New poll hype thread

[amdcpus]

Template: https://docs.google.com/spreadsheets/d/10semu5xRZ-D6gG7J3d19asXVMnoCvh7G2RNuySOv_RI/edit?usp=sharing

Template with my Utah Poll: https://docs.google.com/spreadsheets/d/1bWqN7yncLyttlivU1hc9FX7xN8jZ4o5MguFkPOG4bBk/edit?usp=sharing

Demographics/Turnout Data used: https://docs.google.com/spreadsheets/d/1qsWmPyoLF7q30MoULktph9TjKyqDs-dwpC8Yb5gMInI/edit?usp=sharing

[cinyc]

Template: https://docs.google.com/spreadsheets/d/10semu5xRZ-D6gG7J3d19asXVMnoCvh7G2RNuySOv_RI/edit?usp=sharing

Template with my Utah Poll: https://docs.google.com/spreadsheets/d/1bWqN7yncLyttlivU1hc9FX7xN8jZ4o5MguFkPOG4bBk/edit?usp=sharing

Demographics/Turnout Data used: https://docs.google.com/spreadsheets/d/1qsWmPyoLF7q30MoULktph9TjKyqDs-dwpC8Yb5gMInI/edit?usp=sharing

Thanks!  It looks like you are using the same census data I wanted to use.

[amdcpus]

Template: https://docs.google.com/spreadsheets/d/10semu5xRZ-D6gG7J3d19asXVMnoCvh7G2RNuySOv_RI/edit?usp=sharing

Template with my Utah Poll: https://docs.google.com/spreadsheets/d/1bWqN7yncLyttlivU1hc9FX7xN8jZ4o5MguFkPOG4bBk/edit?usp=sharing

Demographics/Turnout Data used: https://docs.google.com/spreadsheets/d/1qsWmPyoLF7q30MoULktph9TjKyqDs-dwpC8Yb5gMInI/edit?usp=sharing

Thanks!  It looks like you are using the same census data I wanted to use.

No problem. Let me know if you have any questions. I'd also be interested in seeing the raw data too if possible.

[cinyc]

No problem. Let me know if you have any questions. I'd also be interested in seeing the raw data too if possible.

I'll probably provide a link to the raw data for the survey when it's done.  So far, it has 122 of 500 respondents, so I expect it to be complete Monday or Tuesday - though new respondents have slowed to a crawl in the past few hours.  Some of the cross-tabs look a bit wonky (others look pretty good), but that's to be expected with small sample sizes.

As you know, Google calculates weighted results that are supposed to represent a representative sample of the Internet population based on at least age.  Anyone care to guess the South Dakota results so far - or predict the final results?

[amdcpus]

No problem. Let me know if you have any questions. I'd also be interested in seeing the raw data too if possible.

I'll probably provide a link to the raw data for the survey when it's done.  So far, it has 122 of 500 respondents, so I expect it to be complete Monday or Tuesday - though new respondents have slowed to a crawl in the past few hours.  Some of the cross-tabs look a bit wonky (others look pretty good), but that's to be expected with small sample sizes.

As you know, Google calculates weighted results that are supposed to represent a representative sample of the Internet population based on at least age.  Anyone care to guess the South Dakota results so far - or predict the final results?

Google's weighting is utter crap and for some reason they don't even apply most of the time. Weighted correctly, I expect it to be something like

Trump 42

Clinton 30

Johnson 12

Stein 2

Could you PM the current results? You just have to click "share" and change it to whoever has the link, like this: https://www.google.com/insights/consumersurveys/view?survey=ph4cdzact22bfkpwb34ts46io4&question=1&filter=gen%3AMale%2C%2Cage%3A18-24

[cinyc]

Google's weighting is utter crap and for some reason they don't even apply most of the time. Weighted correctly, I expect it to be something like

Trump 42

Clinton 30

Johnson 12

Stein 2

Could you PM the current results? You just have to click "share" and change it to whoever has the link, like this: https://www.google.com/insights/consumersurveys/view?survey=ph4cdzact22bfkpwb34ts46io4&question=1&filter=gen%3AMale%2C%2Cage%3A18-24

Google's weights are more like a sample of "adults" or "Internet users" than registered voters.  Best I can tell, they skew young, which helps a certain 3rd party candidate.

I didn't poll Stein because she's not on the ballot.  I polled Castle because he is.  I also didn't include a don't know/undecided or write-in option because, as I understand how GCS works, undecideds can just skip the question, and few people will actually write-in on election day.  I'd rather have no undecided option unless I could follow up and push the undecideds.  That's too expensive.

Anyway, I think I've hijacked this thread enough with technical mumbo jumbo.  I probably should go back to predictions only.

Check your PMs.  Please PM me with any other technical stuff.

[amdcpus]

Google's weighting is utter crap and for some reason they don't even apply most of the time. Weighted correctly, I expect it to be something like

Trump 42

Clinton 30

Johnson 12

Stein 2

Could you PM the current results? You just have to click "share" and change it to whoever has the link, like this: https://www.google.com/insights/consumersurveys/view?survey=ph4cdzact22bfkpwb34ts46io4&question=1&filter=gen%3AMale%2C%2Cage%3A18-24

Google's weights are more like a sample of "adults" or "Internet users" than registered voters.  Best I can tell, they skew young, which helps a certain 3rd party candidate.

I didn't poll Stein because she's not on the ballot.  I polled Castle because he is.  I also didn't include a don't know/undecided or write-in option because, as I understand how GCS works, undecideds can just skip the question, and few people will actually write-in on election day.  I'd rather have no undecided option unless I could follow up and push the undecideds.  That's too expensive.

Anyway, I think I've hijacked this thread enough with technical mumbo jumbo.  I probably should go back to predictions only.

Check your PMs.  Please PM me with any other technical stuff.

Thanks for the info, and yea I meant Castle, not Stein. If you started the poll yesterday, it should be finished around this time tomorrow. Feel free to ask me any questions about the template when you go to weight it.

[HillOfANight]

I usually only see Google Consumer Surveys on annoying local newspaper websites, so I wonder if that skews the results somewhat.

https://datastudio.google.com/#/org//reporting/0B29GVb5ISrT0TGk1TW5tVF9Ed2M/page/YgS
Also, with Google doing these already (though not as frequent)...

[cinyc]

I usually only see Google Consumer Surveys on annoying local newspaper websites, so I wonder if that skews the results somewhat.

https://datastudio.google.com/#/org//reporting/0B29GVb5ISrT0TGk1TW5tVF9Ed2M/page/YgS
Also, with Google doing these already (though not as frequent)...

GCS gets its poll data from a mix of newspaper websites, other websites and their mobile app.  The current mix in my poll is Mobile App: 52.6%, News: 42.9%, Other: 4.5%.  amdcpus' Utah poll was News: 53.5%, Mobile App: 16.6%, Reference: 4.7%, Other: 25.2%.  That may or may not skew the results.  GCS gives you imputed demographic and some geographic data, so you can weigh it however you wish.

The problem with the polls Google is doing already in small states like SD is that the sample size is tiny (50 voters in SD).  My poll will have a larger sample size - probably around 420 or so when the "I'm not registered in SD" respondents and respondents for which there is no age demo are filtered out.

[dspNY]

So a lot coming up Sunday and Monday

YouGov national
CBS/YouGov states (NV, UT)
NBC/SurveyMonkey national
NBC/WSJ/Marist national (Monday)
Monmouth national (Monday)

CNN has also been long overdue so maybe we hear from them soon

[Xing]

Here are my guesses for the state polls:

NV: Clinton +3
UT: Trump +5

[Tender Branson]

UT is hard to tell, because YouGov generally has a lower Johnson number than other pollsters and McMullings is hard to predict too ...

UT: Trump 34, Clinton 27, McMullings 20, Johnson 8, Stein 2
NV: Clinton 43, Trump 42, Johnson 6

[ApatheticAustrian]

so...there is a new yougov national poll coming today or tomorrow?

[ProudModerate2]

so...there is a new yougov national poll coming today or tomorrow?

Are you asking or saying ?

[ApatheticAustrian]

Are you asking or saying ?

asking - referring to:
https://uselectionatlas.org/FORUM/index.php?topic=244642.msg5320287#msg5320287

[Gass3268]

Are you asking or saying ?

asking - referring to:
https://uselectionatlas.org/FORUM/index.php?topic=244642.msg5320287#msg5320287

NBC/Survey Monkey, Politico/Morning Consult, and YouGov will all have their weekly trackers early this week.

[An American Tail: Fubart Goes West]

I'd love to see some Idaho polls to see if McMullin is cracking 10% up there. Is Dan Jones or a national pollster planning on doing any?

[elcorazon]

NV Clinton +2
Utah Trump + 9

[cinyc]

I'd love to see some Idaho polls to see if McMullin is cracking 10% up there. Is Dan Jones or a national pollster planning on doing any?

I don't know, but you can always buy your own Google Consumer Survey.  If you click on amdcpus' Utah poll link, you might be offered $50 off your first survey, bringing the total cost of a one-question poll to $25 for 500 respondents.

My GCS South Dakota poll is 50% complete.  250 respondents down.  250 more to go.  It will likely be complete tomorrow.

[Ronnie]

My predictions for the Monmouth and Marist national polls tomorrow:

Marist: Clinton +9
Monmouth: Clinton +7

[ApatheticAustrian]

i go for +7 and +5.

including drama dramma drama.

[PresidentSamTilden]

marist +10
MM + 5

[voter1993]

I'd like to see a trump comeback in these polls.. but i can see some surprises tomorrow.

[Fmr President & Senator Polnut]

I'd like to see a trump comeback in these polls.. but i can see some surprises tomorrow.

On what grounds?

I expect we'll see something around that 6-8 lead with MoE playing funny-buggers around the edges.

[voter1993]

Ya i agree could see a little higher in one of the polls too. A dead heat makes a more exciting election just like 2012