MAIN STORY
Four polls reveal an exceptionally narrow gap
Among valid-vote preferences, CNT/MDA measured Lula at 47.8% and Flávio Bolsonaro at 42.1%. Datafolha reported 45% against 42%, Quaest 46% against 45%, and AtlasIntel/Bloomberg 47% against 44.1%. Lula led each first-round comparison, but none established that he would clear the majority threshold outright. Poll figures cannot sensibly be interpreted without their different methods, fieldwork dates and stated margins of error.
Datafolha interviewed 4,006 people on October 2–3 and reported a two-percentage-point margin of error. Quaest questioned 3,702 people over the same period, also with a two-point margin. CNT/MDA interviewed 2,007 respondents between September 30 and October 2, with a 2.2-point margin; AtlasIntel/Bloomberg collected 4,945 responses from September 27 to October 2, with a one-point margin. Even a modest difference between expected and actual turnout could shift the balance.
For the public, these tight estimates carry a straightforward message: there is no confirmed winner. Pollsters use differing ways to assemble samples and adjust results, so their numbers cannot simply be combined into a certain forecast. What the surveys consistently capture is a competitive election, rather than the comfortable mandate that either campaign might hope to project.














