The big story of the night–What’s are the pollsters doing wrong?
Published polls absolutely missed the dynamics of the governor’s race and failed miserably to predict the final 17% spread between Abercrombie and Aiona. In the end, it wasn’t even close, although published polls in the final weeks showed the race close or tied.
Here are the recent poll results as listed out by TMP PollTracker.
![[text]](http://ilind.net/images_2010/pollsoct.jpg)
Similar data were reported by the New York Times.
So what’s wrong with the methodology being used? Is it just that the Abercrombie campaign was different and somehow eluding a methodology that works in predicting more typical campaigns?
You could dismiss the polling failure as just a one-time quirk, except that its a direct repeat of the failure to predict the Abercrombie’s stunning primary victory margin.
Of course, since pre-election reporting is driven in large part by those poll results, which set the stage for reporting, editorializing, and general punditry, those poll errors led to errors in reporting the public mood and the tenor of the electorate.
There really needs to be some serious reassessment, although the sports metaphor that generally dominates election reporting is will make it easier for media to sidestep the issue.
I flagged a number of races that I thought might be close. I think all of them resulted in Republican challengers being rejected. Luckily, I finally got up to see the final results, and Marilyn Lee slipped past her GOP challenger by a handful of votes in a very tight House raise in Mililani. I thought she was losing when I last looked, which I found depressing, but that must have been the fog of political wars. Congrats to Marilyn.
Overall, though, it was a heavily Democratic night in Hawaii, which again stood apart from mainland election dynamics.
Discover more from i L i n d
Subscribe to get the latest posts sent to your email.
I’ve often thought that people should reject participating in polls because we’re better off forcing politicians to guess what we want in the hope that their uncertainty may lead to more honesty on their part. Plus the sheer joy in confounding pollsters and media pundits is not to be underestimated.
It’s going to take a little time to see how the polls’ methods affected the differences between the polls and the results. For now, here is a simpler explanation: turnout was much higher than expected and the people who were not expected to turn out but did were much more likely to be Democrats. Two bits of evidence to support this: 1. T.O. WAS in fact quite high for HI in a mid-term and 2. absentee votes made up a smaller percentage of the total vote than they did in the primary, which suggests that there were more late voters.
I wonder if there is a new group of people that just can’t be contacted using traditional polling methods. And perhaps this group fell more squarely in the Abercrombie camp?
Even if they include cell phone lists — maybe there are many cell numbers just not available?
nonetheless, it was probably a good thing that the polls called things close — because hard grass roots work on the campaigns is a good thing for our democracy — and that is true for both the winning and the losing side
I contend polling has gone entirely into the toilet in Hawaii. Blame arrogant pollsters, the advent of cell phones and consequent loss of land lines or sloppy work. Recent outcomes makes polls not worth reading. I am told some internal efforts were on the mark.
The political scientist types need to get off their duffs and help develop a responsible, dependable model. But keep in mind polling always has been tenuous in this state. How many times did the polls “elect” Frank Fasi as governor? How often did he win?
It turns out that there were some additional absentee votes not counted and Rep. Lee has held on to her seat by about 17 votes!
Great news!
. . . and that’s why the only poll that counts is the official one held on election day.
I was told last night by a core Abercrombie supporter that their internal polls had had him 15 points ahead of Aiona. I don’t know if the internal pollsters are willing to share their secrets by critiquing the others.
The Rasmussen Poll for the Inouye-Cavasso race was cited last night by polling uber-analyst Nate Silver as the Worst Poll ever. This is the poll which had shown Inouye as leading Cavasso 53-40. Inouye won 72-21.
Here’s Nate writing in last night’s NYT election live blog:
“If Mr. Inouye’s margin holds, the 42-point error would be by far the worst general election poll in FiveThirtyEight’s database, which includes all polls since 1998; the previous record was 29 points.”
I think there are at least three likely causes:
1) Hawaii residents have always been difficult to poll accurately on political races
2) Movement away from land lines to cell phones make it more difficult to get a representative sample
3) Translation of raw polling results into predictions of actual votes often involves an assumption about likelihood of voting.
And finally, I am wondering how many people like myself simply refuse to participate in a phone poll anymore, especially with the number of phony polls that are actually attacks by candidates?
Ian, it wasn’t a fog. Marilyn Lee was behind by 63 votes after the 3rd printout came out at about 11 p.m. Absentee votes pushed her ahead by 17 votes in the 4th printout, which came out at about 6 a.m. It’s not the final one though, so who knows what’s in store…
Thanks for that clarification, Micheal. Some of us misunderstood the first printout to include all the absentee voting. It did not. A lot of mail-in early voting was counted later.
The counting is observed by representatives of community groups and th political parties. In a situation like a House race this close, the results would have been handcounted in the early hours of the morning to satisfy all observers the results are fair and accurate. The fourth printout has a timestamp of 5:46 am.
The final results remain uncertified until the pollbooks and other records are audited. But I think these are the final numbers in this race. If the candidate or the GOP is not satisfied with the accuracy of the count, they can challenge the result. But I suspect they have already had their doubts satisfied.
Michael Levine’s explanation helps the rest of us understand the origins of the discrepancy. Thanks.
Given the huge error in the Abercrombie and Aiona poll, maybe the Lord works in mysterious ways? 😉
Or the pollsters are getting kickbacks from the television stations for helping to drum up business?
Reference: Bob Dylan “With God On Our Side.”
The rising number of cell-phone-only households has to be a significant factor. I have mainland family who are cell-phone-only and they never got one robocall or poll throughout 2010 election cycle. For the time being, apparently, mass lists of cell numbers are not available to politicians, charitable organizations & telemarketers.
Assuming that Marilyn’s 17 vote lead holds, the GOP will pick up two new seats in the House in addition to replacing Lynn Finnegan for a grand total of eight: George Fontaine who beat Joe Bertram in HD 11 and Gil Riviere in who handily thumped Larry Sagaysay for the open HD 46 seat.
However, the only Republican now in the Senate is Sam Slom. The loss of SD 25 to DaDonks is a surprise given the local issues in Kailua such as if there should be any economic development at all, the demographics of the likely voters in the district, and the comparative qualifications of the candidate.
Intially, I was also surprised that Pohai Ryan was able to pull out the win in SD25. Upon further review, it isn’t as surprising as one might think for a couple of reasons.
Reason 1: The diverse make up of SD25 includes Kailua (Coconut Grove & Kailua Beach), Lanikai, Keolu Hills, Waimanalo, and Hawaii Kai. The biggest factor was the Waimanlo vote. The precints in District 51 (1-3,5), which includes Waimanalo, accounted for nearly half of Ryan’s votes. I do not want to discount the political knowledge of voters in Waimanalo, but the fact remains that if someone isn’t familiar with either candidate, Hawaiians are more likely to vote for someone with the name Pohai rather then Virginia.
Reason 2: Virginia Enos is not Fred Hemmings.
I think the close polls motivated the Democrats to get to the voting booths. In my community, there was a wait to vote, with every booth full.
The tight polls motivated the Republicans too- but they were already on Red Alert, while many Democrats take a blue Hawaii for granted.
Take a look at the turnout in Massachusetts; once people start to understand what they stand to lose, they are plenty motivated.
I was in the water at Magic Island on Saturday when Charles Djou’s trolley came through the park. Ringing the bell, shouting and waving, it was clear they expected the beach to erupt with cheers. They were met with complete indifference. The people were whelmed- I didn’t see anyone even wave. It was like a high five left hanging. They went on after a pause, ringing and cheering.
I still think Dijou is made of cardboard.
Cardboard? That’s so mid-20th Century.
I suggest Max Headroom animation!
Jon,
I concur.
Nationally, the increased use of cell phones, particularly by younger people and minorities, is assumed to produce about a 4% pro-Republican bias into telephone polls. I ASSUME Hawaii residents have a higher rate of adoption of cell phones than most mainland states. Anyone have data on this?
Yes, some Hawaii residents, for cultural reasons, refuse to answer phone polls.
But if the Abercrombie (and, now I am told, Hanabusa) internal polling was more accurate, how do we account for that?
I think Mahina’s point probably has a great deal of validity. A lot of Democrats were taking victory for granted UNTIL the polls game out showing the major races to be very close. The closing polls may have forced Democratic insiders to stop their squabbling and unite for a final unified push. Hence Mufi’s return for the wilderness with STRONG endorsements for Neil and Ed Case’s ads for Hanabusa.
These factors MAY have tightened up the races, leading to final results different from the polls.
But how the heck did Rasmussen get it so wrong? Can they recover from this? SHOULD they be ALLOWED to recover or just laughed into oblivion?
It was very disconcerting to have cell phones do robo and political calling because I generally answer those calls on my land line. But I did answer the phone once because it said Civil Beat and they had the guts to print a story on Richard Turbin’s illegal vacation rentals. (Mahalo Ian for also verifying the truth too!) But even the Civil Beat poll had Abercrombie winning by only 5%.
Truly enjoyed reading your insights on the election, Ian. I don’t always agree with your assessments, but I can’t fault your investigative work. And the discussions you provoke are much more interesting than those at the SA. I think it won’t be long before your blog is the #1 blog ranked in Hawaii.
The cell phone factor doesn’t really hold water when the polls were more or less right in other races both in the primary and the general.
I mean, if cell phones made that much of a difference, then every poll in every race in the country would have been similarly skewed.
It would, in essence, be the death knell of polling as we know it. Not a bad thing, mind you. 🙂
Maybe it’s polling fatigue. How else to account for the fact that 2% of the people polled say they’re undecided and yet Neil wins by such a large margin?
Maybe it’s the obsession we have with polling so now the polling tail wags the election dog. Hard to believe but when Al Gore first ran for the U.S. Senate in 1985, he did no polling at all.
Here we are 25 years later and we poll incessantly.