Why do polls change almost daily?????

[MeachTheMonster]

?

The candidates, issues, and electorate are the same, and most people identify as democrat/republican and decided.

Are independents and undecideds causing this much variance in the numbers, or am I missing something?

 

[Brown_Pride]

because the sample sizes are so small. They are quasi-educated guesses at best...at absolute best.

How are you going to ask 1000 people a question and then report the response of less than 1% of the population as though it held some legitimate weight?

 

[It Takes A Nation...]

@Brown Pride - very true what you said.

Also there are a lot of flaky, fickle people out there. Some will go with the flow and follow whatever the trend at the moment appears to be. Also there is a distinction between registered voters and likely voters. Plus both campaigns are eager to spin the numbers to their own advantage as both sides believe what will decide the election is which side can get their supporters to show up at the polls.

 

[MeachTheMonster]

So no one has a real answer

 

[Ghost Utmost]

The real answer is because they are bullshyt. Dude was right: they are assuming that since 470 people of 1000 said they like Obama that that's going to equal the percentage out of hundreds of millions. Nice idea, but they are asking a different 1000 people each time.

 

[TRUEST]

polling is pure mathematics.

if u do 100 polls and each time you poll you're polling different people, and the results of those polls always favor one particular candidate, its safe to say that candidate WILL WIN.

i wouldn't say polls are bullsh1t at all.

 

[daze23]

So no one has a real answer

they're constantly polling different individuals. I wouldn't expect them to stay the same

most of them are highly unscientific, and really just give the media something to babble about. I'm not saying they're completely worthless, but I don't think they're accurate enough to act like a change of a few percent here or there really means much

 

[alybaba]

Stastical noise - that's why when it comes to sampling, there are such things such confidence levels and margin of error. Polls also trend in a particular direction due to undecideds going over to one side, voters moving to the registered voter/likely voter column and people switching the candidates who they will vote for (which I assume is rare).

 

[Huellz Santana]

i nor anybody i've ever known has ever been polled

i dont even have my home phone connected
if a nikka with a clipboard asks me a question im ignoring it.

so the question i wonder is... Who is really being polled here?​

 

[AnonymityX1000]

i nor anybody i've ever known has ever been polled

i dont even have my home phone connected
if a nikka with a clipboard asks me a question im ignoring it.

so the question i wonder is... Who is really being polled here?​

i'VE BEEN POLLED SO THERE GOES THAT.

 

[Brown_Pride]

i nor anybody i've ever known has ever been polled

i dont even have my home phone connected
if a nikka with a clipboard asks me a question im ignoring it.

so the question i wonder is... Who is really being polled here?​

i'VE BEEN POLLED SO THERE GOES THAT.

to answer your question harry...morons that's who

i think people call me to poll, but since they call around dinner time when I pick up the phone and they start talking and it's NOT someone i know I hang up or will sometimes just outright ignore the call.

That leaves us with the knowledge that the most likely responders are old people waiting by the phone for human contact.

 

[AnonymityX1000]

I think they call, and my estimation of when these maybe or maybe not calls happen is definitely around dinner which I have at the same time every single night. Now when I pick up the phone for calls that I might or might not get (again not sure), I definitely hang up or not answer at all.
That leaves us with the spurious correlation that responders to poll calls are old people waiting by the phone for human contact.

 

[invincible1914]

Just got polled today +1 for Obama in VA

 

[tru_m.a.c]

Stastical noise - that's why when it comes to sampling, there are such things such confidence levels and margin of error. Polls also trend in a particular direction due to undecideds going over to one side, voters moving to the registered voter/likely voter column and people switching the candidates who they will vote for (which I assume is rare).

yeah you learn this in basic stat

Science Behind The News: Opinion Polls And Random Sampling - Science360 - The Knowledge Network

 

[Cory MBA]

Since the electoral vote is paramount and the popular vote is meaningless, is it possible that these polls are misleading unless they are polling each of the states separately and then weighting by the number electoral votes?

Something I have been thinking about recently...

If most of the people called for the poll reside in NY or California, even if the respondents are 50/50, both states both will be 100% for the democratic candidate.

The stats can be misleading if you do not know the state of residence.

Am I wrong?