Lack of Understanding With the Russians

10/13/2014 - 12h18

The difference between projections and results affected the credibility of polls, as well as that of coverage by newspapers, amply anchored in the surveys. Never have electoral polls been so prominent as they were this year.

Besides this, there were basically two more ingredients: back-and-forth accusations between candidates and complaints about corruption. The first is standard fare in all election campaigns, heating up the polarization, and by the insane microwaves of blogs and social networks.

In the second, the protagonists were federal prosecutors, and newspapers took on the role of publishing headlines almost always constructed from a thread in the news, from leaks done in dribs and drabs by people involved in the investigation.

The lead in stories carried a few drops of something new, and the rest was context and background of what had already been published.

With this discouraging backdrop, the misfortune helped the coverage, and the adversity that Marina Silva faced in the race guaranteed a roller coaster which increased work by polling institutes and helped newsrooms.

In 44 days, Folha credited eight headlines to the Datafolha polling institute and one to the Ibope institute, outside of smaller stories.

By the predominance and the influence of polls in the news - and voting - it was necessary for the newspaper to make a consistent evaluation of the differences between voting results and surveys, addressing the cases one by one trying to provide clear explanations to not allow a lack of confidence concerning manipulation.

That's not what happened. The story published on Tuesday (Oct. 7) sounded like a lecture by the government, with an anodyne headline ("Polling institutes suffer criticism") and subhead saying that the differences generated complaints about politicians.

It wasn't just them, the outcry was general. There weren't only complaints but lack of understanding and a desire to understand.

I observed the inconsistency in my internal critique, thinking that the newspaper should have sought other experts to get independent opinions. It should have deepened the discussion about the methodology adopted and its limits.

The managing editor's office believes that my critique is wrong because part of the presupposition that the results need to be the same as the polls and argued: "When the result from the ballot box is the same as the survey released on the eve of the election, it's not seen as something the institute got right. It's only an indication that nothing relevant to voters changed Saturday and Sunday, when people talk most about the election."

Well, the newspaper is used to celebrating when it gets the numbers right. It should stop doing that. Afterward, the problem is not the presupposition but rather the treatment given to data which should be extensively made relative.

I talked to Mauro Paulino and Márcia Cavallari, responsible for the two main polling institutes, who emphasized the conditions in the surveys. Paulino told me, for example, that two thirds of the interviews by Datafolha were carried out on Friday (3) and one third of them on Saturday (4) morning.

He said that there was a move to change the choice of candidates with the debate on the Globo TV network (in which Marina Silva did poorly) and intensified through the weekend by surveys on Saturday, which showed Aécio Neves of the Brazilian Social Democracy Party rising. The trend changed.

The explanation makes sense, but none of these reservations was contemplated in the story on Sunday nor in the explanations on Tuesday. On the day of the election, there was a long analysis by Datafolha, but these aspects were not mentioned.

There were other differences even more attention-grabbing than the presidential race, but it's not the role of the ombudsman to evaluate the performance of polling institutes. It's that of the newspaper, which gives these surveys top billing in the news.

The managing editor's office asserts that Folha's coverage is cautious and does not treat surveys as an attempt to predict the results in the voting booth. If this is the perception, it's good to flash a yellow light, because the Russians don't understand.

Translated by JOHN WRIGHT

Read the article in the original language