2016 Postmortem
Related: About this forumNEW POLL: Reuters/Ipsos Clinton +6% nationally
Likely Voters:
Clinton 44%
Deplorable 38%
Other 7%
Wouldn't Vote 3%
Don't Know/Refused 7%
LAST Week this poll was 39% to 39%.
Sampled 9/22 through 9/26 (so mostly if not all before the debates)
http://www.realclearpolitics.com/docs/2016/2016_Reuters_Tracking_-_Core_Political_9.27_.16__.pdf
sharp_stick
(14,400 posts)it's been trending the wrong way lately and seems to be a pretty well regarded poll.
mcar
(42,334 posts)The_Casual_Observer
(27,742 posts)Even the stupid have got to be sensing that this clown is a fraud.
Joe941
(2,848 posts)Joe941
(2,848 posts)the sample seems to be pretty heavily weighted with democrats. Is that a representative sample these days?
factfinder_77
(841 posts)1705 in the sample,
752 dem
570 reps
216 indi
1411 reg voters
1042 likely voters
scheming daemons
(25,487 posts)And the pollster has no way of knowing what their real registration is... only what they say in response to the poll.
The thing is.... people who are planning to vote for Hillary will tell pollsters they are Democrats... and people that are planning to vote for Trump will mostly tell pollsters they are Republican. Regardless of what their registration is.
So... when Hillary goes up in the polls... so will the number of people who say they are Ds.
The pollster doesn't decide to "sample" more Democrats than Republicans. The pollsters randomly samples a number of Americans and then reports what they say.
Foggyhill
(1,060 posts)You look at demos or likely voters to see if they match the demos of the actual election.
Registered voters match with population, they don't necessarily note.
So, dems could have a huge advantage there and still lose the election if nobody goes to vote.
The number here tells you that a lot of the others and undecided are likely registered dem voters.
Trump has a very very hard ceiling.
If the dems GOTV, they got this locked up hard, simple as that.
alcibiades_mystery
(36,437 posts)...
scheming daemons
(25,487 posts)The poll asks the respondents what their political affiliation is... They don't "sample" Democrats and Republicans....
They take a random sample of Americans... and report what they say is their political affiliation. If more respondents say they are Democrats than say they are Republicans, that means that likely more people are Democrats than Republicans in the general public as well.
People usually say whatever party aligns with the candidate they're voting for...regardless of what their registration is.
In years when Democrats win the White House, more people self-identify as Democrats than Republicans.
There is no "skewing"... as RWers discovered in 2012. If a poll is D+6... it is probably because the general public is close to D+6.
Joe941
(2,848 posts)geek tragedy
(68,868 posts)Democat
(11,617 posts)A lot of sad Trump supporters making accounts on DU since the debate.
Response to titaniumsalute (Original post)
Post removed
triron
(22,006 posts)was +7 a few days ago.
Foggyhill
(1,060 posts)titaniumsalute
(4,742 posts)Foggyhill
(1,060 posts)Very hard to get a representative sample of the underlying pop with a small poll, which compounds the margin of error
possibly making it completely useless.
A good sample is representative of the population of likely voters (demos match) and if it is large enough, it also is a pretty good gauge
of the actual voter's intent.
titaniumsalute
(4,742 posts)I get the sample and how it effects the MOE.
But the last poll from Reuters/Ipsos was 1,559 respondents. Only about 100 and change fewer respondents. The MOE difference would be infinitesimal.
Again, what is your point?
Foggyhill
(1,060 posts)Got, 30 years of engineering with a crapload of stats/maths,
a MBA and even a comm degree (includes polling).
Do we have to trade business cards now?
My point has nothing to do with the SIZE of the sample or MOE.
It has to do with how representative the poll is in the first place of the targeted population : likely voters.
How much it actually matches the demos of the population of likely voters.
That's the crux of polling, the hard part, which most so called polling operation fail at.
If you poll a million white men and zero of everything else, you're not going to have a margin of error of 0.0001%.
Most poll's margin of error is meaningless no matter how many people they poll because they're sample is not representative
because of methodological problems. For time, money or whatever reasons, they didn't get a good sample; they're mostly polling the most convenient to access rather than the right people.
University/scientific studies often run into the same difficulty in studying students and other people they have ready access too.
Those small sample shoddy method polls are there to get people talking and have no factual value.
The Clinton campaign poller said as much when he compared these polls to their own internal polls.
In small samples, those methodological errors are often made even worse.
Also, I was comparing this poll to a lot of other polls which have samples of 500 - 600 (tons of them in the last few weeks)
I was clear in my first post about that.
So, that is what my initial point was.
Why are most polls bad? Because there is no incentive to produce good ones, seemingly they all get reported in the same way: good or bad.
titaniumsalute
(4,742 posts)But I am saying you don't know what you are talking about when it comes to survey research. Yeah, it is numbers. But as you know the science behind survey research is very unique. And it works.
Of course if you poll a "million white men" it wouldn't be a representative sample. That's why they don't poll a million white men.
It has been proven mathematically that you don't need a huge sample to represent a general opinion. Yes an N=1000 IS good enough, with demographic weighting, to be a representative sample base. Take a look at this from 2012. It is the last 10 or so polls for the 2012 general election. They were all, in aggregate, at about the +- 3%. (3.2% to be exact.) I would say that is pretty close in the grand scheme of things. You could poll 3,000 people but the MOE doesn't change much and the costs start soaring. http://www.realclearpolitics.com/epolls/2012/president/us/general_election_romney_vs_obama-1171.html
Some research companies do employ panel methods that represent the universe demographically up front to reduce weighting to almost zero. But they are typically very expensive as the panelists must be compensated for their continued time and effort.
The poll I mentioned about did take into consideration the demographics and a representative sample by weighting for age, gender, education, and employment. Our company would have also weighted to race and HH income.
So overall most polls are NOT bad. Some are more reliable than others. I do agree that all polls are reported for media's new headline which is bad when it is shoddy internet online unscientific polls.
scheming daemons
(25,487 posts)scheming daemons
(25,487 posts)titaniumsalute
(4,742 posts)It seems like we have some Trump trolls on here today so thanks for clarifying. I added that info to the OP.
Joe941
(2,848 posts)triron
(22,006 posts)at the reuters.com site I saw a "5 day rolling" poll ending Sept 26 showing her lead at 6.4% (2 way)
titaniumsalute
(4,742 posts)You keep saying these things but you make actual statement or point.
Jim Dandy
(358 posts)Cracklin Charlie
(12,904 posts)Terrific numbers!