The role of surveys in the development of Democracy in LATAM
At the beginning of the last century, barely one in a hundred of the people living on the planet lived in a democracy and nowhere was there universal suffrage, only upper class men could vote, and they got their information in print. A century later, more than half the world's population lives under a more or less democratic regime in which people can elect their representatives. The spread of representative democracy over the course of the twentieth century was thanks in part to the media and the use of polls as a scientific tool to find out the opinions and attitudes of people generally, not just those of a minority.
Polls play the role of giving the voter a picture of what his or her fellow voters collectively think: without them, the voter would turn up at the polling station with blinkers on. Nobody in Brazil can say that the 57 million Brazilians who voted for Jair Bolsonaro did so in ignorance of the possible final result, which had been accurately predicted by the polls. And in the United Kingdom, the result of the Brexit referendum was not a surprise to the British themselves because most polls in the previous week were showing a likely win for leave over remain. In the United States, the polls got it wrong in three key states - Michigan, Wisconsin and Pennsylvania - which handed Donald Trump victory in the electoral college despite his defeat in the popular vote as predicted by the polls. In fact, under the electoral system of any country in Latin America, the President of the USA would be a Democrat. We should bear that in mind so as not to confuse surprising results with unforeseen results.
It is true that in periods of change the polls can be too conservative in measuring the scale of change, something that happened at elections between 2013 and 2016 in many democracies. Still, in every case where there has been a surprise result, at the next vote the pollsters have accurately predicted the outcome: I have seen it in more than twelve countries.
But polls have an even more important role: they let those in power know the opinions and attitudes of the voters throughout their term of office and not just when they put their mark on the ballot paper. Democracy is much more that voting every four or five years.
Even so, there is a cloud hanging over Latin American democracies: the tendency to ban the publication of electoral polls in the final stage of the electoral campaign, which is exactly the time when voting intentions are increasingly crystallised. The main reason for the failure of opinion polls in Panama and Chile at the last presidential elections is the very ban on the publication of polling results in the last ten and fifteen days respectively. Those are unthinkable time frames in North America and Europe, where we know that one in four voters decides how to vote in the last week. Voters are deciding how to vote later and later, just as they are deciding what to buy, where to go on holiday, what to do with their lives, later and later: it is one of the hallmarks of our fluid society today. This is what brought us in GAD3 to take a poll on the eve of the first round of voting in the Colombian presidential elections to show that the error in the predicted results of the La Paz Poll was due to changes of voting intentions in the last week, not to the methodology of published polls. The judges of the National Electoral Council of Colombia were able to see that the differences between the vote predicted one week before in the Colombian media and the final vote were down to the expected changes in intentions in the last few days.
The most famous opinion survey fiasco is the victory of Truman over Dewey against every prediction. In its subsequent analysis, Gallup discovered that it was not possible to predict the results on the basis of interviews in people's homes a month before polling day. If that was the case in 1948, imagine how much more so today when there are more mobile phones than voters. In all the reports published on the accuracy of polls one of the essential elements is how close the time information is gathered to what we in the industry call the field day, the day of votes are cast.
Panama is debating the publication of polls in the last stages of campaigning on the basis that in the last presidential elections in 2014, the polls were not able to predict the winner. Instead of allowing publication closer to the vote in the face of the reality that voting intentions are being decided ever later, especially in presidential electoral systems, the ban is being extended to twenty days. It will be the longest ban of any Latin American country. During those twenty days, politicians and pundits can have access to information kept from the voters: their own voting intentions. This doesn't make sense.
Polls are a fundamental tool to give a voice to the population as a whole, there is nothing else that has made those in power and decision-makers take account the views of all the voters as polls do. Nothing has so empowered women, minorities or the least educated as opinion surveys. As I attempted to argue at a conference organised by the Electoral Tribunal of Panama, without surveys there is no democracy.
Add comment