Half of Filipinos still not locked in for 2028

By Francis Allan L. Angelo

Filipinos ranked a woman who says she is not running second in the 2028 presidential field, handed the only declared candidate her commanding numbers in a single region, and left a bloc of voters larger than the third placer entirely uncommitted.

Those three findings sit inside the June 2026 WR National Poll: Philippine Public Opinion Monitor, released on July 31 by WR Numero Research, the polling and survey research arm of WR Advisory Group.

Vice President Sara Duterte led the field with 39.40%. Naga City Mayor Leni Robredo followed with 20.94%. Respondents who had not chosen a candidate accounted for 16.98%, a share larger than the 14.23% recorded by Senator Raffy Tulfo, who placed third.

Senator Bong Go drew 3.58%, Senator Bam Aquino 2.75%, and Senator Risa Hontiveros 2.13%.

Duterte announced her candidacy for the presidency on Feb. 18, 2026. Robredo has said repeatedly that she does not intend to run. The declared candidate leads. The refusenik is second. The survey did not ask respondents whether they knew the difference.

The timing of the fieldwork matters more than any single number in the release. WR Numero interviewed 1,432 adults face-to-face from June 9 to 16, three weeks before the Senate opened the impeachment trial of the Vice President on July 6. The findings were released on July 31, by which point the trial had run 10 hearing days. Every figure in this survey, therefore, describes the 2028 race as it stood before the proceedings that will determine whether the front-runner can hold public office at all.

A second question in the same survey suggests how little the standings should be trusted to hold. Asked whether they might still change their preferred candidate between the survey period and election day in May 2028, 58.3% of respondents said their choice was settled. The remaining 41.7% said they might still change their minds.

Read alongside the 16.98% who have not chosen anyone, that produces an electorate in which something close to half the voting public is either uncommitted or openly persuadable. The two figures cannot simply be added, because the WR Numero release does not state whether the follow-up question was put to undecided respondents as well, and any overlap would reduce the combined total. Even at the conservative reading, the persuadable share dwarfs the gap between second place and third.

The undecided bloc is also unevenly distributed, and its geography is the opposite of the front-runner’s. Uncertainty runs highest in the rest of Luzon at 22.65% and in Metro Manila at 20.86%, and lowest in Mindanao at 5.96%. The Visayas recorded 13.98%, the most decided electorate outside Mindanao.

Area	Duterte	Robredo	Tulfo	Undecided	MOE
Total Philippines	39.40%	20.94%	14.23%	16.98%	±2.6
Metro Manila	24.82%	22.29%	23.29%	20.86%	±7
Rest of Luzon	20.36%	31.27%	16.98%	22.65%	±4
Visayas	35.80%	19.30%	16.78%	13.98%	±6
Mindanao	88.50%	1.02%	1.44%	5.96%	±6

Source: WR Numero Philippine Public Opinion Monitor, Volume 2026, Issue 10. Fieldwork June 9 to 16, 2026. Sample size 1,432 adults. Margin of error at 95% confidence.

The regional breakdown is where the headline number comes apart. Duterte polled 88.50% in Mindanao, her home region. She polled 35.80% in the Visayas, 24.82% in Metro Manila, and 20.36% in the rest of Luzon, where Robredo led with 31.27%.

Applying approximate adult population shares to those four regional results reproduces the published national figures to within half a percentage point, which allows a rough but defensible estimate of what the race looks like with Mindanao removed. Excluding the region, Duterte falls to roughly 25%, Robredo rises to roughly 27%, undecided voters account for roughly 20%, and Tulfo sits at roughly 18%. The gap between the top two in that scenario is smaller than any margin of error, making it a statistical tie rather than a lead.

Daily Guardian’s calculation (not WR Numero’s) rests on population weights of approximately 13% for Metro Manila, 45% for the rest of Luzon, 20% for the Visayas, and 22% for Mindanao, and it should be treated as an approximation because the release does not disclose subsample sizes or the weighting scheme actually used.

The point survives the imprecision. Duterte’s national plurality is not a national coalition. It is one region voting almost unanimously, attached to a plurality everywhere else, and the arithmetic is such that Mindanao supplies close to half of her total national support while holding roughly a fifth of the electorate.

That concentration carries a practical cost that no poll measures. A campaign whose base is geographically dense but regionally confined must spend to build presence in three other areas rather than defend one, and it must do so while its principal is on trial for offenses that carry permanent disqualification from public office on conviction. Campaign finance disclosures for 2028 do not yet exist, and the Commission on Elections statements of contributions and expenditures from 2022 remain the only benchmark. Whether donors are already hedging is a reporting question, not a polling one, and it is the question this survey most conspicuously cannot answer.

The Metro Manila result deserves a separate caution. Duterte’s 24.82%, Tulfo’s 23.29%, and Robredo’s 22.29% fall inside a subnational margin of error of plus or minus seven percentage points. The three are statistically indistinguishable. The WR Numero release nonetheless states that Duterte ranked first in the capital region, a characterization the underlying data does not support.

Mindanao produces the survey’s most doubtful figure. WR Numero reported Hontiveros at 0.01% in the region and Aquino at 0.25%. The stated Mindanao margin of error of plus or minus six percentage points implies a subsample of roughly 270 respondents. One hundredth of one percent of 270 respondents is a fraction of a single person. The figure cannot represent an actual count, and is most plausibly an artifact of weighting or a transcription error. WR Numero has not been asked to explain it, and should be.

For Western Visayas, the more useful finding is what the survey does not disaggregate. WR Numero treats the Visayas as one unit, and reports Robredo at 19.30%, Tulfo at 16.78%, and Duterte at 35.80% across all three Visayan regions combined. Those three names sit within a spread that the regional margin of error cannot separate for second place.

That aggregate sits uneasily against the last actual vote cast in the area. In the 2022 presidential election, Robredo took 149,256 votes in Iloilo City against 85,117 for Ferdinand Marcos Jr., or roughly 64% of the top two vote share. Across Iloilo province and the city combined, Commission on Elections figures recorded 728,781 votes for Robredo against 459,261 for Marcos, a margin of 269,520. Subtracting the city leaves approximately 579,525 votes for Robredo and 374,144 for Marcos in the towns, a provincial margin of about 205,381.

A two-candidate general election result is not comparable to a multi-candidate preference survey taken two years out, and no one should treat 2022 as a forecast. What the comparison does establish is that a single Visayas figure cannot describe an electorate that includes a province where one candidate carried roughly six of every ten votes in the last presidential contest. The regional crosstab that would settle the question has not been published.

The Visayas result carries one finding the release passes over. Aquino recorded 5.37% and Hontiveros 3.89% in the region, their strongest showings anywhere in the country. Together with Robredo, opposition-aligned candidates drew roughly 29% of Visayan respondents, spread across three names rather than consolidated behind one. In the rest of Luzon, the same three drew roughly 37%, with 31.27% of it behind a single candidate.

Several elements of the survey remain unavailable. WR Numero states that the same round measured public assessments of President Ferdinand Marcos Jr. and the Vice President, opinions on the impeachment proceedings, and views on proposed constitutional amendments. None of those figures appear in the release. Impeachment numbers gathered in June, three weeks before the trial opened, would establish the public baseline against which the trial’s effect can eventually be measured, and they have not been published.

The release also omits vice presidential preferences, which WR Numero included in its March 2026 round, and it does not state whether the survey was commissioned or self-funded.

The document carries two different national margins of error. The narrative text states plus or minus three percentage points. The table footnote states plus or minus 2.6. The second figure is the arithmetically correct one for a sample of 1,432 at the 95% confidence level.

WR Numero occupies a specific niche in Philippine polling. It publishes far more frequently than Social Weather Stations or Pulse Asia, which makes it useful for tracking movement over time and less useful as an authoritative single reading. Its March 2026 round, reported at the time, placed Duterte at about 36%, Tulfo at 19%, and Robredo at 16%. If those figures are accurate, the June survey records not merely a set of gains but an exchange of position between second and third place, which the release does not mention.