Democracy Corp provides some data on the difference in candidate preference among people with different telephone habits (cell only, mostly cell, mostly landline, landline only). Democrats made the same argument in 2004 to argue that the polls were overstating Bush’s support. That was probably true then, but not by enough to matter. Eight years later, a lot more people have given up their landlines. Has the phenomenon become statistically significant?
Personally, I think it’s more important that Obama is leading in the polls in enough swing states to win reelection. But, for people wondering who will win the popular vote, the cell phone issue might help them form an informed opinion.
The pre cell phone crowd
Well, Booman, how’s the homestead? Everything secure, everyone safe? You all have been in our thoughts.
Here in SW Ohio we got wet snow overnight, but it’s already melting with the cold rain. We are getting damaging winds, too, but we haven’t seen much tree damage.
We still are getting non-stop phone polling, but we have a land line. We can recognize the poll calls, mostly with caller ID, but also by number. I answered several polls, but now I’m ignoring them. It’s a nightmare on TV, too. Ohio has gotten more political ads than I can ever recall, and now it’s just pissing everyone off.
Good luck with the aftermath. Hope it’s not too bad.
I’m almost embarrassed to say that we didn’t lose power and had no significant flooding or damage. We were incredibly lucky. Unfortunately, we can’t say the same about the people of our region.
Considering that we have multi-day power outages with storms that people never even hear about, I think we don’t need to be enbarassed.
That said, I hope that the advance preparations made for Sandy allow for quick restoral of basic services for everyone who is without power or services. What a mess!
Nothing to be embarrassed about. You could just as easily been out trying to bail floodwater or pull downed trees off your roof. There are no guarantees and no promises when it comes to weather.
Glad you are all safe. Let’s hope those who did take the brunt of the storm get the help and attention they need.
I hope we can assume from this post that everyone at pond central rode out the horrible megastorm safely? thank goodness! It looks like enormous destruction over a vast area.
Forgive me if I’m not going to spend much time thinking about the latest iteration of the cell phone polling issue. Once burned, twice shy, haha. Vis a vis the election, I’m more concerned now with the projected level of vote suppression that is going to naturally result from having a $20 billion unprecedented superstorm hit the main body of Dems in the nation. Power outages projected for a week? Whatever is NJ going to look like?
The reality is that this prez election is up for grabs and will be extremely close. Nothing is going to change that now, whether cellphones are being called or not. Obama is not going to cruise to an easy, predictible victory ala Clinton in ’96 or Reagan in ’84. Far from it. Instead, as incredible as it seems, the disgusting Plutocrat n’ Theocrat team appear to have a good shot at winning the executive branch and basically controlling the gub’mint.
This despite their maintaining (if not openly advocating) rightwing extremist views that our “independents” supposedly abhorr, such as massively lowering taxes on the rich (again), radically and irreversibly altering Medicare and denying the urgency of global warming. And fielding a supposedly highly dislikable plutocrat holding a Bishopric of a crackpot folk religion. So much for attempting to understand the American electorate.
I sure hope that this storm produces a HUGE increase in the number of states in the Sandy-zone who do early voting. What if this was not today, but next Monday, which it easily could be? Shouldn’t we have a system in which the majority of votes can be cast with or without power? And what about the MORONS who are now dependent on ENTIRELY ELECTRONIC systems, which are totally hackable and dependent on power?
We just bought new smart phones? My daughter got a little water on hers, and POOF ! Small dead rock. Totally unsalvagable. We did not buy the extra “cover everything” insurance at $10/phone/month ($50/month total). So, she is now finding a cheaper phone. Water and electronics are a bad mix.
dataguy – have you tried the rice trick for the phone? I’ve used it to bring back a phone, an mp3 player, and even a laptop that were water damaged (the laptop in particular was impressive…it was partially disassembled and directly under a pipe that burst, and was full of water).
Basically, remove the phone cover and the battery, then put them all in a large tupperware container or ziploc bag full of rice. Let it sit for 36 to 48 hours. Then take it out, reassemble, and see if it will turn on.
Doesn’t always work, but $1 worth of rice and a couple days is probably a worthwhile investment.
The “rice trick” worked for me once on an MP3 player. Not so lucky another time.
Why no waterproof electronics? My watch is over 20 years old and I rarely take it off. Keeps on ticking like new.
Maybe I answered my own question.
Haha, well I’m sure that economics does play some part in it. You’re depriving watch companies of their god given right to your money, after all!
But at least as far as phones go, out of necessity they have to have a speaker, phone, some type of interface and / or charging jack. So it would be a lot more complicated to waterproof than a watch.
My ipod has been through the washer twice, still works great. But it’s a mini, so that has a lot to do with it.
I have a little prepaid tracfone which I use as my cellphone.
Now, is my MagicJack ($20 a year!) a landline phone? What about other internet phone services like Vonage? With those, you can often transfer your landline number, but you are no longer a phone company customer.
I guess I fall into the poor/cheap demographic, as well as the graduate degreed, full-time employed, union member category.
I Use Vonage. It is considered a land-line for polling purposes. I also use a (T-mobile) prepaid cellphone, which I spend maybe $25/year on because you only pay for actual usage, which is almost nil for me.
The cell phone report from Democracy Corps is encouraging but I’m waiting for others to weigh in on the topic who know it in depth and have a track record of objective assessments.
What does concern me, though, is that the wingnutosphere is convinced that Romney has a huge lead … as convinced as they are about everything else they believe. The conflict post-election day is likely to be huge.
The conflict could be major, and I would not discount the possibility of violence when Obama wins.
Political violence is a universal feature of fascistic, authoritarian, xenophobic movements.
American “conservatives” have been pretty vocal (and proud) of their endless threats to resort to “second amendment remedies” if they don’t get their way politically. Their “leaders” have irresponsibly engaged in this sort of bomb throwing as well.
Will our rightwing crazies put their ammo where their mouths are? Will their nationwide Noise Machine foment violence should Obama win just the electoral college? Or will “conservative” leaders KKKarl Rover and Boss Rushbo quail when the time for “action” arrives? Who has the tribal authority to counsel calm to the crazies? Rmoney? Ryan? Cheney?
Interesting days ahead….
I don’t know that much about polling, but I can read stats and evaluate the assumptions that go into a set of stats. The goal of any polling organization is accuracy so I think we can adopt a working assumption that the best of these groups work carefully on their methodologies. I found the article fascinating but there are 2 pieces that I wish they had explicitly addressed:
The national polls were there as an example, to show that undersampling cell phones creates a distortion in results. That example should apply to state polls as well.
Some polling organizations do cover cell phones, but as the article stated probably their cell phone samples are too small.
I don’t read Slate as a rule but followed a link to this article:
http://www.slate.com/articles/news_and_politics/victory_lab/2012/10/obama_s_secret_weapon_democrats_
have_a_massive_advantage_in_targeting_and.single.html
This is far more encouraging than anything I’ve read about number of field offices, etc. And my own experience is consistent with that … I receive on average 4 mailers per day from the reich wing on the election now. All they do is make me angry. The GOP should have figured out that although I’m registered unaffiliated here I was registered Democratic in other states and have given to the Clinton and Obama campaigns – but instead they look at my district and income profile and assume GOP. By contrast I’ve received only 2 mailers from the Democrats advising on voting rules and early voting. To be honest I don’t think that any of the GOP mailers offered that – but as I now toss them at first glance I can’t be sure.
They’ve even targeted my 18 year old daughter from the reich wing. They have no data on her voting history but their mailers have to be the worst possible for 18 year old female college student. From the left she’s received one mailing from Planned Parenthood detailing the services they offer that Romney will cut.
Ditto with the calls – just flooded with robo and live calls from the reich – only a few from the Dems (and they always start by thanking me for volunteering in 2008).
The best part of the article is that the reason the GOP is struggling is that these targeting techniques come from academic research into stats and social sciences … it appears that the GOP war on academics, science, and now math has taken its toll on their abilities. It’s kind of like the anti-science Taliban trying to develop advanced weapons – any one with that kind of education wouldn’t be allowed in the tribe.
This caught my eye:
And this:
Manufactured consent is still manufactured. What we have here is a marketing process masquerading as a political process. It does not inform voters so as to make informed decisons; it manipulates voter attitudes to obtain the desired behavior with or without informed consent. That is dangerous in a system that aims to be democratic.
i.e. demagoguery
no… marketing (we cast everything in marketing terms now)
Demagogue:
Guess who dislikes demagogues and demonizes them?
Hmmm! I didn’t know that definition.
I’m surprised that no one has considered whether there might be partisan bias among the “I’ve got caller ID; I don’t answer polls, canvassing calls, or strange callers” crowd. Before caller ID, that demographic didn’t exist.
Actually a lot has been said about it but there is no solution so the pollsters mush on hoping that their miniscule response rates are representative.
In fact this was central to the argument from the wingnutosphere about poll skewing – the idea that the GOPers are too busy working so don’t answer phones as much compared to the lazy welfare bums who vote Dem. We also hear this whenever exit polls indicate a higher number of Dem votes than the official results, which has been very often in the last 12 years.
There is no data to back this up either way. However, Rand and YouGov have come up with what they hope is an improved method which is to identify a large sample base at the beginning of the campaign and keep surveying them. This is similar to the Nielsen Family concept for TV ratings. The strength is that once the large sample base is established they get a steady response not subject to polling swings due to participation variations. As a result, you won’t see the sort of big polling bounces we’ve come to expect after conventions, and neither poll saw much of a debate bounce either. The unknown factors are a) what if the initial base sample was skewed, and b) what impact the act of continual re-surveying might have on the sample so that it ceases to behave like the larger population being measure. For YouGov and Rand, which both show Obama holding a steady lead all year long, this may be good news but we won’t really know until the election is complete.
One other point is that the response situations is worsening. I was surprised to learn that in 2004 Pew found that the average poll response rate was just under 20% and now it is under 10% – less than half of those who responded before. The robocallers have much lower response rates, whereas some pollsters call back the survey list 2 or 3 times to increase the response rate – these are generally considered better quality polls.
What all this tells us is that there is a ton of statistical uncertainty and the better pollsters will be trying everything they can to address this. Someone like Rasmussen, with their robopolling, ignores all this so it’s safe to assume that their purpose is not accuracy but instead to flood the news with as many cheaply-run polls as possible in order to skew the numbers in the polling aggregators.
The “too busy working” issue seems to me to be randomly distributed. Republicans act like not Republican voter has ever been unemployed.
My own guess is that the no answer response likely is also randomly distributed, maybe even the same result motivated by widely different reasons for each side of the ideological divide. The biases might cancel out.
But this is another example of the marketing model of campaigning really not serving the public interest. The political science profession seeks not to understand but to engineer, which makes actual solutions to deep political process issues more difficult.
Of course on the “too busy working” thing. I mean, it’s a stupid belief, but it is core to the tea party. They deeply believe that half of the country is living it up on free welfare and voting Democratic.
Realistically I don’t know that anyone has studied phone answering habits. I’m sure that there are age variations – until the 1970s answering machines were almost unheard of so when you got a call you answered it, period. Corporate voice mail started showing up circa 1990 with full adoption of “call trees” and so forth by the end of that decade. Home and digital cell phone voice mail started showing up in the early 1990s but didn’t become universal until a decade later. Who remembers the early days of Caller ID and the special caller ID device you could buy? What all that suggests is that the earlier you formed your phone answering habits the more likely you are to answer the phone every time.
Then for those who do answer, will they agree to participate? This has to be determined by a wide mix of behavioral factors. Of course most people you know say they never answer polls because the large majority of people either never do or do so rarely. But what are the characteristics of people who do answer and are they representative of the population? I don’t know that the pollsters know.
The pollsters do adjust for demographic factors to weight the polls but except for Rasmussen they don’t adjust for party ID since that is seen as a floating attribute, not solid. Rassmussen, it is pretty clear, designs their polls to lean GOP so that decision is consistent with everything else they do. What this means is that even if a poll sample gets few, say, Latinos, they can adjust for that by increasing the weight given to their few Latino responses. But that also means if they get a lower response rate for some groups then the error factor for those groups becomes much higher, and the very few respondents they get have a greater chance of being non-representative. This could explain some outlier polls showing Romney doing much better than expected with Latinos and Blacks.
All in all there is a ton of uncertainty here. Averaging the polls in an intelligent manner greatly reduces sampling error, but if there is an industry-wide systemic error, such as undercounting certain population segments, averaging won’t account for that.
But a systematic error over all polls can be measured by aggregators and applied to future years’ models–and will work unless is unique to one election. In which case the trend analysis of time series of historical data (of which it becomes one sample) will adjust that in the model.
I think we need to get ahead of the popular vote argument right away by pointing out the regional margins. There’s only one part of the country where Obama trails, and according to Gallup he trails there by over 20 points.
Just guess where that is or what might motivate voters there.
Let me see. Obama trails in Appalachia, the Mountain West (minus New Mexico and Colorado), Alaska, the Great Plains, and all of former Confederacy except for Virginia.
There are indications that North Carolina and Arizona might be in play.
If you want to play the “what motivates” game, you can look at the ring of suburbs around most urban areas and the “big agriculture areas” of states to see this. It’s what keeps GOP organizations going in all states, even the blue states.
I liked it better when you were confident of a landslide.
This last week is going to suck.
Especially here in PA where Romney has decided to join down-ticket Republicans in drowning us in lies.