{"id":1622,"date":"2026-10-20T15:23:10","date_gmt":"2026-10-20T18:23:10","guid":{"rendered":"https:\/\/www.onthenet.com.br\/index.php\/2026\/10\/20\/brazil-runoff-polls-flavio-lula-quaest-datafolha-atlasintel-nexus\/"},"modified":"2026-10-20T20:36:48","modified_gmt":"2026-10-20T23:36:48","slug":"brazil-runoff-polls-flavio-lula-quaest-datafolha-atlasintel-nexus","status":"publish","type":"post","link":"https:\/\/www.onthenet.com.br\/index.php\/en\/2026\/10\/20\/brazil-runoff-polls-flavio-lula-quaest-datafolha-atlasintel-nexus\/","title":{"rendered":"Runoff Polls: Fl\u00e1vio Leads in Seven of Eight Surveys by 0.6 to 4 Points; Only Vox Brasil Shows Lula Ahead"},"content":{"rendered":"\n
\"Electronic<\/figure>\n\n\n\n

Eight national polls released between October 14 and 20, less than a week before the runoff (October 25), show Fl\u00e1vio Bolsonaro (PL) ahead of Lula (PT) in seven of them, by 0.6 to 4 percentage points among valid votes. The exception is Vox Brasil, which has Lula at 51% to 49%. Quaest, Datafolha, BTG\/Nexus and Vox Brasil describe their results as a technical tie. AtlasIntel (a 2.2-point gap against a 1-point margin) and PoderData\/Aya (4 points against a 1.8 margin) show gaps above their reported margins.<\/p>\n\n\n\n

Note:<\/strong> none of the polls was conducted entirely after Band’s debate, held on the night of Sunday (18). PoderData\/Aya and BTG\/Nexus were in the field from October 16 to 18, ending on the day of the debate. AtlasIntel interviewed from the 13th to the 18th, Vox Brasil and DataTrends from the 15th to the 17th. Only Real Time Big Data includes the following day, with fieldwork through October 19. The first rounds likely to come closer to the debate are due Wednesday (21), from Quaest, and Thursday (22), from Datafolha.<\/p>\n\n\n\n

The polls side by side<\/h2>\n\n\n\n
Quaest (Oct 14)<\/text>46%<\/text>44%<\/text>Datafolha (Oct 15)<\/text>47%<\/text>45%<\/text>Vox Brasil (Oct 18)<\/text>44.6%<\/text>46.4%<\/text>BTG\/Nexus (Oct 19)<\/text>46.5%<\/text>45.9%<\/text>AtlasIntel (Oct 19)<\/text>49.8%<\/text>47.6%<\/text>Real Time Big Data (Oct 20)<\/text>48%<\/text>45%<\/text>Fl\u00e1vio Bolsonaro (PL)<\/text>Lula (PT)<\/text><\/svg>
Runoff voting intention (total votes, prompted scenario), for the pollsters that released this figure. Source: polling firms, as published.<\/figcaption><\/figure>\n\n\n\n
Pollster<\/th>Fieldwork (sample)<\/th>Fl\u00e1vio<\/th>Lula<\/th>Valid votes<\/th>Margin<\/th><\/tr><\/thead>
Quaest (Oct 14)<\/td>Oct 10-13 (2,154)<\/td>46%<\/td>44%<\/td>51% x 49%<\/td>2 pts<\/td><\/tr>
Datafolha (Oct 15)<\/td>Oct 13-15 (2,520)<\/td>47%<\/td>45%<\/td>51% x 49%<\/td>2 pts<\/td><\/tr>
Vox Brasil (Oct 18)<\/td>Oct 15-17 (2,100)<\/td>44.6%<\/td>46.4%<\/td>49% x 51%<\/td>2.15 pts<\/td><\/tr>
DataTrends<\/td>Oct 15-17 (2,000)<\/td>\u2014<\/td>\u2014<\/td>51.5% x 48.5%<\/td>2.19 pts<\/td><\/tr>
BTG\/Nexus (Oct 19)<\/td>Oct 16-18 (2,002)<\/td>46.5%<\/td>45.9%<\/td>50.3% x 49.7%<\/td>2 pts<\/td><\/tr>
PoderData\/Aya (Oct 19)<\/td>Oct 16-18 (3,000)<\/td>\u2014<\/td>\u2014<\/td>52% x 48%<\/td>1.8 pts<\/td><\/tr>
AtlasIntel (Oct 19)<\/td>Oct 13-18 (4,998)<\/td>49.8%<\/td>47.6%<\/td>51.1% x 48.9%<\/td>1 pt<\/td><\/tr>
Real Time Big Data
(Oct 20)<\/td>		Oct 15-19 (2,000)<\/td>48%<\/td>45%<\/td>52% x 48%<\/td>2 pts<\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n

The valid votes column shows Fl\u00e1vio x Lula. Electoral court (TSE) registrations: Quaest, BR-08729\/2026; Datafolha, BR-00466\/2026; Vox Brasil, BR-04332\/2026; DataTrends, BR-04454\/2026; BTG\/Nexus, BR-04499\/2026; AtlasIntel, BR-02418\/2026; Real Time Big Data, BR-02952\/2026. Cells with a dash mean the figure was not available in the sources consulted.<\/p>\n\n\n\n

A note on headlines about the debate<\/h2>\n\n\n\n

Some headlines say Fl\u00e1vio “gains 4 points after the debate.” In the PoderData\/Aya poll, what grew was the gap between the candidates, which is now 4 points among valid votes. Fl\u00e1vio went from 50% to 52% (up 2 points) and Lula from 50% to 48% (down 2), compared with the poll released on October 16. In addition, fieldwork ran from October 16 to 18, ending on the day of the debate, not after it.<\/p>\n\n\n\n

Some outlets also describe Real Time Big Data as the first poll to capture the full impact of the debate. The firm interviewed voters from October 15 to 19, and only the last day falls after the debate.<\/p>\n\n\n\n

What each poll shows<\/h2>\n\n\n\n