• Hey, guest user. Hope you're enjoying NeoGAF! Have you considered registering for an account? Come join us and add your take to the daily discourse.

PoliGAF 2016 |OT11| Well this is exciting

Status
Not open for further replies.
C

Cybit

FGC Waterboy
Iolo said:
His AA support dropped from 19% yesterday to 14% today. If this is just a weird sampling problem, I would expect a drop to 6-8% tomorrow, like last bump.
Click to expand...

The LATimes poll surveying the same people over and over does give kind of a cool look at what at least hints (not definitive obviously) of how wishy washy people are about this freaking election. It seems the sample is skewed Trump for sure, but watching the same people switch from Trump to third party to Clinton back to Trump is...intriguing. I'm curious how representative it will end up being when all is said and done.
 
J

Joeytj

Banned
Ugh, those damn Ipsos polls with Trump +3 and Clinton +3 nationally in another one.

I was feeling so hopeful and then I click on 538 and saw those.

Hope we finally get better polls tomorrow and in the run up to the debate. I don't want to go into the debate with only a C +.07 o +1.0 point position in the polls. I'm going to be a nervous wreck starting Sunday night.
 
A

Aaron Strife

Banned
Joeytj said:
Ugh, those damn Ipsos polls with Trump +3 and Clinton +3 nationally in another one.

I was feeling so hopeful and then I click on 538 and saw those.

Hope we finally get better polls tomorrow and in the run up to the debate. I don't want to go into the debate with only a C +.07 o +1.0 point position in the polls. I'm going to be a nervous wreck starting Sunday night.
Click to expand...
I swear they're going to be like the Gallup of this cycle. 2012 looked a lot closer in the aggregate because while most pollsters were showing Obama up 2-3 points they were shitting up the averages with howlers like Romney leading by 7 points or whatever.

Except Gallup actually has a reputation and pedigree.
 
Z

Zukkoyaki

Member
Joeytj said:
Ugh, those damn Ipsos polls with Trump +3 and Clinton +3 nationally in another one.

I was feeling so hopeful and then I click on 538 and saw those.

Hope we finally get better polls tomorrow and in the run up to the debate. I don't want to go into the debate with only a C +.07 o +1.0 point position in the polls. I'm going to be a nervous wreck starting Sunday night.
Click to expand...
Today was a good polling day for Clinton.

Two +5 Florida polls including one from Monmouth

Two +5 national polls

538 just adjusted the shit out of them because TRENDZ. They also included a bunch of Ipsos polls from the past few days that seemed weighted way greater than a daily poll should probably be.
 
C

Cybit

FGC Waterboy
Zukkoyaki said:
Today was a good polling day for Clinton.

Two +5 Florida polls including one from Monmouth

Two +5 national polls

538 just adjusted the shit out of them because TRENDZ
Click to expand...

Actually Monmouth got adjusted far less this time then in August.

Which, hilariously, their adjusted prediction of Monmouth in August was from +9 to +5.

Funny how that works.

That said, I do think by the end of the week the trends from most polls taken in the last week (9/15-9/19) will show a small tilt towards Clinton - but not enough of them to make up for the hit she took earlier in the month yet. Gotta remember polls are usually a week or so back.
 
Z

Zukkoyaki

Member
Cybit said:
Actually Monmouth got adjusted far less this time then in August.

Which, hilariously, their adjusted prediction of Monmouth in August was from +9 to +5.

Funny how that works.
Click to expand...
False.

At the time Monmouth released that poll it was adjusted to +8 on 538. Their model adjusts every single poll in their entire backlog. To meet their current trend.

Remember the +13 Monmouth national poll from early August? It now reads +9 in their adjusted section. It read as +12 back when they originally input it.

Want proof? Take a mental note of this weeks polls and check back in like 3-4 weeks and see what they say.
 
I

Iolo

Member
Cybit said:
The LATimes poll surveying the same people over and over does give kind of a cool look at what at least hints (not definitive obviously) of how wishy washy people are about this freaking election. It seems the sample is skewed Trump for sure, but watching the same people switch from Trump to third party to Clinton back to Trump is...intriguing. I'm curious how representative it will end up being when all is said and done.
Click to expand...

Ignore the Trump skew or whatever you want to call it. The poll is fine for looking at general trends. But Trump's AA support is literally a square wave. It's a low resolution sample whose value is heavily affected by their weighting multiplier. IMO.
 
S

Spiritreaver

Member
Meh, I try not to judge 538 too harshly. I'm not a statistician nor do I really get the secret sauce of their model. Seeing her drop further in their model despite some decent polling raises eyebrows to me, but I just assume I don't know any better, lol.

I mean I'm surprised that other Reuters tracking polling did so much damage. I figured she'd be up over 60% again, lol.
 
W

Wilsongt

Member
Republicans have such a hard on about getting Obama to say radical Islamic extremist. One day if it were to happen again where a guy starts shooting up a planned parenthood in the name of God or the Democrats condemn the way Christians in Africa treat homosexuals, they just come out and say radical Christian extremists. Just so I can see the collective faces of the GOP turn that bright shade of constipation red.
 
C

Cybit

FGC Waterboy
Zukkoyaki said:
False.

At the time Monmouth released that poll it was adjusted to +8 on 538. Their model adjusts every single poll in their entire backlog.

Remember the +13 Monmouth national poll from early August? It now reads +9 in their adjusted section. It read as +12 back when they originally input it.
Click to expand...

Blargh, I fail at English. You're right, meant to say that it adjusted their August trend and the auto-adjustment hit where Monmouth ended up being now. (I am assuming they are not manually adjusting the data outside of their algorithms because that would be, uh, shady)

Still - assuming the adjustment is automatically generated and not dishonestly manually entered - it appears to have at least predicted the current August to September trends well.

Iolo said:
Ignore the Trump skew or whatever you want to call it. The poll is fine for looking at general trends. But Trump's AA support is literally a square wave. It's a low resolution sample whose value is heavily affected by their weighting multiplier. IMO.
Click to expand...

Yeah, that AA support number has always been like "uhhhhhhhh what?" I am curious as to how the breakdown of the voters are (haven't looked too hard for them, not sure if it is provided). I'd laugh if it is in single digits or something bonkers like that.
 
K

kevin1025

Banned
Nazgul_Hunter said:
CNN is like a Curb Your Enthusiasm episode right now. Jeffrey Lord and whats-his-face trying to defend the "black communities are worse off than they've ever ever ever ever been."
Click to expand...

That guy is stuck in the Reagan era. I'd feel kind of bad for him if it were not for his grotesque love of Trump. The best was when he said "Trump called me earlier and ---" and someone called him out for never talking to Trump. He never even fought it. (This was a couple weeks back.)
 
E

eBay Huckster

Member
i guess the main thing that's getting me hot and bothered about the models is i can at least see what makes PEC tick (and to some extent, the upshot)

meanwhile it's not actually settled whether 538 just has an autoadjustment because if you try to reveal the model's innards, ESPN will murder your family at halftime of MNF
 
C

Cybit

FGC Waterboy
AntraxSuicide said:
Trump will always vex me in the same way that ground game stuff used to. We always think a ground game matters, but you can't know unless you test it by running a campaign with a poor ground game (Trump will finally test this, but even this is hard to pin down since he's bad in other ways).

It could very well be that all these scandals have mattered in that every time they happen, Trump's ceiling is reinforced. Maybe if he shut his mouth and ran normally, he'd be gaining right now (not closing the gap at Clinton's expense, but actually gaining voters himself). We'll never know because it hasn't happened and probably won't happen in a way that matters.

Pretend this is a left-wing candidate running with tons of scandals and constant offensive gaffes. I think the main number we'd all be focusing on (and furious about) would be the ceiling of such a candidate. It's damn near made of steel at this point (the best polls of his campaign could only get him to 45%). He just can't stretch his support beyond that.
Click to expand...

Missed this post, but I think you're right about the bolded part. The 1600 folks have talked about how the polarization is harsh enough that anything above a 5 pt victory would be incredible. I think what his gaffes have done is seen in that he is stuck around 40%, but when Clinton drops 5%...he goes up a percent or so IIRC. It's the inability to take those leaning voters and convert them when he should be able to. Which is pretty good news for Clinton, as her floor might just be flat out higher than his ceiling.
 
I

Iolo

Member
eBay Huckster said:
i guess the main thing that's getting me hot and bothered about the models is i can at least see what makes PEC tick (and to some extent, the upshot)

meanwhile it's not actually settled whether 538 just has an autoadjustment because if you try to reveal the model's innards, ESPN will murder your family at halftime of MNF
Click to expand...

Post #2074
Adjusted Post number #2078 (+4)
(following the increasing post number trendline)
 
O

Owzers

Member
whyamihere said:
PPP just said their new poll has Cooper up by 4 or more.
Click to expand...

Anderson is pretty popular, he's doing well standing up to Lord Trump.

Trump's surrogates are morons. "In Trump's opinion it's true." regarding African American communities being the worst they've ever ever ever been. Anderson's response-ish: If i say in my opinion i can fly it doesn't make it true.
 
C

Cyanity

Banned
Seems like the trend lines in polling have been fairly sinusoidal in nature. If the pattern continues, then Hillary should be up in November. Of course none of this matters if she completely bombs the debates, but Trump has proven to be awful on the debate stage so I don't really see that happening. Fingers crossed.
 
P

Plumbob

Member
Cyanity said:
Seems like the trend lines in polling have been fairly sinusoidal in nature. If the pattern continues, then Hillary should be up in November. Of course none of this matters if she completely bombs the debates, but Trump has proven to be awful on the debate stage so I don't really see that happening. Fingers crossed.
Click to expand...

The last dip was clearly because of Hillary's sickness.

Impossible to say where the race will be a week from now without seeing the debate.
 
G

Grief.exe

Member
Cyanity said:
Seems like the trend lines in polling have been fairly sinusoidal in nature. If the pattern continues, then Hillary should be up in November. Of course none of this matters if she completely bombs the debates, but Trump has proven to be awful on the debate stage so I don't really see that happening. Fingers crossed.
Click to expand...

That's why I was cautioning you last week over freaking out over the polls 'tightening.'
 
R

RustyNails

Member
Plumbob said:
The last dip was clearly because of Hillary's sickness.

Impossible to say where the race will be a week from now without seeing the debate.
Click to expand...
I think it was because of that Deplorables line. Cohn tweeted that the last two days of polling during the controversy showed an uptick.
 
I

Iolo

Member
So I guess the RAND poll is not continuous this year, unlike in 2012, but is instead conducted in waves. We should get some new data just after the first debate, and then after the last one. USC Daybreak replaced RAND in the continuous poll slot, though their methodology differs.
 
A

AntraxSuicide

Banned
Cybit said:
But let's take a poll from last week, say, one from a very reputable source

http://www.neogaf.com/forum/showthread.php?p=216864805&highlight=selzer#post216864805

Start at post 15319 and about how the LV screen is off. You know, from the person who had this written about her

http://fivethirtyeight.com/features/selzer/

and even points out an illustrative moment from 2014



You can continue reading how quickly armchair forum goers know more about the LV screen at this point in the campaign than Ann Selzer. (Also, 15549 is where I put my two cents in).

You can go back and look at previous OTs and see this phenomenon happen again and again.
Click to expand...

But this is the point of outliers, and also my point about treating stats as a religion instead of a science. Ann Selzer herself will tell you that she'll produce outliers. It's just how statistics works that everyone gets some outliers. Nothing in that post you quoted is off the mark. A person (and I suspect Selzer herself!) is very right to be suspicious about sub-04 turnout since the trendline (Nate's own favored metric) for that over the last two decades has been an increase in minority turnout.

Cybit said:
Missed this post, but I think you're right about the bolded part. The 1600 folks have talked about how the polarization is harsh enough that anything above a 5 pt victory would be incredible. I think what his gaffes have done is seen in that he is stuck around 40%, but when Clinton drops 5%...he goes up a percent or so IIRC. It's the inability to take those leaning voters and convert them when he should be able to. Which is pretty good news for Clinton, as her floor might just be flat out higher than his ceiling.
Click to expand...

I think the last line is probably pretty true. It's going to take an insane scandal from Clinton plus some great weeks by Trump to get her floor to drop below his ceiling. His only shot would be to tempt the right-leaning Johnson voters to join his camp while simultaneously convincing the left-leaning Johnson voters to stay with Johnson. The smart play would be to correctly identify which camps would rather die than vote for you, to be blunt. No alt-right rural types are voting Clinton, and no minorities are voting Trump. The best plan for both is to drive up their base's turnout, and for Clinton, get as many NeverTrump people to either vote Johnson or vote for her in private.

Trump could technically do the same, but I cannot see how his campaign could ever hope to organize in time to do such political maneuvering. They simply don't have the structure for it. Their strength is in free coverage and the idea that since you're already at rock bottom, nothing hurts you so go all out. If something doesn't work, you're already at your floor, and if it does work, you hope you get enough mileage out of it to matter* in the end.

*And on this note, what constitutes "mattering" in this scenario depends on the two campaign factions I mentioned earlier. A good number of Trump's people don't actually care if he wins (they'd like it, but they also know it isn't happening). They're after the down-ticket and the party's public perception. I don't know if I can recall a campaign where the candidate is being helped by people that probably actively hate him, and honestly would love to see his demise as long as they kept other seats in Congress.
 
H

Holmes

Member
Aaron Strife said:
Willing to bet it's a Dem sweep in NC for statewide elections, Ross, Cooper and Clinton.
Click to expand...
I think so too, but I'm willing to bet Cooper has the largest margin of victory considering he appeals to a certain demographic that the other two don't because he's not a left-wing woman.
 
E

eBay Huckster

Member
Holmes said:
I think so too, but I'm willing to bet Cooper has the largest margin of victory considering he appeals to a certain demographic that the other two don't because he's not a left-wing woman.
Click to expand...

yeah, I can't remember if someone said it here or in the Discord, but the margin's probably gonna be Cooper >> Clinton >= Ross in that order
 
Status
Not open for further replies.

Similar threads

Show us your gaming setup: 2024 Edition
2 3 4
197
29K
Rockondevil replied
  • Locked
Show us your gaming setup: 2023 Edition | 20th anniversary thread
2 3 4 5 6
268
56K
DGrayson replied
  • Locked
PoliGAF 2017 |OT6| Made this thread during Harvey because the ratings would be higher
338 339 340 341 342
17K
2M
UselessNapkin replied
  • Locked
PoliGAF 2017 |OT5| The Man In the High Chair
395 396 397 398 399
20K
2M
RDreamer replied
  • Locked
PoliGAF 2017 |OT4| The leaks are coming from inside the white house
396 397 398 399 400
20K
2M
ShironRedshift replied
Top Bottom