2024 Taiwan Election: Pre-election Telephone Surveys — Research Data

Doublethink Lab

Published in

Doublethink Lab

10 min readJan 19, 2024

Eric Hsu, Researcher, Doublethink Lab

Key findings:

When asked about the source of political information they receive, 71% of respondents stated that it comes from online news media and news channels. It is noteworthy that when combining “online video content creators” and “Key Opinion Leaders (KOLs or influencers) on social platforms” the percentage exceeds 40%, followed by pundits or current affairs commentators at 28.9%.
Despite over 60% of respondents indicating varying degrees of information verification habits, 37.6% of respondents stated that they never search or confirm information received online.
Apart from 40.4% of respondents refusing to answer which camp false information or online manipulation related to the presidential election primarily came from, 26.6% of respondents indicated it was mainly from the Democratic Progressive Party (DPP), followed by 15.2% who believed it originated from the People’s Republic of China/China.
Regarding party support, 30% of respondents stated they do not support any political party. Following this, support for the Democratic Progressive Party was at 28.8%, the Kuomintang at 16.5%, and the Taiwan People’s Party at 18%. As for the least favored political party, 35.5% of respondents stated they do not dislike any party, while specifically disliking the parties, the order was the Democratic Progressive Party at 28.4%, the Kuomintang at 19.9%, and the Taiwan People’s Party at 10.6%.
Over 60% of respondents expressed that “Regardless of the situation, democratic politics is always the best system” but nearly 20% of the public believes that “For me, any political system makes no difference.”
Regarding satisfaction with Taiwan’s current democratic situation, 57.6% of respondents fell within the satisfied range. For trust in Taiwan’s electoral system, 67% of respondents fell within the trust range.
Concerning the perception of a significant increase in prices and the urgent need to address housing justice in Taiwan, 84.7% and 88.4% of respondents, respectively, leaned towards agreeing with these statements.
Regarding allegations such as “The Taiwanese government benefits specific companies, providing citizens with subpar vaccines and contaminated eggs.” and “The current ruling party/Democratic Progressive Party (DPP) has a severe situation of political pork barrel and corruption” 51.7% and 56.1% of respondents, respectively, leaned towards agreeing with these statements.
64.6% of respondents leaned towards disagreeing with the statement “The current ruling Democratic Progressive Party (DPP) is no different from the Chinese Communist Party, and Taiwan lacks freedom of speech.”
Regarding skepticism theories related to the United States, statements such as “The United States only wants to exploit Taiwan…” (48.2%), “The United States uses Taiwan to provoke China…” (37.9%), and “In case of Chinese military aggression against Taiwan, the United States definitely won’t intervene to assist Taiwan…” (40.5%) were areas where respondents tended to agree.

Background:

Based on Doublethink Lab’s previous research, reports and observation, the strategy of information manipulation in Taiwan has shifted towards conspiracy types of narrative making it more challenging to debunk. These conspiratorial narratives also further establish and reinforce cognitive bias with the long term objective to increase citizens’ distrust towards the government and overall polarization in Taiwanese society.

For the 2024 Taiwan presidential election, Doublethink Lab conducted an election survey via telephone interviews and exit poll surveys. The surveys aim to better understand the context in which Taiwan’s society and overall public opinion are subjected to foreign information manipulation and evaluate the scope and effectiveness of its narrative dissemination. Our survey asked questions related to (foreign) information manipulation narratives that we have observed prior to the election, and explored the impact of these information manipulation narratives on the voting intentions, democratic satisfaction in Taiwan and other factors.

This project is part of the Doublethink Lab’s 2024 MHub project. In response to potential Foreign Information Manipulation Interference (FIMI) during the early 2024 presidential election in Taiwan, Doublethink Lab collaborated with domestic fact-checking organizations, research units, and civil society groups. Together, we integrated civilian resources and analytical capabilities through the “Foreign Interference MHub” project, collectively engaging in the observation of foreign influence in elections. For further details about the project, please refer to the public announcement.

Due to Taiwan’s role in global geopolitical affairs makes the outcome of this presidential election a significant demonstration for the development of the global democratic society and against the expansion of authoritarianism.The election will also have profound effects on the political situation and regional stability in the Asia-Pacific region. Below, we will first release the data from our exit-poll survey. We anticipate providing a more comprehensive analysis and report around the Lunar New Year for reference.

Relevant projects in the previous elections:

Since Doublethink Lab’s establishment in 2019, we have not only conducted exit polls for the 2020 presidential election, but also executed similar exit-polls and online surveys during the 2022 local elections. The survey data and analysis results from those election cycles have been provided.

Methodology:

The telephone interviews in this project were conducted in a 7:3 ratio between landline and mobile phone numbers, with a total of no fewer than 1500 valid samples. At a 95% confidence level, the sampling error is within ±2.53%. The survey targeted Taiwanese citizens aged 20 and above with voting rights, conducted between January 1 and January 10, 2024, and employed post-stratification weighting. For landline numbers, residential phone data in Taiwan was used as the sampling frame, employing stratified sampling based on county and city. Mobile numbers were randomly selected using the last five digits in conjunction with the "Status of Allocation of Numbers for Mobile Communication Network Business" published by the National Communications Commission (NCC).

The successful survey samples were weighted according to the latest population statistics from the Ministry of the Interior, adjusting for variables such as gender, age group, education level, and region to ensure the survey samples were representative of the population. The weighting method utilized a multivariate iterative proportional fitting (RAKING) process.

Raw Data (English)：https://docs.google.com/spreadsheets/d/17HvwsVZE8XO6n4nZAEGYPey3U2QVfM4Q/edit#gid=2059943114

The daily progress of the nationwide telephone interviews conducted in this survey is as follows.

Preliminary conclusion:

Through a brief cross-analysis, we have derived the following preliminary conclusions. Further cross-analysis and correlation results will be presented in subsequent comprehensive analysis reports.

In both surveys, we can see that the influence of “Online video content creators” and “Key Opinion Leaders (KOLs or influencers) on social platforms” on politics in Taiwan continues to grow. Compared with the data from our survey of the 2022 local elections (21.8%) increased around 18% to approximately 40%.

The largest proportion of respondents believe the DPP is primarily responsible for the creation and spread of election related information manipulation (Telephone survey: 26.6%, Taoyuan exit poll: 36.55%). Compared with the online survey from our survey to the 2022 local election, it slightly increased by around 3%.

A democratic system continues to be preferred by most respondents as the best system for Taiwan, with more than half of respondents from both surveys responding that they are satisfied with Taiwan’s democracy, and trust the election system. For supporters of each party, the order of both preference for democracy and trust in the electoral system is: DPP, TPP, then KMT. What is worth mentioning is, compared to the proportion that believes authoritarianism might be preferable in certain situations, there has been a slight increase in the proportion that perceives no significant difference between political systems. In contrast to the 8.9% recorded in the 2022 online survey, the current percentages show a slight uptick in both export and telephone interviews, with 10.36% in export and 19.6% in telephone interviews.

In both surveys, over 60% of respondents disagree with the statement that “The current ruling party DPP is no different than the CCP, and Taiwan does not have freedom of speech.” DPP supporters have the highest percentage of disagreement, followed by the Taiwan People’s Party, and then the Kuomintang (KMT).

On average, narratives pertaining to US skepticism are believed by around half of all respondents. For the narrative “the US cannot be trusted, and just wants to hollow out and use Taiwan,” 80% of DPP supporters disagreed, and respectively 80% and 70% of both KMT and TPP supporters agreed.

As the previous point indicates, overall we see significant polarization between DPP and the opposition KMT and TPP parties along the lines of key strategic narratives observed in the CCP’s information manipulation campaigns targeting Taiwan, including US skepticism, and attacks on the DPP for alleged corruption and incompetence.

Limitations:

Sampling limitations

The survey data are weighted estimates based on the latest population statistics from the Ministry of the Interior, using variables such as gender, age group, education level, and regional differences for population estimation to ensure that the surveyed sample is representative of the population. However, there are still limitations to the data. To some extent, random phone sampling shares the nature of voluntary samples, as individuals who are more willing to participate in interviews may have different demographic characteristics or attitudes than those who refuse. This could impact how representative our sample is.

Potential bias and distrust towards the survey planning and implementing organization/interviewers

Telephone interviews have the limitation that respondents cannot see the true identity of the interviewers, which may lead to distrust towards the content of the interview and the survey itself. In addition, while this telephone survey was carried out by a politically neutral polling company, as indicated in the informed consent responsibility, the interview guidance explicitly states at the beginning, “We are Doublethink Lab…,” which may evoke existing bias of the survey planning unit and pose potential challenges to the implementation of this telephone survey.

Acknowledgement:

This project would like to give l special thanks to Dr. Thung-Hong Lin, a full-time research fellow at the Institute of Sociology, Academia Sinica, for providing numerous suggestions and ideas in aspects such as questionnaire design, research logic, and sampling methods during the research preparation stage. Additionally, gratitude is expressed to DSIGroup company, for conducting telephone interviews in accordance with the Doublethink Lab’s planning. All matters arising from this project and report are the sole responsibility of the Doublethink Lab.

Research data:

In this survey, in order to avoid respondents' intuitive selection of options placed at the beginning and to prevent potential biases caused by the order of questions, presents some questions in each section in a random order, and the options within some questions are also randomly sorted.

The sampling and weighting methods designed for this survey are intended to estimate the overall social situation in Taiwan. They may not be entirely the same population that "actually comes out to vote" on the voting day.

Information Habits

Political Preferences

Views on Issues

Impressions of the Presidential Candidates in this Election

Q37. As this question involves open-ended responses, the content provided by respondents is still being organized and will be shared in the comprehensive analysis report later.