Welcome to DU! The truly grassroots left-of-center political community where regular people, not algorithms, drive the discussions and set the standards. Join the community: Create a free account Support DU (and get rid of ads!): Become a Star Member Latest Breaking News General Discussion The DU Lounge All Forums Issue Forums Culture Forums Alliance Forums Region Forums Support Forums Help & Search

RandySF

(58,745 posts)
Wed Jul 4, 2018, 09:57 PM Jul 2018

Why some polls may be underestimating Democrats

(CNN)Republicans dominated the midterm elections in 2010 and 2014 in part because voters who supported Barack Obama in presidential years voted for Republicans or stayed home when he wasn't on the ballot.

The turnout advantage Republicans enjoyed during the Obama years may no longer exist as we head into 2018, and in fact, there are some signs that Democrats may be the one with a decisive turnout advantage this cycle.

This shift in reality may not be reflected in much of the high-quality polling so far this year because much of it has been done among registered voters instead of the more refined category of likely voters. In other words, current polling may not fully reflect Democratic strength heading into the midterms.

Monmouth University has taken into account the gap between registered and likely voters. Monmouth has polled seven House and Senate races so far. In each of them, Monmouth identified "likely voters" based mostly upon a high propensity for voting in the past. Additionally, those voters who had a high-level interest, even if they haven't historically voted in midterms past, are included.

This model I feel is one that captures both the fact that past voting is usually highly predictive of future voting and that there may be some new voters casting ballots this year.

In five of the seven races polled, the Democratic candidate either did better or no worse when Monmouth switched from registered voters to likely voters. And in the two races where the Republican candidate does better with the switch to likely voters, the change was 2 points or less.



https://www.cnn.com/2018/07/04/politics/why-some-polls-may-be-underestimating-democrats/index.html

13 replies = new reply since forum marked as read
Highlight: NoneDon't highlight anything 5 newestHighlight 5 most recent replies
Why some polls may be underestimating Democrats (Original Post) RandySF Jul 2018 OP
That is good news and I will take wjatever I can get. BigmanPigman Jul 2018 #1
Bingo,you noticed what Wellstone ruled Jul 2018 #2
Interesting paleotn Jul 2018 #3
Who believes polls anymore? Meadowoak Jul 2018 #4
Polls were accurate in 2016. RandySF Jul 2018 #5
I do titaniumsalute Jul 2018 #9
The polls I have seen always favor Republicans, is there a good poll to follow? Meadowoak Jul 2018 #11
An aggregate of polls is best titaniumsalute Jul 2018 #13
I wonder if anyone is taking into consideration the shrinking size of the Republican Party. Vinca Jul 2018 #6
Predicting turnout is tricky DeminPennswoods Jul 2018 #7
That's nice Takket Jul 2018 #8
K&R Scurrilous Jul 2018 #10
We have a perfect storm of repug politician sleaze and freakishness, like '06 Mc Mike Jul 2018 #12
 

Wellstone ruled

(34,661 posts)
2. Bingo,you noticed what
Wed Jul 4, 2018, 10:07 PM
Jul 2018

I suspected all along. After doing some polling in the Late nineties and again in the 2006-2010 time frames for the same Polling Agency,what you mentioned is what I understood to be their method of contact. Again,what I found was the use of the same polling contact data base. That of those who voted in the most recent Election,or the most recent registered live voter rolls. Meaning those who bothered to show up at the Voting Polls.

It is all about the methodology in use. And of course the Structure of the asked questions. To damn much push polling in use.

paleotn

(17,911 posts)
3. Interesting
Wed Jul 4, 2018, 10:10 PM
Jul 2018

That's the difference between simply having information and interpreting it. The latter is as much an art as a science.

RandySF

(58,745 posts)
5. Polls were accurate in 2016.
Thu Jul 5, 2018, 01:49 AM
Jul 2018

Finals polls gave Hillary a slight national lead with kry states too close to call.

titaniumsalute

(4,742 posts)
13. An aggregate of polls is best
Thu Jul 5, 2018, 10:49 AM
Jul 2018

Methods vary by company which is healthy. But an aggregate of many polls is good. RealClearPolitics and 538 both do good at the aggregating.

Vinca

(50,261 posts)
6. I wonder if anyone is taking into consideration the shrinking size of the Republican Party.
Thu Jul 5, 2018, 05:46 AM
Jul 2018

Those who remain are the hardcore cult members and will definitely poll according to Dear Leader's orders.

DeminPennswoods

(15,278 posts)
7. Predicting turnout is tricky
Thu Jul 5, 2018, 06:08 AM
Jul 2018

The biggest problem is does a pollster use enthusiasm or past voting history to define a "likely voter"?
So far in elections held since the inauguration, actual results seem to indicate Dems have an enthusiasm edge, but I suspect pollsters at least somewhat lean toward using voting history which is probably a more reliable indicator than the more volatile and harder to define "enthusiasm" measure.

However, having been part of a live interview telephone survey recently done for my local congressional district, I think the interviewers can get a gauge of how likely a person is to vote from just banter with the person being interviewed. For example, there's usually questions using "on a scale of 0 to 10...". Sometimes interviewees jokingly answer "a 100" or "- 10" or they add something extra when answering a question.

Mc Mike

(9,114 posts)
12. We have a perfect storm of repug politician sleaze and freakishness, like '06
Thu Jul 5, 2018, 10:26 AM
Jul 2018

and '12.

We still need to crush them overwhelmingly, because if it isn't close, they can't steal it.

Latest Discussions»General Discussion»Why some polls may be und...