臉書使用與科學公眾參與：一般大眾與科學臉書粉絲專頁使用者之比較 - 政大學術集成

全文

(1)National Chengchi University International Master’s Program in International Communication Studies 國際政治⼤學國際傳播英語碩⼠學位學程 Master’s Thesis 碩⼠論⽂. 學. 臉書使⽤與科學公眾參與：. ‧. ‧ 國. Facebook Use and Public in Science: A 治政 Participation 大 and Users of a Comparison Between the General Public 立 Science Facebook Fan Page. ⼀般⼤眾與科學臉書粉絲專⾴使⽤者之⽐較. n. er. io. sit. y. Nat. al. Ch. engchi. i n U. v. Student: Cheng-Ying Wu 吳承穎 Advisor: Dr. Tsung-Jen Shih 施琮仁教授. 中華民國⼀〇六年七⽉ July 2017.

(2) Acknowledgements First of all, I would like thank my advisor, Dr. Shih, Tsung-Jen. It would be a great difference if I had not met Dr. Shih in my graduate life. He patiently spent time instructing me and guiding me in a humorous way when I felt lost in the “academic sea”. I will always remember how enjoyable the meetings were and how supported I felt after discussion. I appreciate it a lot, thank you, professor TJ. Besides, I would also like to acknowledge my committee Dr. Hsu, Mei-Lin and Dr. Kuo, Win-pin for spending time reading this thesis and kindly giving me useful suggestions and. 政治大. valuable comments on this thesis. I also appreciate Dorothy for proofreading for me. 立. every time without hesitation.. ‧ 國. 學. Thank you, IMICS 2016! We took courses together, experienced every presentation and suffered from the deadlines of papers together. These three years. ‧. could not be more fun and crazy with you guys. I also appreciate Colorunners for. y. Nat. io. sit. making my graduated life more colorful. I liked every meeting and our laughter.. n. al. er. Thank everyone in study room 408 and girls I met from Art and Culture Center. I. Ch. i n U. appreciate everyone I met in National Chengchi University.. engchi. v. I am also grateful to TGIM for every awesome Monday night. The most relaxing moments were when I focused on chasing Frisbees without thinking the course or my thesis. We did have a lot of fun time! Special thanks to Mr. Chu for always comforting me when I felt down and trusting me I could eventually accomplish it. And last, but not least, I must express my very profound gratitude to my family. I appreciate mom and dad for backing me up the whole time and your love and care for me are the treasure in my life. In the blink of an eye, three years have gone by and I still cannot believe today has arrived. Writing thesis was a long and lonely process. What the process taught me i.

(3) is not limited to the academic realm; it is a life-long lesson which I will carry and use in my next journey.. Cheng-ying 2017/08/23 Taipei, Taiwan. 立. 政治大. ‧. ‧ 國. 學. n. er. io. sit. y. Nat. al. Ch. engchi. ii. i n U. v.

(4) Abstract Social networking sites (SNSs) nowadays serve as important platforms for users to engage in public affairs. This study applied the O-S-O-R model to examine the relationship between Facebook and public participation in science and the mediating role of users’ scientific knowledge and self-efficacy. Besides, this study examined the relationship mentioned above by utilizing data collected from two different populations - the general public and the users of the Facebook fan page “Pansci.com,” the biggest scientific online group in Taiwan.. 政治大. Based on a total sample size of 1,960, the result showed that scientific use of. 立. Facebook increased scientific knowledge only for “Pansci.com” users. Besides,. ‧ 國. 學. scientific use of Facebook is positively associated with self-efficacy and public participation in science in both samples. It is noteworthy that the effect of scientific. ‧. use of Facebook on self-efficacy is stronger for “Pansci.com” users but the effect of. Nat. sit. y. self-efficacy on public participation in science is stronger for general users. In. n. al. er. io. addition, self-efficacy serves as a significant mediator in the relationship between. i n U. v. scientific use of Facebook and public participation in science in both samples.. Ch. engchi. Keywords: Facebook, Public Participation, Science Communication, SelfEfficacy, Social Networking Sites. iii.

(5) Table of Contents Acknowledgements ....................................................................................................... i Abstract....................................................................................................................... iii Contents ....................................................................................................................... iv Lists of Tables............................................................................................................... v Lists of Figures ............................................................................................................. v Chapter 1. Introduction .............................................................................................. 1 Chapter 2. Literature Review ..................................................................................... 6. 政治大 Chapter 4. Result ....................................................................................................... 34 立 Chapter 3. Method ..................................................................................................... 24. Chapter 5. Discussion ................................................................................................ 48. ‧ 國. 學. Chapter 6. Conclusion ............................................................................................... 57. ‧. Reference .................................................................................................................... 60. n. er. io. sit. y. Nat. al. Ch. engchi. iv. i n U. v.

(6) Table 1. Descriptions and Correlations of Variables (The “Pansci.com” Sample) ...... 31. 2. Descriptions and Correlations of Variables (The General Sample)................ 32. 3. Descriptions statistics (Means and Standard Deviations) .............................. 34. 4. Age Distribution ................................................................................................... 35. 5. Education Distribution ........................................................................................ 36. 6. Ordinary Least Regression Model with Scientific Knowledge as Outcome Variable ............................................................................................................... 39. 7. Ordinary Least Regression Model with Self-efficacy as Outcome Variable .. 41. 8. Ordinary Least Regression Model with Public Participation in Science as. 9. Mediation Effect Result. Scientific knowledge and self-efficacy as mediator. 政治大 Outcome Variable ................................................................................................ 44 立. ‧ 國. 學. ................................................................................................................................ 47 10 Summary of Facebook motivation scales and individual items ...................... 49. ‧. n. al. y er. io. Figure. sit. Nat. List of Figures. i n U. v. 1. O-S-O-R model of public participation in science ............................................ 23. 2. Interaction Effects on Levels of Self-efficacy by Scientific Use of Facebook of. 3. Different Samples ............................................................................................... 42 Interaction Effects of Scientific Use of Facebook on Public Participation in. Ch. engchi. Science by Level of Self-efficacy of Different Samples ................................ 46. v.

(7) Chapter 1. Introduction As science becomes increasingly complex and intertwined with people’s daily lives, public participation in science turns out to be an important issue in modern societies (Powell & Colin, 2008). In contrast to the more traditional impact assessment model, in which the impact of science on society is evaluated after innovations are developed, scholars nowadays have emphasized the need to involve the public in earlier stages of scientific development (Nistbet & Scheufele, 2009). This idea is called “upstream engagement” (Wilsdon & Willis, 2004; Tait, 2009).. 政治大. Engaging the public is necessary and helpful in two aspects. First, public. 立. participation contributes to a more transparent and accountable process of public. ‧ 國. 學. policy-making. Research has indicated that when risk issues, for example global. ‧. warming, nanotechnology, are discussed only within expert circles (i.e., government officials or scientists), citizens are likely to become more distrustful and skeptical. y. Nat. io. sit. (Powell & Colin, 2008; Eden, 1996; Powell & Colin, 2009). Science is important for. n. al. er. influencing policy decisions on how to deal with those risks. In order to avoid the. Ch. i n U. v. ‘expertization’ of science, it is significant to raise its public profile and strengthen its ability to shape society (Eden, 1996).. engchi. Furthermore, in a democracy, citizens should have a say in decisions about scientific developments that will affect their lives and society in significant ways. Some researchers have pointed out that greater well-being can be obtained through higher levels of democratic engagement among individuals and societies (Hampton, 2011). Moreover, lay citizens may have valuable knowledge and points of view to bring to discussions, and thus their input can contribute to more comprehensive and stronger societal decisions (Powell & Colin, 2008).. 1.

(8) Social networking sites (SNSs) nowadays serve as important platforms for users to instantly interact with one another and engage in public participation (Valenzuela, Park & Kee, 2009; Whiting & Williams, 2013; Kwon, D’Angelo & McLeod, 2013; Raacke & Bonds-Raacke, 2008; Alhabash, Park, Kononova, Chiang & Wise, 2012). First, Nisbet and Scheufele (2009) indicated that the interactive and user-focused digital content of social media could work to fill the information gaps in the area of science. Even users non-interested in specific issues can engage with scientific information due to casual or incidental exposure via social media (Scherman, Arriagada, & Valenzuela 2014).. 治政 Second, social media users are able to distribute 大 a massive amount of 立. information to a large number of contacts on SNSs, which increases the probability of. ‧ 國. 學. reaching critical mass and facilitates mobilization (Scherman, Arriagada, &. ‧. Valenzuela 2014; Chou et al., 2009). Third, SNSs allow users to create and join. sit. y. Nat. groups with common interests and similar ideas, and to eventually organize activities. io. er. (Valenzuela, Park & Kee, 2009). Some features of SNS groups, such as calendar services, allow users to manage events or activity invitations, and thus can support. n. al. i n C organizing activities more concretely et al., 2011). h e(Wohn ngchi U. v. In this study, the O-S-O-R model (Orientations-Stimulus- Orientations-. Response model) is applied. The first O indicates the role of preexisting orientations, for example, socioeconomic status, age, education level. The second O refers to personal-psychological factors in Stimulus that conditions media use. In the study, Facebook is treated as stimulus. Finally, the ultimate effect of media use is Response which is people’s public behavior in science in this research (Jung, Kim, & de Zúniga, 2011; Scheufele, Nisbet, Brossard, & Nisbet, 2004).. 2.

(9) Few studies focused on what mediated communication and public participation and they indicated that future research should explore especially the second O (Jung, Kim, & de Zúniga, 2011). They further pointed out knowledge and efficacy may played a role as mediators in the process leading from communication to participation (Jung, Kim, & de Zúniga, 2011; McLeod, Scheufele, & Moy, 1999). The two concepts are included for further investigation in this paper. Knowledge is highly related to the public involvement in decision making (Wilsdon & Willis, 2004). SNSs can gain knowledge for users in the two ways. First, SNSs can allow users to search and collect information, and obtain useful professional. 治政大most users, SNSs provide expertise (Pi, Chou, & Liao, 2013). Second, although for 立. opportunities for users to gain more knowledge about specific issues which they may. ‧ 國. 學. be not have been interested in originally (Barker, Dozier, Weiss, & Borden, 2013). The. ‧. communication process among citizens also influences people’s civic attitudes and. sit. y. Nat. behaviors by allowing them to exchange information, elaborate on problems facing the. io. de Zúniga, 2011).. er. community and learn about opportunities to participate in civic activities (Jung, Kim &. al. n. v i n As far as the motivator forCparticipating h e n g cinhsuch i Uevents, self-efficacy is a. particularly important concept (Vraga et al., 2015). Self-efficacy is defined as. “people’s beliefs in their capabilities to produce given attainments” (Bandura, 1994, p.307). It determines how people feel, think, motivate themselves and behave. Those with higher levels of self-efficacy are more likely to act (Pinkleton, Austin & Fortman, 1998). There are three reasons why SNS-use can cultivate users’ self-efficacy. First, SNSs provide people with successful experiences of engaging with social issues and communicating information to others in their social network (Vraga et al., 2015; Velasquez & LaRosem, 2015). For example, users may publicly support social. 3.

(10) issues via SNSs and obtain governments’ response to such actions (Vraga et al., 2015). Experience is the most powerful source of efficacy beliefs since it allows individuals to learn firsthand the reach of their own capabilities. Therefore, if these experiences are perceived as successful, users can obtain their own perceived capabilities of what they can achieve (Velasquez & LaRosem, 2015). Second, information-rich SNSs satisfy citizens’ levels of information seeking. SNS users’ acquisition of greater knowledge is positively related to their development of higher perceptions of efficacy (Pinkleton, Austin & Fortman, 1998). Third, SNS users can comment on their friends’ posts and message their friends. Such instant. 治政 interactivity inherent in SNSs also increases the levels 大 of users’ self-efficacy (Vitak et 立 al., 2011; Tedesco, 2007).. ‧ 國. 學. Most of the studies about SNSs focus on one specific sample group (Scherman,. ‧. Arriagada, & Valenzuela 2014; Wohn et al., 2011). For example, a sample may be. sit. y. Nat. limited to one conservative community or one college student population (Zhang,. io. er. Johnson, Seltzer, & Bichard, 2010; Valenzuela, Park & Kee, 2009). Only selecting one sample population has limited previous studies’ ability to generalize research. al. n. v i n findings to area broader context.CFurthermore, h e n g cthehmajority i U of research focused on social media users’ political behavior (Jung, Kim & de Zúniga, 2011; Kushin &. Yamamoto. 2010; Rainie, Smith, Schlozman, Brady & Verba, 2012), while only a few studies examined public participation in science. Additionally, most previous studies focused on the relationship between self-efficacy and traditional media (Hofstetter, Zuniga & Dozier, 2001; Holbert, Lambe, Dudo & Carlton. 2007). For instance, television could provide viewers with an increased sense of efficacy. By giving the feelings of competence, viewers felt they better understood specific issues (Hoffman & Thomson, 2009). The relationship. 4.

(11) between internet use and self-efficacy was also discussed in previous research (Coleman, Morrison & Svennevig, 2008; Shen, Wang, Guo & Guo, 2009). The Internet provided individuals with easy access to information and thus further increased their efficacy (Kenski & Stroud, 2006). Research about SNSs and self-efficacy and their relationship to scientific participation is still rare. To examine whether SNS use certainly influences users’ self-efficacy and public participation, two methods can be used. One method is to investigate the relationship between the two variables within a single dataset, which is the approach taken by most current studies. The alternative method is to compare the. 治政 levels of self-efficacy and behaviors between users and大 non-users of a particular 立 science-related Facebook page, which assembles users with the same scientific. ‧ 國. 學. interest but different backgrounds. The alternative method is applied in this research.. ‧. The Facebook fan page “Pansci.com,” the biggest scientific online group in. sit. y. Nat. Taiwan, is taken as an example. This study contributes to expand the understanding of. io. er. the O-S-O-R model of communication effects and first uncovers the relationship between Facebook, users’ scientific knowledge and level of self-efficacy and public. al. n. v i n C hcombine the two datasets participation, and will then further e n g c h i U – Facebook fan page. users and general users – to compare the differences between specific scientific group member and general users on Facebook. The results of this study can paint a clearer picture about the role of SNSs in facilitating public engagement in science.. 5.

(12) Chapter 2. Literature Review Before laying out the theoretical framework of this study, it is necessary to take a look at some basic information. The following section examines current Internet use and social media use in Taiwan. Based on a survey of broadband Internet usage in Taiwan, 86.3 percent of people have used the Internet in the past six months, and 77.1 percent of them surf on the Internet by mobile phone. The second most frequently used device was the personal computer, as 52.5 percent of people use it to surf on the Internet. 23.1 percent of people use a notebook computers and 18.2 percent use a. 政治大. tablet PC. Additionally, 75.7 percent of people use social networking services and. 立. instant messaging devices (Jeng et al., 2016).. ‧ 國. 學. 2.1 Current Facebook Use in Taiwan. ‧. According to an InsightXplorer Biweekly Report in 2017, in Taiwan, 23.6. y. Nat. sit. percent of Facebook users are 25 to 34 years old, which is the highest percentage of. n. al. er. io. users among every other age level (6-14 years old: 6.3 percent; 15-24 years old: 22.1. i n U. v. percent; 35-44 years old: 21.3 percent; 45-54 years old: 17.2 percent; 55 years old and. Ch. engchi. above: 9.5 percent). The average age of Facebook users is 35 years old. Additionally, Facebook’s number of unique visitors who visit the website is 8.371 million, which is distinctly higher than that of other SNSs. As for the time users spend on Facebook, on average they spend one minute on a page and an average total of 14.3 minutes every time they visit. The average number of pages they visit is 13.9. Compared to other SNSs, the time users spend on Facebook is significantly more (InsightXplorer, 2017, July 17).. 6.

(13) 2.2 Definition of Scientific Use of Facebook Social Networking sites are online services that allow users to create an individual profile, present themselves, articulate their social network, maintain connections with other users – usually people known offline – and navigate through these networks of contacts. Profiles allow users to learn detailed information about their contacts, including personal background, interests, and whereabouts. Users can communicate with people they already know offline or new acquaintances through a variety of tools within the same site, including chatting, sending private message,. 治政 leaving public comments on users’ profiles, linking 大 to outside content and sharing 立. photos and videos (Gil de Zúñiga, Jung, & Valenzuela, 2012; Ellison, Steinfield, &. ‧ 國. 學. Lampe, 2007).. ‧. In this study, Facebook is taken as an example. As of December 2016, monthly. sit. y. Nat. active Facebook users around the world reached 1.86 billion (Facebook, 2017). As. io. er. one of the most popular social networks, Facebook’s sweeping influence cannot be ignored. As one of biggest SNSs around the world, Facebook includes mass. al. n. v i n C husers create an Uaccount Facebook engchi. information. Most of. by connecting with. acquaintances. The functions of commenting, joining/creating a group and liking a fan page allow users to connect with others who share similar ideas and interests with them. Moreover, Facebook provides a diverse selection of information in addition to friends’ posts, for example, topics such as education, traveling and fashion and so on. Among so many topics, Facebook users typically enjoy fun and entertaining content the most (InsightXplorer, 2016, July 15). The massive number of Facebook users indicates that Facebook is not only a social networking site but also a connection. 7.

(14) between various media, advertisers and consumers (InsightXplorer, 2017, July 17). For the reasons mentioned above, and also due to its great popularity, Facebook exemplifies all of the elements of SNSs, in particular promoting public participation, so it is a suitable case for examining how SNSs influence users’ public participation. Ellison, Steinfield & Lample (2007) created an index of Facebook intensity (FBI) to obtain a better measure of Facebook usage than frequency or duration indices. Although the number of Facebook friends and the amount of time spent on Facebook were included, the scale did not include multiple dimensions of use, which provides a more nuanced assessment than only measurement of time spent on site (Smock,. 治政大 assess various dimensions Ellison, Lampe, & Wohn, 2011). For example, some studies 立. of Facebook use by examining different features of Facebook, such as status updates,. ‧ 國. 學. sharing information, posting photos/videos, commenting, joining a group, liking a fan. ‧. page and sending private messages (Vitak et al., 2011; Smock, Ellison, Lampe, &. sit. y. Nat. Wohn, 2011; Pi, Chou, & Liao, 2013). This study took into account the features of. io. al. n. and commenting.. er. Facebook that enable or facilitate communication among users: sharing information. Ch. engchi. i n U. v. 2.3 Definition of Public Participation. In the recent years, involving the public in decision making about science and technology policy has become a trend (Rowe, & Frewer, 2000; Chopyak & Levesque, 2002; Wilsdon & Willis, 2004; Nistbet & Scheufele, 2009). Scientists gradually increasingly involve the public in their work due to “a recognition of basic human rights regarding democracy and procedural justice” (Rowe, & Frewer, 2000, p.5) and in response to perceived citizens’ mistrust of government (Rowe, & Frewer, 2000; Wilsdon & Willis, 2004).. 8.

(15) The concept of upstream engagement was mentioned when public participation was discussed (Joly & Kaufmann, 2008; Pidgeon & Rogers-Hayden, 2007; Kurath & Gisler, 2009; Wilsdon & Willis, 2004). The notion of upstream engagement supports mutual learning and two-way communication. When people discuss how to deal with unique issues, in the early stages of scientific and technological development two-way communication, such as dialogue, has to take place (Joly & Kaufmann, 2008; Pidgeon & Rogers-Hayden, 2007). Not downstream, but upstream engagement suggests a genuine and democratic science-society interaction through involving public knowledge in emerging science and technology related decision-making (Kurath & Gisler, 2009).. 立. 政治大. Controversial issues such nanotechnology, nuclear power, genetically modified. ‧ 國. 學. food or chemical hazards provide wider historical and interpretive contexts, as such. ‧. science and technology is highly relevant to people’s lives. To get research and. sit. y. Nat. development funding and support, policy and science communities try to eliminate. io. er. people’s fears and win the public’s acceptance. Upstream engagement is the best option for avoiding future conflicts over life-science research and its applications, starting. al. n. v i n from the early decision-making C on h basic research funding e n g c h i U (Pidgeon & Rogers-Hayden, 2007; Tait, 2009).. The development of public participation and how scientists involve the public in their work is separated into three phrases: 1. Public understanding of science (PUS), 2. From deficit to dialogue, and finally 3. Moving engagement upstream (Wilsdon & Willis, 2004). Based on the public understanding of science, people generally lacked knowledge of scientific facts, so scientists first embarked on a mission to inform and educate people (Jiang & Chang, 2016; Wilsdon & Willis, 2004). However, this kind of one-way communication caused some problems. When risk issues, for example global. 9.

(16) warming, nanotechnology, nuclear power, and genetically modified food are discussed exclusively within expert circles (i.e., between government officials or scientists), citizens are likely to become more distrustful and skeptical (Powell & Colin, 2008; Eden, 1996; Powell & Colin, 2009). Science is important for influencing policy decisions on how to deal with those risks. In order to avoid the ‘expertization’ of science and the scientific community’s negligence of citizens’ diversity of knowledge, it is significant to raise public profile of science and strengthen the public’s ability to shape society (Eden, 1996; Jiang & Chang, 2016). When policy and science communities started to involve the public into. 治政大public engagement moved decision-making processes, dialogue was initiated and 立 upstream (Wilsdon & Willis, 2004). Lay citizens are believed to have valuable. ‧ 國. 學. knowledge and points of view to bring to discussions, and thus their input can. ‧. contribute to more comprehensive, stronger societal decisions (Powell & Colin, 2008).. sit. y. Nat. Besides, public participation contributes to a more transparent and accountable process. io. er. of public policy-making. (Eden, 1996; Jiang & Chang, 2016; Shih, 2016). Oftentimes the public may be at the receiving end, or lowest level of top-down. al. n. v i n C hof information, such communication or a one-way flow e n g c h i U as attending a speech. At the. highest level, which is characterized by dialogue and two-way information exchange, people actively seek information through mechanisms as consultation exercises, focus groups, and questionnaires, and finally may be selected to take part in exercises that provide them with a degree of decision-making authority (Rowe, & Frewer, 2000). Public participation was defined as “a group of procedures designed to consult, involved, and inform the public to allow those affected by a decision to have an input into that decision” (Rowe & Frewer, 2000, p.6). Another definition was “the practice of consulting and involving members of the public in the agenda-setting, decision-making,. 10.

(17) and policy-forming activities of the organization of institutions responsible for such functions (Powell & Colin, 2008). According to the previous studies, public participation in science in this research was defined as activities attempting to consult, involve and influence decision makers by modifying their alternatives and redefining the social resources available to them (Rowe & Frewer, 2000; Powell & Colin, 2008; McLeod, Scheufele, & Moy, 1999) Citizens can interact with or discuss with others, related government department, and institutions. They can be involved in decision-making process by participating. 政治大. activities such as science cafes, focus groups, scenarios workshops, deliberative polls,. 立. citizen juries, and consensus conferences (Powell & Colin, 2008). Several participation. ‧ 國. 學. methods were categorized: referenda, public hearings/inquires, public opinion surveys, negotiated rule making, consensus conference, citizen’s jury/panel, citizen/public. ‧. advisory committee and focus groups (Rowe & Frewer, 2000). Other previous studies. Nat. sit. y. examined civic participation by measuring the frequency of attending a meeting to. n. al. er. io. discuss problems, purchasing products for the social values of the company they agree. i n U. v. and attending a public hearing and so on (Gil de Zúñiga, Jung & Valenzuela, 2012).. Ch. engchi. Most practical actions were included in the public participation. Some past studies focused on the importance of public participation and the necessary citizen engagement in science (Rowe & Frewer, 2000; Powell & Colin, 2008; Eden, 1996) and evaluation of effective public participation (Rowe & Frewer, 2000). Few studies focused on how people exactly gain the power to involve in public affairs (Shih, 2016) and how to identify the environmental factors that will mediate the effectiveness of the different participation methods (Jiang & Chang, 2016).. 11.

(18) 2.4 Direct relationship The relationship between scientific use of Facebook and public participation in science. Numerous previous studies about how media influence people’s behavior have indicated both negative and positive relationships between traditional media use and public participation. For example, television was found to deflect people from involvement in civic matters (McLeod, Scheufele & Moy, 1999; Wang, 2007). However, some studies pointed out that some people exhibited increased civic behavior due to gaining more information and knowledge through newspapers, and mass media provided people with more mobilizing information. 治政大 (McLeod, Scheufele & Moy, 1999). 立 ‧ 國. 學. After we entered the era of the Internet, mass information was provided on an unprecedented scale. In the era of Web 1.0, the internet allowed audience to select the. ‧. extent of their exposure to information (Wang, 2007). Unlike traditional media, the. Nat. sit. y. internet is full of interactivity, which increases people’s interest and encourages. n. al. er. io. expression of opinion (Wang, 2007). The more frequently users expressed their. i n U. v. opinions through the Internet, the more participation they exhibited (Wang, 2007).. Ch. engchi. The internet increased the accessibility, simplicity and rapidity of the communication process, which allows users to gain mass amounts of information (Wang, 2007; Shah, Kwak & Holbert, 2001). In recent years, the various functions of social media have increased the possibilities for public participation (Gil de Zúñiga, Jung & Valenzuela, 2012). Interactive and user-focused digital content is included as a feature of most SNSs. Research has indicated that the mass amount of information is available on SNSs (Robelia, Greenhow & Burton, 2011). Users are exposed to more information by. 12.

(19) others’ posting and sharing (Bode, 2016). By this, social media fills information gaps in the area of science (Nisbet & Scheufele, 2009). Moreover, social media allows users to obtain information from a network of known others and therefore may be of users’ greater interest (Bode, 2016). SNSs facilitate access to a large number of contacts, increasing the probability of reaching critical mass (Scherman, Arriagada, & Valenzuela, 2014). The social nature of this new media facilitates quicker information distribution. Furthermore, users are influenced by acquaintances they care about and are thus encouraged to display more public participation.. 治政大with common interests and Third, SNSs allow users to create and join groups 立. similar ideas (Valenzuela, Park & Kee, 2009; Scherman, Arriagada, & Valenzuela. ‧ 國. 學. 2014). Besides existing contacts, users can also find new contacts on SNSs or create a. ‧. group or fan page in which they may share and discuss specific topics or ideas that. sit. y. Nat. they are all interested in. After sharing similar ideas with one another, users. io. er. eventually organize activities together. SNSs increase individuals’ connectivity and enable users’ direct participation (Valenzuela, Park & Kee, 2009; Scherman,. al. n. v i n Arriagada, & Valenzuela 2014; C Chou, Beckjord, Moser, & Hesse, 2009). h eHunt, ngchi U 2.5 Indirect Relationship. The O-S-O-R model. Previous research has indicated that people’s behaviors can be directly predicted by different types of media exposure and attention, which can be measured by what is known as the S-R model (Stimulus-Response model) (McLeod et al., 1996; McLeod, Scheufele, & Moy, 1999; Paek, 2008). For example, the particular use of news in a newspaper was a better predictor of people’s level of knowledge and participation in the case of community integration and democratic processes (McLeod. 13.

(20) et al., 1996). However, previous scholars such as Markus and Zajonc also pointed out that the effect of media goes beyond this traditional and simple perspective, and thus the O-S-O-R model was developed (Eveland, Shah, & Kwak, 2003; McLeod, Scheufele, & Moy). Apart from the traditional S-R model, which predicts participation only directly from different types of media exposure and attention, the O-S-O-R model affirms the importance of indirect effects and mediations (McLeod, Scheufele, & Moy, 1999). The first O of the O-S-O-R model signifies the role of preexisting orientations, and the second O refers to personal-psychological factors in Stimulus that conditions media use.. 治政大 Kim, & de Zúniga, 2011; Finally, the ultimate effect of media use is Response (Jung, 立 Scheufele, Nisbet, Brossard, & Nisbet, 2004).. ‧ 國. 學. The O-S-O-R model was generally applied in studies about political. ‧. communication. Because the model equipped people with a better understanding of the. sit. y. Nat. role of the media, it is therefore regarded as an important theoretical model in political. io. er. communication studies (Holbert, 2005). Previous research pointed out the association between pre- and post-orientations, such as attitudes and norms, in studies on. al. n. v i n C hinfluence on people’s antismoking campaigns and their e n g c h i U smoking intentions (Paek,. 2008). Besides, the effect of traditional media and interpersonal communication on political participation was influenced by political knowledge and external efficacy in the case of community integration (McLeod, Scheufele, & Moy, 1999). The O-S-O-R model was applied in this study to examine the effect of social media on public participation and whether the relationship between the two is medicated by scientific knowledge and self-efficacy. The four specific sets of the model will be addressed in the following paragraphs.. 14.

(21) 2.5.1 The first orientations: sociodemographic variables and controlled variables– age, gender, education level, length of use, fan page membership, and scientific exposure to traditional media. In this framework, in addition to the set of structural, cultural, cognitive and motivational characteristics (McLeod, Kosicki, & McLeod, 2002), socioeconomic status, age and gender are included in the first orientations which is the first O in the O-S-O-R model and are important predictors of people’s participation (Jung, Kim, & de Zúniga, 2011). For example, research has presented a positive relationship between people’s age and their participation, and has suggested that the relationship between gender and participation varies across different. 治政大status tend to display higher countries. In general, those with higher socioeconomic 立 media use, knowledge, efficacy and eventually participation (Jung, Kim, & de Zúniga,. ‧ 國. 學. 2011). Besides, exposure stimulates interest and discussion, and can produce. ‧. supplemental effects to participation (McLeod, Kosicki, & McLeod, 2002). In this. sit. y. Nat. study, age, gender, education level, the length of use, the number of fan pages people. io. er. have joined, and their exposure are all included in the first orientations. These variables can affect the other dependent variables in this model, so they are residualized for all. al. n. v i n statistical tests of relationships.C Therefore, study focuses on the process of how h e n gthis chi U communication variables, such as Facebook use, influence people’s scientific participation outcomes without influence of socio-demographic variables.. 2.5.2 Stimulus: the relationship between scientific use of Facebook and scientific knowledge and self-efficacy. Scientific knowledge. If scientists and governments wanted to involve the public in decision-making, the typical first step would be to embark on a mission to inform. From the early 1970s, the US National Science Foundation uncovered the levels of. 15.

(22) people’s understanding of science through a survey of knowledge of scientific facts (Wilsdon & Willis, 2004). Many studies indicated that those who are more cognitively sophisticated are more likely to engage in activities (McLeod, Scheufele and Moy, 1999). The positive relationships between communication tools, or media, and knowledge have been discussed at length. Traditional media, such as television and newspapers, has been viewed tools to close the knowledge gap between people and increase public knowledge (Eveland & Scheufele, 2000; Do & Kincaid, 2006; Vitak et al., 2011). Besides the great amounts of information that it provides, the internet allows. 治政 users not only to seek information at any time, but 大 also to dig deeper into issues 立 through the use of hyperlinks and search engines. However, the current knowledge. ‧ 國. 學. gaps have not disappeared just because the audience no longer only receives content. ‧. via television and newspaper. Any content that users are not interested in could. sit. y. Nat. possibly be skipped over by them (Eveland & Scheufele, 2000). Besides, the internet. io. er. is a medium where individuals largely control their exposure to content. Those who are not interested in scientific issues are likely to not seek out related topics online. al. n. v i n (Kenski & Stroud, 2006). SNSsC may better fieldU for users to gain knowledge for h ebena g chi the following reasons.. A lot of studies indicate that there is an association between SNSs and people’s knowledge (Pi, Chou, & Liao, 2013; Barker, Dozier, Weiss, & Borden, 2013). First, SNSs can allow users to search and collect information, and obtain useful professional expertise (Pi, Chou, & Liao, 2013). It is easier for users to seek information and provide resources to others (Barker, Dozier, Weiss, & Borden, 2013). Second, although for most users, SNSs are used for entertainment and passing the time, users are probably exposed to the information that is shared by their friends. They can view, share and. 16.

(23) follow up on those materials posted by others, and this is how SNSs provide opportunities for users to gain more knowledge about specific issues that they may not have been interested in originally (Barker, Dozier, Weiss, & Borden, 2013). Numerous studies indicate substantial knowledge gain from media use. Although most measures of factual knowledge are often too limited to capture the full picture of what audience members take away from media use, open-ended questions measuring the audience’s cognitive complexity might reflect both personal characteristics and patterns of media use (McLeod, Kosicki, & McLeod, 2002). According to the statements, Hypothesis 1a is proposed:. 立. 政治大. H1a: Scientific use of Facebook will be positively related to the scientific. ‧. ‧ 國. 學. knowledge.. Self-efficacy. Self-efficacy is an obvious motivator of people’s behavior.. Nat. sit. y. Bandura (1995) defined perceived self-efficacy as “beliefs in one’s capability to. n. al. er. io. organize and execute the courses of action required to manage prospective situations”. i n U. v. (p.2). Efficacy beliefs have a great impact on how people think, feel, motivate. Ch. engchi. themselves and act. Bandura separated sources of efficacy beliefs into four types: mastery experience, vicarious experience, social persuasion, and physiological and emotional states. Excluding subjective mastery experience and physiological and emotional states, mass media could provide the functions of vicarious experience and social persuasion. Through vicarious experiences, people see the success of those who are similar to themselves, and believe they have the ability to attain those achievements as well. Social persuasion means that people who are verbally persuaded that they possess the capabilities to master given activities are more likely to mobilize greater. 17.

(24) efforts and sustain them than if they harbor self-doubts and dwell on personal flaws when problems arise (Bandura, 1995). Media effects on efficacy have been described in many previous studies. Traditional media, for example television and radio, cultivated people’s levels of self-efficacy. Concerning media campaigns on television, those who had high levels of exposure had higher levels of self-efficacy (Agha, 2003). However, although watching TV and reading newspapers familiarizes some people with certain knowledge and makes them believe they can make some changes in society, such information no longer reaches the youngest target. Instead, young users turn to the. 治政大 Esser, & Kunz, 2014). By Internet for information searching (Moeller, de Vreese, 立. using the Internet, people gain information and interact without the limitation of time. ‧ 國. 學. and space (Moeller, de Vreese, Esser, & Kunz, 2014). The Internet also enables normal. ‧. citizens to interact with public officials, and many websites are built as platforms to. sit. y. Nat. communicate specific issues and information (Kenski & Stroud, 2006). Moreover,. io. er. diverse perspectives and personal opinions can be found easily on the Internet, and people express themselves according to their knowledge (Shen, Wang, Guo & Guo,. al. n. v i n 2009; Jung, Kim & de Zúniga,C2011). provides an avenue for users not h e nThegInternet chi U. only to exchange opinions and values but also to participate in activities while expending significantly less time, money, and physical efforts (Jung, Kim & de Zúniga, 2011). The more frequently users expressed their opinions through the Internet, the more they had a sense of efficacy (Wang, 2007). The accessibility of information and feeling of full mastery of information made people believe that they could make certain changes (Jung, Kim & de Zúniga, 2011). However, since more knowledgeable individuals could possibly track down more information, the Internet may perpetuate knowledge gaps (Kenski & Stroud,. 18.

(25) 2006). Without the limitation of space and time, users spend a significant portion of the day using SNSs. Sharing and talking online about something users have encountered and are interested in has become the standard way to deal with information. Moreover, in the comments, SNSs users can provide others with hyperlinks to more information about the subject. On SNSs, people can encounter various types of information and knowledge. As friends on SNSs share information about some specific issues they care about, people are exposed to information even if they originally may not be interested in it. With a great amount of knowledge, SNSs stimulate users’ perceptions that they are aware of public affairs, which encourages their efficacious feeling about public. 治政大Pinkleton, Austin, & Fortman, affairs related to science (Jung, Kim & de Zúniga, 2011; 立 1998).. ‧ 國. 學. Besides, social media, with all the advantages of the Internet, provides chances. ‧. to spark conversation between people with different interests. According to Moeller,. sit. y. Nat. de Vreese, Esser, & Kunz (2014), through SNSs, readers can share, comment and. io. er. engage in debate about an article with minimal effort. Citizens’ discussions about science facilitates rational decisions while forming orientations and attitudes. The. al. n. v i n C h and deliberation positive effects of scientific discussion e n g c h i U increase self-efficacy (Jung, Kim & de Zúniga, 2011).. In the case of Facebook, the functions such as posting, commenting, and event invites can involve users in information distribution, issue discussion and even event participation. In the light of the previous argument, Hypothesis 1b is proposed:. H1b: Scientific use of Facebook will be positively related to levels of self-efficacy.. 19.

(26) 2.5.3 The second orientations: scientific knowledge and self-efficacy as mediator. The second O in the O-S-O-R model refers to what is likely to happen between the reception of a message and the response R (Shah et al., 2007), and consists of factors potentially mediating the relationship between stimulus and response (McLeod, Scheufele, & Moy, 1999). In this study, scientific knowledge and self-efficacy were treated as two important factors which might mediate the relationship between Facebook and public participation in science. Scientific knowledge. Knowledge is a predictive of active participation in public affairs (Jung, Kim & de Zúniga, 2011) and some empirical studies have shown that. 治政大participation (Ho, Brossard, scientific knowledge is positively associated with public 立. & Scheufele, 2008). Previous studies indicated that information held a variety of. ‧ 國. 學. behaviors and norms related with a democratic citizenry, and the sources of factual. ‧. knowledge included textbook facts and surveillance facts (Jennings, 1996).. sit. y. Nat. Informational media, such as news programs, can promote public participation by triggering citizens’ mental reasoning and elaboration of news events, which. er. io. promotes. participation in public affairs. The aindividuals’ iv l C n communication process among citizens influences people’s civic attitudes and h e nalso gchi U. n. subsequently. behaviors by allowing them to exchange information, elaborate on problems facing the community and learn about opportunities to participate in civic activities. Besides, knowledge could possibly spur public participation through the feeling of efficacy (Jung, Kim & de Zúniga, 2011). Self-efficacy. Previous studies indicated that self-efficacy can predict levels of public participation (Jung, Kim & de Zúniga, 2011; Shah, et al., 2007). First, through the process of discussion and deliberation inherent in making a rational decision about scientific public affairs, orientations and attitudes that support public participation are. 20.

(27) formed. Self-efficacy, a feeling of confidence and self-qualification, makes people express communicative behavior (Jung, Kim & de Zúniga, 2011). However, a high level of citizens’ sense of efficacy combined with low trust reduces their interest and involvement in public affairs, indicating that they were influenced by one-way communication which was different from two-way communication media forms (Shah et al., 2007). Second, the reason why those with higher levels of self-efficacy are likely to act is that people with higher levels of self-efficacy believe that others are affected more than themself by a given message. The belief is called the third-person effect. People might exhibit this behavior because. 治政 of their predicted influence on others of a given message大 (Pinkleton, Austin, & Fortman, 立 1998).. ‧ 國. 學. According to the aforementioned statement, Hypothesis 2a and 2b are proposed:. ‧. sit. y. Nat. H2a: Scientific knowledge will be positively related to public participation in science.. io. er. H2b: Self-efficacy will be positively related to public participation in science.. al. n. v i n C h samples. A significant 2.6 One pattern in different e n g c h i U amount of research has. indicated that Facebook is positively related to public participation. For example, people with a Facebook account engage in more public activities than those without a Facebook account (Valenzuela, Arriagada, & Scherman, 2012). Users with the same interests and hobbies can discuss topics of interest, share personal expertise and ideas, and provide experience with solving problems to others on SNSs. SNSs such as Facebook provide the function of Group and Fan pages, which offer a specific space for users with common interests to generate discussions and share information. Users. 21.

(28) can easily and efficiently interact and share information with others once they join a group or like a Fan page (Pi, Chou, & Liao, 2013). Furthermore, the Facebook fan page “Pansci.com” has been found to increase people’s public participation (Shih, 2016). Although previous studies have already indicated the positive relationship between Facebook and participation, they investigated only one survey. To understand more clearly the certain relationship between scientific use of Facebook, people’s scientific participation, and their level of efficacy, this study compares the users and non-users of a specific science-related Facebook page. “Pansci.com” users are defined as specific scientific group members. 治政大 was applied in this study. in this study. Figure 1 indicates that how O-S-O-R model 立 The Research Questions are as follows:. ‧ 國. 學. RQ1a: Will the relationship between scientific use of Facebook and scientific. ‧. knowledge vary by different samples?. sit. y. Nat. RQ1b: Will the relationship between scientific use of Facebook and levels of. io. er. self-efficacy vary by different samples?. RQ2a: Will the relationship between scientific knowledge and public participation in. n. al. Ch. science vary by different samples?. engchi. i n U. v. RQ2b: Will the relationship between self-efficacy and public participation in science vary by different samples? RQ3a: Will scientific knowledge significantly mediate the relationship between scientific use of Facebook and public participation in science? Will the relationship vary by different samples? RQ3b: Will self-efficacy significantly mediate the relationship between scientific use of Facebook and public participation in science? Will the relationship vary by different samples?. 22.

(29) “Pansci.com” users vs. General users RQ 1a, RQ1b, RQ2a,. RQ3a. RQ2b, RQ3a, RQ3b. H1a Demographics,. Scientific Use. Controlled. of Facebook. variables. H2a. 政治大. 立. Public Participation in Science. Self-efficacy. H1b. H2b. RQ3b. ‧. ‧ 國. Knowledge. 學 Orientations 2. Stimulus. Nat. y. Orientations. Scientific. n. er. io. al. sit. Figure 1. O-S-O-R model of public participation in science.. Ch. engchi. 23. i n U. v. Response.

(30) Chapter 3. Method Data The analyses of this study were based on two datasets and they were collected by a project under the auspices of the Ministry of Science and Technology (Grant # 103-2511-S-004 -008 -MY2). One dataset is for specific Facebook users who use scientific Facebook fan page “Pansci.com”, and the other is for general Facebook users. The first survey was oriented for specific Facebook users who use the scientific. 政治大 Facebook fan page “Pansci.com”. “Pansci.com” was first administered by Association 立. of Digital Culture, Taiwan (ADCT) but is now an independent company. It is the. ‧ 國. 學. biggest scientific website in Taiwan and currently over 370,000 Facebook users have. ‧. liked its Facebook fan page. Pansci.com provides a forum where scientific researchers,. y. sit. io. er. together.. Nat. scientific educators, science lovers and the general public can discuss scientific issues. “Pansci.com” clarifies and communicates scientific issues and developments in. al. n. v i n Taiwan and around the world. C To h get the public involved e n g c h i U in science and make them. understand more clearly about scientific issues, “Pansci.com” interprets scientific issues and developments in simple and relaxing ways, by video, animation, a short explanation and so on. Furthermore, “Pansci.com” also organizes various offline activities. The first survey collected data from an online survey platform Survey Monkey (http://www.surveymonkey.com), and through many distributive methods including well-known Taiwanese website PPT(telnet://ptt.cc), a sample size of N =1,160 was collected. The survey was conducted from August to December 2014.. 24.

(31) The other analyses in this study for general Facebook users were based on data collected from an online survey platform Eastern Online (EOL) Corporation with a sample size of N = 800. It was conducted for 10 days from May 22 to 31, 2015. The sample of this study shared similar demographic characteristics with the overall Taiwanese population. The cooperation rate was 40 percent. The sample was based on a carefully constructed probability sample that minimized sampling and non-response biases.. Dependent variables. 治政 Public participation in science. In this study, 大 to specifically assess the civic 立. behaviors related to science, five survey questions about consumer behavior and civic. ‧ 國. 學. behavior were applied. These five questions were taken from the report “Americans’. ‧. Actions to Limit Global Warming” (Leiserowitz, Maibach, Roser-Renouf, Feinberg,. sit. y. Nat. & Rosenthal, 2014) published by Yale University and George Mason University. The. io. er. report indicated that consumer behavior has become a way through which people express their values and concerns. Policies and actions of a company may be. n. al. Ch. influenced by consumers’ choices. Furthermore,. engchi. iv n the U report. also pointed out that. people express their values and concerns through their civic behavior, for example, writing a letter or email to related officials or donating money. The variables will be elaborated upon in chapter three. Shih (2016) separated behaviors into two categories of information participation and public participation. In this paper, users’ public participation was needed to use variables about public participation . The variables concerning public participation in science of “Pansci.com” and general samples were measured an additive index of five items with a 5-point scale (1= never; 2=seldom; 3= sometimes; 4=usually; 5= always), 25.

(32) asking respondents about their level of frequency with the following statements: (1) “How often do you participate in scientific, technology related or environmental public hearings or seminars”; (2) “How often do you participate in programs related to environmental protection or ecological rehabilitation?”; (3) “How often do you donate to environmental groups, scientific media, animal protection groups or associated non- governmental institutions?”; (4) How often do you encourage or punish a company for the sake of environmental protection?; (5) “How often do you express your opinions about environmental protection issues by writing a letter, email or calling relevant government offices or representatives?” (“Pansci.com”: M =2.05,. 治政大SD =.72, Cronbach’s alpha SD =.80, Cronbach’s alpha =.78; general: M = 1.93, 立 = .84). . ‧ 國. 學. Mediating variables. ‧. Scientific knowledge. Scientific knowledge is measured as an additive index of. y. Nat. sit. multiple choice quiz-like questions. The questions for “Pansci.com” users were (1). n. al. er. io. Which of the following is a key factor that enables an airplane to lift? (A: Air pressure. i n U. v. beneath the sing is greater than that above the wing.); (2) A farmer thinks that the. Ch. engchi. vegetables on her farm are not getting enough water. Her son suggests that they use water from the nearby ocean to water the vegetables. Is this a good idea? (C: No, because ocean water is too salty for plants grown on land.); (3) How do most fish get the oxygen they need to survive? (B: Using their gills, they take in oxygen that is dissolved in water.); (4) For which reason any people experience shortness of breath more quickly at the top of a mountain than along a seashore? (C: A lower percent of oxygen in the blood.); (5) Which is a better way to make sure whether a new medicine is valid? If 1000 volunteers want to join a scientist’s disease study, he should… (A: He should give the medicine to half of volunteers and not give to the other half of 26.

(33) volunteers.); (6) As part of a laboratory experiment, five students measured the weight of the same leaf four times. They recorded 20 slightly different weights. All of the work was done carefully and correctly. Their goal was to be as accurate as possible and reduce error in the experiment to a minimum. Which of the following is the BEST method to report the weight of the leaf? (C: Average all of the weights that were recorded.). The questions were taken from Science and Engineering Indicators 2010, which is “first and foremost a volume of records comprising the major high-quality quantitative data on the U.S. and international science and engineering enterprises” (Lanzerotti., et al., 2010). Compared to the general sample, the. 治政大STEM (science, technology, “Pansci.com” sample was composed of more people from 立 engineering, and math) fields who were assumed to have more scientific knowledge.. ‧ 國. 學. Therefore, to differentiate the knowledge of “Pansci.com” users, a more sophisticated. ‧. battery of questions was employed. The survey for general users applied questions. y. sit. n. al. er. general users, the questions were from Science and Engineering Indicators. io. For. Nat. from the latest high-quality quantitative data.. i n U. v. 2014 (Bowen., et al., 2014). Scientific knowledge was measured by an additive index. Ch. engchi. of five true/false quiz-like questions: (1) All radioactivity is man-made; (2) Electrons are smaller than atoms; (3) Lasers work by focusing sound waves; (4) The center of the Earth is very hot; and (5) The continents have been moving their location for millions of years and will continue to move in the future. Each correct answer from the respondents were counted as one point. Therefore, the index ranged from 1 to 5, where a higher value indicated a higher level of scientific knowledge (“Pansci.com”: M = 5.5, SD = .76, KR-20 = .31; general: M = 3.53, SD = 1.27, KR-20 =.58).. 27.

(34) Self-efficacy. The questions conducted about people’s efficacy included: (1) “I consider myself to be well qualified to participate in the communication of scientific knowledge or science-related issues”; (2) “I feel that I have a pretty good understanding of scientific knowledge or important scientific issues”; (3) “I think that I am better informed about science than most people”; (4) “Sometimes science seems so complicated that a person like me cannot really understand what is going on”; and (5) “A person like me is unlikely to make much contribution to the dissemination of scientific knowledge.” The respondents answered these questions on the surveys for both “Pansci.com” users and general samples, and these items were measured on. 治政大 3=no comment, 4= agree 5-point scales item from 1= strongly disagree, 2=disagree, 立. to 5= strongly agree. Items (4) and (5) were reverse coded to make the valence of all. ‧ 國. 學. question consistent (“Pansci.com”: M =㖐㗿㖐, SD = .73, Cronbach’s alpha =.78;. ‧. general: M = 2.92, SD = 0.61, Cronbach’s alpha = .76). . n. al. er. io. sit. y. Nat Independent variable . Ch. i n U. v. Scientific use of Facebook. McLeod & McDonald (1985) suggested that. engchi. attention and exposure should be included when a survey assessed media effect. Therefore, scientific use of Facebook was an additive index of five items for both the “Pansci.com” and general samples: (1) “In the past month, how often did you see information about science or technology on Facebook?” (2) “How much attention did you pay to the information on Facebook?” (3) “How often did you read the information you encounter on Facebook?”. Besides, this study also considered the features of Facebook that enable or facilitate communication among users: sharing information and commenting. Therefore, (4) “How often did you share information. 28.

(35) you encounter on Facebook? “(5) “How often did you comment on information you encounter on Facebook?” were also included. The survey on “Pansci.com” users was specifically for users’ scientific use of Facebook fan page “Pansci.com” so the items were (1) “In the past month, how often did you see information from Facebook fan page ‘Pansci.com’ on Facebook?” (2) “How much attention did you pay to the information from Facebook fan page ‘Pansci.com’ on Facebook?” (3) “How often did you read the information from Facebook fan page ‘Pansci.com’ you encounter on Facebook?” (4) “How often did you share information from Facebook fan page ‘Pansci.com’ you encounter on Facebook? “(5) “How often did you comment on. 治政 information from Facebook fan page ‘Pansci.com’ you大 encounter on Facebook?”. For 立 both “Pansci.com” and general samples, these items were measured on 5-point scales. ‧ 國. 學. item from 1= never, 2=seldom, 3= sometimes, 4=usually to 5= always (“Pansci.com”:. ‧. M = 3.18, SD = .56, Cronbach’s alpha = .71; general: M = 2.52, SD = .75,. n. al. er. io. sit. y. Nat. Cronbach’s alpha = .8).. Control variables. Ch. engchi. i n U. v. Socio-demographics. Because this analysis utilized data from two separate data collections, “0” was coded to indicate Facebook fan page “Pansci.com” users and “1” was coded to indicate general Facebook users. Demographics including age, gender, and education level also served as controlled variables in the analysis. Age was measured as a continuous variable (“Pansci.com”: M = 28, SD = 12.42; general: M = 35, SD = 8.26). Gender was a dichotomous variable with female coded as “0” and male coded as “1” (“Pansci.com”: 50 percent females; general: 50 percent females). Education was an ordinal variable and both surveys included five categories, 29.

(36) ranging from “elementary” (coded as “1”), “junior high” (coded as “2”), senior high/secondary specialized (coded as “3”), “college/undergraduates” (coded as “4”) to “graduate and above (coded as “5”). Across the “Pansci.com” and general samples, the median value for education was “college”.. Membership of science fan pages. To know the condition of distributing scientific message on Facebook, one open question was included: “how many Facebook fan pages related to science and technology have you joined?” For both. 治政大(“Pansci.com”: M = 6.68, SD “Pansci.com” and general samples, number was the unit 立 ‧. ‧ 國. 學. = 8.3; general: M = 1.46, SD = 2.81).. The length of time. The length of time included one open-ended question. sit. y. Nat. which was “in the past week, how many hours did you use Facebook including. n. al. er. io. subscribing, reading through and discussing with others and so on per day on. i n U. v. average?”. For both “Pansci.com” and general samples, the unit is hours. For example,. Ch. engchi. a respondent who used Facebook for three hours and thirty minutes in the past week, a value of 3.5 would be recorded . (“Pansci.com”: M = 5.4, SD = 3.56; general: M = 2.73, SD = 2.48).. Scientific exposure to traditional media. To know Facebook users’ levels of concern about scientific information on traditional news and media, an additive index of two items was used for both samples: (1) “How often do you see the news about science and technology in the newspaper, in a magazine, on television and radio, or 30.

(37) other sources except the Internet?”; (2) “How often do you notice the news about science and technology in the newspaper, in a magazine, on television and radio, or other sources except the Internet?”. For both “Pansci.com” and general samples, these items were measured on a 5-point scale from 1= never, 2=seldom, 3= sometimes, 2.96, SD = .8, r = .76). The study took account of all dependent variables, socio-demographic variables, controlled variables and mediating variables. All of the descriptive statistics and bivariate correlations were summarized separately for the “Pansci.com” sample and. 政治大. general public sample in Tables 1 and 2 respectively.. 立. ‧ 國. 學. Table1. ‧. Descriptions and Correlations of Variables: Demographics, Other Controlled Variables, Scientific Knowledge, Self-Efficacy and Public Participation in Science in. y. Nat. al 3 Ch. n. 1 Age. 2. 4. 5. engchi. i n U 6. v. 7. 1 .246**. 1. 3 Gender. .009. .051. 1. 4 Time. -.022. .004. .007. 1. 5 Fan Page. .020. .033. -.023. .078**. 6 Exposure. .150**. .026. -.128** -.059* .076**. 7 Scientific. .159**. .051. -.141**. 2 Education. er. io 1. Correlations. sit. the “Pansci.com” Sample. Level. .028. 1 1. .113** .252**. use of Facebook 31. 1. 8. 9. 10.

(38) 8 Scientific. .026. .044. -.158** -.105**. .030. .124** .168**. 1. knowledge 9 Self-efficacy. -.021. .114** -.174** -.073* .108** .352** .374** .284**. 10 Public. .059*. .176**. .006. .053. .106** .198** .208**. .008. 1 .239**. 1. participation. **. Correlation is significant at the 0.01 level (2-tailed). *. Correlation is significant at the 0.05 level (2-tailed).. 立. 政治大. ‧ 國. 學. Table2. ‧. Descriptions and Correlations of Variables: Demographics, Other Controlled Variables, Scientific Knowledge, Self-Efficacy and Public Participation in Science in. 2 Education. al 3 Ch. 4. 5. 6. engchi U. 1 -.059. 1. .025. .128**. 1. -.168**. -.036. -.070*. Level 3 Gender 4 Time. 1. 5 Fan Page. -.110** .116** .170** .160**. 6 Exposure. .108** .141** .244**. .067. 1 .224**. 32. 1. sit er. 2. Correlations. n. 1 Age. io. 1. y. Nat. the General Sample. v ni. 7. 8. 9. 10.

(39) 7 Scientific. -.115**. .014. .090*. .162** .226** .128**. 1. use of Facebook 8 Scientific. -.103** .236** .236**. .009. .103** .266**. .038. 1. .042. .244** .436** .201** .237**. knowledge 9. -.036. .159** .341**. -.007. .119** .134** .191** .353** .306** .238**. 1. Self-efficacy 10 Public participation **. Correlation is significant at the 0.01 level (2-tailed). *. Correlation is significant at the 0.05 level (2-tailed).. 立. 政治大. ‧. ‧ 國. 學. n. er. io. sit. y. Nat. al. Ch. engchi. 33. i n U. v. .014. .352. 1.

(40) Chapter 4. Results Data Comparison Table 3 shows the means and standard deviations of all the variables in the “Pansci.com” and general samples. All means of all the variables in the “Pansci.com” sample are higher than general sample. Notably, the educational level of “Pansci.com” users was significantly higher than general public users (t = 5.56, df = 1865.78, p< .001). Besides, the “Pansci.com” users also indicated a significantly longer time spent on Facebook (t = 19.48, df = 1856.82, p< .001) and more memberships of. 政治大. science fan pages (t = 19.88, df = 1515.61, p< .001) on Facebook than the general. 立. users. Moreover, the “Pansci.com” users demonstrated more scientific knowledge (t =. ‧ 國. 學. 39.82, df = 1181.54, p< .001), self-efficacy (t = 12.41, df = 1881.17, p< .001) and public participation in science (t = 3.44, df = 1834.14, p< .001) than the general. er. io. sit. y. Nat. Table 3. ‧. public users.. al. n. v i n Ch and Outcome Variables in The “Pansci.Com” General Public Samples e n g cAnd hi U. Descriptions Statistics (Means and Standard Deviations) of Control, Independent,. Median Gender Education level. Age The length of. Pansci User. General User. (N=1,160). (N=800). Percentage. Median. 50% Female. Percentage 50% Female. 4. College. 4. College. M. SD. M. SD. 28. 12.42. 35. 8.26. 5.4. 3.56. 2.73. 2.48. 6.68. 8.3. 1.46. 2.81. time (hr/week) Membership of. 34.

(41) science fan page Scientific. 3.53. .74. 2.96. .80. 3.18. .56. 2.52. .75. 5.5. .76. 3.53. 1.27. 3.3. .73. 2.92. .61. .80 1.93 政治大. .72. exposure to traditional media (1-5) Scientific use of Facebook Scientific knowledge (1-5) Self-efficacy (1-5) Public. 2.05. participation in. 立. science (1-5). ‧ 國. 學. To understand the specific age distribution and educational level distribution,. ‧. the study lists specific percentage. Table 4 shows that 67.8 percent of “Pansci.com”. y. Nat. sit. users were under 30 year-olds; however, only 30 percent of general public users under. n. al. er. io. 30 year-olds. The table shows that younger users is more in “Pansci.com” sample than. i n U. v. general public sample. As for education level, in Table 5, 89.2 percent of “Pansci.com”. Ch. engchi. users have college or undergraduates above degree and 69.8 percent of general public users have college or undergraduates above degree. The “Pansci.com” users have averagely higher educational level than the general users.. Table 4 Age Distribution. ≦19. “Pansci.com” Sample (N=1160) 15.1%. General Sample (N=800) 0%. 20-24 25-29. 32.8% 19.9%. 14.0% 16.0%. 35.

(42) 30-34 35-39 40-44 45-49 ≧50. 11.4% 7.4% 5.9% 4.8% 3.1%. 19.0% 17.0% 17.0% 17.0% 0%. Table 5 Education Level Distribution. Elementary. “Pansci.com” Sample (N=1160) .4%. General Sample (N=800) .1%. Junior high. 1.0%. .6%. Senior high/Secondary Specialized. 9.4%. 12.0%. 立. sit. y. Nat. Analysis. 2.5%. ‧. ‧ 國. 35.4%. 67.3%. 學. College/ undergraduates Graduate and above. 政 53.8%治大. n. al. er. io. In this study, an ordinary least squares (OLS) regression was used to test models. i n U. v. predicting each of the three following dependent variables: 1) scientific knowledge, 2). Ch. engchi. self-efficacy, and 3) public participation in science, separately. The OLS regression tested the relationships modeling. The independent variables were entered in blocks according to their assumed causal order, so as to examine their relative explanatory power. In this analysis, the following independent variables were entered in blocks: 1) Demographics and other controlled variables (age, gender, education level, membership of science fan pages, the length of time) 2) Scientific exposure to traditional media and Scientific use of Facebook 3) Scientific knowledge and self-efficacy. 36.

(43) To examine interaction effect, the two samples were combined into a sample. In the combined sample, interactions were calculated in order to test: 1) whether the relationship between scientific use of Facebook and scientific knowledge will vary by different samples; 2) whether relationship between scientific use of Facebook and levels of self-efficacy will vary by different samples; 3) whether the relationship between scientific knowledge and public participation in science will vary by different samples; 4) whether the relationship between self-efficacy and public participation in science will vary by different samples; 5) whether the relationship between scientific use of Facebook and public participation in science mediated by. 治政大6) whether the relationship scientific knowledge will vary by different samples; and 立 between scientific use of Facebook and public participation in science mediated by. ‧ 國. 學. self-efficacy will vary by different samples. As outlined earlier, statistical interactions. ‧. represent the notion that the relationship between two variables (e.g., the link between. sit. y. Nat. scientific use of Facebook and public participation in science) differs according to. io. er. Facebook users’ scientific knowledge and level of self-efficacy.. Each of the interaction terms was constructed by multiplying the standardized. n. al. Ch. values of the main effect variables. This was. engchi. iv n done U to. prevent any possible. multicollinearity problems that might arise between the interaction term and its components. The interactions between 1) data and scientific use of Facebook, 2) data and scientific knowledge, and 3) data and self-efficacy, were discussed in the final section of this study to predict the dependent variable. Table 6 showed the first set of ordinary regression models of factors predicting scientific knowledge. For the “Pansci.com” sample, females (β=-.16, p<.001) and those with scientific exposure to traditional media (β=.07, p<.05) exhibited more scientific knowledge. However, the length of time (β=-.11, p<.001) was negatively. 37.

(44) associated with scientific knowledge of “Pansci.com” users. The relationship between scientific use of Facebook (β=.13, p<.001) and scientific knowledge was significantly positive. Those with higher levels of scientific use of Facebook had more scientific knowledge. This result corresponded to Hypothesis 1a: Scientific use of Facebook will be positively related to scientific knowledge. Socio-demographics and other controlled variables accounted for 3.5 percent, and scientific exposure to traditional media accounted for 2.5 percent. The overall regression model accounted for 6.4 percent of the total variance in predicting scientific knowledge for the “Pansci.com” sample.. 治政大those with higher education For the general sample, males (β=.21, p<.001) and 立. levels (β=.20, p<.001) recorded more scientific knowledge. Moreover, the. ‧ 國. 學. relationship between scientific exposure (β=.22, p<.001) and scientific knowledge. ‧. was positive. The relationship between scientific use of Facebook and scientific. sit. y. Nat. knowledge was insignificant. Hypothesis 1a was not supported in the general sample.. io. er. Socio-demographics and other controlled variables accounted for 10.9 percent, and scientific exposure to traditional media accounted for 4.1 percent. The regression. n. al. Ch. model accounted for 15 percent of the total knowledge for the general sample.. engchi. iv n variance U in. predicting scientific. Research question 1a asked whether the relationship between scientific use of Facebook and scientific knowledge will vary by different samples. The answer was no. Socio-demographics and other controlled variables accounted for 50.5 percent, and scientific exposure to traditional media and scientific use of Facebook accounted for 2.1 percent.. 38.

(45) Table 6 Ordinary Least Regression Model with Scientific Knowledge as Outcome Variable “Pansci.com”. General Sample. Combined sample. Sample. (N=800). (N=1960). .01. -.09**. -.04*. -.16***. .21***. .04. Education level. .05. .20***. .10***. Membership of. .03. -.03. .03. (N=1160) Block 1 Age Gender (1=male). 政治大 .01 立-.11***. The length of time. --. 4%***. 10.9%***. 50.5%***. .07*. .22***. .14***. ‧ 國 Nat. Block 2. Scientific exposure. al. n. media. io. to traditional. er. Incr. R2. Scientific use of Incr. R2. Ch. .13***. Facebook. sit. (0=”pansci.com”; 1=general). -.65***. 學. --. ‧. Data. -.04*. y. science fan pages. engchi. i -.02 n U. v. .05**. 2.5%***. 4.1%***. 2.1%***. --. --. -.03. --. --. .1%. 6.4%. 15%. 52.6%. Block 3 Data x Scientific use of Facebook Incr. R2 2. Final R. Notes. Coefficients are standardized ordinary least squares regression coefficients. ***p<.001, **p<.01, *p<.05 From Table 7, for “Pansci.com” users, females (β=-.18, p<.001), respondents with education level (β=.13, p<.001), respondents with membership of science-related 39.

臉書使用與科學公眾參與： 一般大眾與科學臉書粉絲專頁使用者之比較 - 政大學術集成

臉書使用與科學公眾參與：一般大眾與科學臉書粉絲專頁使用者之比較 - 政大學術集成