Quaest’s first poll since Brazil’s first round shows a closer contest between Flávio Bolsonaro and Luiz Inácio Lula da Silva than the current Votendi polling average. It adds evidence of a competitive runoff, but does not, on its own, establish that the race is tightening.
The survey, released on October 14, puts Bolsonaro at 46% of total voting intentions and Lula at 44%. Blank or null responses account for 9%, with 1% undecided, according to UOL’s report of the Quaest results. The reported margin of error is two percentage points.
After adding the survey, Votendi’s Brazil runoff tracker contains 11 published polls and displays a weighted average of 52.2% for Bolsonaro and 47.8% for Lula on a valid-vote basis. These figures are a snapshot checked on October 15; the live tracker will change as further surveys are added.
What Quaest actually measured
Quaest interviewed 2,154 voters aged 16 and over between October 10 and 13. The survey was commissioned by TV Globo and O Globo and registered with Brazil’s electoral court under BR-08729/2026. Its stated confidence level is 95%.
Bolsonaro’s two-point lead in total voting intentions is within the survey’s reported margin of error. It should not be read as a secure advantage or a prediction of the October 25 result.
Why 46–44 becomes approximately 51.1–48.9
Comparisons require the same denominator. Quaest’s 46% and 44% include all respondents in the base. The runoff tracker compares candidate shares on a valid-vote basis, excluding blank, null and undecided responses.
UOL calculated approximate valid-vote shares from the rounded published figures: 46 divided by 90 gives 51.1% for Bolsonaro, while 44 divided by 90 gives 48.9% for Lula. Quaest did not itself publish those valid-vote percentages. A calculation using the underlying unrounded data could produce slightly different values.
Votendi records this distinction in the Quaest poll detail page. The converted figures allow a comparison with other runoff surveys, but should retain their calculation note whenever reproduced.
One poll versus the broader polling picture
| Measure | Flávio Bolsonaro | Lula | Bolsonaro lead |
|---|---|---|---|
| Quaest: total voting intentions | 46% | 44% | 2 points |
| Quaest: UOL’s calculated valid-vote shares | 51.1% | 48.9% | 2.2 points |
| Votendi weighted average: 11 polls | 52.2% | 47.8% | 4.4 points |
On the comparable valid-vote measure, Quaest places Bolsonaro’s lead 2.2 percentage points below the lead in Votendi’s average. The survey therefore sits on the closer-race side of the current polling picture.
The tracker weights surveys by recency, sample size, pollster grade, methodology and editorial weight. It is not a simple arithmetic mean, and its 11 surveys are not 11 equally weighted or necessarily independent measurements. See Votendi’s methodology for how the tracker is constructed.
Why this is not yet proof of a narrowing race
This is Quaest’s first survey since the first round. Comparing it with a different institute’s result does not isolate a change in voter preferences: differences can also reflect fieldwork dates, interviewing methods, sample composition and recurring differences between pollsters.
Adding a relatively close poll can reduce the aggregate lead mechanically. That is different from demonstrating that voters have moved toward Lula over time. Stronger evidence would come from comparable repeat surveys showing a consistent movement across several institutes.
For now, the useful conclusion is limited but clear: Quaest shows a narrower Bolsonaro advantage than the broader average, while both leave room for a competitive runoff. Neither the individual survey nor the average is a win-probability forecast.
Follow the Brazil polling tracker for the full survey list, fieldwork dates and updated averages, and read our runoff guide for the election context.
