2016 Postmortem
Related: About this forumThe Monmouth poll that has Clinton ahead in Iowa is horrendously flawed.
The methodology used in the newly released Monmouth University poll that shows Clinton ahead of Sanders in Iowa, is so flawed--the poll should be ignored.
Two recent polls had Clinton up by single digits (Quinnipiac Clinton by 9; CBS YouGov Clinton up by only 6). It would be easy to say that Monmouth is simply an outlier. It's not. It's junk and it should be deemed unreliable and unscientific.
I read through the poll methodology described on Monmouth's website. You can read more at the link at the bottom of this post. Their skewed methods reveal over-the-top sampling flaws--and every single flaw favors Hillary Clinton.
Here are the four primary failures of this poll:
1. AGE
A whopping 69 percent of the all respondents were older than 50; only 12 percent polled were between the ages of 18-34. How convenient to outpoll senior citizens and barely represent younger voters. A complete bias toward Clinton.
2. GENDER
56 percent of those polled were female; 44 percent male. Again, bias favors Clinton.
2.) LAND LINE/CELL PHONES
Of the 405 people polled, 72 percent were polled from a landline; 28 percent were reached on a cell phone.
Again, a bias toward Clinton with older and more conservative demographics having land lines and younger voters generally reached on cell phones.
3. VOTER BEHAVIOR
Those polled were Democrats who had voted in at least one of the last two Iowa state primaries (2008 and 2004); and in both the 2012 and 2014 General Elections.
*First, Iowa has an open primary. The night of the caucus, many Independents and some Republicans will cross over, register and caucus with the Democrats. It happens EVERY caucus. Many Bernie supporters will be Independents (and maybe some Republicans) who will not be registered Democrats until the night of the caucus. All of these people were left out of the polling sample.
*Secondly, those two Iowa state primaries (2004/2008) were 7 and 11 years ago. This would completely omit younger voters who were not old enough to vote in those primaries.
*Lastly, voting in both the 2012 and 2014 General Elections skews to older generations as well. Turnout during non-Presidential election years (2014 in this case) favors an older, more seasoned voter.
This confluence of ridiculous biases and multiple flaws should render this poll meaningless.
If you want to peruse the exact wording of the Monmouth methodology, here it is from their website:
"The Monmouth University Poll was sponsored and conducted by the Monmouth University Polling Institute from December 3 to 6, 2015 with a statewide random sample of 405 Iowa voters drawn from a list of registered voters who voted in at least one of the last two state primary elections, voted in both the 2012 and 2014 general elections, or have registered to vote in the past year.
Results in this report are based on voters who say they are likely to attend the Democratic presidential caucuses in February 2016 (including voters already registered as Democrats and voters who say they will declare themselves as Democrat for the caucuses). This includes 294 contacted by a live interviewer on a landline telephone and 111 contacted by a live interviewer on a cell phone, in English. Monmouth is responsible for all aspects of the survey design, data weighting and analysis. Final sample is weighted for age and gender based on state registration list information on the pool of voters who participate in primary elections. Data collection support provided by Braun Research (field) and Aristotle (voter list). For results based on the total sample, one can say with 95% confidence that the error attributable to sampling has a maximum margin of plus or minus 4.9 percentage points (unadjusted for sample design). Sampling error can be larger for sub-groups (see table below). In addition to sampling error, one should bear in mind that question wording and practical difficulties in conducting surveys can introduce error or bias into the findings of opinion polls.

(FULL LINK)
http://www.monmouth.edu/assets/0/32212254770/32212254991/32212254992/32212254994/32212254995/30064771087/9a435a5e-2c3d-4b5d-b0ea-a45c5192af0a.pdf
saturnsring
(1,832 posts)CoffeeCat
(24,411 posts)...to think this poll is valid, after reading about the methodology.
I nearly laughed reading through it.
reformist2
(9,841 posts)saturnsring
(1,832 posts)shenmue
(38,506 posts)Wow.
HerbChestnut
(3,649 posts)It's recognizing the fact that 33% of the people polled in this survey were 65+. Another 36% were between 50-64. That leaves the other 31% below age 50, of which 20% was between 35-49. This means that, once again, millenials are being under polled.
leftynyc
(26,060 posts)get off their asses and votes like the over 50+ crowd, THEN you'll have a valid point. Also, the gender voting in 2012 was 52% women, 47% men so there's a slight advantage to Hillary in that poll. They're obviously looking for LIKELY voters (measured by who actually votes and not those who say they will).
jkbRN
(850 posts)obviously not, because you used an ad hominem for your rebuttal.
The OP is right
If you wanna talk stats, I'm all ears, but attacking a group of people makes you look like you lack the knowledge to have a real discussion in regards to the content of the post.
saturnsring
(1,832 posts)yardwork
(61,722 posts)Would you care to discuss the supposed flaws in this poll's methodology? I'm not seeing much, certainly not enough to warrant the hysterical claim that the poll is hopelessly flawed.
jkbRN
(850 posts)Did you fail? They claims are pretty clearly stated by the OP. Also, the use of a 8-9% MOE on demographics (age) is shameful.
If not, please enlighten me..
yardwork
(61,722 posts)To describe this poll in these hysterical terms is silly.
forest444
(5,902 posts)Last edited Tue Dec 8, 2015, 09:02 PM - Edit history (1)
And the OP spared no detail in explaining why, point by point.
Proof once again that it's not who votes that counts, as Stalin might say, but who counts the votes.
TheOnlyOpinionYouNee
(1 post)There is a problem with the sampling frame, not the methodology.
I have no doubt the results are accurate given the sampling assumptions they made. Are those assumptions accurate? Personally I don't think so, but that's an empirical question come caucus day.
yardwork
(61,722 posts)The sampling is based on assumptions about who will vote in the caucuses. Is the model accurate? As you say, we will see when the votes are actually cast.
workinclasszero
(28,270 posts)the ones Bernies behind in.
Everybody clear on that?
SMH
Buzz Clik
(38,437 posts)http://elections.huffingtonpost.com/pollster/2016-iowa-presidential-democratic-primary
The average of at least five polls.
CoffeeCat
(24,411 posts)...and reliability and validity are foundational with polls.
This poll has neither.
Keep spewing your worthless, hollow talking points though...instead of adding something intelligent to the discussion.
Buzz Clik
(38,437 posts)All giving the same result.
CoffeeCat
(24,411 posts)Quinnipaic 11/16-11/22--Clinton 51; Sanders 42
CBS News YouGov 11/15-11/19; Clinton 50; Sanders 44
Those show Sanders within single digits.
Not the 24-point spread that Monmouth shows.
Buzz Clik
(38,437 posts)Variance in the results is to be expected.
Talk to me when Monmouth is the only poll showing Hillary with a substantial lead.
CoffeeCat
(24,411 posts)...and when you look at who was polled--that variance makes sense.
Older voters; more female voters; not counting younger Democrats who hadn't voted in one of the last two primaries. All of that skews toward Hillary.
The result of the poll makes perfect sense, when you peek behind the curtain.
yardwork
(61,722 posts)The Monmouth results are at the high end, but they fit with the overall trend of recent polls, as shown by Buzz Clik's post.
I studied statistics.... I'm not seeing the severe problems you cite.
SCantiGOP
(13,874 posts)Clinton has better than a 2 to 1 lead in SC. I expect she will end up with 60%+ which will make her win in Iowa - even if you're right and she only wins by 9% - not terribly significant.
DCBob
(24,689 posts)The most recent (just a few days old) CNN and Monmouth have Hillary up big.
You are grasping.
Kalidurga
(14,177 posts)Let them smell the coffee after breakfast is over.
CoffeeCat
(24,411 posts)...have one purpose--to make Sanders supporters give up and feel that he can't win.
It's a parlor trick.
I think the last two polls that had Sanders within single digits of Hillary (exactly where Obama was before he won the Iowa caucuses, and Hillary came in third) have caused a bit of fear in the Clinton camp.
Buzz Clik
(38,437 posts)Fortunately, only the votes matter.
Fuddnik
(8,846 posts)Buzz Clik
(38,437 posts)Kalidurga
(14,177 posts)We talk about this at meetings all the time. We don't tell people to ignore the polls we tell them to caucus or go vote in the primary.
JaneyVee
(19,877 posts)CoffeeCat
(24,411 posts)...behind that poll.
I'm afraid of what I might find.
There are so many hours in the day that I have to unravel the junk science leveraged and touted by the Clinton camp.
riversedge
(70,358 posts)CoffeeCat
(24,411 posts)And if the CNN poll isn't flawed, that's fine.
However, the Monmouth poll is.
I'm interested in truth.
cascadiance
(19,537 posts)... even if it seems that such criteria for the poll is "neutral" when it isn't in the way it is being conducted.
It shows that so many of the other polls out there that don't expose the details on how they are measured could be doing the same things too.
It perhaps EXPLAINS why the polls in 2008 were so FULL OF IT when they basically almost predicted a Clinton victory, when the voters wanted something different than what those polls then "measured".
And I think many of us need to step back and ask ourselves who most likely funds and directs how these polls are conducted and where is that money coming from, and what interests do these polls serve in terms of who is funding it.
Corporate interests have funded our politicians heavily (and giving away that money would be pretty unprofitable for them to do as "good businessmen" if they didn't get some "favors" back in return).
Well it isn't really any secret which candidate is more likely to repay "favors" to who is funding their campaign for the wealthy in this campaign, even more so than in 2008! So its not hard to see why those doing the polls and those funding who's doing the polls want an outcome that helps push someone in that is more likely to repay favors to them. And that person that is repaying them favors is definitely NOT likely to be Sanders in this election, who has far more average voters supporting him than any other candidate as opposed to big Wall Street or other corporate donors.
riversedge
(70,358 posts)Oh knock yourself out--un-skewing yet ANOTHER recent poll. Have a good day
Hillary Clinton Leads in 2016 Match-Ups as Ben Carson, Marco Rubio Run Closest
http://www.nbcnews.com/meet-the-press/clinton-leads-2016-match-ups-carson-rubio-run-closest-n474591?cid=sm_twitter_feed_firstread
by Mark Murray
Hillary Clinton leads the Republican presidential field in hypothetical general-election match-ups, with Ben Carson and Marco Rubio running the closest to her, according to a new national MSNBC/Telemundo/Marist poll.
And with just one exception, the margin of Clinton's lead among Latino voters determines just how competitive each match-up is.
Clinton's biggest lead is against current GOP frontrunner Donald Trump: She's ahead of him by 11 points among all voters, 52 percent to 41 percent, and a whopping 42 points among Latino voters, 69 percent to 27 percent.
..................
.............. The MSNBC/Telemundo/Marist poll also finds that a generic Democratic presidential is tied with a generic Republican at 45 percent to 45 percent, while that lead expands to 20 points among Latino voters, 56 percent to 36 percent.
To put these Latino poll numbers into perspective, Barack Obama defeated John McCain among Latinos by 36 points in 2008, 67 percent to 31 percent.
And he beat Mitt Romney by 44 points in 2012, 71 percent to 27 percent..................
cascadiance
(19,537 posts)You aren't responding to the note in this post that shows how polls are rigged, and therefore to put too much stock in them is not getting us anywhere. Especially when those interests that control the small oligopoly of media interests are better served by pushing those candidates that better serve reinforcing the corporate oligarchy for the wealthy elite's interests, not ones like Sanders who wants to tear down this oligarchy to serve the interests of average Americans more.
NurseJackie
(42,862 posts)I think that these professional polling organizations know what they're doing. It's unlikely that they've never considered the possibility that new first-time voters are constantly entering the stream. And I'm sure they know enough to consider cell phones vs land lines, and they know how to weight the results appropriately. They aren't idiots, you know?
Many Bernie supporters will be Independents (and maybe some Republicans) who will not be registered Democrats until the night of the caucus.
Why should anyone assume that those independents and cross-over folks are automatically Bernie's?
I know these are difficult and frustrating times for Bernie's fans. But, on the bright side, this will all be over for him by the end of March.
CoffeeCat
(24,411 posts)...but having an open primary makes a huge difference. You can literally walk into your caucus, ten minutes before it begins, and register as a Democrat and caucus.
There is a very large contingent of Independents in Iowa, who do not want to be affiliated with any party. There are many reasons for that, but many in Iowa can't stand the endless mountain of political junk mail and phone calls with which we are inundated during the caucus years.
I love politics, but 5 phone calls a night; and your mailbox clogged with multiple flyers, can be annoying.
This is a "thing" in Iowa. And it is but one of the many reasons that the poll is flawed. I put that reason last, because I think the age, cell phone and gender flaws are even worse.
NurseJackie
(42,862 posts)wildeyed
(11,243 posts)help even out the outliers a bit. I looked and the poll is a bit high, but not way outside what others show. It gets evened by a couple of lower, earlier polls to +14 if you round.
http://www.realclearpolitics.com/epolls/2016/president/ia/iowa_democratic_presidential_caucus-3195.html
Every single time I see people claiming that the electorate looks waaaay different than the polling says, they are wrong. I did it during GWB vs. Kerry. We said the same stuff.... too many olds in the sample, landlines, etc. We were wrong. My disappointment was huuuuge, but I learned the hard way. Romney did it last go 'round. To himself, sadly. Experienced politicians should know better.
Obama did win Iowa, and by more than expected, but he was inside the margin of the aggregate polls by this date if I read it right. And there was huge buzz about his ground game the month or so before. I am not hearing this about Sanders. Also, Edwards was a legit threat at that point, so the race was really different. http://www.realclearpolitics.com/epolls/2008/president/ia/iowa_democratic_caucus-208.html
Does anyone know what the voter registration process for Iowa caucuses? I know you can change party affiliation at the caucus, but do people need to register to vote with the Board of Elections ahead of time? If so, what is the cut off date? If Iowa is seeing a huge wave of NEW registrations, that leaves the possibility of surprise turnout.
My state saw huge gains in young and dem registrations in both '08 and '12, but Obama only won here in '08 and by a sliver, so no guarantees either way. And SC looks like the undecideds are breaking for Clinton, at least according to the graph at Real Clear Politics.
NurseJackie
(42,862 posts)Outliers can be fun and exciting (or depressing and shocking) ... but the trends tend to give a more realistic idea of what's happening.
MaggieD
(7,393 posts)I'm not a pollster, but imagine they are trying to tailor it to who usually shows up at caucuses. Either way, I feel very confident she will win the nomination. I surely do not believe ALL the polls are wrong.
Segami
(14,923 posts)Attorney in Texas
(3,373 posts)n8dogg83
(248 posts)from 10/22 - 10/25, Bernie was down by 41%. So taking the poll at face value would indicate that Bernie has made significant gains over that period of time.
brooklynite
(94,801 posts)...you'd know that Monmouth did two extra runs, with larger estimates of voter turnout (the "secret" voters)...Clinton came out on top of both.
CoffeeCat
(24,411 posts)Polling more women, polling an inordinate number of senior citizens and polling a large contingent who were answering land lines.
What is astounding about all of this--is that one of these flaws would make the poll questionable.
When you put all of this together--this poll should be relegated to some kind of junk-science poll museum. You know, spotlights on it, velvet ropes around it. The whole sha-bang.
brooklynite
(94,801 posts)I re-ran the aggregator with Monmouth omitted...
Clinton is consistently above 50% and rising.
Feel better?
riversedge
(70,358 posts)CoffeeCat
(24,411 posts)I still have problems with the Monmouth poll, and I think it should be nuked from orbit.
What's important to me is that the information is accurate.
I don't think Monmouth is accurate, so we'll just have to agree to disagree.
No one has ever re-aggregated for me, so this is quite special.
yardwork
(61,722 posts)taught_me_patience
(5,477 posts)"Final sample is weighted for age and gender based on state registration list information on the pool of voters who participate in primary elections."
They take the random sample and weight it based on the demographics of those that participate in primaries. You understand nothing about polling.
DemocratSinceBirth
(99,716 posts)2008 exit polling indicates women represented 57% of Iowa caucus voters:
http://abcnews.go.com/images/PollingUnit/IADemHorizontal.pdf
upaloopa
(11,417 posts)then there is no control over who is selected.
There is no control over age and gender.
If the pollsters selected people by age and gender it would not be a random sample and would therefore be meaningless.
The pollsters also have no control over whether the people selected had a land line or cell phone.
Your points are not valid
What you are saying is that the pollsters should introduce bias into their polls.
This is so much like 2014 with the repubs saying the polls are skewed.
Politicalboi
(15,189 posts)I can predict that Hillary will be ahead in all polls ONLY because she's Hillary. People know who she is, and that is ALL. Us Bernie supporters should ONLY be concerned about getting Bernie message out there, and voting for him. Don't respond to these bullshit posts that have Hillary already at 1600 Pennsylvania Ave.
Let's hope Hillary's people pull a Karl Rove come primary night.
misterhighwasted
(9,148 posts)You are referring to what exactly?
Politicalboi
(15,189 posts)Rmoney could still win, and he had a melt down.
CoffeeCat
(24,411 posts)...and she came in third.
It wouldn't be the first time that she went in polishing her tiara and came out gobsmacked.
brooklynite
(94,801 posts)...I know, I asked Robby Mook personally.
By the way, do keep up the Clinton insults. The more emotional your view of the Primary is, the more likely your efforts for Sanders will be misdirected.
CoffeeCat
(24,411 posts)...and I will say this. My primary reasons for supporting Bernie are rooted in his policy stances. However, I am also motivated, in part, to make sure that Hillary Clinton is not the next President.
I have always been fearful of her neocon tendencies and her cozy relationship with Wall Street and the defense industry. I haven't cared for her policies, for years. It's not personal. I don't know Hillary Clinton.
But some of her supporters have taken my disapproval of her to a new level. The snark, the hubris, the constant throwing of cold water on any Sanders supporter who posts something positive about Sanders. I don't know who these people are, but they are only motivating Sanders supporters to new levels.
I was an absentee-ballot courier for Kerry, and a precinct captain for Obama at our caucuses. However, I never did the hard work--the canvassing, the phone calling, etc. I am doing those things for Bernie now.
So really, the emotion that has been stoked in here by the rudeness and "sorority girl" mentality of many of her online supporters has given me that extra push to work hard for Bernie. When I walked into Bernie's campaign office, I was greeted with kindness and professionalism. Quite the stark difference from the middle-school vipers that nest in here.
It may be that the anonymity of the internet causes people to become bullies and jerks. Whatever it is, it is only emboldening Sanders supporters to do more to get him elected. It isn't deflating us. And the behavior in here only reinforces the mistrust that many of us have for Clinton.
taught_me_patience
(5,477 posts)"Final sample is weighted for age and gender based on state registration list information on the pool of voters who participate in primary elections."
Unskewing polls is really grasping at straws.
NurseJackie
(42,862 posts)leftofcool
(19,460 posts)CoffeeCat
(24,411 posts)...but I was hoping that people might be interested in the flawed methodology.
I don't think the people who conducted this poll are intelligent enough to engage in any kind of conspiracy.
They polled 56 percent females and 69 percent of their respondents are older than 50.
It's amazing they know how to tie their shoes!
...but then again, maybe you know something that I don't.
taught_me_patience
(5,477 posts)They take the results of the random sample and weight it to the demographic of typical caucus goers. Is says so right there in the methodology.
"Final sample is weighted for age and gender based on state registration list information on the pool of voters who participate in primary elections."
You are not seeing the results of the random sample. You're seeing the weighted results. These people are experts in their field and you clearly do not know what you're talking about.
CoffeeCat
(24,411 posts)...but what I don't understand is how a pollster determines what a "typical caucus goer" is for an upcoming 2016 caucus for which there are so many unknowns?
Wouldn't caucus demographics change from election to election, making it difficult to aggregate who exactly a "typical caucus goer is" for a caucus that hasn't happened?
I'm sure the "typical caucus goer" was very different in 2008, than it was in 2004, for sure.
And another question:
If polls weight their results to the "typical caucus goer" could that be the reason why Obama's 2008 caucus win (and Hillary's third place) was a somewhat unexpected--all of the candidate camps?
In 2008, an unprecedented number of younger people showed up to caucus for Obama. So, if the Iowa polls (taken before the 2008 caucuses) were weighted for the "typical caucus goer" as you said, then those polls would have given less weight to younger voters in 2008, correct? This could partially explain the results, which were surprising to many. Am I smoking my socks here?
I'm listening and very interested. I'm not a numbers gal. I'm a science journalist who worked in PR for many years. I'm interested in information.
NurseJackie
(42,862 posts)It is frequently the case that the people who answered your poll are not fully representative of the region you were polling over. Weighting is a technique to adjust answers to account for over- and under-represented groups.
For example, suppose the population in an area is 55% female; but 70% of my survey answers come from women. When reporting my survey results, I'd want to to boost the importance of men's answers a bit, and reduce women's answers. Weighting is the process to do this.
BlueCheese
(2,522 posts)But the gender split seems completely reasonable for the Democratic Party.
MohRokTah
(15,429 posts)oasis
(49,430 posts)workinclasszero
(28,270 posts)It's flawed, soooo flawed people!
NurseJackie
(42,862 posts)Ferd Berfel
(3,687 posts)I don't dive into the demographics of these polls too often but I have done it a few times recently. Every poll I have dipped into does that same thing. They don't poll 18-34 crowd. Also are they polling via land line or Cell phone?
SmittynMo
(3,544 posts)Don't ever believe a poll ever again. There are far too many variables. Too many ways to skew numbers in 1 direction.
Keep the faith in your candidate, and vote on judgement day. All the rest is BULLSHIT. PERIOD!!!
gcomeau
(5,764 posts)And for 3 out of the 4 items on your list I will do so again as they are not really important. (There are usually more women voters than men, get over the cell phone/land line thing people harp on that every election to no effect, etc...)
However, the age breakdown actually *is* way out of whack. I cannot fathom a justification for it. That is completely out of line with voter demographics. 69% over 50??? Expanding that to even over 45 the 2012 breakdown never broke 60%, let alone approached 70%. That's ridiculous.
DCBob
(24,689 posts)Hillary up by 19 points and stabilizing
http://elections.huffingtonpost.com/pollster/2016-iowa-presidential-democratic-primary#!mindate=2015-08-01&estimate=custom
DCBob
(24,689 posts)Sounds like typical polling methodology to me.
aelroy
(1 post)You know what CoffeCat? This is all for the best. The greater the fabrication the more of a devastation initial primary losses will be for Hillary. I wouldn't worry about this too much. The numbers on the ground is what is going to matter.
CoffeeCat
(24,411 posts)I just find the fabrication interesting, from a PR perspective.
It will be interesting to see the ground games. Being in Iowa, I get a front-row seat and I'll definitely be watching closely.
Thinkingabout
(30,058 posts)Godhumor
(6,437 posts)The gender breakdown is almost dead on to what was observed in 2008 which was 57% female to 43% male.
Cellphones being at 28% is pretty norm for the course and not an attempt to bias.
Likely voter screens, almost by definition, have to use past history as a predictor.
Age ranges are different from 2008. 60% of Caucus voters were 45+. 17% were 24 or under. Monmouth looks to have used a screen assuming the unprecedented youth turnout in 2008 for Obama would not happen again in 2016. So looking back to 2004, people 45 and older did represent 68% of voters. Only 17% were under 30.
3 of your 4 points I would say aren't causes for concern. Age could be debatable based on belief in youth turnout.
For the record, I think Monmouth is too high on Clinton. I think she is probably around +12 or so with likely voters.
DemocratSinceBirth
(99,716 posts)Scuba
(53,475 posts)NCTraveler
(30,481 posts)It renders your whole point useless.
WI_DEM
(33,497 posts)Response to CoffeeCat (Original post)
Name removed Message auto-removed