2016 Postmortem
Related: About this forumThe newly released Marist/NBC poll is flawed. Excludes first-time IA caucus goers/younger voters
I just read the methodology used by the Marist/NBC poll that shows Clinton leading Sanders in Iowa by a few points. I was willing to accept this poll. I think it's normal for polls to bounce around.
I'm sure Clinton supporters will laugh at me and mock me, "Haahahaha...Oh she's unskewing the polls again! Unskew! Unskew!" I'm just looking for reliable data.
Outlined in "Marist Poll Methodology for Iowa" is the following (link below):
426 likely Democratic caucus-goers defined by a probability turnout model which determines the likelihood respondents will participate in the 2016 Iowa Republican/Democratic Presidential Caucus based upon their chance of vote, interest in the election, and past election participation.
In short, their polling sample of "likely Democratic caucus-goers" was determined, in part, by past election participation. So, those who have never participated in an Iowa election or caucus are excluded.
The effect?
--First-time Iowa caucus-goers/voters (regardless of age) are excluded from the poll.
--Younger voters are excluded from the poll.
Marist doesn't say if "past election participation" means participation a previous caucus (which was 8 years ago) or previous voting (which was four years ago). The poll, for sure, excluded everyone under the age of 22; and possibly everyone under the age of 26--depending on how they defined "past election participation". Flawed on both accounts.
Ann Selzer--who does the Iowa Poll--predicted Obama's 2008 Iowa caucus victory. She's the "gold standard" of Iowa polling and Nate Silver gave her an A+ rating for her stellar methods. She has repeatedly stated that polling from a sample of Iowans who have voted/caucused before is foundationally flawed. Because you're excluding those younger voters and first-time caucus goers.
Selzer starts with registered voter lists, then asks FIRST, "Do you plan to attend the Iowa caucuses?" This allows for the most accurate, science-based results. This NBC/Marist poll does not.
The Quinnipiac Poll that came out yesterday, showing Sanders ahead by 4, used the same methodology that Selzer uses. That was a sound poll.
Ann Selzer's Iowa Poll--the last one before the Iowa caucuses--comes out this Saturday. Count on that result being extremely close to the final result.
Link to Marist methodology:
http://maristpoll.marist.edu/wp-content/misc/IApolls/IA160124/NBC%20News_WSJ_Marist%20Poll%20Iowa%20Tables%20of%20Likely%20Republican%20Caucus-Goers_January%2028%202016.pdf#page=1
liberal N proud
(60,335 posts)You see the thing about early primaries and specifically the caucus, they can surprise the hell out of you.
You still have to get those "young" voters to attend the caucuses and the weather may play a big part in that.
The young voters do not have a very good record for turnout in the past few elections so who knows.
CoffeeCat
(24,411 posts)and regardless of the weather and whatever anyone thinks might happen--this is about a poll that excludes everyone in Iowa that was younger than 22 and anyone who is a first-time caucus goer.
That's not reliable data, regardless of what you--or anyone else is guessing--about the behavior of those groups.
So your points are moot.
Given all of that--Younger voters turned out in droves in 2008 for the Iowa caucuses. The weather will not be a factor on Monday. Clear skies and warmer weather predicted for Iowa (However, a huge snowstorm is predicted to begin on Tuesday morning, hours after the caucuses are finished).
Jarqui
(10,126 posts)41% men, 59% women is a significant problem with it as well. Put those two together and Bernie is probably ahead.
I do not know what it is this year but nearly all the polls I've checked have favorable demographics for Clinton. Even the Quinnipiac Poll yesterday that had Sanders +4 - should have been closer to +6 or so if you project a "correction" for the demographics. I don't know why but it's to the point I'm a little (not gravely) suspicious.
CoffeeCat
(24,411 posts)They do not use flawed polling samples as this NBC/Marist poll does. So that's good. The accuracy is better.
It's possible that they skewed more toward women, because more women tend to participate in the Iowa caucuses.
That's not as egregious as what NBC/Marist did by excluding everyone under 22 and first-time caucus-goers.
But it seems odd to me that a pollster would use opinion/predictions about what they think will happen in their sample. Predictions about who will show up aren't always accurate.
Godhumor
(6,437 posts)Women should be higher in the screen.
Jarqui
(10,126 posts)Age demographics also favored Hillary
so both factored.
It wasn't tons but in a tight race, a couple of points is a couple of points.
in_cog_ni_to
(41,600 posts)Why am I not surprised?
Thanks CoffeeCat! You're saving my sanity.
PEACE
LOVE
BERNIE
still_one
(92,217 posts)relying on the polls.
It is also more difficult because it is a caucus rather than a primary.
The only poll that means anything is the actual caucus results themselves.
HerbChestnut
(3,649 posts)Iowa is a toss up right now, and anybody who clings to one single poll as proof that their candidate is ahead is just spinning their wheels.
CoffeeCat
(24,411 posts)You just never know what the turnout will be, until you get into that room.
I am seriously dying over here. I cannot stand the wait.
So, I look to these polls to give me any sort of reliable info. Which is why I read heavily into how they achieved the numbers they did.
With that said, there is one poll that I would consider very reliable; almost predictive. Ann Selzer's Iowa Poll. And the last one before the Iowa caucuses is released this Saturday. I am on pins and needles waiting for that poll. She predicted Obama's 2008 win and many other Iowa races. She was accurate on many races, when other pollsters were not.
Democrats and Republicans alike consider her the Goddess of Iowa polling.
And I have no idea how that poll will land. It's terrifying waiting for it. I almost wish that an accurate poll, such as hers, didn't exist. She's so accurate that whatever she predicts will most likely be extremely close to the final result. I am excited to caucus for Bernie, but I just want this DONE.
cascadiance
(19,537 posts)... and my bet that this also skews results if many of them hadn't participated in a past caucus election, if this is measuring participation in primary/caucus elections rather than general election behavior.
CoffeeCat
(24,411 posts)The largest voting block in Iowa is Independents. They are 42 percent of all registered voters in Iowa. That's huge.
And they skew heavily toward Bernie.
Ann Selzer describes how Obama won with a "trilogy" of support: First-time caucus goers, younger voters and Independents. That's exactly what Bernie has coalesced. Exactly.
The question remains...are there enough of those voters to pull a win? And are those voters spread around the state? He's got to pick up voters in the more conservative/red parts of the state. Bernie has been campaigning heavily in those areas and in small towns. He's got offices in areas where Clinton did well in Iowa in 08. His rally attendance is also very impressive in those areas. So, there's definite enthusiasm.
He's also doing extremely well in Polk County. I don't see a lot of people discussing this. Polk County has 17 percent of the state's total delegates. It's very liberal, and it's become more liberal and "younger" since 2008. Nearly 20,000 new young people have moved into the downtown area since 2008. It's been completely revitalized--literally transformed. Most of these new people are young professionals.
cascadiance
(19,537 posts)... who arguably had an even more progressive message than Obama's "Hope and Change" message did last time too. Don't think that O'Malley will get quite the same voter support that Edwards did then The question is who are those voters that voted for both Obama and Edwards going to vote for this time? Maybe some for Hillary, but I'd bet more of them go to Bernie.
Godhumor
(6,437 posts)Again, all in the demographics
CoffeeCat
(24,411 posts)...and that's probably an accurate reflection of what will be in the room on Monday.
Godhumor
(6,437 posts)Right in the methodology on page 4 is the line breakout for past participation.
Likely caucus goers:
66% of sample were previous participants
34% are first timers
CoffeeCat
(24,411 posts)Did you read that at the link that I provided? I'm interested in reading what you have found.
What I read was on one page...not four. I'd love to know what you read.
And thank you!
Godhumor
(6,437 posts)38% newcomers
51% not on Democratic verified voter lists.
16% are 18-29 years old
CoffeeCat
(24,411 posts)Thank you.
Godhumor
(6,437 posts)Marist actually does a pretty good job of lading or their demographic subcategories. Much better than a lot of pollsters.
book_worm
(15,951 posts)apparently only polls that have Bernie ahead are accurate. As for all of these polls they are really meaningless at this point. We will see what happens on Monday night.
hedda_foil
(16,375 posts)They"get it" much more than any other age cohort because they know both intellectually and emotionally that they and their children will bear the brunt of the consequences. Because of that, coupled with their level of support for Bernie, I think their turnout may be even bigger than it was for Obama.
nolabels
(13,133 posts)The Younger people and their movement showed up to put Obama over the top, I wonder if they can do it again.
This will be interesting to watch and a very close finish will make things even more complex The establishment hopes for bad weather and low turn out.
I would say though, if six months ago you would of said things would be like this then...
a lot of people would be saying 'I like our chances'
NCTraveler
(30,481 posts)Until she puts out a scathing unskewing video I will take the poll as highly accurate reflection of where the race currently stands.