The Potential Association Between Fake News and Political Identity
Tsunglin-Wu
2026-05-06
Origin of the Research
The origin of this study stems from the researcher’s firsthand observations of the contemporary, everyday social media environment in Taiwan. In today’s network society, platforms such as Threads, Instagram, and Facebook have become the primary channels through which the public receives information and expresses political stances. However, driven by algorithms, these social platforms exhibit severe phenomena of “group stratification and echo-chamber polarization.”
The researcher observed that whether in the comment sections of Facebook or on Threads—a platform that has risen in popularity in recent years with text-based exchanges at its core—whenever posts regarding the “governing party” or “specific political candidates” appear, the comment sections beneath them almost instantly degenerate into fierce political flame wars, overwhelmed by emotional antagonism and invective.Even more concerning is that while a few rational users in the comment sections occasionally attempt to engage in “popular science explanation” or fact clarification, further tracing the core issues driving these netizens’ disputes reveals a shocking reality: the events causing the public to engage in frenzied opposition and endless fighting are, at their core, deliberately manipulated fake news or rage-bait posts.
This social media spectacle of “falling into collective polarization over non-existent non-issues” prompted the researcher to reflect deeply: In an era where fake news is so rampant, how exactly does disinformation trigger the emotions of individuals with specific party identifications? And how do citizens’ political leanings influence their identification of and reaction to false information?
Taking this social media reality as its starting point, this study embarks on an in-depth investigation into these questions.
Research Motivations
Based on the aforementioned everyday observations, the primary motivation of this study centers on examining how individuals’ social backgrounds and political identities interactively influence their information perception and partisan attitudes when confronted with rampant disinformation.First, the study aims to explore media literacy blind spots across different social strata and life experiences. While the contemporary internet ecosystem has rendered the spread and prevalence of fake news more severe than ever, the public does not receive or perceive this information equally. This study seeks to investigate whether individuals’ ability to detect disinformation—circulating via mobile phones, online networks, traditional print media, or foreign forces—is profoundly shaped by the interaction of their personal life experiences, generational cohorts, and social classes. Intuition in past research often assumed that groups with higher education, income, or social status possess superior information defense mechanisms, or that older adults linger in blind spots due to the digital divide. However, could wealth and social status also breed a form of intellectual overconfidence, thereby inversely skewing their self-assessment regarding susceptibility to fake news? Moreover, to what extent do closed social circles across different income brackets create disparities in the dissemination and fact-checking of false information? A key motivation of this study is to elucidate the potential media literacy blind spots experienced by various age and class cohorts within an increasingly clamorous social media environment.Second, the study seeks to deconstruct the filtering mechanisms of “party identification” and “national identity” in disinformation perception. The unceasing political flame wars on modern social media reflect how voters’ inner political orientations deeply color their cognitive processing of false information. This study explores how voters with diverse political leanings, varying levels of favorable or unfavorable attitudes toward specific political parties, and differing degrees of Taiwanese versus Chinese identity filter and interpret information once it has been consolidated by platform algorithms. For instance, does the proliferation of fake news universally erode public trust in political figures, or does it trigger defensive psychology among voters of specific camps, thereby generating reverse interpretations? Furthermore, beyond the black-and-white political battles on social platforms, what is the sensitivity toward fake news among voters situated in the middle or gray areas of Taiwanese identity and party preference? By leveraging empirical data, this study aims to clarify how national identity and party preference function as “optical filters” that shape citizens’ interpretive logic when processing false information.
Research Methods and Data
This study employs Multiple Correspondence Analysis (MCA) as the core analytical method for processing categorical data. The aim is to uncover latent variables underlying the survey items through dimensionality reduction techniques, while examining the internal consistency and relational structure among variables. During computation, MCA converts complex cross-tabulations of categorical variables into a low-dimensional perceptual map. By calculating the relative distances and coordinates of category points in space, MCA visually depicts the degree of association among variables, where proximity indicates a higher statistical relationship.In terms of model interpretation, this study delineates respondents’ cognitive characteristics and domain dynamics based on spatial quadrants constructed by coordinate axes. To ensure analytical quality, the study refers to the squared cosine (\(\cos^2\)) values to evaluate the contribution of each category to the dimensions, supported by scree plots for dimension selection. Considering that the number of response categories may affect total inertia, the variable design stage aimed to balance options across items to mitigate bias effects. Although MCA manages high-dimensional data in social sciences where the proportion of explained variance may remain modest, the researcher adheres strictly to the statistical logic outlined above for in-depth interpretation, thereby presenting the underlying structure and academic implications of the data comprehensively (Liu, 2016).The data selected for this study comes from the survey dataset titled “Fake News, Echo Chambers, and Polarization in Taiwan’s Social Media: An Innovative Human-Machine Integrated Research Approach,” principaled by Professor Yu-tzung Chang of the Department of Political Science at National Taiwan University. Funded by the Ministry of Science and Technology, the survey targeted Taiwanese citizens aged 20 and above as the population and collected data via an online survey method. Executed between November 1, 2021, and February 25, 2022, the survey gathered 1,058 valid samples.Three primary reasons guided the selection of this dataset. First, the survey topic aligns closely with this study, focusing on social media usage, disinformation transmission, echo-chamber effects, and political attitudes—critical factors influencing public information reception and political behavior in contemporary digital society. Second, the survey covers rich demographic background variables and political attitude indicators (including age, education, income, party identification, national identity, media consumption habits, and political participation), enabling multi-faceted analysis of respondent traits and attitudes to enhance explanatory power. Finally, using an online survey method ensures the sample primarily comprises digital citizens actively engaging with social media, reflecting the actual environment of disinformation spread and echo chamber formation. Furthermore, the survey period coincided with a highly active period in Taiwanese politics when public attention to social issues was elevated, making it ideal for observing the relationships among political attitudes, information reception behaviors, and social media usage (Chang & Shih, 2021). From this survey, 18 items most relevant to political orientation and disinformation perception were selected for MCA analysis (see Appendix for the full list of items and options).
Research Findings
This indicates that Dimensions 1 and 2 contribute the highest percentage of variance; therefore, Dimensions 1 and 2 are utilized as the primary analytical dimensions.
The data reveals an intuitive phenomenon: believing that newspapers and
magazines never spread fake news (Print: Never), believing that mobile
phones never spread fake news (Mobile: Never), believing that mobile
phones seldom spread fake news (Mobile: Seldom), and believing that fake
news has not altered trust in political figures (Trust Govt: Same)
cluster together.
Regarding this phenomenon, I argue that what it reflects behind the scenes is not a rock-solid, rational belief in the democratic system among respondents, but rather a “logical closed loop resulting from a lack of media literacy.”Looking at media literacy and environmental awareness, against the modern background of a highly digitized society rife with information warfare, these respondents subjectively believe that traditional print media or mobile phones “never” or only “seldom” disseminate fake news. In terms of objective reality, this exposes their extreme lack of basic digital media literacy. They remain entirely unaware of the rampant misinformation, content farms, or deliberately manipulated public opinion around them, essentially living in a severe “information blind spot.”Furthermore, looking at the causal logic, this perfectly explains why they chose “fake news does not change their trust in the democratic system.” Since individuals do not subjectively perceive the existence of fake news in mainstream media or on their mobile phones, fake news is, for them, a “non-existent non-issue.” Given that no threat is perceived, it is naturally impossible for fake news to exert any substantive impact or shock on their political trust or institutional identification with democracy. In short, this group’s “unaffected institutional trust” is essentially an immunity effect brought about by ignorance. Lacking the intellectual capacity and skill to identify false information, they view a noise-filled media environment as peaceful and harmonious. This extreme obtuseness toward their actual environment ultimately manifests in the data, rather ironically, as a form of “blind institutional stability.”
Having this set of intuitively observable data with potential connections reinforces the credibility of the subsequent analysis on underlying associations.
Generations, Local Identity, and Alertness Blind Spots in Digital New Technologies
First, the data demonstrates a potential correlation between perceiving mobile phones as frequently spreading fake news (Mobile: Often), being middle-aged (Age: Middle), holding a strong Taiwanese identity (TW ID: Strong), possessing a moderate Chinese identity (CN ID: Moderate), and maintaining a neutral stance toward the Taiwan People’s Party (TPP: Neutral).
Regarding this phenomenon, this study provides the following analysis. In terms of age and media literacy, the middle-aged group (aged 50 to 64) consists of native-born Taiwanese. During their upbringing and the formation of their subjective identity, local Taiwanese identity gradually deepened, which explains why they demonstrate a localized consciousness characterized by “strong Taiwanese identity and moderate Chinese identity.” Furthermore, modern electronic products and smartphones were not yet prevalent during this group’s youth. It was not until around 2021, when this survey was conducted, that the widespread adoption of smartphones and social media made it highly likely that they or their friends had personal experiences of “being misled by fake news on mobile phones.” Because this generation lacked systematic media literacy training during their formative years, they exhibit greater alertness and vulnerability when bombarded with fake news from new technologies. By contrast, older generations may be less proficient in using smartphones or remain unaware even after being deceived, making them less sensitive to this variable than the middle-aged group.
In terms of party identification and political attitudes, the prevailing neutral and wait-and-see stance toward the Taiwan People’s Party (TPP) among this group is primarily related to the political context of that time. When this survey was administered in 2021, the TPP was still an emerging political party, and its core figure at the time, Ko Wen-je, had not yet been involved in subsequent legal cases. Compared to the two traditional major political parties (the DPP and KMT), the middle-aged group tended to maintain a cautious and reserved attitude toward this new party. Moreover, the TPP at that time was highly centered around Ko Wen-je as an individual, and the party’s overall long-term direction and platform remained ambiguous and uncertain, which further led the middle-aged group to adopt a watchful position rather than rushing to take a stand.
Knowledge Arrogance and System Defense Mechanisms Among High Socioeconomic Status Individuals
Next, a potential correlation exists among self-assessed invulnerability to fake news (Impact: V.Low), high social status (Status: High), and a unfavorable attitude toward the Taiwan People’s Party (TPP: Dislike).
Regarding information identification and the psychological traits of high socioeconomic status (SES) individuals, high-SES individuals typically have received higher education. This equips them with keener discernment and critical thinking skills when processing information, leading them to confidently believe that they are “not affected by fake news at all.” However, this trait can also be accompanied by a degree of intellectual arrogance. Due to their past life experiences and high academic background, even when these individuals are unwittingly misled by sophisticated fake news, their overconfidence in their own judgment often makes them more stubborn—remaining firmly convinced even after being deceived—which further solidifies their self-assessment of being immune to influence.
Furthermore, the reasons behind high-SES individuals’ aversion to the Taiwan People’s Party (TPP) can be analyzed from two perspectives. The first involves social class and label identification. As political discourse has evolved, the term “Little Grass” (Xiao Cao) in contemporary Taiwanese political context has gradually shifted into a label with derogatory connotations and grass-roots, populist colorings. For high-SES individuals positioned at the top of society who value their class image and intellectual taste, their inherent class superiority (or arrogance) creates a strong psychological sense of distance from this group, leading them to refuse association with it—a sentiment directly projected as dislike toward the TPP. The second perspective involves system defense and resistance from vested interests. The TPP’s political platform primarily appeals to opposing the existing system and breaking established structures, thereby attracting a large number of young supporters dissatisfied with the status quo. However, high-SES individuals are inherently “vested beneficiaries” of the current social system, and the stable operation of the existing framework best aligns with their class interests. When an emerging political party attempts to challenge, disrupt, or even deconstruct the very system from which they benefit, the high-SES group naturally develops an institutional aversion and rejection toward it, driven by defensive psychology and interest-based calculations.
Media Framing and Subjective Threat Construction Among Upper-Middle Socioeconomic Status Individuals
Furthermore, a potential correlation exists between upper-middle social status (Status: Mid-High) and the perception that China or foreign forces frequently spread fake news (Foreign: Often).
Looking through the lens of information acquisition and media framing, upper-middle status individuals typically possess better educational backgrounds and actively consume a broad range of international and cross-strait political and economic news in their daily lives. However, the information this group receives is often heavily influenced by long-standing media framing effects. Mainstream Taiwanese media frequently employ specific labels and stereotypes when shaping the image of China, such as overemphasizing its “underdeveloped aspects (e.g., public restrooms without doors)” or focusing strictly on its institutional features (e.g., internet censorship, pervasive wiretapping, and total surveillance). Over time, this highly selective news rendering subtly deepens the upper-middle SES group’s established perception of China as a primary threat source and a generator of fake news.
Examining this phenomenon through the lens of social constructivism further underscores the “subjective constructivism” of national security and threat perceptions. Constructivism posits that threats in international politics are not determined entirely by objective material capabilities, but rather depend on how actors “understand” and “define” one another. From this standpoint, the image of a “frightening China proliferating fake news” in the eyes of the upper-middle SES group is largely a product woven together by media discourse, historical memory, and collective social anxiety.
In other words, while China’s objective threat level certainly exists, the extent to which its terrifying image is amplified and fixed in the minds of the Taiwanese public might be a manifestation of “fear generated from within.” It is precisely the internal anxiety within Taiwanese society over the potential subversion of its free and democratic system that, projected through the media lens, inversely shapes China into a pervasive shadow constantly disseminating fake news day and night. Ultimately, this subjectively constructed fear translates into the statistically observed association between upper-middle SES individuals and a heightened vigilance toward foreign forces.
Belief Defense and Reverse Interpretation Mechanisms Among the Elderly
Subsequently, the data demonstrates a potential correlation between the elderly demographic cohort (Age: Senior) and the perception that fake news paradoxically increases trust in political figures (Trust Govt: Up).
Looking through the lens of media literacy and educational background, older generations generally lacked training in modern digital media literacy during their formative years, and their average education levels were not as widespread as those of younger generations. This objectively renders them a high-risk group vulnerable to being misled and deceived when confronted with complex, hard-to-verify modern information warfare.
However, when these seniors chose “fake news increases trust in political figures” on the survey, the underlying psychological mechanism was not blind gullibility, but rather a “reverse interpretation stemming from self-perceived insight into the truth.” Specifically, when seniors see a news report on media that is unfavorable to a political figure they support and are subsequently told it is “fake news,” their thinking leaps to another level. They angrily conclude: “This so-called ‘fake news’ is actually ‘another, even bigger fake news’ deliberately fabricated by opposing politicians to smear and suppress the person I support.” Under this psychological projection of “politicians smearing each other,” seniors do not lose faith in the political figure; instead, they develop strong sympathy and cohesion because they feel “he is being persecuted by the opposing camp using fake news.” Consequently, this heightened defensiveness and political attachment subjectively deepen their trust in that politician. What lies behind the data is precisely the arrogance of seniors who believe they have seen through politicians’ tricks, alongside political biases that have grown even more entrenched in the era of rampant misinformation.
Constitutional Identity Countering Binary Opposition and the Choice of Implicit Centrist Voters
The data indicates a potential correlation between a moderate Taiwanese identity (TW ID: Moderate) and a neutral stance toward the Kuomintang (KMT: Neutral).
First, it is essential to dismantle the binary fallacy of black-and-white identity. In Taiwan’s political context, a “moderate Taiwanese identity” is often intuitively misconstrued as a “strong Chinese identity.” However, this statistical association effectively overturns such a simplified dichotomy. Although these respondents feel a relative distance from the primary label of “Taiwan,” this does not mean they emotionally shift their identification toward the People’s Republic of China. Instead, they are more likely to project their national affection onto the constitutional framework and historical legitimacy of the “Republic of China” (ROC). This fluid and mediating state between “Taiwanese identity” and “Chinese identity” explains why they exhibit a mild and non-radical stance on cross-strait relations and identity issues.
Furthermore, the reasoning behind their neutral stance—rather than strong support—toward the Kuomintang (KMT) can be attributed to cognitive divergence regarding party positions on national identity (Taiwan vs. ROC). On state orientation, these voters hold a fundamental cognitive divergence from the Democratic Progressive Party’s (DPP) strong emphasis on “Taiwanese subjectivity and Taiwanese sovereign independence” in recent years. Yet, the reason they have not strongly gravitated toward the KMT is that the KMT’s cross-strait discourse and defense of the “Republic of China” often appear multifaceted and ambiguous, failing to fully satisfy the expectations of voters whose core identity rests on the ROC. Consequently, while the DPP’s “Taiwan” narrative alienates them, the KMT’s current posture fails to win their complete buy-in. Caught in the awkward gap between a “moderate Taiwanese identity” and the national identity trajectories of the two major parties, these voters ultimately make a rational and cautious choice to maintain an impartial, “neutral” wait-and-see position toward the KMT.
The Free Speech Cost Theory and the “Tearful Support” Motive Among Lean Green Voters
Additionally, the data presents a highly provocative combination: a potential correlation exists between leaning toward the Pan-Green camp (Lean: Lt Green) and the perception that fake news paradoxically increases trust in the democratic system (Trust Demo: Up).
From the perspective of an objective awareness of tolerance within democratic systems, the reason these lean-green voters believe disinformation increases democratic trust is that they keenly observe the proliferation of fake news in daily life; however, instead of reacting with panic, they interpret this phenomenon as concrete proof that the nation enjoys a high degree of freedom of speech. In their cognitive logic, precisely because Taiwan is a genuine democracy, the government has not resorted to heavy-handed restrictions on speech, internet censorship, or the suppression of public discourse to control opinion. Thus, the presence and circulation of fake news paradoxically corroborate the openness and tolerance of the democratic framework—“Noise exists, but it is the price of freedom.” This intellectual awareness subsequently transforms into a deeper trust in Taiwan’s democratic institutions. This aligns closely with the long-standing party brand crafted by the DPP. Within Taiwan’s historical political context, the DPP has long projected an image of “fighting for freedom, breaking authoritarianism, and representing local democracy.” When these respondents subjectively perceive that the nation is practicing such democratic values—protecting freedom without censoring speech—this positive institutional experience naturally projects onto the Pan-Green camp.
However, this partisan attachment has faced severe tension and challenges from real-world governance in recent years. Undeniably, Green supporters have accumulated internal dissatisfaction during major policy controversies (such as Ractopamine pork imports), occasionally experiencing disillusionment with the ruling party’s perceived heavy-handedness or mismanagement. Intriguingly, even when disappointed or angered by specific policies, these voters still view the Pan-Green camp as the option that better represents core democratic values regarding “national orientation” and “free speech frameworks.” In other words, despite frustration with recent concrete governance, witnessing the state’s refusal to restrict free speech leads them to maintain a baseline trust in democratic institutions and the Pan-Green camp. This complex voter psychology—“criticizing misgovernance on one hand while tearfully supporting out of a desire to defend freedom on the other”—ultimately manifests in the data as the phenomenon where perceived disinformation paradoxically increases democratic trust, leaning toward light-green rather than deep-green alignment.
Breaking Through Central Tendency Defenses and the Traits of Stigma Resistance
Self-assessed poverty (Income: Poor) and self-assessed comfortable status (Income: Comfortable) exhibit a potential correlation.
Looking through the lens of survey-response psychology, the general public in most questionnaire surveys tends to succumb to central tendency—conservatively claiming their household income is merely “average” due to social norms, modesty, or defensive psychological mechanisms. By contrast, for the genuinely top-tier “wealthy” class, superior affluence is an undeniable objective reality. However, respondents who neither choose “average” to follow the crowd nor belong to the genuinely elite “wealthy” class, yet explicitly and candidly locate themselves at the “poor” or “comfortable” levels, demonstrate a unique personality trait in itself.
Regarding their psychological traits and expressive capacities, respondents in these two groups typically possess higher self-awareness and are better at articulating their true selves to the outside world. They refuse to blindly conform to the collective mask of “considering oneself average.” Whether candidly confronting their own financial hardship (self-assessed as poor) or holding confidence in their accumulated resources (self-assessed as comfortable), both exhibit highly clear and distinct perspectives on their personal circumstances.
Therefore, the statistical association between Income: Poor and Income: Comfortable arises not because their material conditions are similar, but because they share a common underlying personality foundation: they are both highly self-aware, outspoken respondents within the population who are willing to express their true viewpoints. Ultimately, this psychological candor during survey completion transforms in the data into an invisible thread, curiously linking these two seemingly opposing economic strata together.
Conclusion
The contemporary political information environment in Taiwan is undergoing severe structural polarization. This uncertainty no longer remains at the macro level of Pan-Blue versus Pan-Green confrontation, but directly transmits into micro-level “information filtering mechanisms” and “degrees of institutional trust” through individuals’ digital media perceptions and class backgrounds. In Taiwan’s political ecosystem, “national identity” and “party inclination” complement each other, forming the core filter through which voters screen political information (Hsu, 2002).
Exploratory Data Analysis (EDA) in this study reveals that actors who perceive mobile phones as often spreading fake news (Mobile: Often) are significantly associated with middle-aged groups (Age: Middle), strong Taiwanese identity (TW ID: Strong), non-strong/moderate Chinese identity (CN ID: Moderate), and a neutral attitude toward the Taiwan People’s Party (TPP: Neutral). This phenomenon can be explained by referencing Liu’s (2016) concept of the “Circuit pattern.” When voters construct a strong localized national identity during “symbolic processing,” their ideology functions like an electrified defense network, causing them to exhibit high vigilance toward mobile messages that might undermine Taiwanese sovereignty. As for this group’s neutral stance toward the then-emerging TPP, it further reflects the cautious evaluative posture adopted by “closet partisans” when confronting ambiguous political symbols to avoid interpersonal tension brought by a polarized social media environment (Liu & Chu, 2012).
However, public defensive psychology toward the information environment and party rejection are not solely driven by national sentiment; the capital class an individual occupies within the social structure also profoundly shapes their intellectual self-confidence and inclination toward system defense. Exploratory data analysis demonstrates that high social status individuals (Status: High) self-assess as being not at all affected by fake news (Impact: V.Low) and are highly correlated with disliking the TPP (TPP: Dislike). This can be explained by Pierre Bourdieu’s theory of “Distinction” (1984), where high status individuals, possessing abundant cultural capital, tend to display strong intellectual self-confidence and cultural superiority in the information market. Such confidence often makes them stubborn due to overconfidence in their own judgment, even when facing intricately crafted fake news. Moreover, high status individuals are typically vested beneficiaries of the current system, exhibiting an institutional rejection toward emerging political parties that seek to challenge or deconstruct established structures—a reaction that is essentially an actor’s instinctual response to maintain social structural stability (Bourdieu, 1984).
It is worth noting that subjective economic capital does not necessarily translate into the critical thinking of cultural capital; mere wealth accumulation may instead construct another closed information blind spot. The analysis points out a positive association between subjective wealth (Income: Wealthy) and being affected by fake news (Impact: High), revealing that simple economic capital accumulation does not automatically come with rigorous logical reasoning training (Bourdieu, 1984). Wealthy groups usually inhabit highly trusted and homogeneous closed interpersonal networks—so-called “identity circles.” When fake news infiltrates such circles and is forwarded among peers, the credibility of the message is blindly amplified based on excessive trust in same-stratum connections (Liu, 2016). This collective blind belief within the circle often leads them to lower their guard, making them more susceptible to being misled (Allcott & Gentzkow, 2017).
Furthermore, the ambiguity of class perception itself is demonstrated in the data: the correlation between self-assessed poverty (Income: Poor) and self-assessed comfortable status (Income: Comfortable) suggests that people’s perception of their position in social space is often influenced by their “trajectory effect” (Bourdieu, 1984). Class identity does not depend entirely on objective income, but rather on how actors understand and define their relative position in resource distribution, thus presenting as a clustering phenomenon of class identity in the data.
The subjective construction of the information environment is further drawn by the interaction between mass media framing effects and generational habitual domains. The data indicates a high correlation between upper-middle social status (Status: Mid-High) and believing that China or foreign forces frequently spread fake news (Foreign: Often). As high consumers of information, the upper-middle socio-economic group’s cognition is deeply affected by long-term mass media “framing effects” (Na et al., 2023). Mainstream media’s selective rendering of China’s image (such as surveillance and threats) subjectively constructs a concrete threat image in this group’s mind. From a social constructivist perspective, this reflects a deep-seated anxiety within Taiwanese society regarding the potential subversion of the free democratic system, which, projected through the media lens, inversely shapes China into a pervasive source of fake news (Lee, 2023).
Conversely, at the level of generational differences, the elderly group (Age: Senior) exhibits a unique tendency to believe that fake news increases trust in political figures (Trust Govt: Up). This phenomenon can be viewed through the “Habitual Domains” theory: the long-accumulated information decoding habits of older generations, when facing pervasive media smearing, often trigger a “backlash effect.” This leads them to view fake news as political persecution by political opponents, which instead arouses sympathy and inversely reinforces sympathy for and cohesion around that political figure (Botambu Collins et al., 2021).
Finally, this subjective perception of the information environment ultimately feeds back into shaping and constructing the public’s overall confidence in the nation’s democratic system. The data shows that leaning toward the Pan-Green camp (Lean: Lt Green) is significantly associated with believing that fake news paradoxically increases trust in the democratic system (Trust Demo: Up). This provides strong empirical evidence for “partisan-motivated reasoning,” where Pan-Green voters, driven by support for the status quo and governing preferences, tend to believe that democratic institutions possess self-correcting and fact-checking capabilities (Liu & Tsai, 2016). When fake news is exposed, they view it as positive proof of a healthy system in operation, thereby strengthening their confidence in the institution; this positive reasoning result is essentially a psychological defense to align with their political preferences (Leeper & Slothuus, 2014).
However, on the other side of partisan motivation, the analysis also presents a special phenomenon of “immunity through ignorance”: the group believing that newspapers, magazines, or mobile phones “never” or only “seldom” spread fake news (Print: Never, Mobile: Never, Mobile: Seldom) displays unchanged trust in the democratic system (Trust Govt: Same). This confirms the structural existence of a “media literacy blind spot.” Although theory suggests that education levels and media consumption should correlate positively with discernment capabilities (Allcott & Gentzkow, 2017), when respondents mistake a noise-filled environment for being “peaceful and harmonious,” fake news becomes a non-existent non-issue to them. This sense of institutional stability arising from ignorance inversely shapes a unique “immunity effect,” causing this group to automatically disappear from the radar of information warfare and become a collective blind spot in the defense of democratic institutions.
References
張佑宗（2022）。臺灣社群媒體中的假新聞、同溫層與極化現象：一種人機整合創新的研究途徑（E10875）【原始數據】。取自中央研究院人文社會科學研究中心調查研究專題中心學術調查研究資料庫。
張佑宗、施琮仁（2021）。社群媒體使用與政治極化：以 2020 年臺灣總統大選為例。
徐火炎（2004）。台灣結、中國結與台灣心、中國情：台灣選舉中的符號政治。選舉研究，11（2），1-41。
劉正山（2016）。世代之間政治認同差異的圖像：以多重對應分析結合習慣領域視野進行的探索。習慣領域期刊，7（2），27-50。
劉正山、朱淑華（2012）。不中間的中間選民：以質性方法初探有政黨傾向選民隱藏政黨傾向的原因。東吳政治學報，30（4），177-233。
劉正山、蔡艾真（2016）。中間選民的迷思與隱性選民之政黨傾向初探。台灣選舉與民主化調查研究（台政刊），65-123。
徐夏春祥（主編）（2023）。傳播研究與實踐，13（2）。
Allcott, H., & Gentzkow, M. (2017). Social media and fake news in the 2016 election. Journal of Economic Perspectives, 31(2), 211-236.
Bourdieu, P. (1984). Distinction: A social critique of the judgement of taste ( R. Nice, Trans.). Harvard University Press.
Collins, B., Hoang, D. T., Nguyen, N. T., & Hwang, D. (2021). Trends in combating fake news on social media – a survey. Journal of Information and Telecommunication, 5(2), 247-266.
Olan, F., Jayawickrama, U., Arakpogun, E. O., Suklan, J., & Liu, S. (2024). Fake news on social media: The impact on society. Information Systems Frontiers, 26, 443-458.
Appendix
Fake News Variables
Do you think foreign forces frequently fabricate fake news? (x) <numeric>
# total N=1058 valid N=1058 mean=2.11 sd=0.78
Value | Label | N | Raw % | Valid % | Cum. %
--------------------------------------------------------
1 | Foreign: Daily | 250 | 23.63 | 23.63 | 23.63
2 | Foreign: Often | 461 | 43.57 | 43.57 | 67.20
3 | Foreign: Seldom | 330 | 31.19 | 31.19 | 98.39
4 | Foreign: Never | 17 | 1.61 | 1.61 | 100.00
<NA> | <NA> | 0 | 0.00 | <NA> | <NA>Do you think newspapers and magazines frequently spread fake news? (x) <numeric>
# total N=1058 valid N=1058 mean=2.39 sd=0.73
Value | Label | N | Raw % | Valid % | Cum. %
------------------------------------------------------
1 | Print: Daily | 125 | 11.81 | 11.81 | 11.81
2 | Print: Often | 419 | 39.60 | 39.60 | 51.42
3 | Print: Seldom | 487 | 46.03 | 46.03 | 97.45
4 | Print: Never | 27 | 2.55 | 2.55 | 100.00
<NA> | <NA> | 0 | 0.00 | <NA> | <NA>Do you think the internet frequently spreads fake news? (x) <numeric>
# total N=1058 valid N=1058 mean=1.66 sd=0.66
Value | Label | N | Raw % | Valid % | Cum. %
---------------------------------------------------------
1 | Internet: Daily | 466 | 44.05 | 44.05 | 44.05
2 | Internet: Often | 494 | 46.69 | 46.69 | 90.74
3 | Internet: Seldom | 92 | 8.70 | 8.70 | 99.43
4 | Internet: Never | 6 | 0.57 | 0.57 | 100.00
<NA> | <NA> | 0 | 0.00 | <NA> | <NA>Do you think mobile phones frequently spread fake news? (x) <numeric>
# total N=1058 valid N=1058 mean=1.95 sd=0.73
Value | Label | N | Raw % | Valid % | Cum. %
-------------------------------------------------------
1 | Mobile: Daily | 296 | 27.98 | 27.98 | 27.98
2 | Mobile: Often | 523 | 49.43 | 49.43 | 77.41
3 | Mobile: Seldom | 231 | 21.83 | 21.83 | 99.24
4 | Mobile: Never | 8 | 0.76 | 0.76 | 100.00
<NA> | <NA> | 0 | 0.00 | <NA> | <NA>Does fake news increase or decrease your trust in the democratic system? (x) <numeric>
# total N=1058 valid N=1058 mean=2.40 sd=0.65
Value | Label | N | Raw % | Valid % | Cum. %
---------------------------------------------------------
1 | Trust Demo: Up | 100 | 9.45 | 9.45 | 9.45
2 | Trust Demo: Same | 437 | 41.30 | 41.30 | 50.76
3 | Trust Demo: Down | 521 | 49.24 | 49.24 | 100.00
<NA> | <NA> | 0 | 0.00 | <NA> | <NA>Does fake news increase or decrease your trust in political figures? (x) <numeric>
# total N=1058 valid N=1058 mean=2.62 sd=0.65
Value | Label | N | Raw % | Valid % | Cum. %
---------------------------------------------------------
1 | Trust Govt: Up | 98 | 9.26 | 9.26 | 9.26
2 | Trust Govt: Same | 202 | 19.09 | 19.09 | 28.36
3 | Trust Govt: Down | 758 | 71.64 | 71.64 | 100.00
<NA> | <NA> | 0 | 0.00 | <NA> | <NA>Are you affected by fake news? (x) <numeric>
# total N=1058 valid N=1058 mean=2.53 sd=0.67
Value | Label | N | Raw % | Valid % | Cum. %
-------------------------------------------------------
1 | Impact: V.Low | 70 | 6.62 | 6.62 | 6.62
2 | Impact: Low | 386 | 36.48 | 36.48 | 43.10
3 | Impact: High | 569 | 53.78 | 53.78 | 96.88
4 | Impact: V.High | 33 | 3.12 | 3.12 | 100.00
<NA> | <NA> | 0 | 0.00 | <NA> | <NA>Over the past year, when encountering fake news, have you reported it in the comments? (x) <numeric>
# total N=1058 valid N=1058 mean=1.66 sd=0.47
Value | Label | N | Raw % | Valid % | Cum. %
----------------------------------------------------
1 | Report: Yes | 363 | 34.31 | 34.31 | 34.31
2 | Report: No | 695 | 65.69 | 65.69 | 100.00
<NA> | <NA> | 0 | 0.00 | <NA> | <NA>Political Orientations, Demographic Variables, and Feeling Thermometers
Which political party do you currently feel most inclined to support? (x) <numeric>
# total N=1058 valid N=995 mean=3.07 sd=1.22
Value | Label | N | Raw % | Valid % | Cum. %
-------------------------------------------------------
1 | Party: DPP | 203 | 19.19 | 20.40 | 20.40
2 | Party: KMT | 96 | 9.07 | 9.65 | 30.05
3 | Party: TPP | 128 | 12.10 | 12.86 | 42.91
4 | Party: Neutral | 568 | 53.69 | 57.09 | 100.00
<NA> | <NA> | 63 | 5.95 | <NA> | <NA>Which political camp do you consider yourself closer to (Pan-Blue or Pan-Green)? (x) <numeric>
# total N=1058 valid N=1058 mean=2.72 sd=0.76
Value | Label | N | Raw % | Valid % | Cum. %
---------------------------------------------------------
1 | Lean: Deep Blue | 64 | 6.05 | 6.05 | 6.05
2 | Lean: Lt Blue | 303 | 28.64 | 28.64 | 34.69
3 | Lean: Lt Green | 558 | 52.74 | 52.74 | 87.43
4 | Lean: Deep Green | 133 | 12.57 | 12.57 | 100.00
<NA> | <NA> | 0 | 0.00 | <NA> | <NA>What is your age? (x) <numeric>
# total N=1058 valid N=1058 mean=1.90 sd=0.85
Value | Label | N | Raw % | Valid % | Cum. %
----------------------------------------------------
1 | Age: Youth | 384 | 36.29 | 36.29 | 36.29
2 | Age: Adult | 455 | 43.01 | 43.01 | 79.30
3 | Age: Middle | 160 | 15.12 | 15.12 | 94.42
4 | Age: Senior | 59 | 5.58 | 5.58 | 100.00
<NA> | <NA> | 0 | 0.00 | <NA> | <NA>On a 1-to-10 scale, where would you place your subjective social status? (x) <numeric>
# total N=1058 valid N=1058 mean=1.97 sd=0.69
Value | Label | N | Raw % | Valid % | Cum. %
---------------------------------------------------------
1 | Status: Low | 248 | 23.44 | 23.44 | 23.44
2 | Status: Mid | 611 | 57.75 | 57.75 | 81.19
3 | Status: Mid-High | 181 | 17.11 | 17.11 | 98.30
4 | Status: High | 18 | 1.70 | 1.70 | 100.00
<NA> | <NA> | 0 | 0.00 | <NA> | <NA>What is your average monthly household income? (x) <numeric>
# total N=1058 valid N=1058 mean=2.27 sd=0.74
Value | Label | N | Raw % | Valid % | Cum. %
------------------------------------------------------------
1 | Income: Poor | 142 | 13.42 | 13.42 | 13.42
2 | Income: Average | 531 | 50.19 | 50.19 | 63.61
3 | Income: Comfortable | 339 | 32.04 | 32.04 | 95.65
4 | Income: Wealthy | 46 | 4.35 | 4.35 | 100.00
<NA> | <NA> | 0 | 0.00 | <NA> | <NA>Feeling toward the Democratic Progressive Party (DPP) (x) <numeric>
# total N=1058 valid N=1058 mean=1.83 sd=0.79
Value | Label | N | Raw % | Valid % | Cum. %
-----------------------------------------------------
1 | DPP: Dislike | 439 | 41.49 | 41.49 | 41.49
2 | DPP: Neutral | 363 | 34.31 | 34.31 | 75.80
3 | DPP: Like | 256 | 24.20 | 24.20 | 100.00
<NA> | <NA> | 0 | 0.00 | <NA> | <NA>Feeling toward the Kuomintang (KMT) (x) <numeric>
# total N=1058 valid N=1058 mean=1.60 sd=0.74
Value | Label | N | Raw % | Valid % | Cum. %
-----------------------------------------------------
1 | KMT: Dislike | 586 | 55.39 | 55.39 | 55.39
2 | KMT: Neutral | 307 | 29.02 | 29.02 | 84.40
3 | KMT: Like | 165 | 15.60 | 15.60 | 100.00
<NA> | <NA> | 0 | 0.00 | <NA> | <NA>Feeling toward the Taiwan People's Party (TPP) (x) <numeric>
# total N=1058 valid N=1058 mean=1.78 sd=0.76
Value | Label | N | Raw % | Valid % | Cum. %
-----------------------------------------------------
1 | TPP: Dislike | 446 | 42.16 | 42.16 | 42.16
2 | TPP: Neutral | 396 | 37.43 | 37.43 | 79.58
3 | TPP: Like | 216 | 20.42 | 20.42 | 100.00
<NA> | <NA> | 0 | 0.00 | <NA> | <NA>How strong is your identification as Taiwanese? (x) <numeric>
# total N=1058 valid N=1058 mean=1.81 sd=0.67
Value | Label | N | Raw % | Valid % | Cum. %
--------------------------------------------------------
1 | TW ID: V.Strong | 346 | 32.70 | 32.70 | 32.70
2 | TW ID: Strong | 581 | 54.91 | 54.91 | 87.62
3 | TW ID: Moderate | 119 | 11.25 | 11.25 | 98.87
4 | TW ID: Weak | 12 | 1.13 | 1.13 | 100.00
<NA> | <NA> | 0 | 0.00 | <NA> | <NA>How strong is your identification as Chinese? (x) <numeric>
# total N=1058 valid N=1058 mean=3.21 sd=0.75
Value | Label | N | Raw % | Valid % | Cum. %
--------------------------------------------------------
1 | CN ID: V.Strong | 23 | 2.17 | 2.17 | 2.17
2 | CN ID: Strong | 142 | 13.42 | 13.42 | 15.60
3 | CN ID: Moderate | 483 | 45.65 | 45.65 | 61.25
4 | CN ID: Weak | 410 | 38.75 | 38.75 | 100.00
<NA> | <NA> | 0 | 0.00 | <NA> | <NA>Variable Data Distributions
The total number of people after cleaning: 995 Foreign Print Internet Mobile Trust_Demo Trust_Govt Impact
"factor" "factor" "factor" "factor" "factor" "factor" "factor"
Report Party Lean TW_ID CN_ID Age Status
"factor" "factor" "factor" "factor" "factor" "factor" "factor"
Income DPP_Like KMT_Like TPP_Like
"factor" "factor" "factor" "factor" Foreign Print Internet Mobile Trust_Demo Trust_Govt Impact Report Party
1:234 1:120 1:441 1:281 1: 91 1: 93 1: 64 1:331 1:203
2:440 2:393 2:467 2:486 2:411 2:189 2:359 2:664 2: 96
3:306 3:455 3: 83 3:222 3:493 3:713 3:540 3:128
4: 15 4: 27 4: 4 4: 6 4: 32 4:568
Lean TW_ID CN_ID Age Status Income DPP_Like KMT_Like TPP_Like
1: 59 1:322 1: 20 1:352 1:234 1:134 1:414 1:537 1:422
2:297 2:551 2:136 2:435 2:578 2:502 2:335 2:298 2:370
3:516 3:114 3:462 3:151 3:169 3:318 3:246 3:160 3:203
4:123 4: 8 4:377 4: 57 4: 14 4: 41 
Social Circle Blind Belief and the Decoupling of Cultural Capital Among the Wealthy
Finally, this study uncovered an additional potential correlation between being affected by fake news (Impact: High) and high household income (Income: Wealthy). While this result may appear surprising at first glance, it proves entirely logical upon closer theoretical examination.
First, highly affluent individuals are often the primary targets for various types of disinformation and precision scam operations. More importantly, wealthy groups typically inhabit specific, closed social circles (wealthy elite circles). When a piece of disinformation (such as a specific investment rumor, insider claim, or political/economic myth) successfully infiltrates this circle and circulates among peers, its credibility is blindly amplified based on trust within the same social stratum. This group blind belief within elite circles often causes them to lower their guard and skip independent fact-checking steps, making them more vulnerable to collective deception.
Second, it is necessary to distinguish the essential difference between “economic wealth” and “high social status.” Monetary affluence does not directly equate to high cultural capital or elevated social standing. The wealthy cohort in this variable may include traditional business owners, landholders, or emerging commercial investors. Although these individuals have accumulated vast financial capital, their backgrounds may not have included rigorous training in critical thinking or media literacy. Therefore, their ability to discern and defend against fake news when facing ambiguous digital information cannot necessarily be equated with that of high-status (high cultural capital) intellectuals.
Despite potential deficiencies in media literacy, the confidence derived from vast wealth makes their hubris and arrogance toward information no less than that of high-SES individuals. This overconfidence spawned by wealth ultimately becomes an intellectual blind spot in the digital age, leaving the objective reality of their “susceptibility to fake news” exposed plainly in the data.