Gallup Gave Up. Here’s Why That Sucks.

[Algernon Moncrieff]

LINK to full article

We’re worse off without Gallup

Gallup uses rigorous polling methodologies. It employs live interviewers; it calls a lot of cell phones; it calls back people who are harder to reach. More than that, it took the criticism it received after the 2012 election seriously, even bringing in outside help to figure out what went wrong. Gallup rates as solidly average in FiveThirtyEight’s pollster ratings in large part because of those techniques. It’s had two bad elections recently, but it’s never a good idea to judge a pollster on just a couple of election cycles; Gallup has also had good years.

Polling consumers are far better off in a world of Gallup’s than in a world of Zogby Internet polls and fly-by-night surveys from pollsters we’ve never heard of. There is plenty of shadiness in the polling community, and Gallup seemed to be opening its doors.

Gallup says it will still conduct issue polling, but here’s the problem: Elections are one of the few ways to judge a pollster’s accuracy. And that accuracy is important: We use polls for all kinds of things beyond elections. How do Americans feel about the economy? Do elected leaders have the trust of the public? Is there support for striking a deal with Iran? By forgoing horse-race polls, Gallup has taken away a tool to judge its results publicly.

Still, Newport told me that there are other ways to check the accuracy of Gallup’s polls, including by comparing its results to government surveys. “We were able to track after the Affordable Care Act went into effect the drop in the uninsured, which turned out to be very close not only to other polls but the government polling when it finally came out,” he said. Newport also suggested that Gallup might conduct horse-race election polls without publishing the results.

[Fred Sanders]

1. What if polls were banned, say two weeks before an election? Then maybe the issues would be the

focus and not the inaccurate to begin with daily "horse race".

I am good with Gallup abandoning an unproven model that failed as much as it succeeded...why not just toss a coin? Hopefully the other "daily tracking" of the horse race polls will be looked at less as serious and more as entertainment.....which apparently the public can not get enough of.

Surveys and on-line polls are for entertainment only, but when they are taken seriously as scientific polls then the entertainment becomes politicized and entire system of statistical random sampling is as corrupted as the system being sampled.

[SheilaT]

3. Great idea, which unfortunately

won't happen. Even if it weren't a first amendment issue (which it would be) all the news organizations in the world would go absolutely nuts. Their reporting model seems to depend heavily on the "horse race" conception almost exclusively. It's why they're so delirious with the current large pool of Republican candidates, and why they are so keen on Joe Biden entering the Democratic race.

There is some one poll that got it very wrong in a Presidential election, mainly because they stopped polling about a week before the election. Not sure which organization or which election, although 1948 seems the most likely to me. I'm trying to figure this out by cruising the internet, but either I'm not putting in the right search terms or my terms just aren't coming up with the result.

It's completely impossible for me to understand how anyone can possibly still be undecided about their choice for President even two weeks before the election, let alone a day or two, but I consider myself a high information voter who is highly involved in the election. Hence my participation on DU.

For your edification, please read "Franchise" by Isaac Asimov. Here's a link:
http://www.5novels.com/ScienceFiction/Asimov41/27325.html

[Eleanors38]

6. Bans, bans, bans. What entity/element has that power?

[Gormy Cuss]

2. That Gallup is giving up on candidate polling tells me that the organization is in chaos.

They've had several years to correct the methodology flaws that made them so far off the mark in the last election. Interesting. I wish I knew someone in the organization.

[Fred Sanders]

4. According to the article Gallup believes the entire theory of daily tracking polls is fatally flawed

This would be the statistical scientists they employ, the best in the business it is said, concurring on that.

[Gormy Cuss]

7. Which article? The OP link doesn't assert that.

Neither does this imbedded link:
http://www.politico.com/story/2015/10/gallup-poll-2016-pollsters-214493

I've only scanned the intro and finding in Gallup's internal report
http://www.gallup.com/poll/162887/gallup-2012-presidential-election-polling-review.aspx

and see no reference to daily tracking polls being flawed. What Gallup did suspect was the following flaws in their methodology:


---" likely voter" estimation was inaccurate

-- regional quota groups insufficiently controlled

---bad screener for race and ethnicity

---wrong sample frame on landlines because they relied on listed numbers rather than a list-assisted RDD approach.

[sufrommich]

8. This. This is what happens to polling organizations

that lose their ability to poll correctly.

[muriel_volestrangler]

5. Candidate polls are the worst way of describing a contest, except all the others

There's a real value in having the best estimate possible of how popular a candidate is - especially in a primary. A (possibly the) central feature of the Democratic primary is whether Bernie Sanders could beat Republican candidates in the general election; and that's fairly typical - do you go for a candidate that the party likes, or one you feel confident will win in the whole population?

The polling organisations vary in their accuracy, at different times; but they're better than listening to some pundit put forward their personal guess, or reading internet forums, or thinking about what your acquaintances say. They make an effort to ask the whole population.