The Death of Polling?
Perhaps the most annoying aspect of modern political coverage (and believe me, that says a lot) is the media's ridiculous obsession with polling. Everyday a slew of new polls is released and journalists, pundits, and bloggers puzzle endlessly over the meaning of the numbers. Who's up? Who's down? What does it all mean?
Mainstream journalists in particular love to discuss the latest poll numbers. There are a number of reasons for this--the most obvious being that it's very easy to talk about poll numbers. Any yahoo can read the latest Zogby poll results and offer some sort of banal analysis about what it means. Discussing poll numbers doesn't require any substantive knowledge and it gives journalists a chance to do what they love to do: offer their own bits of armchair political analysis and pop psychology. There's another factor too, I think. Unlike covering the horserace, covering substantive issues is much harder to do under prevailing journalistic conventions. It's a lot easier to maintain an aura of "neutrality" when discussing ostensibly objective data. When you wade into substantive issues, you're bound to anger one side or the other.
Unfortunately, this obsession with polls puts Heisenberg's uncertainty principle on full display. By attempting to measure public sentiment over and over again--and then discussing it endlessly--the news organizations end up altering what they're attempting to measure. And perhaps even more perversely, the news cycle begins to revolve more around momentary fluctuations in poll data than the actual events that are supposed to be driving that data.
Thus this week, as we enter the final days before the election, the lead stories in every major paper are about polls, not actual events. We're now so used to this poll-driven coverage that we forget how totally ass backwards it is.
So what can be done about it? The short answer is--not much. Polling is an integral part of modern political coverage and no amount of whining is going to make it go away, especially as long as pollsters' predictions continue to be borne out by actual elections results most of the time.
But therein lies my hope. The increasing popularity of cellphones, caller ID, and other technology may make it harder for future pollsters to get the kinds of representative samples they need to be at all accurate. For instance, there's just no way that a pollster would ever get a hold of anyone at my house. We don't answer the phone unless we recognize the number. That's what voicemail's for, right? I have a hunch this is a generational thing. If so, it's eventually going to be a real problem for polling companies.
I hope that's the case. If polls start to become unreliable, then maybe, just maybe, the news media won't drop what they're doing every time a new poll is released. One can always hope.
Mainstream journalists in particular love to discuss the latest poll numbers. There are a number of reasons for this--the most obvious being that it's very easy to talk about poll numbers. Any yahoo can read the latest Zogby poll results and offer some sort of banal analysis about what it means. Discussing poll numbers doesn't require any substantive knowledge and it gives journalists a chance to do what they love to do: offer their own bits of armchair political analysis and pop psychology. There's another factor too, I think. Unlike covering the horserace, covering substantive issues is much harder to do under prevailing journalistic conventions. It's a lot easier to maintain an aura of "neutrality" when discussing ostensibly objective data. When you wade into substantive issues, you're bound to anger one side or the other.
Unfortunately, this obsession with polls puts Heisenberg's uncertainty principle on full display. By attempting to measure public sentiment over and over again--and then discussing it endlessly--the news organizations end up altering what they're attempting to measure. And perhaps even more perversely, the news cycle begins to revolve more around momentary fluctuations in poll data than the actual events that are supposed to be driving that data.
Thus this week, as we enter the final days before the election, the lead stories in every major paper are about polls, not actual events. We're now so used to this poll-driven coverage that we forget how totally ass backwards it is.
So what can be done about it? The short answer is--not much. Polling is an integral part of modern political coverage and no amount of whining is going to make it go away, especially as long as pollsters' predictions continue to be borne out by actual elections results most of the time.
But therein lies my hope. The increasing popularity of cellphones, caller ID, and other technology may make it harder for future pollsters to get the kinds of representative samples they need to be at all accurate. For instance, there's just no way that a pollster would ever get a hold of anyone at my house. We don't answer the phone unless we recognize the number. That's what voicemail's for, right? I have a hunch this is a generational thing. If so, it's eventually going to be a real problem for polling companies.
I hope that's the case. If polls start to become unreliable, then maybe, just maybe, the news media won't drop what they're doing every time a new poll is released. One can always hope.
posted by A.L. at 12:51 AM
11 Comments:
I agree with you AL that the coverage of polls is become somewhat ridicules. I'm also often befuddled how people can claim a "dramatic" change in opinions based on polling data from different polling data sets. Demographics, partisan make-up, and wording of the questions can often sway the results as anyone who has seriously analyzed these polls would know.
But as you say, its become easier to just say "Well the latest polls show..." rather than seriously consider the issues behind that data.
It'll take a while. First thing pollsters will do is adjust the margin of error. Then they'll try different techniques to ensure fairly accurate results. This is their life's-blood. They won't just quit.
That said, I really wish the MSM were pushing the ideas, policies and positions instead of the polls. The public doesn't get any useful info out of how the electorate is leaning. They can get info, useful in the polling booth, if the ideas, policies and positions are spelled out for them.
But then, Diebold will probably negate their votes anyway. C'est la vie.
"If polls start to become unreliable" ????
Sorry, but that just tickled me. I think polls started to become unreliable long ago. The 1994 election was noted as not being predicted by polling.
They say the most reliable polls are the exit polls -- but as we've seen in recent elections, even those aren't all that reliable. Or maybe it's just the elections that have become unreliable.
What with people not participating (I have refused to give answers to polls for 20 years), cell phones, caller ID, people who simple don't answer their telephones at all, and the ones who lie, statistical sampling has met its match and then some. Throw in the "submerged" voters, those who will vote for the first time tomorrow, and there's just no way for pollsters to win.
I notice that many of these polls are samples from about 2000 respondents. This may be fine, ordinarily, but statistics says that, if the variance goes up, to get a good result you will need more samples. What could cause the variance to go up? Anger? More new voters than expected? How about the big push to do early and absentee voting?
I'd like to be a pollster. Scratch that, let me rephrase: I'd like to have been a pollster for the last several weeks. But if I were, I'd be on a plane for somewhere out of the line of fire tomorrow morning. Early.
IMHO, you miss an important point - THE LYING LIARS IN THE MSM HAVE NOTHING BUT POLLS TO REPORT.
They have not actually provided meaningful coverage of most every other issue for the rest of the year - so why would they start now. Yes - it is ass-backwards, but its really just the same ol' game - distractions and "tune in tomorrow to see the color of our feces" style journalism.
That said - the math concepts behind polling are as sound as the laws of physics - most don't understand them. If polls are "wrong" it is because the math underlying them is being dishonestly or mistakenly applied (you can guess which one).
You raise good points, but I have to take exception when we play the repug/neocon game and start declaring polls don't matter and are always wrong - THAT IS THE MEME THAT WAS USED TO COVER UP THE THEFT OF THE 2004 ELECTIONS.
Exit polls should be an important part of our election process - we regularly use them to assess the honesty of elections in other countries, BUT THE MSM JUST POO-POOs THEM HERE.
A.L., we need polls because we don't have free, fair, open, and verifiable elections. DO YOU WANT TO JUST GIVE THEM BLANK BALLOTS AND THEN ACCEPT THEIR RESULTS UNQUESTIONABLY?
Oh, and an afterthought: I don't really care that much if the newsies blather on about meaningless polls. Or much of anything else. I don't remember the last time they covered anything truly meaningful in a sensible way anyway, and, unfortunately, the only way we can "un-elect" them is by not watching. Since they do manage, apparently, to entertain a fair number of people (apparently not watching to be informed, but merely entertained) -- that's a systemic problem.
What bugs me more is the use of polls by politicians for "triangulation." If everybody would participate in making the polls hopelessly unreliable, maybe we could put an end to that -- and it would be a tremendous boon. Without rapid turnaround from polls, zero-integrity politicos like Joe Lieberman and Norm Coleman would never have managed much of a career, and wouldn't that have been an improvement? They'd be left to run on their record or their principles.
So please, everyone: when the pollsters call, simply decline to answer. And if you want to play with them and give false answers to every question, I, for one, will not blame you.
As for exit polling, I think it's much more important to have a verifiable, recountable, voting method. We shouldn't have to rely on something as inaccurate as exit polling to verify election results.
Rather than invoking Heisenberg, I prefer to think of excessive bloviating on polls as a feedback howl which drowns out everything else.
I agree that the pundits talk about the polls because they provide a politically neutral subject -- no one can accuse you of liberal bias if you only talk about the numbers.
I also agree with anonymous above that exit polls should be an integral part of the election process. Testing and calibrating your equipment is just good engineering.
Just to clarify: my post was refering to pre-election telephone polls, not in person exit polls. I think exit polls serve as an important cross-check of the integrity of the overall process and have little, if any, Heisenberg effect on the vote itself. Plus, because they're done in person, they won't face the same obstacles telephone polls will.
Charles: the one thing I disagree with you about is the accuracy of polls. Pollsters can and have been off in the past. But they are surprisingly accurate much of the time, especially when taken in the aggregate.
This is one of the reasons I bail on The Daily Kos about three weeks before every election. Markos is a polling "expert," so fucking what. He even titles his posts with them!: "Polls, polls, polls." The guy quits being a writer and a liberal, jesus christ with chicken bisquit crackers I just can't read it. Plus the diaries become rabid goon squads of Democratic Party Correctness, it's disgusting.
We don['t talk about wages, we don't talk about unions, we don't talk about daycare, we don't talk about pensions or disability, all we talk about is our dick-in-the-hand poll numbers.
By the way, Al, you get paid for excellent political writing like this by becoming a fucking pundit, man. You need some heroes, well, they'd be Ezra Klein and Mathew Yglesias, time to live on dogs wages in a profession in total disgrace. Does that sound good to you? They please stay the fuck on course, I'm positive you're doing precisely the right thing with perfect timing.
Anyway, polling is a curse that will never leave us. It should be 20% of what we talk about, not the 90% we use now. The world can always change, it's fuckin' true, we'll just have to try. It's fun bitching about it if nothing changes anyway.
I love Markos Moulitsas, okay, it's just a peeve of mine. Markos has made enormous sacrifices for us all and I'm eternally grateful to him. I could not have done what he's accomplished.
I dunno what the hell has gotten into me, I'm jumpy'n snappy as hell. We'll win at least the House tomorrow and it is going to be one real bitch.
The country is still in rank denial about losing and might just stay that way until 2008. Bush and the press will do everything they can to make the Democrats fail. We need the A team of fucking A teams out there'n jesus save us, please forgive me for sayin' it, but the last six years have shown us we've got nothing but chumps, we're lucky if they can find their dicks in the morning. Pelosi, thank God, is not so afflicted, but can she get Murtha as whip? She could crash and burn with that dickless whip she has now.
We and the country could be so fucked. Two years of constitutional crash and burn, toothpick Dick Cheney said just yesterday subpoenas mean nothing to his felon ass. Of course not. And then do we really want these felons in charge when we have to flee to Kuwait? No. Nor can we wait another day. Fucked again.
It's bad, it's really bad. We just cannot look our men and women in the face in Iraq and say wait for 770 days, we might get our political act together.
Bush has no succession, even, to worry about for 2008. He could do anything, I mean anything, the next two years, he is a lying bloodthirsty freak, who will be angry and petulant.
Oh well, here we go. Good luck and godspeed, Nancy Pelosi, you are so going to need it.
Andy Says:
Here's a poll I'd like to see. How many people vote for X
because they read or heard that people of their type vote for X?
I guess, based on introspection, that it is a big part of
lots of people's votes.
I want to see a poll on why people talk to the pollsters and whether or not they actually give any meaningful information.
Post a Comment
Links to this post:
Create a Link
<< Home