Poll Inaccuracy's could lead to unforseen landslide?

Mr. Incognito

Okay,

Most of the polls that are conducted are over landlines.

Considering Obama's base is students who wouldn't have a land phone (mobiles) , and black minorities/majorities? who would most likely be unregistered in a lot of states and as such would not be polled, added to the fact you can bet your bottom dollar that a lot of black people WILL vote in this election for the first time and the demograph of America, >40% of the popultation are in the 18-35 bracket does anyone else think that this race isn't as close as the polls suggest?

A lot of the polls showed Hiliary ahead in States that Obama won comfortably in the primaries - the difference? Student mobilisation networks and the black vote.

I have a feeling Obama could win in a landslide.

Am I alone in thinking this?

The_Minister

They have already ajusted polls to reflect Obama's effect on demographics already. They had to do it after several of the primaries had unexpected results.

Mr. Incognito

Oh, adding imaginary tweak numbers to actual poll results.

How.........um......accurate?

I can see an overhall on polling practices after this election.

is_that_so

The primaries are not a good place to look seeing as they were pitching to their own people. And then Hillary ignored the threat until it was too late and Obama had momentum.
The problem with that base is that they may not show up to vote, as Gore discovered in 2000. It is also possible that some people, could be lying so as not to appear racist. Do some research OP, landslides in US elections are rare. This one will be tight.

banquo

I think they merge the numbers with internet polls or something.

Ludo

Have a read of this:

http://electoral-vote.com/evp2008/Pres/Maps/Sep11.html

about half way down under the heading: Polling and Partisan Identification

Very good site that actually.

Mr. Incognito

Granted, but I mean the conventonal things that would affect it like low voter turnout for students and blacks is going to be negated in this election.
The student vote for the last round of voting in Congress was up 400% in places and don't tell me black people aren't going to be beating a trail down to the booths.

Mr. Incognito

Have a read of this:

http://electoral-vote.com/evp2008/Pres/Maps/Sep11.html

about half way down under the heading: Polling and Partisan Identification

Very good site that actually.

Good link that but it doesn't really adress my original point- they aren't polling young urban voters or unregistered blacks

Ludo

SetantaL wrote: »

Good link that but it doesn't really adress my original point- they aren't polling young urban voters or unregistered blacks

True. Just shows the multiple ways polls can be "biased"

Bubs101

The way they conduct these surveys is by telephone banks, which are banks of numbers and demographics. The surveyors choose very very carefully where to call and when to do it. An example would be not calling during the day because you will get too many women. I can guarantee you that, unless they want a biased survey for headlines, they will have carefully selected their demographic's so they can get a fair representation of the expected voters. Any changes will be tiny, 1-2% max from the preceding poles barring a massive surprise.

Also, keep in mind that the Democrats nearly always have the black vote and the blacks tend to be concentrated in certain states (Illinois, Maryland etc.) so it might not have that big of an impact on the country as a whole

BenjAii

OP, there is the well documented Bradley Effect to consider.

Pollsters reporting far higher support for black candidates then they actually get, because some people don't want to seem racist when talking to someone else.

It seems there are a still many undecided voters and it is these people who will decide the election, so what polls are saying now won't predict what will happen.

Mr. Incognito

The surveyors choose very very carefully where to call and when to do it. An example would be not calling during the day because you will get too many women. I can guarantee you that, unless they want a biased survey for headlines, they will have carefully selected their demographic's so they can get a fair representation of the expected voters. Any changes will be tiny, 1-2% max from the preceding poles barring a massive surprise.

I find this hard to believe. I'm sorry but I don't think ringing people up and asking them how they are going to vote is an accurate indication of the actual vote. A gauge with peramaters sure, but saying that's it's an accurate science is just too much for me.

Also, this doesn't address the whole point of the thread. The peole I think are going to turn up and vote for Obama do not have landlines so your poll data is worthless and they are slipping outside the net. Great if they don't turn out to materially affect the difference between polling middle-class white americans polled and that demographic subsequently voting along those lines, but when a block turns up, en masse, that fell outside the polling paramaters in the first place the polls are worthless.

The question is- how worthless. Blacks are 12% of the population, the young vote conservatively 25% that's 40% of the country right there. Although that don't mean ****e if they don't actually vote.

Also, keep in mind that the Democrats nearly always have the black vote and the blacks tend to be concentrated in certain states (Illinois, Maryland etc.) so it might not have that big of an impact on the country as a whole

Aye, the blacks that bothered to vote in the past from a process they felt disenfranchised from, but now that it's a black man for president, you can quadruple the black turnout.

Ludo

Bubs101 wrote: »

The way they conduct these surveys is by telephone banks, which are banks of numbers and demographics. The surveyors choose very very carefully where to call and when to do it. An example would be not calling during the day because you will get too many women. I can guarantee you that, unless they want a biased survey for headlines, they will have carefully selected their demographic's so they can get a fair representation of the expected voters. Any changes will be tiny, 1-2% max from the preceding poles barring a massive surprise.

But to get a fair representation of the demographics you almost need to know the result of the poll before starting it. I.E. to make it fair you need roughly 53% dem leaners and 47% rep leaners...but you need to poll to find out this, but to make that fair you need to know the percentages...aaargh my head hurts.

Just stop polling so damn much. Let the people decide the outcome and not the polls!

dan_d

And statistics will prove anything anyway.They can pretty much be manipulated to prove whatever you want, unless it's an extremely controlled environment with very few/no outside influences.Really, they'd need to tell you the exact number of people polled, the method of contact (so you can discount those who aren't contactable by a particular method), the time of day polled at, the location of those polled, their gender and the exact questions asked, probably their race, occupation,...have I forgotten anything?I'm getting confused...all of these things are influencing factors in the final result.

Ludo

http://www.youtube.com/watch?v=2yhN1IDLQjo

Class stuff

A Primal Nut

I think its possible that one way or the other, this race might not be as close as people expect. Although the popular vote might be close, I can see Obama getting enough of the newly-registered black vote to bring him over the line in enough of the battleground states to give him a landslide in the electoral college votes. Or it could go the other way. Its a fascinating race.

I was in Chicago for three months over the summer; a bad example as Illinois is a Democratic state and Obama is a senator there. However, support for Obama across all demographics was huge, as in everyone supported him. Once, there were suggestions on the radio Obama, confirmed by one of his advisors, that they would campaign in Texas, perhaps the most well known Republican base of all, although they admitted it was unlikely they would win there. This race might not be as close as people think.

The opinion polls are a complete guess and didn't do too well at predicting the primaries, whether Hillary won or not.

amacachi

Of course the blacks as well as everyone else could get sick of Obama by polling day. He may have oversaturated too early.

Mr. Incognito

Of course the blacks as well as everyone else could get sick of Obama by polling day. He may have oversaturated too early

That is quite possibly the silliest thing I have ever read on boards and in the politics forum too.

As far as I know the only person who ever got sick of being black and was able to do something about it was Michael Jackson.

wiredup

SetantaL wrote: »

I have a feeling Obama could win in a landslide.

Am I alone in thinking this?

Yeah, I've been feeling it for months. The reason is obvious, if the republicans get in again after what they have done you can only conclude the citizens of the USA are among the dumbest on the planet. I don't believe that myself which is why Obama will win with a landslide.

brianthebard

SetantaL wrote: »

Aye, the blacks that bothered to vote in the past from a process they felt disenfranchised from, but now that it's a black man for president, you can quadruple the black turnout.

Lolz I'd like to see your poll results for that figure. As already pointed out landslides don't happen in US politics.

Benedict XVI

wiredup wrote: »

Yeah, I've been feeling it for months. The reason is obvious, if the republicans get in again after what they have done you can only conclude the citizens of the USA are among the dumbest on the planet. I don't believe that myself which is why Obama will win with a landslide.

I really can’t see where people are getting all this optimism about Obama.

It is probably because all the media around here seems to be leaning towards the Dems.

In reality he hasn’t a hope. On this thread people seem to be saying the black vote will get in in, and that all the eligble blacks in America will register and vote for him.

That will not happen, may blacks will not vote for him because they may see him as elitist, which he is, he has little in common with your average black man in a poor inner city neighborhood

Also mentioned are the young people, these are probably the same young people that were supposed to deliver John Kerry, well do you know what happened to them, they didn’t even bother voting.

At the same time people in the south and the white working class will be out voting for McCain because they see him as a safe pair of hands and an experienced politician.

I see Obama conceding this one earlier on the morning after the vote rather than later, that is how much McCain will win by.

Hobbes

Funny cause I see Obama winning.

I mean are we sharing different worlds or something? Banks are tanking in the US and McCain is going as if everything is fine.

https://www.youtube.com/watch?v=6reQLzgywzk

As for polls. mobile phones are polled. But polls can be flawed if you don't read the questions asked.

For example:
http://www.politico.com/blogs/bensmith/0908/Jewish_voters_complain_of_antiObama_poll.html

Manic Moran

Quoting myself from the other thread.

Although this is true, there is a factor which cancels out the amount of cell-phone-only people who may vote Obama: The several hundred thousand overwhelmingly republican people serving in the US military who are not at their home addresses right now to pick up their 'phones. Not just the ones in Iraq and Afghanistan, but those in Korea, Germany, Italy, on sea duty on ships...

It works both ways.

NTM

Mr. Incognito

That will not happen, may blacks will not vote for him because they may see him as elitist, which he is, he has little in common with your average black man in a poor inner city neighborhood

Also mentioned are the young people, these are probably the same young people that were supposed to deliver John Kerry, well do you know what happened to them, they didn’t even bother voting.

One important difference, people weren't voting for Kerry as opposed to voting AGAINST Bush. Kerry didn't even oppose the war ffs. If he had the balls to demand pulling the plugs on the troops in 2004 I reckon the young vote would have been mobilised.

Now, as well as voting against Bush young people, aye and Black people have an idol bordering on hero worship for a lot of these supporters. Having a choice between two guys with very similiar polocies in 2004 is poles apart from the idealogogies that McCainan and Obama are touting now.

Although this is true, there is a factor which cancels out the amount of cell-phone-only people who may vote Obama: The several hundred thousand overwhelmingly republican people serving in the US military who are not at their home addresses right now to pick up their 'phones. Not just the ones in Iraq and Afghanistan, but those in Korea, Germany, Italy, on sea duty on ships...

Aye, but it's thousands verses MILLIONS. I think the marine corp only has like 50,000 troops or that, maybe less.

At the same time people in the south and the white working class will be out voting for McCain because they see him as a safe pair of hands and an experienced politician

There's a lot of Black votes in the south you know.

I see Obama conceding this one earlier on the morning after the vote rather than later, that is how much McCain will win by.

I don't think so. Paddy Power agrees with me.

amacachi

SetantaL wrote: »

That is quite possibly the silliest thing I have ever read on boards and in the politics forum too.

As far as I know the only person who ever got sick of being black and was able to do something about it was Michael Jackson.

The blacks I'm referring to are the huge numbers who never bother voting. This year there's a "black" candidate who they were all very excited about. By now they, along with many white people must be sick of hearing about "change". I'm certainly not saying that they'll go vote for McCain, but that they may not vote at all. Obama may have peaked too early is what I'm trying to get at.

Ludo

amacachi wrote: »

The blacks I'm referring to are the huge numbers who never bother voting. This year there's a "black" candidate who they were all very excited about. By now they, along with many white people must be sick of hearing about "change". I'm certainly not saying that they'll go vote for McCain, but that they may not vote at all. Obama may have peaked too early is what I'm trying to get at.

That is a very valid point about Obama peaking too early. Not really his fault though. McCain wrapping up the republican nomination so early meant he was out of the public eye for a long time while Obama had to keep in it to see off Clinton. It reached saturation point and then McCain came back onto the scene like a new candidate. I believe that has helped him immensely.

amacachi

Yeah, I thought it was a valid point but apparently it came across that I was suggesting that black people are sick of being black.

I can actually see McCain swing one of the Northeastern states this year, not saying he will, just that I wouldn't be surprised.

Ludo

amacachi wrote: »

Yeah, I thought it was a valid point but apparently it came across that I was suggesting that black people are sick of being black.

Nah...the person who replied read something into your post that simply wasn't there. Misunderstanding I imagine as you made perfect sense to me.

amacachi wrote: »

I can actually see McCain swing one of the Northeastern states this year, not saying he will, just that I wouldn't be surprised.

New Hampshire being the obvious one!

amacachi

SetantaL wrote: »

...but now that it's a black man for president, you can quadruple the black turnout.

So there's gonna be a turnout of over 200% of black voters apparently, scary stuff. http://www.census.gov/prod/2006pubs/p20-556.pdf

New Hampshire being the obvious one!

Aye, keep an eye on New Jersey as well.

Bubs101

SetantaL wrote: »

I find this hard to believe. I'm sorry but I don't think ringing people up and asking them how they are going to vote is an accurate indication of the actual vote. A gauge with peramaters sure, but saying that's it's an accurate science is just too much for me.

Also, this doesn't address the whole point of the thread. The peole I think are going to turn up and vote for Obama do not have landlines so your poll data is worthless and they are slipping outside the net. Great if they don't turn out to materially affect the difference between polling middle-class white americans polled and that demographic subsequently voting along those lines, but when a block turns up, en masse, that fell outside the polling paramaters in the first place the polls are worthless.

The question is- how worthless. Blacks are 12% of the population, the young vote conservatively 25% that's 40% of the country right there. Although that don't mean ****e if they don't actually vote.

.

In fairness Setanta, I studied this, did an exam on it and read three extensive essays on different types of surveying from the history to how it's carried out for my essay. Riveting stuff. And it was part of my political science degree so it is an actual semi-science believe it or not. Keep in mind that not all surveys are done over the phone, just the cheep ones and I can guarantee you that Obama and Co. won't be skimping.

It's also a little niave to assume that every white block gets polled while every black block get's ignored. They do the surveying based on the demographics of the area so say for example they were doing Iowa, they wouldn't need to survey virtually any black people whereas if they did New York they would strain themselves to survey the correct proportion of blacks, whites and hispanics while factoring in how much each demographic votes. A.K.A., less time wasted on hispanics because they tend to be lazy come voting day

Black Swan

Students tend to speak louder and vote less in the USA. The only time in recent US history that the student vote really made a difference was during the Viet Nam War, which had a military draft, and sufficient numbers of them were being drafted and killed or injured.

The States has a serious problem with racism, and African Americans seem to get the worst treatment among all the different races, ethnic differences, and those of different national origin. For example, Los Angeles is a checkerboard of racially segregated communities. Oh, it's PC not to show this, but how many will act differently when they have the anonymity of the polling booth?

Your criticism about the methods of polling may have some value, if in fact most polling is done by telephone in the States (I'm uncertain of their non-telephony validation methods). A very large percentage of people now use mobile phones, where returning a call on a poll costs them money, a disincentive to reply. Further, mobile phones and land lines have voice mail or answer machines to screen calls. I just asked my boatmate if she would return a polling call, and she said why bother? I said the same. How this may bias the polling methodology is uncertain. I would hope that they have methods to measure and account for this potential source of error.

In recent times there has been evidence of polling saturation in the States. People get tired answering polls and refrain from doing so. To what extent would a poll be biased by those still enthusiastic to answer a political poll (McCain-Palin or Obama-Biden advocates), verses those in the general population that may be saturated?

I am very suspect of online polls conducted of people that are active on the web. They represent a biased segment of the population, typically more educated than the general population, and unique in many other ways that may differentiate them from the general population.

Ludo

Blue_Lagoon wrote: »

I am very suspect of online polls conducted of people that are active on the web. They represent a biased segment of the population, typically more educated than the general population, and unique in many other ways that may differentiate them from the general population.

Fully agree with this as no doubt most people do. Online polls are completely meaning less. Homer Simpson man of the century? a Nation Once Again the best song ever?

Although I don't know about the more educated than the general population...reading some stuff on boards.ie shoots down that argument

Bubs101

Blue_Lagoon wrote: »

I am very suspect of online polls conducted of people that are active on the web. They represent a biased segment of the population, typically more educated than the general population, and unique in many other ways that may differentiate them from the general population.

Online surveys and polls have absolutely no respect unless you are going for a certain demographic. For one, they tend to completely miss out on O.A.Ps who are the biggest voters in percentage terms. Apparently they don't trust the ineternet and are afraid to give out details. Most internet surveys are normally specifically designed to target males from 16-30.

As for the market being oversaturated and Americans getting tired of answering polls, this may be true normally but I'd assume in the run up to the election participation goes through the roof, especially one which has caught the nation's attention like no other

brianthebard

Latest polls have Obama ahead by 2 points, which is as meaningful as the poll last month with McCain ahead by 5-not much.

Ludo

National polls are pretty meaningless anyway with the way the system works. State polls are much more important/imformative.

Black Swan

Ludo wrote: »

National polls are pretty meaningless anyway with the way the system works. State polls are much more important/imformative.

You might want to rethink this? Federal law in the US requires each citizen (or head of household) to participate in the US Census that occurs once every 10 years. The US census is not intended to be a sample, but rather a total enumeration of the US population. Although participation is required by law, there is a wide variation in response rates by the 50 states in the US, ranging from a low of 56% (Alaska) to 76% (Iowa), with an overall average response rate of 67% reported by the US Census Bureau.

A census is a form of polling (or surveying) the US population. When a large segment of a state's population refuses to participate, even when it is required by law, to what extent would these same people refuse to participate in a voluntary poll or survey? What is different about these people that fail to participate from those that do? Do they think or believe differently? At the state level, 44% of the Alaska population failed to participate for some reason. The essence of the problem is that we don't know much about the non-responders? How does this impact on the validity and reliability of state polls (or surveys), as well as the conclusions suggested by them?

Ludo

Nah...we are talking about a presidential election poll. As each state is winner takes all, an overall national poll, while giving an estimate of feelings, does not necessarily reflect how the election will turn out.

I find this to be the best site for polling information. Updated daily.

http://electoral-vote.com/

Manic Moran

Resurrecting this one because I saw an interesting little factoid in the fine print of the latest NBC/WSJ poll.
http://msnbcmedia.msn.com/i/msnbc/sections/news/080924_NBC-WSJ_Released.pdf

Apparently, a number of the polled persons were called on their mobiles. If they were called on their mobile, but also had a landline, they were discarded from the pool. This would indicate that the pollsters are specifically seeking out persons who only use mobiles, thus reaching some of the Obama base theorised as missed out because they don't own landlines.

NTM