Three Polls in Ceará, Three Different Races. Somebody Is Badly Wrong
Brazil · POLITICS
Key Facts
- —Ipsos-Ipec, 28 July Ciro Gomes 43%, Elmano de Freitas 31%, Eduardo Girão 6%. Fieldwork 23–26 July, 800 voters, margin of error 3 points.
- —Datafolha, 13 August Ciro 52%, Elmano 33%. Fieldwork 10–12 August, 1,022 voters across 51 municipalities, margin of error 3 points.
- —Real Time Big Data, 20 August Ciro 44%, Elmano 44%. Fieldwork 15–19 August, 1,600 voters, margin of error 2 points.
- —The spread Nineteen points to nothing in seven days, between two institutes both registered with the electoral court.
- —The 6×1 bill The Chamber approved the constitutional amendment moving Brazil to a five-day week on 28 May. The Senate president sent it to the constitutional affairs committee on 19 August.
- —Where Caiado stands At a transport industry forum in Brasília on 19 August he declined to take a position, calling the proposal “populista, eleitoreira” and saying “Eu não voto nessa matéria.”
Ipec had it at twelve points. Datafolha had it at nineteen. Real Time Big Data has it at nothing. The candidates did not change.
Ceará governor polls have produced three different races in four weeks. Ciro Gomes led Elmano de Freitas by twelve points in late July, by nineteen in mid-August, and by nothing at all on 20 August. All three institutes are registered with the electoral court. At least two of them are describing a state that does not exist.
What the three Ceará governor polls say
Ipsos-Ipec went first. Fieldwork on 23–26 July, published 28 July, 800 voters, margin of error three points: Ciro Gomes of the PSDB on 43%, Elmano de Freitas of the PT on 31%, Eduardo Girão of Novo on 6%. In a runoff, Ciro 50% to Elmano 41%.
Datafolha went next. Fieldwork 10–12 August, published 13 August, 1,022 voters across 51 municipalities, margin of error three points: Ciro 52%, Elmano 33%. Runoff, Ciro 55% to Elmano 37%. That is a nineteen-point first-round lead — wider than Ipec had it two weeks earlier, not narrower.
Then Real Time Big Data. Fieldwork 15–19 August, published 20 August, 1,600 voters, margin of error two points, self-funded: Ciro 44%, Elmano 44%. Runoff, Elmano 46% to Ciro 45%.
Seven days separate the end of Datafolha’s fieldwork from the start of Real Time’s, and in those seven days a nineteen-point lead becomes a tie. Nothing happened in Ceará in that week that would do that.
How to read polls that disagree this badly
The first rule is the one people skip: treat any gap smaller than twice the margin of error as no gap at all. On that test, Real Time’s 44–44 is a genuine tie, and both Ipec’s twelve points and Datafolha’s nineteen are real leads. The tests do not reconcile. This is not statistical noise; it is methodological divergence.
The usual suspects are sampling frame and the treatment of undecideds. An institute that pushes leaners harder produces sharper margins. One that weights by past vote in a state where the PT has governed for sixteen years produces different numbers from one that weights by demographics alone.
The second rule is to look at direction rather than level. Two of the three have Ciro ahead. One has the race level. If you are betting on the shape of the contest rather than the number, the honest reading is that Ciro leads by an amount nobody has pinned down, and that the race has tightened since July rather than that it has flipped.
The third rule is to distrust anyone quoting one poll. Including the campaigns, both of which spent this week quoting exactly one.
The 6×1 bill, and what Caiado actually said
Brazil’s move away from the six-day working week stopped being a campaign and became a bill some time ago, and the coverage has not caught up. The Chamber approved the constitutional amendment shifting to a five-day, two-day-off schedule on 28 May, with a transition rule. It is now in the Senate, where the Senate president signed the order sending it to the constitutional affairs committee on 19 August.
That makes what presidential candidates say about it consequential rather than rhetorical, which is why Ronaldo Caiado’s answer at the transport confederation’s forum in Brasília on 19 August is worth quoting rather than summarising. Asked about the measure, the Goiás governor declined to take a position: the proposal, he said, had “um objetivo simplesmente populista, eleitoreira, de voto.” And then, plainly: “Eu não voto nessa matéria. Não posso influenciar uma matéria que está com viés eleitoreiro.” The agenda would become his, he added, when he took power.
That is a candidate declining to be pinned down on the largest change to Brazilian working time in a generation, five weeks before an election. It is a defensible position — he is not a senator and has no vote — and it is also an answer that tells you the polling on this issue is not where the business audience he was addressing would like it to be.
If it passes, a shift from six days to five reaches almost every hourly worker in the country, and every employer of them.
Two other things said out loud this week
Romeu Zema, the Novo candidate and governor of Minas Gerais, opened his campaign in Montes Claros on 16 August predicting his own runoff and a united right in it: “Estarei no segundo turno e, tal como aconteceu no Chile, toda a direita vai estar unida.” The Chile comparison is doing the work in that sentence, and he chose it deliberately.
Flávio Bolsonaro, four days later, on a livestream, complained about the ethanol blend rising to 32% — “A gasolina virou etanol” — and said the quantity would have to be discussed. In the same broadcast, and on a separate subject, he complained that people who call themselves centre-right spend their time attacking him instead of the president: “Estou sozinho.”
Put Zema and Flávio side by side and you have the shape of the problem on that side of the ballot. One is campaigning on the assumption that the right consolidates behind whoever reaches the runoff. The other is publicly complaining that it is not consolidating behind him.
Why the Ceará governor polls matter beyond Ceará
Ceará is one of the PT’s most reliable states and Ciro Gomes is running against it from the PSDB, having spent a career in and around the party he is now opposing. A PT loss there would be read nationally, and both campaigns know it.
For anyone living in Brazil, the practical point is narrower and more useful: between now and October you will be shown a great many numbers, and the spread between three registered institutes in one state in four weeks is nineteen points. Check the fieldwork dates, check the sample size, and check who paid.
And if you employ people here, the file to watch is not the presidential race but the constitutional affairs committee of the Senate. That is where the working week is decided.
Frequently Asked Questions
Who is leading the Ceará governor race?
It depends which poll you read. Ipsos-Ipec had Ciro Gomes 43–31 over Elmano de Freitas on 28 July, Datafolha had it 52–33 on 13 August, and Real Time Big Data had it level at 44–44 on 20 August. Two of the three show a clear Ciro lead.
Has Caiado backed ending the 6×1 work week?
No. At the CNT forum in Brasília on 19 August 2026 he declined to take a position, called the proposal “populista, eleitoreira” and said “Eu não voto nessa matéria,” adding that the agenda would be his once he took office.
Where does the 6×1 constitutional amendment stand?
The Chamber of Deputies approved it on 28 May 2026. It is in the Senate, where the Senate president signed the order sending it to the constitutional affairs committee on 19 August 2026.
Connected Coverage
Sources: Ipsos-Ipec; Datafolha; Real Time Big Data; Poder360; Carta Capital; O Povo; InfoMoney; Câmara dos Deputados; Senado Federal; Confederação Nacional do Transporte.
This article was produced by The Rio Times’ automated newsroom system. How we use AI · Report an error
Read More from The Rio Times