2016 Postmortem
Related: About this forumGravis Marketing Has Willard Romney Winning The African American Vote In CO 60%-40%
http://www.realclearpolitics.com/docs/2012/Gravis_CO_1006.pdfPlease go to page eleven.
If you "unskew that poll" to the more likely 90% -10% margin the president will enjoy among African Americans in that state the poll results would show the president winning by a percent or two instead of trailing by three.
If I was conducting a survey and discovered such counterintuitive results I would surely check my sample.
Jennicut
(25,415 posts)Sigh. I hate when all polling firms fudge their #'s. I like to look at every poll in order to decide where the race is going and I like to average them all together. I even look at Rasmussen though I know they robo poll and are R-leaning. This is so bogus and so blatant that it makes me more sad then angry. Doug Kaplan is an asshole.
DemocratSinceBirth
(99,710 posts)According to Gravis Marketing, Obama is winning Latinos in the state of Colorado by 2-1 and losing among African Americans 2-1.
Does that make sense?
fugop
(1,828 posts)Some really weird stuff in the crosstabs. I hate how these flash-in-the-pan pollsters get picked up to drive the narrative. It's ridiculous. I am so disgusted with how our media reports polls as if they're all equal, all the same. Just as they compare, for example, one month's poll with a Dem-skewing sample to the next month with a complete switch to a GOP-leaning sample and then trumpet how the poll went from Dem to GOP.
Just. So. Sick. Of. All of them.
DemocratSinceBirth
(99,710 posts).
Drunken Irishman
(34,857 posts)The math suggests blacks made up about 75 voters in this poll. That means 30 said they're voting Obama and 45 said Mitt Romney. With such a low margin, you're going to get funky math like that ... especially when you account for the fact A) margin of error B) maybe even people lying to the pollsters.
It's why you should never take one poll as gospel. If other polls next week, including PPP, show Colorado shifting to Romney, then we should be concerned. But yeah, right now, this poll is flawed.
DemocratSinceBirth
(99,710 posts)Than His Fellow White Coloradan.
DemocratSinceBirth
(99,710 posts)You are on to something. It's an automated poll. Perhaps white Romney supporters are pretending to be black to make it appear Obama is doing poorly among blacks,ergo:
David MarcusBerkeley, Ca
"One of those "strong swing state polls for Romney" is in Colorado: http://www.realclearpolitics.com/docs/2012/Gravis_CO_1006.pdf. But looking at p. 11 of that poll, it shows Romney winning the black vote by 58-40. If the actual black vote were 80-20 in favor of Obama, and everything else in the Gravis poll were exactly the same, the results would change from a 3.49% margin in favor of Romney to a 1.07% margin in favor of Obama.
I've observed the same very counter-intuitive results for black voters in Gravis polls of FL, NC, and MI in the last month. Worthy of some 538 review and analysis? "
Wow. How messed up are Republicans if they are doing that.They are sick, sick, sick...
titaniumsalute
(4,742 posts)Even if a human was interviewing if the respondent says they are black they are logged as black. Just because someone may or may not "sound" black is not decided by the interviewer. Now there may be some validity that if a human was interviewing the respondents people would be less likely to lie for fear of getting caught.
DemocratSinceBirth
(99,710 posts)I have seen (R)asmussen's internals on another site. Even when Obama is doing well there is a gender gap but the other way. He leads among men and trails among women. I believe some yahoos just press the wrong button to mess with the robocaller's results.
anneboleyn
(5,611 posts)manipulating results in such obvious ways -- just to drive a narrative and/or make it appear that there is more support for Romney (black voters 60% to 40% in favor of Romney, and these "very counter-intuitive results" are happening in multiple swing states?? That is ridiculous).
Doctor Jack
(3,072 posts)I think we need to wait until some more credible pollsters get to work before we can conclude what is happening. So far we have complete crap, like Rasmussen, We Ask America, or whatever it is called, and now this.
DemocratSinceBirth
(99,710 posts)A poster at 538 did.
I smell a rat..Talking about rats it seems indisputable Willard got a bounce. But Republicans pollsters are manipulating poll results to inflate the bounce and create a narrative.
Doctor Jack
(3,072 posts)Win by any means necessary
BlancheSplanchnik
(20,219 posts)DavidDvorkin
(19,473 posts)geek tragedy
(68,868 posts)TroyD
(4,551 posts)As is Rasmussen, We Ask America, Purple Strategies etc.
So what? What we think doesn't matter - it's what the public thinks and how this affects the narrative.
The point is that we had the lead in all these polls and the momentum, and Obama may have thrown that away in the debate.
Time to stop taking things for granted.
Now it's time to hit back at Romney by exposing him as a pathological liar as Ted Kennedy did in 1994, and by making sure the September Jobs Report becomes BIG NEWS and overshadows Romney's lies.
"We KNOW Gravis is untrustworthy"
...that's the case, why the hell are people, including Nate Silver, giving these polls credibility? Doesn't he at least owe an explanation in his commentary? At least offer a caveat. A bullshit poll that shows Romney winning the black vote 60 to 40 should be flagged.
Here's Nate:
http://fivethirtyeight.blogs.nytimes.com/2012/10/06/oct-6-romney-maintains-poll-momentum/
The poll is utter bullshit. Still, if they want to take these absurd polls seriously an increase Republican delusion, so be it.
"Fair and balanced" is creating the impression that psychos rule this country. You see, Republicans are complaining about the polls so every poll has to be given merit.
Lying is Genius. Greed is good. Vague is brilliant. Cheating is American.
mzmolly
(50,985 posts)dsc
(52,155 posts)and lets assume they are no more or less likely than anyone else to vote so they make up 4.3% of the electorate as well. He is losing 40 to 58 according to that sample among Colorado blacks. That translates to 1.72% to 2.49% or a -.77%. Assuming that he really wins 90 to 10. That would be 3.87% to 0.43% which is + 3.44 or a total difference of 4.21% in margin. Instead of leading by 3.5% as Romney is in this poll he should be losing by 0.71% which is not exactly a major difference given the MOE.
TroyD
(4,551 posts)That's what I don't get.
Why would they do it this way?
Why not show a huge increase for Romney amongst some other demographic?
No way is Romney going to be winning over African American voters. I don't have a single AA friend that believes that.