New Pew Survey - American Voters

[Cheerful Charlie]

http://www.people-press.org/2016/09...election-two-coalitions-moving-further-apart/
[h=1]The Parties on the Eve of the 2016 Election: Two Coalitions, Moving Further Apart[/h]
---

A large study of American voters, Democrats and Republicans. Probably more than you want to know about the demographics for this election cycle. Short version. Democratic voters are better educated, younger, less religious and more ethically diverse.

 

[prideandfall]

Democratic voters are better educated, younger, less religious and more ethically diverse.

short version: welcome to the last 40 years, glad you could make it? heh.

 

[untermensche]

Hey, Republicans have moved from 93% white to 86%.

We are making progress.

 

[ronburgundy]

Hey, Republicans have moved from 93% white to 86%.

We are making progress.

Not really as rosy a picture as it appears. The only group that has shifted more toward the Republicans in the last 8 years are whites, and by a large % from 46% in 2008 to 54% now. Asians have moved away from Republicans, while Hispanics and Blacks support the GOP at their same low rates.
So, the increase in Republicans that are non-white is an artifact of Hispanics being an overall larger % of the total population, which includes Cubans who typically go Republican and strongly religious Hispanics for whom promoting sexism, homophobia, and theocracy are their driving issues.

 

[untermensche]

Hey, Republicans have moved from 93% white to 86%.

We are making progress.

Not really as rosy a picture as it appears. The only group that has shifted more toward the Republicans in the last 8 years are whites, and by a large % from 46% in 2008 to 54% now. Asians have moved away from Republicans, while Hispanics and Blacks support the GOP at their same low rates.
So, the increase in Republicans that are non-white is an artifact of Hispanics being an overall larger % of the total population, which includes Cubans who typically go Republican and strongly religious Hispanics for whom promoting sexism, homophobia, and theocracy are their driving issues.

I wasn't saying it is rosy.

Sarcasm doesn't translate well in print.

 

[Horatio Parker]

I doubt Cubans have much to do with that number. Cuba is the whitest country in the Caribbean, and I'll bet Cuban Trump supporters overwhelmingly self identify as white.

 

[ronburgundy]

I doubt Cubans have much to do with that number. Cuba is the whitest country in the Caribbean, and I'll bet Cuban Trump supporters overwhelmingly self identify as white.

The results are not merely about Trump, but about identification with the GOP more generally. 27% of self-identifies "Hispanics" identify or lean toward the GOP now as they did in 2008, and Cubans, especially those in Florida tend to go mostly in favor of the GOP. They likely comprise a large portion of that 27%.

 

[Horatio Parker]

I doubt Cubans have much to do with that number. Cuba is the whitest country in the Caribbean, and I'll bet Cuban Trump supporters overwhelmingly self identify as white.

The results are not merely about Trump, but about identification with the GOP more generally. 27% of self-identifies "Hispanics" identify or lean toward the GOP now as they did in 2008, and Cubans, especially those in Florida tend to go mostly in favor of the GOP. They likely comprise a large portion of that 27%.

The problem is when you conflate Hispanic with non-white. Hispanic is not a race. Many Cubans identify as white hispanic.

 

[ronburgundy]

The results are not merely about Trump, but about identification with the GOP more generally. 27% of self-identifies "Hispanics" identify or lean toward the GOP now as they did in 2008, and Cubans, especially those in Florida tend to go mostly in favor of the GOP. They likely comprise a large portion of that 27%.

The problem is when you conflate Hispanic with non-white. Hispanic is not a race. Many Cubans identify as white hispanic.

I think that "white hispanic" is counted as "hispanic" in the OP poll. "White" typically is limited to only "white non-hispanic".

So, that still supports Cubans being the majority of the 27% "Hispanics" in the data who support the GOP, and since their numbers are growing relative to non-hispanic "whites", the GOP appears to be getting less "white", but no in any positive way that reflects the GOP is any less bigoted (they are moreso).

 

[Horatio Parker]

The problem is when you conflate Hispanic with non-white. Hispanic is not a race. Many Cubans identify as white hispanic.

I think that "white hispanic" is counted as "hispanic" in the OP poll. "White" typically is limited to only "white non-hispanic".

So, that still supports Cubans being the majority of the 27% "Hispanics" in the data who support the GOP, and since their numbers are growing relative to non-hispanic "whites", the GOP appears to be getting less "white", but no in any positive way that reflects the GOP is any less bigoted (they are moreso).

Yes.

I'm just pointing out the distinction.

I doubt many Cuban Americans would agree they're racially diversifying the GOP.

 

[Loren Pechtel]

I wonder how accurate polls are these days.

The sheer number of polls this election cycle has no doubt driven a lot of people to not answer.

 

[Elixir]

I wonder how accurate polls are these days.

The sheer number of polls this election cycle has no doubt driven a lot of people to not answer.

How do the pollsters ensure a representative cross-section, when most millennials only have cell phones, everyone has caller ID, and they can only poll people who are bored enough to answer phone calls from unknown callers?

 

[whichphilosophy]

I wonder how accurate polls are these days.

The sheer number of polls this election cycle has no doubt driven a lot of people to not answer.

The election results can be the best test of opinion polls.

 

[ronburgundy]

I wonder how accurate polls are these days.

The sheer number of polls this election cycle has no doubt driven a lot of people to not answer.

How do the pollsters ensure a representative cross-section, when most millennials only have cell phones, everyone has caller ID, and they can only poll people who are bored enough to answer phone calls from unknown callers?

Good points.

PEW apparently is up to 75% of its samples being via cell phones. They call a number at least 7 times at varied days and times to try and get a response. Still, response rates are down to about 10% due to ID, screening, etc..

Low response rate doesn't always mean a non-representative sample, unless the views being asked about are correlated in some way to the willingness to answer.
They have been doing various studies to assess how representative their samples still are. One way is comparing current polls on things that would not be expected to change in the last 20 years with polls from 20 years ago when response rates were higher. If they get a similar result of opinion distribution, it is evidence the low current response rate is still largely capturing a random sample of people.

Here is what PEW has to say.