At the time this article was written
Richard Dresner was Chairman of Dresner, Wickers and Associates of New York. This
article is based on remarks to a seminar organized by the Library of Parliament
on November 21, 1996.
Much has been written on various aspects
of the public opinion polling that finds its way into the headlines of daily
newspapers and television. But less is known about the kind of private polling
done for political candidates. This article sheds some light on this facet of
There are certain distinctions between
public opinion polling and strategic polling for politicians. For many years I
was a public opinion pollster. I did work for ABC, for Louis Harris, and for
various government departments. The purpose of such polling is to compile
"objective" information about what people are thinking about a
specific topic. Then I became more of a political junkie and began to work as a
strategist for political candidates.
When I run a poll it is to discover what
should be in the commercial or the direct mail or the speech. We are not
concerned as much about objectivity as we are about winning and of course, we
do not make our polls public since that would reveal our strategy.
As candidate pollsters we do not ask people
what is the number one issue facing the country. If we are any good, we know
that already. We are hired to put together commercials, to do direct mailing
and to advise on what messages are having the most effect in a campaign. We try
to find out what issues our opponents are going to bring to the campaign and
what issues we should bring.
We give people a list of ten or twelve items
and ask which of these would influence their vote. (We do the same thing in
jury selection work. We present several versions of the same case to
prospective jurors and by asking five or six questions we determine which is
the strongest and most likely to influence a group of jurors)
When I worked for Tom Foley, former Speaker
of the House of Representatives he said: "Everyone tells me I am going to
lose my seat. I am not hiring you to tell me that." Everyone agreed that
he did not vote with his district. In fact, he voted against them on
everything. He said: "Your job is to find an area where I have something
in common with my constituency." We discovered that he did not care about
gun control and he became a quick member of the National Rifle Association Hall
of Fame. That is what saved his seat in the 1978 election. It was the NRA
sending out letters to conservative voters, saying do not believe Tom Foley is
a liberal, we support him, so that means he is a conservative. He used polling
to find the one area where he would have something in common with conservatives
in his district.
Candidate polls are now fairly universal
since in most countries the voters are looking for someone who actually agrees
with them, or is going to do something they like. What strategic polling does
is explain to the candidate or the party exactly where the common ground is,
and what the best common language might be. So if you are a candidate or a
party, hopefully you are doing polling that is extremely detailed.
Everybody knows that in the last US
election, American voters were concerned about taxes. You do not have to poll
to find out that people care about taxes. What you use polling for is to
determine how they react to a proposal to reduce taxes by fifteen percent. Is
this believable or not? I guess Bob Dole did not pay enough attention to his
internal polling and neither did his campaign people. The Clinton
administration polled every day and polled every conceivable topic and theme
even the notion of a "bridge for the 21st century". They came up with
five or six lines, and they ran with them.
When I worked for Boris Yeltsin, we took
some of his endless speeches and played them before small audiences to find out
what would be popular and what would not. In between being comatose, the
audiences would respond every so often to a line. We literally copied that line
and then tested it in surveys. We then went back to President Yeltsin and his
campaign and said, "You say X,Y and Z." The people really liked X, so
it would be great if you would repeat it." So the polling was very
detailed and very task-oriented.
A question that came up in Russia was
"what role should the wife and daughter play publicly." We had a list
of 10 or 15 different things they might do. We tested each and everyone of
them. Then the election team took it back to the campaign and said here are the
2 or 3 things Mrs. Yeltsin ought to be involved with. It was no accident that
she set up charitable organisations. It was no accident that she talked about
home life almost from the perspective of an average housewife. Russians saw her
that way. But everyone knows, she is no average housewife.
Another issue with simulated polling is
about those evil "push questions". I do not know why anyone would
consider them evil. What they basically do is say to voters if you learned that
candidate X was for the death penalty and candidate Y was against who would you
vote for. People will tell you whether they will switch their votes based on
those things. Again you go back to the campaign and say "if we use this
issue, it is going to mean a lot of votes.
It took us two years to convince Pete Wilson
to champion proposition 187, the immigration issue in California, and then he
did it only when it became a fiscal issue. When you could say it cost the State
of California 3 billion dollars a year to provide services to illegal aliens,
only then was he willing to use it. It was an issue that took on a life of its
own. We never planned a referendum or an initiative on this. We are going to use
it to make certain executive changes and to espouse the issue. It became the
key dividing issue in California politics. Even when we started 30 points
behind, we realised as long as Kathleen Brown was on the other side of the
immigration issue, there was no way we were going to lose. It was just a matter
of time to catch up. So what polling does, it gives the candidate a sense of
saliency on issues. This issue was going to produce a much tougher reaction
than some other issue even if the other issue is more popular.
Of course sometimes things do not work out.
We did some work in Romania last year. We were using push questions and doing a
tremendous amount of polling. Then we started using telephone polling in some
of the bigger cities. The push question so upset the opposition that they
raided and trashed our phone banks and accused us of being terrorists. The
polling organisations that we had hired in Romania decided that they were not
going to ask any more push questions because people might come and trash their
Polling allows a voter to say to the
politicians, " Tell me about something that you are going to do that I
have not been thinking of every single day and that I am going to like when I
hear it." In California it was affirmative action, immigration and now
even putting money into education. In Russia it was the notion that government
was actually going to pay back salaries to government employees, to raise
pensions and to ease the impact of inflation on pensions. People wanted an end
to the war in Chechnya, one way or another.
I think that when you use polls for other
purposes or start publishing polls, then other factors come into play. One is
money, especially at the initial stages of a campaign. When a poll shows that
someone is trailing by 34 points, nobody wants to give money anymore. In the
last gubernatorial election in California there was no doubt in our minds that
election was going to be close, no matter what happened. Yet we had to struggle
for almost a year and a half in raising money because every poll that came out
showed us between 17 and 25 points behind. This becomes crucial when private
polls are published early on in campaigns. You have to ask " Are they
using this to raise money or not".
I tell my clients if we have enough money to
get our message out, it does not matter a great deal how much the other guy
spends. So, adequate money is important.
While money is important there are certain
misconceptions about the importance of money in American elections. The
landscape of American politics is littered with multi-millionaires (Ross Perot
and Steve Forbes being a couple of recent examples) who thought that if they
had enough money they could buy victory or at least make a respectable showing.
But just spending money does not work. You have to tap into something that is
of concern to people.
There are certain exceptions; circumstances
where money is decisive. In the last election the AFL CIO targeted some 30
Congressmen and spent 60 million dollars in these campaigns. They managed to
defeat 14 incumbents which is phenomenal by American standards. I once worked
for a very large company that pumped 7-10 million dollars into a limited number
of congressional races in a ten-day period. We won 15 of the 18 seats. So you
have to select the races very carefully but in those select cases money can
make a difference.
One final point about money. It can have
negative effects as well if people become concerned about how much you are
spending or where the money is coming from. When you are in an election race
you have to balance many factors including the personality of the candidate,
policies, party allegiance and money. In any given election all of these
factors will be in play and one may turn out to be the swing factor. In another
election it may be a different factor. If I was allowed to ask only one
question to determine how a person would vote in a given election it would
still be "What party do you support?"
Another thing that has happened in the name
of polling is illustrated by a study done in the 1970s at Yale University. It
found that calling people and polling them increased voter turnout. They took
people who, based on voting records, had not voted in two or three years and
just called them to ask: "who are you voting for in the mayoralty election
in New Haven?". The result was something like a 15 or 20 percent increase
So now people who are trying to persuade
voters have created all kinds of new uses for telephone surveys. If they get a
call on behalf of the XYZ campaign, wanting to know 5 things or 10 things,
people just ignore it completely. But what if you get a call from somebody who
says, I am with the XYZ’s polling company and we would like to know who you are
voting for? Now they have targeted you because you made the mistake of telling
them you are an undecided voter.
If you are a decided voter one way or
another, they then give you what is a thinly disguised message. "Would you
still be undecided if you know that so and so had not paid taxes in 20 years.
Or, they are against welfare reform, or they voted to raise your taxes. You
answer these push questions and while you think you are part of a survey, what
is really happening is that you are being persuaded to vote a certain way.
Without market research in this world, it is
very tough to run a campaign. There are examples of candidates who do it but
usually they are sitting on the crest of issues and transcend the process .
Another trick is used by single issue
operations. You get a call in advance that targeted you as an undecided voter
or somebody that is interested in the environment or in crime. The next thing
you know, you are getting a call from the police chief’s association or the
Sierra Club or somebody like that, delivering a message that you have been
pre-screened for. I think that is not an ethical use of polling.
I would not want to leave the impression
that strategic polling is an exact science. It is still imprecise and not
everything works. Sometimes you go out and do all the surveys and spend 3
million dollars of someone’s money and nothing happens. That is my worse
nightmare but it happens. Then I wonder why anyone is still speaking to me.
Maybe we missed the context, maybe the other side was smarter, maybe we underestimated
the intelligence of the voter. There is a limit on how far you can take the
people on the issues. The best strategic polling can do is give you an idea of
where the salient points are and how to best present them.