Internal poll megathread

[GeorgiaModerate]

Georgia D-Internal poll from Lake Research Partners has Clinton +1 (Clinton 41, Trump 40, Undecided 16) with a sample of 600 LV.  No three-way numbers are listed (Stein isn't on the ballot here).  No crosstabs.

Link: https://cmgajcpolitics.files.wordpress.com/2016/07/lrpmemo-georgiademocraticparty-f-072016.pdf

Favorability:

Hillary Clinton 42/53  (-11)
Donald Trump 39/55  (-16)
Barack Obama 51/45 (+6)
Michelle Obama 61/27 (+34)
Bill Clinton 53/39 (+14)

[GeorgiaModerate]

Kevin Yoder (R) internal poll of KS-03 shows Yoder +17 and Clinton +6 per Politico's Scott Bland (https://twitter.com/PoliticoScott/status/763792785057148928).  Romney won the district by 10.

[GeorgiaModerate]

Florida's gonna be reallllly tight this year

I'm gonna bank on everyone underestimating Puerto Rican participation and turnout and it not being as close as everyone expects.

http://www.politico.com/states/florida/story/2016/09/rubio-leads-by-2-in-internal-murphy-campaign-poll-105537

Hillary Clinton leads Donald Trump 46-43 percent in Florida, according to the poll of 800 likely voters by Global Strategy Group.

Sept. 6-11; 40% of the survey respondents were Democrats, 38% Republicans and 22% independents — a generally representative breakdown of turnout by party in Florida presidential elections.

At the end of the article it says that they purposely over-sampled Hispanics

Oversample generally means they poll more of a subgroup to get a lower margin of error, but their share of the vote is not overstated. For example, a national poll of 400 might only get 50 Hispanics to respond, which is a huge margin of error. So they could additionally focus poll 200 Hispanics, use the R/D split, and then fit it back to the 50/400.

You are such a hackish poster. And you absolutely work for the HRC campaign.

That's an absolutely unjustified insult.  Oversampling is a recognized technique to reduce the margin of error within the subgroup.  It does not change the overall result.  Here's a description of the technique from the well-respected pollster Pew Research (emphasis mine):

"Oversampling

For some surveys, it is important to ensure that there are enough members of a certain subgroup in the population so that more reliable estimates can be reported for that group. To do this, we oversample members of the subgroup by selecting more people from this group than would typically be done if everyone in the sample had an equal chance of being selected. Because the margin of sampling error is related to the size of the sample, increasing the sample size for a particular subgroup through the use of oversampling allows for estimates to be made with a smaller margin of error. A survey that includes an oversample weights the results so that members in the oversampled group are weighted to their actual proportion in the population; this allows for the overall survey results to represent both the national population and the oversampled subgroup."

This is from http://www.pewresearch.org/methodology/u-s-survey-research/sampling/.

Perhaps you should consider not insulting people based on your false assumptions.

[GeorgiaModerate]

Hmmm.. ME-02 should definitely be a Tossup IMO.

At least a Tossup.  Lean R is probably closer to the current situation.

[GeorgiaModerate]

https://twitter.com/reidepstein/status/784559054349799424

You must be logged in to read this quote.

[GeorgiaModerate]

You must be logged in to read this quote.

https://twitter.com/jmartNYT/status/785552978727972864

[GeorgiaModerate]

NBC moved the Dakotas and Montana to Lean Republican on their battleground map...do they have access to polling that we don't have to make that conclusion?

http://www.bloomberg.com/politics/articles/2016-10-04/which-states-can-gary-johnson-and-jill-stein-spoil

Clarity Campaign Labs is projecting Montana to be a 1.1 Trump win. A Johnson surge or Trump collapse could move it in her favor.

They're dreaming.  I can't see the third party vote breaking double digits anywhere except Utah, NM, and maybe Alaska.

[GeorgiaModerate]

Senate republicans have shared internal polls to the National Review and some of the results are interesting

According to these polls
Clinton +14 in NH
Clinton +11 in PA
Clinton nar­rowly beats Trump in AZ
Trump +2 in MO

And senate

You must be logged in to read this quote.

What are your thoughts? Belieavable?

Question is, will NH and PA be called at poll closing? If they are, we know what type of night it's going to be.

For President, probably both will be called immediately.  For Senate, neither.

[GeorgiaModerate]

Erick Erickson ‏@EWErickson  4m4 minutes ago
Washington Post has Clinton up 4 in Georgia.  Internal GOP polling has her up 5 in Georgia.

I really wanna believe it... but the Clinton campaign has clearly signaled that the state isn't worth the time and effort. I'd like to see a quality public pollster release something.

Anecdotally: I voted after work today in Forsyth County.  There was no wait, and perhaps a third of the voting machines were occupied.  The poll worker I asked said they'd had a steady stream, but it never got busy enough to require a line.  For comparison, when I voted early in 2012 there was a very long backup (same county, but different location).

[GeorgiaModerate]

Erick Erickson ‏@EWErickson  4m4 minutes ago
Washington Post has Clinton up 4 in Georgia.  Internal GOP polling has her up 5 in Georgia.

I really wanna believe it... but the Clinton campaign has clearly signaled that the state isn't worth the time and effort. I'd like to see a quality public pollster release something.

Anecdotally: I voted after work today in Forsyth County.  There was no wait, and perhaps a third of the voting machines were occupied.  The poll worker I asked said they'd had a steady stream, but it never got busy enough to require a line.  For comparison, when I voted early in 2012 there was a very long backup (same county, but different location).

Well that's quite interesting. If Trump can't hold Romney's margins in the Atlanta exurbs, then he's probably at risk of losing Gwinnett and maybe Cobb County. There are enough potential Clinton voters to turn the state blue, but it's pretty late in the game to a make serious effort to make it happen.

I think they should have been here earlier, but it just feels on the ground like she can win. I'm thinking Cherokee/Gwinnett will lose a few GOP voters (college educated/evangelical), but he's going to lose a horrific amount of support in Rubio country (Fulton/Dekalb). Even though that's not his base, it's a lot of votes. Lots of moderate Republicans I know refuse to vote for Trump.

http://elections.sos.ga.gov/Elections/voterabsenteefile.do
Kind of hard to use since they split the counties, but...below are the vote in person numbers as well as D-R margin and vote count. I'll try and figure out how to make a script to combine them to be easily pivotable, but for now, just went into each file...

(Obama-romney) 2012 votes

7.0K in Fulton (64-35) 390K
7.6K in Dekalb (78-21) 305K
1.6K in Gwinnett (45-54) 295K
3.7K in Cobb (43-55) 309K
3.5K in Forsyth (18-81) 82K

It does seem like Forsyth is punching above its weight relative to its low population.

One factor is probably that Forsyth makes it very easy to vote early.  There are five early voting locations (compared to one in Gwinnett, which has four times the population) with extensive hours: 8am-6pm Mon-Fri this week and next week, then 7am-7pm on Saturday the 29th and the following Mon-Fri.

[GeorgiaModerate]

It's a disgrace what they're doing in Gwinnett, just 1 in Lawrenceville for the whole county (not surprisingly, most of the early voters are from Lawrenceville).

http://www.tomnash.eu/how-to-combine-multiple-csv-files-into-one-using-cmd/
So I figured out how to combine the CSV files and did some analysis on it. Simply, ballot status A (accepted), comparing 2016 vote totals to 2012 final count. Obviously simple analysis that doesn't take into consideration of current voter registration.

Fulton is at 12.6K (now that I counted accepted mail), 3.2% of 2012
Dekalb at 4%
Cobb at 3.3%
Gwinnett at 2.3% (disastrous 1 voting site)
Forsyth is 6%
Clayton's at 3.4%
Muscogee is 4.1%
Columbia is 4.8%
Henry is 3.1%
Hall is 4.5%
Cherokee is 2.8%.

Overall, IF access is there, Republicans are doing better (Forsyth). Otherwise kind of mixed.

Also, I took each county's 2012 D/R margin, applied it to current county vote totals and came up with Trump 55% Clinton 44%. Hopefully as in person early voting ramps up this gap closes.

I think error will creep in because of the bolded part, although I confess I don't know what else you could do that's better.  I expect the proportions to change significantly in some counties, partly due to demographic shifts and party due to Trump's unique (non-)appeal to certain groups.  Forsyth was 80/18 for Romney, but I really doubt that Trump will hit 70 here.

[GeorgiaModerate]

I think I figured out why these internals are so pro-Democratic, relative to the other polls.  Basically, the Democrats are operating normally, showing polls favorable to them.  But, elite Republicans almost want a disaster to prove how bad Trump is, so they are leaking their worst internals.

They also want to retain the Senate and the House. Leaking these type of numbers don't help that cause.

Maybe not.  It might be a subtle way of distancing themselves from Trump without explicitly disavowing him, which would cause problems with his base.  This sends the message that Trump is going to lose, with the implicit point that GOP voters need to stick with their down-ballot candidates to act as a check on President Clinton.

[GeorgiaModerate]

margins favorable to Hillary, i suppose?

That's what he said in a later reply, but he refused to give details when asked.

[GeorgiaModerate]

https://twitter.com/7im/status/790673637946732544

Fascinating.  How do we apply the 5-point rule in a 3-way race?

[GeorgiaModerate]

Is this result for the district or statewide?

Just the district.

Can we extrapolate a statewide result from that?

If you add +5 since it's a D internal. It would be Trump +15 so Trump maybe is one or two points behind Romney

What do you do with McMullin?  They have him tied with Trump. 

[GeorgiaModerate]

Not sure if these will be public polls so posting them here...

PublicPolicyPolling ‏@ppppolls  7m7 minutes ago Raleigh, NC
Just got out of field in 2 Midwestern states we tracked a week ago- margin unchanged in one, Hillary actually doing 2 points better in other

Kind of as a followup to the above: I'm not sure where else to put it; maybe we need a PPP tweets thread?

You must be logged in to read this quote.

[GeorgiaModerate]

gop operative from NH claims great homecoming, fast closing gap
https://mobile.twitter.com/rebeccagberg/status/793855147252285444

But the counterpoint from  a GOP strategist is:

You must be logged in to read this quote.

[GeorgiaModerate]

PPP says their polling has Clinton ahead in NH, for what it's worth.

I wonder if that means she's like a point ahead or solidly ahead.

Very few tweets by PPP in the last five or so days.

They're likely pretty busy with their private clients.

I believe they said not long ago that this election is the busiest they've ever been.

[GeorgiaModerate]

John Harwood

"Trump strategist at last weekend: "MI/PA/WI break our hearts. We win OH/NC/IA; NH/FL toss-up. We win those, LONG night waiting for NV/CO/AZ""

"Clinton strategist on lead entering final weekend: "Solid. Things improved a bit last two days. basically same as we were a week or two ago""

The truth is (as usual) in the middle.

[GeorgiaModerate]

@GlennThrush

Trump wants to win. But two people in his orbit have told me in last week that decisively beating Romney's 206 EVs is increasingly a goal.

Hmm if they're under 206, they're likely behind in Ohio or Arizona.

Matching Romney's 206 is a very plausible outcome.  Swap NC/AZ for IA/OH.  Both pairs add up to 26 EV and cancel each other out.

[GeorgiaModerate]

@GlennThrush

Trump wants to win. But two people in his orbit have told me in last week that decisively beating Romney's 206 EVs is increasingly a goal.

Hmm if they're under 206, they're likely behind in Ohio or Arizona.

Matching Romney's 206 is a very plausible outcome.  Swap NC/AZ for IA/OH.  Both pairs add up to 26 EV and cancel each other out.

Ohio + Iowa = 24 EV.

So it is.  Don't know what I was thinking.  Thanks.

[GeorgiaModerate]

John Harwood ‏@JohnJHarwood  6m6 minutes ago
senior GOP Senate strategist #2: lose IL/WI/NV/PA/NC. Win OH/NH/FL/IN. win MO by an eyelash (but wouldn't surprise me if Kander won). 50-50

Surprising to see them more bearish on NC than MO and NH.

[GeorgiaModerate]

John Harwood ‏@JohnJHarwood
Trump campaign chief Bannon, when I asked for his Election Day results forecast: "We're too busy winning this baby. I don't want to jinx it"

He should have said "we'll have the best results."