In the closing weeks of the 1972 presidential election, I was a
witness to a bizarre attempt by the Nixon campaign to pressure the
Harris Poll, in order to influence our published numbers. Chuck
Colson, one of the Nixon aides who later served time in prison
because of Watergate, called to tell us that the peace negotiations
with North Vietnam were at a very critical stage. The North
Vietnamese, he assured us, were following the polls very closely. If
they saw Nixon's lead slipping, they would probably wait on the
possibility of a McGovern victory. If Nixon maintained his big lead,
there was a good chance they would agree to peace terms with Nixon
before the election. Tough pressure on a pollster. Fortunately,
Nixon's lead did hold up, so we didn't have to feel guilty for his
failure to get a peace treaty before the election.

Unfortunately, attempts to manipulate and corrupt the polls are now
a serious worldwide phenomenon. The spread of democracy and free
elections around the world has brought to power a new generation of
political leaders who are learning a painful truth: What democracy
giveth it can also take away. Those who win by the ballot box can
also lose by it, and freely elected governments are often defeated in
the next free election. Because of this, some of them are tempted to
tilt the electoral playing field, to manipulate the press, to make
elections less free and fair, and, on occasion, to stuff ballot boxes
and steal elections.

However, it is harder for incumbents to steal elections if there are
honest, accurate, and publicly reported pre-election opinion polls
and exit polls that show someone else to be well ahead. One
surprising and alarming new trend is that in many countries,
governments, politicians, business interests, and even the media are
using their power to manipulate and suppress the publication of
honest opinion polls. To their surprise, many pollsters find
themselves in the unexpected role of champions of civil rights and
bastions of democracy. This also puts honest pollsters at great risk
from those who want to corrupt the political process. Those who play
along get rich; those who don't may get badly hurt. The pressures
they face make Richard Nixon's earlier attempts at control look like
softball.

As an example, consider Mexico. On a recent visit there, I discussed
this problem with several potential presidential candidates, senior
members of the three main political parties, a senator, two
governors, pollsters, and two very influential journalists. Most
confirmed, and none denied, that all of the following have occurred
within the last year:

* The media quoted poll results that were wildly inaccurate, either
because the numbers were changed or because the polls had never
actually been conducted. Some poll clients are willing to pay
handsomely for phony polls.

* Honest, independent poll findings were suppressed by the media
because they displeased the powerful.

* Polling firms who did not provide their clients with poll numbers
they liked (and who were unwilling to change real numbers to
fictional ones) were sometimes not paid.

* Contracts for multiple polls--for both the media and
politicians--with honest polling firms were canceled because the
clients did not like the numbers shown in the early polls and the
polling firms would not change them.

* Some courageous media executives suffered (for example, one lost
his job) because they had published honest polls.

This is not just hearsay; I have seen some of the canceled contracts
and the polls that were suppressed. Nor are these attempts to mislead
the public with phony polls and the censorship of honest ones unique
to Mexico--I have heard similar reports from many other countries,
most recently from sources I trust in Taiwan, the Philippines, and
Ecuador.

In fairness I should note that similar abuses have occurred in
Western countries in the recent past. In a famous case in the 1970s,
a French cabinet minister persuaded a leading French polling firm to
reduce the number of people who were reported to be hostile to
immigrants. In Britain in the early 1970s, Labour leader Harold Wilson
was guilty of a mean-spirited intervention that affected me
personally. My firm, which I had recently sold to Louis Harris and
Associates, had launched what I believe was the first regular poll
jointly sponsored by a television channel and a newspaper (the
CBS/New York Times poll followed later). Our ITN/Times (of London)
poll was a great success; both our clients were delighted with it.
But Harold Wilson was furious. How dare ITN, the news channel he
trusted much more than the BBC, get into bed with the Times (then
edited by William Rees-Mogg, whose editorials encouraged the
soon-to-follow departure of Roy Jenkins and the "Gang of Four" from
Labour)! Wilson sent Labour mp Gerald Kaufman to persuade the ITN
management to end the affair; they refused. Wilson then nobbled two
ITN board members who persuaded the board to vote to end it. As a
result, a high quality, well-regarded--and truly independent--regular
poll bit the dust.

After the 1972 election, President Nixon asked two of his top aides
to "influence" the Gallup and Harris Polls, particularly their data
on Vietnam and Watergate. Fortunately, there is no evidence that
Nixon succeeded. Both polls showed the public swinging strongly
against Nixon on both issues, and these polls may well have had some
influence on critical congressional votes against him.

In 1994, Frank Luntz, a well-known Republican consultant, persuaded
the American media to run stories saying that his polls found 60
percent of the public supporting every element of the Contract with
America. Last June, long after the election, it emerged that there
were no such polls, and Luntz was formally censored by the American
Association of Public Opinion Research. This rebuke will probably
have no impact on him or the use of his services by political
candidates.

In the United States, Canada, and Europe, there is less need for
concern about the manipulation of election-related polls than there
is in the case of new democracies. But there should be much concern
about the increasing use of "advocacy polls" in the West. Trade
associations, special interests, lobbying groups, and even individual
companies frequently commission polls designed not to inform but to
persuade. The questions are often tailored to produce the answers
clients want, and the replies are cherry-picked with only the most
supportive data being released.

For those who consider opinion polls to be a disagreeable newcomer
to the political process, who believe they have a malign effect, or
who dismiss them as frivolous or insignificant, all this may be
treated as of no consequence. But the truth is that polls now play an
important part in the political process--otherwise people would not
try to manipulate them. Contrary to much popular belief, there is
firm evidence that polls do not cause a bandwagon effect, with voters
rushing to support the front-runner. But they do influence the
political agenda, the financial support candidates can amass, the
media coverage of issues and candidates, and votes in the relevant
legislature. A president or prime minister with high poll ratings can
influence the legislature much more readily than one with low
ratings. Members of Congress are avid poll readers. Indeed, a
worldwide review suggests that the frequent publication of
independent polls, not controlled or influenced by the government or
powerful interests, makes a valuable contribution to the democratic
process. Where corrupt polls mislead, good independent polls inform.

It is no surprise, then, that military dictatorships, communist
governments, and other authoritarian regimes have never allowed free
independent political polls in their countries. The truth would hurt
them. Most real democracies allow opinion polls complete freedom
(although, to their shame, thirty countries, including France and
Italy, have banned their publication in the closing weeks or days of
electoral campaigns). As for countries that are becoming more
democratic, such as Mexico, the main defense against the corruption
of opinion polling lies with the media. If local media can expose
abuses, they can sharply reduce them. And if the media in such
countries are themselves subject to intimidation, the international
media, particularly those of the United States, can exert
considerable influence. Most governments and politicians care about
how the international media portray them. They want to be viewed as
true democrats, not as corrupt officials clinging to power through
manipulation and fraud. Local leaders and the media will repeat
reports from leading U.S. and European newspapers. The same dynamic
even works for fully fledged democracies. It was, for example, a
British, not a French, paper that exposed the above-mentioned
manipulation of the polls in France.

One thing is certain: If the media do not report the manipulation of
the polls things will get worse. And, sadly, the local media in many
newly democratic countries will be too timid to report such
manipulation unless the American media and international media make
it an issue first.

Humphrey Taylor is chairman and CEO of Louis Harris and Associates.