Ladies and gentlemen, may I have your attention,
Moderator Michael Dukakis has just called tonight's
meeting to order.
Welcome, ladies and gentlemen, and thank you for joining
us here in Boston's historic Faneuil Hall on The Advocates. During the weeks
preceding the presidential election on November 7, a wide variety of public
opinion polls appeared to show Senator McGovern far behind President Nixon,
and certainly the results on election day bore out those predictions. Did
the publication of those polls have an effect on the outcome of the
presidential election, and does the publication of candidate preference
polls impair or distort the democratic process? That's the underlying
question which we debate tonight. And we turn first to Mr. Fisher. Mr.
Fisher, the floor is yours.
Thank you, Mr. Dukakis. Last week on election day,
millions of Americans watched the Today show on NBC. Less than two minutes
of that broadcast were on substantive issues. More than 39 minutes were
devoted to polls and predictions. Once again, people were told how they were
going to vote. That day, 45 percent of the Americans of voting age stayed
home and did not go to the polls.
A detailed study has just been made of the TV evening news
coverage of the primary campaigns, the Democratic primaries from last
November to last June. During that period, NBC evening news had 316 stories
about the campaign. Less than one-half of those stories made any mention of
any substantive issue. The other networks were about equally bad. This
impartial study concludes that any citizen who wished to know what the
candidates had to say about the issues and what their qualifications were
for high office had very little help from television news. Yet TV is the
principal source of news for more than 60 percent of Americans.
Tonight, we consider how the media should allocate their
resources, how they should spend their time. I believe that they should
concentrate on the issues and on the competence, the ability, and the
experience of the candidates themselves. Polls on issues help politicians
understand their electorate, but one kind of poll that contributes very
little to the democratic process is what we call the head-to-head poll, the
candidate preference poll. It's the "Who's ahead now?" kind of poll, and
publication of these horse race figures do little good and cause great
Now, pollsters themselves are becoming increasingly
concerned about the media treatment of their polls, irresponsible use of
their polls. We have one pollster with us tonight, Dr. Herschel Shosteck.
Would you take the stand?
Welcome to The Advocates, Dr. Shosteck.
Mr. Shosteck is President of Herschel Shosteck
Associates, a Washington based public survey firm, and he has consulted on
political campaigns during the last six years. What kind of polls do you
take, Mr. Shosteck?
I undertake a large number of public opinion polls
focusing on issues and mass media analysis, radio and television
Do you take candidate preference polls?
Do you publish them? Are they published?
No, they are not published.
Do you believe that such polls should be
No, I do not think they should be published; indeed, I
think it's a disservice to the electorate to publish them.
Well, for three reasons. First, they tend to be
inaccurate. Secondly, they tend to be misrepresented. And, thirdly, and most
importantly, they diffuse the debate from the basic issues to this horse
race kind of phenomenon, who's ahead in the polls.
Now, the old Literary Digest polls . . .
Excuse me just a second. Mr. Shosteck, when you take
these preference polls, who do you take them for? Are you working for
I take them for individual candidates.
They are not published in the media.
They are not published in the media. I recommend that
they not be published in the media.
Are most of the questions you ask, “who’s ahead?"
No, we look fundamentally at issues. What are the basic
issues. The public knows what the issues are. You ask them, they'll tell
you. And the candidate is the one who has to be educated.
Now, I thought the Literary Digest poll in the Landon
days was inaccurate. But aren't polls accurate today?
Some are. The major polls by Harris and Gallup are
fairly accurate. But during an election year, these are not the major polls
that are published. The major polls that are published are the straw polls
by the newspapers, publishers who have very little competence in the polling
field, who don't know what they are doing, who are interested primarily in
getting a dramatic headline, and, as such, could care less what the polls
are really saying.
You think a large number of newspaper polls are
And even the accurate ones, you say, are
Yes. The polling process, the electoral process is an
information process. The public is trying to get information about the
candidate. They're educating themselves. And to ask them what they think
about a candidate, where a candidate stands, before this education process
is completed, before election day, is the same as asking a jury to come in
with a decision before they've heard a trial. They're not in position to do
that yet, and they know it, and they don't want to.
Besides being misunderstood and some inaccurate, you
said they distorted the public debate.
A very brief answer, please, Mr. Shosteck.
Very much so. Basically, they focus away from the
fundamentals of any campaign, and that is who is running and what are their
capabilities and what are the basic issues.
All right, gentlemen, let me interrupt, and let's go to
Mr. Miller who's now going to ask you some questions in
Mr. Shosteck, you say the major polls, such as Gallup
and Harris, are generally accurate. Aren't those the very polls which the
major media then would not publish? That is, all we'd have are these other
polls. We'd have no Gallup or Harris poll in national media or
I don't understand your question.
Well, let me put it another way. There will still be
polls taken, won't there?
Labor unions will take polls.
Businessmen will take polls.
Of course, they will. I mean people who contribute to
the campaigns are still going to take polls.
And so we're going to have all of these private polls
floating around. Suppose, for example, labor unions say to their members,
"We've taken a poll. I'm not going to tell you what it is, but we've taken
it. And our poll shows us that Candidate X is leading." Now, a business
group takes another poll that says Candidate Y. What you've done is
prevented the media from publishing the only genuinely accurate source,
What I'm saying is that we shouldn't publish the polls
because they divert from the basic issues. The basic issue is what are the
issues and what are the capabilities of the campaigns.
No, but they'll still be taken, and presumably they'll
be available to people who want to purchase them. If you want to subscribe
to a pollster's newsletter, you'll subscribe and get the polling
So the net result is that only those who can afford it
will have the privilege of being misinformed.
That's incorrect. That's entirely incorrect. The people
who hire my services or hire Mr. Field's services or hire Gallup’s services
are those who are interested primarily in organizing their campaign in a
prudent manner. And the outcome of this is that they are very much aware of
their need to know what the public opinion is, what the voter's opinion is,
on the basic issues of their district.
But doesn't this proposal go opposite to a major thrust
in American society; that is, today, when we talk about information that
Presidents have or people making decisions have, we want more disclosure. If
a President, like President Johnson, is going to make decisions based on the
Vietnam War on a poll, we'd like to know how good that poll is. If a
candidate is going to key his campaign to certain issues, we want more
disclosure, not less, don't we?
I think disclosure is very fine on what the fundamental
issues are. Indeed, that is something I am very much in favor of. What I am
against and what I feel is a disservice to the public is the
irresponsibility of publishers in focusing entirely on these horse races
which are, in many cases if not most cases, distortion; and in so doing,
they are taking away from the basic issues; they are not focusing on the
But do you assume that there will be these other polls.
They'll constantly be in circulation, and therein simply be rumor, innuendo,
and no source to check them out.
Well, Mr. Miller, let's distinguish between polls on
issues, which, I take it, . . .
No, I'm talking about polls on candidates. I'm talking
about . . .
You're suggesting that some people commission private
horse race polls, if you will, . . .
Well, of course, they will.
…and it's those that will be floating around. Let's keep
our discussion on that.
Major campaign contributors ...
In point of fact, my experience has been that they
don't. They are far more interested in what the issues are, and this, I
think, is very laudable.
They don't, and they won't?
My personal experience and that of my colleagues with
whom I have discussed this is that the least important part of any private
poll is the horse race part of it, particularly when you get into a
campaign. Now, that doesn't ...
And candidates do not leak their horse race polls to the
press and attempt to get them published?
A lot of them do leak them.
And I think that, one, they're ill-advised because I
think it has a negative effect. Secondly, more importantly, I think that it
is again diffusing the proper focus of a campaign away from the competence
of the candidate and the basic underlying issues, so the public focuses on
this horse race.
You think before there were polls, people didn't focus
on the candidates and the horse race.
What I'm saying is that, in my opinion, the publishers
are little more than pimps for the pollsters who are for . . .
That says something about the pollsters that may not be
It says something about the publishers also. They are
focusing away from the basic issues. They have trivial statistics, and by
focusing on these trivial statistics, they are abdicating their journalistic
responsibilities to educate the public, and I think they are cheating the
American people in so doing.
Well, let me ask you this. Do you support tonight's
I think that the horse race type poll, in which you have
Candidate A as two points ahead of Candidate B and Candidate B as three
points behind Candidate C, is an insult to the American people. I think the
American voters deserve more than that. I think that the publishers are
cheating them. And I think that the polls should not be published. I think
that the publishers should voluntarily refrain from this and go on about the
business of journalism which is to educate and to inform the public...
Educate and inform the public by your standards of what
education and information is to . . .
…on meaningful topics.
Mr. Miller, I'm going to let you follow up on this
question in just a moment, but I want to go back to Mr. Fisher for just one
last question, Mr. Fisher.
Well, basically your notion is that the basic facts of
publishing these figures contributes no benefit and does have this
diversionary . . .
Well, it not only has no benefit but it's divisive. Not
only does it focus away from the issues, it can be used to subvert the
democratic process because early in the game, if a candidate will leak a
poll and his poll is structured either honestly or fraudulently to show that
he's ahead and the other candidate's behind; in a primary election, the
effect of this is to dry up campaign funds. And by so doing, this means that
the candidate at the disadvantage cannot get money, he cannot communicate
with the people, he cannot present his capabilities, he cannot present his
stance on the issue, and this is, I submit, a subversion of the democratic
All right, Mr. Shosteck, Mr. Miller was waving a piece
of paper. I don't know if he's going to continue to wave it . . .
Well, you think that not only the ordinary voter but the
person who contributes major sums of money to political candidates in the
neighborhood of 10, 50, or 100 thousand dollars simply is too naive to
understand what a poll really means.
Polls are very technical; and in many cases he's too
naive, and in many cases he's sold a faulty bill of goods.
Mr. Miller, no more questions. Mr. Shosteck, it's been a
pleasure to have you on The Advocates.
Thank you very, very much. All right, Mr. Fisher.
To testify from his own experience on the harmful
effects which polls can have on a campaign, I call on Congressman William
Welcome to The Advocates, Congressman Green.
Congressman Green has conducted six successful
Congressional campaigns and one unsuccessful campaign to be Mayor of
Philadelphia last year. Now that you've had seven campaigns, Mr. Green, in
your judgment, does the publication of horse race head-to-head polls help
I don't think it contributes a thing to the democratic
process, no. And I think it's harmful, as a matter of fact.
What happened in your Philadelphia race?
Well, I ran last year for Mayor of the city of
Philadelphia; and shortly before the end of that campaign, one of the
newspapers published a poll which showed me running a poor third. The fact
of the matter was I was running a close second.
What happened after the poll was published?
It had a catastrophic effect, first of all on
contributions in the campaign. The situation was such that there were three
candidates in the race. Actually there were four. One dropped out and
endorsed me. The Governor of the state came into the city and endorsed me,
and the campaign was really rolling. And then this poll came out, and the
thing that happened, really, was that for the next three or four days
virtually nothing was discussed in the campaign except that poll. And in a
weekend prior to that we had raised about $40,000 in the campaign, and
during that time we didn't raise a dime.
Now, you say the poll was inaccurate. How do you know it
It was inaccurate because the Philadelphia Daily News
that published the poll, after four days of examination, research, study
calling pollsters around the country to test the sample they had taken,
printed the entire front page to apologize to my campaign.
Full front page apology for an inaccurate poll.
Did that pick up the morale of your staff and put you
back in the...
It put us back in the ballgame, yes.
Would an accurate poll 10 days earlier base made a
I tend to think it would have had a profound effect
upon the money raising in the campaign and had a profound effect upon a
discussion of the issues, which I consider to be the most basic thing.
Did the publication there and in other cases divert the
public from discussing the merits of the candidates?
I would say the top line on the media, radio and
television programs and newspapers, was that inaccurate poll for several
Does the publication of such polls, head-to-head polls
do anyone any good?
I suspect it does a great deal of good just for the
newspaper that published it and perhaps for the bookies or betters.
All right, Congressman, let's turn to Mr. Miller. Mr.
Miller, some questions.
I'm glad you ended on the note of bookies or betters,
Congressman, because before we had polls, we still had estimates of who was
going to win. Let me read you a headline from the November 1, 1920, New York
Times, the Harding-Cox election. Headline reads, "Harding odds still six to
one. Even money bet that he will carry New York by 300,000." Do you prefer
Jimmy the Greek to Gallup in formulating these…
No, I don't. In fact, I think that people before these
polls . . .
Congressman, I'm going to claim a point of personal
privilege, if we have any more references to my ethnic ancestors.
Obviously, he had to be Greek. The fact of the matter
is that these polls are published today as if they were scientific and under
that label. You know, the contest you referred to or Jimmy the Greek's
figures are considered sort of an off-the-cuff, almost comic
Let me ask you what your standards are for freedom of
the press. Now, it's not only polls, for a moment focusing on the inaccurate
polls. There may be inaccurate crime reporting because some crimes are
Well, I tell you, as a matter of fact, I…
Let me finish the question. Should papers publish no
Well, first of all, I'm the first Congressman in the
country to introduce a bill to provide for a national criminal statistics
center because I think that the way we record statistics is very inaccurate.
And I don't see what bearing that has upon elections. It may do something
for the crime problem. Secondly, let me say that I do not advocate, nor do I
think anyone should advocate, a law to outlaw polls being published. But I
feel that the democratic process is best served is we concentrate on the
real issues in the campaign, be they drugs or gangs or crime or whatever,
rather than having three or four days of a campaign headlined as to who's
going to win or not win. That's a decision for the people. And the real
poll, the poll of the people, will be the one that decides.
Well, now there's no question but that the democratic
process is best served by focusing on the real issues. Sometimes there are
real issues for polls. In primary campaigns, for example, the Democratic
Party will have to choose a Democratic nominee in 1976. Isn't it helpful to
the average voter, not the one who can pay $25 a month for a voting service,
but to the average voter to know which of the candidates of the Democratic
Party may be more popular and be able to win?
I don't think that that's particularly helpful, and I
think that's one of the tragedies, both in fund raising. The people have a
tendency to look to see who it happens to be that is most popular rather
than whom, in justice and in right and what's good and everything else, they
People who contribute funds will still have these
polls. They don't have to appear in the New York Times for them to have
I'm not so sure of that. That's not my
Let's take it this way. What else should newspapers not
publish? You say you don't want them barred from publishing this
information. Should they not publish foreign affairs articles because
they're complicated and misleading?
I don't think that there is anything that newspapers
should be prohibited from publishing. I think that they do, in many
instances, exercise a great deal of self-restraint.
Well, what is your position on tonight's proposal now.
Your position is newspapers should not publish these candidate preference
I think they should exercise self-restraint in
Well, of course, they should exercise self-restraint.
They do in all sorts of news.
Well, they don't when they publish polls.
But if they have an accurate poll in the midst of
inaccurate information flying around from everyone else's poll, a newspaper
has an accurate poll that it's checked out that meets its standards, it
shouldn't publish that poll?
The fact of the matter is that they don't check them
out. And I'm a perfect case...
Answer the question. Answer the question, now.
No, you didn't. A newspaper has an accurate poll.
There's all sorts of misinformation fly . . .
Well, who determines whether or not the poll is
It meets the methodological standards, it's scientific,
you've got the right sample…
The Philadelphia Daily News thought they had
No, well, you see, I want the question answered. In the
midst of misinformation, a newspaper is satisfied that the poll it has, not
Jimmy the Greek odds and not a candidate's leak, a poll . . .
The Daily News was satisfied that its poll was
accurate. Now, let's presume for the moment that their poll was
Yes, and you would still prohibit them from publishing
I wouldn't prohibit anybody from publishing
Would you still have them not publish it?
I would think that it would serve the best interest of
discussion of the real issues, if they left the horse race aspects out of it
and dealt with drugs, crime, gangs, education, housing, and...
Now, just a moment. There are all these things all over
the place. Labor union leaders are telling their members someone is ahead,
and that's an issue in the campaign.
Well, I'm not sure that that's the case at all. I don't
agree with that assumption. I heard that line of questioning earlier, and I
don't agree with it.
You don't think other people will take polls.
That's not been my experience in seven campaigns.
There are no private polls now, the Sindlinger Poll or
other polls, that people subscribe to?
Oh, sure, there are many private polls.
I think people take polls. I've taken polls
Gentlemen, we'll have an opportunity to continue this,
but let's go back to Mr. Fisher for a quick question.
Mr. Green, do you think that it's an abridgement of
freedom of the press for newspapers not to publish confessions in advance of
trials? Mr. Miller's suggestion that if a newspaper doesn't publish a
category of things, this somehow abridges the freedom of press. I take it
that most newspapers, exercising responsibility, do not publish confessions,
purported confessions of defendants, even though they think they're
Well, newspapers don't publish a great many things that
are sometimes kicked around in. a campaign which, perhaps, don't belong in
it, the private lives of candidates and things like that. The only
suggestion I make tonight is that constant discussion or headline discussion
of who's where in the campaign and who stands where in the polls just
diverts from the more important discussion that both the people and the
newspaper should be concerned with, and that's the issues of the
Mr. Miller said . . .
No, Mr. Fisher, I'm not going to let you ask another
one because Mr. Miller…
Of course, we're not talking about self-restraint in
their private lives, we're talking about self-restraint in their public
standing. And since public standing is an issue that people look to, who to
give money to, why shouldn't the average voter have the information that the
wealthy contributor can get by paying for it?
Well, first of all, I don't know that the wealthy
contributor has any advantage over the average voter in that situation. You
kept talking about labor unions, and you virtually said a minute ago,
yourself, you thought everybody took polls. I don't agree with that. The
question is should our newspapers or our television stations or our radio
stations, which are a fundamental source of the information that the voter
receives, be the conduit for this guestimate, this projection.
No, the question is should they publish no polls. Your
position is, in exercise of self-restraint they should publish no
My position is that the public interest would be best
served if they devoted as much space as possible to the discussion of the
Well, what is your position on this question, though?
Do you think polls are so harmful they should never publish them? That's the
question we're talking about, whether this is such an enormous evil that
people don't have a right to hear or see an accurate poll ...
First of all, I don't interpret that to be the
question. I interpret the question to be whether or not they should publish
them. And I think they shouldn't. I did not say that I thought they were
absolutely eating away at the roots of our democracy.
All right, on that note, gentlemen, I'll have to excuse
Congressman Green. Thank you very much, Congressman, for being with us. Mr.
Let me remind you, in case Mr. Miller is misleading
you, no one is talking about prohibiting anything. We are suggesting that
the mass media, like television and radio and mass newspapers, refrain from
publishing a particular kind of horse race poll which is, "If the vote were
taken today, who would you vote for?" because this is usually interpreted as
who is ahead without the subtlety and background of changes in position and
what they are. To tell us that democracy can survive very nicely without
publishing head-to-head polls, I call on Mr. David Anderson from Victoria,
Welcome to The Advocates, Mr. Anderson. It's nice to
have you with us.
Mr. Anderson is the leader of the Liberal Party in the
province of British Columbia and a member of the legislature. He is a former
member of the Parliament in Ottawa. Now, did newspapers and radio use to
publish candidate preference polls in British Columbia?
And how were things then in British Columbia?
Well, they were much as they are elsewhere. This is
prior to '53, they were published in the normal sense, and people decided
that it put too much emphasis on the horse race, who was ahead, and not
enough on Issues, so the legislature, in 1953 passed an amendment to the
election act and prohibited polls during the electoral period, which is the
38 days immediately prior to the election.
They are prohibited during the campaign, which, in
Canada, runs five weeks and a little bit . . .
Provincially, it runs five weeks in British
How are things in the Province now in provincial
Well, it's hypothetical. We're dealing with the
situation now as opposed to the situation prior to '53 But in my view,
there is more emphasis upon issues, more emphasis upon what I think should
be the real questions facing electorates, and less upon the who's ahead
today, the discussion of why so-and-so's ahead in terms of the day to day
influence upon the electorate of media and other things advertising and
other things . . .
Do the people still have an interest in the horse race
Oh, heavens, yes. By golly. We have, by the way, in the
federal elections, we have the horse race element there, and we have polls
taken. But there is one thing that shows, I guess, that they are interested.
We have a poll fair every year in British Columbia, and it often coincides
with the provincial election, and there is a gentleman, very enterprising
hamburger salesman, who sells hamburgers with the names of the party
leaders. And this is legal. There are Andersonburgers. There are . . .
I didn't do very well. In any event . . .
Are the results of hamburger sales published in the
Yes, they are, and this is considered to be legal
because they are not testing their political preference. They are simply
saying what hamburger would you like to buy. The hamburgers happen to be
identical. I think the man could be charged, but I'm not sure about
Nobody's prosecuted him yet, I take it.
Nobody's prosecuted him, but would it be better if the
press refrained from publishing those crazy figures?
Well, fortunately, and this is where we get back to the
question of the private polls, fortunately, it's considered to be
inaccurate. It doesn't have the backup of the scientific polling system
behind it. It doesn't give the impression of being the gospel truth. It's
just simply John Dice (?) and his hamburgers down there at the P and
Now, in Canada in the federal elections, polls are
permitted and published.
They are, right.
And you, in British Columbia, have had a chance to
compare your last August election with this October election.
Right. We had a provincial election on the 30th of
August, the federal election on the 30th of October.
And which practice do you think is better?
I prefer the provincial. I think that the federal
election showed that the Prime Minister, Prime Minister Trudeau, was ahead
for a good part of the time; indeed, he wound up ahead in the end. But, the
publishing of polls led to the horse race effect, not enough emphasis upon
issues, and indeed it also, in my mind, led people to simply not bother to
vote on the grounds that it was a foregone conclusion.
All right, gentlemen, let's turn to Mr. Miller. Bear in
mind, Mr. Miller, that in that hamburger poll, fish eaters are, quite
I was going to say I thought the hamburger poll tested
the voter's gut reaction to the election.
That's probably an irrational way to test their
Yes, well, fortunately the polls that we have today, as
opposed to the ones that were around in ‘53 are more accurate than simply
testing that reaction.
I should say you can cook the results. There were a
good number of hamburgers purchased for committee rooms in the last days of
Now, let's do what Mr. Fisher likes and switch to the
issue here. Tell me about the British Columbia statute. It's not only
head-to-head candidate polls that are prohibited. In fact, candidates are
prohibited from taking any poll, even from finding out what their electorate
may think on an issue.
Well, the law has not been tested. In actual fact
people do. It's interpreted to be what party would you vote for or how would
Well now, it says that no person shall take votes which
will prior to the election distinguish the political opinions of the voters
in any election.
Yes, but, in other words, it's quite possible to say,
"Are you concerned about crime in the street?" and if the reply is, yes, of
course that could be published.
That's OK. Are you concerned about crime in the
street?" And which candidate most people think take a position on crime in
the street, that is not a…
Well, not if it's interpreted in terms of how would you
vote were the election tomorrow.
Let me ask you if you've had this problem in British
Columbia because there are other countries that have tried this. The
Germans, of course, tried once to ban polls in the late 1960's in one of
their federal elections. What they found was that out-of-country newspapers,
the London Times took a poll that was widely distributed inside Germany. Has
anything like that happened in British Columbia?
No. This is entirely for provincial elections, and
there may have been some Eastern Canadian papers that tried to do this, but
I'm not aware of it if there are.
And suppose there were polls taken. Are you
recommending this British Columbia system for the United States?
No, I'm not. Your system is different from ours. Your
Constitution is different from ours. We do have a legislative prohibition
against polling. We are here discussing whether editors…
Are you recommending that, that editors ought not
publish any polls at all?
Well, my view is that it would be helpful during the
five weeks prior to the election.
Well, during the five weeks. Of course, that's one
major difference, not only size and habit, but one presidential candidate
announced 22 months before the election.
Gentlemen, we're getting close to time. Would you
recommend this policy, do you think, for most states and provinces?
I think you can set up a reasonable length of time
without any trouble, and I think that it would be helpful to focus attention
In other words, it's OK to find out at a certain time
but not at a later time?
When the election draws into that last two months, I
think you should . . .
All right, Mr. Anderson . . .
. . . thanks for coming all the way from British
Columbia to be with us. Thank you. All right, Mr.- Fisher.
Thank you, Mr. Anderson. The issue is one of the
benefits of a particular kind of poll, the kind which shows the horse race
figures. Most pollsters spend most of their time on gathering other
information, and editors have great restraint in publishing other types of
information. It's a narrow question. Do not let Mr. Miller deceive you into
thinking we're passing a law to prohibit studying the state electorate or
what's on their minds. Thank you.
Thank you, Mr. Fisher. We’ve now reached the midway
point in our program, and for those of you who may have joined us late, Mr.
Fisher has argued that the democratic process would be improved if
newspapers and magazines, radio and television, would refrain from
publishing candidate preference polls. Now, it's Mr. Miller's turn, and Mr.
Miller, you're on.
We're not publishing a law, but we are saying that
because polls are too complicated or because someone thinks they don't state
the real issue they ought not to be published. That's the approach of
dealing with complexity, with complicated issues - by secrecy. It's a
fundamentally elitist view, which concept tells us that not only the average
voter but also the substantial contributor simply can't be trusted to deal
with this information; and therefore that we must protect people from
information. I categorically reject that view.
Of course, polls are complicated. So are many other things
we deal with. Taxes are complicated. Defense is complicated. Education is
complicated. There's misreporting on all of those things, difficult things
to understand, inaccuracies. Does that mean that newspapers should publish
nothing that falls in an area where there may be some inaccuracy or some
misleading things? Of course, not.
Now, there are disappointments with elections. But that
disappointment and the disappointment with the polls cannot be elevated to
this policy because the problem is that the mind that thinks this is too
difficult or that makes the judgment that it's simply not relevant will also
think that other things are too difficult or are not relevant. The spirit
that is afraid of the public having this information will be afraid of the
public having other information that someone decides. The instinct that
calls for a suppression of information that people want and that will be
around in other forms is basically the instinct that distrusts democracy
-and your ability to decide.
In the end, as we began, the surest way to deal with any
problem is not through this kind of calculated ignorance but through
openness and candor that will keep us free. To talk to us about why we need
that openness and candor in polling as elsewhere, we have with us tonight
one of the nation's most respected pollsters, Mr. Mervin Field.
Welcome to The Advocates, Mr. Field. It's nice to have
you with us.
Mr. Field is the founder and director of the highly
respected California Poll. Tell me, Mr. Field, what happened before polls?
Were people not interested in which candidate was going to win?
No, I think you have to go back 25 centuries. It was
the Greeks who decided to start things by balloting. I think when that first
election was scheduled in Athens, maybe 2500 years ago, that somebody wanted
to find out how the campaign was going. And I'm sure that the Delphic oracle
at that time had a lot of requests.
Well, tell me, Mr. Field, do you simply make polls for
the media, or, in fact, are most polls made even on candidates and elections
for private and other sources?
No, our California Poll is supported by media. It's a
syndicated service. We do not work for candidates whose races we are
covering in the California Poll.
And what would happen if the media could not publish a
highly respected poll like yours or Mr. Gallup's or Mr. Harris'? Would other
people take polls just the same?
And what would happen to those polls? They'd be talked
about . . .
What would happen? They'd be talked about, and they
would be given more credence than, I think, the published polls because the
public has come to accept published polls. They're part of the process, and
when they create a vacuum, I think the unstructured, the unsupported, the
unsystematic poll then gets more credence than it should. And one of the
problems in polling is to get more and more disclosure. There's no mystery
in polling. It's a very simple process. There are tens of thousands of
people who are now coming out of school who know how to take polls, and it's
nothing mysterious at all. It's just a very systematic way of going about
measuring public opinion.
Now, we've heard a lot of talk about polls being
inaccurate. But, in fact, how does the well run poll do it? Is that more
accurate or less accurate than odds or random rumor?
Well, there's always a problem in trying to assess
accuracy. When we talk about accuracy, all that we can say is that if we
take a poll as a measurement at one point in time and we go through a
certain system, the kind of system that other professional people can
replicate by going through it, then we feel that we have an accurate poll.
One of the misconceptions about polls is that people believe, and this is a
widespread belief, that a poll taken at one point in time, no matter whether
it's one week, one month, one year before, is an automatic forecast of how
elections are going to turn out. And what they fail to see is that there's
always a lot of interventions. The campaign is going on. The candidates are
trying to sway public opinion. Their campaign managers are turning out the
right voters and avoiding so-called wrong voters. New events come in. A
campaign is a very dynamic process.
And do you think the concept of a poll and what it
measures and what it's good for and what it's not good for when it's well
done is simply too difficult a concept for the person reading a newspaper to
fro. This is the kind of information they want, and
what we have found is that there is one segment of the public that is
concerned about issues, but a large majority of the public are interested in
personalities. And whether you agree with it or not, we are a nation of
score keepers. People want to know who's ahead. And the reason they want to
know who's ahead, not so much to bet or to satisfy their own perception, but
rather to see is my view of that candidate shared by other people. Now, 100
years ago, when we all lived in communities where we could go to our
neighbors or go to our ward boss, we could get some reaffirmation of that.
But now, we are a very mobile society. We need some idea, some
reaffirmation, do other people perceive the candidate the way I feel. And
so, when they these poll reports and they see them change, they are getting
a measure of public opinion research in action, public opinion research
which has value in many other ways other than candidate polls.
All right, gentlemen, let's turn to Mr. Fisher, who has
got some questions for you now, Mr. Field.
Thank you. Mr. Field, let's see if we can narrow the
area of disagreement. You agree that there are problems about publishing
Problems in what way?
Impact on the electoral process. For example, that they
make it more difficult for a new politician to get into the race.
I don't think so. There are some politicians who think
that it might be difficult.
Suppose I read to you the sentence, "There is
demonstrable validity to the charge that polls limit the opportunity of
comparative newcomers to the political field." Would you recognize what you
That's a statement I made some time ago, yes.
Would you agree with this statement you made?
Now it's less viable. We've had many newcomers who were
behind in the polls, have found new ways to finance campaigns.
How long ago did you write this statement?
I think I first wrote it in '52 or '54.
Most recently wrote it this year?
I don't recall. I may have repeated it.
Do you agree that polls often cut off campaign
contributions to a candidate - publishing a poll - as we've heard from Mr.
What they cut off are the money from the fat cats, the
people who are self-appointed financial people for candidates, and what we
have now demonstrated, particularly in the Senator McGovern campaign, new
ways of financing. Senator McGovern probably was hurt by the polls, but he
was able to organize campaign financing. His campaign ended in the black.
They had a target of $25 million and were able to raise it exclusive of the
traditional fat cat contributors.
They raised less than half or about half as much as the
Their target was $25 million.
Oh, knowing what the polls were, they may have had a
realistic target. You do agree that it does tend to cut off the money,
though, for whatever reason, and that does limit a candidate's
It cuts it off to the ignorant financial people who do
not know how to read polls, the unsophisticated people.
Do you think the public are likely to misunderstand
candidate preference polls during the last part of the campaign?
Some may misunderstand.
You think some may. Would you recognize the statement
that "in recent years, nearly all responsible opinion polling organizations
have tried to emphasize that their political polls are not predictions.
Despite these efforts, nearly everybody (may I repeat that, nearly
everybody), and that includes pollsters themselves, in unguarded moments,
view a late opinion poll as a forecast of events to come." Do you agree that
most people misread the polls?
That's my statement.
You agree they misunderstand the polls.
No, I don't. No, I . . .
You state that most people construe a late poll in the
campaign as a prediction, and that is not the fact, and you say they
misunderstand it. Do I understand you correctly?
Well, let's let him finish his answer, if we might, Mr.
I say there is widespread misunderstanding of what a
poll does, and all segments of the public, general public, politicians,
lawyers . . .
I agree. Wide misunderstanding of what that kind of
Now, what proportion of all polls that are broadcast on
radio or TV or published in newspapers would you say are vouched for by
reputable professional organizations, professionals, and are described in
enough detail to permit them to assess the validity. What proportion of
published polls are reputable and described in enough detail to permit an
assessment of their validity.
I think as a result of my activities and activities of
other people in a professional society that now the Harris Poll, the Gallup
Poll, the California Poll, the Minnesota Poll, the Iowa Poll, and I'd say
all the public polls that have had a demonstrated reputation, who have been
continuously polling now do disclose ...
What percent of the newspaper polls, television polls,
or company…When you give a report . . .
Mr. Fisher, hold on just a second. Do you have an
opinion as to…
Yes, the majority of the polls now do disclose.
Now, do disclose. When you issue a poll to one of your
subscribers, one of your media, you attach an explanation to that poll, to
each sheet of that poll or each report, which includes a vast amount of
And you do that why?
Because we want to disclose. We want the public to
understand. We want the media to understand just what's in the polls.
And do the media attach such a report to each TV
They do. They do describe the sample size. They do
describe how the poll was taken.
All the things that you suggest ought to be
I heard 39 minutes of the Today show on election
morning. I don't believe I heard a single reference to sample size or method
of polling. I just heard overwhelming…
I didn't hear the show. I didn't hear the show. I might
also say that we deposit all our polling data at two repositories, all our
poll data, so that anybody can get access to questionnaires, IBM cards,
full access to it, and there is now an increased understanding of polls just
by those efforts.
I agree with you that your polling is fine. What I'm
concerned is what the media do to your polls after they are published
because . . .
Gentlemen, you're going to have an opportunity to
explore that at a little greater length. Let's go back to Mr. Miller for an
additional question. Mr. Miller.
Mr. Field, I suppose the question is whether the public
would have more information and be better off if the highly respected polls
were no longer published by the major media. What would be the state of
affairs if they were not? Would there be more or less
I think there would be more misunderstanding. I think
if you denied the public the polls that had acceptance, the bad polls would
drive out the good polls. And if you drive out the good polls, they're just
going to be supplanted by bad polls.
All right, let's go back to Mr. Fisher. Mr.
You've suggested that the best guarantee against abuse
is for self-restraint on the part of editors, and editors not to publish any
polls that are not vouched for by a reputable professional, and not
described in enough detail to permit one to assess their validity. Now, if
they shouldn't publish those polls, should the viewer on TV, can he judge
the Philadelphia Poll?
What's the Philadelphia Poll?
You suggested . . .
The poll that Congressman Green referred to.
There was a newspaper poll that came out that was all
over the front page . . .
You see, the California Poll, the Gallup Poll, the
Harris Poll are reliable professional polls. Anybody that wants to have
access to how they took the sample, to examine sample size, sampling error,
tabulation schemes, can do that. They are vouched for because they've been
in business and they have acceptance.
How does your publication of that data improve the
selection of the quality of the candidates we have in public office.
Because the public wants to know how the candidate is
faring. He wants to see how the candidate manages his campaign in success
and adversity. And that's how it improves it, because most people do not
vote according to issues, they vote according to personality, and they want
to see this personality react to good news and to bad news.
Gentlemen, I have to interrupt at this point. I'm
sorry, Mr. Fisher. Mr. Field, thanks very much for being with us on The
All right, Mr. Miller, go ahead.
Not only does tonight's proposal drive out the good
polls for the bad but it also, of course, violates our fundamental ideas of
freedom of speech. To talk to us about that, I've asked to join us tonight
Senator Charles Goodell.
Welcome to The Advocates, Senator Goodell. It's nice to
have you with us.
Senator, you had an unhappy experience with polls in
the 1970 Senate race, didn't you?
Yes, my experience, I think, was very much like the one
that Bill Green described earlier. I was reported as running third in a
three-way contest. It did have some adverse effect, I think, upon my
campaign. It's questionable exactly what the impact was. There were many
other things involved in the campaign. But I certainly feel that the public
had a right to know what the situation was. In fact, I think they even have
a greater right when there's a three-way candidacy for one office. If the
Number One candidate in the polls is someone that the people who are
favoring Number Three don't like, they ought to switch to Number Two if they
feel that way.
Let me ask you what is the fundamental question in
tonight's discussion. I suppose the same litany could have been made about
all sorts of information, that people may misunderstand and that some of it
may be inaccurate, as has been made tonight. The question is do we deal with
those problems by newspapers suppressing that information and not publishing
Well, I certainly think that if you start saying that
people don't understand polls because they're too complicated you are
opening up to some very dangerous things. I do not think that the press
should try to restrain the public from having information that they think is
controversial or complex. If the press wishes to print a poll and they think
it's inaccurate, I think they can put in an evaluation themselves. They can
talk about it in the editorial columns. I think others can do that. If you
exclude the press from publishing particularly the reputable polls that are
approved by the professional standards, you're going to have a lot of
garbage. Candidates are going to be going out handing out pamphlets of their
own polls. They'd be totally unreliable.
Well, let me ask you, on many polls, should we perhaps
take a totally contrary approach; that is, men in public office, the
President, men in the Senate, use polls without disclosing how the poll was
made. They use it to frame their own judgments about candidates and issues,
politics by secrecy. In fact, instead of preventing publication, shouldn't
we require publication of polls that are used to make judgments about
candidates or other things.
Yes, I believe that. There have been many instances
where public officials have claimed that they were taking a given position
on an issue because their polls indicated that's what the American people
wanted, or the people of their district wanted. When any reference of that
nature is made, I think they should come forward and give the full
information about how that poll was taken. Congressmen are well known, for
instance, in saying that I sent out my newsletters, and these are the polls
I got. Well, of course, they have a very limited number of people receiving
You've been in the Senate, you've been in the Congress,
you've wrestled with this issue, you've been hurt by a poll, you've seen
good polls and bad polls, Do we deal with this problem of complexity, of
complicated information, of some misunderstanding, do we deal with it by
cutting it off? Or do we get more education, more learning, more
Well, having been hurt by publication of polls,
according to a number of people, at least, I must say that I think the
people in New York State had a right to information about where the various
candidates stood in that particular campaign. I happen to believe that the
particular straw vote involved was not accurate. There were other polls
available that were accurate and should have been published along with the
Daily News straw vote, and I think the people have a right to know it even
if it's inaccurate.
Of course, the essence of the free speech value, which
we so often forget, is that everyone can be in favor of accurate
information, but what do we also have to defend -the right to publish . .
One of the great arguments in any election campaign is
the whole issue of what is accurate.
Tell me, Senator Goodell, in terms of more information,
what kind of more information would you require. You'd want people to know
what about polls in order to evaluate them.
Well, the pollsters have worked out the standards for
professional polls that are considered to be reliable. They involve the
training of the individuals who are asking the questions, the choice of the
sampling, the size of the sampling, the time of the sampling, a whole
variety of things of that nature. If that kind of information is given,
people can perhaps begin to evaluate polls better. I think the answer is to
educate them on how to evaluate them.
And is there any virtue . . .
Gentlemen, I'm going to have to break in. I'm sorry,
Mr. Miller, but we'll get back to you. Senator Goodell, Mr. Fisher has got
some questions for you.
Senator Goodell, do you agree that, the publication of
polls causes some harm and may do some benefit, we're weighing that? Do you
agree that publishing a head-to-head may cause harm?
I agree that publishing almost any information can
And you went so far as to draft legislation proposing
the prohibition on publishing head-to-head candidate polls during a certain
period before the campaign. You considered such legislation, did you
I considered it and rejected it. My proposal was that
they be required to give the full information on how the poll was taken, so
the people could evaluate it.
At the time, you told in testimony why you rejected the
proposal. Do you remember the reasons you gave why you rejected it?
I rejected it for a variety of reasons, including that
I didn't think it was constitutional..
That was the first reason you gave, that it was
not constitutional to pass a law prohibiting it, and I agree with you on
that. I'm not disagreeing at all. The second one was you thought the voter
had a right to vote, irrationally, if he wanted to, on the data that might
be there. Do you believe that voters ought to vote on the basis of polls as
to who's ahead or who's behind?
I, personally, would hope that under normal
circumstances they would not. In every campaign I've run, I've spent the
time telling people the issues that I think they ought to consider
important, and I would tell them that I don't think polls are important in
most cases. However, I think if they disbelieve me and they think an issue
is very important, they should have that information to vote on.
If you were an editor, and you thought that to publish
polls was going to divert the campaign from the issues and was going to
cause voters to vote irrationally, would you publish such figures?
Yes. I don't think I would feel that I was in a
position to make a judgment about whether people were going to vote
irrationally on the basis of information that I had given them.
You have $5000, Senator Goodell, to spend on a
particular campaign, on a reporter to go in. Will you commission someone to
go out and create some news by asking people questions which will be news
when they answer it, or would you try to commission someone to write a piece
about the issues? This is an allocation of resources as to whether that last
$5000 is better spent helping the public understand the candidates or is it
better spent by the media producing figures that will lead to irrational
votes, that you and I both consider irrational.
I think that's an irrational choice. I don't think it's
a choice that's normally made by a newspaper. . .
A TV station has two minutes to discuss something. They
can discuss substantive question; they can discuss the latest head-to-head
poll. Which would you, as a producer or editor of a television program, put
I would want them to discuss the issue, but I certainly
wouldn't come down on either side to say that they cannot discuss the issue
or the information about the issue or the poll. . .
Let's let him finish. Let's let him finish.
I certainly wouldn't tell them they shouldn't publish
information about a poll that people are interested in.
No one here has suggested that they be prohibited or
barred. . .
I apologize. I would not suggest that you were
saying that. . .
We are suggesting that editors exercise their
responsibility. Now, suppose an editor found that his readers would be
fascinated by a survey of the private sex lives of public officials. Would
you think that a responsibly journalist might decide, as most newspapers
have, that, despite the interest in reading about the sex lives of
candidates, they will refrain from commissioning investigators to gather
I don't think it's a responsible question. I think
basically the responsible issues . . .
The freedom of information?
Let him finish, Mr. Fisher. Go ahead, Senator.
…the responsible issues should be published. I think
the people should have them, the information about them. I don't think that
we should ask newspapers to restrain from publishing any information they
think is important to the judgment of people. . .
We are giving . . .
It's quite possible, as a matter of fact, given your
question, that people would be very interested in the sex life of the
I take that by hypothesis, and I say that I believe
that the media. The question is responsible. I think a responsible editor,
wisely restrains from commissioning studies of the sex habits, and I think a
wise editor would refrain from commissioning this particular kind of poll
because in each case it would divert the attention. Would you commission
Well, those of us who are in public life certainly
approve of your position, I think. We would prefer that they not do that.
And I think there is some question of privacy in public officials’ lives.
And I would not commission a poll on the sex life of a man in public
Even though it's of interest, even though a lot of
voters would like to do it, you would not regard that as infringement of the
freedom of press for an editor to' responsibly decide . . .
I would not consider it to be an infringement of the
freedom of press even if you restrained them.
Clearly not. Well how about a poll on the innocence or
guilt of a man pending trial. Would you think the newspapers ought to
I don't see that that's related. There, we're
protecting the rights of an individual who has not been adjudged guilty yet.
In this instance . . .
How about a civil trial?
. . . we're talking about a public election where
people need information. The public should not judge the guilt or innocence
of a man who's coming to trial. They should judge who is the better
candidate running for public office. They should have the information.
The theory in each case is that they should vote on the
merits and not how their colleagues think. We don't want the jurors to vote
on the public opinion poll on who's going to win a traffic accident. Do you
think the election of Congressman or Senator is less important than
litigating a civil traffic accident?
Certainly not less important…
…but also I must say it's the kind of a situation where
it's perfectly proper and relevant for the people to know how the election
is going and how other people feel about it.
All right, gentlemen, let's go back to Mr. Miller. One
additional question, Mr. Miller.
Well, Senator Goodell, we do not have polls on the sex
lives of public officials today, perhaps because of the difficulties of a
random sample. We do not have . . .
Somebody out there likes that, Mr. Miller. How about a
The question is are we talking about the same problem.
For 25 or 30 years now we've had highly respected polls in this area,
information the public wants. Does a responsible editor cut off information
the public has had, wants, and serves a purpose?
I don't believe it does, and I believe the evidence is
that most of the reputable polls now are very accurate as to the situation
at the time they're taken. They are reliable information for people to judge
by if they wish to use this in their voting patterns. And if you restrained
yourself from publishing those good polls, they would be replaced by all
sorts of abuses that are not really legitimate polls.
And is there any virtue
I'm sorry, I'm going to have to interrupt, Mr. Miller,
and the moderator is going to ask a question. Senator, supposing there was
some way that we could inform people by two o'clock in the afternoon on
election day what the vote was as of that moment so as to help the people
who hadn't voted make up their minds, would you favor that kind of procedure
and the publication of that information?
It is somewhat a different issue. You're talking about
the ability now to put a small amount of information into a computer,
perhaps, and predict how the election is going to come out.
I'm saying that at two o'clock in the afternoon, if we
could do it, should we publish what the results of the election are at that
time for the benefit of the people who haven't voted yet?
I would favor it if it's information that they wish to
hear and the people think it's reliable. That's a judgment everyone would .
It's certainly a lot more reliable than a poll, isn't
it, if we're actually counting . . .
Not necessarily. If you publish at two o'clock in the
afternoon when the polls are still open, you're probably doing it on a very
small sample from a very few districts.
Mr. Fisher, I'm sorry I can't let you ask another
question. I'm sorry I took your question away from you. Senator, thanks very
much for begin with us on The Advocates.
All right, Mr. Miller, your closing argument,
I suppose what we need are men who are intelligent
enough to understand the polls and courageous enough to disregard them. What
we need, with regard to the polls though, is not less disclosure but more
disclosure, not a cutting back of information but a feeding in of more
information. Listen to what you've heard tonight, that newspaper editors
should not publish legitimate information that people are concerned with
because they divert the issues of the campaign. Who decides that the issues
of the campaign are diverted? Roger Fisher? Who will decide next week?
You should beware of people who wish to protect you from
yourselves, who wish to protect you from gathering information you can't
handle. They succeed only too well. In the end, the men who have been
redeemed in this area are the men who have trusted and not been suspicious,
not the men who have been fearful of information but who have welcomed it.
Thank you, Mr. Miller. All right, Mr. Fisher, your
The issue is one of self-restraint, not freedom of the
press. We're asking that the editors decide what is legitimate and what is
not. We're asking that they decide with this as they do today in not
publishing the private sex lives of public officials. It's not a question of
a poll. It's a question of choosing in your responsible judgment not to
publish information which, Mr. Field says, almost everybody misunderstands.
They think a poll is a prediction.
What are the benefits we've been offered? Mr. Field says,
well they're interested, they like it, it's good entertainment, it competes
with entertainment programs, they love it. And he also says that it helps
them judge a candidate. They can see how a candidate squirms as the people
misinterpret the polls that are published and they can judge his stability.
That's some benefit, I suppose. The cost, no one disputes. Senator Goodell
has written about the destruction of the debate of the campaign when issues
are diverted. It causes people to think about who's ahead and not who will
make a good officer. It's exactly the wrong question.
We are not talking about punishing or prohibiting. We are
saying that editors here as elsewhere should exercise self-restraint, I hope
tonight you'll vote for self-restraint. Thank you.
Thank you very much, gentlemen. We've now been debating
the question of public opinion polls for almost an hour, and it's time for
you here with us in Faneuil Hall in Boston and for those of you at home to
let us know how you feel about this issue. Those of you with us in Boston
have had ballots distributed to you. Please mark them, indicating your
preference and leave them in the ballot boxes which you will find at the
door as you leave Faneuil Hall. And for the thousands of you watching on
your television sets at home, it's time now for you, too, to get in on the
act and let us know how you feel about this issue. Should the news media
refuse to publish candidate preference polls? How do you feel? Send us a
letter or post card with your yes or no vote and mail it to The Advocates,
Box 1972, Boston 02134. What do you think is important? We'll tabulate your
votes, and we'll make them known to the members of Congress, to our leaders
in government, and to leading newspapers in cities and towns all across the
country. So let us know, send us your vote. That address again: The
Advocates, Box 1972, Boston 02134.
And now with thanks to our two very able advocates and to
their distinguished witnesses and to you, our audience, we conclude
Next week The Advocates travels to Madison, Wisconsin
to join an important symposium on prison reform. Governors, ex-prisoners,
and prison wardens will join us then when we debate the question, would we
be safer from crime if we closed down most of our prisons.
The Advocates, as a program, takes no position on the
issues debated tonight. Our job is to help you understand both sides more
This program was recorded.