Michigan, site of one awful stadium rally, and some truly abysmal polling.
As we continue to decompress from what wound up being a highly enjoyable Election Night in America, the balance of 2012 will likely be dedicated to postmortems of a truly interesting electoral cycle.
Down the line, when the official numbers have been tallied from state-to-state, we will look at how the pollsters, as a collective, performed. Indeed, the only criticism I would make of many of the (quite good) polling analyses I have seen post-election is that the data is (even more than 10 days after the fact) incomplete.
For example, if you examined California polling on Election Night, you were looking at a 57-41 race. As of last Thursday evening, it was 59-38, with the very real prospect it will be 60-38 when all is said and done. That's a six-point shift since Nov. 6, which is not insignificant.
So, patience on the "big picture" polling analysis. Those polls (all 2,900 of them, for the two-year cycle) aren't going anywhere.
What we can do, for now, is place a spotlight on the truly craptacular. It took a lot of deliberation, and the honorable mentions list is as long as it is marvelously undistinguished. At last, though, what follows is one man's opinion of the five worst polls conducted in the final hours of the 2012 campaign marathon.
(Continue reading beneath the fold.)
First, the sole limit to the "contest" here: The polls had to be be included in our nightly Polling Wrap in either October or November of 2012. Honestly, it simply isn't fair to ding a polling firm for a lousy April poll for a November election. Let's face it: Nine times out of 10, the predictive value of a poll that far out is probably not worth much of a debate.
As long as a poll was publicly released, it was eligible for consideration. Only polls for federal offices or gubernatorial races were under consideration. With apologies to those whose hearts were broken (or spirits were lifted) by a horribly errant state Senate poll, there are limits to how many races the Polling Wrap could cover at once.
To make the top five, you had to do more than simply be way off the fairway. Lots of pollsters completely crap out when polling a race (for example, damned near everyone at the end in the Nebraska Senate race blew that one). There had to be awfulness that rose above and beyond the call of duty. The final quintet that you'll read about below fit that standard.
But, first, just to take a pleasant stroll down memory lane, let's recap some of the most criminally errant topline results of the 2012 election cycle. Almost all of these are campaign internal polls (making the huge errors somewhat more understandable), but here are a series of polls released in the final month, nearly all of which have the common bond that they missed the final tally by double digits:
- Merriman River (Hawaii President, 11/2/12): In their defense, it is tough to get a proper margin for a blowout election. But they had it Obama +27, and he wound up winning by over 40.
- Kimball Consulting (R) (Massachusetts Senate, 11/2/12): They were bullish on Brown throughout, but their final poll (Brown +2) wound up being off by a full 10 points.
- Wenzel Strategies (R) (Ohio President and Senate, 11/1/12): Romney up three in Ohio?! Josh Mandel up by 5 points?! Bwahahaha!!
- SurveyUSA (NV-04, 10/31/12): SUSA actually had a damned good cycle, but they (and lots of folks) underestimated the Democratic base turnout in the newly drawn 4th. That's why they had Republican Danny Tarkanian leading throughout. And that led them to be off by double digits in this key hold for the Democrats.
- OnMessage (R) (MN-08, 10/29/12): Republican Chip Cravaack's pollster had him up 10 with a week to go. He lost by eight points. Not. Even. Close.
- Victory Enterprises (R) (IA-01, 10/26/12): Republican challenger Ben Lange came close to Democratic Rep. Bruce Braley in 2010, and his poll with less than two weeks to go had him up by a pair. He lost by a 57-42 margin.
- McLaughlin and Associates (R) (NY-06, 10/22/12): The winner for the single most errant poll of the cycle? It goes to McLaughlin and Associates (who also makes an appearance in the top 5), which missed by a mere 34 points with this poll, which showed Republican Dan Halloran down just three points to Democrat Grace Meng in the newly created New York 6th district. She wound up winning by a quite unambiguous 68-31 margin. There are plenty of adequate reasons to make excuses here: Turnout in NYC was unpredictable due to Sandy, an open seat in a wholly reconfigured district is tough to predict, etc. And all of that is true. But it can hardly excuse being off by 34 points, right?
With those polls that drove it deep into the rough now off of our plate, let's proceed to the five polling air balls that set themselves apart from the pack, with a brief explanation of what made them so gloriously poor.
5. Finkelstein and Associates (R) (CA-36: 10/16/2012)
Okay, okay, one uproariously bad polling call had to make the cut, if only because it is reasonable to assume that it served as a microcosm for the poll-driven catastrophe that was visited on Republicans this cycle.
This internal poll for the campaign of now-former Rep. Mary Bono Mack put her up 55-38 over Democratic physician (and Congressman-elect) Raul Ruiz. One has to believe that, to get a number as absurd as that one (off by a mere 21 percent, at last check), Team Bono must've neglected to poll any Latinos, a key constituency in this Inland Empire district.
And why were GOP internal polls, and those of certain public pollsters (cough ... Rasmussen! ... cough) so far off the mark? Because they assumed a far more pale electorate than the one that actually cast votes in this election.
The glorious kicker was this quote from one of Bono Mack's campaign flacks:
“Our polling has consistently shown Congresswoman Bono Mack with a strong lead despite the unrelenting attacks of Nancy Pelosi’s allies and liberal groups,” said Marc Troast, political director for the Bono Mack campaign. “I would say the Democratic polling looks more like wishful thinking.”
The Democratic poll Troast was attacking? Ruiz +3. Actually a little more than a point under where he actually performed.
4. Foster McCollum White/Baydoun (Michigan President: 11/4/2012)
This Democratic autodial firm was off the fairway the entire cycle, to be frank.
FMW/B offered us a late Summer poll that would have easily qualified as the whopper of the year had pre-October polls been factored into the mix.
Yes, polling fans, this was the notorious "Romney +14 in Florida" pollster.
And they were also the pollster that, on Election Eve, had Mitt Romney fractionally ahead in Michigan. Of course, when all was said and done, Romney lost the state of his birth, and it really wasn't all that close (Obama +9).
FMW/B had one recurring problem that they just couldn't shake—their samples were always comically old. Unweighted, their sample was almost wholly comprised of respondents over 50 years of age (nearly 82 percent!).
But, even when they weighted, they still rather drastically overshot the age thing. Even winnowing down the 50+ slice of the electorate, they were still dramatically higher than the exit polls showed. FMW/B assumed that 29 percent of the electorate would be 65 years of age or older. It wound up being ... 17 percent.
The only age group Romney carried in Michigan on Election Day? Those 65 years of age or older. Well ... that explains that.
3. McLaughlin and Associates (R) (Indiana Senate, 11/2/12)
As is often (accurately) noted, a shrewd reader of polls has to take a campaign's internal polling with a several grains of salt. For one thing, the construction of the poll can (though a truly professional pollster would be loath to do so) influence the topline numbers: If issues-based leading questions precede the trial heat question, for example. Not only that, but a campaign might sit on several lousy internal polls, and then elect to release the one (and perhaps only) positive poll for their campaign. They release them at their own discretion, and that, too, is cause for alarm.
My beef with the McLaughlin polls in the Mourdock-Donnelly race was not that the pollster had Mourdock still leading (46-44) on the eve of the election. It was the pollster's insistence that Mourdock's notorious "God's will" comment had barely moved the race.
The same pollster had been in the field in Indiana earlier in October, and found a three-point Mourdock lead (45-42) over Donnelly.
So, were we really supposed to believe that one of the most controversial statements of the entire 2012 campaign cycle ... such a prodigious verbal air ball that it got a ton of national press...was only worth a single point in the electoral fortunes of the candidate which uttered it?
Team Donnelly did not agree. Their previous polling showed a coin flip, but their post-debate polling showed Donnelly +7. Similar movement was also detected by a public poll (put out by a Democratic-Republican polling partnership) for the Howey Report, which took the race from a dead heat to a double-digit Donnelly edge.
As it happens, the electoral results proved otherwise, as well. Far from a one-point shift, Mourdock lost the Senate race by six points, even as Republican presidential nominee Mitt Romney carried the state by double digits.
2. North Star Campaign Systems (R) (MI-09, 10/3/12)
This internal poll for Republican candidate and Some Dude Don Volaric, were it not for the truly epic fail that tops the list, would've been the worst poll of the cycle in a landslide.
While we recall the caveat written above on internal polling, this one was so bad that someone should have called it out before it saw light of day.
Unless, of course, you bought stock in this poll's finding that Levin, a veteran Democrat, was really losing by 10 points to his underfunded GOP challenger among black voters. Or unless you believed that one of the most veteran Jewish members of the House of Representatives was really losing the Jewish vote to his GOP challenger.
By 22 percent.
There are so many demographic goodies in this one, that you really need to digest them all, not the least of which is the fact that they had Barack Obama, who had carried the new 9th district with 59 percent of the vote in 2008, losing to Mitt Romney by five points.
In what may have come as a shock to the Volaric campaign, but certainly wasn't to the rest of us, Levin nearly doubled the Republican up, cruising to a 62-34 victory on Election Day.
1. Mason-Dixon (UT-04, 11/2/12)
My guess is that Brad Coker, the head honcho at the venerable Mason Dixon polling firm, is thanking the lord on a daily basis for fellow pollster David Paleologos.
It was Paleologos (the head at Suffolk's polling unit), after all, that stepped on his johnson in a spectacular way by going on Fox News a month out from the election and conceding Florida, Virginia and North Carolina to the Romney campaign, saying his team would not even bother with polling those states in the future.
Coker, a bit more quietly (though he did have this notable verbal catastrophe regarding Obama's prospects in Florida), had a far worse cycle than Paleologos.
Suffolk, for the pro-Republican bluster of its director, actually had only a modestly Republican lean in most of its polling, and was at least on the fairway more often than not.
It is harder to find a poll that Mason Dixon got right than it is to find one where they got it wrong. Just ask "Senators" Denny Rehberg and Rick Berg, two Republicans forecast by M-D to be narrow winners on Election Night, only to lose narrowly to their Democratic opponents. While M-D's miss in North Dakota was not substantial in nature (they had it Berg +2, it wound up being Heitkamp +1), their error in Montana was way higher, a full eight points off of their final result.
But their worst poll of the cycle was also way off: by a full eight points. Or thirteen points. It depends.
Confused? So was Mason Dixon, apparently.
The firm made a huge splash in the final week of the campaign when they released a media-sponsored poll in Utah showing veteran Democratic Rep. Jim Matheson losing by double digits to Republican upstart Mia Love. With that poll release (and a subsequent release by local pollsters Dan Jones and Associates showing Love up by a more modest 5 points), House pundits began to shovel dirt on the prospects of the sole Democratic member of the delegation.
But then, in the final hours of the campaign, Mason Dixon offered differing numbers on the race. It was an almost unprecedented step: a revision of a previously released poll. Apparently, M-D had weighted the poll using a statewide assumption of party ID, when the 4th district (centered largely in Salt Lake County) was quite a bit more Democratic than the state at-large. Thus, they recalibrated, and found Mia Love up by seven points (50-43) instead. She wound up losing by a single point, meaning that even in their final, last-ditch attempt at "unskewing" themselves, M-D still fell way off.
* * * * * * * * * * * * *
But, hey, with over 1,200 polls during the time frame of consideration alone, it's possible I might have missed some truly deserving nominees. Feel free to comb throug the Polling Wrap archives (which you can find here and here). If you find a more suitable and deserving bit of numerical catastrophe, make your case in the comments. And enjoy the upcoming Thanksgiving week!