As an interested external observer, there sounds like there is constant blanket polling going on, by many different companies, for all the political races, on each and every day of the election cycle (and perhaps even continually between elections). So, how many times would the average American citizen receive a call from a pollster? And how much would this figure increase in an important swing state like Ohio?
I’m really wondering if, at this stage in the cycle, and if the above number is ‘a lot’, whether people have polling hatred/fatigue, and basically decide to mess with the stats with silly answers. Would 20 people within a polling number of 800 responses be able to effectively skew results sufficiently? Or would it need to be 100 out of 800 answering falsely to make a statistical difference? Is ‘lying’ by participants accounted for in the standard models?
Okay, that was lots of questions. Thanks for any insight.
It’s possible it could happen, but since polling is still pretty accurate, they keep on doing it. if they ever reached a point where respondents became useless, it would show in wildly inaccurate polls and they’d have to change tactics or throw their hands up.
I was polled three times last night (and that wasn’t unusual) and I live in the Chicago suburbs. Illinois is by no stretch of the imagination a swing state.
I can’t imagine what the people of Ohio, Virginia, Colorado, et al are going through.
Econometrics just refers to the use of statistical methods by economists. See psychometrics, biometrics, chemometrics … we in such fields can’t feel smart about ourselves without complicated-sounding terminology. E.g., “Who do you think you are calling me a glorified applied stats person? I’m an econometrician!”
Iowa voter here. Certainly a swing state. All I can say is, Thank goodness for caller ID.
Just during the time frames when we are home, we probably get about 6-12 calls that are almost certainly pollsters every day. (I say almost certainly because we don’t answer unfamiliar numbers, and the same numbers pop up time and time again.) Admittedly, some of those are repeat calls that we wouldn’t get if we answered the first time, but … the phone rings an awful lot in the Jocko household these days.
If only we did respond with our preferences, Obama would probably be up by ten or twelve in the likely voter numbers, so all is well, fellow Democrats! The numbers you see are skewed because I’m not answering the phone in Iowa!
Heeee! Oh, really? All of a sudden, we should assume the best result is wrong? Oh, oh, and you *think *another pollster has Obama up by 4? Somethings tells me you know *precisely *what pollster has Obama up. You guys are on *top *of these polls like a sumnabitch.
Ya’ll killing me. The ultimate poll is that ballot drop. If it aint Obama, Imma be knockin’ back a scotch, shedding a tear. In the mean time, I’m cracking up watching folks get polled in the ass. With no Vaseline.
[QUOTE=Uncle Jocko]
Just during the time frames when we are home, we probably get about 6-12 calls that are almost certainly pollsters every day…
[/QUOTE]
Thanks for the replies. Another follow-up - if you did answer and complete their survey accurately, will they mark your number as ‘answered’ and give you a break from other calls, or will it add you to a ‘happy to respond’ list that will only increase the number of future calls you receive?
I imagine that polls would get more representative the more (different) people that answer, so a repeat call to a known ‘answerer’ might not be a suitable or accurate way for a scrupulous pollster to proceed.
Let’s say each poll polls 1000 people. Let’s say there are 100 polls a day. 170M or so registered voters. So it would take about five years of incessant polling like that for the polls to reach every registered voter, if you make sure that no dupes occur.
As mentioned a single poll doesn’t mean much, it’s best to look at an average of multiple polls like the RCP average or TPM Polling Average which have Romney up by 0.5 and 1 point nationally.
Since it’s the electoral college that matters, you really need something like Nate Silver’s model to figure out who’s winning, right now his Obama winning percentage is 66%. I would also take a look at different betting markets. Predictwise has a nice pagewhere they average out the winning percentages from three political markets, they have Obama winning at 66.5%. I think an average of 538 and Predictwise is the closest you can get to a single number that summarizes the state of the race. Purely as a matter of personal opinion I think Obama’s chances are a quite a bit higher, around 75-80% but that’s just my subjective take on the race.
Okay, fine, if Romney is up seven, then he should be leading in all the states that matter, right? Maybe I or someone can find their latest state-by-state numbers?
The reason it’s hard to believe is that it means that Barack Obama has been a terrible President for three and a half years and suddenly people are figuring this out a few weeks before the election.
If Obama was as obviously terrible as you’re saying, wouldn’t people have known it all along and given Romney a solid lead throughout his campaign? And if Obama has had some means of concealing his poor performance all this time, why would it suddenly stop working?
A more likely explanation is that most people don’t see either candidate as being as horrible as their opponents believe. Overall, people appear to have seen Obama and Romney as being approximately equivalent with Obama having a slight but solid lead. Recently, Romney has appeared slightly better and Obama has appeared slightly worse and this has narrowed the gap between them - perhaps to the point where Romney is now preferred.
Rasmussen has him up by 2 today, and IBD/TIPP has it tied. I’d guess that nationally it’s somewhere around the Rasmussen figure.
OTOH, Obama’s still got an edge in OH polling, even today. That’s really the state that matters the most, and it could be that his auto bailout is paying real dividends for him here.
Not necessarily. Romney could just be improving his lead in states where he was already the preferred candidate. If he went, for example, from having a sixty percent lead in Texas to having a seventy percent lead in Texas, that would bump up his overall national numbers but it wouldn’t mean there had been any change in Ohio.
