EDITORIAL: Political Surveys

POLITICAL SURVEYS or public opinion polls have become both an important and notorious tool during every election. Its use or intention has been both a bane and boon, depending on which side the subject of such survey is on, or more specifically on whose client such survey is serving.

Because here in the Philippines, liars and cheats, and fake news gurus are viciously more than the norm than the exception, especially on election period, such surveys and opinion polls become even more vicious and evil, plunging the whole exercise into a deadly instrument of deceit and political machination. In short, no matter what the purpose of such poll is, credibility about them has long been lost since politics in the Philippines, or its instrumentalities, has gone from bad to worse.

But let’s talk about polls and surveys on their truer form and thus at least give us a breath of fresh air and save us from eternal damnation.

Public opinion polls are meant to be viewed as snapshots. It’s not always obvious what they are showing, as a literature on news reporting would tell us. It appears to uncover mysteries here and there, or unwillingly provides opportunities for misrepresenting the reality given in numbers.

And there are important things that we, as readers, should remember.

A national census is the ultimate poll. Polls related to election or politics, which register the opinion of the voter or would-be voter are close second.

It is admitted that comparing a poll on party popularity with most recent election result is sound practice. But when comparing one poll with earlier polls on the same topic, one should not assume that similar results make the polls more accurate because they don’t.

Polls, like snapshots, show a moment in time. By the day after the questions were asked, people have already begun to change their minds. It makes it wrong to suggest, therefore, that a poll result can predict how people will behave. An example would be an opinion poll conducted in, say January, to forecast an election result that is bound to happen in May yet of the same year.

And there’s a big difference between “per cent” and “percentage points.” The poll is accurate with 2.5 percentage points, not per cent. A political party whose support has apparently gone from a level of 16 per cent to 24 per cent has registered an increase of 8 percentage points but its support has increased 50 per cent.

The sponsor of a poll may influence the questions asked and other aspects. The story should say if the poll was paid for by a political party, lobby group or someone else with a vested interest in the outcome.

Another important item is the so-called “margin of error.” It needs to be spelled out and its significance explained. Where changes in comparative figures are smaller than the margin of error, public opinion cannot be said with any certainty to have moved up or down.

But regardless of its notorious reputation, or as an affront to it, public opinion poll had served former Camarines Sur Representative Leni Robredo its purpose. Virtually unknown, Robredo phenomenally rose -- as surveys showed -- from zero percent in popularity to become our rightful Vice President. How did she do it? By simply campaigning hard and exposing herself to more areas to prove the surveys wrong.