THE USC DORNSIFE / LA TIMES PRESIDENTIAL ELECTION "DAYBREAK" POLL

LA Times ^ | 6 Nov 2016

[mandaladon]

Trump 48.2

Hillary 42.6

[hoosiermama]

Woo Hoo. Thanks Larry!
Answered prayers!

[Pearls Before Swine]

Thanks.

So, the systematic error I worried about lies in whether people who are reachable through an internet poll might be more likely to vote Democrat or Republican. To be determined.

[justlurking]

Sorry, I erred: it is weighted. But, it's not weighted by party ID.

The details are here: https://uas.usc.edu/documents/uas/UAS%20Weighting%20Procedures.pdf

[justlurking]

So, the systematic error I worried about lies in whether people who are reachable through an internet poll might be more likely to vote Democrat or Republican.

Actually, they do address that. They provide a tablet and broadband access if necessary, once a respondent is selected.

More info here: https://uasdata.usc.edu/recruitmentoverview/1

[Pearls Before Swine]

Wow. That is quite detailed.

Whether it is an accurate procedure for removing data bias ... we’ll see. I do like it that the vote is accompanied by a demographic questionnaire.

[Impy]

Iowa Congessional Polls

IA CD-3 (Loras): David Young (R) 44%, Jim Mowrer (D) 39%

IA CD-1 (Loras): Rod Blum (R) 47%, Monica Vernon (D) 41%

GOP still ahead, establishment GOP Young in swing seat, conservative Blum in dem leaning seat.

This same pollster has Trump DOWN 1 in Iowa, kinda leads to conclude these seats aren’t in danger.

‘Nother Iowa Poll

P2016 IA (Des Moines Register): Trump 46%, Clinton 39%, Johnson 6%, Stein 1%

CO USSEN (Gravis/Breitbart): Michael Bennet (D) 47%, Darryl Glenn (R) 44%

This race has closed, major.

NY Congressional Polls

NY CD-1 (Siena): Lee Zeldin (R) 51%, Anna Throne-Holst (D) 38%

NY CD-3 (Siena): Tom Suozzi (D) 47%, Jack Martins (R) 40%

NY CD-19 (Siena): John Faso (R) 48%, Zephyr Teachout (D) 42%

Teachout is the nut who ran against Cumomo in the rat gov primary.

NY CD-22 (Siena): Claudia Tenney (R) 38%, Kim Myers (D) 34%, Martin Babinec (Ref) 16%

[Ernest_at_the_Beach]

Well....then the gap could grow even larger.....

BIGGEST EVER????

[Stevenc131]

A 3% advantage is supposed to be a lock for an electoral victory, at least that’s what the pollsters used to say. But with all the push polls and the tricks used today, who knows?

[phoneman08]

Trump only down 3 with Hispanics? Wish that were realistic. Is a small sample group, so large swings happen within the sub groups.

That said, the past several days have had a large percentage movement among self identified Hispanics in this poll for the first time in the life of this poll. Strange.

I wonder if the spirit cooking story all over social media has something to do with this.

I do hope this poll turns out close to the actual vote. 5+ points in the national vote would be an absolute route. Not to mention the satisfaction we all would get when someone ask the “experts” why they ignored the LA Times poll that was one of the most accurate polling methodologies in 2012.

[GLDNGUN]

Whether it is an accurate procedure for removing data bias ... we’ll see

How accurate was it 4 years ago?

[phoneman08]

The team responsible for the Daybreak Poll four years ago developed the successful RAND Continuous Presidential Election Poll, which was based on the same methodology.

The Daybreak Poll is based on an internet probability panel survey. Daybreak Poll members are participants in the ongoing UAS internet probability panel of about 5,500 U.S. residents who were randomly selected from among all households in the United States. Members of recruited households that did not have internet access were provided with tablets and internet service. The UAS panel is still growing. We project it will reach about 6000 members in the coming months.

[roostercogburn]

It was very accurate 4 years ago with the final tally.
They called Obama with a 3.35 point win, I think actual was 3.85.

[Cheerio]

Don't be too shocked upon waking Wed morning to find the RATS challenging every state they can. Even a landslide win by the Trumpster will not go down easily as the RATS scurry off to their favorite "fixing venue", the Courts. Never forget Al Franken's stolen victory in 2008.

EXCLUSIVE: Virginia Gov. Pardons 60,000 Felons, Enough To Swing Election

When 1,099 Felons Vote In A Race Won By 312 Ballots

[GilGil]

At this point it is safe to assume that Trump is 50+ and Hillary is under 40!

[jimnm]

This poll could have a response bias as it is requires a bit more work on the part of a respondent to register, compared to a one time phone survey, perhaps balanced by the financial reward (the reward would also be a potential issue). They then continue to receive a financial reward each time they participate (which would be 20 dollars once a week).

They allow all members 18+ from a survey’s household to participate. This will multiply the financial reward and could lead to households with more than 1 member being more likely to register to begin with.

In any case the trends are great, any polling is imperfect, and traditional phone surveys have their own issues. It’s also interesting because it tracks the same people over time, once per week, rather than being a new random sample each day. I’m looking forward to a Trump victory and a liberal meltdown!

[rocklobster11]

The key thing about this poll is that they sample from the same pool of people every day, so you can get an idea of what the trend is regarding changing opinions as opposed to just sampling error.

The poll has been going since July, and Trump has been up for 50-70% of that duration, but there have been swings and Hillary was virtually tied or ahead From end of July until Sep 11 when she had the medical problem. At that point, the poll swung heavily to Trump until the Access Hollywood tape. Hillary took a brief lead, but the poll has been moving continuously to Trump since the week before the FBI reopening the email case.

Trump now has close to his biggest lead, so he has peaked at the right time.

It also matches historical trends of undecideds going to the challenger 2:1 in the closing days of an election.

[NKP_Vet]

NBC/WSJ and of course ABC tracking poll has the bitch from hell up 5 points this morning. Both polls say the bitch from hell is getting the majority of undecideds and momentum started swinging to her AFTER Comey reopened the investigation. Fox News poll from 1-3 November has leading Trump 44-43. Any poll that has Trump leading is summarily ignored by the media, including Fox News.

[chad_in_georgia]

I wish I had your confidence. Just seems like we have too many needles to thread to pull it off.

[Gay State Conservative]

Don't be too shocked upon waking Wed morning to find the RATS challenging every state they can...

Every state that DJT carries by less than,say,1% will be challenged by the Rats.And with SCOTUS at 4-4 there's no telling *what* could happen.

[Maelstorm]

If this poll is right as it was in 2012 then its going to be a good night Nov 8th. The good trends I see is that Trump is closing the hispanic gap and the woman gap while surging higher with men. Those criticizing this poll don’t understand the science of polling. The important thing is having representative samples of the electorate. It doesn’t matter if you constantly change the members of the sample or not. Polls in the wild have big problems. Its increasingly hard to get respondents and its easy to get very skewed samples on occasion plus when you throw in human interviews and push poll questions as nearly all the major network polls do you tend to inject bias into the poll. The LaTimes poll isn’t like polls in the wild because it isn’t calling random people unannounced interrupting their lives and it has a population of nearly 3000 respondents.

Now the critics could argue that the sample is wrong and it could very well be and the weighting could be wrong but we won’t know that till election day. Last time this methodology worked very well and did a far better job of predicting Obama’s margin of victory.

https://alpdata.rand.org/?page=election2012#election-forecast