One of the most talked about topics in WhatsApp groups on Sunday night, Sep. 2, was the difference in performance between what the leading research institutes pointed out and the results of the polls.
During the campaign, Bloomberg Línea published the results of four institutes/associated companies.
Datafolha, Ipec (ex-Ibope), Quaest/Genial Investimentos, and FSB/BTG Pactual), based on the existence of a history of conducting surveys that would serve as a reference.
None of them captured the strength of current president Jair Bolsonaro, who ended the first round with 43.47% of the valid votes.
The closest one was the Quaest/Genial Investimentos, which indicated that the president could reach 40%, counting the margin of error.
One possible explanation is that Bolsonaro maintained or even accelerated in the final stretch of the first-round campaign the upward trajectory demonstrated in the polls.
But this time, they all got the vote intention right for former president Lula da Silva, who reached the final phase of counting with 48.12% of the valid votes.
The closest poll was the FSB/BTG Pactual one, which indicated that Lula would reach 48% – conducted a week ago before the others.
As a result of the polls, the difference between Lula da Silva and Bolsonaro was around five percentage points.
Even with the margin of error, the four institutes mentioned above were wrong.
Datafolha and Ipec indicated that the difference could have reached ten percentage points. At the same time, Quaest/Genial and FSB/BTG suggested that the difference between the candidates could go seven percentage points.
See what each institute indicated in their latest poll:
DATAFOLHA
- Lula da Silva: 50% of the valid votes (48% to 52% with the margin of error)
- Bolsonaro: 36% (34% to 38% with margin of error)
Datafolha surveyed 12,800 people aged 16 and older in 310 municipalities on Friday, Sep. 30, and Saturday, Oct. 1.
The margin of error was two percentage points higher or lower.
IPEC
- Lula da Silva: 51% of the valid votes (49% to 51% with the margin of error)
- Bolsonaro: 37% (35% to 39% with margin of error)
The latest Ipec (formerly Ibope) poll heard 3,008 people between Thursday, Sep. 29, and Saturday, Oct. 1.
The margin of error was two percentage points higher or lower, and the confidence index was 95.45%.
QUAEST/GENIAL INVESTIMENTOS
- Lula da Silva: 49% of the valid votes (47% to 51% with the margin of error)
- Bolsonaro: 38% (36% to 40% with margin of error)
Quaest heard 3,600 people in their homes between Friday, Sep. 30, and Saturday, Oct. 1.
The margin of error was two percentage points higher or lower, and the confidence index was 95%.
FSB/BTG PACTUAL
- Lula da Silva: 48% of the valid votes (46% to 50% with the margin of error)
- Bolsonaro: 37% (35% to 39% with margin of error)
The survey interviewed 2,000 people by telephone between Sep. 23 and 25.
The margin of error was two percentage points higher or lower, and the confidence index was 95%.
Outside the four institutes mentioned above, Atlas Intel, in partnership with the political consultancy Arko Advice, came closest to the president’s result.
In a survey released on Friday, Sep. 30, Atlas/Arko pointed out that Lula da Silva had 50.7% of the valid votes (49.7% to 51.7% with the margin of error), while Bolsonaro would have 41% (40% to 42% with the margin of error) – that is, still without hitting 43.26%.
The Atlas/Arko survey collected responses from 4,500 people over the Internet between September 24 and 28. Respondents were recruited organically during routine web surfing in geolocated territories on any device (smartphones, tablets, laptops, or PCs).
The margin of error for the survey was one percentage point. The survey had a 95% confidence level.
WHAT EXPLAINS THE DIFFERENCES?
It is not the first time a right-wing candidate has had his vote estimated downward in the polls.
Four years ago, Bolsonaro arrived on the eve of the first round with 40% of the valid votes, according to Datafolha, while the runner-up, Fernando Haddad (PT), had 25%.
The counting showed superior performance by both, especially Bolsonaro: 46%, versus 29.3% for Haddad.
Before the first round of voting, some specialists and executives from research institutes pondered issues that could lead to or prevent a greater accuracy of the polls.
One is the difficulty of defining sample groups of voters less close to the Brazilian social pyramid by income and education but more similar to those who go to the polls.
In theory, lower class strata tend to direct their votes to leftist candidates, but their presence in the polls is smaller than the rest of the population.
With information from Bloomberg