|
Genuine Opposition senatorial candidate John Osmeña
has beefed about his slipping ranking in a poll survey of
senatoriables in the May 14 election. He could not accept the
explanation of the polling agency that his political ad showing him
tearing up the E-Vat law had contributed to his downward plunge.
On the contrary, he said, the ad
could have generated more popular support for him, especially from
the poor burdened by the increased value-added tax.
Actor-candidates Richard Gomez
and Cesar Montano have also questioned poll survey results that they
were progressively distancing from the Magic 12 in the senatorial
contest and that there was a slim chance they could recover.
To be sure, some senatorial
hopefuls in the administration’s Team Unity were similarly in
disagreement with surveys showing the GO candidates dominating the
win column.
We do not blame them for their
grievances. They have the moral right to refute the survey results
which, to be sure, were not the product of an exact science.
In fact, two US polling agencies
could not live down their major embarrassment when they wrongly
predicted in their surveys the presidential winners in the 1936 and
1948 elections.
In 1936 the Literary Digest,
which had conducted a survey on the presidential poll announced
wrongly that Alf Landon, Republican, would win against Franklin D.
Roosevelt, Democrat. Twelve years later, most US polling agencies
made the grievous mistake of predicting the victory of Republican
candidate Thomas Dewy over Democrat Harry S. Truman.
In both instances, the error
arose largely because of the overrepresentation of the rich who were
presumed for Landon and Dewey and the underrepresentation of the
poor voters, who were for their opponents, in the sampling procedure
of the pollsters.
Since then, rapid advances have
been made in the methods and techniques of measuring the attitudes,
perspectives and preferences of people toward events and issues of
great importance.
“Public opinion polling,”
says the Funk and Wagnalls Encyclopedia, “involves procedures to
draw a representative sample of the population under study.” It
adds that if a polling agency, for instance, seeks to survey the
attitudes of all adults in a country, “it should draw up a list of
the entire population and select at random a sample to be
surveyed.”
People have questioned how a
pollster can show the preference of an entire voting population for
certain candidates based only on a sampling of a few hundreds of
people.
Funk and Wagnalls explains that
“when proper techniques are used and the sample is large
enough—1,000 to 1,500 people—the results obtained are likely to
be very close to the results one would get if the entire population
were surveyed.”
It says that “if the 60 percent
of the sample say it approves of the president’s policies,
statistical theory shows that if the entire population were
surveyed, the probability is 95 percent that between 58 percent and
62 percent of the people would express the same approval as the
sample.”
“The criterion of excellence in
a sample is representativeness, not size,” it concludes.
Opinion polling is generally
accepted worldwide as a scientific undertaking in gauging the
attitudes of people toward events, people, circumstance and even
products. In the local context, the Social Weather Stations and
Pulse Asia have been actively engaged in public opinion polling,
whether in politics, business or academic research.
By their past track records, they
have acquitted themselves creditably, especially in measuring voters
preferences in an election. In past elections, SWS and Pulse Asia
forecast almost precisely not only the winning candidates but also
their rankings. For instance, in the 1998 poll, the late statesman
Blas F. Ople was predicted to be No. 8 in the win column for
senators. That was precisely his place when the official election
results were announced.
There are reasonable questions
though that need to be addressed. Would election results have turned
out differently, for example, if the poll surveys are not published
in the media? In other words, do survey results of polling companies
influence in any manner the preferences of voters, such that the
official results tend to validate the forecasts?
Some citizens have called for a
halt to the publication of election surveys by the country’s major
polling agencies. They may have a good reason to do so.
Information that some candidates
may not be doing well in the surveys could discourage voters to
support them. Most likely, they would vote for the winning bets as
predicted in the surveys.
|
