Editore"s Note
Tilting at Windmills

Email Newsletter icon, E-mail Newsletter icon, Email List icon, E-mail List icon Sign up for Free News & Updates

May 8, 2007
By: Kevin Drum

THE CULT OF THE POLLSTER....In The Nation this week, Ari Berman writes about the various conflicts of interest posed by Mark Penn's polling work for Hillary Clinton while remaining CEO of one of the world's biggest PR firms. Matt Yglesias responds:

What Ari doesn't get into is whether, all that notwithstanding, Penn is just such a brilliant pollster that we should all be thrilled to have someone of his stature working for a leading Democrat. I would say "no."

Can someone please explain to me the cult of the pollster in big league politics? It seems like it all dates from Jimmy Carter's 1976 campaign, when Patrick Caddell became the first celebrity pollster, generally portrayed in gushing media reports as a cross between Rasputin and Wernher von Braun. Is that right? Or am I too young to remember genuine celebrity pollsters before that?

In any case, I can see how having a rocket scientist pollster on board might have been helpful back then. Polling was still fairly limited, and having somebody to do your own work and really think hard about what it meant might give you a real leg up. But today? Surely any politician with an IQ in three digits is pretty well aware of what people think and how they vote. You can hardly avoid knowing it given the tidal wave of polling information available at the national level these days. And the number of people who know how to read and interpret that tidal wave of data is way bigger than it was 30 years ago.

So in terms of raw polling skill, what does Penn bring to the table? Anything? I don't get it.

Kevin Drum 1:13 PM Permalink | Trackbacks | Comments (39)

Bookmark and Share
 
Comments

"Can someone please explain to me the cult of the pollster in big league politics?"

It's just an archetype re-establishing itself. The seer reading the entrails...

Posted by: David on May 8, 2007 at 1:34 PM | PERMALINK

My reading of history says that you are correct that Caddell was the first celebrity poller.

But I think you are incorrect in thinking that pollster is a low-level job.

In many campaigns, the pollster has a big chunk of responsibility for developing a candidate's message.

Posted by: Petey on May 8, 2007 at 1:34 PM | PERMALINK

I think the answer lies somewhere in the fact that this "pollster" is the CEO of a firm that does PR.

Connect the dots carefully...

Posted by: craigie on May 8, 2007 at 1:34 PM | PERMALINK

Pollsters are like astrologers: their talent lies not in interpreting the stars or some objective reality, but in helping their clients to interpret their own intuitions. It is a useful service to the extent that it gives them confidence in their decision-making, as it makes them think their decisions are grounded in objective fact.

Ulimately, however, it is a practice grounded in bad faith that yields nothing but bad results.

Posted by: lampwick on May 8, 2007 at 1:40 PM | PERMALINK

Nuance, Kevin. And triangulation. Being able to carve out a coalition or platform that will shave a voter or two off of supporting one of the other 'dwarves' running without risking those already committed to you for your already stated positions.

Posted by: Maeven on May 8, 2007 at 1:41 PM | PERMALINK

It's like anything. There are thousands of highly skilled lawyers. But some people have to hire the one "name" guy. Call it a form of risk-aversion -- you don't want to hire someone not universally heralded (as of yet) simply because if you end up making a mistake, the press and your supporters will saying "why didn't you hire x." If Hillary had hired someone else and stumbles, and folks will second guess her message operation and point out that "proven" polling guru Mark Penn was available. The guy on the bench always looks better if you make a mistake, never mind that guy might've brought even worse results. I think the fear of being second-guessed was why folks kept hiring Bob Shrum long after any sentient being would have figured out the guy's a pure mediocrity.

Posted by: Ryan on May 8, 2007 at 1:48 PM | PERMALINK

Politics is becoming like the movie "Groundhog Day". Everyone is doing the exact same thing in the exact same way. Over and over and over again.

Something that worked once will be slavishly replicated until something better comes along. Then everyone will flock to that idea like pigeons to a pile of breadcrumbs.

I'm not so sure if a pollster is needed in most of these campaigns. They just need someone who can convincingly spin the candidate's lousy numbers as a good thing without being laughed out of the room.

Posted by: Joshua Norton on May 8, 2007 at 1:51 PM | PERMALINK
In any case, I can see how having a rocket scientist pollster on board might have been helpful back then. Polling was still fairly limited, and having somebody to do your own work and really think hard about what it meant might give you a real leg up. But today? Surely any politician with an IQ in three digits is pretty well aware of what people think and how they vote.

Today its probably important because there is a flood of conflicting raw polling data from different sources, and you either need (1) an expert in polling to sift through and resolve the reasons for and meaning of the conflicts, or (2) an expert in polling to do your own polling and get decent answers that don't suffer from the conflicts. The obvious things ("the war in Iraq is rather unpopular") aren't the interesting ones.

And the number of people who know how to read and interpret that tidal wave of data is way bigger than it was 30 years ago.

Politics is competitive. The number of basically competent people may be larger (though its not a static field, so maybe not), the number of people that are better than the people working for your major competitors is probably about the same or even less (and in any class, a very small number), as the importance of polling has become more widely recognized.


Posted by: cmdicely on May 8, 2007 at 1:51 PM | PERMALINK

And let's not forget finding those great pools of untapped voters and shaping the message to mobilize them.

"Pollster" as usually used really is a kind of shorthand for "political strategist." Rove used to be the current ideal-- he not only knows all congressional districts and voting records down to the census enumeration block level, he also knows the turnout percentage of every possible slice of adult American society. Or so said the legend. "The Math" and all that. 'Course, it didn't work out so well for him last time.

Reporters love focusing on pollsters because that saves them the bo-o-o-o-ring task of trying to work out what the candidates actually stand for and want to do and what that will do to American society. Horse races are always more fun for most of them than wonkiness.

Posted by: Altoid on May 8, 2007 at 1:54 PM | PERMALINK

having somebody to do your own work and really think hard about what it meant might give you a real leg up.

And still does. Although there are many polls, all of them have their own flaws, and having someone dedicated to thinking about them is advantageous. Seldom are polling results clear and applicable; usually they have biases, ambiguities and contradictions.

Posted by: MatthewRmarler on May 8, 2007 at 1:54 PM | PERMALINK

Caddell was famous but a generation earlier Kennedy's polster--Harris I think--got a lot of attention as a smart guy. This is probably one of those things where the older you are the further back you can push it. I bet my father could have named some genius on public opinion back in FDR's day.

Posted by: bbolles on May 8, 2007 at 1:56 PM | PERMALINK

Kevin,

Actually it dates back to Roosevelt and George Gallup. Lew Harris and JFK were also a big item.

On the other hand, you may be right that Pat Caddell was the first "celebrity" pollster who became a big name from the political polling. Gallup and Harris were already household names due to their high-profile non-partisan polls.

Posted by: Joel on May 8, 2007 at 1:58 PM | PERMALINK

hiring Bob Shrum long after any sentient being would have figured out the guy's a pure mediocrity.

You're absolutely correct. Wes Clark Jr. was talking about how politicking is done now-a-days. The first thing backers want to know is "who are your people"? They want to hear the old stand-by names. Most of the time they're oblivious to the fact that those "big names" may have devolved into an ineffective bunch of yahoos who haven't had a new idea since Nixon resigned.


Posted by: Joshua Norton on May 8, 2007 at 1:58 PM | PERMALINK

Altoid's last point is exactly right. The media will talk about anything as long as it isn't related to issues that affect the country.

However, good pollsters are important, and there is always new data. In 1960, some pollsters told Kennedy that he should focus on Catholics in the South. On its face, the advice didn't make sense because there are not a lot of Catholics in the South. In reality, it did make sense because many of them were swing voters in swing states who voted for Kennedy once he got them riled up.

Similarly, pollsters for candidates don't just focus on issues--they figure out where the swing voters are, and that changes from election to election, and even changes at different phases of an election. During the final two weeks, you have to appeal to people who live in swing states and only vote occasionally, and you need a pollster to tell you who those people are in this particular election and what they care about.

Posted by: reino on May 8, 2007 at 2:05 PM | PERMALINK

MatthewRmarler: Yes, I agree. My question is whether your pollster has to be a super high-powered celebrity like Penn. These days, aren't there lots and lots of people who can do a pretty good job of interpreting polls and explaining what they mean? People who would just be mid-level staffers, not top advisors? That's the part I don't quite get. Frankly, I'd rather have my pollster be a technician who I can trust to just be a technician, rather than a big-time advisor whose own agenda probably colors his polling advice.

Posted by: Kevin Drum on May 8, 2007 at 2:05 PM | PERMALINK

The value of a talented pollster will increase in races where the pollster's employer is competing with several candidates appealing to the same voters in very similar ways. Both the Democratic and Republican Presidential races this year fit this description.

The theory is that very minor differences, in positioning on issues, tone, or even physical appearance can be decisive with voters who could just as easily vote for one candidate or one of several others. And this doesn't increase the perceived value only of pollsters, either. Strategists, TV ad designers, online consultants, and every other kind of campaign professional become more valuable when very small changes in a candidate's approach are thought to, potentially, make a decisive difference -- and, of course, when the candidates themselves do not disagree on very much besides who should be the next President.

It also makes a difference how much candidates think like campaign professionals. Today's candidates -- especially the Bushes and Clintons -- think more or less in precisely the same terms. Nothing in their professional lives matters as much as campaigning, so having people for whom campaigning is a profession is deeply reassuring.

Posted by: Zathras on May 8, 2007 at 2:10 PM | PERMALINK

For the first time I'm ashamed of my fellow commenters, who are usually much more informed and incisive than me. This is a pretty famous event in American politics.

George Gallup became a real celebrity with a political poll. Before then he was a low rent Ripley's. Here's a PBS site describing his ascent:
http://www.pbs.org/fmc/segments/progseg7.htm

Gallup guaranteed newspapers he'd refund their subscription fee if his column was wrong about who'd win the presidential election of 1936. In fact, he guaranteed that he'd be more accurate than the most respected presidential poll of the time, Literary Digest.

Literary Digest mailed 10 million questionnaires to 1/3 of American households and would get around 2 million responses. They'd been mass polling since 1916 and had called the winner every time. Gallup took a random sample and wanted about 3000 responses.

Literary Digest called the race for Republican Alf Landon. Gallup called the race for FDR. I'd guess George was much more famous in 1936 than Patrick was in 1976.

We can now return to subjects where you know more than me!

Posted by: Gene Ha on May 8, 2007 at 2:23 PM | PERMALINK

Though there had been, as others have said, use of polling going back to FDR, the difference with Pat Caddell is, he was the first pollster who was effectively promoted as being more crucial to the outcome of a campaign than the candidate. This may have been true, or may simply have been DC insiders' assumption that someone they didn't know (Jimmy Carter) couldn't possibly have won on his own.

Caddell's reputation maintained into '84, when he ran Hart's surprisingly effective insurgent campaign. On TV, George Will snidely said to Hart, We wanted to talk to the head of the camapign, but Pat wasn't available. (Hart couldn't conceal his disdain)

Caddell's evolution -- from McGovern to Carter to Hart to "I'll trash any Democrat you want" pundit -- is an object lesson in putting too much faith in analysts. Which is a problem I have with Mark Penn, as well. I think it's come to the point where he knows the result he wants, and he seems to massage his polling to bring the candidate along. It's the only reason I can think of why he'd always get the same, Lieberman-like findings, despite changing environments.

Posted by: demtom on May 8, 2007 at 2:24 PM | PERMALINK

Interpreting polls of average voters would seem to require a skill and a knowledge that Dem pollsters/consultants who have lived for decades in simply fabulous Georgetown mansions and derive a lot of their income from major corporations probably lack. I suspect these people don't know a damned thing about how average working people live and think and will vote.

Posted by: Chrissy on May 8, 2007 at 2:29 PM | PERMALINK

"...Politics is becoming like the movie "Groundhog Day". Everyone is doing the exact same thing in the exact same way. Over and over and over again.

Something that worked once will be slavishly replicated until something better comes along. Then everyone will flock to that idea like pigeons to a pile of breadcrumbs."
Posted by: Joshua Norton on May 8, 2007 at 1:51 PM

Hmmm. Just like movies, music, news, X.., has become? We direly need some true creativity in politicians' thinking, not a more effective way to push all the right buttons using ever more clever analysts.

Posted by: Doc at the Radar Station on May 8, 2007 at 2:33 PM | PERMALINK

It all has to do with the fact that most modern political candidates no longer have any principles to speak of. They will say and do anything to get elected, regardless of what they actually believe.

Therefore, the role of pollster becomes crucial to a campaign, since the candidate needs to be told precisely what to say so as to attract the most votes. Without a good pollster most candidates would be lost, since they seem to have the political instincts of a bull in a china shop and the backbone of a jellyfish.

You don't think Hillary, or Rudy, or Romney actually belive in anything, do you? How naive! They believe in whatever their pollsters tell them will get them votes!

Posted by: mfw13 on May 8, 2007 at 2:36 PM | PERMALINK

"Pollster" is just his title, Kev. As someone up above said, "connect the dots".

Posted by: brendan on May 8, 2007 at 3:11 PM | PERMALINK

Pat Caddell became a bigger celebrity 'pundit' during the Clinton impeachment brouhaha. He used to appear on a local cable show with David Horowitz. It was hard to discern which was the biggest foaming-at-mouth lunatic. Caddell turned up on Fox with their other celebrity pollster, Dick Morris. The most successful is Frank Luntz whose work turned the business into one of framing and linguistic twisting. Every statement from his company is Orwellian and not in a good sense he claims to use.

Posted by: Mike on May 8, 2007 at 3:39 PM | PERMALINK

Joel--I don't think Gallup worked for FDR the way that Caddell worked for Carter.

In fact, wasn't it Gallup who thought Alf Landon would win (or at least not be crushed) in 1936? IIRC from subsequent reports, his mistake was using the telephone to poll likely voters, not taking into account that the population of people without telephones in 1936 skewed sharply Democratic.

Posted by: Henry on May 8, 2007 at 4:14 PM | PERMALINK

I think the issue is that the campaign pollster conducts polls on a wide variety of subjects not covered by the herd, and continues to craft the image based upon the findings. Hence, the move to the middle in so many areas is probably based upon the nature of the questions, the number of polls taken to shape the issues, and the natural tendency of complex polls on obscure subjects to skew to the median.

Posted by: Ernie Fazio on May 8, 2007 at 4:16 PM | PERMALINK

Henry, it was the now-defunct Literary Digest that took the infamously bad poll in '36 -- they actually had Landon winning by a wide margin, and the reason for the error was, as you say, only using people with telephones or other signs (at the time) of affluence. Gallup was at the time a fledgling pollster, but he got it right. (His faux pas, of course, came in '48)

Posted by: demtom on May 8, 2007 at 4:18 PM | PERMALINK

"Surely any politician with an IQ in three digits is pretty well aware of what people think and how they vote."

This is truly an ignorant statement, Kevin. You obviously have no idea how polling is used to develop message and strategy in campaigns. The detailed info needed to guide a campaign is decidely NOT available publicly.

Just to cite a few quick obvious examples, if you were Hillary's cmapaign trying to figure out the most effective ways to go negative on Obama, would you be able to find comparative public data on, say, 20 different specifically worded attacks on Obama and exactly how well each one popped with 50 different demographic subgroups? Would there be public data available that compared different versions of those messages down to the single word level to determine exactly which wording works better? Is there public data on the many different private ads that Hillary's media consultant has made, in close consultation with her pollster, that allows you to figure out exactly which of Hillary's ads, down to the single word level, work better than others?

The answer is no.

Posted by: Disutrbance on May 8, 2007 at 4:58 PM | PERMALINK

what Disturbance said

Plus, you need someone to call up potential voters and ask "how would it affect your vote if you knew that X has been humping sheep? Without a condom?"

Posted by: thersites on May 8, 2007 at 5:45 PM | PERMALINK

George Gallup. He was BIG. Used to hear his name nightly from radio comics and sitcoms. Everybody in America who had a radio knew his name and what he did.

Posted by: buddy66 on May 8, 2007 at 7:06 PM | PERMALINK

The accepted and politicially correct statements for using pollsters is that they allow candidates to better utilize scarce funds and resources. However, with campaigns for State Senate campaigns going over $1,000,000 I do not know of any campaigns with scarce resources or funds.
The difference between Republican and Democrat pollsters for a long time were that the GOP pollsters were better able to read the numbers then the Democrats. That changed over time not because the Dems became better at reading the polls, but the GOP became worse at it.
The very good pollsters from the 70s and 80s left and went on to money making ventures. Now what you have are people cashing in on national trends.
In 1980 and 1981 GOP consultants were making fortunes based on the results from the 80 elections. In 1992 the Dems started making a lot of money, and in 1994 the GOP came back with what every one said were the best and the brightest of pollsters and consultants. Now the big craze after the voting last year are the Dems. The advent of cable and the 24 hour news cycle have made voting on all levels national events. I fear the expression that "all politics are local" will soon be a quaint saying of a distant past. Polling has become a self fulfilling prophecy. Not only saying what people think and feel and are passionate about but also leading them to think that way.

Posted by: Paras on May 8, 2007 at 7:43 PM | PERMALINK

I once had the privilege of studying under the direction of Dr. Yasuyuki Kuroda, Professor of Statistics at the University of Hawaii, and considered by many in the western U.S. to be one of the best all-around pollsters in his field.

Dr. Kuroda performed his work quietly, thoroughly and effectively, without any public fanfare, and his accuracy was amazingly uncanny. Even so, he shunned the idea that he should parlay his craft into a career in punditry.

would never have thought of himself as a celebirty

Posted by: Donald from Hawaii on May 8, 2007 at 8:08 PM | PERMALINK

Excuse me, I mistyped -- Dr. Kuroda never would have thought of himself as a celebrity.

Posted by: Donald from Hawaii on May 8, 2007 at 8:14 PM | PERMALINK

The disturbing thing about what Disturbance said is that its so true. These days political campaigning has descended so far that pollsters are now telling candidates exactly what to say so as to best appeal to a specific demogrpahic.

God forbid a candidate actually say something they truly believe!

Posted by: mfw13 on May 8, 2007 at 8:20 PM | PERMALINK

Why doesn't someone, someday, just go with their gut feelings? They might be surprised with the results.

On the other hand, fear of being unable to predict the public's reaction may well be the main reason that politicians don't go with their gut instincts. One of the main purposes of the major political parties, after all, is to suppress public opinion instead of expressing it.

Posted by: dr sardonicus on May 8, 2007 at 8:31 PM | PERMALINK

Someone upthread talked about the 1936 polling error of using telephones to do polling, but missing the correct prediction by not polling those without telephones. Couldn't today's polling be in error by not being able to poll the increasing number of voters who only use cell phones?

Posted by: Donna on May 8, 2007 at 9:20 PM | PERMALINK

As Petey and Altoid noted, "pollster" is just shorthand for master political strategist. Reading polling data is just a part of the job.

The importance of polling in this job more or less coincides with the emergence of popular votes used to select party nominees. It wasn't that long ago when candidates were chosen by the delegates at the convention and primary elections were just straw polls. In those days, the equivalent of today's "pollster" was the guy who knew how to work the delegates at the convention.

In that sense, the precursors to Pat Caddell are guys like Harry M Daugherty and Mark Hanna.

Posted by: mdl on May 9, 2007 at 12:01 AM | PERMALINK

Dr Sardonicus asks:

Why doesn't someone, someday, just go with their gut feelings? They might be surprised with the results.

There have been many many candidates who do exactly that, including some in this year's crop. Invariably, such candidates do poorly. They're never among the front-runners, and they tend to either drop out early or become irrelevant.

Occasionally they'll win praise from the political commentariat (as Ron Paul has done recently), but they don't win any elections.

Posted by: mdl on May 9, 2007 at 12:08 AM | PERMALINK

Google Edward Bernays Freud's cousin who basically invented p.r. and polling and blatantly called it propaganda. Bernays first used p.r. to help get the U.S. into WWI (sound familiar now?). It was also a big part of the yellow journalism of the day.

This is all covered in an excellent fashion focusing on Clinton and Blair's campaigns ion the BBC documentary century of the self.

Posted by: Matt Rogers on September 16, 2007 at 1:54 PM | PERMALINK

Good Day!!! www.washingtonmonthly.com is one of the best informational websites of its kind. I take advantage of reading it every day. www.washingtonmonthly.com rocks!

Posted by: payday loans canada on December 18, 2009 at 2:38 PM | PERMALINK




 

 

Read Jonathan Rowe remembrance and articles
Email Newsletter icon, E-mail Newsletter icon, Email List icon, E-mail List icon Sign up for Free News & Updates

Advertise in WM



buy from Amazon and
support the Monthly