Taipei Poll Puts Shen Puma 17 Points Ahead of Chiang Wan-an, but Turnout Models Show Close Race

A big polling lead meets a much tighter forecast

Democratic Progressive Party candidate Shen Puma has taken a 17 percentage point lead over Taipei Mayor Chiang Wan-an in a survey released on October 20, 2026. Yet the same polling exercise projects a much closer election, with some turnout scenarios giving the incumbent a narrow victory.

The survey commissioned by the Taipei Xinmin Cross-Strait Research Association puts Shen at 50.3% and Chiang, of the Kuomintang, at 33.3%. A month earlier, Chiang led by 41.7% to 40.5%. The gap between the two candidates has therefore moved 18.2 percentage points toward Shen between the association's September and October surveys.

Poll designer and analyst Tai Li-an offers a more cautious picture of what could happen at the ballot box. His estimates, which account for voting intention, strength of commitment and previous election results, put Shen's possible vote share at 49.2% to 52.9% and Chiang's at 46.9% to 49.9%.

The association's detailed presentation describes outcomes ranging from a Chiang victory by 0.7 percentage points to a Shen victory by about six points. Association founding chairman Huang Ching-lung described the range more broadly as Shen leading by roughly seven points or trailing by roughly one. Both descriptions put the projected contest much closer than the headline support figures.

The central issue is who answers a survey, who openly states a preference and who actually votes on November 28. Those groups can overlap without being identical. That helps explain why a candidate can hold a large lead among respondents while facing a competitive election.

What changed between September and October?

Shen gained 9.8 percentage points from his September support level of 40.5%, while Chiang lost 8.4 points from 41.7%. Huang described October as Shen's first clear polling lead since entering the race in May.

The remaining responses show that a substantial share of respondents still stood outside the two main camps. Another 12.2% gave no clear answer, while 3.0% said they would abstain or cast an invalid ballot. Four other candidates each received less than 1%: independent Lin Chih-cheng had 0.7%, Kuo Hsi 0.4%, and Tang Hsin-min and Hsiao Wen-chien 0.1% each.

The campaign's recent sequence is:

• September 16 to 18: The association conducted its first survey, interviewing 1,076 people.
• September 22: Results showed Chiang at 41.7% and Shen at 40.5%.
• October 14 to 16: Interviewers conducted the second survey, completing 1,075 interviews.
• October 20: The association released the new results, with Shen at 50.3% and Chiang at 33.3%.
• November 28: Voters are scheduled to cast ballots in Taiwan's local elections.

Shen responded on October 20 by saying public opinion could change quickly during the 39 days remaining before voting. He said his team was more concerned with whether residents connected with its policies than with support or favorability figures.

Where Shen gained beyond the DPP

Taiwan People's Party supporters and voters without a firm party preference account for an important part of Shen's improvement, according to Huang. Both groups moved toward him between the two surveys, although Chiang still led among Taiwan People's Party supporters.

Shen's support among respondents identifying with the Taiwan People's Party rose from 1.6% in September to 19.0% in October, an increase of 17.4 percentage points. Chiang's support in that group fell from 88.5% to 63.8%, a decline of 24.7 points.

Among respondents who described themselves as neutral or said they chose candidates rather than parties, Shen rose 8.3 points to 34.2%. Chiang fell 8.8 points to 22.9%. Shen consequently held an 11.3 point lead in that group.

Huang attributed much of Shen's growth to neutral voters and some people with weaker attachments to the Kuomintang camp. Tai similarly said Chiang's strongest partisan base had changed little, while support among more moderate voters leaning toward that camp was slipping.

District results show pressure on Chiang across Taipei

Shen led in all six groupings used for Taipei's city council elections. His largest advantage was in Zhongshan and Datong, where he received 56.2% against Chiang's 23.6%, a gap of 32.6 percentage points.

Songshan and Xinyi, described as a traditional Kuomintang stronghold, showed another large change. Chiang fell from 48.7% to 34.9%, while Shen rose 24.4 points to 52.7%. In Zhongshan and Datong, Chiang lost 15.7 points and Shen gained 15.1 points.

The October results across the six district groupings were:

• Zhongshan and Datong: Shen 56.2%, Chiang 23.6%.
• Zhongzheng and Wanhua: Shen 51.7%, Chiang 33.1%.
• Songshan and Xinyi: Shen 52.7%, Chiang 34.9%.
• Neihu and Nangang: Shen 49.6%, Chiang 32.3%.
• Beitou and Shilin: Shen 46.6%, Chiang 33.3%.
• Daan and Wenshan: Shen 47.6%, Chiang 39.2%.

Daan and Wenshan produced the smallest gap, at 8.4 points. The spread across the city gives both campaigns specific places to concentrate their efforts: Chiang faces losses across several areas, while Shen must turn geographically broad survey support into actual ballots.

Young voters move sharply, while some groups favor Chiang

Respondents aged 20 to 29 recorded the largest age group movement. Shen rose from 42.8% in September to 71.6% in October, while Chiang fell from 37.3% to 10.5%. Shen's increase was 28.8 percentage points; Chiang's decline was 26.8 points.

Shen also led among people aged 40 to 49, with 60.2% against Chiang's 32.5%. Among those aged 30 to 39, the figures were 46.5% and 36.9%. Shen reached 46.6% among respondents aged 70 or older after an increase of more than 10 points.

Chiang retained an advantage among respondents aged 50 to 59, receiving 42.4% against Shen's 37.9%. He also led among people with senior high school or vocational school education, at 46.4% to 40.8%.

Women favored Shen by 50.0% to 31.8%. Compared with September, his support among women increased 11.9 points, while Chiang's declined 10.5 points. These figures describe smaller groups within the survey, whose estimates carry greater sampling uncertainty than the full sample.

Why enthusiasm matters to the forecast

Nearly four in five respondents, 79.7%, said they would definitely vote, and another 12.9% said they might. Together, 92.6% expressed at least some intention to participate. Huang estimated that actual turnout could exceed 70%.

Among respondents saying they would definitely vote, Shen received 56.7% support and Chiang 35.2%, a gap of 21.5 percentage points. That makes the association's much tighter election projection especially striking.

Tai explained that people willing to answer political surveys often have greater political interest and a stronger desire to express an opinion than people who are not interviewed. His projection therefore considers participation and people outside the responding sample, as well as respondents' stated preferences.

Huang and Tai also described a change in willingness to speak openly. They argued that the political atmosphere had made supporters of the DPP camp more eager to declare themselves, while some Kuomintang supporters had become quieter.

They called this a spiral of silence, meaning that people who feel their political side is under pressure may become less willing to state their preference. The survey still obtained a declared position from more than four in five respondents. The size of any quieter electorate, and how it will vote, remains uncertain.

A dispute over how to weight the sample

Kuomintang figures challenged the survey's composition. New Taipei city council candidate Teng Kai-hsun said respondents aligned with the DPP camp outnumbered those aligned with the Kuomintang camp by 15.6 percentage points. He criticized the use of adjustments for sex, electoral area, age and education without an adjustment for party preference.

Weighting means changing how much influence different responses have in the final calculation. For example, a survey can adjust an age group's contribution to better match its share of the population. The argument here is whether party preference should also be matched to a chosen benchmark.

A calculation using the party preference proportions in two TVBS surveys conducted that year reversed the headline result, putting Chiang seven points ahead. That is a different calculation from the turnout estimates that produced a close contest.

Tai argued that weighting by party preference could distort public opinion and lead to a mistaken reading of the race. Party identification can itself change during a campaign, making the choice of a fixed benchmark consequential.

The association commissioned Tai to design the questionnaire and analyze the results, while Beacon Market Research conducted the interviews. The October survey covered Taipei residents aged 20 or older whose household registration was in the city. Its 1,075 interviews comprised 753 residential telephone interviews and 322 mobile telephone interviews, approximately 70% and 30% respectively. The stated maximum sampling error was plus or minus three percentage points.

Suitability, party change and the campaigns' responses

Shen also led when respondents were asked whether each candidate was suitable to serve as mayor. A total of 56.6% considered him suitable and 27.5% unsuitable. Chiang received 41.2% suitable and 46.8% unsuitable.

Compared with September, Shen's suitability rating rose 10.4 percentage points, while Chiang's fell 8.8 points. Huang said judgments about suitability and voting preferences were related but measured different choices.

Asked about which party should govern Taipei, 48.7% preferred giving the DPP a turn and 31.7% preferred continued Kuomintang administration. In September, those figures were 42.0% and 38.0%. Among definite voters in October, the preference for DPP administration reached 54.7%, against 32.6% for the Kuomintang.

Kuomintang representatives responded by calling on Shen and DPP New Taipei mayoral candidate Su Chiao-hui to resign their legislative seats to demonstrate commitment. Su had led by 13 points in an earlier association survey. New Taipei council candidate Ho Yuan-kai accused the association of using polling to build DPP confidence, discourage Kuomintang supporters and influence neutral voters.

Chiang said he would maintain his campaign's plans and pace. Taichung Mayor Lu Shiow-yen, who chairs his campaign headquarters, said when asked about weak polling that she had not seen the information and would comment when she had more to share.

Enthusiasm spreads beyond the capital

DPP campaign assessments describe a wider boost from enthusiasm surrounding Shen, nicknamed the Yang current after a character in his Chinese name. Internal estimates put the positive effect at at least six percentage points across several local contests.

Those assessments describe Su's New Taipei campaign as moving toward a steadier lead, with races in Keelung and Taichung approaching even contests. They also describe improving conditions in Taitung, Taoyuan and Hsinchu City, and a widening advantage for DPP candidate Lai Jui-lung in Kaohsiung. The six point estimate comes from internal campaign assessments rather than the Taipei telephone survey.

DPP planning calls for candidates from other areas to appear alongside Shen in Taipei, allowing him to focus on the capital instead of traveling constantly. President Lai Ching-te reminded party officials at an October 14 meeting that polling numbers and final votes were different, while party secretary general Hsu Kuo-yung urged caution and humility.

A voluntary online ballot conducted from October 11 to 16 also recorded enthusiasm for Shen: he received 2,833 votes, or 56.1%, against Chiang's 2,092 votes, or 41.5%. Participants chose to join the ballot, which was open beyond Taipei, so its percentages describe those participants. The association's telephone survey instead sought a sample of registered Taipei residents.

November 28 is the next decisive date. Shen's task is to retain the neutral and opposition leaning respondents who have moved toward him. Chiang's is to recover support and bring less vocal supporters to the polls. The association's own modeling leaves both campaigns with a competitive race to fight.

Key Points

• The October telephone survey put Shen at 50.3% and Chiang at 33.3%, reversing Chiang's narrow September lead.
• Turnout estimates produced a closer race, including a possible Chiang victory by 0.7 percentage points.
• Shen gained among neutral voters, Taiwan People's Party supporters, women and younger respondents.
• Chiang retained advantages among respondents aged 50 to 59 and Taiwan People's Party supporters.
• A dispute over party preference weighting produced sharply different calculations of support.
• The survey interviewed 1,075 Taipei residents from October 14 to 16, with a stated maximum sampling error of three percentage points.
• Voting is scheduled for November 28, 2026.