It has already become a classic in recent years, increasingly exaggerated in its forms. Polls are never right about what is going to happen, and not only are they not right, but they are wrong by ridiculous margins. There are just a few examples, which serve as precedents for this Sunday the 26th. In 2019, in the PASO in which Mauricio Macri was buried by an abysmal difference of 47 to 32 percent against Alberto Fernández, no consulting firm could foresee such a blow, and some even told the then president that he was one or two points ahead of his rival. In the general elections of that same year, no poll predicted that the PRO leader would narrow the difference to 48 to 40. In the PASO of 2023, the one the polls did not take into account was Milei: no one predicted that he would come out first in those primaries. Just as no one saw it coming that in the first round that followed weeks later the Peronist Sergio Massa would win, and only half of the pollsters gave Milei the winner in the runoff. And finally, this year, no consulting firm saw the difference of 14 points that occurred in the elections in the Province, and none – not a single one – was correct in predicting that La Libertad Avanza would reverse that result last Sunday and reach more than 40 points at the national level.
So, if they always screw up, why continue consuming surveys? And why don’t those responsible for making them take responsibility for the disaster? Are they still credible after so many repeated errors?
One of the precursors of today’s pollsters, the now legendary Artemio López, once told me several of the dirtiest secrets of the profession. One of Artemio’s maxims was that, when consultants measure image and not voting intention, “they let their imagination fly.” Of course, because the positive or negative image is not something that will later be ratified in an election, it is “undemonstrable.” Another warning from López was that sudden shifts of votes in the hours before an election do not exist. They are simply the pollsters’ excuse to explain why the predicted result did not occur: “There was a landslide in the last few hours.” Another revelation from Artemio was that many of his colleagues used to “sharpen their pencil” in their last polls before an election. Only there did they reveal the real measurement they managed, while in the previous weeks they “operated” in favor of the client who hired them and inflated their numbers. The Kirchners’ favorite pollster in the 2000s – renamed “Artemiópolis” by Jorge Asís, in reference to that imaginary country in which Néstor and Cristina pierced the ceiling of 80 popularity points and it was all good news – also told me revealing scenes. Like that of a young Alberto Fernández, Chief of Staff at the time, who asked one of the hired pollsters, Analía del Franco, why they didn’t close his numbers when he added up the percentages of the different candidates measured: “Analía, here the count gives me more than a hundred, how can it be?” And then the pollster corrected the slip, produced by creative interventions to favor those who paid for that work. No, the job of pollsters was not an exact science by any means.
Today, the underlying question is whether, in addition to incompetence, there is bad intention on the part of professionals in the field in the dissemination of misleading numbers that reality, time and again, ends up denying. Are pollsters wrong on purpose? Or did they just stop taking the trouble to measure meticulously and rigorously, which, of course, costs money? The increasingly smaller samples of its measurements – in some cases, less than a thousand cases -, the persistence of obvious methodological errors such as telephone surveys or the sum of tricks such as those revealed by Artemio López point to a general crisis of the profession and explain why it lost credibility.
Surveys stopped being a mirror of reality and increasingly became a business.
