HomeLatest ThreadsGreatest ThreadsForums & GroupsMy SubscriptionsMy Posts
DU Home » Latest Threads » Forums & Groups » Main » General Discussion (Forum) » Are POLLS reliable these ...

Sun Nov 9, 2014, 03:36 PM

Are POLLS reliable these days?

Could I get some thoughts on the validity of POLLS?

I keep hearing folks say so-n-so is leading in the polls.
Or, polls show America leans one way or another.

I've has a statistics course or two so I understand the results...
But the fact that most polls are land-line phones, and rarely
include geographical data, I wonder if a poll of 500 or 1000
people has any bearing or what the MAJORITY of people think?
And lastly, most people I know hang up on cold callers.
So does that mean the people who DO complete poll questions
are not representative of the majority?

56 replies, 3022 views

Reply to this thread

Back to top Alert abuse

Always highlight: 10 newest replies | Replies posted after I mark a forum
Replies to this discussion thread
Arrow 56 replies Author Time Post
Reply Are POLLS reliable these days? (Original post)
Cosmic Kitten Nov 2014 OP
peacebird Nov 2014 #1
Cosmic Kitten Nov 2014 #2
peacebird Nov 2014 #15
yeoman6987 Nov 2014 #17
Cosmic Kitten Nov 2014 #19
yeoman6987 Nov 2014 #21
Cosmic Kitten Nov 2014 #24
yeoman6987 Nov 2014 #25
Cosmic Kitten Nov 2014 #18
spanone Nov 2014 #3
Cosmic Kitten Nov 2014 #5
hrmjustin Nov 2014 #4
Cosmic Kitten Nov 2014 #8
hrmjustin Nov 2014 #13
Cosmic Kitten Nov 2014 #16
hrmjustin Nov 2014 #23
Cosmic Kitten Nov 2014 #26
hrmjustin Nov 2014 #27
Cosmic Kitten Nov 2014 #30
hrmjustin Nov 2014 #31
Cosmic Kitten Nov 2014 #35
hrmjustin Nov 2014 #39
Cosmic Kitten Nov 2014 #50
DemocratSinceBirth Nov 2014 #6
Cosmic Kitten Nov 2014 #9
DemocratSinceBirth Nov 2014 #12
Cosmic Kitten Nov 2014 #20
merrily Nov 2014 #7
Cosmic Kitten Nov 2014 #10
Tierra_y_Libertad Nov 2014 #11
hrmjustin Nov 2014 #14
Cosmic Kitten Nov 2014 #22
Savannahmann Nov 2014 #28
Cosmic Kitten Nov 2014 #32
CK_John Nov 2014 #29
Cosmic Kitten Nov 2014 #33
CK_John Nov 2014 #44
Cosmic Kitten Nov 2014 #52
Spazito Nov 2014 #34
Cosmic Kitten Nov 2014 #37
Rex Nov 2014 #36
Cosmic Kitten Nov 2014 #40
Rex Nov 2014 #42
Cosmic Kitten Nov 2014 #47
Lady Freedom Returns Nov 2014 #38
Cosmic Kitten Nov 2014 #43
Lady Freedom Returns Nov 2014 #49
Cosmic Kitten Nov 2014 #53
Yo_Mama Nov 2014 #41
Cosmic Kitten Nov 2014 #45
Yo_Mama Nov 2014 #51
Cosmic Kitten Nov 2014 #54
Yo_Mama Nov 2014 #55
roamer65 Nov 2014 #46
Cosmic Kitten Nov 2014 #48
valerief Nov 2014 #56

Response to Cosmic Kitten (Original post)

Sun Nov 9, 2014, 03:43 PM

1. Polls can be reliable but not polls about an election 2yrs in the future

Reply to this post

Back to top Alert abuse Link here Permalink


Response to peacebird (Reply #1)

Sun Nov 9, 2014, 03:47 PM

2. No doubt about that.

Two years out only shows who has NAME RECOGNITION.

I guess I just imagine polls only indicate people
who still have land-lines, or are bored enough
to talk with a pollster?

I was only ever called ONCE for a political poll...
in my whole life.

Reply to this post

Back to top Alert abuse Link here Permalink


Response to Cosmic Kitten (Reply #2)

Sun Nov 9, 2014, 04:05 PM

15. This is another reason polls are less reliable, many folks only have cell phones not land lines

So a lot of polls are under counting millenials

Reply to this post

Back to top Alert abuse Link here Permalink


Response to peacebird (Reply #15)

Sun Nov 9, 2014, 04:15 PM

17. I don't think they undercounted millenials this election

 

Over counted maybe.

Reply to this post

Back to top Alert abuse Link here Permalink


Response to yeoman6987 (Reply #17)

Sun Nov 9, 2014, 04:21 PM

19. Yeoman, why do you think that?

Reply to this post

Back to top Alert abuse Link here Permalink


Response to Cosmic Kitten (Reply #19)

Sun Nov 9, 2014, 04:27 PM

21. I was being snarky.

 

Failed again.....

Reply to this post

Back to top Alert abuse Link here Permalink


Response to yeoman6987 (Reply #21)

Sun Nov 9, 2014, 04:36 PM

24. Don't be so hard on yourself :)

I'm a noobie, so it may take a while
till I get to know folks and get the hang of this place.

Reply to this post

Back to top Alert abuse Link here Permalink


Response to Cosmic Kitten (Reply #24)

Sun Nov 9, 2014, 04:41 PM

25. Cool and thank you!

 

Welcome by the way. You will love it here!!!!!!

Reply to this post

Back to top Alert abuse Link here Permalink


Response to peacebird (Reply #15)

Sun Nov 9, 2014, 04:16 PM

18. Thats definitely an issue with sampling.

Plus, millennials may or may not be as willing
to complete a poll versus older generations?

That would invalidate a poll, yes?

Reply to this post

Back to top Alert abuse Link here Permalink


Response to Cosmic Kitten (Original post)

Sun Nov 9, 2014, 03:49 PM

3. pollsters are hired by people who want to poll high in the polls. do not let clients down.

1st rule of business....

Reply to this post

Back to top Alert abuse Link here Permalink


Response to spanone (Reply #3)

Sun Nov 9, 2014, 03:52 PM

5. That's kinda why I'm curious.

So when the MSM does a poll, who's the client?

Reply to this post

Back to top Alert abuse Link here Permalink


Response to Cosmic Kitten (Original post)

Sun Nov 9, 2014, 03:51 PM

4. Polls are a snap shot of right now.

 

When I say that Hillary is supported by a majority of Democrats, that is at the moment.

The primary fight will move the numbers and we shall see in the end where things end.

Reply to this post

Back to top Alert abuse Link here Permalink


Response to hrmjustin (Reply #4)

Sun Nov 9, 2014, 03:56 PM

8. How do you define MAJORITY of DEMOCRATS?

That's kinda the essence of my question.
If a poll shows Hillary as being "supported
by a majority of Democrats" which Democrats, where?

To say it's the "majority" seems pretty broad brushed?
Was it 1000 white, senior citizens, in Florida?

Reply to this post

Back to top Alert abuse Link here Permalink


Response to Cosmic Kitten (Reply #8)

Sun Nov 9, 2014, 04:03 PM

13. Scientific polling from different polling firms showing registered democrats support her for the

 

nomination at this point. Doesn't mean that she will win but that right now she is favored.

I think some of those polls include democratic leaning independent voters in xome states that can vote in the primaries.

Reply to this post

Back to top Alert abuse Link here Permalink


Response to hrmjustin (Reply #13)

Sun Nov 9, 2014, 04:14 PM

16. YES, I get that she is leading for now...

But are those polls reliable?

Can we say that a snapshot or unknown "registered democrats"
has any bearing on what the MAJORITY thinks?
How many registered democrats are needed to create
a reliable sample that extends to the majority?

And does geography have any influence?
Democrats in the south are very different from the northeast or west coast.
Thats's why I'm curious about the validity of POLLS generally?
Can we reliably say that 1000 retirees in Florida represent ALL Democrats?

Reply to this post

Back to top Alert abuse Link here Permalink


Response to Cosmic Kitten (Reply #16)

Sun Nov 9, 2014, 04:30 PM

23. We should take all polls for what they are worth.

 

Remember that no matter what they say now a primary will move the numbers.


Polls are done nationwide and statewide.

As for the validity every polling firm I would say every firm that has publicly posted their results have Hillary ahead big.

But this is mainly name recognition and this will change as the primary goes forth.

Reply to this post

Back to top Alert abuse Link here Permalink


Response to hrmjustin (Reply #23)

Sun Nov 9, 2014, 04:45 PM

26. Yes, take all polls for what they are worth...

That's what I'm wondering about???

It just seems right now that NO POLLS are worth much.
People are giving lots or reasons why they are biased

Reply to this post

Back to top Alert abuse Link here Permalink


Response to Cosmic Kitten (Reply #26)

Sun Nov 9, 2014, 04:48 PM

27. Well some are better than others.

 

But they all are consistent on Hillary being ahead in the nomination fight.

Reply to this post

Back to top Alert abuse Link here Permalink


Response to hrmjustin (Reply #27)

Sun Nov 9, 2014, 05:31 PM

30. That's why I'm confused!?.

What does that mean?
What am I supposed to think?

Are pollsters trying to please Hillary?
Did Hillary pay for these polls?
Are the polling groups favoring Hillary?
Would Republicans benefit from Hillary leading the nomination?

HALP!

Reply to this post

Back to top Alert abuse Link here Permalink


Response to Cosmic Kitten (Reply #30)

Sun Nov 9, 2014, 05:39 PM

31. No these pollsters are not trying to please people.

 

The fact is someone or a group of candidates other than Hillary will emerge and change the numbers


The numbers are big for Hillary now because of name recognition, likability, and the fact they think she is most electable.

The numbers are where you would expect them to be.

Reply to this post

Back to top Alert abuse Link here Permalink


Response to hrmjustin (Reply #31)

Sun Nov 9, 2014, 06:03 PM

35. Hmm, seems reasonable.

Indeed, other candidates will change the numbers.
But won't polling companies pushing Hillary as the leader
influence how people see her?
Just thinking that less informed voters may begin to favor
Hillary simply because "scientific pollsters" say she is most electable?

In that example polls aren't taking a "snapshot"...
they are setting the scene, yes?

I think you are totally right about her name recognition.
EVERYONE in the USA knows of Hillary Rodham Clinton.

But how do you know she is likable or electable from these polls?
If the polls are mostly of Democrats and Democrats likely to vote
where does likability and electability figure in?
Wouldn't we need to know what "others" think of her?

Do the polls account for the Clinton "baggage"?
I KNOW many women who do NOT like Hillary based on how
she "stood by her man" when we learned Bill Clinton is a sexual predator!
This is scuttlebutt and not relevant to the OP, sorry.

Reply to this post

Back to top Alert abuse Link here Permalink


Response to Cosmic Kitten (Reply #35)

Sun Nov 9, 2014, 06:11 PM

39. Remember we are atalking about multiple polling firms so I doubt it is to curry favor.

 

And remember that she was ahead early in the 2008 fight and that changed as the primary went on.

My feeling is that her lead is more solid this time.

Reply to this post

Back to top Alert abuse Link here Permalink


Response to hrmjustin (Reply #39)

Sun Nov 9, 2014, 06:39 PM

50. She is undoubtably in "the lead"

But how can we be sure polling firms
are NOT trying to curry favor?

Lets put on our "business hats".
As a polling firm we want business.
We want the biggest business with the deepest pockets, right?

So, KNOWING the upcoming Presidential election
will be the most expensive ever, in part due to "Citizens United",
whose money should we seek out?
The biggest players, right?
There is NO ONE bigger than Hillary.

She has Wall st, Defense contractors, Agriculture
and more ready and willing to support her campaign.
So I am skeptical that any "business" would not compete
to earn her business. In marketing, "free samples"
go a long long way to getting the "consumer" to buy.

In other words, it seems like a missed opportunity
if they didn't each try to curry favor. Our shareholders
would be aghast if we didn't get those millions in profits.

Reply to this post

Back to top Alert abuse Link here Permalink


Response to Cosmic Kitten (Original post)

Sun Nov 9, 2014, 03:53 PM

6. Most reputable pollster include cell phone users and make sure their samples measure ...

Most reputable pollsters include cell phone users and make sure their samples measures what they are intended to measure. That being said they are cameras and capture a moment and that's all they capture.

Reply to this post

Back to top Alert abuse Link here Permalink


Response to DemocratSinceBirth (Reply #6)

Sun Nov 9, 2014, 03:58 PM

9. DO reputable pollsters explain sampling techniques?

Reply to this post

Back to top Alert abuse Link here Permalink


Response to Cosmic Kitten (Reply #9)

Sun Nov 9, 2014, 04:02 PM

12. Some pollsters are more transparent than others.

Nate Silver's site is a good place if you're interested in such...He writes about which pollsters are the most accurate or reliable. He also goes into some detail about their methodology...

I'm not a polling guru but I know how hard it is to get a representative sample.

Reply to this post

Back to top Alert abuse Link here Permalink


Response to DemocratSinceBirth (Reply #12)

Sun Nov 9, 2014, 04:24 PM

20. A representative sample!

That's it!
I never hear anyone who uses polling data
ever explain WHY or HOW their poll is representative.
It always seems fiat that a poll is valid just because, numbers.

Reply to this post

Back to top Alert abuse Link here Permalink


Response to Cosmic Kitten (Original post)

Sun Nov 9, 2014, 03:53 PM

7. Depends. When a politician wants to know what's what, he or

she does an internal poll. Then general public does not hear about it unless the politician's camp decides to leak the results.

I think most pollsters are hired with an agenda and, like anyone else in business, they try to please the client.

Phone polls have also suffered, as you noted, from the land line vs. cell phone divide.

Reply to this post

Back to top Alert abuse Link here Permalink


Response to merrily (Reply #7)

Sun Nov 9, 2014, 04:01 PM

10. So, if pollsters are hired with an agenda...

that would seem to invalidate the poll?

Actually, it seems more like a propaganda technique
from what you are suggesting! Yikes!!!

Reply to this post

Back to top Alert abuse Link here Permalink


Response to Cosmic Kitten (Original post)

Sun Nov 9, 2014, 04:02 PM

11. There are lies, damned lies, and statistics. Mark Twain

 

Reply to this post

Back to top Alert abuse Link here Permalink


Response to Tierra_y_Libertad (Reply #11)

Sun Nov 9, 2014, 04:03 PM

14. All polling firms?

 

Reply to this post

Back to top Alert abuse Link here Permalink


Response to Tierra_y_Libertad (Reply #11)

Sun Nov 9, 2014, 04:28 PM

22. Nice ;~/

Facts just twist the truth around
Facts are living turned inside out

Eno/Byrne

Reply to this post

Back to top Alert abuse Link here Permalink


Response to Cosmic Kitten (Original post)

Sun Nov 9, 2014, 04:48 PM

28. Polls now include cell phones

 

Polling showed the Republicans winning 52 Senate seats, they're probably going to have 54. Polling gives you a glance at things, and the one thing to watch is historical trends.

Polling showed the Democrats at a huge disadvantage, a truth that most people ignored until the week or at most two weeks before the election. The results were pretty close generally speaking.

I have no idea how the Democrats got the internal poll numbers they did. As near as I can tell they were questioning people working in the campaign headquarters. "Good news, the latest poll is in and you're the favorite by twenty points! We asked everyone in the building and they all said they're voting for you."

Polling is especially useful when you start looking at issues. The big issue ignored by Democrats was the economy. Some here will tell you that the people should have been told of the great accomplishments in the economy. The problem is you don't have to tell people the economy is going well, they see that on their own. The thing you have to tell them is your plans to improve the economy, when a majority think it isn't going well.

Yes, Polling is reasonably accurate, and yes, it is useful. If nothing else it lets us know what the people are thinking, and what they think is important. The ultimate poll is the election. The fact that we lost there, as well as in the polls shows how much we should be keeping an eye on them.

Reply to this post

Back to top Alert abuse Link here Permalink


Response to Savannahmann (Reply #28)

Sun Nov 9, 2014, 05:40 PM

32. Oooooh, Savannahmann. Good info, I think?

So this 52-54 thing...
Do you think that a poll showing "my" candidates
losing would depress turnout?
I vote no matter what, but some might not be so committed.
I do know that when people feel helpless they will give up.
Do you think polls could diminish turnout?

Totally agree with you on the "big issue" thing.
Show me, don't tell me.

BUT...
How is polling "accurate"?
Exit polls don't seem to be accurate?

And many polls only tell us the result of a poll,
not always the specific questions asked.
If I don't know how a poll question is framed
how can I know if my interpretation of the results
reflects the actual questions asked?
That might impact the accuracy/interpretation?

Reply to this post

Back to top Alert abuse Link here Permalink


Response to Cosmic Kitten (Original post)

Sun Nov 9, 2014, 05:12 PM

29. I've posted several times about the "fear of change" factor. IMO, Polls are fairly accurate when

they are taken, since they now include cell calls but ....between 5-7% will change their vote right in the voting booth for the incumbent. They were not lying to the pollsters but human nature has a great fear of change.

Also if there is no incumbent those 5-7% people will not vote for anyone in that slot.

Reply to this post

Back to top Alert abuse Link here Permalink


Response to CK_John (Reply #29)

Sun Nov 9, 2014, 05:46 PM

33. I'm still not clear about HOW polls are accurate?

I'm sure the actual statistics part is "accurate"
The standard deviations, margin or error, etc.

Wait... I kinda remember from statistics that
the "margin of error" can INVALIDATE the results???
So a poll with a margin of error above a specific amount
say, 4+/- would mean that the same poll done again
could lead to a different result?
IS that right?

Does margin of error play any part in the statistical significance?

Reply to this post

Back to top Alert abuse Link here Permalink


Response to Cosmic Kitten (Reply #33)

Sun Nov 9, 2014, 06:23 PM

44. The most important thing about a poll is who paid for it. Society only needs polls because

9 out of 10 people are afraid to make a decision before they vote, they want to know who their peers are going to vote for.

IMO, margin of error is just to keep pollsters from being sued and having to give back the buyers money.

Reply to this post

Back to top Alert abuse Link here Permalink


Response to CK_John (Reply #44)

Sun Nov 9, 2014, 06:50 PM

52. I think I understand you?

People will vote according to the herd?

Because life is complicated, not everyone has
the luxury to research the candidates.

When faced with low information, voters rely on
the MSM or polls so they know which way
the wind is blowing?

Then when the herd is at the water cooler
they can have an opinion, and know who to vote for,
even if it is based simply on polls?

Fear is a great motivator!

Reply to this post

Back to top Alert abuse Link here Permalink


Response to Cosmic Kitten (Original post)

Sun Nov 9, 2014, 05:54 PM

34. Less and less reliable as technology advances while polling methodology continues...

to rely predominantly on land-line calls given more and more of the public are using their cell phones and have given up their land-lines. The pollsters, themselves, have admitted the difficulty and increasing unpredictability of polling results.

Reply to this post

Back to top Alert abuse Link here Permalink


Response to Spazito (Reply #34)

Sun Nov 9, 2014, 06:07 PM

37. Is that so?!

Huh, I did not know that.

It's refreshing when the "experts" admit that
what they do is not so absolute.

The reliance on technocracy often clashes with reality.
Thanks for the thoughts

Reply to this post

Back to top Alert abuse Link here Permalink


Response to Cosmic Kitten (Original post)

Sun Nov 9, 2014, 06:04 PM

36. After this last election, sure doesn't seem like it.

 

What are the odds of 8 darkhorse wins for the SAME party?

Reply to this post

Back to top Alert abuse Link here Permalink


Response to Rex (Reply #36)

Sun Nov 9, 2014, 06:12 PM

40. Maybe someone's thumb is on the scale?

However, that is why I'm so curious about polling.

Did polls and pundits repeating that republicans
were ahead and going to win inevitably impact the elections?
Or maybe it was just voter purges, apathy, weak campaigns,
weak candidates, or black box voting

Reply to this post

Back to top Alert abuse Link here Permalink


Response to Cosmic Kitten (Reply #40)

Sun Nov 9, 2014, 06:19 PM

42. I think you have the answer right there - voter purges, apathy

 

but IMO the biggest contributor to how we vote is NOT the candidate or the campaign. It is the news station that tells people 'what is going on'...we have to just assume they are telling us the truth. We have to hope they are unbiased and won't try and make up voters minds for them.

Sadly that is all the M$M does now...that and make money off of drug commercials.

Reply to this post

Back to top Alert abuse Link here Permalink


Response to Rex (Reply #42)

Sun Nov 9, 2014, 06:26 PM

47. ...the news station that tells people 'what is going on'...

Sad, but probably true!
Makes me feel the way your avatar looks.

Reply to this post

Back to top Alert abuse Link here Permalink


Response to Cosmic Kitten (Original post)

Sun Nov 9, 2014, 06:10 PM

38. It all depends on the Questions.

Like with the Mass Governor race. Many a poll had Charlie Baker in a land slide. However MassINK Polling had it being very close.

The final result?
Charlie Baker48.46%
Republican1,041,640

Martha Coakley46.58%
Democrat1,001,279

Why? Because of asking the right questions.

Reply to this post

Back to top Alert abuse Link here Permalink


Response to Lady Freedom Returns (Reply #38)

Sun Nov 9, 2014, 06:21 PM

43. Certainly how and what questions are asked...

will impact the poll results.
And, what precincts are polled.

The gerrymandering probably adds another dimension as well.
Coakley looks to be out roughly 40,000 votes?
Not a whole lot out of 2 million?

Wouldn't be hard to swing that election with some effort.
I know I was given inaccurate info at the voting booth
about voter ID and what is accepted.

A "nice" lady told me if someone didn't have photo id
they could go home to find it and then come back.
Photo ID is NOT REQUIRED in my state!
The reality was that voter could simply show a utility bill,
or swear an affidavit and then vote.
I am confident some people were deprived of their votes
because of such misinformation.

Reply to this post

Back to top Alert abuse Link here Permalink


Response to Cosmic Kitten (Reply #43)

Sun Nov 9, 2014, 06:28 PM

49. It is like here in Arizona.

If you don't have I.D with you, you can still Vote. But your ballot will be put aside till you bring the I.D to the Recorderís Office, the deadlineto due this is 5:00 p.m. on the 5th business day after election day

Reply to this post

Back to top Alert abuse Link here Permalink


Response to Lady Freedom Returns (Reply #49)

Sun Nov 9, 2014, 07:07 PM

53. That's ok, but still onerous on voters

Sorry to hear your state makes voting so difficult.
It is saddening that people such as the elderly
would have such an unrealistic demand put on them.

My grandmother wouldn't even be allowed to vote.
She has no "official" photo ID!

This was my first vote in a new state so I did some QandA.
In my state, a sworn affidavit is all that's required, PERIOD.
I asked about 5 poll workers, and NONE of them told me that.
I asked if they had a copy of the voting statute that defined
the voting requirements.
They did not know or have the statute.

Although, there was a poster issued by the elections board
on the front window of our polling place that
indicated what was acceptable to verify voter identification.
None of the workers directed me to that poster/information.
Afterwards I looked up the law at the state.gov website.

Reply to this post

Back to top Alert abuse Link here Permalink


Response to Cosmic Kitten (Original post)

Sun Nov 9, 2014, 06:18 PM

41. Turnout can really change matters

"Likely voter" polls are the most reliable, but they rely on being able to assess whether someone is really going to vote. A lot of people say they will vote when they are speaking to someone due to social pressure, but don't actually vote.

Reply to this post

Back to top Alert abuse Link here Permalink


Response to Yo_Mama (Reply #41)

Sun Nov 9, 2014, 06:23 PM

45. Why is a "Likely voter" poll the most reliable?

Reply to this post

Back to top Alert abuse Link here Permalink


Response to Cosmic Kitten (Reply #45)

Sun Nov 9, 2014, 06:47 PM

51. Because, when you survey, you inevitably talk to a lot of registered people who don't vote

Voter turnout in 2012 was 57.5%. Sometimes it is around 60%, but that only occurs in presidential election years. So if you can't screen out the percentage of people who won't show up at the polls (or mail in the ballot), your poll is bound to be off.

But voter participation changes for reasons that are not always obvious to pollsters. Every polling firm has an independent algorithm, which the firm tries to rebalance each cycle based on what was off the previous cycle. So minor changes in voter sentiment can change the vote tallies quite a bit from polling. It is impossible to get likely voterhood exactly right, no matter how many polling questions you include.

Also, people who are embarrassed about not voting or not having voted might lie to the pollsters.

Also, likely voter assessments become much more accurate late in the election cycle.

Reply to this post

Back to top Alert abuse Link here Permalink


Response to Yo_Mama (Reply #51)

Sun Nov 9, 2014, 07:23 PM

54. So "likely" voters is better than "registered" voters...

because they have "committed" to vote,
versus people who could vote but might not?
It's about the voter making a commitment and then following thru?
That seems like a reasonable assumption.

MY curiosity though lies not in algorithms
or tweaking the machinery.

MY curiosity is in how a small sample,
say hundreds or a few thousand "likely voters"
from specific geographic areas can be extended to
make assumptions about MILLIONS of voters?

Why should we believe that say 1000 people in any area
would be representative of millions of likely voters?
Also how can we know that polling data is not impacting voter behavior?
Might not that idea explain why likely voter assessments
become much more accurate late in the election cycle?

What if polling data drives voter opinion significantly
and ultimately impacts election results?
IS polling valid or is it a campaign technique?

Reply to this post

Back to top Alert abuse Link here Permalink


Response to Cosmic Kitten (Reply #54)

Sun Nov 9, 2014, 07:43 PM

55. Ah, what they do is take the characteristics of the people they spoke to

And then rebalance the raw results by counting the results in proportion to the overall population, not just the people they reached. And polls now do sample cell phone only users.

So, they will try to pick a geographically and socially balanced sample, but then after they get the results, if they find the over-sampled oldsters in relation to the younger groups, they will up the weight of the younger groups they did sample. If they find they have too many men, they'll up the weight of the women they do have.

So the actual output of a poll nowadays is not the results of the poll, but the results of the poll reworked to become more representative.

But in reality, the margin of error is always at least 2 or 3 percent, so it can't predict the outcome of a close election. End results of 52-48 might seem like a nice solid win, but the best poll before hand might have had 50-49 or even 49-50. It's not an exact science and never will be, and even getting the exit polls right is difficult, because you can have sample error there just through adverse selection (early voter turnout different than late voter turnout; willingness to participate lower among pressured workers just zipping by, etc).

The other funny thing about polling is that the longer you make the poll (information and double-checking for enhanced accuracy), the more you skew the results because not everyone will take the time. It's kind of the Heisenberg uncertainty principle of polling. The exit polls are also less valid than they used to be because of early voting.

Reply to this post

Back to top Alert abuse Link here Permalink


Response to Cosmic Kitten (Original post)

Sun Nov 9, 2014, 06:24 PM

46. Elections aren't reliable these days. N/t

Reply to this post

Back to top Alert abuse Link here Permalink


Response to roamer65 (Reply #46)

Sun Nov 9, 2014, 06:27 PM

48. Geez, yeah, cut right to the chase.

You cynic

Reply to this post

Back to top Alert abuse Link here Permalink


Response to Cosmic Kitten (Original post)

Sun Nov 9, 2014, 08:55 PM

56. Probably as reliable as our votes, so no, no fuckin' way.

Reply to this post

Back to top Alert abuse Link here Permalink

Reply to this thread