Hispanics backing GOP in Georgia

[JRP1994]

http://www.11alive.com/story/news/politics/2014/09/29/politics-georgia-carter-deal-perdue-nunn-hispanic-voters/16446329/

[Dixie Reborn]

lol

[Lief 🗽]

> making conclusions about voting patterns from a subsample of 40 people

[Eraserhead]

That article was so dumb it gave me a headache.

[IceSpear]

> making conclusions about voting patterns from a subsample of 40 people

[KCDem]

> making conclusions about voting patterns from a subsample of 40 people

"Journalists" are idiots these days.

[Bacon King]

It's a well known fact that almost no pollster can ever get their Hispanic subsample anywhere near accurate. Does anyone know if that one organization that only surveyed Hispanics in 2012 is doing that again this year?

[Adam Griffin]

Yeah, the samples are always too small to make any reasonable estimate from, and it's likely that Republican-leaning Latinos are going to be easier to contact than Democratic-leaning ones, anyway. Though it has always perplexed me how just about every poll shows Latinos either evenly-split or leaning GOP in the state.

I thought I posted this somewhere a while back, but this was something I did with NGPVAN access for three congressional districts in Georgia. Basically, I was able to assess 700,000 registered voters' (who voted in 2012) individual likelihoods of voting D/R, and then take the aggregate to produce these results. Think of it as instant polling, without all the fuss and cost. It's my experience that VAN data can be off considerably on one or a few people, but when you're dealing with hundreds of thousands of people and averaging them all out, it's phenomenally accurate. Anyway, this is what it showed for Latinos in the presidential election:

[Adam Griffin]

Relevant.

http://www.11alive.com/story/news/politics/2014/10/10/georgia-politics-poll/17054641/

You must be logged in to read this quote.

[eric82oslo]

It's a well known fact that almost no pollster can ever get their Hispanic subsample anywhere near accurate. Does anyone know if that one organization that only surveyed Hispanics in 2012 is doing that again this year?

Latino Decisions you mean?

Their polling page: http://www.latinodecisions.com/recent-polls/

They have already polled Colorado, Florida, New Mexico, Arizona and Texas so far this year.

[Fmr President & Senator Polnut]

It should be noted that nationally, English-language only polling led to under-polling Obama by about 10% among Latinos.

[Bigby]

Yeah, the samples are always too small to make any reasonable estimate from, and it's likely that Republican-leaning Latinos are going to be easier to contact than Democratic-leaning ones, anyway. Though it has always perplexed me how just about every poll shows Latinos either evenly-split or leaning GOP in the state.

I thought I posted this somewhere a while back, but this was something I did with NGPVAN access for three congressional districts in Georgia. Basically, I was able to assess 700,000 registered voters' (who voted in 2012) individual likelihoods of voting D/R, and then take the aggregate to produce these results. Think of it as instant polling, without all the fuss and cost. It's my experience that VAN data can be off considerably on one or a few people, but when you're dealing with hundreds of thousands of people and averaging them all out, it's phenomenally accurate. Anyway, this is what it showed for Latinos in the presidential election:

Interesting how Asians supported Romney in northern Georgia despite being mostly Democrat nationally. Granted, they didn't go for Romney as much as whites, but he still had a lead. Part of me is starting to wonder if Asian voters mimic White voters, and the Democratic dominance in the Asian voting base is because most of them live in the Pacific region. Then again, I don't feel comfortable making that kind of call without enough evidence.