Two New Polls Just Redrew the Governor Races in Ceará and Pará

Brazil · Politics

Key Facts

Ceará is tied.Real Time Big Data puts Elmano de Freitas and Ciro Gomes on 44% each in the first round.
And tied in a runoff.A simulated second round gives Elmano 46% to Ciro’s 45% — inside the margin of error.
The method.1,600 voters interviewed 15–19 August, margin of error 2 percentage points, 95% confidence, registered with the electoral court as CE-08223/2026.
Pará is not close.AtlasIntel has Hana Ghassan on 48.6% against Dr. Daniel Santos on 40.5% in the first round, from 1,197 interviews.
Presidential coat-tails.In Ceará, Lula leads Flávio Bolsonaro 66% to 27% in a runoff and 65% to 21% in the first round.
Still to come.Fresh Quaest rounds are expected in Pernambuco and Rio de Janeiro; no registration numbers or release dates had been published as of 20 August.

A week ago Datafolha had Ciro Gomes 19 points clear in Ceará. The newest poll has the race level at 44 apiece — which tells you as much about polling as about the candidates.

If you were reading Brazil governor polls a week ago, you would have concluded that Ciro Gomes was running away with Ceará. The poll published on 20 August says the race is exactly level. Nothing dramatic happened in between. This is simply what happens when two different firms, using two different methods, measure the same close contest — and it is worth understanding before October, because Brazil’s state races will shape how the next federal government governs.

Ceará: 44 to 44

Real Time Big Data interviewed 1,600 voters in Ceará between 15 and 19 August and found Elmano de Freitas, the sitting governor, and Ciro Gomes, the former minister and four-time presidential candidate, on 44% each. The margin of error is 2 percentage points at 95% confidence, and the survey is registered with the electoral court as CE-08223/2026.

A simulated runoff between the two produces 46% for Elmano and 45% for Ciro. That is a one-point gap in a poll with a two-point margin — statistically, it is a coin flip.

Compare that with the Datafolha poll published on 13 August, commissioned by O Povo, which had Ciro on 52% to Elmano’s 33% from 1,022 interviews on 10–12 August. Neither firm is wrong. They weight their samples differently and they reach respondents differently, and in a race this tight those choices move the result by far more than the gap between the candidates.

Pará: a clearer picture

Pará is the opposite kind of race. AtlasIntel interviewed 1,197 voters between 14 and 19 August, with a three-point margin of error, registered as PA-04533/2026. It was the first poll published after a leaked private video involving Dr. Daniel Santos. Hana Ghassan of the MDB takes 48.6% in the first round against 40.5% for Dr. Daniel Santos of Podemos. Everyone else is under 1.5%.

A simulated second round barely moves it: Ghassan 49.1%, Santos 41.8%. Blank and spoiled ballots come to 6.4% and 2.7% say they do not know.

Pará matters beyond its own borders. Belém hosted the COP30 climate summit in late 2025, and the state is where Brazil’s arguments about the Amazon, oil at the mouth of the river and federal money for the North all land at once.

What the presidential numbers do to the state races

The same Real Time Big Data fieldwork also tested the presidency — registered separately as BR-08791/2026 — and the numbers there are lopsided. Lula leads Flávio Bolsonaro 66% to 27% in a simulated runoff in the state, and 65% to 21% in the first round.

That is a regional result, not a national one — the Northeast has been Lula’s strongest base for two decades. But it complicates Ciro Gomes’s position. He is running against a governor from Lula’s party in a state where Lula is enormously popular, which means every vote he wins has to be won on local terms rather than national ones.

Reading Brazil governor polls without getting fooled

Three habits will save you from most polling mistakes here. Check the field dates, not the publication date. Check the registration number with the TSE, the electoral court; every legitimate poll has one and the registration names the client — both polls above were commissioned by the firms themselves. And treat any gap smaller than twice the margin of error as no gap at all.

Applied to Ceará, that last rule says the race is tied and has been tied for some time, and that the apparent swing from the earlier Datafolha poll is probably some mixture of method and noise rather than a real collapse in Ciro’s support.

What lands next

Two more states are due. Fresh Quaest rounds are expected in Pernambuco and Rio de Janeiro, but as of 20 August neither registration number nor release date had been published, so treat any circulating numbers with care until they are.

In Pernambuco the last published numbers, from a Genial/Quaest survey fielded 22–26 July with 900 respondents (PE-09649/2026), had Raquel Lyra on 43% and João Campos on 37%. In Rio, a Genial/Quaest poll fielded 21–25 July with 1,200 respondents (RJ-02671/2026) put Eduardo Paes on 38%, with Douglas Ruas and Anthony Garotinho on 10% each. Rio is currently being run on an interim basis by Ricardo Couto, president of the state court of justice, after Cláudio Castro’s resignation.

For foreign residents and investors, state governors are not a footnote. They set state VAT policy, sign infrastructure concessions and control the police. Whoever wins in Recife and Rio will matter to your electricity bill and your commute long after the presidential result stops being news.

Frequently Asked Questions

Who is leading the Ceará governor race?

Nobody, on the latest numbers. Real Time Big Data’s poll published 20 August has Elmano de Freitas and Ciro Gomes on 44% each, and a simulated runoff at 46% to 45% — both results well inside the two-point margin of error.

Why do Brazil governor polls disagree with each other?

Different firms use different sampling and weighting methods, and interview at different times. In a race where the candidates are within a point or two of each other, those technical choices move the headline number more than any real shift in opinion does.

What are the latest numbers for Pernambuco and Rio de Janeiro?

The most recent published figures are from July. Genial/Quaest had Raquel Lyra on 43% to João Campos’s 37% in Pernambuco, and Eduardo Paes on 38% in Rio. Newer rounds are expected but had not been registered or released as of 20 August.

Connected Coverage

Sources: Poder360 — AtlasIntel Pará; Valor — Quaest Pernambuco registration; Tribunal Superior Eleitoral — poll registry

This article was produced by The Rio Times’ automated newsroom system. How we use AI · Report an error