異質性資訊、討論和態度明確性：選擇性迴避的調節效果 - 政大學術集成

全文

(1)國立政治大學國際傳播英語碩士學位學程 International Master’s Program in International Communication Studies College of Communication National Chengchi University. 碩士論文. 治政 Thesis Master’s 大. 立. ‧. ‧ 國. 學. 異質性資訊、討論和態度明確性：選擇性迴避的調節效果. y. Nat. er. io. sit. Heterogeneous Information Exposure, Discussion and Attitude. n. al Certainty: The Moderating Role of Selective i v Avoidance Ch. n U engchi. Student: 易淳敏 Chun-min Yi Advisor: 施琮仁教授 Dr. Tsung-jen Shih. 中華民國 110 年 01 月 January 2021 DOI:10.6814/NCCU202100303.

(2) 國立政治大學國際傳播英語碩士學位學程 International Master’s Program in International Communication Studies College of Communication National Chengchi University. 碩士論文 Master’s 治政 Thesis. 大. 立. ‧. ‧ 國. 學. 異質性資訊、討論和態度明確性：選擇性迴避的調節效果. y. Nat. er. io. sit. Heterogeneous Information Exposure, Discussion and Attitude. n. Certainty: The Moderating Role of Selective Avoidance a. iv l C n hengchi U. Student: 易淳敏 Chun-min Yi Advisor: 施琮仁教授 Dr. Tsung-jen Shih. 中華民國 110 年 01 月 January 2021. DOI:10.6814/NCCU202100303.

(3) 異質性資訊、討論和態度明確性：選擇性迴避的調節效果 Heterogeneous Information Exposure, Discussion and Attitude Certainty: The Moderating Role of Selective Avoidance. Student: Chun-min Yi. 研究生：易淳敏. Advisor: Dr. Tsung-jen Shih. 指導教授：施琮仁. 國立政治大學. 學. 國際傳播英語碩士學位學程. ‧. io. sit. y. Nat. 碩士論文. n. al. er. ‧ 國. 立. 政治大. Ch. engchi. i Un. v. A Thesis Submitted to International Master’s Program in International Communication Studies National Chengchi University In Partial Fulfillment of the Requirement For the degree of Master of Arts. 中華民國 110 年 01 月 January 2021 DOI:10.6814/NCCU202100303.

(4) Acknowledgements First and foremost, I would like to show my biggest appreciation to my parents for supporting me financially to complete my master’s study, for their love, caring, and sacrifices for educating and preparing me for my future.. I would like to express my sincere gratitude to my advisor Prof. Tsung-jen Shih (施琮仁), for his patience and invaluable guidance throughout this research.. 政治大 me the methodology to carry out the research, explained obscure theories with 立 His immense knowledge, vision, and motivation have inspired me. He has taught. ‧ 國. 學. clarity, and helped me to sort out my thoughts as clearly as possible. I would also like to thank him for his friendship, encouragement, and great sense of humor.. ‧. Besides my advisor, I would like to extend my thanks to the rest of my thesis. sit. y. Nat. committee: Prof. Chiung-wen Hsu (許瓊文), and Prof. Wen-ping Kuo (郭文平).. n. al. er. io. They have raised critical questions on this dissertation and gave me insightful. i Un. v. feedback and encouragement to motivate me to finalize this research.. Ch. engchi. Furthermore, I am particularly thankful to Prof. Shefong Chung (鍾適芳) and her colleagues for showing their support and offering me a place to stay whenever I need to be in Taipei during this time. There were many obstacles and self-doubt throughout writing this research paper that I did not think I would get done. It was the constant encouragement, ceaseless company, and genuine help of my close friends and IMICS friends Eva, Yuanling, Wen-wei, Des, and Wayne that gave me the strength to complete my study.. i. DOI:10.6814/NCCU202100303.

(5) Abstract The prevalence of social media has transformed the way people gather information and hold interactive debates on public affairs. Since public opinion on several social issues is quite divided in Taiwan in recent years, social networking sites (SNS) might form an echo chamber for users, resulting in the extremity of attitude. Recent studies have different discussions on how heterogeneous information exposure plays a role in individuals’ attitude certainty, which is crucial to make SNS a more tolerant field of deliberation. Thus, this study aims at investigating the association between. 政治大. heterogeneous information exposure and attitude certainty. Besides, scholars. 立. highlighted the indirect effect of heterogeneous discussion on affecting attitude. ‧ 國. 學. certainty. By adopting the cognitive dissonance theory, previous research pointed out the potential effects of selective avoidance. Employing the dataset of the Taiwan. ‧. Institute for Governance and Communication Research (TIGCR) in 2018, the result. y. Nat. sit. demonstrates a positive relationship between heterogeneous information exposure and. n. al. er. io. attitude certainty on controversial social issues. Apart from the main effect, the. i Un. v. interaction effect of selective avoidance is also identified. When individuals’ selective. Ch. engchi. avoidance on SNS is low, the effect of heterogeneous information exposure on attitude certainty is stronger when it is of high-level selective avoidance. The findings contribute to the understanding of cognitive dissonance theory in the context of SNS and provide future practical suggestions for designing persuasion messages in communicating controversial social issues.. Key words: Social networking sites, social media, heterogeneous discussion, information exposure, controversial social issues, attitude certainty, selective avoidance. ii. DOI:10.6814/NCCU202100303.

(6) 摘要. 社群媒體的普及改變了人們汲取資訊、溝通公共事務或討論社會議題的方式。近年，民眾對於爭議性社會議題各持己見，在社群網站上意見分歧導致「回音室」（echo chamber）與網路同溫層的形成，從而導向極端態度。針對社群媒體上的異質性資訊曝露之於態度明確性扮演的角色，過去研究亦有不同論點。有研究認為異質性資訊能夠提高民眾對於不同態度的理解，進而調節民眾對於議題的態度明確性。然而，有研究則指出，根據認知失諧理論（Cognitive dissonance theory），. 政治大而會使人們更加確信自身對於議題的態度。本研究旨在探討異質性資訊曝露與態立人們傾向接納態度一致的資訊已取得認知平衡。因此，暴露在異質性資訊中，反. ‧ 國. 學. 度明確性的關聯性。異質性討論的中介效果及選擇性迴避的調節作用也是本研究探究之重點。本研究採用台灣政經傳播研究中心 2018 年調查資料，結果顯示社. ‧. 群網路異質性資訊曝露和爭議社會議題態度明確性為正相關，而選擇性迴避在此. sit. y. Nat. 過程中具有調解作用。然而，異質性討論對於異質資訊暴露與態度明確性的中介. io. al. er. 效果不顯著，選擇性迴避對於中介效果也無調節作用。本研究啟發了在社群媒體. v. n. 環境下如何理解認知失諧理論，針對未來溝通爭議性社會議題的說服訊息設計，提供一些實際建議。. Ch. engchi. i Un. 關鍵字：社群網絡、社群媒體、異質性資訊曝露、異質性討論、態度明確性、爭議性社會議題、選擇性迴避. iii. DOI:10.6814/NCCU202100303.

(7) Table of Contents Chapter 1. Introduction ··········································································1 Chapter 2. Literature Review ···································································6 2.1 Controversial Social Issues in The Taiwanese Context ....................................... 6 2.2 Social Media as A Communicative Role of Controversial Social Issues .......... 10 2.3 Attitude Certainty and Opinion Polarization on Social Media .......................... 13 2.4 Heterogeneous Information Exposure and Attitude Certainty ........................... 16 2.5 Mediating Effect of Heterogeneous Discussion on Social Media...................... 21. 治政 Chapter 3. Methodology ······································································· 28 大立 3.1 Hypothesis and Research Question .................................................................... 28 2.6 Moderating Role of Selective Avoidance on SNS ............................................. 23. ‧ 國. 學. 3.2 Survey Design and Sampling ............................................................................. 29 3.3 Measurement ...................................................................................................... 31. ‧. 3.3 Analysis Strategy................................................................................................ 35. y. Nat. sit. Chapter 4. Results ··············································································36. n. al. er. io. 4.1 Descriptive Analysis .......................................................................................... 36. i Un. v. 4.2 Hypotheses and Research Question Testing ...................................................... 42. Ch. engchi. Chapter 5. Discussion ··········································································47 5.1 Discussion .......................................................................................................... 47 5.2 Limitations and Future Suggestions ................................................................... 52 Chapter 6. Conclusion and Contribution ····················································56 Reference ······················································································62 Appendix. ······················································································72. iv. DOI:10.6814/NCCU202100303.

(8) Lists of Tables and Figures Tables. Table 1. Sample demographics .................................................................................... 37 Table 2. Variable characteristics................................................................................... 39 Table 3. Sample characteristics of outcome variable ................................................... 40 Table 4. Correlation matrix of the variables involved ................................................. 41 Table 5. Estimates of variables predicting attitude certainty and association with heterogeneous discussion ............................................................................ 43. 政治大. Table 6. Mediation effects of heterogeneous information exposure on attitude. 立. certainty through heterogeneous discussion. ............................................... 44. ‧ 國. 學. Table 7. Moderated mediation effect of selective avoidance ....................................... 44. ‧. Figures. sit. y. Nat. n. al. er. io. Figure 1. Research model ............................................................................................ 28. i Un. v. Figure 2. The conceptual diagram of Model 8 from Andrew Hayes’ process macro. Ch. engchi. (Hayes, 2018, p.13) .................................................................................. 35 Figure 3. The moderating effects of selective avoidance on the association between heterogeneous information exposure and attitude certainty .................... 45 Figure 4. The path of the moderated mediation model predicting attitude certainty ... 46. v. DOI:10.6814/NCCU202100303.

(9) Chapter 1. Introduction In the past decades, social media and social networking sites (SNS) have emerged as essential public discourse channels. Even though there are some arguments about SNSs algorithms filtering out dissimilar information (Brannon, Tagler & Eagly, 2007; Wojcieszak & Price, 2010), it is evident that social media provide a convenient forum for informational exposure and seeking. SNSs are especially valuable for its ability to provide attitudinally challenging information, which is referred to as heterogeneous information. Scholars believed that heterogeneous information can broaden people’s. 政治大. understanding of different perspectives, moderate individuals’ pre-existing attitude, and,. 立. therefore, enhance the political tolerance with dissimilar others (Papacharissi, 2002;. ‧ 國. 學. Lee, Kim & Kim, 2014; Chen, Ping & Chen, 2015; Liu & Su, 2017; Lee & Choi, 2019).. ‧. In addition to information exchange, discussions of political and public affairs. y. Nat. sit. nowadays also shifted to SNSs such as Facebook and Twitter. Public opinion on several. n. al. er. io. social issues has been quite divided in Taiwan in recent years. In 2018, Taiwan held a. i Un. v. referendum regarding controversial social issues that have been debating over decades,. Ch. engchi. including pension reform, the implementation of the nuclear power plant, same-sex marriage legalization, making them essential topics of interpersonal discussions on SNSs. During the days leading up to the referendum, people went on social media to express and exchange various opinions. Both the supporters and opponents of these issues have thrived to defend their perspectives and dispute “the other side.” Speaking to interpersonal communication, discussion with people who have different opinions is likely to indirectly affect the relationship between heterogeneous information exposure and individuals’ attitude formation (Lee & Choi, 2019). Past studies also demonstrated that various information is available in a heterogeneous network composed of users of 1. DOI:10.6814/NCCU202100303.

(10) different races, political identification, occupation, and educational level. Being exposed to diverse information and viewpoints is more likely to facilitate cross-cutting discussion among users, resulting in possible attitude change (Lee et al., 2012). Therefore, it is noteworthy to examine the relationship between heterogeneous information exposure and attitude formation and the indirect effect of heterogeneous discussion.. To date, social media has been highly penetrated in Taiwan and served as the primary pipeline for public discourse. Individuals nowadays rely more on shared. 政治大. information from SNSs than traditional media outlets (Turcotte et al., 2015). According. 立. to the “Global Digital Report” issued jointly by We Are Social and Hootsuite in 2019,. ‧ 國. 學. 20.82 million out of a total population of 23.73 million people in Taiwan are online. ‧. users. Above all, it is indicated that up to 95% of the population surf the internet every day. The average time spent on consuming and searching through digital devices is 7. y. Nat. io. sit. hours, 39 minutes. Besides, approximately 89% of the online citizens are active social. n. al. er. media users, mostly aged between 16 to 64 (We Are Social, 2019). The percentage. Ch. i Un. v. remains at a similar level in 2020 (We Are Social, 2020). Based on the statistic, each. engchi. internet user has 8.4 accounts on social media, and the average time spent is 1 hour 52 minutes per day. SNSs such as Facebook and Twitter have increased their impacts on public and social changes by enabling online users to disseminate firsthand news, engage in various discussions, and alter opinion climate (Chen, Ping & Chen, 2015; Matakos, Terzi, & Tsaparas, 2017). Especially, information or debates about controversial social issues on SNSs often encompass diverse, or even polarized perspectives, which differ from those of the mainstream media (Liu & Su, 2017). In other words, online audiences have more opportunities to be exposed to diverse information and incongruent opinions (Benkler, 2006; Kim, 2014; Wang, 2018). Hence, 2. DOI:10.6814/NCCU202100303.

(11) it seems promising to make online opinion more diverse, making society overall more democratic and open-minded.. Nevertheless, SNSs use might lead to selective avoidance and turn out to be a negative effect. Based on the cognitive dissonance theory (Festinger, 1957), individuals are inclined to seek out information consistent with their pre-existing perspective (Knobloch-Westerwick & Meng, 2011; Stroud, 2008, 2010). With the algorithm on SNSs, it accelerates the formation of echo chambers and filter bubbles. Instead of accepting the inconsistent perspectives and mitigating the differences, individuals. 政治大. usually reinforce their existing opinions, leading to increased attitude extremity. 立. (Bakshy et al. 2015; Bessi et al. 2016; Matakos, Terzi, & Tsaparas, 2017). Empirically,. ‧ 國. 學. several studies have been done regarding how social media intervenes in public opinion. ‧. on controversial social issues (Chung, 2018). However, whether SNSs use influence individuals’ attitude certainty thus mitigates the polarized opinion has been under. y. Nat. io. sit. controversy. Some argued that SNSs facilitate polarization by filtering out disagreement. n. al. er. and make people “wall themselves” (Sunstein, 2001; Stroud, 2010). While other. Ch. i Un. v. findings have revealed that social media expose individuals to divergent views,. engchi. increases the frequency of heterogeneous discussion, thereby making public attitudes toward conflict issues more tolerant (Bimber, 2008; Hampton et al., 2011). Past studies have examined how heterogeneous discussion plays a role in users’ opinions towards social issues (Shih, Scheufele & Brossard, 2012; Shi, 2016). In addition, research regarding online friendship formation and dissolution behavior has been conducted (Boyd & Ellison, 2007; Sibona & Walczak, 2011; Sibona, 2014). Surveys pointed out that 14% of Facebook users have ever unfriended others or hidden posts because they disagreed. Selective avoidance is considered to influence the heterogeneous information exposure and discussion on SNSs, further affect people’s attitude certainty 3. DOI:10.6814/NCCU202100303.

(12) on public issues. Such mixed findings raise the need to expand our comprehension of the influencers to attitude certainty, especially in the social media environment.. So far, most relevant literature has been conducted either under the Western context or the field of political communication (Mutz, 2002, 2011; Mutz & Mondak, 2006; Lee et al., 2014). Less is known about whether heterogeneous discussion mediates the relationship between information exposure on SNS and attitude certainty in the Taiwanese context. Neither does much research have been investigated about the impact of selective avoidance on SNSs. Therefore, this study aimed to bridge the gap.. 政治大. Realizing the underlying position and arguments can potentially help users formulate. 立. more robust perspectives on controversial issues, thus fostering their attitude formation.. ‧ 國. 學. Understanding the relationship between heterogeneous discussion and attitude certainty. ‧. would make SNS a more democratic field of deliberation.. y. Nat. io. sit. As a result, the purpose of this study is to explore the association between. n. al. er. heterogeneous information exposure and attitude certainty towards controversial social. Ch. i Un. v. issues. Furthermore, the indirect effect of heterogeneous discussion and the moderating. engchi. role of selective avoidance were investigated. Four controversial social issues of the 2018 Taiwanese referendum, including the pension reform, legalization of same-sex marriage, implementation of the nuclear power plant, and air pollution led by the increasing coal-fired power generation, were selected in the current research. This study conducted a secondary data analysis collected by the Taiwan Institute for Governance and Communication Research (TIGCR). The fieldwork of a national survey on lay citizens was implemented from July 9, 2018, to November 23, 2018, regarding their perceptions towards controversial social issues, political polarization, political communication, and voting behavior. The findings demonstrated the positive 4. DOI:10.6814/NCCU202100303.

(13) association between heterogeneous information exposure and attitude certainty in the Taiwanese context and expanded the understanding of selective avoidance on influencing attitude certainty.. 立. 政治大. ‧. ‧ 國. 學. n. er. io. sit. y. Nat. al. Ch. engchi. i Un. v. 5. DOI:10.6814/NCCU202100303.

(14) Chapter 2. Literature Review 2.1 Controversial Social Issues in The Taiwanese Context In 2018, Taiwan launched a national election and referendum on ten cases regarding several disputable topics. Among all the referendum subjects, issues regarding pension reform policy, same-sex marriage, go green with nuclear, coal-fueled and thermal power generation has drawn much attention online. Those issues have been under frequent controversial disputation for many years. Before the referendum took place,. 政治大 information, launch discussions, 立 and hold online and offline activities. With the social networking sites provided individuals and organizations places to forward. ‧ 國. 學. ubiquity of SNS, people could confront diverse points of view, find consensus with the attitudinally consistent others, and have several relevant discussions. Social media not. ‧. only served as an information transmitter, but a communicative role between two sides. Nat. er. io. sit. y. of different attitudes.. The pension reform policy has been a complicated and disputed issue since it was. n. al. Ch. i Un. v. first launched in 2012 by the former president Ma Ying-jeou ( 馬英九 ). As the. engchi. Democratic Progressive Party (DPP) has held the rein of government in Taiwan in 2016, they took over the pending pension reform policy. Subsequently, the government passed the draft, ensuring a sustainable pension system. During President Tsai Ingwen’s (蔡英文) term in office, the Cabinet and legislator continuously negotiated and held conferences for the pension bill amendment. The two main bills passed in 2017 were known as “eliminating 18 percent annual interest on savings in 2.5 years” and “decreasing the replacement rate” (Hsieh, Yu & Lin, 2018/06/30). Nevertheless, this action provoked a series of protests among veterans, civil servants, and teachers who oppose pension reform. In order to safeguard their rights and interests, those antagonists 6. DOI:10.6814/NCCU202100303.

(15) created public groups on Facebook, which were “The Pension Reform Oversight Coalition (監督年金改革行動聯盟)” and “800 Warriors (八百壯士捍衛權益)” to recruit alliance. Several protests have taken place with an estimated 100,000 participants against the policy. By the year 2018, the veterans advanced to social media platforms, recruited virtual communities, and continuously called on rallies against the pension reform bill from enacting into law. Attributed to SNSs, members could share relevant information simultaneously, live-stream the all-night sit-ins and hunger strikes surrounding the Legislative Yuan, raise online opinion polls, comment on the posts, and facilitate the discussion with different perspectives (CNA English News, 2018).. 政治大 The other controversial social issue that went viral on SNS is the legalization of 立. ‧ 國. 學. same-sex marriage. On May 24, 2017, Taiwan’s high court ruled that excluding samesex couples from marrying violated the Taiwanese constitution and gave the legislature. ‧. two years to amend the Civil Code to legalize marriage equality (The Guardian,. sit. y. Nat. 2018/11/22). To disseminate and advocate their claims, both sides massively spread. n. al. er. io. information, launched online and offline discussions to emphasize their attitudes, and. v. organized a series of campaigns on social media. The “Happiness Alliance” (下一代幸. Ch. engchi. i Un. 福聯盟) and “Family Guardian Coalition (護家盟),” which in favor of drawing up a separate law to extend equal marriage rights to same-sex couples, have created online communities on Facebook. Those groups inculcated the family value of heterosexuals by making videos and slogans, and reinforcing their aims by posting articles and bringing up discussions (Isabella Steger, 2018; Chen, 2019). On the other hand, the advocates who support the amendment of current marriage law and stress the importance of gender equality education also formed Facebook fan pages, as “The Taiwan Alliance to Promote Civil Partnership Rights (TAPCPR, 台灣伴侶權益推動聯盟)” and “Equal Love Taiwan ( 婚姻平權大平台 ).” They expressed opinions. 7. DOI:10.6814/NCCU202100303.

(16) supporting their attitudes, shared relevant information, and attempted to negotiate with the opposite. When the voting for the referendum finally took place on November 24, 2018, the result indicated a denial of ensuring same-sex couples’ marriage rights within the Civil Code attained 765 million votes (Central Election Commission, 2018). Ultimately, by May 17, 2019, lawmakers passed a specialized bill that guarantees samesex marriage rights and allows marriage registration with government agencies. Taiwan thus became the first country in Asia to legalize same-sex marriage; however, obstacles and controversies to the issue have still existed.. 政治大 on the topic of nuclear power. People who support the referendum of “go green with 立. Besides, the proliferation of social media enabled multidimensional conversation. ‧ 國. 學. nuclear” utilize the Facebook group as a means of persuasion. They convinced others that nuclear power is irreplaceable, and it is essential to call for the resumption of. ‧. Gongliao Nuclear Reactor No.4. They also assembled people via Facebook group such. sit. y. Nat. as “Nuclear Mythbusters (核能終結留言)” and “Green Nuclear Vote community (以. n. al. er. io. 核養綠－缺電公民自救會)” to sign a current petition and to march on the street.. i Un. v. According to the Bureau of Energy, Ministry of Economic Affairs (2019), industries. Ch. engchi. consumed 54% of overall national electricity (264.3 billion kWh) in 2018. The domestic power generation was 273.6 billion kWh in total, with 46% of coal-fueled power (126.6 billion kWh) and 35% of gas-fueled power generation (94.8 billion kWh). Only 5% of the total power generation was contributed by renewable energy (12.7 billion kWh). Taiwan lacks of natural resources due to its geographical environment, and most of the energy resources come from imports. With the increasing electricity demand and the difficulty of finding a primary alternative for base-load power generation, Taiwan is at high risk of suffering a power shortage from 2021 (World Nuclear News, 2019/02/01).. 8. DOI:10.6814/NCCU202100303.

(17) Taiwan’s anti-nuclear movement has progressed for forty years since Taiwan Power Company (Taipower) proposed the Lungmen nuclear power plant’s construction project in 1980. The enormous expense of the Lungmen project and the accidental conflagration in Maanshan nuclear power plant in 1985 has triggered a public discussion on the prerequisite of nuclear power (Nownews, 2019/02/01; Wang, 2012/11/28). Due to the two major nuclear disasters in Chernobyl, 1986, and in Fukushima, 2011, an increasing number of people turned to oppose nuclear power. Since thermal power generation has been considered as alternation, people were aware of the air pollution triggered along. In 2016, the DPP won the presidential election and. 政治大. announced a policy of building a “nuclear-free homeland” by 2025. Under this prospect,. 立. the DPP government passed an amendment to the first paragraph of Article 95 of the. ‧ 國. 學. Electricity Act, which stated, “the nuclear-energy-based power-generating facilities. ‧. should phase out by 2025”. Moreover, the DPP proclaimed that the energy policy by 2025 would be a mix with 50% of natural gas, 30% of coal, 20% of renewables, and. y. Nat. er. io. sit. 0% of nuclear energy (Liu, 2019; World Nuclear News, 2019/02/01; Chen & Chung, 2019/05/07). However, the continual shortage of electricity and air pollution came to. n. al. Ch. i Un. v. light have made people suspicious about the current energy policy. The adoption of the. engchi. “Go Green with Nuclear” referendum on November 24, 2018, showed that people accepting nuclear energy in Taiwan had increased significantly, placing energy deficiency on nuclear energy. Moreover, the budget and techniques for developing renewable energy and nuclear food import make nuclear energy a contested and sophisticated issue in Taiwan.. 9. DOI:10.6814/NCCU202100303.

(18) 2.2 Social Media as A Communicative Role of Controversial Social Issues Mass media, such as television, newspapers, and radio broadcasting, have been deemed a dominant channel for citizens to gain information about public affairs or social issues. While some scholars have discovered the significance of interpersonal communication in mass media as a factor influencing people’s attitudes. Lazarsfeld and others (1944) first proposed that interpersonal communication impacted individuals’ political attitudes and behaviors, thus developing the two-step flow of communication model. This paradigm supposed that instead of receiving the information directly from the mass. 政治大. media, most individuals gain the perspectives from interpersonal conversations with. 立. well-informed, politically interested opinion leaders within their social networks. It is. ‧ 國. 學. convinced that individuals at least periodically look for information and discuss political issues with friends, family members, acquaintances, and co-members of social. ‧. groups. Hence, individuals can elaborate and reconsider their understanding of political. y. Nat. io. sit. issues through interpersonal communication (McLeod et al., 1999; Shah et al., 2005).. n. al. er. Social media use not only directly affect political expression (Gil de Zúñiga et al., 2014). Ch. i Un. v. but play an influential role in the association between exposure to diverse viewpoints. engchi. and discussion network heterogeneity (Kim, 2011; Kim, Hsu, & Gil de Zúñiga, 2013; Lee et al., 2014). The prevalence of Web 2.0 technologies has revitalized public spheres, forming virtual communities, and expanding participation among individuals (Sánchez-Villar, J. M., 2019). New ways of collective actions have arisen out of social media platforms, such as protest websites, crowdfunding campaigns, online petitions, or coordinated boycotts (Van Earl & Kimport, 2011). By definition, social media are web-based platforms, including social networking sites (SNSs) (e.g., Facebook), content-sharing. 10. DOI:10.6814/NCCU202100303.

(19) channels (e.g., YouTube), microblogging services (e.g., Twitter), and blogs. It enables information exchange, social interaction, content generating, personal elaborations, and public discussions (Boyd & Ellison, 2007; Addawood & Bashir, 2016; Skoric et al., 2016; Wen, N., & Wei, R., 2018). Apart from the political issues, researchers started to examine whether media consumption are associated with people’s attitudes toward controversial social issues such as same-sex marriage (Lee & Hicks, 2011; Zhang & Counts, 2015; Oeldorf-Hirsch & McGloin, 2017; Gibson, 2018). For instance, Zhang and Counts (2015) analyzed. 政治大 characterize citizens’ ideology, and finding that “level of engagement, emotional 立. relevant posts on Twitter between 2011 to 2014, using social media expressions to. ‧ 國. 學. expression, and moral framing is the most predictive of attitude change on marriage equality.” Besides, People’s attitudes toward controversial topics such as same-sex. ‧. marriage will be moderated through informational access or mediated communication. sit. y. Nat. (Gibson, 2018). SNS enables users to share personal attitudes, connect to like-minded. n. al. er. io. others without geography boundary, ultimately leverage the bandwagon effect of viral. i Un. v. content. Chung (2017) also utilized secondary data analysis to explore factors. Ch. engchi. associated with supporting same-sex marriage. The result pointed out that active SNSs users online over 4 hours every day are more likely to support marriage equality. As a result, it has been demonstrated an association between SNSs use and individuals’ attitude transformation. As a democratic country with freedom of speech, Taiwan has fostered online and offline spaces for people to express their opinions and discuss controversial social issues unrestrictedly. Most theorists and scholars have acknowledged that citizens can access diverse arguments through interactive discussion, and subsequently, facilitate their understanding of counter-attitudinal arguments (Dewey, 1954; Gutmann & 11. DOI:10.6814/NCCU202100303.

(20) Thompson, 1996; Katz, 1997; Ho & Mcleod, 2008). Previous studies also indicated that the emergence of new media technology enables individuals to come across various views and have more moderate attitudes on political and public affairs (Bimber, 2004; Papacharissi, 2002; Lee, Kim & Kim, 2014). In sum, social media gradually replaces traditional media as a communicative role that might affect individuals’ attitudes on political and social issues.. 立. 政治大. ‧. ‧ 國. 學. n. er. io. sit. y. Nat. al. Ch. engchi. i Un. v. 12. DOI:10.6814/NCCU202100303.

(21) 2.3 Attitude Certainty and Opinion Polarization on Social Media Sorting out from the social psychological literature over years, an attitude is defined as one’s evaluation of an object, while attitude certainty refers to a sense of conviction that people hold their attitudes (Festinger, 1954; Gross, Holtz, & Miller, 1995; Rucker & Petty, 2004; Petrocelli, Tormala & Rucker, 2007). Attitude certainty is a secondary cognition attached to a pre-established cognition (Petty, Briñol, Tormala, & Wegener, 2007). It has drawn considerable attention in the recent attitudes and persuasion literature. The researcher found out that as certainty increases, attitudes become more resistant to counter-attitudinal information. Over the years, many studies have been. 政治大. accumulated regarding the antecedents and consequences of attitude certainty.. 立. According to the previous research, attitude certainty is influenced by individuals’. ‧ 國. 學. direct and indirect experiences with an attitude object (Fazio & Zanna, 1978; Wu &. ‧. Shaffer, 1987). It is also associated with the ability to resist persuasion and predict. sit. y. Nat. behavior. In terms of consequences, high attitude certainty is evidenced to yield greater. io. er. correspondent behavior (Rucker & Petty, 2004; Tormala & Petty, 2004a, 2004b; Bizer. al. et al., 2006), increased attitudinal persistence, and increased resistance to persuasion. n. iv n C (Tormala & Petty, 2002). That is tohsay, as people’s attitude e n g c h i U certainty towards issues. increased, it is more likely to form an attitude extremity or opinion polarization (Petty & Krosnick, 2014; Matakos, Terzi, & Tsaparas, 2017). With social media’s structural characteristics, it can also gradually form a homophily online sphere where users are predominantly exposed to information by like-minded others and align with the similarity (Garrett, 2009; Bakshy et al., 2015; Bessi et al., 2016). For instance, algorithms embedded in social media can help users decide which information to include in the database and how personalized content will be transmitted to users (Lee & Choi, 2019). Filter bubbles that exclude dissenting. 13. DOI:10.6814/NCCU202100303.

(22) opinions were thus created (Sunstein, 2001). Through the features of “unfriending” or “unfollowing” on social media, it accelerates the development of a huge “echo chamber” (Neo, 2015). On the other hand, attributing to the cognitive dissonance theory and confirmation bias, when individuals confront different attitudes, they are drawn to information that supports their attitude to balance psychological conflicts (Festinger, 1957). In this sense, individuals shift towards more extreme positions (Myers, 1982; Isenberg, 1986; Nie et al., 2010; Stroud, 2010). Once it happens at both sides of the attitude spectrum, a self-reinforcing dynamic may amplify initial attitude differences, further intensifying the possibility of attitude extremity and opinion polarization. 政治大. (Dandekar, Goel, & Lee, 2013; Mäs & Flache, 2013; Maes & Bischofberger, 2015).. 立. ‧ 國. 學. Opinion polarization on social media has been observed over various issues and topics, ranging from common ongoing discussion to the decisive debate politics and. ‧. controversial social issues. Nevertheless, there is still a wide spectrum of opinions on. sit. y. Nat. whether social media platforms serve as a breeding ground for more polarized opinions.. n. al. er. io. Social media and social networking sites create an open sphere for information. i Un. v. exchange, public affairs discussion, and opinion collision; it could be viewed as a two-. Ch. engchi. sided sword. On the one hand, social media exposes individuals to heterogeneous information, which can moderate their pre-established attitude and make the online sphere more tolerant. On the other hand, social media can also limit exposure to attitude-challenging information, further leading to the rise of attitude certainty. Based on the previous studies, attitude certainty will lead to individuals’ selective exposure, eventually causing the extremity of attitudes (Knobloch-Westerwick &Meng, 2009), which is likely to shape the polarization of opinion. Theoretical modeling work shows that opinion distributions polarize when individuals are exposed selectively by media sources with partial ideologies. That is, 14. DOI:10.6814/NCCU202100303.

(23) the media might influence individuals’ attitude certainty, and foster observed opinion polarization tendencies (Mckeown & Sheehy, 2006). However, some scholars hold different perspectives. They consider social media a heterogeneous network that allows users to diverse information than face-to-face situations (Barnidge, 2015; Barnidge, 2017). Therefore, it effectively modifies people’s attitude on inconsistent viewpoints, thus attenuating opinion polarization. According to Bimber (2004), the internet provides “a flexible and adaptable set of opportunities for communication.”, which the blurred and porous boundaries in online space facilitated exposure to uncongenial viewpoints (Brundidge, 2010). It can be seen that social media play a positive role in. 政治大. counter-attitudinal information exposure and the level of network heterogeneity (Kim,. 立. 2011). Social media use enables individuals to encounter more diverse perspectives and. ‧ 國. 學. to have moderate attitudes towards conflict issues (Papacharissi, 2002; Bimber, 2008;. ‧. Hampton et al., 2011). It benefits communication and understanding of the nonlikeminded others and creates a more tolerant online forum. In line with this argument,. y. Nat. er. io. sit. research has raised evidence stating that social media has the potential to advance exposure to cross-cutting views, thus forming a heterogeneous network (Kim et al.,. n. al. Ch. i Un. v. 2013; Lee et al., 2014). The arguments described above conflict with the discourse. engchi. addressing that social media help forward selective exposure and foster attitude extremity. Particularly in Taiwan, social media is highly penetrated in people’s daily life. Users not only search for information and news but express their perspectives and have discussions with others. Hence, the inconsistent findings raise the necessity to delve into how different information exposure on social media related to people’s attitude certainty.. 15. DOI:10.6814/NCCU202100303.

(24) 2.4 Heterogeneous Information Exposure and Attitude Certainty Opinion polarization is found associated with information seeking and attitude strength. It has been reported to have a detrimental effect on communities, societies, and democracies. It separates individuals into confronting sides with little or no communication, and even reinforce their pre-existing perspectives. Receiving only attitudinally consistent information and interacting exclusively with like-minded others could turn the online network into more fragmented. Compared to traditional media, social media has been discovered to mitigate opinion polarization on online networks (Gentzkow & Shapiro, 2010). Traditional media such as television, broadcasting, and. 政治大. newspapers provide individuals with particular content and viewpoint on different. 立. channels. However, social media are based on the internet, where users are more easily. ‧ 國. 學. exposed to all sorts of information on a single platform without boundaries and. ‧. limitations. Joint research launched in 2014 demonstrated that television and. sit. y. Nat. broadcasting are the media facilitating polarization among users, instead of social. io. al. er. media (Veenstra, Hossain, & Lyons, 2014). Even though algorithms on SNSs mostly. n. present the congenial information to users, it is not possible to completely filter out attitude-challenging information.C h. engchi. i Un. v. As a result, studies have focused on ways to decrease attitude extremity and opinion polarization. For instance, raising awareness of the opposite arguments and exposing individuals to different perspectives can moderate extreme attitudes and intensify network heterogeneity (Mutz & Mondak, 2006; Matakos, Terzi, & Tsaparas, 2017; Lee & Choi, 2019). Heterogeneity is defined as the extent of exposure to disagreements, including different races, gender, religions, political ideologies, and opposing arguments (Choi et al., 2013; Lee & Choi, 2019). It is convinced to effectively moderate the existing attitudes and reduce polarization by exposing users to dissimilar. 16. DOI:10.6814/NCCU202100303.

(25) information (Munson & Resnick, 2010; Munson et al., 2013; Vydiswaran et al., 2015; Matakos, Terzi, & Tsaparas, 2017). Some studies also concentrated on proposing mechanisms that will expose social media users to contents that are not primarily aligned with their prior beliefs. Informational use on SNSs referred to searching for information and news, posting messages, and exchanging information (Lee et al., 2014). The default settings of SNS allow users to share their opinions and to comment on others’ posts. Information can be handily forwarded and responded to by others. According to the findings of Lee. 政治大 involvement through information seeking, cross-cutting information exposure, and 立. and Myer (2016), SNSs use is significantly related to political view change and issue. ‧ 國. 學. discussion with disagreements. However, the positive role of heterogeneous information exposure in decreasing attitude certainty remains disputable. Some have. ‧. argued that individuals are motivated to seek out attitudinally consistent information. sit. y. Nat. due to the cognitive dissonance theory (Festinger, 1957). Some even end up in a. n. al. er. io. confirmation bias, and show more certain about their pre-existing attitudes after being. i Un. v. exposed to diverse information, thereby aggravating opinion polarization (Taber &. Ch. engchi. Lodge, 2006). In terms of controversial topics such as gun control (Taber & Lodge, 2006), sexual minority rights (Wojcieszak & Price, 2010), and stem cell research (Shih, Scheufele & Brossard, 2012), individuals are more aligned with their predilections after confronting disagreements. From a rational perspective, individuals are supposed to consume all available information relevant to the topic owing to the information utility (KnoblochWesterwick & Meng, 2009). However, based on the cognitive dissonance theory, people are inclined to seek consistent information while encountering attitudechallenging viewpoints. If individuals confront highly counter-attitudinal arguments, 17. DOI:10.6814/NCCU202100303.

(26) dissonance may thus increase, resulting in avoidance of the information. This phenomenon has been traditionally termed as selective exposure and subsequently labeled as confirmation bias (Jonas et al., 2001; Brannon, Tagler & Eagly, 2007). Selective exposure is found more pronounced for stronger attitudes than weaker attitudes. People with definite attitudes prefer to receive information that supports their viewpoints and are resistant to change their attitudes and behavior (Krosnick & Petty, 1995; Mutz & Martin, 2001; Brannon, Tagler & Eagly, 2007). Owing to the confirmation bias, attitude certainty is likely to lead to individuals’ selective exposure, thereby fostering extreme attitudes and opinion polarization in the long term.. 政治大 Otherwise, theorists proposed that exposure to disagreements can stimulate 立. ‧ 國. 學. individuals’ understanding of the opposite. “Ambivalence and confusion stem from the challenging viewpoints will lead people to access the conflicting arguments in a. ‧. balanced way” (Sniderman, 1981; Lee & Choi, 2019). It makes people realize that there. sit. y. Nat. is not only a dichotomous view of an issue (Mutz, 2002). Through a deliberative process,. n. al. er. io. individuals are expected to respect different perspectives and eventually reach a social. i Un. v. consensus (Lowell, 1913). Several empirical findings are in line with this argument. In. Ch. engchi. an investigation of political discussion by Mutz and Mondak (2006), they discovered that exposure to heterogeneous information could prompt an understanding of disagreements and raise political tolerance. According to surveys conducted in Germany, Spain, and the U.S. (Barbera, 2014), being exposed to various information will ease political polarization. Furthermore, researchers suggested that even though most individuals prefer viewpoint-reinforcing information, they do not demonstrate a systematic bias against the attitude-challenging information (Garrett, 2006). The findings infer that it is common for individuals to seek out information with a consistent attitude, whereas constant avoidance of dissimilar viewpoints is rare (Garrett, 2006).. 18. DOI:10.6814/NCCU202100303.

(27) Individuals may even seek out disagreements in a few circumstances. For instance, when individuals are not certain about their attitude, they are willing to look roughly at dissimilar information (Knobloch-Westerwick & Meng, 2009). In an online behavioral investigation, Jang (2014) also found out that apart from the existing religious belief and prerequisite information, people tend to check out the scientific information that challenges their pre-existing positions. Consequently, social media nowadays is highly penetrated in people’s daily life in Taiwan. Discussion and information exposure regarding controversial social issues. 政治大 to be pervasively exposed to heterogeneous information on SNSs. Drawing from the 立 have gradually transferred to social media platforms. Individuals have the opportunity. ‧ 國. 學. past studies, informational use and exposure have been proven to facilitate political attitude change. Therefore, it is predictable that heterogeneous information exposure on. ‧. SNSs will influence individuals’ attitude certainties on controversial social issues.. sit. y. Nat. People who insist on their pre-existing attitudes might be moderated after approaching. n. al. er. io. the dissimilar information, resulting in a decrease of attitude certainty. While people. i Un. v. with ambivalent attitudes could be more certain within the exposure of diverse. Ch. engchi. viewpoints. On the other hand, according to the cognitive dissonance theory, individuals prefer to be aligned with attitudinally consistent information. In this way, attitude certainty might increase because people are more clear with their attitudes on a specific issue. Since there were conflicting arguments in existence, it still lacks enough empirical evidence in the Taiwanese context. Hence, exploring the association between heterogeneous information exposure on SNS and individuals’ attitude certainty on controversial social issues becomes noteworthy. The author thus proposed the following research question.. 19. DOI:10.6814/NCCU202100303.

(28) RQ1. How is heterogeneous information exposure on SNSs associated with attitude certainty on controversial social issues in Taiwan?. 立. 政治大. ‧. ‧ 國. 學. n. er. io. sit. y. Nat. al. Ch. engchi. i Un. v. 20. DOI:10.6814/NCCU202100303.

(29) 2.5 Mediating Effect of Heterogeneous Discussion on Social Media During the last decade, mass media have been losing their credibility among the audience. Social media technologies feature the ubiquity of the interaction between users and sources (Quan-Haase & Young, 2010) have transferred users’ access to public affairs. The informational use and discussion of SNSs have become more common. In Lee and Myers’ study (2016), they indicated that social media users’ information seeking, heterogeneous exposure, and cross-cutting discussion are significantly related to political attitude change and issue involvement. Drawing from the previous research (Lee & Myers, 2016), a significant association between SNS use for information. 政治大. seeking and cross-cutting discussion and attitude change is discovered.. 立. ‧ 國. 學. Discussions enable citizens to voice their arguments, reconsider initial,. ‧. unreflective notions, and cultivate an understanding of alternative viewpoints. sit. y. Nat. (Gutmann & Thompson, 1996). Frequent SNS users are likely to have more. io. er. heterogeneous discussions because they are exposed to diverse information and. al. exchange ideas with distinct individuals. Exposure to conflicting viewpoints broadens. n. iv n C people’s understanding of a specific h incident and facilitates e n g c h i U discussion on public affairs (Eveland, 2001; Shah et al., 2005; Lee et al., 2014). Drawing from the data collected. by the Pew Research Center (2012), 25% of respondents have been more involved in a political issue after discussing it or reading different posts about it on SNS, and 16% of users have even changed their attitudes. The heterogeneous discussion is necessary to make attitude change and issue involvement possible, further raising understanding among dissimilar others. When engaging in heterogeneous discussions, individuals might be persuaded by uncongenial others and gradually modify their attitudes. Individuals’ pre-existing attitudes might also be influenced by others, further become. 21. DOI:10.6814/NCCU202100303.

(30) more certain, or they might remain ambivalent (Kim, 2011; Kim, Hsu, & Gil de Zúñiga, 2013; Lee et al., 2014). While attitude changing towards political or controversial social issues is not common, discussion with attitudinally inconsistent others does enable individuals to reconsider their perspectives (Knobloch-Westerwick & Meng, 2009).. In conclusion, several studies have focused on individuals’ political predisposition (Abramowitz & Saunders, 2008) and their media use frequency (Nie et al., 2010; Stroud, 2007, 2010) as predictors of attitude extremity and polarized opinion. A few research also emphasized on the relationship between the characteristics of SNSs and attitude. 政治大. certainty (Huckfeldt et al., 2004). In other words, discourse with different perspectives. 立. becomes an essential path in forming a pluralistic society (Mutz & Martin, 2001).. ‧ 國. 學. Hopefully, it will moderate people’s attitude certainty, thereby mitigate opinion. ‧. polarization on social media. Thus, the author assumed that heterogeneous discussion on SNSs would influence heterogeneous information exposure, which, in turn,. y. Nat. n. al. er. io. is written as the following:. sit. decreases individuals’ attitude certainty on controversial social issues. The hypothesis. Ch. engchi. i Un. v. H1. Heterogeneous discussions mediate the relationship between heterogeneous information exposure and individuals’ attitude certainty on controversial social issues.. 22. DOI:10.6814/NCCU202100303.

(31) 2.6 Moderating Role of Selective Avoidance on SNS Social media has transformed our online environment with an increased likelihood of heterogeneous information exposure and opinion expression with inconsistent attitudinal others (Kim, 2011; Messing & Westwood, 2014). Yet, people tend to prefer attitude-reinforcing information over attitude-challenging ones to eliminate cognitive dissonance (Garrett, 2009a; Knobloch-Westerwick & Meng, 2009, 2011). Social media characteristics strengthens homogeneous exposure by allowing users to hide posts and comment that they are not in favor of, unfollowing, unfriending, and blocking other accounts. Furthermore, such behaviors reference social media algorithms to filter. 政治大. content based on individuals’ preferences, resulting in an echo chamber (Bakshy,. 立. Messing, & Adamic, 2015; John & Dvir-Gvirsman, 2015).. ‧ 國. 學 ‧. According to the literature review, selective avoidance on social media is known. sit. y. Nat. as actively removing the present and future exposure to dissonance by cutting off. io. er. connections with dissenting information. Researchers have also conceptualized selective avoidance as “shielding oneself from dissonant views by sorting out unwanted. al. n. iv n C information and breaking social tieshthat transmit dissimilar e n g c h i U information (Zhu, Skoric, & Shen, 2016). Unfriending, blocking, or hiding posts are classified as selective avoidance behaviors. People who have extreme attitudes or political engagement are more likely to conduct selective avoidance (John & Dvir-Gvirsman, 2015; Skoric, 2015). Additionally, individuals that serve as weak social ties with heterogeneous information and perspective are the primary target of those disconnecting behaviors (Rainie & Smith, 2012; Sibona, 2014). Approximately 95 % of people’s Facebook connections are formed as casual friends, which could be considered weak ties (Stutzman & Kramer-Duffield, 2010).. 23. DOI:10.6814/NCCU202100303.

(32) In general, people are prone to build up friendships with those who share certain similarities, including age, ethnicity, religion, education, occupation, characteristics, and so forth (McPherson et al., 2001). Apart from family members, the largest portion of friendships is established through organizational structure. Schools, work, and geographic location have been proposed as the major elements to form a relationship. They may also be factors in how dyadic connections are created on SNSs (McPherson et al., 2001; Sibona, 2014). Boyd and Ellison (2007) categorized thirteen common reasons for sending and accepting friend requests on SNSs. For instance, when the dyad is actual friends, acquaintances, or have similar interests or values, it is easy to form the. 政治大. online connection. Nevertheless, online friendships established on SNSs can be fluid. 立. and delicate, where the connections can be dissolved with a click of a bottom (Sibona,. ‧. ‧ 國. 學. 2014).. The default function of “unfriend” on SNSs enables users to end a relationship. y. Nat. io. sit. with the other person without any permission. Moreover, the unfriended person will not. n. al. er. even receive notifications in most cases (Sibona, 2014). Hence, “unfriending” has. Ch. i Un. v. become a widely-used feature of SNSs. According to Pew Internet, it has revealed that. engchi. 63% of users unfriended at least one member of their online social networks in 2011 (Madden & Smith, 2012). Drawing from a survey in 2018, 44% of teenagers unfriend or unfollow people on social media sometimes, with 14% who do it quite often. Conversely, there are still young people who rarely (39%) or never (14%) unfriend or unfollow people on social media. The word “unfriend” was defined as “removing someone as a ‘friend’ on social networking sites” by the New Oxford American Dictionary for the word of 2009. Unfriending on social networking sites such as Facebook is different from ending a friendship in the real world. Sibona (2014) defined unfriending on Facebook as a conscious behavior by an individual to end the 24. DOI:10.6814/NCCU202100303.

(33) bidirectional relationship and explained it “manifests itself through the removal of a link between the dyad.” The other acts of selective avoidance on Facebook refer to “unfollowing” or “hiding,” by which one filters out another without terminating their Facebook friendship. Both practices remove certain content from a user’s News Feed (the flow of stories Facebook presents to its users). Once they have been unfriended or hidden, users will no longer receive any notifications from Facebook (John & Gvirsman, 2015).. Regarding the cause of unfriending on Facebook, Sibona and Walczak (2011) have. 政治大. discovered four common online reasons and two offline reasons. The formers were. 立. frequent or unimportant posts, different attitudes or opinions (politics and religion),. ‧ 國. 學. inappropriate posts, and daily life posts that urge people to unfriend others. In. ‧. comparison, the two offline causes included disliked behavior and changes in the relationship. The finding indicated that 55% of people had unfriended the others for. y. Nat. er. io. sit. their online posting behavior, 28% due to the offline factors, and 17% unsure. An empirical study by Rainie and Smith (2012) reported that when users encounter. n. al. Ch. i Un. v. disagreement on social media, between 18% and 26% of American SNS users have. engchi. blocked, unfriended, or hidden the contacts. Only 9% of American SNS users unfriended or hid someone’s post because of political disagreements in their entire usage histories. Whereas most of the others (66%) usually ignore the posts. Drawing from the findings in an Asian context, 15.6% of respondents hide others’ posts or unfriend people with political disagreements during the Umbrella Movement in Hong Kong (Zhu, Skoric, & Shen, 2016). In a similar vein, the Pew research center (2018) investigated the reasons that teens at least sometimes unfriend or unfollow people on social media. The survey showed that 54% of teens have disconnected from someone who posted too often or irrelevant. A smaller percentage of teens have unfollowed 25. DOI:10.6814/NCCU202100303.

(34) others because they act differently online than in person (43%) or posting political views they disagree with (22%). A few months before the 2018 referendum in Taiwan, individuals and groups expressed their incongruent attitudes on controversial social issues on SNSs, which led to confrontation and arguments. Some users debated with non-likeminded acquaintances on SNSs, while some even blocked contact with dissimilar others. It is in line with the previous research that people disconnected from those who hold dissimilar attitudes.. Disconnecting disagreements on Facebook filter out diverse opinions and. 政治大. gradually lead to a homogeneous network. If individuals have been in a politically. 立. homogeneous SNS, they were more likely to unfriend people with inconsistent attitudes. ‧ 國. 學. and fortify echo chambers (Jamieson & Cappella, 2008). It can also be explained by. ‧. Social Identity Theory (SIT), assuming that individuals maximize differences between the group they psychologically belong to and the opposition (Yoo, Ng & Johnson, 2018).. y. Nat. io. sit. Unfriending people holding different attitudes will reduce exposure to conflicting. n. al. er. information, definitizing people’s inherent attitudes on political issues (John &. Ch. i Un. v. Gvirsman, 2015; Yoo, Ng & Johnson, 2018). As individuals increasingly performed. engchi. selective avoidance, it is likely to facilitate their attitude certainty (Mutz, 2006; Sunstein, 2009; Stroud, 2010).. Prior studies regarding friendship dissolution were largely based on intimate relationships such as close friends or romantic partners (McPherson et al., 2001). Nonetheless, research that delves into digital selective avoidance remains scant so far (Light, 2014; Light & Cassidy, 2014). Unfriending on Facebook is due to different reasons mentioned above, while it may simply happen to avoid context collapse triggered by the ubiquity of information access on SNSs (Sibona, 2014). Unfriending 26. DOI:10.6814/NCCU202100303.

(35) or hiding posts from others who stress dissimilar attitudes are likely to filter out heterogeneous exposure and cross-cutting discussion on SNS. The reduction of heterogeneous information exposure will minimize the network heterogeneity and lessen the opportunity for cross-cutting discussion. Furthermore, the decrease of heterogeneous information exposure will lead to selective exposure for individuals, strengthening their attitude towards certain incidents. In sum, selective avoidance on SNSs, conceptualizing as unfriending and hiding posts or comment, is susceptible to moderate the effect of heterogeneous information exposure and cross-cutting discussion on attitude certainty. Hence, the hypotheses were raised as below:. 立. 政治大. H2. Selective avoidance on SNS moderates the mediating effects of heterogeneous. ‧ 國. 學. discussion in the relationship between heterogeneous information exposure and. ‧. attitude certainty.. y. Nat. er. io. sit. H3. Selective avoidance on SNS moderates the relationship between heterogeneous information exposure and individuals’ attitude certainty.. n. al. Ch. engchi. i Un. v. 27. DOI:10.6814/NCCU202100303.

(36) Chapter 3. Methodology 3.1 Hypothesis and Research Question The purpose of the study was to see how heterogeneous information exposure on SNSs related to individuals’ attitude certainty on controversial social issues. The role of heterogeneous discussion and selective avoidance in affecting the attitude certainty were also examined. Based on the literature reviews, this study aimed to answer the following hypotheses and research question. The research model was thus formulated. 政治大 RQ1. How is heterogeneous立 information exposure on SNSs associated with attitude below (see Figure 1).. ‧ 國. 學. certainty on controversial social issues in Taiwan?. H1. Heterogeneous discussions mediate the relationship between heterogeneous. ‧. information exposure and individuals’ attitude certainty on controversial social issues.. Nat. sit. y. H2. Selective avoidance on SNS moderates the mediating effects of heterogeneous. al. n. certainty.. er. io. discussion in the relationship between heterogeneous information exposure and attitude. Ch. engchi. i Un. v. H3. Selective avoidance on SNS moderates the relationship between heterogeneous information exposure and individuals’ attitude certainty.. Figure 1. Research model 28. DOI:10.6814/NCCU202100303.

(37) 3.2 Survey Design and Sampling. To test the proposed hypotheses and research question, this study employed a national survey conducted by the Taiwan Institute for Governance and Communication Research (TIGCR). TIGCR is funded by the “Higher Education Sprout Project’s Part Two: The Feature Areas Research Center Program” of the Ministry of Education. To delve into the perspective of democratic governance, the TIGCR concentrated on investigating the “Political attitude, Policy-making, and Governance Communication in Taiwan.” Examining aspects contain political attitude formation, the distribution of. 政治大. government bureaucrats and Taiwan citizens, the impact of digital technology on public. 立. opinion and behavior, and the interaction between the communication model of political. ‧ 國. 學. information and attitude. Data archive includes the citizen’s panel survey data through Face-to-Face interview and the government bureaucrats’ panel survey data through an. ‧. online survey.. sit. y. Nat. n. al. er. io. Given the goal of exploring the influences of SNSs information exposure and. i Un. v. heterogeneous discussion on individuals’ attitude certainty on social issues, the 2018. Ch. engchi. national survey of lay citizens, including questions for several controversial social issues, was suitable for this study. The fieldwork of a national survey on lay citizens was implemented from July 9, 2018, to November 23, 2018, at the National Political and Economic Research Center of the National Cheng-chi University. The questionnaire survey focused on Taiwanese’ viewpoints of critical issues, such as violation of human rights during the authoritarian regime, nuclear power generation, air pollution led by the increasing coal-fired power generation, same-sex marriage, pension reform, and so on. Political opinion polarization, governance satisfaction, political partisanship, network community size, social media usage, interactive 29. DOI:10.6814/NCCU202100303.

(38) behaviors on SNS, including posting, commenting, and debating with others, were examined in the survey with 2478 representative participants aged 18 and above (Male = 47.3%, Female = 52.7%). As for the data analysis of this study, 1330 respondents who used SNSs to receive information about political and public issues were taken as the sample. Sample respondents were filtered by the determined questions of “Do you use laptops, tablet computers or mobiles to surf the Internet?”, “On average, how many days in a week do you use social media?” and “How often do you receive information about political or public issues from social media (forwarded or shared by friends are also included)?”.. 立. 政治大. ‧. ‧ 國. 學. n. er. io. sit. y. Nat. al. Ch. engchi. i Un. v. 30. DOI:10.6814/NCCU202100303.

(39) 3.3 Measurement 3.3.1 Outcome Variable Adapted from the past research, attitude certainty was measured by asking the respondents how certain they were of their attitudes toward an issue or objects (Fazio & Zanna, 1978; Tormala & Petty, 2002). In Petrocelli, Tormala and Ruckers’ research (2007), they conducted an experiment by using a single item to measure attitude certainty on specific issues. Respondents were asked to rate the extent to which they are certain about a specific issue with a nine-point scale (1 = not certain at all; 9 = very certain). Previous research operationalized the degree of certainty to measure attitude. 政治大. certainty. However, the current research aimed to examine whether individuals can. 立. develop a certain attitude. Therefore, this study employed the conceptualization of. ‧ 國. 學. attitude certainty by asking respondents which statements do they agree with the. er. io. sit. y. Nat. Attitude Certainty. ‧. selected issues.. Attitude Certainty (Kr20 = .557, M = 3.55, SD = .82) can be evaluated by implicit. al. n. iv n C or explicit measures. An explicit method was implemented by asking respondents how hen gchi U certain they are with their attitudes (Clarkson et al., 2011). In comparison, an implicit method was measured with structural parameters (Ryffel et al., 2014). In the current study, attitude certainty was measured by to what extent the participants can develop a. certain attitude. Participants were asked which statement they consider to agree or disagree with four controversial social issues. Issues included nuclear power generation, air pollution led by the increasing coal-fueled power generation, marriage for homosexuality, pension reform for retired teachers and veterans. The response categories were: 1=agree with the issue, 2=disagree with the issue, 3= agree with both. 31. DOI:10.6814/NCCU202100303.

(40) sides, 4=disagree with both sides, 97=I don’t know. Responses for agreement and disagreement were coded as “1”, which referred to a certain attitude. On the contrary, ambivalent responses such as “both agree,” “both disagree,” “don’t know” were recoded as “0”, which represented those attitudes were uncertain. Attitude certainty was then measured by summing those four items.. 3.3.2 Independent variable Heterogeneous Information Exposure Heterogeneous information exposure on SNS (M = 2.82, SD = .88) served as the. 政治大. only independent variable in this study. Respondents were asked, “How often do you. 立. see the information of political or public issues on social media that differ from your. ‧ 國. 學. position.” A five-point scale (1=never, 2=seldom, 3=sometimes, 4=usually, 5=always,. ‧. 94=it depends, combined with 3) was used to indicate the level of heterogeneous information exposure on SNS.. n. al. er. io. sit. y. Nat 3.3.3 Mediator. Heterogeneous discussion. Ch. engchi. i Un. v. To measure heterogeneous discussion (M = 1.54, SD = .75), this study adopted the operational measurement of cross-cutting discussion based on the previous research (Shih, Scheufele & Brossard, 2013; Kim, Hsu & de Zuniga, 2013). One critical question was chosen by asking, “During the past six months, when you saw your friend’s Facebook post or comment have different positions on political or public issues, how often do you raise your own opinions?” Responses were on a five-point scale, with higher points indicating higher frequency (1=never, 2=seldom, 3=sometimes, 4=usually, 5=all the time, 94=it depends, combined with 3).. 32. DOI:10.6814/NCCU202100303.

(41) 3.3.4 Moderators Selective avoidance on SNS According to the previous research, selective avoidance (M = 1.3, SD = .57, r = .4) was assessed by unfriending and hiding others’ posts and comments (Zhu, Skoric, & Shen, 2016). The former referred to aggressively removing the online connection between the dyad without informing the other person. Simultaneously, the latter served as a passive act to hide content from Facebook News Feed without terminating the relationship. These two items were thus combined to a scale. To measure these concepts (John & Gvirsman, 2015; Yoo, Ng & Johnson, 2018), participants had to rate the. 政治大. following statements respectively on a five-point scale (1 = never to 5= all the time).. 立. Questions were listed below.. ‧ 國. 學. (1) “During the past six months, when you saw your friend’s Facebook post or. ‧. comment have different positions on political or public issues, how often do you hide the post or comment?”. y. Nat. io. sit. (2) “During the past six months, when you saw your friend’s Facebook post or. n. al. er. comment have different positions on political or public issues, do you unfriend or delete that friend?”.. Ch. engchi. i Un. v. 3.3.5 Control variables To avoid a confounding effect, demographic variables such as age, gender, and educational level were included as controls in the analysis. Firstly, to measure age, respondents were asked to indicate their birth year. In terms of gender, male was coded as 1, female as 0. The ratio of men to women was near 1:1, with slightly more women (52.7%). Educational level was calculated with a seven-point scale (1=none or selfstudy, 2=elementary school, 3=vocational/ junior high school, 4=senior high school and cadet school, 5=open university, junior college of military/police, and junior college, 33. DOI:10.6814/NCCU202100303.

(42) 6=university, college of military/police, 7=master’s degree and above) (M=5.2, SD=1.2).. Besides, drawing from the observation in Chen’s research (2019), it is evidenced that people who do not hold any religious beliefs and whose partisanship falls to the “pan-green” (DPP) revealed higher supportiveness to controversial social issues such as same-sex marriage. Therefore, party identification and religion were included for control purposes in this research. Party identification was distributed in three categories: “pan-green”, “independent”, and “pan-blue”. Since this variable only consisted of three. 政治大. categories, which were not suitable to view as a continuous variable. For the purpose. 立. of data analysis, party identification was created as two dummy variables, with “pan-. ‧ 國. 學. green” as the reference group. Religious belief was recoded as a nominal variable, with. ‧. 0 indicating no religion, 1 referring to any kind of religious beliefs.. n. er. io. sit. y. Nat. al. Ch. engchi. i Un. v. 34. DOI:10.6814/NCCU202100303.