2016 Postmortem
Related: About this forumTrust the pollsters
Too often in this forum, I see people bashing pollsters in the same way that the right wing bashes the "liberal media" and the "intellectual elite." Whether we're talking about Gallup, Rasmussen, PPP, or the analysis of Nate Silver or the National Journal, we are talking about people who are trying to get it right. They may want Obama, they may want Romney, but they are professional, educated, intellectual people and organizations that try to be fair. Neither Gallup nor Silver are changing their methodology to accommodate their political beliefs. They have a methodology; they think it makes sense; and they let the cards fall where the cards fall.
You can argue that Newsmax/Zogby puts Obama ahead in Florida to get him to waste money or that Gallup gives Rmoney a large lead because they want to discourage Democrats, but that simply isn't true. They have no motivation to commit that kind of fraud. Those who design the polls take pride in their knowledge of statistics and their ability to model the electorate. Those who hired the modelers want to have a good reputation so their polling firms gets hired by private business later to help with their marketing efforts.
The reality is that none of the pollsters are dishonest. Most of them will come pretty close to getting it right, but in their day-to-day sampling of anywhere from a few hundred to a few thousand people, they will run into anomalies. Chance alone will cause them to reach households where the people lie about their intentions, where they run into a disproportionate number of Republicans for Obama or vice-versa, or where people like about demographic information. Some polls will be outlier, and it's no one's fault. In my opinion, places like RealClearPolitics, HuffPost pollster, and Nate Silver (who I like best) have the best shot at getting it right because they have analyzed the models of the different pollsters and are aggregating those results in a logical way.
Some pollsters may be wrong, but the fact is with just about every one, if they are right, then Obama wins the election. They are honest people, like journalists and intellectuals, and deserve to be trusted. We can be disappointed in their results, examine their methodology and raise valid questions, but to believe they have some ulterior manipulative motive strikes me as just the kind of thing we hear from the right when it comes to journalists, professors, and writers.
dchill
(38,527 posts)"The reality is that none of the pollsters are dishonest."
I call you on that.
Maximumnegro
(1,134 posts)and the post lost any credibility then and there.
dchill
(38,527 posts)Haven't been on the same planet for - well, forever.
BainsBane
(53,056 posts)I just go there for recent poll numbers. I like that they link to the actual poll. They include some polls and not others, but I don't worry about that. It's just their polling average. I just go there for the basic rundown and links. I don't see any reason to get aggravated about it.
Onlooker
(5,636 posts)They are a marketing pollster attempting to do political polls. They may not have the qualifications, but I doubt they are simply making shit up. They had Obama up by one in Florida, which aligns with other polls, and show the same results for Ohio as the Democratic-leaning PPP. Gary Johnson scoring so high in OH certainly makes their poll an outlier, but every polling firm has outliers. Gravis Marketing does not seem like a high powered political pollster, but is instead applying traditional marketing polling techniques to the electorate.
You should tell that to Nate Silver.
http://www.democraticunderground.com/10021579317
krawhitham
(4,647 posts)dchill
(38,527 posts)Pulled out of butt numbers.
graham4anything
(11,464 posts)and yes, at the very end, their polls will be accurate, or when people are watching them, they are accurate.
all these new rightwing pollfirms started only after ras was outed as bad.
then all of a sudden, on some days ras is correct now.
but they are just 100% propaganda
Yes it does exist.
and yes, EVERYONE has an ulterior motive-their candidate winning.
just some people want to do it honestly, and that lets out the republicans.
DemKittyNC
(743 posts)Trust me, if you can buy candidates you can easily buy a poll. To think otherwise is very naive.
Welcome_hubby
(312 posts)Are you a psychologist? Did you predict Research 2000 and Strategic Vision would make numbers up before they did?
Onlooker
(5,636 posts)But, that wasn't my point. My point is simply that the polls being debated are driven by people more interested in getting it right than in manipulating the election. And I'm disappointed at the anti-intellectual mentality in this forum, which in my opinion parallels the anti journalist and anti-intellectual mentality of the right, where they assume that anyone who reports or writes something they don't agree with is simply elitist or is trying to manipulate the public.
CJCRANE
(18,184 posts)This OP is an example of the rhetorical technique I call an "appeal to pure motive":
"{P}ollsters...are honest people, like journalists and intellectuals, and deserve to be trusted". That assertion is absurd.
Since when should we trust all journalists and intellectuals without reservation? Have you missed the last decade of discussions on DU?
If you're going to trust, you should "trust but verify" at the very least. Otherwise you come across as incredibly naive.
bemildred
(90,061 posts)What is this, come to Jesus for pollsters?
Are you out of your mind?
Have you never heard of push-polls and tailoring polls?
Most polls are commercial operations, and it shows.
Onlooker
(5,636 posts)Push polls are done by campaign firms, and no serious website that attempts to analyze the election is including the results of push polls.
Yes, polls are commercial operations, and those who do best are most likely to get work for corporations and others who require polling.
bemildred
(90,061 posts)We only need to talk about the ones that you trust?
And you think we should all be gullible like you?
Are we hurting the pollster's feelings?
The only ones I pay the slightest attention to are the academic ones, and they don't need to be defended, because they are not trying to brainwash anybody, and their methodologies do not suck.
liberal N proud
(60,340 posts)Ghost of Tom Joad
(1,356 posts)began to 'tighten up' not long after republicans began to complain about them.
Frumious B
(312 posts)I think the majority of pollsters do an honest job, but it's silly to say there are no bad apples. I also suspect that some use some "secret sauce" via weighting, "likely voter" models and such to sexy up the numbers a little, not necessarily out of bias toward a particular candidate or party, but in order to get attention or help drive a particular narrative.
Just for example, the NBC poll yesterday was tied at 47% with "likely voters" with President Obama up 49-44% with "registered voters." I looked through the cross tabs and they didn't reveal the exact specifics of how someone gets classified as a "likely voter". Obviously the folks on the right will crawl over broken glass to try and vote Obama out of office. I would have crawled over broken glass to vote Bush out in 2004. I'll crawl over broken glass to vote to keep Obama IN office this year even though I live in Georgia, which is dark red on every map you care to check. It's hard for me to believe that in a Presidential year, with the Obama campaign's skills at getting out the vote and tapping into non-traditional "likely" voters that the actual turnout won't be at least somewhat more favorable to Obama than the picture that the NBC poll paints. If that happens then Obama wins. I also think he wins a near tie in the national popular vote due to having a slightly more favorable Electoral map. Maybe that's just wishful thinking on my part. We'll see.
DemocratSinceBirth
(99,711 posts)We know there have been pollsters in the past who have just made up numbers and didn't do any actual polling. We also know that a lot of mathematical assumptions go into the building of a poll model.
That's why it is important to look at poll averages.