do our best to help our candidates win, but most of the time we're just reporting a snapshot in time when public opinion can change.”
"Polling 3-4 weeks out is not predictive!"
"So how about this newly leaked polling from one week out?"
Ouch.
Head down below the fold as I break down this latest laugher response from this (very-typical) GOP pollster.
“As for Virginia I said: ‘Our last internal poll was over a week out and the leads were lost in that week.’" This seems to be factually correct. Wasn't October 30 over a week out from the November 6 election?
That final McLaughlin poll had Romney and George Allen leading by six. Romney lost Virginia by four, and Allen lost the Senate race by six.
So is he really arguing that in a single week, there was a 10-point swing in the presidential race, and a 12-point swing in the Senate race?
"Hurricane Sandy hit Virginia and we didn't track there after that. I did hear reports that other Republican organizations did poll later right before election day and it wasn't good.”
Hurricane Sandy glanced by Virginia. There were two hurricane-related fatalities (compared to 63 in New York) and only 180,000 customers lost power, most in Northern Virginia, that well-known hotbed of conservatism, right? (NoVa is heavily Democratic.) In fact, Republican Gov. Bob McDonnell
said, after the hurricane, that Virginia had been "spared a significant event".
But even on October 30, when McLaughlin had the two top-of-the-ticket Republicans up six in the state, the public polling showed an Obama lead. The polling aggregate had a 48.5-46.9 Obama lead, a margin narrowed by junk GOP pollsters. Check this out:
Among Democratic pollsters, PPP had the race at 49-46, a three-point edge close to the final 3.88-point margin of Obama's victory, Garin-Hart-Young was Obama +3. Mellman was also +3 Obama. Good job, Democratic pollsters! The unbiased media folks were also good. Washington Post had Obama +4 on 10/26. CBS had Obama +2 on 10/28.
But on the GOP side there was Zogby with a +1 Romney lead, Rasmussen +2 Romney, Fox +2 Romney. Gravis had it tied. Go out another week, and Wenzel (polling for Citizens United) had Romney +2. You get the picture.
So either the Republicans were all right a week out, and the Democratic and media pollsters were all wrong, or Frank Luntz is right, and GOP pollsters really DO suck.
To be clear, it's the latter.
“Your source appears to be proving my statements to you to be not only accurate, but also honest,” continued McLaughlin. “If your source thinks we should have won these races in a year when Republicans lost the Presidency, lost seats in the Senate and the House, the source should go on record and make their case.
Oh, so despite having polling showing the two Virginia Republicans winning by six, there was no chance of Republicans actually winning those races? In other words, there's no way they could've won, because they lost? Interesting logic.
But also bunk, because we know for a fact that two days before the election, McLauglin was convinced that Romney was headed toward victory. How do we know that? Because he told us.
Me on November 4, 2012, two days before the election:
To have any hope for Tue, GOPers have to ignore the data. Don't worry, they have plenty of practice doing that.
— @markos
McLaughlin's immediate response:
@markos when Romney wins will u fire PPP like u did Research2000? #markosskewshispolls Republicans coming out #howsyourearlyvote? Not2008
— @jmclghln
So I was an unskewer who needed to fire my pollster because Romney was going to win, but there was no way that Romney could've won because Romney lost. With logic like that, it's no wonder he keeps getting Republicans to hire him, including the House Majority Leader.
And by the way, Politico's "source" in all of this? All she or he did was leak McLaughlin's final polling memos in Virginia and Indiana. So if McLaughlin wants to debate the source, he just needs to debate himself, which is actually what he appears to be doing.
McLaughlin 1: Romney and Allen are headed toward GLORIOUS VICTORY!
McLaughlin 2: That's crazy talk! In a week's time, Romney and Allen will lose, so how could you argue they will win?
McLaughlin 1: Bah, that's just liberal UNSKEWERS, who will all have to fire their pollsters as a result of our reality bending data.
McLaughlin 2: I CHALLENGE YOU TO A DUEL!
"We do our best to help our candidates win, but most of the time we're just reporting a snapshot in time when public opinion can change.”
McLaughlin's snapshot in time was wrong, contradicted by all real pollsters. Sure, it helped reinforce a conservative bubble convinced of Romney's impending victory, but self-reinforcing bullshit is still bullshit.
What's amazing is that McLaughlin is still attempting to rationalize away reality. But perhaps it's not so surprising. He built an entire career around it.
The Daily Kos staff sat around laughing at all of this, and Jed Lewison summarized McLaughlin's argument best:
Look, if somebody wants to make the case that they believed Republicans were going to win these races, then they should make that case. As for us, our polling speaks for itself. Clearly, Republicans were doomed, and the results prove that.
Brilliant, all of this!