Copyright by Ori Tenenboim 2020
The Dissertation Committee for Ori Tenenboim Certifies that this is the approved version of the following Dissertation:
News Engagement Logics: Examining Practices of Media Outlets and Their Audiences on Social Networking Sites
Committee:
Stephen D. Reese, Supervisor
Gina Chen
H. Iris Chyi
Thomas J. Johnson
Natalie J. Stroud
News Engagement Logics: Examining Practices of Media Outlets and Their Audiences on Social Networking Sites
by
Ori Tenenboim
Dissertation Presented to the Faculty of the Graduate School of The University of Texas at Austin in Partial Fulfillment
of the Requirements for the Degree of
Doctor of Philosophy
The University of Texas at Austin December 2020
Acknowledgements
When I was a child and my parents thought I was working on something too hard or too long, they would say: “Ori, this is not a PhD dissertation.” In November 2020, I successfully defended my PhD dissertation. Bringing my work to fruition would not have been possible without the support of mentors, colleagues, friends, and loved ones. It would also not have been possible without interviewees who took the time to share their insights with me. To all of them, I am indebted and express my gratitude.
As part of my work and PhD program at the University of Texas at Austin I got to interact with inspiring people who taught me a great deal. First, I would like to thank my dissertation adviser, Stephen Reese, for his mentorship. Dr. Reese taught me how to think conceptually by encouraging me to explicate concepts and to identify connections between them. By so doing, he made it possible for me to offer theoretical advancement. He also encouraged me to pursue my own ideas and to clearly state what I do and why it matters. He has been highly patient and supportive throughout my academic journey.
I express my thanks to the Director of the Center for Media Engagement and my dissertation committee member, Talia Stroud, for her investment in me. Dr. Stroud gave me the opportunity to be part of the Engaging News Project, which later became the Center for
Media Engagement, and to contribute to research that has practical implications. She helped me think about the role of journalism in a democracy and taught me to consider both democratic and business implications of media-related research. She also taught me to do meticulous work and has been incredibly generous to me.
I thank the Associate Director of the Center for Media Engagement and my dissertation committee member, Gina Masullo, for her support. I was fortunate to be part of
Dr. Masullo’s research group, Group of Online Media Engagement, and to later work closely with her in the Center for Media Engagement. I learned from her a great deal—from insights
iv on incivility in the digital media environment to properly conducting experiments and focus groups. She also taught me how to write in an accessible way and has been very generous to me.
I express my thanks to my dissertation committee member, Iris Chyi, for her mentorship. Dr. Chyi guided me through the process of developing my first major research project at UT. I later learned from her about the business of journalism and had the privilege to study with her newspapers’ online and print readership and newspapers’ price hikes. Our work was even featured in leading media outlets, such as Politico and The Guardian. Dr.
Chyi encouraged me to challenge conventional thinking and has been greatly supportive throughout my academic journey.
I thank my dissertation committee member, Tom Johnson, for his support. In Dr.
Johnson’s Content Analysis course I learned how to conduct a content analysis of social media posts, and as part of his Digital Media Research Program I learned to develop survey questions. He offered valuable feedback on my papers and encouraged me to pursue my ideas and to clearly articulate them. I also appreciate his great generosity.
At the start of my PhD journey, I had the privilege to work closely with Regina
Lawrence, the former director of Annette Strauss Institute for Civic Life. Thanks to Dr.
Lawrence I gained a deeper understanding of political communication and civic engagement and learned to ask better questions. I am grateful for her tremendous support and for greatly inspiring me. At Strauss I was later fortunate to work closely with Roderick Hart. In each conversation with him I learned something new—for example, how hope is different from optimism, and why reader letters to newspapers are important. I am grateful for Dr. Hart’s words of wisdom, as well as for his help, kindness, and patience.
I also had the great fortune to be mentored by friends and colleagues who were students when I arrived at UT and are now professors in other institutions. I express my
v thanks to Rachel Mourão, Shannon McGregor, and Magdalena Saldaña for their tremendous guidance, help, support, and generosity, as well as for the fun time we spent—whether it was at a party in Austin or on a boat in Prague.
I am grateful for friendships and collegial relationships formed in UT School of
Journalism and Media, including with (in alphabetical order) Lourdes Cueva Chacón, Víctor
García, Danielle Kilgo, Kyser Lough, Shuning Lu, Paro Pain, Martin Riedl, Jiyoun Suk, and
Miki Tanikawa, who have been a source of encouragement. I am also indebted to friends and colleagues related to the Center for Media Engagement for their support and encouragement, including (in alphabetical order) Melody Avant, Jessica Collier, Alex Curry, João Gonçalves,
Yujin Kim, Ivan Lacasa, Taeyoung Lee, Ashley Muddiman, Caroline Murray, Per Oleskog
Tryggvason, Christian Staal Bruun Overgaard, Cynthia Peacock, Josh Scacco, Emily Van
Duyn, María Victoria-Mas, Kelsey Whipple, Claudia Wilhelm, Tamar Wilner, and Marc
Ziegele.
Additional faculty and staff members at UT Austin (in the present or past) deserve thanks and praise for their help, support, and insight, including R.B. Brenner, Kathleen
McElroy, Sharon Jarvis, Mary Bock, Susan Nold, Kathy Warbelow, Fred Zipp, David
Donaldson, Clare Boyle, Chaz Nailor, Chuck Courtney, Melissa Huebsch-Stroud, and Sylvia
Edwards. Furthermore, as a teaching assistant I had the privilege to teach and interact with
UT journalism students. I was impressed by their creative storytelling and thank them for being a source of inspiration.
I express my thanks to Akiba Cohen and Idit Manosevitch of Israel for mentoring me and allowing me to gain research skills before I came to UT Austin. My first published academic work is with Dr. Cohen, and the development of ideas related to participatory journalism and online deliberation was made possible as part of my work with Dr.
Manosevitch.
vi
I am also grateful for the meaningful time I spent with older and newer friends and colleagues in different parts of the world. In particular, I want to acknowledge my friends
Yigal Harkavy, Boaz Anin, Ely Kovetz, Kobi Barkan, and Yoel Kornreich, who have enriched my life since childhood.
A major source of strength throughout my life has been my amazing family. In particular, I thank my parents and sister for their love. Thanks to the unwavering support and dedication of my mother, I have been able to pursue higher education. Thanks to the lessons from my father, I learned to never give up. Thanks to my sister, I have been inspired and learned that even pursuing a PhD is an achievable goal. I also express special thanks to my niece for her help and to my family members Hana and Marcel for their support and for making my visits on the East Coast unforgettable.
Last but not least, I thank my partner Alysia whose love, patience, help, and persistent encouragement have made this dissertation possible.
vii
Abstract
News Engagement Logics: Examining Practices of Media Outlets and Their Audiences on Social Networking Sites
Ori Tenenboim, PhD The University of Texas at Austin, 2020
Supervisor: Stephen D. Reese
In an attempt to build relationships with audience members in the digital media environment, news organizations operate beyond their proprietary platforms. On non- proprietary platforms, such as Facebook, they occupy spaces that are termed here triple-party news-spaces: digital spaces that involve a news publisher, a platform owner, and users. The proposed dissertation seeks to identify and explicate the underlying logics of media production and usage in these spaces. On the production side, it draws on 28 interviews to investigate how 14 news organizations in the United States of America and Israel produce messages for triple-party news-spaces. On the media usage side, it employs a content analysis of 1,600 messages and an analysis of engagement metrics for 157,962 messages to examine to what extent and how news organizations’ messages differ in the modes of engagement they generate: commenting versus sharing versus liking/reacting. By examining media production and usage in triple-party news-spaces, the dissertation develops conceptually and empirically news engagement logics that are employed in these spaces—logics by which news organizations act to evoke audience interaction with their content, and audience members actually interact with it. While audience engagement is generally important for news organizations, it is particularly important on social networking sites where algorithms prioritize posts that generate engagement. In developing news engagement logics, the
viii dissertation uses the theoretical construct of media logics, news value theory and literature on engagement enhancers, and the participatory paradigm in audience research, suggesting that certain content characteristics are associated with each of the examined modes of engagement in more than one country and other content characteristics are associated with particular modes of engagement. The dissertation also suggests that the news organizations under study strive to balance between perceived journalistic imperatives or standards and perceived rules of the social media “game” by combining older and newer logics in selecting content, deciding when to post it, choosing expression style, signaling which content deserves more attention, and determining the organizations’ approach toward user-generated content.
Business implications for news organizations and democratic implications for civic life are discussed.
ix
Table of Contents
List of Tables…………………………………………………………………………....xiii
List of Figures………………………………………………………………………….. xiv
CHAPTER 1 ...... 1
Introduction: From Newspapers to Facebook News Feed ...... 1
Study Purpose ...... 3
Justification ...... 6
Structure of the Dissertation ...... 7
CHAPTER 2 ...... 9
Theoretical Background: Journalism, Media Logics, and Engagement in a Networked Environment ...... 9
From Media Logic to Media Logics ...... 10
Older and newer journalistic norms and practices ...... 13
News organizations on social networking sites ...... 17
Audience Attention and Engagement ...... 19
Explicating engagement ...... 20
The importance of engagement...... 22
Commenting, sharing, and liking ...... 24
The Participation Paradigm, News Values, and Other Engagement Enhancers ...... 27
The Hierarchy of Influences and Different Social Systems ...... 34
Research Questions and Hypotheses ...... 36
CHAPTER 3 ...... 38
Methods: A Mixed Approach to Understanding Media Production and Usage ...... 38
Interviews...... 41
Analyzing posts and engagement ...... 43
Coding and variables ...... 47
Analysis ...... 50
x
CHAPTER 4 ...... 52
Findings: Logics of Media Production in Triple-Party News-Spaces ...... 52
Purposes of Using Social Networking Sites ...... 53
Organizational Structures and Policies ...... 58
Key Dimensions of Media Production Logics ...... 65
Item selection: "It's about finding that nugget in a story" ...... 66
Temporality considerations: Posting every 15…30…40…60 minutes ...... 77
Hierarchy considerations: "There are ways to show importance" ...... 86
Expression style: It's OK to be "a little bit more playful" ...... 91
Approach toward UGC: "Crowdsourcing stories as a way to build community" ...... 101
Summary ...... 108
CHAPTER 5 ...... 111
Findings: Logics of Media Usage in Triple-Party News-Spaces ...... 111
Commenting, Sharing, Liking, and Reacting: Correlations and Overlaps ...... 113
Analysis of Commented-On, Shared, and Liked or Reacted-To Content ...... 128
Older media logics: Topics and news values ...... 129
News values ...... 139
Newer media logics: Evaluations, emotions, and structural features ...... 147
Emotions ...... 153
Structural features ...... 158
Combining older and newer media logics ...... 161
Summary ...... 166
CHAPTER 6 ...... 170
Discussion: News Engagement Logics in Triple-Party News-Spaces ...... 170
Key Findings and Contributions ...... 171
Business and Democratic Implications ...... 177
xi
Limitations and Ways Forward ...... 181
Appendices ...... 183
Appendix A – Interview Guide ...... 183
Appendix B – Content Analysis Codebook ...... 185
Appendix C – Inter-Coder Reliability ...... 197
Appendix D – Frequencies of Content Characteristics ...... 199
References ...... 202
xii
List of Tables
Table 1: Number of Posts on Facebook Pages of U.S. News Organizations ...... 80
Table 2: Number of Posts on Facebook Pages of Israeli News Organizations ...... 80
Table 3: Pearson Correlations Between Engagement Metrics for U.S. News Organizations ...... 116
Table 4: Pearson Correlations Between Engagement Metrics for Israeli News Organizations ...... 117
Table 5: Percentage of Overlap Between Highly Engaging Posts of U.S. News Organizations ...... 119
Table 6: Percentage of Overlap Between Highly Engaging Posts of Israeli News Organizations ...... 120
Table 7: Relationships Between Topics and Modes of Engagement for News Organizations' Posts ...... 130
Table 8: Relationships Between News Factors and Modes of Engagement for Media
Outlets' Posts ...... 141
Table 9: Relationships Between Evaluative Aspects and Modes of Engagement for Media Outlets' Posts ...... 149
Table 10: Relationships Between Emotions and Modes of Engagement for News Organizations' Posts ...... 154
Table 11: Relationships Between Structural Features and Modes of Engagement for Media Outlets' Posts ...... 159
Table 12: Predicting Modes of Engagement With News Organizations' Posts on Facebook: Multivariate Negative Binominal Models ...... 163
Table 13: Inter-Coder Reliability for the Examined Variables ...... 197
Table 14: Frequencies of Content Characteristics Among Posts by Media Outlets in the U.S. and Israel ...... 199
xiii
List of Figures
Figure 1: The Main Facebook Page of The New York Times ...... 44
Figure 2: U.S. President Trump's State of the Union Address as Reflected in a
Facebook Post and the Print Front Page of The New York Times on
February 6, 2019, the Day After the Address ...... 73
Figure 3: A Facebook Post by Vox About a U.S. Political Debate on June 27, 2019 .74
Figure 4: The Words "Breaking News" or "BREAKING" in Facebook Posts by ynet
(On the Left) and USA TODAY (On the Right) ...... 89
Figure 5: Subjective Expression in a Facebook Post by Haaretz ...... 99
Figure 6: UGC on the Facebook Page of Yedioth Ahronoth ...... 103
Figure 7: A Highly Shared Facebook Post About Health ...... 136
Figure 8: A Highly Commented-On Facebook Post Related to Entertainment ...... 138
Figure 9: A Facebook Post With a Surprising or Unusual Element ...... 145
Figure 10: An Engaging Facebook Post Focusing on a Woman Giving Birth ...... 147
Figure 11: A Facebook Post Evoking Anger ...... 155
Figure 12: A Facebook Post Evoking Sadness ...... 156
xiv
CHAPTER 1
Introduction: From Newspapers to Facebook News Feed
“We’re live with Stephen Colbert,” The New York Times posted on its Facebook page in August 2018, generating a high volume of user comments and likes. The post contained a live video of a conversation between the newspaper’s culture reporter, Sopan
Deb, and the Emmy Award-winning comedian and TV host. “Get ready for laughs, and if you have any questions, leave them in the comments and Deb will ask some,” the text above the video read (The New York Times, 2018). A day earlier, the Israeli newspaper Haaretz used a different way to prompt audience interaction on Facebook: mocking the prime minister of
Israel, Benjamin Netanyahu. The newspaper posted a link to a news story according to which right-wing Netanyahu said that left-wing protesters against a new law declaring Israel the nation state of the Jewish people “want to turn Israel into an Israeli-Palestinian state.” Rather than merely reporting the prime minister’s remark, the post also included the following words: ”For the attention of Facebook user Benjamin Netanyahu: In a democratic country, different people want all sort of things” (Haaretz, 2018). These two posts demonstrate untraditional editorial choices by quality newspapers in the United States of America and
Israel that sought to engage audience members on Facebook.
News organizations take advantage of non-proprietary platforms, such as Facebook, to offer content and to connect with audience members, expanding journalism beyond traditional media outlets and designated news websites (Hermida, 2016; Westlund &
Ekström, 2018). These organizations are active where a substantial number of people are present (e.g., Facebook, 2020a; Shaban, 2019): social networking sites that allow people to connect with family, friends, and other users, to get content of interest, to interact with it, as well as to create their own content. A large number of people get their news on these sites
(Newman et al., 2020; Shearer & Matsa, 2018), where they are also offered opportunities to
1 engage with it. Facebook users, for example, can share news-related content with their friends, comment on it, and use reaction buttons to indicate that they like it or to express other emotions (e.g., anger).
On these non-proprietary platforms, news organizations occupy what I term here triple-party news-spaces: digital spaces that involve a news publisher, a platform owner
(that is not the news publisher), and users. Although news publishers select and construct messages for these spaces, platform owners can decide which messages will be more visible to users. For example, messages that generate greater user engagement—broadly described in this introduction as digital ways of interacting with content—can gain greater visibility
(Bromwich & Haag, 2018; Oremus, 2017). Thus, users play a role in determining which messages would be prioritized by platform owners and reach other users.
The production of messages for triple-party news-spaces is seen in this study as largely geared toward digital engagement, and media usage is seen as encompassing different manifestations, including different modes of engagement—e.g., commenting, sharing, and liking/reacting. These spaces contain underlying logics of media production and usage, which can be understood as “rules of the game of particular media” (Klinger & Svensson,
2018, p. 4656)—norms, procedures, and processes for constructing messages within particular media (Altheide, 2014; Altheide & Snow, 1979) as well as for media use (Klinger
& Svensson, 2015, 2018). Drawing on Chadwick’s notion of a “hybrid media system” (2013) where older and newer media logics interact, the underlying media logics of triple-party news-spaces combine older and newer logics—mass media logic and network media logic
(Klinger & Svensson, 2015). Different combinations of “old” and “new” may exist in different contexts (Chadwick, 2013)—e.g., cultural/national and organizational contexts.
Journalism scholars and practitioners have addressed challenges that social
2 networking sites pose to the news industry (e.g., Bell, 2016; Newman, 2011; Newman et al.,
2020), and a large body of research has examined how individual journalists use these sites
(e.g., Lasorsa, Lewis, & Holton, 2012; Lawrence, Molyneux, Coddington, & Holton, 2014;
Mourão & Molyneux, 2020; Tenenboim, 2017). While news organizations’ use of social networking sites has also been explored (e.g., Bullard, 2015; Hågvar, 2019; Hille & Bakker,
2013; Lischka, 2018), limited scholarly attention has been paid to the ways in which different news organizations select and construct messages for these sites in different cultural/national contexts. Scholars have also made limited attempts to identify what in these messages generates different modes of engagement—commenting versus sharing versus liking/reacting—rather than one mode (e.g., only commenting or sharing), or aggregated engagement.
Study purpose
The purpose of the dissertation is to identify and explicate the underlying logics of media production and usage in triple-party news-spaces and to examine them comparatively in different cultural/national and organizational contexts. On the media production side, I investigated through 28 interviews with media workers how 14 news organizations apply older and newer media logics in the production of messages for these spaces on social networking sites. I analyzed what constitutes the media production logics, addressing dimensions such as item selection, style of expression, and considerations related to the time of posting. I also examined the application of these logics in two countries—the U.S. and
Israel—and among two types of organizations—newspapers and digital-native outlets. These two countries were selected to capture divergent cultural features while holding other features constant. The U.S. and Israel have democratic features—e.g., in terms of the electoral process and pluralism (Economist Intelligence Unit, 2020), as well as media systems that are similar
3 in their market orientation and professional journalistic ethos (Neiger & Tenenboim-
Weinblatt, 2016; Sheafer & Wolfsfeld, 2009). Yet, unlike the U.S. culture, the Israeli culture is characterized by “straight talk” that favors simplicity of expression and explicitness of purpose (Katriel, 1986). Furthermore, in terms of journalistic culture, Israeli journalists tend to be more interventionist in their role perceptions, that is, they are more likely to perceive the pursuit of particular missions and promotion of certain values as important (e.g., a mission to decrease the cost of living for the middle class), compared with their U.S. colleagues (Hanitzsch et al., 2019; Tenenboim-Weinblatt, 2014). As I will demonstrate, these divergent features are also relevant to the production of messages for triple-party news- spaces. Taking the same approach of combining similar and different, newspapers and digital-native news outlets were selected because in both of them written text plays a role in storytelling, but these older and newer outlets may differ in production practices.
On the media usage side, I looked at digital audience engagement with messages posted by the organizations under study. Specifically, I identified through a content analysis of 1,600 posts and an analysis of engagement metrics for 157,962 posts from a one year period (July 1, 2018–June 30, 2019) to what extent and how news organizations’ posts on
Facebook differ in the modes of engagement they generate, that is, commenting, sharing, and liking/reacting. The dissertation shows which content characteristics (e.g., a topic, or a surprising element) are associated with commenting, which with sharing, and which with liking/reacting. This investigation is distinct from research focusing on associations between certain content characteristics and only one mode of engagement (e.g., Valenzuela, Piña, &
Ramírez, 2017; Weber, 2014). The dissertation also shows if content characteristics related to different modes of engagement differ across the examined cultural/national and organizational contexts. This part of the study focuses on triple-party news-spaces on
Facebook because this platform is dominant worldwide, including in the U.S. and Israel
4
(Newman et al., 2020; Perrin & Anderson, 2019; StatCounter, 2020), and is used by all of the organizations under study.
By investigating media production and usage in triple-party news-spaces, I propose what I term here news engagement logics—logics by which news organizations act to evoke audience interaction with their content, and by which audience members actually interact with it. Through a comparative study of countries and organizations, the dissertation identifies what in these logics may be the same regardless of cultural/national or organizational contexts, and what is different.
In revealing news engagement logics, the dissertation makes an important theoretical contribution by drawing not only on the theoretical construct of media logics, but also on the participation paradigm in audience research that asks how people engage with, accede to or contest media (Livingstone, 2013), news value theory that looks at what makes content newsworthy (e.g., Galtung & Ruge, 1965; Harcup & O’Neill, 2017), and literature on engagement enhancers (e.g., Amit-Danhi & Shifman, 2020; Berger & Milkman, 2012). It posits that certain news values or factors play a role not only in determining newsworthiness, but also in eliciting different modes of audience engagement (e.g., Trilling, Tolochko , &
Burscher, 2017; Weber, 2014). Furthermore, I suggest that certain content characteristics— e.g., a focus on politics and a reference to a human emotion—are associated with each of the examined modes of engagement (commenting, sharing, and liking/reacting) in more than one country, whereas certain content characteristics are associated with particular modes of engagement and not with all of them. Thus, the dissertation helps think about content units as more engaging or less engaging than others, or as engaging in different ways (e.g., prompting shares but not comments). It also makes a contribution by providing a nuanced account of which older and newer norms, procedures, and processes are layered into each other and how
5 in spaces where news organizations try to balance between perceived journalistic imperatives or standards and perceived rules of the social media “game.”
Justification
Developing conceptually and empirically news engagement logics that are employed in triple-party news-spaces is important, because these logics can help better understand how news organizations and their audiences behave on platforms non-proprietary to the news media (Westlund & Ekström, 2018). These platforms are owned by technology companies that “now control what audiences see and who gets paid for their attention, and even what format and type of journalism flourishes” (Bell & Owen, 2017, p. 9). Thus, Facebook has even been described as “swallowing” journalism (Bell, 2016). A better understanding of how news organizations and their audiences behave on platforms they do not control would allow a better understanding of journalism in the digital media environment.
Although audience engagement is generally important for news organizations and can occur on different platforms, it is particularly important on non-proprietary platforms where algorithms seem to prioritize and make more visible posts that generate engagement
(Bromwich & Haag, 2018; Oremus, 2017). The study focuses on the modes of engagement of sharing, commenting, and liking/reacting, which are important because they can be seen as manifestations of engaged citizenship (Dalton, 2016) or actualizing citizenship (Bennett,
2008)—new styles of citizenship that are rooted in self-actualization through social expression (Bennett, Wells, Freelon, 2011). As will be explained later, sharing, commenting, and liking/reacting can also be seen as different civic acts. Thus, the proposed dissertation helps better understand digital civic engagement and its drivers by providing a nuanced account of which content characteristics drive different acts of civic engagement. From a business perspective, identifying types of content that are more likely to prompt shares,
6 comments, and likes/reactions may help news organizations decide what to post and how to construct it in an attempt to reach more people.
Although the proposed contributions may give the impression that audience engagement is a benign thing, it is worth bearing in mind that engagement can also have dark sides, such as hateful speech (Quandt, 2018) and viral deception (Jamieson, 2017). This led
Quandt (2018) to offer the concept of dark participation. However, participation does not have to be dark, and social networking sites do not have to be a cesspool of deception. While it should not easily be assumed that these sites are a net positive (Lewis & Molyneaux, 2018), they have a potential to play an important role in the relationships built between news organizations and audience remembers.
This study is also important because it extends the understanding of digital journalism beyond the more often researched U.S. or U.K. context by comparatively investigating news outlets in the U.S. and Israel. It shows what transcends cultural/national contexts and what does not. Moreover, considering that audience engagement with news- related content has also been explored in other national contexts in recent years—e.g., Chile
(Valenzuela et al., 2017), Germany (Karnowski et al., 2020), and Sweden (Larsson (2018)— the present study adds to the understanding of audience engagement as a cross-national phenomenon.
Structure of the dissertation
The dissertation consists of six main parts. After the introduction on these pages
(Chapter 1), Chapter 2 sets out the theoretical argument and provides a literature review. It explicates the key concept of media logics and identifies older and newer journalistic norms and practices in the networked media environment. This chapter also explicates the key concept of engagement and its importance, presents the foundations of the participation paradigm in audience research, and addresses news values theory, engagement enhancers,
7 and the Hierarchy of Influences. Based on the proposed review, research questions and hypotheses are presented.
Chapter 3 describes and explains the mixed-method approach employed in this study: qualitative interviews with media workers and a quantitative content analysis of news-related posts on Facebook. The chapter addresses the selection of news organizations in the U.S. and
Israel, the recruitment of interviewees, ethical issues, and the process of interviewing and analyzing the transcripts. This chapter also provides information and explanations about the collection of data from Facebook, sampling posts for analysis, the process of content analysis, and the key variables in the analysis.
Chapter 4 focuses on logics of media production in triple-party news-spaces. Based on the interviews with media workers, I identify and explicate how news organizations apply older and newer logics in the production of messages for these spaces. I present and analyze the key dimensions of media production logics and the application of these logics in different cultural/national and organizational contexts.
Chapter 5 focuses on logics of media usage in triple-party news-spaces. It first shows the extent to which different news engagement metrics correlate and the degree of overlap among the most commented-on, shared, and liked/reacted Facebook posts by news organizations. Subsequent subchapters show, based on binominal regression analyses, which content characteristics of the posts are associated with different modes of engagement and how these characteristics differ across cultural/national and organizational contexts.
Chapter 6 offers a final discussion. It reviews the key findings and addresses the theoretical and practical contributions of this study. It also considers business and democratic implications, points to study limitations, and offers avenues for future research, including the investigation of engagement and its outcomes in additional digital spaces.
8
CHAPTER 2
Theoretical Background: Journalism,
Media Logics, and Engagement in a Networked Environment
The problem I am confronting in this study is how to best understand the ways in which news organizations and their audiences behave on digital platforms non-proprietary to the news media—platforms controlled by technology companies, such as Facebook
(Westlund & Ekström, 2018). These companies play a key role in shaping the distribution of content, or in determining who gains prominence and what people see in the digital media environment. They have been highly dominant to the extent that Facebook has been described as “swallowing” journalism (e.g., Bell, 2016). With the power of the technology companies in mind, I investigate how news organizations produce messages for social networking sites and how audience members use them. Theoretically, I do so by using the construct of media logic (Altheide, 2014; Altheide & Snow, 1979) and examining hybridity of older and newer media logics (Chadwick, 2013; Klinger & Svensson, 2015) which characterize the production of messages by news organizations. The production of these messages is largely geared toward audience engagement, which I explicate. To examine engagement, I am guided by the participation paradigm in audience research (Livingstone, 2013) that asks how audience members engage with different types of posts on these sites. Based on news value theory
(e.g., Galtung & Ruge, 1965; Harcup & O’Neill, 2017) and literature on engagement enhancers (e.g., Amit-Danhi & Shifman, 2020; Berger & Milkman, 2012), certain content characteristics play a role in eliciting audience engagement, but more research is needed to identify which characteristics elicit different modes of engagement—sharing versus commenting versus liking/reacting—in different cultural/national and organizational contexts. Drawing on the Hierarchy of Influences (Shoemaker & Reese, 1996, 2014), I address factors at the level of media organizations (e.g., organizational policies) and the level
9 of social systems (e.g., culture) that can contribute to shaping media content. At the level of media organizations, the dissertation examines newspapers and digital-native outlets for different markets (e.g., the mass-market newspaper USA TODAY and upper-market newspaper The New York Times). At the level of social systems, the dissertation focuses on the U.S. and Israel to capture divergent cultural features while holding other features constant, as I will explain. Through a comparative study, I seek to identify which behaviors of news organizations and their audiences on platforms non-proprietary to the news media may be similar across organizations in more than one country, and which behaviors may be different. Toward the end of the literature review, I present research questions and hypotheses that are designed to help understand these behaviors.
From Media Logic to Media Logics
The rise of digital technologies and cultures has contributed to the expansion of journalism beyond traditional media outlets and designated news websites (Anderson, 2013;
Hermida 2016; Papacharissi & Oliveira, 2012). News-related content is created and disseminated on multiple sites, such as social networking sites, by an ensemble of actors
(e.g., news organizations, bloggers, and activists), actants (material objects, such as algorithms, and content management systems), and audiences (that can be found on different platforms and devices). According to Lewis and Westlund (2015), these actors, actants, and audiences are intertwined through activities of cross-media news work, such as production, distribution, and interpretation. The news work occurs in a networked media environment where interconnections among diverse units (such as organizations, individuals, and objects) challenge “more hierarchical closed structures of traditional news outlets” (Russell, 2016, p.
150).
A key concept that can help analyze and understand journalism in the networked
10 media environment is media logics. Altheide and Snow (1979) introduced the concept of media logic in the singular, defining it as “a form of communication; the process through which media present and transmit information” (p. 10). Addressing its principles and features, Altheide (2014) explicated (p. 22):
A basic principle of media logic is that events, actions, and actors’ performances
reflect information technologies, specific media, and formats that govern
communication. A related principle is that communication guidelines become
institutionalized and taken for granted, serve as an interpretive schema, guide routine
social interaction, and thereby become integral in creating, maintaining, and
changing culture… Media logic refers to the assumptions and processes for
constructing messages within a particular medium. This includes rhythm, grammar,
and format.
In the context of media and politics, Strömbäck and Esser (2014) understood logic as appropriate behavior that is consistent within the rules and norms of the institutional context
(p. 381). In a journalistic context, Esser (2013) saw media logic as encompassing “all those imperatives” that guide news production. It combines professional aspects (such as news values for determining newsworthiness), commercial aspects (economically motivated rationales for news production), and technological aspects (medium-specific technological conditions) (p. 166–167).
The concept of media logic has been contested by several scholars. Lundby (2009) questioned the existence of single and linear media logic. Couldry (2008) criticized the tendency to identify “one single type of media-based logic that supersedes older logics across the whole of social space” (p. 378). In addition, Krotz (2009) asserted that there was no media logic independent of cultural and societal contexts.
Heeding the criticism, Klinger and Svensson (2015, 2016) talked about logics in the
11 plural, acknowledging that different logics—in the media and other realms—coexist. In a similar vein, Chadwick (2013) as well as van Dijck and Poell (2013) talked about media logics in the plural. Chadwick (2013) defined logics as technologies, genres, norms, behaviors and organizational forms. Klinger and Svensson (2018) understood media logics as
“rules of the game of particular media”—the specific norms, procedures, and processes that drive the production and distribution of content, as well as media use (p. 4656). While all the above-mentioned definitions inform the dissertation, I use Klinger and Svensson’s definition because it acknowledges the importance of media use in addition to production and dissemination.
In analyzing the current media environment, a distinction can be made between mass media logic and network media logic (Klinger & Svensson, 2015) or between mass media logic and social media logic (van Dijck & Poell, 2013). Klinger and Svensson’s analysis suggests that the two media logics differ in the inherent communication norms and practices pertaining to media production, distribution, and usage. According to mass media logic, content is generated by professional gatekeepers based on news values, transmitted to readers or viewers/listeners, and is passively consumed by them. However, network media logic suggests that content is generated by (lay) users based on individual preferences and attention maximizing, distributed by users who are like intermediaries within networks of like-minded others, and is geared toward user interaction.
These logics can be layered into each other, or be hybridized. According to
Chadwick (2013), hybridity “captures heterogeneity and those things that are irreducible to simple, unified essences,” eschews “either/or” in favor of “not only, but also” patterns of thought, and draws attention to flux, in-betweenness, and the liminal (p. 4–8). Although hybridity was a factor in media environments long before the emergence of social networking sites and may not sufficiently explain the complexity of the journalistic field (Witschge,
12
Anderson, Domingo, & Hermida, 2019), the concept of hybridity can help think in a more nuanced way about shifts in this field. Chadwick (2013) saw the media environment as a
“hybrid media system” in which older and newer media logics interact. He suggested that the key to understanding how they interact is “to try to be as specific as possible about the combination of media logics in flow in any given event, process, or context” (p. 210).
Chadwick investigated and encouraged further investigation of combination of media logics in the context of political communication, but this can also be done in journalism studies. The dissertation aims at doing so, focusing on combinations of media logics in news organizations’ spaces on social networking sites. By doing so, it seeks to develop a theoretical framework for understanding the underlying logics of media production and usage in these spaces.
Older and newer journalistic norms and practices
Research has identified long-standing journalistic norms and practices in news production. These include, among others, heavy reliance on official/bureaucratic sources
(Fishman, 1980; Gans, 1979; Tuchman, 1978) that are perceived as authorized knowers in society (Fishman, 1980), and presentation of conflicting possibilities in news reports—“X said A” and “Y said B”—in an attempt to appear impartial (Tuchman, 1972). Over the late
19th and early 20th centuries, the norm of objectivity started to take hold in U.S. journalism, guiding journalists to report “the facts” and to separate facts from values (Schudson, 2001).
More than actually separating facts from values, journalists developed a set of procedures— e.g., source attribution and the use of direct quotation—to protect themselves from criticism and to allow autonomy in news selection (Gans, 1979; Tuchman, 1972). In reporting issues of political controversy, this meant non-partisanship—a disinclination to take sides (Singer,
2005) or to appear as taking sides.
Other practices are related to the journalistic role of gatekeeping—“culling and
13 crafting countless bits of information into the limited number of messages that reach people each day” (Shoemaker & Vos, 2009, p. 1). Journalists select and construct messages based on various factors (e.g., Gans, 1979; White, 1950), including news values for assessing newsworthiness—e.g., conflict, surprise, and proximity (Galtung & Ruge, 1965; Harcup &
O’Neill, 2017). Evaluating newsworthiness is an important part of journalists’ work, but their role as gatekeepers has been challenged in the digital media environment, where various users can create and disseminate messages.
An involvement of media users who are not professional journalists in news-related processes can lead to hybridization of older and newer media logics. Meraz and Papacharissi
(2013) theorized networked gatekeeping: a “process through which actors are crowdsourced to prominence through the use of conversational, social practices that symbiotically connect elite and crowd in the determination of information relevancy” (p. 21). Meraz and
Papacharissi demonstrated this process on Twitter during the 2011 uprisings in Egypt.
Journalists were not the only ones who curated and created reports about the uprisings.
Rather, journalists and other actors, such as bloggers and activists, co-curated and co-created flows of information, inverting prior top-down models of media gatekeeping.
Other combinations of older and newer media logics can also exist on Twitter.
Lasorsa et al. (2012) and Lawrence et al. (2014) found a significant amount of opinion expression in journalists’ tweets, indicating a deviation from journalists’ role as nonpartisan, impartial information providers. But journalists were less likely to share their traditional gatekeeping role by retweeting other users’ posts. As for sourcing practices: Hermida, Lewis, and Zamith (2014) suggested that Twitter may allow broadening the range of voices in the news. They demonstrated how National Public Radio’s Andy Carvin gave considerable prominence to unofficial sources in his tweets while being a central node in the information network on the Arab Spring uprisings. However, Tenenboim (2017) showed that in wartime,
14 journalists who were members of one of the conflicting parties heavily relied on official sources—an “old” journalistic practice—though they had agency that manifested in “new” practices: retweeting critical messages and conversing with people outside official power circles. As previously suggested, journalists may both adopt features of a relatively new medium (such as Twitter) or a new media form (such as a blog) and adapt these features to their existing norms and practices (Singer, 2005; Hermida et al., 2012). That being said, journalistic norms and practices can also shift (Chadwick, 2013). What exactly shifts and to what degree may vary in different contexts.
Adapting to the digital media environment is challenging, as ethnographic research indicates. In his first ethnography, Boczkowski (2004) examined the process of innovation in the early evolution of online newspapers, showing that journalists were “reactive, defensive, and pragmatic” when faced with change (p. 48). In a later ethnography focusing on
Argentinian newsrooms, Boczkowski (2010) demonstrated dynamics of monitoring and imitation that led to homogenization of news products. Studying U.S. metro newspapers,
Ryfe (2012) found that journalists did not adapt adequately to the digital environment and failed to offer substantial innovations. Efforts to change deeply engrained habits triggered identity crises that were difficult for journalists to work through. In her study of The New
York Times, Usher (2014) showed how journalists tried to negotiate the challenges of the encounter between the print and the web. She identified three core values of online journalism—immediacy, interactivity, and participation—that were points of tension and change.
The value of immediacy points to the importance of temporality in news work
(Tuchman, 1978; Zelizer, 2018). Journalists seek to report events in close proximity to their occurrence, but whereas immediacy is common on TV and on the radio, it is not always possible in a print newspaper. Although immediacy has become a hallmark of online
15 journalism (Tenenboim-Weinblatt & Neiger, 2018; Usher, 2014), on social networking sites news organizations may also post content at frequency or time that take into consideration the algorithm of a given platform in an attempt to reach more users. As part of my analysis later,
I will demonstrate the latter practice and suggest it reflects newer media logic, whereas posting about events in close proximity to their occurrence (which news organizations also do on non-proprietary platforms) can be regarded as reflecting older media logic in the examined context.
According to Usher (2014), the value of participation was understood at The Times as being present on social networking sites. Journalists had accounts—mainly on Twitter—and used them for information gathering and brand building. Although they built their own brand, they were also seen at The Times as sub-brands of the news organization. This is why reporters were guided to be careful not to editorialize in their posts. There were also journalists who did not know how to utilize social networking sites, and some top editors saw these sites as an interruption to the print product. The Times had its own Facebook pages and
Twitter accounts, and these were used to push out content (Usher, 2014). However, the ways of producing messages for these pages/accounts were not the focus of Usher’s study. Further, the study was conducted in 2010, when “the whole idea of participation was simply so new that there were no tried and true ways to think about the process” (p. 197). Even about five years later, The Times’s editor for social media and community, Cynthia Collins, indicated that the organization’s strategy on social networking sites was still developing. At the
International Symposium on Online Journalism, she talked about “real aggressive, aggressive testing strategy on social” (ISOJ, 2015, p. 15). Although the strategy of The Times and other news organizations may still change and develop, the dissertation seeks to identify “rules of the game” that drive the production of news-related content for social networking sites and audience engagement with it.
16
News organizations on social networking sites
In an attempt to reach and attract audience members in the digital media environment, news organizations built presence on non-proprietary platforms (Westlund & Ekström,
2018)—platforms they do not own and control, such as social networking sites. They created fan pages (e.g., on Facebook) and accounts (e.g., on Twitter) and appointed “social media editors” (Silverman, 2009) to operate in what is termed here triple-party news-spaces: digital spaces that involve a news publisher, a platform owner, and users. Some organizations also created guidelines and undertook training programs on how to use these spaces while protecting their principles and brands (Newman, Dutton, & Blank, 2012).
As elaborated below, research has explored how news organizations use social networking sites or how journalists perceive this use (e.g., Bullard, 2015; Hågvar,
2019; Hille & Bakker, 2013; Jenkins & Nielsen, 2018; Lischka, 2018; Welbers &
Opgenhaffen, 2019). A survey of senior editors of U.S. print, broadcast and online news outlets (Bullard, 2015) found that the top reasons for using social networking sites very frequently were promoting stories or work (69%), reporting news in real time (59%), engaging or interacting with audiences (58%), growing readership (56%), and building brand
(53%). As for actual practices: Most of the editors said their staffs very frequently posted links to coverage on Facebook (56%) or Twitter (53%), and fewer said their staffs very frequently posed questions to users (30%), responded to Facebook comments (29%) and responded to tweets (22%). Focusing on Facebook pages of Dutch news organizations, Hille and Bakker (2013) found that they were mainly used for content promotion by posting links to stories, short summaries and announcements of upcoming broadcasts or new issues. Some organizations also invited users to post comments or to appear on a TV show. Hille and
Bakker suggested that future research should include interviews with editors-in-chief and social media editors to further understand the use of Facebook by news media.
17
Through interviews and a survey of social media editors in Finland and
Switzerland, Lischka (2018) addressed the selection of content for Facebook. Social media editors said that entertainment news was posted more often on the Facebook pages of their outlets than on the outlets’ websites. According to their evaluation, foreign politics and economic news was posted less often on Facebook. Further, social media editors perceived certain news characteristics, such as emotional and surprising elements, as more important for
Facebook news than for website news. What seems to explain this is editors’ attempt to comply with user preferences and the logic of Facebook’s News Feed algorithm that gives prominence to certain posts over others. However, editors also strived to adhere to traditional journalistic standards.
Focusing on the Facebook pages of Dutch and Flemish newspapers, Welbers and Opgenhaffen (2019) conducted computational text analysis to compare the content of the messages that present news articles to the headlines and leads of these articles.
They found that the messages contained more subjective and positive language than the headlines and leads. Based on the findings, Welbers and Opgenhaffen proposed that the use of subjective language is one of the ways in which social media logic manifests in content.
They raised the possibility that this is because news is used on social networking sites as a form of self-expression. In line with their study, Hågvar (2019) found that stating subjective points of views was one of the rhetorical strategies used by Norwegian news organizations on
Facebook, at least when stories were not highly controversial. Subjective expression in
Facebook messages by news organizations is yet to be examined in additional contexts (e.g., cultural/national and organizational). Furthermore, Welbers and Opgenhaffen (2019) offered to investigate more closely how this subjective expression takes shape and how it affects the audience. This dissertation aims at doing so as part of an investigation of how older and newer media logics are applied by news organizations in their production of messages for
18 social networking sites, and what types of messages are associated with different modes of engagement.
Audience Attention and Engagement
To elicit audience interaction with their content, news organizations first need to draw attention to the content. Capturing people’s attention can be a major challenge in the digital media environment, where various players operate and content far exceeds what people can consume or be exposed to (Chyi, 2009). The race to capture it has been dubbed as the “attention economy,” where players compete for people’s cognitive focus (Webster,
2014; Taylor, 2014). Attention is a valuable resource as it is scarce and can be monetized through advertising, subscriptions, or nonprofit funding (Stroud, 2017).
In their race for audience attention, news organizations need to face the digital dominance of major technology companies, such as Google, Facebook, and Twitter.
Emily Bell (2016), the director of the Tow Center for Digital Journalism at Columbia’s
Graduate School of Journalism, described this dominance in a speech entitled “The end of the news as we know it: How Facebook swallowed journalism.” According to her assessment, news organizations have lost control over the distribution of content to the major technology companies. The latter have become “extremely powerful in terms of controlling who publishes what to whom, and how the publication is monetized” by filtering news through algorithms and platforms that are opaque and not always predictable (Bell, 2016, para. 9).
However, news organizations wish to be present where audience members are.
Facebook, for example, had about 2.70 billion monthly active users as of June 30, 2020
(Facebook, 2020a). According to Pew, 68% of surveyed U.S. adults said they at least occasionally got news on social media. Furthermore, 67% of surveyed Facebook users reported they got news there (Shearer & Matsa, 2018).
When users of social networking sites not only get news there, but also engage
19 with it—for example, by commenting or sharing—the news may reach a wider audience. This is because the algorithms of sites such as Facebook and Twitter seem to prioritize and make more visible posts that generate engagement (Bromwich & Haag, 2018;
Oremus, 2017). Engagement has thus become a media industry buzzword and is also used in scholarly work (Lawrence, Radcliffe, & Schmidt, 2018), but there is a lack of consensus over the meaning of engagement.
Explicating engagement
Engagement in the context of media can be perceived in different ways.
According to Malthouse and Peck (2014), engagement is “the collection of experiences that readers, viewers, or visitors have with a media brand.” Further, it “causes people to ‘use’ the media brand by reading, viewing, or interacting with it” (p. 4).
Engagement can also be understood as a continuum. Ksiazek, Peer, and Lessard (2016) proposed a continuum ranging from mere exposure to interactivity, where “more (quantity) and better (quality) ways to interact with content and with other users indicate deeper engagement” (p. 505). Oh, Bellur, and Sundar (2018) conceptualized user engagement as a continuum that “begins with users’ preliminary assessment of, and interaction with, interactive media interfaces, followed by deeper absorption with media content and behavioral outcomes” (p. 739). Considering both psychological and behavioral dimensions of user engagement, they perceived it as a form of user experience that includes four components: physical interaction (e.g., scrolling, swiping, sliding, or zooming in/out an object), interface assessment (evaluating the extent to which the interface is intuitive and easy to use), absorption (deep involvement with the media content), and digital outreach
(distribution and management of online media content).
Focusing on news, scholars have also used the term news engagement (Ha et
20 al, 2018; Livingstone & Markham, 2008). It can be defined as “the involvement in news content for either personal or social purposes, which can be indicated by the effort made in obtaining and utilizing the news content among the audience” (Ha et al., 2018, p. 720). Ha and her colleagues conceptualized news engagement as a hierarchy of four levels (listed from lowest to highest): the amount of time in the consumption of news, the seeking of news in diverse media sources, exchanging of news content among audience members (e.g., sharing news stories on social networking sites), and participation in news media—for example by creating content.
The concept of news engagement can be linked to other concepts that emphasize the active role of audience members in news-related processes: participatory journalism (Bowman & Willis, 2003), citizen journalism (Allan & Thorsen,
2009), reciprocal journalism (Lewis, Holton, & Coddington, 2014), engaged journalism
(Batsell, 2015), and more. While citizen journalism may be produced entirely by everyday citizens, the other types of journalism seem to involve both professional journalists and everyday citizens. For example, Singer et al. (2011) saw participatory journalism as involving people inside and outside the newsroom who are “engaged in communicating not only to, but also with, one another. In doing so, they all are participating in the ongoing processes of creating a news website and building a multifaceted community” (p. 2).
Audience engagement may occur at different stages of the news production process: access/observation, selection/filtering, processing/editing, distribution, and interpretation (Domingo et al., 2008; Singer et al., 2011). Yet, while audience members have opportunities to engage in information-gathering (access/observation) and in distribution and interpretation, the stages of deciding what news is and how to construct it
(selection/filtering and processing/editing) are not as directly open to audience engagement
(Singer et al., 2011). This reflects a tension that journalists have struggled to reconcile
21 between what Lewis (2012) called professional logic of control over content and participatory logic of free engagement.
The dissertation focuses on audience engagement with news-related content.
In this context, I see engagement as involvement in news-related content, drawing on
Ha and her colleagues (2018). However, the dissertation particularly addresses engagement at the stages of distribution and interpretation, the final two stages of the news production process (at these stages, journalists’ control seems to be limited). Engagement at these stages is perceived here as digital ways of interacting with news-related content—e.g., sharing, commenting, and liking/reacting. By identifying news characteristics that evoke different ways of interaction, the dissertation seeks to develop news engagement logics.
The importance of engagement
Audience engagement in the forms of shares, comments, and likes/reactions can be important for both business and democratic reasons. On the business side, engagement may help increase audience reach (as mentioned earlier, engagement may make posts more visible on social networking sites) and brand loyalty (Vujnovic et al., 2010). Audience engagement in distribution of content may also be regarded as a form of “free labor” (Jönsson &
Örnebring, 2011), helping news organizations disseminate their content. On the democratic side, audience engagement offers potential for public deliberation and civic empowerment
(Manosevitch & Tenenboim, 2017).
Although news organizations tend to focus on the business side of audience engagement (Lawrence et al., 2018), its democratic side is also worthy of attention.
Sharing, commenting, and liking can be seen as manifestations of new styles of citizenship that are rooted in self-actualization through social expression (Bennett, Wells, Freelon, 2011): engaged citizenship (Dalton, 2016) and actualizing citizenship (Bennett, 2007, 2008).
According to Bennett (2008), there is a generational shift in most of the postindustrial
22 democracies from a dutiful citizen (DC) model to an actualizing citizen (AC) model. The first model is based on the notion that citizenship is a matter of duty and obligation. The ideal dutiful citizen votes in elections, participates in government-centered activities, becomes informed about public affairs by following mass media, and joins civil society organizations or expresses interests through parties. The second model (AC) is based on a diminished sense of obligation to the government. The actualizing citizen has a high sense of individual purpose, sees voting as less meaningful than more personally defined acts—e.g., volunteerism and consumerism, trusts the media and politicians to a limited extent or mistrusts them, and favors loose networks of community action that are “often established or sustained through friendships and peer relations and thin social ties maintained by interactive information technologies” (p. 14). A citizen is not necessarily either dutiful or actualizing. He or she can combine elements from both models. Focusing on citizenship in the digital environment, Bennett et al. (2011) suggest that actualizing citizenship is largely manifest in social expression: Individuals produce content and share it over peer networks that tie personal identity to engagement.
In congruence with Bennett and his colleagues, Dalton (2016) argued that the modernization of American society has transformed the norms of citizenship that shape people’s expectations and behaviors. Norms of duty-based citizenship—e.g., voting in elections, paying taxes, serving on a jury, and belonging to a political party—have declined, while norms of engaged citizenship have emerged. The latter type emphasizes “a more assertive role for citizens and a broader definition of the elements of citizenship to include social concerns and the welfare of others” (p. 6). It includes participation centered in non- electoral activities, such as being active in civil society groups and consuming products for political reasons; keeping watch on government; and helping worse-off in the U.S. and the world. In addition, engaged citizenship is characterized by “a concern for the opinion of
23 others, potentially an expression of support for a more deliberative style of political activity”
(p. 33). Dalton also mentioned that the digital environment facilitates the expansion of citizen involvement in politics, for example through online political discussions.
Based on these models of citizenship, forms of audience engagement in the digital media environment can be regarded as acts of civic engagement. Drawing on
Zuckerman’s distinction (2014) between different types of civic acts, commenting, sharing, and liking tend to be thin rather than thick, and expressive rather than instrumental. However, as discussed below, each of these acts can matter.
Commenting, sharing, and liking
Online comment sections provide a space for individuals to have their say, learn the positions of others, and interact with one another. According to a survey by the Center for
Media Engagement (Stroud, Van Duyn, & Peacock, 2016), 55% of surveyed Americans
(aged 18 or older) reported they had left an online comment on social media, a product review site or app, an online news site or app, or somewhere else. More than three quarters of respondents (77.9%) said they had read online comments at some point. Just over half
(50.7%) of respondents said they had not read news comments or left comments on news sites.
While some news organizations disabled comments for their content due to incivility and hate speech (Ellis, 2015; Kennedy, 2020; Montgomery, 2016), commenting is a widely offered avenue for audience engagement (Stroud, Scacco, & Curry,
2016; Netzer, Tenenboim-Weinblatt, & Shifman, 2014). Audience members can comment on various issues in the news, though they are particularly likely to comment on political and social issues (Stroud, Van Duyn, & Peacock, 2016; Tenenboim & Cohen, 2015).
Citizen discussions about politics and public affairs have long been perceived as vital
24 for a healthy democracy (Barber, 1984; Dahlgren, 2005; Dewey, 1939), as they can help gain political knowledge (Eveland & Thomson, 2006) and better understand important issues. As
Delli Carpini, Cook, and Jacobs (2004) put it, talking in public with other citizens “arguably provides the opportunity for individuals to develop and express their views, learn the positions of others, identify shared concerns and preferences, and come to understand and reach judgement about matters of public concern” (p. 319).
Online commenting can provide this opportunity, and it may also be linked to offline engagement in civic and political life. A recent meta-analysis shows how political expression on social media is related to offline engagement. Boulianne (2019) conducted an analysis of
133 cross-sectional studies examining the magnitude of the relationship between social media use and offline engagement (e.g., voting, talking politics, participating in street marches, signing petitions, boycotting, and volunteering), as well as the different effects of three types of social media use: getting information (e.g., reading about political issues), networking
(e.g., number of political officials or organizations a user likes/follows) and political expression (e.g., discussing politics). Boulianne found that social media use for political expression was more likely to produce a significant effect and a larger effect, compared with other types of social media use. Political expression via digital channels can thus be a driver of engaged citizenship (Shah, 2016).
Sharing and liking/recommending are other important forms of audience engagement. Users of social networking sites can share with their friends and followers content pertaining to issues of public concern, and indicate that they like a content unit or recommend it. According to a survey by Pew in the U.S. (Mitchell, Gottfried, Barthel &
Shearer, 2016), 49% of social media news consumers said they shared or reposted news stories often (11%) or sometimes (38%), and 58% of consumers said they “liked” news stories often (16%) or sometimes (42%).
25
The act of sharing allows people to play a role in determining what will reach other people and what the latter will know. When people “share” a piece of content, they increase the chance that their online friends and followers will be exposed to this piece. Even the most avid news consumers may not reach this piece of content unless their friends share it, and many people are not avid news consumers to begin with, but they see what their friends share. Indeed, Hermida, Fletcher, Korell, and Logan (2012) found that two fifths of surveyed social media users in Canada said they received news from people they follow online. Further, most online news consumers in Pew’s survey (Mitchell et al., 2016) said they mostly came across news online without actively seeking it out. Hermida and his colleagues
(2012) suggested that sharing has become central to the way people experience the news.
As with commenting, sharing can also be a form of political expression that drives engaged citizenship. Gil de Zúñiga, Molyneux, and Zheng (2014) found that social media news use in the U.S. had effects on political participation mediated via political expression.
To create a measure of political expression on social media, the researchers constructed a five-item averaged scale that included “posting personal experiences related to politics or campaigning”, “friending a political advocate or politician”, “posting or sharing thoughts about politics”, “posting or sharing photos, videos, or audio files about politics”; and
“forwarding someone else’s political commentary to other people” (p. 618).
News organizations in particular attach importance to user sharing. As Olmstead,
Mitchell, and Rosenstiel (2011) put it, “understanding not only what content users will want to consume, but also what content they are likely to pass along may be a key to how stories are put together and even what stories get covered in the first place” (p. 1). The researchers further suggested that “if searching for news was the most important development of the last decade, sharing news may be the most important of the next” (p. 10).
That said, user liking or recommending is also worthy of attention. This form of user
26 engagement can be seen as an act of civic engagement in several ways. First, by “liking” or recommending a content unit pertaining to a public matter or public figure, users explicitly express their support. They show to others that they see the message or person in a favorable way. While they may be exposed to various content units, they vote in favor of this message or person. However, Facebook now allows users to press a thumbs-up sign or another emoji, such as a crying face or a laughing face (Krug, 2016). Thus, users can explicitly express how they feel about a certain message. Liking or recommending can be seen as an act of civic engagement in another way: By pressing “Like” users may play a role in determining what will command public attention. As Dvir-Gvirsman (2019) has shown, popularity cues such as
“likes” can influence people’s attention allocation and news selection. Content units that are more “liked” may generate more attention. Further, by rewarding certain types of messages through “likes,” users may encourage other users to produce such messages. For example, using millions of user comments from The New York Times, Muddiman and Stroud (2017) found that partisanship and incivility increased user recommendations. It is plausible that the high number of recommendations may encourage users to post partisan and uncivil messages by raising their expectation to be rewarded for the messages.
The Participation Paradigm, News Values, and Other Engagement Enhancers
In view of the importance and the distinct nature of commenting, sharing, and liking, the dissertation seeks to identify to what extent and how news-related messages differ in the modes of engagement they generate. To do so, it draws on the participation paradigm in audience research (Livingstone, 2013), news value theory (e.g., Galtung & Ruge, 1965;
Harcup & O’Neill, 2017), and literature on engagement enhancers (e.g., Amit-Danhi &
Shifman, 2020; Berger & Milkman, 2012). Livingstone recognized the scale and significance of the present focus of audience research on participation, suggesting that an important task of researchers is to understand the changing conditions of communication that underpin
27 participation. After addressing several paradigms in communication research (such as the behavioral paradigm that encompasses media effects and uses and gratifications, and the incorporation/resistance paradigm that typifies Stuart Hall’s work on encoding/decoding), she proposed the participation paradigm in audience research (p. 6), whereby
The questions are, on the one hand, what modes of participation are afforded to
people by the particular media and communication infrastructures which mediate
social, cultural or political spheres of life? And, on the other hand, how do people
engage with, accede to, negotiate or contest this as they explore and invent new ways
of connecting with each other through and around media?
The second question can be addressed through an analysis of engagement metrics and an analysis of content that people engage with. The dissertation performs such analyses building on Tenenboim and Cohen (2015), who showed that 40–59% of the heavily clicked stories on a news website were different from the highly commented-on stories—a pattern that was consistent over time. The study suggested that different types of content generated different expressions of interest— consumption (clicking) versus discussion (commenting). The acts of clicking and commenting are distinct in the degree of visibility associated with them (Stroud & Muddiman, 2019): Whereas user clicks are typically not seen by other users, user comments may be seen by others in a given digital space (though aggregated data about clicks or content generating clicks may also be available). Although shares and likes/reactions can also be visible, the acts of commenting, sharing, and liking/reacting are distinct in the degree of expression they allow users.
Comments are open-ended in that they allow users to use their own words (Manosevitch,
2011). Likes/reactions are closed-ended in that they are selected from existing options
(pressing a thumbs-up sign or another emoji) and do not include users’ own words. Shares allow users to add their own expression to the reposted content, though content can merely be
28 reposted without additions. Based on Tenenboim and Cohen’s approach (2015) and considering the distinct nature of commenting, sharing, and liking/reacting, it is plausible that different news-related messages engage users in different ways. To examine that, the dissertation analyzes the extent to which comments, shares, and likes/reactions correlate and the degree of overlap among highly commented-on messages, highly shared messages, and highly liked/reacted messages on Facebook, a popular and dominant platform worldwide
(Facebook, 2020a; Newman et al., 2020). Subsequently, I examine which content characteristics of messages are associated with different modes of engagement— commenting, sharing, and liking/reacting. This is an important investigation, because it can help understand what drives different acts of engagement, allowing the development of news engagement logics.
Drawing on news values theory (e.g., Galtung & Ruge, 1965; Harcup &
O’Neill, 2017) and literature on engagement enhancers (e.g., Amit-Danhi & Shifman,
2020; Berger & Milkman, 2012), I identified five types of content-related characteristics that may be associated with modes of engagement: news topics (e.g., politics and health), news values or factors (e.g., a prominent figure and a surprise), evaluative aspects (e.g., an overall tone and expressing criticism), triggers of emotions (e.g., evoking anger or sadness), and structural features (e.g., incorporating a video, or having a written message in different lengths). Research has typically focused on relationships between some of these characteristics and a particular mode of engagement (e.g., Karnowski et al., 2020;
Valenzuela, Piña, & Ramírez, 2017; Weber, 2014), but more attention is needed to possible relationships between different types of content-related characteristics and different modes of engagement to gain a nuanced understanding of audience engagement and its enhancers.
In a media environment where different modes of engagement are offered, used, and possibly contribute to the visibility of content, such understanding is important. The
29 dissertation examines which content-related characteristics are associated with commenting, sharing, and liking/reacting, and it also tests several hypotheses about specific associations.
News topics and news values or factors can be seen as reflecting older media logics. In making news judgements, journalists prioritize content about certain topics, events, or people (e.g., Galtung & Ruge, 1965; Shoemaker & Cohen, 2006). For example, journalists may see public affairs—e.g., politics, economics, and international topics— as particularly important and may pay more attention to such affairs than audience members do
(Boczkowski et al., 2011; Wendelin et al., 2017). Journalists also tend to prioritize content that contains certain factors, such as a surprising element. According to news value theory, stories generally satisfy one or more of certain requirements if they are to be selected for publication. These requirements can be regarded as news factors or news values (Galtung &
Ruge, 1965; Harcup & O’Neill, 2017; Schulz, 1982). Galtung and Ruge (1965) proposed a list of values, such as reference to persons, and cultural proximity. Other scholars (e.g.,
Harcup & O’Neill, 2001; 2017) revisited the list and offered their own values, including conflict and surprise (what Galtung and Ruge called unexpectedness) among others.
Research has suggested that audience members may apply similar selection criteria when consuming news items or interacting with them (e.g., Boczkowski &
Mitchelstein, 2012, 2013; Karnowski et al.,2020; Tenenboim & Cohen, 2015; Trilling et al.,
2017; Weber, 2014; Ziegele, Weber, Quiring, & Breiner, 2018a). Although stories on public affairs are not necessarily highly clicked-on compared to other stories, e.g., entertainment, sports, crime, and weather, (Boczkowski et al., 2011), the latter may be highly-commented- on (Boczkowski & Mitchelstein, 2012, 2013). Research has found that politics / government items were linked to a high volume of comments (e.g., Larsson, 2018; Tenenboim & Cohen,
2015). Larsson (2018) also found that health issues were linked to a high volume of shares on
Facebook pages of Swedish newspapers. This finding is in line with Boczkowski and
30
Mitchelstein’s work (2012, 2013) that suggested that a distinct attribute of highly emailed news items was their useful character, though Boczkowski and Mitchelstein did not examine shared items on social networking sites. Based on the existing research, I hypothesize:
H1: The news topic politics / government is positively associated with
commenting, while the news topic health is positively associated with sharing.
Research focusing on news sites also suggested that one of the attributes of highly commented-on news stories can be a mention or reference to a controversy or conflict (Boczkowski & Mitchelstein, 2012, 2013; Tenenboim & Cohen, 2015). An experiment also linked a high level of controversy in an article to commenting on it (Ziegele,
Weber, Quiring, & Breiner, 2018a). Focusing on sharing, Rudat, Buder, and Hesse (2014) found that students were more likely to share on Twitter news that contained the factor of unexpectedness. In a similar vein, Boczkowski and Mitchelstein (2012, 2013) suggested that another attribute of highly emailed news items was their bizarre character. Furthermore,
Tenenboim and Cohen (2015) identified the factor of surprise as a key curiosity-arousing element that can generate clicks but not necessarily comments, though they did not examine its relationship with shares. The dissertation proposes:
H2: The news factor conflict is positively associated with commenting, while the news
factor surprise is positively associated with sharing.
Whereas news topics (particularly topics related to public affairs) and news factors are seen here as reflecting older media logics in that journalists have long made news judgements (e.g., what to report, and how prominently to display a story) based on them, evaluative aspects (e.g., an overall tone and expressing criticism), triggers of emotions (e.g., evoking anger or sadness), and multimedia features (e.g., incorporating a video or a photo) can be regarded as reflecting newer media logics. Describing things as
“new” or “newer” in this study does not necessarily mean that they did not exist. Rather, they
31 have been brought to the foreground in the era of digital media and fit affordances—broadly understood as possibilities for action (Evans, Pearce, Vitak, & Treem, 2017)—or constraints of social networking sites. Related to the earlier discussion about the norm of objectivity in
U.S. journalism, the boundaries between professional journalists and everyday citizens as well as between public and private are more blurred in the digital media environment, giving rise to a more subjective or emotional expression (Hermida & Mellado, 2020; Lasorsa et al.,
2012; Molyneux, 2015; Wahl-Jorgensen, 2016, Papacharissi & Oliveira, 2012). Although subjectivity and emotionality are not new in journalism, they are particularly prominent and may be a vital force in enabling newer forms of engagement on digital platforms, such as social networking sites (Wahl-Jorgensen, 2016). As Papacharissi and Oliveira (2012) suggested, these sites can contain affective news streams that are constructed out of subjective experience, opinion, and emotion (p. 279). In addition, news-related messages on these sites can be multimedia messages, including written texts, images, videos, and links.
The dissertation examines how evaluative aspects of news-related messages, triggers of emotions, and structural features are related to modes of engagement.
Evaluative aspects are understood here as including subjective language by news workers beyond the fact of an occurrence or issue (Lasorsa et al., 2012; Lawrence et al., 2014), criticism toward someone or something (Tenenboim, 2017, p. 3505–3506), emotional language—an explicit reference to a human emotion (Ziegele et al., 2018b), and an overall tone. Although Berger and Milkman (2012) found that news items were more likely to be emailed the more positively valenced they were, a recent study of political actors’ Facebook posts suggested that a negative tone increased the number of shares and comments (Heiss,
Schmuck, & Matthes, 2019). The relationship between the overall tone of news organizations’ posts and different modes of engagement on Facebook requires investigation.
Another type of content-related characteristics is triggers of emotions, such as
32 anger, sadness, and anxiety (Berger & Milkman, 2012). Studies have linked anger- evoking content to sharing (Amit-Danhi & Shifman, 2020; Berger & Milkman, 2012) and liking (Nave, Shifman, Tenenboim-Weinblatt, 2018). However, the study about liking was conducted before Facebook globally allowed users the opportunity to use an angry face in reaction to posts rather than only a thumbs-up sign (Krug, 2016). The research linking content to emotions shows that a content analysis is a possible way to identify relationships between triggers of emotions and engagement. Drawing on this research, the dissertation hypothesizes:
H3: A trigger of anger is positively associated with sharing and reacting.
Berger and Milkman (2012) also suggested that content evoking sadness was less likely to be emailed. However, Schultz et al. (2018) found that sad imagery did not influence message processing, although they hypothesized that sad imagery would increase message engagement via depth of processing. The relationships between news organizations’ messages that attempt to trigger sadness and modes of engagement are also investigated in the dissertation. Drawing on Berger and Milkman (2012), relationships between a trigger of anxiety and engagement are examined as well.
The final type of content characteristics is structural features, such as a video, a photo, and a link to a webpage. Multimedia features were found to be positively associated with commenting (Ksiazek, 2016), though the presence of a video or photo in a
Facebook post did not yield a statistically significant relationship with sharing (Valenzuela,
Piña, & Ramírez, 2017). The dissertation further investigates relationships between structural features and different modes of engagement.
The different types of content characteristics are analyzed in Facebook posts by different news organizations in more than one country. Thus, it is possible to understand which characteristics are associated with modes of engagement across
33 cultural/national or organizational contexts and which associations are limited to a particular context.
The Hierarchy of Influences and Different Social Systems
Media content can be shaped by various factors. Shoemaker and Reese (1996, 2014) proposed the Hierarchy of Influences model that addresses factors at five levels of analysis: individuals, routines, organizations, social institutions, and social systems. This model “takes into account the multiple forces that simultaneously impinge on the media and suggest how influence at one level may interact with that at another” (Shoemaker & Reese, 2014, p. 1).
The individual level focuses on the characteristics of specific media workers—their personal attitudes, values and beliefs; as well as their professional backgrounds and roles. The routines level addresses rules and repeated practices or forms that media workers use. The level of media organizations focuses on factors that are tied to specific organizations, such as the ownership, goals, roles, policies, bureaucratic structure, and economic viability. The level of social institutions addresses forces outside the formal media structures, such as government policy and control, media sources, advertising and public relations, and the commercial marketplace. Finally, the level of social systems focuses on the larger society as a system and on how ideological, political, economic and cultural contexts shape media performance. This level of analysis is also appropriate for cross-national comparisons (Shoemaker & Reese,
2014).
The dissertation addresses factors at different levels of analysis—for example, news values at the routines level, and the commercial marketplace at the level of social institutions—but it particularly focuses on two levels: organizations, and social systems. At the level of organizations, it looks at newspapers and digital-native outlets for different markets (e.g., the mass-market newspaper USA TODAY and upper-market newspaper The
New York Times). Newspapers and digital-native news outlets are both outlets in which
34 written text plays a role in storytelling, but these older and newer outlets may differ in production practices. At the level of social systems, the dissertation looks at two countries:
U.S. and Israel. The selection of countries for comparative studies can be based on three main approaches (Collier, 1993; Sheafer & Wolfsfeld, 2009): 1) the “most similar” systems— comparing countries that are similar on key variables of interest; 2) the “most different”— comparing countries that are highly diverse, though similar processes of changes may occur there; 3) a combination of the “most similar” and the “most different” approaches. This approach allows capturing divergent features while holding other features constant. The choice of the U.S. and Israel falls within the third approach (Sheafer & Wolfsfeld, 2009).
Both countries have democratic features—e.g., in terms of the electoral process and pluralism
(Economist Intelligence Unit, 2020). In addition, they have media systems that are similar in their market orientation and professional journalistic ethos (Neiger & Tenenboim-Weinblatt,
2016)—characteristics of Hallin and Mancini’s Liberal Model (2004). That said, while the
U.S. media system follows the Liberal Model, the Israeli media system also bears marks of other models (Peri, 2012). For example, the Israeli media system is characterized by a strong development of mass circulation newspapers (a mark of the Democratic Corporatist Model) and some level of political parallelism (a mark of the Polarized Pluralistic Model) (Hallin &
Mancini, 2004). It can thus be seen as a hybrid of models (Neiger & Tenenboim-Weinblatt,
2016; Tenenboim-Weinblatt, 2014). The Israeli journalistic culture combines Western- oriented journalistic values (Hanitzsch et al., 2019) with interventionist tendencies
(Tenenboim-Weinblatt, 2014), that is, tendencies to pursue particular missions and to promote certain values (Hanitzsch, 2007; Hanitzsch et al., 2019). Further, the Israeli culture in general is characterized by “straight talk” that favors simplicity of expression and explicitness of purpose (Katriel, 1986). As I will demonstrate, the two distinct features
35 relevant to my investigation of media production are the Israeli “straight talk” and Israeli journalists’ interventionist tendencies.
The proposed dissertation investigates ways in which different news organizations in two different social systems produce messages for triple-party news-spaces, as well as audience engagement with these messages. By so doing, it identifies and explicates media logics that underlie these spaces.
Research Questions and Hypotheses
Based on the review regarding journalism and media logics in a networked media environment, audience attention and engagement, the participation paradigm, news values and other engagement enhancers, and the Hierarchy of Influences and different social systems, the following research questions are examined:
RQ1: How are older and newer media logics applied by news organizations in their production of messages for social networking sites?
Specifically:
RQ1a: What constitutes the media logics applied by news organizations in
their production of messages for social networking sites?
RQ1b: How are media production logics for social networking sites applied in
different national and organizational contexts?
RQ2: To what extent and how do news organizations’ messages on Facebook differ in the modes of engagement they generate?
Specifically:
RQ2a: To what extent do different engagement metrics correlate?
RQ2b: What is the degree of overlap among the most commented-on, shared,
and liked/reacted messages?
36
RQ2c: Which content characteristics of messages are associated with different
modes of engagement?
RQ2d: Do the content characteristics associated with different modes of
engagement differ across cultural/national and organizational contexts?
The above questions are examined with respect to news topics, news values or factors, evaluative aspects, triggers of emotions, and structural features.
The dissertation also tests research hypotheses that were presented earlier and are summarized here:
H1: The news topic politics / government is positively associated with commenting, while the news topic health is positively associated with sharing.
H2: The news factor conflict is positively associated with commenting, while the news factor surprise is positively associated with sharing.
H3: A trigger of anger is positively associated with sharing and reacting.
The other content characteristics are examined in an exploratory manner due to a limited or inconsistent literature on relationships between them and modes of engagement.
By examining these research questions and hypotheses, I gain an understanding of the ways in which news organizations and their audiences behave on digital platforms non-proprietary to the news media. Through a mixed-method investigation
(as explained in the next chapter), the questions and hypotheses lead to findings that help reveal the underlying logics of media production and usage in triple-party news-spaces.
37
CHAPTER 3
Methods: A Mixed Approach to Understanding
Media Production and Usage
To examine the underlying logics of media production and usage in triple- party news-spaces, the dissertation employs a mixed-method approach. It is based on qualitative interviews with media professionals and a quantitative content analysis of news- related posts on Facebook. This mixed-method approach was designed to gain insight into interrelations between aspects of news-related content, its production, and audience engagement with it (Loosen & Schmidt, 2016). Specifically, interviews were conducted to elucidate lived experiences and views (Tracy, 2013) from media professionals, allowing me to identify how older and newer media logics were applied by news organizations in the production of messages for triple-party news-spaces (RQ1). A content analysis was conducted to investigate news organizations’ messages that were posted on Facebook, allowing me to examine relationships between these messages and audience engagement metrics they elicited. Through the content analysis and additional quantitative analyses of engagement metrics, I identified to what extent and how news organizations’ messages on
Facebook differed in the modes of engagement they generated—commenting, sharing, and liking/reacting (RQ2, H1-H3). The qualitative and quantitative approaches enriched each other. In interpreting the meaning of the quantitative findings, I was assisted by the qualitative data that provided context and examples. In addition, quantitative data helped demonstrate statements by interviewees. Thus, the combination of qualitative and quantitative approaches provided a rich insight into triple-party news-spaces. These approaches are detailed in the following pages after explaining the selection of news organizations for the study.
To be able to identify and explicate media logics, I selected a group of news
38 organizations. They are all for-profit organizations that are prominent in different markets in terms of circulation, traffic, or perceived status of the outlets they produce. In addition, they are all active on social networking sites. By talking with media professionals who work for these organizations, examining messages posted by the organizations, and analyzing audience engagement metrics, I discerned patterns in triple-party news-spaces.
As part of this study, I also examined the application of media logics in different cultural/national and organizational contexts. The study focuses on news organizations in the U.S. and Israel, allowing a comparative perspective beyond the typically nation-centered studies of the press (Schudson, 2000). As explained in the previous chapter, these countries were selected to capture divergent cultural features while holding other features (e.g., democratic and commercial features) constant. Taking the same approach of combining similar and different, the types of organizations selected are newspapers and digital-native news outlets—outlets that were born and grown online (Wu, 2016). In both of them written texts play a major role in storytelling (unlike television and radio), though they also tend to use audio-visual materials. However, these older and newer outlets may differ in production processes. Newspapers firms have seen their activity on digital platforms as crucial in view of declining print circulation and advertising revenue (see review in Chyi &
Tenenboim, 2017 and further information in Pew, 2019).Yet, adapting to digital platforms has been challenging for newspapers, as indicated earlier. For digital-native news outlets, activity on different digital platforms has not required a transition from one medium to another.
For each country, the dissertation addresses up to eight outlets, including outlets for different markets and outlets with divergent political leanings. Outlets with different political leanings were included to account for the wide landscape of news publishers in the examined countries even though the research questions and hypotheses did
39 not address difference attributable to the political leanings of outlets. The U.S. outlets include
The New York Times, USA TODAY, Los Angeles Times, The Dallas Morning News, South
Florida Sun-Sentinel, Vox, Slate, and The Daily Caller. These outlets consist of five newspapers (the five outlets mentioned first) and three digital-native news outlets (the outlets mentioned last). The former are major daily newspapers that were among the 100 most circulated newspapers in the U.S., according to the Editor & Publisher Newspaper Databook
(Editor & Publisher, 2018). The latter are prominent publishers of news-related content in the
U.S. digital news media landscape (Ad Fontes Media, 2020; Alexa, 2020) and were estimated to draw millions of visitors per month (Website IQ, 2020). Sites that rank political leaning of
U.S. media outlets ranked USA TODAY as relatively centrist, whereas most of the other outlets were either to its left or right (Ad Fontes Media, 2020; Media Bias / Fact Check,
2020). The New York Times, Los Angeles Times, Vox, and Slate were identified as leaning center-left or left. The Dallas Morning News was ranked as leaning center-right, and The
Daily Caller was identified as leaning right. South Florida Sun-Sentinel was regarded as relatively centrist and editorially endorsed both Republican and Democratic presidential candidates (Media Bias / Fact Check, 2020).
The Israeli outlets include Haaretz, Yedioth Ahronoth, Maariv, Israel Hayom,
Makor Rishon, ynet. Walla, and Srugim. The first five outlets are among the major daily newspapers in Israel (Verthaim, 2018), and the last three outlets are digital-native news outlets. Walla and ynet are popular news websites in Israel (Avraham, 2013; Benzaquen,
2017). Although ynet website and Yedioth Ahronot newspaper are owned by the same publisher, they operate separately. Srugim news website gained popularity mainly among religious or conservative audience (see Website IQ, 2020 for web traffic estimates). Haaretz was identified by its publisher as a liberal newspaper (Haaretz, 2016), and Makor Rishon was marketed as a newspaper for a right-leaning audience (Makor Rishon, 2019). Israel Hayom
40 was described by its editor-in-chief as a right-wing newspaper (Israel Hayom, 2017), and
Srugim website appeals to a religious audience (Srugim, 2020). The other outlets may be considered as more centrist or leaning center-left (though they tend not to explicitly self- identify).
The first part of the study is based on interviews with media professionals related to 14 of the above-mentioned news outlets—seven U.S. outlets and seven
Israeli outlets. The second part of the study is based on analyzing content and engagement data of 16 outlets—the 14 outlets, as well as The Daily Caller (U.S.) and Srugim (Israel).
These two outlets were added to further diversify the selected news organizations.
Interviews
The first part of the study was interviewing social media editors or other key actors who took part in shaping social media strategies (e.g., managers, and editors in chief) to identify how older and newer media logics are applied in the production of messages for social networking sites (RQ1). I conducted semi-structured interviews with 28 media professionals—14 professionals from seven U.S. news organizations and 14 related to seven
Israeli news organizations. The number of interviewees per organization ranged from one to three. The 28 interviewees consisted of 16 men and 12 women. The interviews were conducted in-person (17 interviews) or by phone (11 interviews). Each interview approximately lasted between 30 and 80 minutes.
The Institutional Review Board (IRB) of the University of Texas at Austin approved this part of the study in the summer of 2018, and the interviews were conducted after getting the approval—in the remaining of 2018 and in 2019. Potential interviewees were identified based on contact information on the organizations’ websites or recommendations of contact people inside or outside the examined organizations. The potential interviewees were typically contacted via email, Facebook messenger, or phone. In
41 certain organizations, an interview with one person (e.g., a director) led to interviews with other people (e.g., social media editors) to get additional perspectives. The participants were told that their participation in the study was voluntary and that they could decline to answer any question and had the right to withdraw from participation at any time. They provided their consent to participate in the study and did not receive any payment for participating.
An interview guide (see Appendix A) was prepared to ask the media actors about the organizational use of social networking sites. Specifically, they were asked about the purposes of using these sites, organizational policies or guidelines, production practices, interactions with other people within the organization, and the process of decision- making with regard to constructing and publishing messages on social networking sites. They were encouraged to share—and indeed shared—their perceptions, experiences, and dilemmas.
My background as an Israeli who studied journalism and media engagement in the
U.S. and as a former news editor and head of the news desk at the Israeli website Walla helped me create a rapport with interviewees. I spoke “their” language—English in the U.S. and Hebrew in Israel—while using some terms of an insider (e.g., Israeli editors may say “the main,” referring to the main headline), shared with them some journalistic or research experiences when deemed appropriate, and showed understanding of experiences and dilemmas they shared. Although interviewees were generally cooperative, people I contacted in the process of recruiting interviewees were not always cooperative or were willing to communicate with me only off the record (they were not included in the study), particularly in some U.S. news organizations (e.g., The Wall Street Journal).
All of the interviews were audio recorded and later transcribed. The transcripts were coded in MaxQDA, a software for organizing and analyzing data. I marked with codes sentences in each transcript, identifying different topics and elements. Subsequently, I
42 narrowed down and refined the categories I had created, with the aim of identifying key dimensions of media logics applied by news organizations in the production of messages for social networking sites. The development of categories was based on interplay between patterns emerging from the qualitative data and insights from literature pertaining to news media production (e.g., Domingo et al., 2008; Galtung & Ruge, 1965; Harcup & O’Neill,
2017; Singer et al., 2011; Tuchman, 1978). This approach draws on the idea of grounded theory while acknowledging that developing categories from data is not a purely inductive process (see Kelle, 2007 for approaches in grounded theory, and see Reichertz, 2007;
Strübing, 2007 for ideas of abductive reasoning). The analysis I conducted resulted in five key dimensions of media production logics (RQ1a). The application of these logics in different cultural/national and organizational contexts (RQ1b) was also identified in this process.
To protect the participants’ anonymity, pseudonym names were assigned to them and are used in the presentation of findings. The selected names were typical in the examined cultural/national contexts.
Analyzing posts and engagement
To examine to what extent and how news organizations’ posts on Facebook differed in the modes of engagement they generated (RQ2) and to test hypotheses regarding associations between content characteristics and modes of engagement (H1-H3), an analysis of posts and engagement data was conducted. This part of the study focuses on
Facebook because it is a dominant platform around the world, including in the U.S. and Israel
(Newman et al., 2020; Perrin & Anderson, 2019; StatCounter, 2020), and was used by all of the news organizations under study. Posts and engagement data from July 1, 2018, to June
30, 2019—a one-year period—were collected. During this period, interviews with media actors were conducted (for the first part of the study). Therefore, I was able to connect issues
43 addressed in the interviews to the content analysis. A one-year period was also selected to allow capturing various issues in the news.
To collect data, the main Facebook page of each news organization was first identified. Whereas a Facebook profile is typically created for users’ personal needs, a
Facebook page allows individuals or organizations to connect with fans or customers. Those who like or follow a page can get updates from the page in their News Feed (Facebook,
2020b). A Facebook page contains an image—e.g., a company logo—at the top, and information, such as an email address and a phone number. Managers of a page can post on it written and audio-visual materials, and users can interact with these materials by commenting, sharing, and liking/reacting. Users can also send messages to the managers of a page.
Figure 1
The Main Facebook Page of The New York Times
44
The news organizations under study had one or more Facebook page: a main page that could include content from different sections of the newspaper or news website (e.g., books section), and possibly one or more pages that represented specific sections. For example, The
New York Times had a main page with this name, as shown in Figure 1, and other pages—The
New York Times Books, and The New York Times Opinion Section. This study focused on the news organizations’ main pages, which were considered by these organizations as their flagship pages and typically attracted the highest number of likes or followers.
Data about posts that appeared on the news organizations’ pages were collected. A post is understood in this study as a message that can contain a written text, audio-visual elements (such as a video or a photo), and/or a link. For example, The New
York Times’ page shown in Figure 1 contains a post with an embedded link to a story that appeared on the Times’ website. By embedding the link, the creator of the post was also able to include a photo as part of the post. In addition, the post contains a written text that introduces the story (or an accompanying text). Throughout the study I use either the word posts or messages to refer to the same units of content.
The collection of data about news organizations’ posts was conducted through
CrowdTangle, a social media monitoring tool owned and operated by Facebook
(CrowdTangle Team, 2020). The tool has access to historical Facebook data, including data about all the posts that appeared on a Facebook page in a given time period. A list of data can be obtained by providing a page username (e.g., nytimes) and a selected time period. For each news organization, data about the posts published in the second half of 2018 (from July
1 to December 31) and the first half of 2019 (from January 1 to June 30) were separately collected (for each period). The data included the content of posts, links to the posts, time of posting, the number of likes the Facebook page had at the time of posting, and engagement metrics (the numbers of comments, shares, likes, and other reactions) for each post. The data
45 also indicated if a post was sponsored, that is, if the content was an ad in the form of a post that someone paid to publish on a news organizations’ page. Sponsored posts were excluded from the data.
In total, data about 157,962 posts were left—115,207 posts from eight U.S. organizations and 42,755 posts from eight Israeli organizations. For U.S. news organizations, the number of posts in a given time period (the second half of 2018 or the first half of 2019) ranged from 4,668 (The Dallas Morning News, 2019) to 11,093 (The New York
Times, 2018). For Israeli news organizations, the number of posts in a given time period ranged from 312 (Srugim, 2018) to 4,961 (Maariv, 2018).
To examine the extent to which different engagement metrics correlated
(RQ2a) and the degree of overlap (expressed as the percentage of the same items that appeared in two examined groups) among the most commented-on, shared, and liked/reacted posts (RQ2b), correlational analyses were conducted, and percentages of overlap were calculated. Detailed explanations about the analyses are provided in the findings section
(Chapter 5).
To examine associations between content characteristics of posts and modes of engagement (RQ2c and H1-H3) and to identify whether content characteristics associated with different modes of engagement differed across cultural/national and organizational contexts, a content analysis was carried out. This analysis was conducted on a sample of 1,600 posts, consisting of posts published by the 16 organizations under study. To sample these posts, two steps were taken. First, a constructed week sampling was used:
“Sample dates are stratified by day of the week to account for systematic variation due to day of week” (Riffe, Aust, & Lacy, 1993, p. 133). Based on the approach that for online news content at least two constructed weeks are needed in a six-month period (Hester & Dougall,
2007), four constructed weeks were created—two for the last six months of 2018 and two for
46 the first six months of 2019. The total number of posts published on the days of the constructed weeks was 12,191: 8,969 posts by the U.S. organizations and 3,222 posts by the
Israeli organizations. Second, systematic random sampling (selecting every nth post after selecting a random starting point) was used to select 800 U.S. posts (for eight organizations) and 800 Israeli posts (for eight organizations). To get this equal number of posts, the systematic random sampling was separately conducted among the U.S. posts and among the
Israeli posts while using different intervals and random starting points.
Coding and Variables
The unit of analysis was a Facebook post, and coders were asked to address only the post’s content as described above and not the content of user comments attached to it. The analysis was based on a detailed codebook, as can be seen in Appendix B.
The posts were analyzed by the author and three trained native speakers of English or
Hebrew. Inter-coder reliability was examined for 150 of the U.S. posts (18.75%) and 150 of the Israeli posts (18.75%). The inter-coder reliability was examined using Krippendorff’s alpha, and the coefficient alpha for the variables included in the analysis ranged from .708 to
1.000. The coefficient was lower than .700 for three variables (frequency, awe, and humor), which, as a result, were disqualified and not included in the analysis presented in the findings section (Chapter 5). Krippendorff’s alphas for the examined variables in the U.S. sample and the Israeli sample (calculations were conducted separately) are shown in Appendix C. An additional measures of inter-coder reliability—Cohen’s kappa—is also shown in that appendix for comparison.
The variables included in the analysis and their measurements are described below. The detailed frequencies in the general sample (n = 1,600, US + Israel), US sample (n = 800) and Israeli sample (n = 800) appear in Appendix D. Frequencies also appear in the findings section (Chapter 5).
47
To examine news topics, coders were asked to identify the main topic of a given post. A list of topics was created building on Tenenboim and Cohen (2015) and
Tewksbury (2003). The list includes topics such as politics / government, military / defense, economy / business, crime, health, education, disasters / accidents, arts / entertainment, sports, and food. Coders were also presented with the options of “other” (and were asked to specify what other meant) and “hard to determine.”
Coders were also asked to identify news factors in the posts under study
(present / absent): conflict, surprise, prominent figure, reference to persons, impact, damage, success, and proximity (Galtung & Ruge, 1965; Harcup & O’Neill, 2017;
Tenenboim & Cohen, 2015; Weber, 2014; Ziegele et al., 2018a, 2018b). Drawing on
Burscher, Odijk, Vliegenthart, de Rijke, and de Vreese (2014) and Valenzuela and his colleagues (2017), the coders were asked to determine if a given post “reflects disagreement between parties, individuals, groups, or countries” (conflict) (Valenzuela et al., 2017, p. 197).
Building on Harcup and O’Neill (2017) and Tenenboim and Cohen (2015), the coders were asked to determine if the post contained an element of surprise or something unusual. The coders were told that the coding should be yes if, for example, the post reported that a political candidate who had been trailing in the polls was now leading, or it addressed a terrorist attack in Oslo—a place that was perceived as peaceful, or it mentioned a rude statement by a person who was considered polite.
Furthermore, the coders identified if the post mentioned a prominent figure, such as a president or an actress (Tenenboim & Cohen, 2015), and if the post contained a reference to persons (Galtung & Ruge, 1965) by using a person’s name or showing a person’s face.
To examine impact, the coders determined whether the post addressed an
48 event or issue that had consequences for one or more social categories of groups, such as nations, workers, or students (Weber, 2014).
Adopting Ziegele and his colleagues’ measures of damage and success
(2018b), the coders identified if the post referred to a negative consequence of an event, such as a death, injury, illness, or material damage; and if the post referred to a positive consequence of an event, such as an improvement of the standards of living, breakthrough in health research, or raising money to help the elderly.
The coders were also asked to determine proximity (Weber, 2014) to the U.S. and Israel. For example, coders of U.S. posts were asked if the post addressed an occurrence or occurrences in the U.S. only, a country or countries other than the U.S., the
U.S. and at least one country, or other option.
In addition to identifying news topic and news factors, coders were asked to assess evaluative aspects, triggers of emotions, and structural features. As explained earlier, news topics and news factors are seen in this study as reflecting older media logics (journalistic decisions have long been based on news topics and factors), whereas evaluative aspects, triggers of emotions, and multimedia features are seen here as reflecting newer media logics.
To examine evaluative aspects of the posts under study, coders assessed the use of evaluative language, criticism, emotional language, and tone. Coders assessed whether a given post contained an “evaluative language or… unattributed commentary beyond the fact of an occurrence or issue” (Lawrence et al., 2014, p. 10) (yes / no). They were asked to code the post as containing criticism when it included “evaluative language that expresses disapproval of someone or something” (Tenenboim, 2017, p. 3505–3506). The coders were asked to identify the post as containing emotional language if it explicitly referred to a human emotion (Ziegele et al., 2018b). Furthermore, the coders were asked to
49 determine the overall tone of the post: positive, negative, or neutral. The option of “hard to determine” was also provided.
The coders were also asked to identify posts that attempted to evoke anger, sadness, and anxiety. In their study, Berger and Milkman (2012) provided coders with a Wikipedia link to the definition of each emotion and asked them to rate articles on a five- point Likert scale (e.g., how much anxiety an article evokes). Building on this approach, the codebook contains definitions of emotions as they appear in the online Oxford dictionary.
However, rather than asking coders to rate a given post on a scale, they were asked to determine whether it evoked each emotion (the three categories to choose from were yes, no, or hard to determine). This approach was taken to increase the chance that satisfactory inter- coder reliability would be reached (also see Amit-Danhi & Shifman, 2020 who examined the presence of emotional engagement enhancers).
Finally, the coders addressed structural features of the posts (Ksiazek, 2016,
Lewis, Zamith, & Hermida, 2013): a video, a photo, a link, and a written text. The coders determined whether a given post contained a video player, at least one photo, and a link to a webpage. If a post contained a link, the coders assessed if the link was to a news report, commentary or opinion piece, magazine feature, or other. The coders also addressed the length of a written text in a post by assessing if the post contained a sentence or less than a sentence, more than a sentence and not more than a paragraph, more than one paragraph, or no written text.
Analysis
Relationships between the content characteristics described above (from news topics to structural features) and modes of engagement (commenting, sharing, and liking / reacting) were examined. The content characteristics were seen as the independent variables, whereas the modes of engagement were seen as the dependent variables. The type
50 of statistical analyses conducted was regression analyses. Specifically, negative binominal regression analyses were performed as they are suitable for over-dispersed count data. Count data, such as the number of comments or shares per post, consist of non-negative integers {0,
1, 2…}, and over-dispersion means that the conditional variance is greater than the conditional mean. Negative binominal regression can accommodate over-dispersion in the data (Cameron & Trivedi, 2013). This type of regression was previously used for predicting sharing on social media (Valenzuela et al., 2017) and participation in an article’s comments section on a news site (Weber, 2014). The statistical procedures used in this study and their interpretation are presented in the findings section (Chapter 5).
By analyzing the interviews with the producers of messages, as well as analyzing the content of published messages and the relationships between elements in these messages and user engagement metrics, I gained insights into both the production and usage of messages.
These insights allow me to propose news engagement logics in triple-party news-spaces. The next chapter (Chapter 4) focuses on the production side based on the interviews. It is followed by a chapter focusing on the usage side based on the quantitative analyses and additional insights from the interviews (Chapter 5). The final chapter offers an integration of the production and usage sides of news engagement logics.
51
CHAPTER 4
Findings: Logics of Media Production in Triple-Party News-Spaces
This chapter focuses on the underlying logics of media production in triple-party news- spaces. These logics can be understood as the norms, procedures, and processes (Klinger &
Svensson, 2018) that drive the production of messages posted in the spaces under study. In this chapter, I identify and explicate how news organizations apply older and newer logics in the production of messages for triple-party news-spaces (RQ1). “Old” and “new” are layered into each other or hybridized (Chadwick, 2013), constituting the logics I investigate.
Describing things as “new” or “newer” in this study does not necessarily mean that they did not exist. Rather, they have been brought to the foreground in the era of digital media and fit affordances or constraints of social networking sites. I present and analyze key dimensions of the investigated media production logics (RQ1a) and the application of these logics in different cultural/national and organizational contexts (RQ1b). Their application reflects attempts by different news organizations in two countries to strike a balance between perceived journalistic imperatives or standards and perceived rules of the social media
“game.”
Based on interviews with 28 media actors—14 actors from seven U.S. news organizations and 14 actors related to seven Israeli news organizations—I identified five dimensions of the media production logics underlying triple-party news-spaces. These dimensions include item selection, temporality considerations, hierarchy considerations, expression style, and approach toward user-generated content. They emerged from the analysis of the interviews in which I asked various questions about the ways news organizations used social networking sites (see Appendix A). Transcripts of the 28 interviews with social media editors or other key actors who took part in shaping social media strategies were coded in MaxQDA, a software for organizing and analyzing data. I marked with codes
52 sentences in each transcript, identifying different topics and elements. Subsequently, I narrowed down and refined the categories I had created, with the aim of identifying key dimensions of media logics applied by news organizations in the production of messages for social networking sites. The development of categories was based on interplay between patterns emerging from the qualitative data and insights from literature pertaining to news media production (as specified later). In addition to the five key dimensions of media logics identified through this process, the analysis yielded two contextual categories that are important for understanding these dimensions: (a) organizational purposes of using social networking sites; and (b) organizational structures and policies (e.g., who is in charge of posting on social networking sites, and does the organization have a code of conduct).
In what follows, I first address what interviewees said about organizational purposes of using social networking sites, and about organizational structures and policies. I then present and analyze the key dimensions of media logics and their application in different cultural/national and organizational contexts. Finally, I offer a summary of the key findings.
Purposes of Using Social Networking Sites
To understand media logics applied by news organizations in their production of messages for social networking sites, it is important to first understand what news organizations seek to accomplish through these sites. News organizations use social networking sites to reach people. When users of these sites are exposed to a news organization’s piece of content, they may click on it and get to the organization’s site. Users may also stay on the social networking sites and interact with the news organization’s content there—e.g., by sharing, liking, or commenting on it. When they do so, they may increase the visibility of this content, which could be prioritized by the algorithm of a given social networking site (Bromwich & Haag, 2018; Oremus, 2017) and reach more users (some of whom may also click on it or engage in other ways). But users who are exposed to a news
53 organization’s piece of content on social networking sites may also choose not to do anything beyond seeing it there. For most of the news organizations under study, it is a top priority to attract users of social networking sites to their sites. This priority reflects organizational hopes to build brand loyalty and to be appealing to advertisers. However, even if many users stay on social networking sites (rather than visiting news sites), the examined news organizations see their own activity on social networking sites as helpful for raising brand awareness or shaping the organizations’ image.
Interviewees suggested that driving web traffic (web users who visit a site) from social networking sites to the site of a given news organization was a primary organizational purpose of using social networking sites. As Lisa, an audience editor at USA TODAY, put it:
“We need to be using social media because we need referral traffic... We have pretty high social referral goals. So we’re trying to drive people from social media back to our website…
Our website is free, so we’re pretty successful at that.” Patricia, an audience engagement editor at Los Angeles Times, mentioned percentages of traffic: “I know at any given time,
25% to 30% of our total traffic (comes) from social media, although that’s not just from our pages,” she said. “That’s from everybody sharing our stories on Facebook and Twitter.”
According to David, a director at the Israeli newspaper Haaretz, 15% to 22% of its digital platforms’ traffic comes from Facebook. At the Israeli digital native outlet ynet, most of the traffic does not typically come from social networking sites, but traffic from Facebook alone can sometimes reach 50%, according to Dvir, a member of the outlet’s new media team.
Social networking sites may allow news outlets to reach not only people who are already loyal consumers of the outlets, but also other people. According to Daniel, a director at Slate, “a lot of what we’re doing on social platforms is trying to find new readers or infrequent readers and ideally expose them to good stories that they like, and then hopefully convince them to become slightly more frequent readers.” Emily, a senior editor at The New
54
York Times, said that one of the organizational purposes of using social platforms was “to share the best of our journalism with people who may not know The New York Times or come to us directly.”
Itai, a director at the Israeli newspaper Yedioth Ahronoth, saw social networking sites as spaces for attracting younger readers:
It’s no secret that the audience of the print newspaper is getting older. On Facebook,
through (social) networks we manage to reach people the newspaper no longer
reaches. On Facebook, our strongest audience is 24 to 35 years old… mostly women.
It’s an excellent audience from a commercial marketing perspective. On Instagram,
the audience is even younger… So you’ll now ask me: ‘What do you get out of it at the
end of the day? Do more people buy newspapers thanks to Facebook?’ I don’t know…
I do know that it exposes the brand and the newspaper to hundreds of thousands more
people…
According to Linda, an audience engagement producer at The Dallas Morning News, to reach younger readers the organization is active not only on Facebook, where “a lot of our older readers are,” but also on other platforms, such as Instagram and the video-sharing platform TikTok. “The reason why we have all of these accounts,” she said, “is just to meet our audience where they are, meet them halfway because they can’t be doing all of the work.
They can’t be the ones looking for the stories. We kind of have to put them out there for them to find.”
Many of the news organizations under study also value audience interaction with their content on social networking sites—e.g., in the forms of shares, likes, and comments. Ido, a director at the Israeli newspaper Israel Hayom (Israel Today, in English), explained the importance of such interaction while addressing the social aspect of social networking sites:
55
What does it mean social? We hear this word, and it sounds to us that people are
talking with each other... If it’s possible to spark discussion, it’s a wonderful thing
because the algorithm also likes discussion. When I have comments, the post is sent to
more people. More people will see it. It’s also likely that more people will click on the
story if there’s a link to the story… or engage, that is, comment, like, share, anything
that can be clicked-on as part of Facebook. This post becomes more viral. The more
viral it is, the more people will see it. The more people see it, the more likely people
will see my next posts, or like my page, or will become part of my target audience.
Pleasing Facebook’s algorithm and reaching a wide audience on this platform are challenging. According to Uri, a social media editor at the Israeli newspaper Haaretz, when a post is published (e.g., by a news publisher), Facebook first shows it to a sample of users and decides how to further disseminate it based on the reaction of this sample. In an attempt to increase the exposure to certain posts on Facebook, Uri sometimes includes in them a relatively long segment of text. Although it may not drive users to Haaretz website, it could help generate attention and interaction. Tamar, a member of the new media team of the Israeli outlet ynet, identified a tension in this regard between attempts to entice user interaction on
Facebook and attempts to drive web traffic to ynet. The site’s interest is to have high traffic,
“but at the same time, user engagement is important to us,” she said. “It is important to us that our Facebook page will grow and continue to be in the top three or five biggest news pages in the country.” From Tamar’s perspective, this requires a delicate balancing act between providing enough content to engage users on Facebook and leaving users hungry enough to get to the news site.
News organizations also perceive social networking sites as spaces where they can shape their image as important or unique actors in the media landscape. Furthermore, in these spaces they have opportunities to listen to people. According to Emily, the senior editor at
56
The New York Times, “It’s really driving and owning the conversation – like the biggest news stories of the day.” She mentioned “listening to our readers and audience” as another important purpose of using social networking sites. In a similar vein, Karen, a director at
Vox, explained why the digital native outlet is active on social networking sites:
We really see being on socials as an opportunity to be in the conversation with our
audience. Socials have moved very quickly. It’s where stories break, where people are
talking about good things or bad things that are happening in the news… we want to
be in those conversations, and we see our accounts as being valuable in that space…
we have lots of great planners on our website or videos that people can find if they’re
on those sorts of platforms, or if they go to our website, but social media is—you’re
reaching people where they already are and sharing in that way.
Daniel, a director at Slate, talked about the importance of the outlet’s presentation on social platforms. “That is the place where we let our voice shine,” he said. “We know that there are a lot of people whose main exposure to Slate will be how we present ourselves on those social media feeds. And so we think about our voice, and making sure we’re true to who we are, is one of the main goals.”
Interviewees also acknowledged that different social networking sites could serve different purposes. For example, Israeli news organizations see Facebook as particularly important for driving traffic to their sites given the popularity of this platform. They do not expect to get much traffic from Twitter, but they do want to be present among elite actors who are on this platform. U.S. news organizations generally seemed to attach more importance to Twitter than Israeli organizations, though both U.S. and Israeli organizations tended to see Facebook as highly important for reaching people due to the popularity of the platform. The different approach toward different platforms is also reflected in the responsibilities of social media editors, as further addressed in the section below.
57
To conclude, interviewees suggested that their organizations used social networking in an attempt to attract users to the organizations’ sites, to entice interaction with their content on social networking sites, to increase brand awareness and to shape the organizations’ image as important or unique actors, and to listen to users. In what follows, I address structures and policies that are designed to help organizations achieve at least some of their goals.
Organizational Structures and Policies
Media production is shaped by various factors, including the structure and policies of media organizations (Shoemaker & Reese, 1996, 2014)—e.g., what are the roles and division of responsibilities within the organizations, and which principles guide the organizational conduct? Most of the news organizations under study adapted their structure and developed policies for using non-proprietary platforms (Westlund & Ekström, 2018). Acknowledging the importance of these platforms, the examined organizations created accounts on different platforms, assigned people to manage and perform activities there, and set basic guidelines.
Newer job titles, such as an “off-platform editor,” “audience editor for emerging platforms,”
“social media editor,” “audience engagement editor,” and “engagement manager,” are indicative of the organizations’ attention to these platforms.
The New York Times, for example, created an “off-platform team,” which focuses on platforms that are not owned or controlled by the news organization. Team members post content on Facebook, Twitter, Instagram, Reddit, LinkedIn, as well as other platforms.
According to the estimation of Emily, a senior editor, the team included 12 people. When physically present at the Times’ headquarters, team members sit in the newsroom, not far from the “on-platform” people whose work focuses on the Times’ website, newsletters, or push notifications. Emily said that the “off-platform team” was “immensely” involved in the
Times’ everyday activity. “We’re regularly pulled into the (journalistic) investigations because there’s a lot of lead time that we need, especially on really sensitive or complex
58 stories,” she explained. Referring to organizational meetings, Emily added: “We present a daily audience update at our morning news meeting. So that’s a look back on our audience and readership and coverage for the past day and kind of a look ahead to what we’re watching that’s trending on my search (engines) or social (media).”
The New York Times also developed written “social media guidelines for the newsroom” (The New York Times, 2017). The guidelines call for impartiality and fairness in social media posts by workers and address other issues as well. As an illustration, one of the
16 key points reads as follows: “Always treat others with respect on social media. If a reader questions or criticizes your work or social media post, and you would like to respond, be thoughtful. Do not imply that the person hasn’t carefully read your work” (para. 13).
Another major U.S. news organization that adapted its structure for the digital media landscape is USA TODAY. The organization created an “audience team” that seeks to digitally meet audience members where they are—mobile devices or traditional computers, social networking sites or email accounts, for example. According to the estimation of Nancy, a senior manager, her team consisted of about 17 people when she talked with me, including people who focus primarily on social networking sites. “We are very much in the heartbeat of everything that’s happening,” she said. One manifestation of this is the inclusion of team representatives in daily editorial meetings. Lisa, an editor who focuses on social networking sites, shared her experience about the early morning meeting she had attended hours before her meeting with me at USA TODAY headquarters:
I start at seven, immediately go to the seven fifteen meeting… I hear from every
section about what’s coming, and we also have the network (of USA TODAY). So, for
example, today we were talking a lot about whether tornadoes are happening in
different parts of the country… we heard from properties who are covering in Iowa,
Kansas City, Arkansas, Oklahoma... So they give us the information pretty early, and
59
we can start to plan our day around that seven fifteen meeting. There’s obviously
communication happening before that too, as early as overnight. We’re twenty four
hours…
Using USA TODAY’s accounts on social networking sites requires careful training. As
Lisa explained, “no one is ever allowed to touch an account that is USA TODAY branded until they have talked through the process of what’s good for the accounts, how we manage them, what our rules and regulations are. So we hold those keys pretty tight.” In an attempt to maintain a reputation of being a fair and balanced news outlet, USA TODAY also has a written code of conduct for workers using personal accounts on social networking sites.
Although there are commonalities among the news organizations under study, they vary in medium (a newspaper or a digital native outlet), size, resources, and priorities. This is also reflected in the teams that manage and perform organizational activities on social networking sites. For example, U.S. digital native outlets Vox and Slate had smaller teams— about 5–6 people in each organization, according to interviewees’ estimate. These teams also used an automatic tweeting option: When certain types of content appeared on the organizations’ websites, they were also sent out on Twitter. Vox and Slate developed a set of guidelines and best practices for using social networking sites, but the expectations from all workers seem to be more general and less strict compared with The New York Times’ code of conduct. As Daniel, a director at Slate, put it:
We do not have a (written) policy... our guidelines are essentially to increase marks
and know that you’re journalists and representing the organization, but… we trust
you. If you ever have questions or encounter information, come to me or members of
the audience team or come to senior editorial leadership and we’re happy to talk that
through, but generally we trust you and we’re happy to help teach you things…
Vox briefly addresses social networking sites in its “editorial ethics and guidelines”
60 for workers (Vox Staff, 2020, para. 16):
We implore all staff to wield their tool/voice with extra care. Behavior on social
media is a reflection of both one’s personal and professional self and, as a voice of
Vox Media and our editorial networks, we expect social posts to apply the same
journalistic integrity required in our editorial content.
According to interviewees, members of Vox’s “engagement team” and Slate’s
“audience development team” work closely with other people in their respective newsrooms, sharing ideas and coordinating content promotion (e.g., how to promote a podcast on social media).
U.S. newspapers that are not nationally circulated—Los Angeles Times, The Dallas
Morning News, and South Florida Sun-Sentinel in this study—also attach importance to organizational activity on digital platforms they do not own or control. Los Angeles Times has an “audience engagement team,” which considerably increased in size in recent years. Team member Patricia estimated that it consisted of 12 people, including people who post content on social networking sites. The team distinguishes between “core audience editors” who handle Los Angeles Times’ main accounts on digital platforms (such as Facebook, Twitter, and Instagram), and “embedded audience editors” who focus on typically smaller accounts where content from a specific newspaper section is posted (such as Los Angeles Times
Entertainment and Los Angeles Times Sports accounts). Los Angeles Times also developed ethics guidelines for the newsroom that mention social media: “Staff members should avoid public expressions or demonstrations of their political views, whether on bumper stickers, lawn signs, blog posts, social media or online comment” (Edgar, 2014). The Dallas Morning
News has an “audience team” that is in charge of posting content on the website homepage and social platforms. Team member Linda estimated that it included 6–7 people at the time she met with me. “We are central to the newsroom because the mindset… is that once you
61 publish a story, it doesn’t go anywhere until it reaches the audience team,” she said, meaning that a story did not appear on the website homepage or social platforms before her team placed it there. “We are the ones who open the gates and say, ‘Hey, guys. Look at this story.
Come read it.’” At South Florida Sun-Sentinel, newsroom editors and sometimes reporters are assigned to post content on the website homepage and social networking sites at given times, but this is not what they are exclusively designated to do as newsroom workers. In addition, stories published on certain channels of the news website are automatically posted to Twitter through a third-party platform (e.g., the title of a given story and a link to it are copied from the news website when they are published there and are then posted to Twitter).
Compared with U.S. news organizations under study, the examined Israeli organizations had smaller teams to manage and perform activities on non-proprietary platforms. For example, the digital native outlet ynet had a “new media team” that consisted of four members. The number of team members in other organizations did not seem to be higher than that. In most of the organizations, team members are assigned to rotating shifts— usually from early in the morning until late afternoon and from late afternoon until late at night. They typically sit in the newsroom, not far from other website editors, and are aware of stories the newsroom is working on. In two newsrooms I visited—at the headquarters of ynet and Haaretz—I saw the social media editor sitting close to the homepage editor of the news website. This proximity allowed the former to easily communicate with the homepage editor about content promotion on Facebook (e.g., which stories to post and when).
Yair, a manager at ynet, pointed toward the news desk and said: “The morning editorial meetings take place here, next to the homepage editors and the Facebook position.”
This position is where an editor sits and posts content – mainly on Facebook. Referring to the editor who sits there in the morning, Yair added: “He hears more or less what’s planned for the day. Each channel (on the news website) comes and gives us an update… and we simply
62 know how to get ready.” Tamar, who occupied the “Facebook position” while I was there, addressed her proximity to the homepage editor and other editors: “I’m just trying to be connected… I also ask the home page (editor) all the time what is interesting the users now on the home page.”
At Haaretz, interviewees emphasized that the social media editors were senior and experienced editors in the newsroom, suggesting that the position of a social media editor was considered important. As Eitan, a senior editor at Haaretz website, put it:
We attach importance to the fact that the editors who are there are pretty senior
editors… Uri, who is now on duty, was the homepage editor as well as the deputy
editor of the website… Dina, who is now the homepage editor, is also assigned to
work on Facebook in other times. Because we know that overall our reach there (on
Facebook) is very high in Haaretz terms… not everyone does it. Those who post
(content) are very specific people…
Not all Israeli organizations designated newsroom workers to exclusively post content on social networking sites at a given time. At Walla, a digital native outlet, homepage editors were in charge of posting content on Facebook in addition to editing the website homepage.
The organization also had a worker who provided guidance on the use of social networking sites and helped produce some content for these sites. Makor Rishon newspaper outsourced activity on social networking sites to marketing people who did not sit in the newsroom
(though a senior editor had to approve messages before they were posted). This outsourcing may suggest that managers at the newspaper perceived social networking sites as a marketing tool more than an extension of the newspaper. It also seems that the newspaper’s activity on social networking sites was late in the game, and managers felt they needed an external help.
63
Although the Israeli news organizations under study are active on different platforms non-proprietary to the news media, they focus most of their efforts on Facebook, where they typically reach the largest audience (compared with Twitter, Instagram, and other platforms).
Haaretz and Walla automated most of their posting on Twitter: When they publish news items on their website, the items are sent out on Twitter. Ynet also uses this automatic option during the weekend.
Israeli news organizations developed guidelines and best practices for using social networking sites, but they are less likely than U.S. organizations to have formal written policies or codes of conduct. According to Itai, a director at Yedioth Ahronoth newspaper,
There’s no policy dictated from above… It’s important for me to meet certain
standards, also in the wording… writing in good Hebrew and talking at eye level. We
don’t patronize our users. I never fight with them in the comments. I almost don’t
block, almost don’t delete… I’m committed to freedom of expression. Even if things
are written against us, we leave them. In cases of harsh incitement by one user
against another I delete… but not beyond that.
Moshe, a director at Maariv newspaper, said: “We have no written contract…
There’s an unwritten code, a certain ethic, what’s allowed and what’s not allowed, and every employee knows (that).” Other interviewees addressed issues of journalistic integrity and fairness. In addition, Yishai, a director at Israel Hayom, mentioned that the newspaper had “a clear political line, center-right” and was less sensationalist compared with other outlets.
These characteristics may also manifest in the newspaper’s digital activity, he suggested.
The described structures, principles, and purposes of news organizations in the U.S. and Israel provide a helpful background to the next section, in which I identify and explicate key dimensions of the media logics applied by these organizations in their production of
64 messages for social networking sites. I also examine these dimensions comparatively in different cultural/national and organizational contexts.
Key Dimensions of Media Production Logics
Being highly active on social networking sites can be regarded as relatively new for news organizations, particularly organizations that have existed for decades. However, not everything in their activity there is equally new. In producing messages for social networking sites, news organizations apply older and newer media logics. This section addresses the first research question (RQ1), which asks how older and newer media logics are applied by news organizations in producing messages for these sites. I present and analyze what constitutes the media production logics (RQ1a) and how they are applied in different cultural/national
and organizational contexts (RQ1b).
I suggest that five key dimensions constitute these media production logics: item selection, temporality considerations, hierarchy considerations, expression style, and approach toward user-generated content (UGC). Furthermore, I suggest that along three of these dimensions—item selection, temporality considerations, and hierarchy considerations— there are considerable commonalities among news organizations, but along the two other dimensions—expression style, and approach toward UGC—there are stark differences among organizations. U.S. news organizations under study were less likely than Israeli organizations to perceive unattributed evaluative language as appropriate in their messages for social networking sites. In addition, some of the examined news organizations were more likely than others to highlight UGC. I offer possible explanations for these differences.
The proposed analysis focuses on two levels of the Hierarchy of Influences model
(Shoemaker & Reese, 1996, 2014): the level of media organizations and level of social systems. The former level focuses on factors that are tied to specific organizations, such as goals, bureaucratic structure, and policies. The latter level focuses on the larger society as a
65 system and on how ideological, political, economic and cultural contexts shape media performance. This level of analysis is examined here through cross-national comparisons
(Shoemaker & Reese, 2014).
In what follows, I present each key dimension of media production logics and consider it in the cultural/national and organizational contexts under study. Doing so can help understand what in these logics may be the same regardless of the country and organization of production and what is different.
Item selection: “It’s about finding that nugget in a story”
News media production inherently involves content selection (e.g., Domingo et al.,
2008; Galtung & Ruge, 1965; Harcup & O’Neill, 2017; Singer et al., 2011; Tuchman, 1978).
One key dimension of the media logics applied by news organizations in their production of messages for social networking sites is item selection—the choice of units of content for these sites from all the content that appears in their outlet (the newspaper or website, according to this study). News organizations that produce a considerable amount and breadth of content do not typically post each unit of content on social networking sites (or at least they do not post it on all of the sites they use). They select what or about what to post. They may also select on which site to post it.
According to older media logic, when editors perceive something as “breaking news” or a major development, they are likely to post about it on a social networking site because they think it is important and highly newsworthy. Breaking news can be understood as “news that is arriving about events that have just happened” (Oxford Learner's Dictionary, n.d.), such as a shooting, a tornado, or a death of a prominent figure. Major developments that are not necessarily breaking news may include a policy change, an action taken by a political leader, and a corruption investigation, among other things.
66
According to newer media logic or what may be perceived as such, editors try to asses which stories or story components would generate more attention and engagement on social networking sites – and post them. Stories or story components can vary when selected for a news organization’s main account on such a site (rather than an organizational account of a particular section, such as culture or sports). They may address public affairs, such as politics, economics, and international topics; or nonpublic affairs, such as entertainment, sports, crime, and weather (Boczkowski et al., 2011). They can focus on specific people or even animals; take different forms—e.g., a news report or a magazine feature; or be a particular element, such as a photo or a quote.
In editors’ eyes, some units of content that are important or newsworthy can also be attention-grabbing or engaging, though editors acknowledge that the most important or newsworthy content is maybe not the most attention-grabbing or engaging content. In their selection of content for social networking sites, news organizations hybridize older and newer media logics. “Old” and “new” can coexist in the same messages posted on social networking sites, that is, the messages include content that is perceived as both important/newsworthy and engaging. “Old” and “new” can also coexist as separate messages (in the same space), that is, message A reflecting an older logic is posted and afterwards message B reflecting a newer logic is posted. Although different news organizations may prioritize different content for social networking sites, the described combination of “old” and “new” was identified cross-nationally and among divergent organizations.
Interviewees talked about the significance of newsworthiness in the selection of items for social networking sites. For example, Lisa, an audience editor at USA TODAY, said that
“first and foremost, the news matters… More than anything, our job is to inform. So maybe our audience is exhausted by the Mueller investigation, but guess what? The fact that he’s speaking, it’s news. We have to get it out there.” Lisa referred to remarks by special counsel
67
Robert Mueller that were delivered at the Justice Department in Washington, D.C., the same day she met with me at USA TODAY headquarters in McLean, Virginia. For the first time,
Mueller publicly addressed the investigation he led into alleged Russian interference in the
2016 U.S. presidential election, generating wide media coverage. That day (May 29, 2019), the audience team at USA TODAY posted content pertaining to Mueller and the investigation on social networking sites. However, other content was also posted in an attempt to generate audience engagement. As an illustration, the newspaper’s main Facebook page included posts about burgers (“We put plant-based burgers to the (taste) test”), a 97-year-old woman who won prom queen at her first prom (“She didn’t stop smiling the whole night!”), and the expected opening of Star Wars land at Disneyland (“With just a few steps through an arch,
Disneyland visitors are transported to a long time ago and a galaxy, until now, impossibly far away”), among other posts. As reflected in the words of Nancy, a director at USA TODAY, the two following things can coexist: “doing something for complete social play” and publishing “the news that we need to get out whether or not it’s going to be the best thing that we posted on Facebook today.”
The new media team at the Israeli native digital outlet ynet also made an explicit distinction between important news and other content. As team member Dvir put it:
As a team policy, we look at items as roughly divided into two groups. The first group
is the things you can’t delay… urgent news. And there is a second group that is shelf
items… they have more of a magazine style… we know it may be possible to turn them
into a draft (of a post) and keep them for later. For that matter, any news item that is
important and is an emergency by definition… an initial report of a stabbing incident
in Afula… will be posted on all possible platforms…
Dvir referred to an apparent terror incident in a northern Israeli city, in which an 18-year-old woman was stabbed and seriously wounded. Whereas the new media team at ynet thought it
68 was essential to post about this incident and other newsworthy events on social networking sites, the team did not see all items in the “shelf items” group as essential for posting. Team members selected a limited number of items from that group. According to their assessment, such items had the potential to engage audience members. Team member Tamar said that
“bizarre stories will usually do well because they are funny… things we know will generate emotions among users because what interests us is user engagement. It’s reactions, tags, shares.” She added that a story related to a controversial issue or a failure to do something could also provoke engagement.
In an attempt to engage audience members, news publishers may focus on particular topics or people. For example, Linda, an audience engagement producer at The Dallas
Morning News, focused on items related to immigration:
Let’s say an immigration story (is) posted. I would try to get that story up on
Facebook as soon as I could because immigration is a big topic here in Texas. We’re
right at the border. I would put that at any time because usually people are scrolling
through Facebook either at work or at home or anything like that. The reason why I
would choose it is because it would drive engagement.
A combination of issues people care about (whether the issues involve developments that editors perceive as highly newsworthy or not) and a mention of a prominent figure
(which can add to the newsworthiness of a story, e.g., Harcup & O’Neill, 2017) may be particularly attractive to editors and audience members. Patricia, an audience engagement editor at Los Angeles Times, provided the following explanation:
I think of it as Olympic rings of interest, where there are big circles of interest and
then overlap, and the stuff that is in those rings does well, and the stuff that’s in the
overlap does particularly well. Our audience loves news about the environment. They
love news about the Trump administration. They love news about healthcare and
69
things like that. If I see something that’s about the Trump administration is making
new environmental rules that could affect people’s health, I know that's the story
that’s going to do super, super well.
In Israel, interviewees made the point that engaging content—particularly on
Facebook—may be “soft” or sensational. They suggested that the proportion of “hard news” on the main Facebook page of their news organization was lower compared with the website homepage of their organization. “Harder” and “softer” stories may be determined based on what subjects the stories are about (e.g., public affairs or nonpublic affairs) and how they are told (e.g., straight news or feature style) (Boczkowski & Peer, 2011).
Eitan, a senior editor at Haaretz website, pointed out that even “hard news” stories his organization was known for did not appear on Facebook as much as they did on the news website:
Our basic assumption is that news materials work less effectively on Facebook
because those who want news don’t go to Facebook. They go to other places…
Haaretz deals a lot with important news stories, say (the Israeli) occupation (of the
West Bank)… It’s not that we won’t post them (on Facebook), but… if you compare
the content mix to the website, obviously on Facebook there’s much less of that…
Facebook is a much softer place… soft or sensational… Donald Trump is not soft
things, but because he’s sensational, I think our most successful post today was about
Melania…
Eitan referred to a story regarding U.S. president and his wife, first lady Melania Trump.
President Trump revealed details about the medical condition of his wife, saying that she underwent a “big operation” that kept her from travelling abroad with him (Rogers, 2018).
According to Yonatan, a senior editor at the digital native outlet Walla,
70
We know there are things that regularly work on Facebook. They are usually more
sensational… They are more superficial a lot of times. Not always, but a lot of times.
And they get a relatively high prominence compared with the website. Things that
won’t get prominence on the website will get prominence on Facebook. On the other
hand, things we’re doing on the website, which mostly serve a news agenda… won’t
be on Facebook…
In a similar vein, Yishai, a director at Israel Hayom, said that compared with the newspaper and website, Facebook “will focus more on the interesting, less on the important… If I split the newspaper, Facebook is the back pages. It’s not the front pages.”
Ido, another director at the newspaper, further explained which type of items is likely to be selected for Facebook:
It should be something interesting. Interesting doesn’t have to be just gossip. It can be
gossip. It can be something really sensational, such as talking about the Big Brother
(TV show) or Bar Refaeli (a model). But it can also be something concerning the
pocket of every consumer in Israel, such as a big economic change or a sudden gas
price hike… You may now see (on the website) a main headline, a second item, a third
one – and the fourth one… that is put there to have color, I’ll take (to Facebook). To
me, these are the diamonds.
U.S. interviewees were less likely to perceive the website homepage and the main
Facebook page of their news organization as considerably different in the proportion of certain topics or stories. One possible explanation for that is the greater importance that U.S. organizations under study may attach to their reputation or image as reflected on Facebook and other social networking sites. That said, both U.S. and Israeli organizations applied older and newer logics in the selection of content for these sites by choosing to post content deemed important or newsworthy and content deemed engaging.
71
Both U.S. and Israeli interviewees also conveyed the notion that selected items can be story components—elements that exist in stories or are related to them, but are not necessarily prominent on the news outlet. A selected element can become the focus of a post on a social networking site. Put another way, item selection is not only about choosing item A over item B, but can also be choosing certain elements from items A and B. Appealing elements can also be “diamonds,” to use Ido’s word in the quote above (for the Israeli outlet
Israel Hayom). Such an appealing element can also be called a “nugget”—the word used by two U.S. interviewees, one from The New York Times and the other one from USA TODAY.
This is how Emily, a senior editor at The New York Times, explained the logic guiding her “off-platform team” in the selection of items:
One, we’re focused on what are the biggest news stories. And then two, it’s what is a
good social story… So it’s about understanding what makes a good social story, how
our audiences will react to it, and kind of… just finding that nugget in a story, or
explaining in a story to make it really engaging and compelling on social.
For example, such a “nugget” can be a photo, as shown on the left side of Figure 2.
The photo was taken in February 2019 during Donald Trump’s State of the Union address, the annual message delivered by the President of the U.S. to a joint session of Congress. In an unusual way, Trump said that “we must reject the politics of revenge, resistance and retribution and embrace the boundless potential of cooperation, compromise, and the common good” (Fandos et al., 2019, para. 13). Then, Nancy Pelosi, the speaker of the House of Representatives and a member of the rival Democratic Party, clapped at him “the way you might applaud a little kid who wants credit for behavior that warrants none or a dad proudly announcing that he has finally changed a single diaper,” as Wesley Morris wrote at The New
York Times (Morris, 2019, para. 3). Photos of Pelosi applauding Trump were circulated on social networking sites and generated attention and engagement. Although photographer
72
Doug Mills of The New York Times also captured that moment, the photo did not appear on the print front page and the website homepage of the Times the day after State of the Union address, but it was the focus of posts by the Times on social networking sites. The posts included a link to a “side story” regarding this photo and were highly engaging. Figure 2 shows the Times’ post on Facebook (on the left) alongside the newspaper’s front page (on the right) the day after the State of The Union Address.
Figure 2
U.S. President Trump’s State of the Union Address as Reflected in a Facebook Post and the Print Front Page of The New York Times on February 6, 2019, the Day After the Address
Interviewees talked about how an appealing element is found. Lisa, of USA TODAY, said: “We take the story, read it, obviously, before you share it out… and find that like nugget of information that is gold, that you think people are going to share, but also would want to click into the story and learn more.” Lisa suggested that this could be a surprising figure or fact, for example. Eitan, of Haaretz, expected social media editors to “scan” an item properly and get the “interesting stuff” out of it. This, he said, required creativity and approaching the
73 text “from a completely different point of view, from a point where you say not necessarily what’s important in the text but what will make people read it.”
Karen, a director at Vox, provided another example of an appealing element: a quote.
A televised debate among candidates who sought to become the nominee of the Democratic
Party for U.S. president included Senator Kamala Harris of California, who said: “America does not want to witness a food fight, they want to know how we’re going to put food on the table.” Although this quote was not the headline of a Vox report about the Democratic debate, it was the focus of a Vox post about it on Facebook. A link to Vox coverage was included in the post, which appears in Figure 3.
Figure 3
A Facebook Post by Vox About a U.S. Political Debate on June 27, 2019
74
Referring to the attention-grabbing quote, Karen said: “We’re sort of infusing this extra emotion of what a big deal this was.” She added that the selection and prominence of this quote were not meant to convey the message “Go Kamala Harris.” “It's just us saying like,
‘She was not messing around when she said this.’”
News organizations or individual workers developed practices that help in the selection of items for social networking sites. First, those who manage and perform activities on these sites check which items have been published or are expected to be published on the news outlet. They monitor the outlet and get internal digital updates (for example, through
Slack, a professional app for internal communication). Those who sit in the newsroom may also hear what people are working on and communicate with them. Second, those in charge of activities on social networking sites attend regular organizational meetings. In one type of meetings, they hear from representatives of different sections or channels within the organization which stories are expected. Another type may be planning meetings, in which attendees discuss a specific project or preparations for a certain event (For example, a Vox interviewee mentioned a planning meeting ahead of Robert Mueller’s hearing in U.S.
Congress). Third, those who manage and perform activities on social networking sites get recommendations on what to post from representatives of different sections or channels of the news outlet (As an illustration, at USA TODAY content is pitched every two hours during the day through the chat-based collaboration platform Microsoft Teams. At ynet, workers use the instant messaging app Telegram, where a “recommendations group” was created for pitching content. At Haaretz, social media editors get emails with recommendations from other workers). Finally, workers get and monitor data about the digital performance of posted stories. They use tools or software, such as CrowdTangle and Chartbeat, to see what is resonating with audience members. Reports with web traffic data or audience engagement
75 data (e.g., which stories were highly shared) are regularly issued within most of the organizations under study. Managers and editors say they check these reports, try to learn from concrete examples and to identify trends, and may talk about what they learned in meetings.
In the process of evaluating which content is engaging or important and choosing what to post, editors may also choose on which platforms to post it. For example, visually appealing content will possibly be posted on Instagram, and content related to complicated policy issues can be posted on Twitter but not on Facebook that may seem appropriate for content with a broader appeal. Unique content for which news organizations want to be known (a column by a certain writer, an investigative reporting piece, a podcast, a magazine feature, etc.) could be posted on different platforms. Organizations that produce more content or a wider breadth of content than other organizations have more to choose from (e.g., the major organizations The New York Times and USA TODAY versus the smaller organizations
Vox and Slate in the U.S.). In addition, some organizations do not seem as strong as other organizations at digitally reporting the most recent developments (e.g., Yedioth Ahronoth and
Makor Rishon newspapers are not as strong as ynet and Walla digital native outlets in Israel), and this may reflect in the types of content (e.g., magazine features and commentaries) they choose for different platforms.
Interviewees indicated that over time they got a sense of what resonated with audience members, but they also acknowledged that item selection involved trial and error.
Data about the performance of items were available and monitored, though as Lisa of USA
TODAY said: “…there’s just so much information; it’s just really hard to drill down, ‘Am I making the right decision?’”
To briefly summarize the dimension of item selection, news organizations apply older and newer media logics in choosing units of content for social networking sites by assessing
76 what is important and what is engaging or attention-grabbing. Stories that are “breaking news” or stories that satisfy criteria of newsworthiness are likely to be perceived by editors as important. Some of these stories may be seen as engaging or attention-grabbing as well, but other stories (whether they are “important” or not) can also be perceived as engaging or attention-grabbing. Although there are some differences among the organizations under study in the U.S. and Israel, hybridized logics were identified cross-nationally and among different organizations.
Temporality considerations: Posting every 15… 30… 40… 60 minutes
What to post is only one key dimension of media production of messages for social networking sites. Another one focuses on when to post or how frequently to post. Issues related to time are central to news media production (e.g., Tenenboim-Weinblatt & Neiger,
2018; Tuchman, 1978; Usher, 2014). As journalism and the digital media environment evolve, answers to the questions of when or how frequently may also evolve. Drawing on the notion that “journalism is constituted as a primarily temporal enterprise” (Zelizer, 2018, p.
111), I see temporal considerations as a key dimension of the media logics applied by news organizations in their production of messages for social networking sites. Temporal considerations are understood here as considerations related to the time or frequency of posting messages on these sites.
According to older media logic, when news organizations see something as “breaking news” or a major new development, they are likely to post about it as soon as possible.
Reporting events in close proximity to their occurrence is common on TV and on the radio, and it has become a hallmark of online journalism (Tenenboim-Weinblatt & Neiger, 2018).
As Usher (2014) put it, immediacy is one of the “core values” of online journalism. News organizations apply this value on social networking sites as well. That said, if these organizations seek to post content that “can wait” (and does not “have to” be posted
77 immediately), they also take into account other temporal considerations, such as the habits of their audience (e.g., at what time do audience members typically work).
However, according to newer media logic, news organizations also try to publish posts in a frequency that takes into consideration the algorithm of a given platform as they understand it. If posts are published too frequently, they may draw lower audience engagement individually, it is feared. Thus, they may be less visible to other audience members. On the other hand, if content is not posted frequently enough, news organizations’ brand may not be as visible as they want it to be. Therefore, news organizations try to strike a balance between posting frequently enough and not posting too frequently.
When it comes to temporality considerations, news organizations combine older and newer media logics in posting messages on social networking sites. They post messages based on the older media logic, as well as messages based on the newer media logic. “Old” and “new” coexist in the same space. Organizations differ in the number of messages they post per day and in timing certain messages to specific hours, but the combination of “old” and “new” was identified among most of the organizations under study in the U.S. and
Israel.
The general guideline or norm in most of the examined organizations was posting content on the organizational main Facebook page at least once an hour on a weekday, not including late night hours, according to interviewees. Some interviewees even sought to post there content every 30 or 40 minutes (during the day) if possible. On Twitter, the posting could be more frequent—e.g., every 15 minutes. The posting frequency on these platforms and other ones was not set in stone and could change based on the news the organizations produced. As encapsulated in the following words, “If it’s breaking news, it gets up immediately” (Lisa, USA TODAY), or “If there’s breaking news, we will publish immediately” (Linda, The Dallas Morning News).
78
Dvir, of the Israeli outlet ynet, addressed the importance of frequent posting on social networking sites and described what his new media team was doing:
I treat it like yeast. If you don’t give it the proper treatment, it will just die at some
point. You must keep a site alive… (We post) at least once an hour on all platforms.
For that matter, on Instagram we’re less precise about it. We’ll settle for three posts a
day or a shift. In an eight-hour shift at least eight Facebook posts, at least eight
tweets on Twitter, if not much more than that… It can be three tweets an hour, which
is a completely standard day.
Patricia, of the Los Angeles Times, described a more frequent posting routine of her audience engagement team:
…what we generally do is we try to post two posts on Facebook every hour between 6
a.m. and 9 or 10 p.m., depending on how much news there is… and four tweets every
hour during those active times. And then on the overnight, 10 p.m. to 6 a.m. when it’s
all a little slower, it’s typically one Facebook post and two tweets per hour.
To get a sense of how frequently different news organizations posted content on a major social networking site, I examined the average number of posts per day published on
Facebook by each of the organizations under study. Data about the posts published on the main Facebook page of each news organization in the second half of 2018 (from July 1 to
December 31) and the first half of 2019 (from January 1 to June 30) were separately obtained (for each period) through the Facebook-owned social media monitoring tool
CrowdTangle (CrowdTangle Team, 2020). Based on these data, the average number of posts per day was calculated. The results for each organization are presented in Table 1 (U.S. organizations) and Table 2 (Israeli organizations), showing that the average number of posts published by an organization in one time period was similar to the average number in the other time period.
79
Table 1
Number of Posts on Facebook Pages of U.S. News Organizations
Average Number of Posts per Day Jan 1,2019– Jul 1, 2018– Jun 30, 2019 Dec 31, 2018 The New York Times 60 60 USA TODAY 42 44 Vox 32 35 Slate 35 37 The Dallas Morning News 26 27 Los Angeles Times 34 28 South Florida Sun Sentinel 43 44
Table 2
Number of Posts on Facebook Pages of Israeli News Organizations
Average Number of Posts per Day Jan 1,2019– Jul 1, 2018– Jun 30, 2019 Dec 31, 2018 Haaretz 20 21 Yedioth Ahronoth 15 14 ynet 18 19 Walla 15 15 Israel Hayom 17 16 Maariv 25 27 Makor Rishon 4 3
The results also show that U.S. organizations published more posts than Israeli organizations. Whereas the number of posts on the Facebook page of a U.S. organization ranged from 28 to 60, the number of posts on the Facebook page of an Israeli organization ranged from 3 to 27 (though only one organization had a single-digit number of posts per day). Possible explanations for this difference are the amount of content produced by the organizations, the resources put into their activity on social networking sites, and the perceived conventions in a given market.
That said, interviewees from different organizations—both in the U.S. and Israel— indicated they were taking into consideration Facebook algorithm (as they understood it) in posting content there. Betty, a member of the audience development team at Slate, explained
80 the rationale of posting frequency that can manifest in a different number of posts for different organizations:
The logic is that we want to maintain a steady presence. So we want to have some
kind of regular intervals… if they’re super close together, the Facebook algorithm…
will kind of create too much noise and nothing will rise to the top. So you want to
space them out enough that each one can get enough engagement on its own…
Mark, a digital producer at The Dallas Morning News, predicted that the frequency of posting on Facebook would be somewhat different in different organizations, but he also thought that the organizations had something in common:
I imagine everyone you are going to talk to is going to give you a different answer
and everyone is going to have a different reason for it but, ultimately, it’s just they are
trying to appease this algorithm… it’s this weird balance of where you are trying to
appease this thing that you have no control over, but you also don’t know what it is
wanting from you.
Indeed, interviewees from other organizations said they experimented with
Facebook’s algorithm and had changed the volume of posting there over time. They did not have a “magic formula” because they were not sure how exactly the algorithm operated. They did reach a formula that worked for them, but they also acknowledged it was not necessarily stable due to possible changes in how the algorithm operated.
Michael, a senior editor at South Florida Sun-Sentinel, shared what his organization was doing:
I think people do a lot of guessing, and Facebook’s algorithms change. So we’ll
experiment and we’ll purposely go up on volume. There was one point where we were
up close to 70 (posts per day) and felt like we were not seeing any change in reach. In
fact, some of those posts were being throttled back. And also that means a waste of
81
time because we’re spending resources to write and post those. So we throttled that
back…
Similar experiences were shared in Israel. According to Yonatan, a senior editor at
Walla, his organization did not want to oversaturate audience members on Facebook, but also figured it should have “some double-digit number” of posts per day. “We did not find the formula. It’s impossible to crack it, really,” he said. “We tried less, we tried more… we settled on some number. I don’t know to tell you that it’s optimal, because… Facebook doesn’t give you the real insights on what works.” At Haaretz, the pace of posting had accelerated over the years (before the periods in the table), as reported by Uri, a social media editor. “At first, we used to post once every two hours,” he said. “Then we posted once an hour… today between half an hour and fifty minutes, approximately. If something came out fifty minutes ago, I know I already have to put out a new post.” Uri added that after posting a newsworthy item that did not generate a lot of interest, he would sometimes post a different item shortly afterwards (e.g., ten minutes later). On the flip side, Moshe, a director at Maariv, indicated that an engaging post—particularly a post that generated comments—could get more time than usual to entice engagement. “You let Facebook run with it… give it an hour, as long as it takes,” he said.
At yent, although the general guideline was to post content at least once an hour, the new media team was also encouraged to experiment with the frequency of posting. As Yair, a manager at the news outlet, explained:
With all the changes we have on Facebook right now, we can’t afford to remain with
the same number of posts… Trial and error has always been the way on social media,
today more than ever… So I encourage the team to do experiments. For that matter,
one day say, ‘I’ll upload a post now, and in seven more minutes I’ll upload another
82
one.’ I’ll see if they affect each other. Or on another day – to choose only our
strongest items, even if a post comes up once every two hours.
However, the perception at Makor Rishon was that posting on Facebook should not be that frequent. Itamar, one of those who ran the Facebook page of the newspaper, said:
If you saw a person posting uninteresting things every 40 minutes, what would you
think about him? That he’s annoying and has no life. I hide his posts. He no longer
interests me. He floods my (Facebook) feed. It’s the same with a media institution. It
should be interesting and sharp and shouldn’t waste my time.
Temporality considerations include not only the frequency of posting content, but also the time of posting. News organizations timed certain content to specific hours or days based on what they perceived as habits or preferences of audience members, as well as on organizational possibilities or constraints (e.g., when a certain piece of content was available). For example, The Dallas Morning News typically posted a business item on
Facebook at 5 a.m., the first morning slot, assuming that “a lot of business people wake up incredibly early” and may scroll through Facebook, to use Linda’s words. Early morning business stories “usually perform very well,” she said. Other routines at the newspaper were posting a food story at noon “because that’s when people are hungry” and another food story or entertainment story at 7:30 p.m. The midnight slot was marked as “the big thing that people need to know before they go to bed.” As an illustration, this could be a story related to
Dallas City Council that is produced late in the day.
In a similar vein, when Haaretz and Walla had new content related to food, they were likely to post it on Facebook around noon, according to interviewees. Such content can be a report about a restaurant opening, a restaurant review, a recipe for something, a story about a new food product, or another story. As Ariel, of Walla, put it, timely food-related content is
83 intended to stimulate people’s taste buds when they may be thinking or talking about food and to offer them useful options.
Another useful content may be weather reports. At Haaretz, Walla, and South Florida
Sun-Sentinel, noteworthy weather reports were likely to be posted on Facebook early in the morning before people may leave for work. For example, editors perceive weather reports as worth posting when it is unusually cold or hot, or if it is going to rain for a week.
Social media editors may choose to post other types of content in the morning or evening. Dvir, of ynet, said that based on his experience, people liked to start their morning with “good news” if possible. Therefore, when a “positive” and typically a soft story was available on ynet website—e.g., content about a newly born giraffe, a headstand of the British queen’s great-granddaughter, or air force pilots who got married—Dvir tended to post it in the morning unless he needed to post more newsworthy content. However, when content contained sexual references, Dvir tended to post it late in the evening—possibly when children were less likely to consume it. In a similar vein, Ido, of Israel Hayom, indicated that he was less likely to post content on a scary movie in the morning. Uri, of Haaretz, also thought that morning hours were not ideal for certain types of content. “You won’t drop on someone a 4,000-wordtext about the Holocaust at 8:15 a.m. with the morning coffee,” he said. Uri added that he was always trying to think about his audience—what people may be doing and how much attention they could pay to certain content at a given time.
News organizations may also time content to specific hours in an attempt to reach more people. For example, Betty, of Slate, assumed that at late hours (e.g., from 7 to 10 p.m.) content produced by her organization could perform relatively well on Facebook. Thus, the posting of certain pieces of content was timed to these hours when it was possible. At ynet,
Dvir took advantage of a commercial break during the evening newscasts of the main TV channels in Israel to post content on the Facebook page of his organization. “I have a grace
84 time of 3.5 minutes where Facebook’s algorithm may identify traffic,” he explained, suggesting that stories could perform relatively well when they are posted during this time.
This example illustrates that in addition to taking into consideration what audience members may be doing, news organizations are affected by behaviors of other media outlets. This is also one of the reasons for which it is particularly important to certain organizations to post breaking news earlier than competing organizations if possible, so that they would reach people first.
News organizations also time content to specific days. As an illustration, posting magazine features is particularly common during weekends, when the news cycle may be slower and people may be able to devote more time to reading. In addition, weekly columns or commentaries are often posted the day or the day after they appear on the organization’s website.
Timing content to specific times could be challenging. First, it is not clear when the user of a given social networking site would actually get to see the content. Second, some of the organizations under study seek to reach people in different time zones. As Emily, of The
New York Times, pointed out, “about 30% of our audience is international, so while the majority of our audience is in the U.S. time zone, there are stories that are targeted to people around the world that we will target to those times… peak morning times in those places.”
To briefly summarize the dimension of temporality, news organizations apply older and newer media logics in determining the time and frequency of posting messages on social networking sites. They seek to post about events in close proximity to their occurrence, but also attempt to post content at frequency or time that may help reach more users of social networking sites. Although the organizations under study post content at different frequency or time, the combination of “old” and “new” was identified among different organizations in the U.S. and Israel.
85
In the production of messages for social networking sites, news organizations consider not only what (dimension 1) and when (dimension 2) to post, but also how to post.
Considerations related to how are the focus of the next two dimensions.
Hierarchy considerations: “There are ways to show importance”
News media production involves considerations of how prominently to display content, or how to signal that certain content is deemed more important or worthier of attention than other content. I call these considerations hierarchy considerations and suggest they constitute a dimension of the media logics applied by news organizations in their production of messages for social networking sites.
News outlets or newscasts typically have a hierarchical structure—a structure in which items are organized into different levels of importance or prominence. As an illustration, a newspaper or news website usually displays a main headline and other items on its front page. Additional items appear on other pages. According to older media logics, editors decide which items of those selected for publication to display more prominently.
Item A may be more prominently displayed than item B, which may be more visible than item C, and so forth. The different prominence of different items signals to audience members which items are perceived by the news organization as more important or as deserving more attention than other items at a given time. Offering a content hierarchy may be particularly valuable in a media environment where content far exceeds what people can consume (see Chyi, 2009, for example).
However, according to newer media logics, when news organizations post content on certain platforms, such as social networking sites, they do not necessarily determine that item
A would be more prominent than item B and that the latter would be more visible than item C and so forth. They may post one item at a time, and the algorithm and users of a given platform may determine what would get more prominence.
86
Nevertheless, news organizations hybridize older and newer media logics on social networking sites by singling that certain content they post on these sites is more important or deserves more attention (in their view) than other content they post there. This singling can be done in different ways, including explicitly labeling a unit of content as “breaking news,” posting a series of connected messages (a Twitter thread), and posting on the same issue or even posting the same content more than once. Although news organizations under study experimented with different ways of singling out content, a combination of “old” and “new” was identified among different organizations in the U.S. and Israel.
Interviewees made the point that prominence of content on social networking sites was not the same as prominence of content on their news outlets. As Emily, of The New York
Times, put it:
It’s hard to say what prominent means on social, right? Because social is about your
feed and when you might see it; it’s not necessarily about me saying the system’s most
important thing. There are ways for us to show that importance by perhaps doing a
Twitter thread or something like an Instagram story versus a feed post, but outside of
that… one of the challenges and opportunities of social is that things are flattened…
that’s why it’s about discovering questions that your community, your friends and
family (are interested in), whereas the homepage is serving a much different purpose
which is to say like, “These are the things we think are most important on work.”
Indeed, managing a homepage of a news website is different from managing a news page or account on a social networking site. First, whereas multiple stories can be placed on a homepage at a given time, a single story at a time is typically posted on a social networking site. Second, whereas an editorial hierarchy of stories (e.g., a lead story, second story, third story, etc.) is presented on a home page, the editorial status of stories is not necessarily presented on a social networking site.
87
Focusing on the first point, Dvir, of ynet, talked about homepage editors versus social media editors:
The advantage of the home page is that it can use so much content at any given time.
That is, on the hour they (editors) can upload seven items to the homepage, and all of
them will gain some prominence. For us, it doesn’t exist… the essential difference is
that I have to pick one item at a given moment. This means that the bet would better
be successful. It should be an educated bet.
According to Uri, of Haaretz, the editorial hierarchy of stories is what distinguishes a news website homepage from a Facebook page. “On the homepage, it’s more a matter of status,” he said. “Does this thing deserve to be chief? On Facebook there’s no hierarchy… So you’re not saying to yourself: ‘Does it deserve to be a post?’ Everything deserves to be a post.”
However, news organizations can signal that not all of their posts have the same status. One way to do so is to label certain posts as “breaking news.” For example, such labeling can be done by including in a post an image with the words “Breaking News,” or starting the text of the post either with these words or similar ones. Figure 4 shows (on the left) a Facebook post by ynet with a prominent red image that contains the words “Breaking
News” in Hebrew. The post reports that sirens went off in southern Israeli cities during a visit in one of them by the Prime Minister of Israel Benjamin Netanyahu (December 25, 2019).
The sirens indicted that rockets were launched form the Palestinian Gaza Strip toward these
Israeli cities. Figure 4 also shows (on the right) a Facebook post by USA TODAY that starts with the capitalized word “BREAKING.” The post reports that U.S. President Trump canceled a meeting with Russian leader Vladimir Putin after Trump’s former lawyer had pleaded guilty to making false statements to U.S. Congress (November 29, 2018).
88
Figure 4
The Words “Breaking News” or “BREAKING” in Facebook Posts by ynet (On the Left) and USA TODAY (On the Right)
In addition, Facebook allows certain news organizations to tag a limited number of items as “breaking news.” According to Facebook, the company ran a test that allowed more than 100 news publishers from North America, Latin America, Europe, India, and Australia, to use a Facebook label to highlight urgent news stories (Rhyu, 2020a). The selected publishers could use the label once a day and to set how long the story was marked as breaking for (up to six hours). The publishers also had an additional pool of five labels per month they could use (Rhyu, 2020b).
Michael, of South Florida Sun-Sentinel, said that his organization was using this
Facebook feature. He thought it was a valuable tool:
That’s a significant tool for us because those posts get a lot more reach than a
regular organic post without that tagging on it. Sometimes many times more than it
would without the tagging… We have guidelines that Facebook provides for us for
how we can use it. So it has to be a local breaking news story. We couldn’t put a
breaking news tag on a post about Trump impeachment hearings, for example.
89
A different way to signal that certain issues or events are deemed particularly important is to have multiple posts about them. For example, on April 9, 2019—the day Israel held a general election—the vast majority of the Facebook posts by the Israeli news organizations under study were related to the election. Based on Facebook data obtained through CrowdTangle, I identified about 68.4% to 100% of the posts by a given organization on this day as related to the election.
News organizations can also have more than one post with the same piece of content published at different times. According to Betty, of Slate, her organization published more than once the same piece of content (e.g., a story, or a commentary) on Twitter if the piece got at least 100 clicks in the first round or was considered “very valuable or newsworthy.” By doing so, the organization sought to reach more users and to draw more attention to its important or interesting pieces of content. In a similar vein, Patricia, of the Los Angeles
Times, said that her audience engagement team developed a relatively new guideline: “if a story does well, tweet it again. If a story is number one on Chartbeat, make sure it gets three to four tweets during the day and another couple on the overnight. If stuff does well, keep pushing it.” However, this practice was not pursued on Facebook. “Even if a story is doing great, we only post it once on Facebook during the day because we know Facebook dings you for posting the same links within the same 72 hours,” Patricia explained. “If a story does great, we can’t post it three times to Facebook that day and assume it will do well. Each time it will do progressively worse,” she added.
Indeed, in posting content and highlighting it news organizations need to consider the possibilities and constraints of different platforms. On Twitter—unlike Facebook— organizations can also post a thread of messages, that is, a series of connected tweets. The
New York Times (as mentioned earlier) and other organizations post such threads on occasions deemed significant. For example, after Special Counsel Robert Mueller had
90 submitted his report on President Trump and Russia to the U.S. attorney general (March 22,
2019), the Times posted on Twitter a thread of 10 messages with different links. After two shooting incidents in Texas and Ohio, the Times posted a thread of eight tweets that started with the words: “In less than 24 hours, 2 mass shootings in the U.S. have left at least 29 people dead. The killings are the latest in a particularly brutal week for gun violence in
America” (August 4, 2019). A different way to signal importance, which can be used on different platforms, is to broadcast live video or include a link to such a video. Walla and yent, for instance, seemed to be using this way on Facebook and Twitter at least once on a weekday.
To briefly summarize the dimension of hierarchy considerations, news organizations apply older and newer media logics in displaying content on social networking sites. To some extent, they offer an editorial hierarchy of content, though this hierarchy is typically more flattened compared with the one in a news outlet. Although different news organizations use different ways to signal certain content as more important or noteworthy, the combination of
“old” and “new” was identified across organizations in the U.S. and Israel.
Expression style: It’s OK to be “a little bit more playful”
How to post content on social networking sites manifests not only in hierarchy considerations, but also in the style of expression—a dimension along which considerable differences were found based on cultural/national and organizational contexts
News media production involves decisions on how to word or convey messages— decisions about the expression style used by news organizations. These organizations may use formal or informal speech, as well as varying degrees of subjective expression.
Expression style—the focus of this section—is a key dimension of media production logics for social networking sites.
91
According to older logics of mainstream media in the U.S. and other countries (not all of them), journalists use or should use minimal subjective expression in news reporting
(Schudson, 2001). Journalists developed a set of procedures, such as source attribution and using direct quotation, to come across as “objective.” They believed that conforming to these procedures would help protect themselves from criticism or attack (Gans, 1979; Tuchman,
1972). In addition, as part of older media logics, messages are transmitted to audience members and are passively consumed by them. Therefore, the style of expression used to construct messages may not be geared toward audience interaction with them.
However, according to newer media logics, social networking sites and other platforms non-proprietary to the news media afford considerable latitude for self-expression, at least in cultural/national contexts where public expression is not severely restricted. On such platforms, the lines of separation between journalists and other users are more blurred and journalists are not subject to the same layers of editing in their home outlets. Thus, journalists may be more inclined to use explicit subjective expression there (e.g., Lasorsa et al., 2012; Lawrence et al., 2014). In addition, as part of newer media logics, messages are distributed by users within networks and are geared toward user interaction. The expression style used to construct messages examined here supports that.
News organizations blend older and newer media logics in their expression style on social networking sites. They can try to avoid the appearance of taking sides when presenting news and use a style geared toward audience interaction with the content—e.g., a more informal, conversational, or playful when deemed appropriate. But news organizations can also combine a higher degree of subjective expression and an engaging style. Different combinations of “old” and “new” were identified in this study based on cultural/national and organizational contexts. In general, both U.S. and Israeli organizations tried to develop a style of expression geared toward audience interaction with the content. However, the perceptions
92 of interviewees from mainstream media in the U.S. reflected more caution about using unattributed subjective expression on social networking sites compared with interviewees from Israel. The former thought they should avoid or minimize such expression in the presentation of news, particularly hard news. They did post opinion pieces, commentaries, and reviews (in addition to news), but labeled them as such. In Israel, organizations were generally more willing to express opinions—at least on certain issues. This may reflect the
Israeli cultural characteristic of more direct talk (Katriel, 1986), as well as the tendency of
Israeli journalists to be more interventionist in their role perception, that is, they are more likely than their U.S. colleagues to think it is one of their roles to promote certain agendas
(Tenenboim-Weinblatt, 2014). That said, Israeli news organizations also varied. Mass-market outlets, which sought to attract a wide audience with diverse views, were more cautious about opinion expression.
News organizations under study—particularly in the U.S.—were concerned that unattributed subjective expression on controversial issues might undercut their journalistic reputation. A stark illustration of this concern is The New York Times’ “social media guidelines for the newsroom” that stipulate: “In social media posts, our journalists must not express partisan opinions, promote political views, endorse candidates, make offensive comments or do anything else that undercuts The Times’s journalistic reputation” (The New
York Times, 2017, para. 7). The guidelines go on to say that “our journalists should be especially mindful of appearing to take sides on issues that The Times is seeking to cover objectively” (para. 8) and that “these guidelines apply to everyone in every department of the newsroom, including those not involved in coverage of government and politics” (para 9). As demonstrated earlier, other codes of conduct also addressed—one way or another—the importance of journalistic integrity on social networking sites.
93
However, as interviewees indicated, avoiding an appearance of taking sides could coexist with accessible style of expression suitable to social networking sites. For example,
Emily, of The New York Times, said that “journalism is the thing that’s completely across everything we do… and then it’s from there that you adapt the stylistic things, the voice, but it will always have that highest kind of journalism integrity.” Emily further talked about the importance of adapting the expression style of the Times to social networking sites:
It’s about accessibility… one of the most important things about these platforms is
that these are people who may not come directly to the Times, so what are the
shortcuts that we can take? What do we need to, perhaps explain more? What can we
do to sound much more acceptable and like native to those platforms? You don’t want
to be on a platform and sound like you’re just unaware of how people talk there. It’s
like going into a party and being… not dressed for the occasion and speaking another
language.
An expression style suitable to social networking sites may be more informal or conversational. Karen, of Vox, explained:
Our voice is definitely conversational and friendly without talking down to people.
The guidance that we give ourselves for the social team is to think about how you
might tell, explain something to your friend at a bar or to your dad on the phone...
We want you to sound like I just learned this really interesting thing, ‘Did you know
that like X percent of X helps Y?’ And like that would be a good share copy because
that’s exactly how I would tell my friend about a story… There are times where you
could be a little more emotional or add a little bit more energy into a post, but in
general we’re pretty even keeled and unbiased… We are all trained to know there are
certain writers that have a little more hot takes, or there are certain topics that are
94
more intense, so in those cases we don’t lean away from that emotion, but we credit it
and couch it in somebody saying, ‘This is somebody else’s thoughts.’
According to Michael, of South Florida Sun-Sentinel, the wording of posts on social networking sites could be less restrained as long as journalistic standards are met:
This is not something that’s written as policy but just something that we discuss
frequently is that we can be a little bit looser with our language on social media as
long as we’re following the same journalistic guidelines that we follow for the
website. So if it’s not something we would allow on the website, we would not allow it
on social media. Now that’s not to say that we wouldn’t be more playful or use a
headline on social that doesn’t work so much on the website, but it does have to
follow the same principles and guidelines of our website.
The word playful or similar words came up in interviews with actors from different
U.S. organizations, describing the expression style used by their organizations on social networking sites when deemed appropriate. Mary, a member of the audience engagement team at Los Angeles Times, talked about taking a “playful approach” that could manifest in a funny or “kind of cheeky” headline “with the right sort of story.” Daniel, of Slate, said: “We have some humor. We try not to be flippant, and we try not to be like reactionary, but we’re at times trying to be witty… we could take a serious tone at the very serious (news), but we like to have a lighter touch in a lot of ways.” According to Lisa, of USA TODAY, being playful could manifest in the use of humor, emojis, graphics, multimedia, adjectives, and stronger verbs. For example, her audience team may be “amping up the language a little bit” by using the word “stifling” instead of saying “it’s hot.” She explained that “obviously it needs to match what is being said in the story, but sometimes we do this a lot with our language.” As another illustration, Lisa thought that a message such as “Robert Mueller holds press conference after investigation” was not engaging enough. “We want to make sure that
95 we’re packing that with some adjectives, some good strong verbs, making people want to know what he’s saying,” she said.
Nancy, of USA TODAY, further explained the rationale guiding the audience team, acknowledging the possible tension between older and newer norms and practices:
Obviously we want to be conversational and engaging, but free from personal bias or
opinion… (We are) thinking through when it makes sense to have a little bit more
voice. This comes up quite a bit... It’s a gray zone for sure, but I could see an example
of someone writing social chatter to go along with the story that says something like
“we see you”… something that’s just a little bit more playful.
Nancy suggested that this was more likely to happen with softer content, such as stories from the entertainment section of the newspaper.
News organizations’ expression style may be tailored not only to stories, but also to platforms. According to Lisa, of USA TODAY, “there’s certain lingo that people use on
Twitter versus how posts are displayed on Instagram or Facebook… Twitter is the one where we like to have a little bit more fun.” In a similar vein, interviewees from The Dallas
Morning News indicated that they were using more humor on Twitter than on Facebook.
Mark suggested that although audience members on Facebook could appreciate a joke, “you don’t have to put your funniest stuff on there. And sometimes your funny stuff will go over people’s heads on Facebook, so… most of the time I play it straighter on Facebook.”
According to Linda, The Dallas Morning News’ audience on Facebook was generally older and more serious compared with other platforms. Therefore, a playful expression style did not always work there, but it could engage audience members on Twitter.
Yet, interviewees from different organizations indicated that a practice they performed regardless of platform was clearly marking opinion pieces or editorials as such.
For example, Jennifer of Los Angeles Times, shared that her audience engagement team had
96 an entire section in its guidelines about how to label pieces of content that were not news reports. “We try to make a very big point of that,” she said, suggesting it was important to the newspaper to come across as impartial.
In Israel, interviewees thought it was important to indicate who wrote an opinion piece when posting a link to it, but not all of the organizations under study shied away from using evaluative language when presenting news, including news related to public affairs.
Thinking about subjective expression as a continuum, one end is minimal subjectivity and the other end is maximal subjectivity. In-between, there are various possibilities. As an illustration, The New York Times sought to use minimal subjective expression in the presentation of news on social networking sites. However, interviewees from Haaretz showed inclination toward a considerably higher degree of subjective expression—possibly the highest degree among the news outlets under study. In the Israeli context, Haaretz is considered an important and highbrow newspaper (Tenenboim-Weinblatt, Hanitzsch, &
Nagar, 2016). In that sense it bears a resemblance to the Times. I would even argue that among Israeli newspapers, Haaretz is the most equivalent to the Times. Yet, the difference between these two newspapers when it comes to expression on social networking sites is striking.
According to interviewees at Haaretz, it was acceptable for journalists to express an opinion as part of a post on Facebook or another platform. As posts by other news organizations, a post by Haaretz often included a link to a unit of content on the news website and an accompanying text. The social media editor who constructed the post at Haaretz was allowed to include a subjective expression in the accompanying text. Including such an expression was seen as a possible way to draw attention and to engage audience members.
Although the subjective expression of social media editors at Haaretz did not have to be minimal, editors did not get a free hand. Senior editor Eitan explained that expressing an
97 opinion was acceptable when it was in line with liberal values reflected in the newspaper’s editorials:
It’s allowed to express an opinion. But this opinion should be on issues toward which
the newspaper’s approach is very clear… there are things over which a black flag
hovers and are so clear. There’s no doubt, for that matter, what the position is…
surely with the values this newspaper represents… For that matter, some kind of
misogyny or something about Betzalel Smotrich (a conservative politician) and his
views about gays… There’s no problem (expressing an opinion against them).
Social media editors also did not shy away from criticizing the Israeli Prime Minister
Benjamin Netanyahu when presenting news related to him. For example, after a general election in Israel, Netanyahu accused his rival, the leader of Kahol Lavan party Benny Gantz, of “conducting negotiations with members of Knesset who support terror organizations and want to destroy the country” (Lis, 2019, para. 2). Netanyahu referred to members of the Arab
Joint List party and suggested that Gantz was trying to form a government based on their support (and the support of other parties). On Facebook, Haaretz posted a link to a news report about what Netanyahu had said – along with this text: “While this is a far-reaching statement, it may have been Netanyahu’s most inciting speech against Arabs ever.” Readers were invited to judge for themselves, though the following words were added in parentheses:
“Just keep in mind that you may feel the need to shower later.” Figure 5 shows another example. A Facebook post by Haaretz (on March 18, 2019) included a link to a news report related to Netanyahu’s investment in shares in a Texas-based steel factory that was managed by his cousin. When making this investment, he was not the prime minister of Israel.
According to the report, Netanyahu changed his version of events: Earlier he said the investment had been made when he was a private citizen, but later he claimed it had been made in 2007. However, in 2007 Netanyahu was the opposition leader. The social media
98 editor who posted the link to this report added a critical or sarcastic text loosely translated as
“Of all the people who might do a complete 180 from one moment to the next – (would you believe) Netanyahu would?”
Figure 5
Subjective Expression in a Facebook Post by Haaretz
Interviewees from Haaretz indicated that the newspaper’s expression style on
Facebook was different from the style used in the newspaper and on the news website. David used the words “young” and “kicking” to describe the style of expression on Facebook. He said the newspaper was testing limits there. According to Eitan, Haaretz did not want to come across only as elitist, but also as “very connected to what’s happening” and “cool and amusing sometimes.” He also said that humor was encouraged on Facebook “as long as it is in good taste.”
Interviewees from other Israeli news organizations also indicated that their organizations’ expression style on social networking sites was more informal or conversational compared with their news outlets. Humor could also be used when deemed appropriate. Yet, organizations differed in their willingness to opine on social networking
99 sites. Walla and Makor Rishon seemed more inclined to opine compared with Yedioth
Ahronoth and Maariv that sought to attract audience members with different views. Israel
Hayom and ynet seemed to be somewhere in-between, though interviewees also indicated that these organizations were cautious about more controversial issues.
For example, interviews at ynet suggested that members of the new media team were more likely to express views on issues related to LGTBQ than issues related to elections or the Israeli-Palestinian conflict. Team member Tamar recalled that she had once published on the organizational Facebook page a post about an incident involving Jews and Palestinians in the West Bank. Tamar presented the story in a way she perceived as balanced, but users severely criticized ynet for the wording. As a result of that, she became even more cautious when addressing such issues and tried to stick to the reporting language that appeared in the news stories. But she did not shy away from favorably addressing LGTBQ issues even if there was no consensus about them in the Israeli society. As Yair and Dvir indicated, the team had a pro-LGTBQ agenda.
Makor Rishon and Israel Hayom, two newspapers that can be considered right- leaning, had somewhat different approaches toward expression on their Facebook pages.
Itamar, one of those who ran the Facebook page of Makor Rishon, said: “We have no problem expressing an opinion… if the opinion is clear and (reflects) a consensus where we come from, which is right… Zionist right… We have a lot of opinion columns, so we express
(these opinions).” Ido, of Israel Hayom, talked against mixing news with views and indicted that the social media team was cautious about its expression on political and controversial issues. But the team also sought to have a certain vibe coming out of the newspaper’s page on
Facebook: “We love Israel,” as Ido put it.
Maariv and Yedioth Ahronoth seemed even more cautious in their expression on
Facebook. For example, Moshe, of Maariv, said that his newspaper was not taking a stance.
100
“We give the facts as they are. (We) give everyone a say,” he asserted. Itay, of Yedioth
Ahronoth, wanted his newspaper to have a broad appeal on Facebook and said he was attentive to “the voice of the public.”
To briefly summarize the dimension of expression style, news organizations apply older and newer media logics in wording or conveying messages on social networking sites.
“Old” and “new” can coexist in the same messages or as separate messages in the same space. News organizations try to use a style accessible to users of these sites—e.g., a more informal or conversational style—while using either a minimal degree or a higher degree of subjective expression. The perception of which degree of subjectivity was acceptable varied based on cultural/national and organizational contexts, suggesting that professional journalism on social networking sites has more than one language.
Approach toward UGC: “Crowdsourcing stories as a way to build community”
What, when, and how to post content on these sites are all important questions, but another question worthy of consideration is who produces the content or gets attention.
Although the dimensions of media production have so far addressed mainly content produced by news organizations, these organizations can also use or interact with content produced by users. User-generated content (UGC) is the focus of the last dimension.
In the digital media environment, content is produced and made available not only by news workers, but also by other users. News organizations’ approach toward UGC is another dimension of media production logics for social networking sites. Approach toward UGC refers to news organizations’ inclination or lack thereof to solicit, highlight, or interact with
UGC.
According to older media logics, news workers produce content, and audience members consume it. Audience members could also express views, but spaces to do so in the
101 media environment were limited before the Web 2.0 era (e.g., sections of “letters to the editors” in newspapers, and call-in radio programs).
According to newer media logics, news workers and audience members produce and consume content. News workers and audience members can also co-curate and co-create messages. In addition, more opportunities are offered to interact with content—e.g., commenting, sharing, and liking/reacting (Hermida, 2016).
News organizations can combine older and newer media logics by producing messages to be consumed by users of social networking sites and soliciting, highlighting, or interacting with UGC on these sites. Some of the organizations under study highlighted UGC by posting it on their page / timeline. Some organizations replied to commenters.
Organizations also solicited UGC occasionally. Generally, mass-market outlets (e.g., Yedioth
Ahronoth) seemed more inclined to post UGC. Although the different organizations under study sought to entice audience engagement with their content, it seems that the potential for organizations’ engagement with UGC was not fully realized, at least not in public spaces (as opposed to non-public spaces, such as closed Facebook groups).
One way to elevate voices of audience members is to post UGC on news organizations’ page / timeline on social networking sites. Two Israeli outlets, Yedioth
Ahronoth and ynet, posted such content on their Facebook pages, in an attempt to engage audience members and to come across as outlets of the people. One type of content was
Facebook posts copied from users’ pages to the outlets’ pages (after getting permission). As part of the copied posts, editors tagged those who originally created them. Selected posts could focus on ordeals, happy moments, or other human experiences. They could also contain a plea for help (e.g., asking to donate a bone marrow). The outlets welcomed posts by patients, soldiers, parents, and others. For example, the new media team at ynet collaborated with a group that helped cancer patients. Posts on patients who needed help or fought cancer
102 were taken from the Facebook page of this group. In addition to posting messages (written texts and visuals) as they appeared on other Facebook pages, ynet and Yedioth Ahronoth posted audio-visual materials that audience members shared with them. Figure 6 shows a photo of a sunrise over Agamon Hula, a lake in northern Israel. Yedioth Ahronoth posted the photo with a brief text: “Enchanting: There’s nothing like starting Saturday morning with sunrise at Agamon Hula” (November 22, 2019). In addition, the name of the person who took the photo is tagged in the post.
Figure 6
UGC on the Facebook Page of Yedioth Ahronoth
Interviewees at Yedioth Ahronoth and ynet indicated that posting UGC was a common practice. Itai, a director at Yedioth Ahronoth, said: “I put a lot of emphasis on user-generated content… because it’s important to build this community and that people would feel it speaks to them.” Itai also thought that by promoting UGC calling people “to do something good,”
Yedioth Ahronoth showed it was “a newspaper with a heart.” Yair, of ynet, suggested that the content his team shared about cancer patients generated people’s interest because it was
103 strong and moving. According to team member Dvir, such content showed people’s greatness in coping with a difficult disease. “We are happy to spread their stories,” he said.
UGC can also be highlighted in spaces dedicated to this type of content. At the time of my interviews with USA TODAY workers, the organization had accounts called “Your
Take” on social networking sites. According to the account description on Facebook, for example, “Your Take is the news through your eyes. Share your original photos or videos with us here, or on Twitter or Instagram. You could be featured on USA TODAY.” Nancy, a director at USA TODAY, said the content on these account included “stunning photographs.”
In addition, she mentioned that audience members were invited to email photos of their pets.
The “Pet of the Week” was featured on the main Instagram account of USA TODAY.
Another practice mentioned by interviewees is retweeting messages by users. As an illustration, Slate retweeted messages not only by its journalists, but also by people who wrote something positive about Slate and its products (as long as these people were not politicians). Slate interviewees hoped that such messages would encourage audience members to consume the organization’s products and to even join its membership program.
News organizations under study occasionally solicited content from audience members. According to Karen, of Vox, “We do crowd source stories and story ideas as a way to build community. We do it mostly around big projects like a big video series, or we just did a yearlong reporting series on emergency room bills where we asked people to submit their bills, and then we report based on what we found.” For a Los Angeles Times’ project on breakup songs, audience members were asked what was a song they could not hear without thinking of their ex or someone from their past. Names of selected songs and people’s explanations of their affiliation with these songs were later published, and a Spotify playlist was offered.
104
A prominent type of UGC is comments attached to posts. Because comments are not always civil and well-reasoned, “commenting can get a bad rap,” as Mary, of Los Angeles
Times, put it. But the news organizations under study sought to entice audience comments on their posts. A high volume of comments was desired in cognizance of their potential to help increase the visibility of posts (e.g., Facebook’s algorithm may prioritize highly commented- on posts). Although the examined organizations wanted their posts to generate different modes of engagement—commenting, sharing, and liking/reacting—interviewees did not perceive these modes as equally valuable. Commenting and sharing were perceived as more valuable than liking/reacting. To like/react, one only needs to press a button, but to comment, one needs to say something. Some interviewees even suggested that comments could be more valuable than shares. Yair, of ynet, said he understood from Facebook that for the algorithm, comments and shares were linked together, but comments carried more weight. In other words, it was “better” to get more comments than more shares. Itai, of Yedioth Ahronoth, said that recently Facebook’s algorithm put more emphasis on discussions within posts. In other words, based on his understanding, the algorithm prioritized posts that generated discussions among commenters. Other interviewees thought that shares were more valuable than comments because the former allowed the spread of content.
One way or another, comments on a post have a notable aspect: it is possible to interact with them within the post. The possibility of interacting with audience comments in open spaces drew mixed reactions from interviewees: saying that they were interacting with commenters at least to some extent, expressing reluctance to interact, and expressing willingness to have more interaction. For example, David, a director at Haaretz, said that social media editors were encouraged to reply several times a day to commenters on the newspaper’s main Facebook page (while using the user name of Haaretz). Indeed, social media editor Uri shared that he was replying to commenters up to three times during an eight-
105 hour shift and trying to avoid getting into arguments. On Walla’s main Facebook page, replying to commenters was done occasionally, according to interviewees. Ariel, of Walla, thought it would be worthwhile to interact with commenters more frequently because interaction could help increase audience engagement. Yedioth Ahronoth and ynet, which highlighted UGC in other ways, seemed reluctant to reply to commenters on Facebook and tried to avoid controversies there. The approach of these mass-market outlets was starkly different from the approach of the upper-market outlet Haaretz, which was unlikely to post
UGC on its Facebook page but interacted with commenters there to some extent.
Interviewees at Vox and Slate indicated that replying to commenters rarely happened on the organizations’ public pages and was more likely to occur in closed groups. Karen, of Vox, explained:
We reply to comments if there’s an error or if we tweet out a job opportunity and
someone says, ‘Hey is this remote?’ We might respond to that, like a legitimate
question. But where we most interact with people is in closed, moderated communities
that we sort of own and can set up guidelines and expectations… in our Twitter
mentions and in our Facebook page comments, I can't control who’s commenting,
who’s saying what…
USA TODAY also experimented with closed groups, but interviewees indicated they were open to more interaction with commenters on public pages in the future. Lisa thought her audience engagement team would do well to have “a comments editor,” a person dedicated to moderating comments and interacting with commenters when appropriate.
Nancy suggested it was not easy to determine when and to what extent to converse with commenters. As she put it, USA TODAY could respond to people’s questions, but what happens if there is “a conversation that’s just going in a downwards spiral”? At which point would USA TODAY get involved, and what would that serve? Nancy said that a conversation
106 about these issues was happening at USA TODAY and across the industry and that an interaction with audience members was “very much something that we care about and are looking at.” In a similar vein, interviewees related to Makor Rishon and Israel Hayom indicated they would be happy to have more interaction with commenters on the newspapers’
Facebook pages, though they also acknowledged challenges it might involve.
At the Los Angeles Times, interviewees also thought it was important to converse with audience members. The organization used a relatively new commenting system on its website and developed a newer policy: a limited number of items on the website were open to commenting, and journalists were encouraged to interact with commenters. Such interaction was perceived as more complicated on social networking sites, where each post is typically open to commenting. However, Jennifer expressed support for two-way communication with audience members:
I’ve been trying to push more and more the metrics of things like meaningful
conversations in the comments sections, trying to retain people that are already
subscribers to make them feel that someone’s looking out for you, or even for people
who might not be subscribers who are thinking about subscribing and just holding
them onto LA Times content and telling them, ‘Hey, we’re here. We’re seeing what
you’re saying.’
To briefly summarize the dimension of approach toward UGC, news organizations blended older and newer media logics by producing messages to be consumed by users of social networking sites and soliciting, highlighting, or interacting with UGC on these sites.
However, the “old” generally seemed more dominant than the “new.” In addition, considerable differences were identified among organizations. In particular, mass-market outlets seemed more inclined than other outlets to post UGC.
107
Summary
This chapter has examined how older and newer media logics are applied by news organizations in their production of messages for social networking sites (RQ1). I have identified and explicated what constitutes the media production logics (RQ1a) and how these logics are applied in different national and organizational contexts (RQ1b). Social networking sites are platforms non-proprietary to the news media—platforms that have become dominant in the digital media environment. On these platforms, news organizations occupy triple-party news-spaces, which involve a news publisher, a platform owner (that is not the publisher), and users. Indeed, this chapter has shown that in the production of messages for the spaces under study, news organizations take into consideration perceived behaviors of platform owners and users.
I have suggested that the media production logics underlying these news-spaces are comprised of five key dimensions: item selection, temporality considerations, hierarchy considerations, expression style, and approach toward UGC. Along each dimension, news organizations combine “old” and “new“ by either applying them in the same messages, that is, “old” and “new” coexist, or applying them in separate messages, that is, message A reflecting an older logic is posted and afterwards message B reflecting a newer logic is posted. By so doing news organizations strive to balance between perceived journalistic imperatives or standards and perceived rules of the social media “game.” In their selection of items, news organizations try to assess which stories or story components are important and which ones would generate more engagement. In determining the time and frequency of posting, news organizations seek to post about events in close proximity to their occurrence, but also attempt to post content at frequency or time that may help reach more users based on how the algorithm of a given platform is understood to be working. In displaying items, organizations signal the importance of certain content, though the hierarchy they offer is
108 typically more flattened compared with the one in a news outlet. In wording or conveying messages, organizations try to use a style accessible to users—e.g., a more informal or conversational style—while using either a minimal degree or a higher degree of subjective expression. Finally, organizations produce messages to be consumed by users, and solicit, highlight, or interact with UGC.
In the application of “old” and “new,” I identified commonalities and differences among the examined organizations in the U.S. and Israel. Considerable commonalities were found along three dimensions: item selection, temporality considerations, and hierarchy considerations. However, considerable differences were identified along the other two dimensions: expression style, and approach toward UGC. With regard to expression style,
Israeli organizations were generally more willing than U.S. organizations to use unattributed evaluative language in posts—at least on certain issues. Possible explanations for this difference are the Israeli cultural characteristic of more direct talk and the tendency of Israeli journalists to be more interventionist in their role perception. With regard to approach toward
UGC, mass-market outlets seemed more inclined than other outlets to post UGC possibly because they tried to come across as connected to “the people.”
The unpacked media production logics underlying triple-party news-spaces are important as they elucidate how news organizations operate on platforms that play a key role in shaping who gains prominence and what people see. Analyzing the application of these logics in different cultural/national and organizational contexts is also important because it helps understand what transcends these contexts and what may be contingent upon them.
Media production is one part of the picture. Another part is media use. News organizations produce content geared toward audience engagement. The next steps of this dissertation involve a systematic examination of the produced content and the actual engagement it generates. Furthermore, as the discussion on organizations’ approach toward
109
UGC has suggested, certain modes of engagement may be perceived as more valuable than other modes. Therefore, it is also important to understand to what extent different types of content generate different modes of engagement (e.g., commenting versus sharing versus liking/reacting) and which types of content generate different modes of engagement. The next chapter addresses these questions, drawing on quantitative analyses and using the qualitative interview data for context or examples. The combination of qualitative and quantitative approaches allows me to gain understanding of both production and use of messages in triple- party news-spaces and to reveal news engagement logics that underlie these spaces.
110
CHAPTER 5
Findings: Logics of Media Usage in Triple-Party News-Spaces
This chapter focuses on user engagement with messages by news organizations in triple-party news-spaces. Specifically, I examine to what extent and how news organizations’ posts on Facebook differ in the modes of engagement they generate (RQ2). Facebook is a dominant platform worldwide and is used by all of the news organizations under study. It also offers different modes of engagement: commenting, sharing, and liking/reacting (using a thumbs-up sign or another emoji, such as a laughing face or a crying face). As suggested earlier in the literature review, different modes of user engagement can be regarded as different acts of civic engagement. Briefly, commenting allows users to have their say and converse with others about matters of public concern. Through sharing users can distribute content about such matters, and through liking/reacting users can indicate how they view or feel about matters at hand. As indicated in the previous chapter on logics of media production, news publishers seek to evoke comments, shares, and likes/reactions, but do not necessarily perceive all of these modes of engagement as equally valuable. Interviewees thought that Facebook’s algorithm gave more weight to comments and shares than to likes/reactions. Shares were seen as helpful in spreading content, and comments attached to posts were perceived by some interviewees as contributing even more to the visibility of posts. Because commenting, sharing, and liking/reacting are different acts of engagement, it is important to understand not only which messages are generally more engaging than other messages, but also which messages prompt different acts of engagement.
In this chapter, I offer three main insights about Facebook posts that are produced by news organizations and prompt user engagement—or three key principles of logics of media usage in the triple-party news-spaces under study. First, although highly commented-on posts, highly shared posts, and highly liked or reacted-to posts can overlap, they only
111 partially overlap. Each group of posts can differ from one another to a notable degree.
Second, certain content characteristics are linked to each of the examined modes of engagement in more than one country. For example, a reference to a prominent figure, such as a president or an actor, and a surprising element or something unusual, were positively associated with sharing, commenting, and liking / reacting among the examined posts by both U.S. and Israeli organizations. Third, certain content characteristics are linked to one or more modes of engagement, but not to all of them. As an illustration, posts about health were more likely to be shared than non-health posts particularly in the U.S., but the former posts were not associated with commenting and liking / reacting. Posts with an overall positive tone were more likely to be liked than posts without this tone in both the U.S. and Israel, but the former posts were not more likely to prompt comments. These principles of logics of media usage in the examined spaces help think about content units as more engaging or less engaging than others, or as engaging in different ways.
Drawing on literature regarding news content, I identified five possible content- related components to examine: news topics (e.g., politics and health), news factors (e.g., a prominent figure and a surprise), evaluative aspects (e.g., an overall tone and expressing criticism), triggers of emotions (e.g., evoking anger or sadness), and multimedia features
(e.g., including a video or a photo). As mentioned earlier and explained in this chapter, news topics and factors are seen in this study as reflecting older media logics, whereas evaluative aspects, emotions’ triggers, and multimedia features are seen as reflecting newer media logics in that they have been brought to the foreground in the era of digital media and fit affordances or constraints of social networking sites. The findings presented in this chapter demonstrate that particular content-related characteristics reflecting older or newer media logics are associated with commenting, sharing, or liking/reacting.
112
The chapter is divided into four sections. The first one shows the extent to which different engagement metrics correlate (RQ2a) and the degree of overlap among the most commented-on, shared, and liked/reacted posts (RQ2b). It is based on correlational analyses of numbers of comments, shares, and likes/reactions attached to posts by 16 news organizations—eight U.S. organizations and eight Israeli organizations. Separate analyses were conducted for each organization, and metrics for 157,962 posts were analyzed in total.
In addition, the degree of overlap (expressed as a percentage) among the most commented- on, shared, and liked/reacted posts was examined for each organization. The subsequent sections show which content characteristics of the posts are associated with different modes of engagement (RQ2c and H1-H3) and whether these characteristics differ across cultural/national and organizational contexts (RQ2d). These sections are based on a content analysis of 1,600 posts by the organizations under study. I first address news topics and factors (components reflecting older media logics). I then address evaluations, emotions, and structural features (components seen as reflecting newer media logics). Finally, I present models that include the different content-related components (reflecting older and newer media logics) and evaluate which components are stronger predictors of modes of engagement.
To better understand the quantitative findings throughout this chapter, I include examples of posts, as well as quotes from media professionals about user engagement that did not appear in the chapter on logics of media production. The combination of quantitative findings and qualitative elements allows the identification and explication of the logics of media usage underlying the triple-party news-spaces under study.
Commenting, Sharing, Liking, and Reacting: Correlations and Overlaps
This section addresses the research questions on the extent to which different engagement metrics correlate (RQ2a) and the degree of overlap among the most commented-
113 on, shared and liked/reacted posts (RQ2b). Investigating these questions allows assessing if different posts by news organizations engage users in different ways. Data about the posts published on the main Facebook page of each news organization in the second half of 2018
(from July 1 to December 31) and the first half of 2019 (from January 1 to June 30) were separately obtained (for each period). Sponsored posts that appeared on the news organizations’ pages were excluded from the data. Adopting Tenenboim and Cohen’s approach (2015) to analyzing the relationships between clicking and commenting on news items, Pearson correlations were examined between audience engagement metrics of
Facebook posts: comments and shares, comments and likes, comments and reactions (likes + other emojis), shares and likes, and shares and reactions (likes + other emojis). The unit of analysis was a post, and the correlations were tested for each news organization separately in each of the two time periods. Two time periods were examined to help identify possible patterns. In addition, the degree of overlap (as a percentage) was examined between highly commented-on and highly shared posts, highly commented-on and highly liked posts, highly commented-on and highly reacted-to posts, highly shared and highly liked posts, and highly shared and highly reacted-to posts. The overlap was tested for both the top 10% and the top
5% of posts in each group—e.g., the top 10% of the most commented-on posts versus the top
10% of the most shared posts, as well as the top 5% of the most commented-on posts versus the top 5% of the most shared posts—for each news organization in 2018 and 2019. Highly engaging posts are given close attention in this study because they may be more visible to
Facebook users. Facebook’s Like button existed years before the relatively new emojis and is equivalent to features on other platforms (e.g., Twitter’s Like button). Therefore, likes were analyzed both separately and as part of “reactions” (likes + other emojis).
The news organizations under study include the 14 organizations that were mentioned in the chapter on logics of media production, as well as two additional organizations—the
114
U.S. organization The Daily Caller and the Israeli organization Srugim. These are two digital- native outlets that lean conservative (Hagey, 2020; Srugim, 2020). They were added to further diversify the selected news organizations. For U.S. news organizations, the number of analyzed posts in a given year ranged from 4,668 (The Dallas Morning News, 2019) to
11,093 (The New York Times, 2018). For Israeli news organizations, the number of analyzed posts in a given year ranged from 312 (Srugim, 2018) to 4,961 (Maariv, 2018). In total,
157,962 posts were analyzed—115,207 posts from U.S. organizations and 42,755 from Israeli organizations.
Table 3 shows Pearson correlations between engagement metrics for the examined
U.S. news organizations, and Table 4 presents Pearson correlations between the same engagement metrics for the Israeli news organizations under study. The examined engagement metrics were positively correlated (p < .001), though the correlations were not necessarily strong. Specifically, the reported correlations between comments and shares were usually not higher than .700. They ranged from .251 to .698 for 28 of the 32 coefficients.
Two of the 32 coefficients were higher than .700 (.844 in the U.S. and .815 in Israel), and two other coefficients were particularly low (.116 in the U.S. and .169 in Israel). In a similar vein, the correlations between comments and likes were usually not higher than .700. They ranged from 218 to .648 for 30 of the 32 coefficients. Two of the 32 coefficients were higher than
.700 (.733 in the U.S. and .703 in Israel). However, the reported correlations between comments and reactions ranged from .435 to .907. For most of the news organizations, these correlations were higher than the correlations between comments and likes. Most of the correlations (27 out of 32) between shares and likes ranged from .338 to .673, and five correlations were above .700 (three in the U.S. and two in Israel). The correlations between shares and reactions varied, ranging from .281 to .931. For most of the news organizations, the correlation between shares and reactions in a given time period was higher than the
115
Table 3
Pearson Correlations Between Engagement Metrics for U.S. News Organizations
News Commenting and Commenting and Commenting and Sharing and Liking Sharing and Organization Sharing Liking Reacting Reacting 2019 2018 2019 2018 2019 2018 2019 2018 2019 2018
The New .527 .594 .375 .370 .702 .737 .616 .550 .761 .723 York Times
USA TODAY .844 .403 .733 .401 .907 .579 .801 .543 .931 .667
Vox .607 .415 .579 .406 .707 .523 .706 .712 .611 .573
Slate .607 .532 .332 .328 .741 .728 .565 .537 .818 .755
The Dallas .698 .556 .631 .526 .799 .706 .393 .584 .845 .643 Morning News
Los Angeles .592 .603 .448 .511 .666 .677 .634 .645 .783 .777 Times
South Florida .116 .544 .538 .614 .651 .722 .461 .494 .281 .731 Sun Sentinel
The Daily .391 .373 .218 .223 .820 .715 .564 .673 .538 .595 Caller
Note. All correlations are significant at p < .001. The number of analyzed posts ranges from 4,668 (The Dallas Morning News, 2019) to 11,093 (The New York Times, 2018).
116
Table 4
Pearson Correlations Between Engagement Metrics for Israeli News Organizations
News Commenting and Commenting and Commenting and Sharing and Liking Sharing and Organization Sharing Liking Reacting Reacting 2019 2018 2019 2018 2019 2018 2019 2018 2019 2018
Haaretz .815 .557 .703 .578 .756 .679 .769 .612 .797 .716
Yedioth .327 .520 .420 .431 .446 .435 .649 .560 .658 .599 Ahronoth ynet .389 .299 .570 .486 .592 .488 .499 .432 .529 .481
Walla .559 .313 .478 .451 .539 .493 .418 .437 .520 .449
Israel Hayom .251 .169 .401 .478 .512 .500 .665 .546 .657 .526
Maariv .433 .470 .432 .561 .455 .625 .408 .338 .444 .431
Makor .603 .575 .648 .635 .740 .666 .684 .726 .751 .757 Rishon
Srugim .367 .339 .309 .392 .429 .446 .408 .585 .551 .591
Note. All correlations are significant at p < .001. The number of analyzed posts ranges from 312 (Srugim, 2018) to 4,961 (Maariv, 2018).
117 correlation between shares and likes in the same time period. That said, none of the presented correlations indicate a perfect relationship (Pearson’s r = 1) between engagement metrics.
More stable patterns were found among highly engaging posts. Table 5 shows the percentage of overlap among highly engaging posts of the examined U.S. news organizations, and Table 6 presents the percentage of overlap among highly engaging posts of the Israeli news organizations under study. The percentages indicate that although groups of highly engaging posts can contain the same posts (e.g., posts can be both highly commented-on and highly shared), these groups also differ from one another to a notable degree. The degree of overlap between the top 10% of commented-on posts and the top 10% of shared posts of a given organization ranged from about 39% to 60%. Put differently, between 40% (100 – 60) and 61% (100 – 39) of the most commented-on posts were not the same as the most shared posts. For almost all of the organizations (except for one), the degree of overlap between the top 5% of commented-on posts and the top 5% of shared posts in a given time period was even lower than the degree of overlap between the top 10% of the engaging posts in the same time period. Focusing on the top 5% of engaging posts, the degree of overlap between commented-on posts and shared posts ranged from about 30% to 55%. Conversely, between
45% and 70% of the most commented-on posts differed from the most shared posts.
Similar patterns were found regarding commented-on and liked posts. The degree of overlap between the top 10% of commented-on posts and top 10% of liked posts ranged from about 21% to 55%. Put differently, between 45% and 79% of the most commented-on posts were not the same as the most liked posts. For all of the organizations under study, the degree of overlap between the top 5% of commented-on posts and the top 5% of liked posts in a given time period was lower than the degree of overlap between the top 10% of the engaging posts in the same time period. Focusing on the top 5% of engaging posts, the degree of
118
Table 5
Percentage of Overlap Between Highly Engaging Posts of U.S. News Organizations
News Highly Commented-On Highly Commented-On Highly Commented-On Highly Shared and Highly Shared and Organization and Highly Shared Posts and Highly Liked Posts and Highly Reacted-To Highly Liked Posts Highly Reacted-To Posts Posts Top 10% Top 5% Top 10% Top 5% Top 10% Top 5% Top 10% Top 5% Top 10% Top 5% 2019 2018 2019 2018 2019 2018 2019 2018 2019 2018 2019 2018 2019 2018 2019 2018 2019 2018 2019 2018
The New 57.3 57.0 43.1 41.8 40.1 40.5 23.5 21.3 59.0 60.3 48.1 44.3 57.3 56.7 47.2 41.6 75.6 73.9 66.2 65.1 York Times
USA 53.4 54.3 43.8 42.6 49.4 47.6 41.7 34.2 61.7 56.1 53.0 47.0 60.1 62.7 50.7 52.5 72.6 74.5 66.0 69.1 TODAY 49.1 47.6 39.9 40.9 38.7 36.2 30.6 29.9 59.0 60.3 43.4 42.5 56.6 60.1 39.9 52.5 68.6 66.4 61.8 59.4 Vox 53.8 50.7 42.9 42.1 35.2 32.8 22.7 20.3 57.9 53.4 42.6 44.7 55.4 55.6 45.7 44.7 74.8 73.2 69.4 67.1 Slate 56.8 60.2 46.8 48.0 49.5 51.7 40.8 38.3 60.6 65.4 52.4 50.4 57.2 56.9 46.8 47.2 72.2 70.0 64.0 61.3 The Dallas Morning News 49.8 53.7 42.0 48.7 43.9 41.1 30.5 31.8 54.4 54.9 41.0 46.7 60.2 57.0 53.4 47.9 76.1 73.4 69.5 68.2 Los Angeles Times 53.6 46.4 39.1 37.9 51.9 52.1 40.9 46.9 63.2 60.0 56.0 52.9 61.6 59.0 48.6 50.6 56.8 67.6 64.3 81.8 South Florida Sun Sentinel 51.9 43.9 40.5 30.1 21.2 21.1 13.1 14.7 51.1 35.9 43.1 25.1 48.0 49.2 41.1 44.8 70.7 59.5 61.5 53.3 The Daily Caller
119
Table 6
Percentage of Overlap Between Highly Engaging Posts of Israeli News Organizations
News Highly Commented-On Highly Commented-On Highly Commented-On Highly Shared and Highly Shared and Organization and Highly Shared Posts and Highly Liked Posts and Highly Reacted-To Highly Liked Posts Highly Reacted-To Posts Posts Top 10% Top 5% Top 10% Top 5% Top 10% Top 5% Top 10% Top 5% Top 10% Top 5% 2019 2018 2019 2018 2019 2018 2019 2018 2019 2018 2019 2018 2019 2018 2019 2018 2019 2018 2019 2018
Haaretz 54.9 55.6 47.0 49.0 53.2 53.3 42.2 40.1 57.8 55.4 51.9 48.4 64.1 66.3 54.6 54.7 68.9 71.0 59.5 65.1
Yedioth 42.7 44.4 32.8 31.8 44.2 44.4 38.8 35.7 50.2 46.7 40.3 35.7 55.8 55.6 50.8 56.6 59.6 60.7 54.5 57.4 Ahronoth ynet 43.2 48.5 37.8 42.1 38.3 39.5 36.0 35.4 42.6 45.1 37.2 34.8 48.9 47.3 46.3 40.5 54.7 54.6 50.6 48.3
Walla 57.4 50.7 51.2 46.7 48.7 45.2 42.8 39.3 53.2 51.9 48.1 41.5 55.9 54.4 53.4 49.6 63.5 62.2 58.8 54.1
Israel Hayom 54.0 51.0 44.9 41.8 43.5 42.5 35.9 38.4 48.2 46.6 37.8 41.1 46.3 50.0 38.5 45.9 56.2 58.6 47.4 53.4
Maariv 51.1 52.8 42.3 45.6 55.1 51.2 51.5 40.7 59.9 55.7 53.3 44.8 45.4 43.4 33.9 37.9 47.6 48.0 37.0 43.2
Makor 53.0 51.6 54.6 43.8 51.5 53.1 48.5 40.6 60.6 56.3 57.6 46.9 66.7 67.2 54.6 56.3 69.7 71.9 63.6 59.4 Rishon
Srugim 43.9 38.7 39.3 31.3 37.9 35.5 35.7 18.8 45.5 35.5 42.9 18.8 45.5 48.4 42.9 56.3 51.1 45.2 46.4 56.3
120
overlap between commented-on posts and liked posts ranged from about 13% to 52%, that is, between 48% and 87% of the most commented-on posts differed from the most liked posts.
The degree of overlap between the top commented-on posts and the top reacted-to posts in a given time period was typically higher, or at least not lower, than the degree of overlap between the top commented-on posts and the top liked posts in the same time period.
The degree of overlap between the top 10% of commented-on posts and the top 10% of reacted-to posts ranged from about 36% to 63%. Put differently, between 37% and 64% of the most commented-on posts were not the same as the most reacted-to posts. For all of the examined organizations, the degree of overlap between the top 5% of commented-on posts and the top 5% of reacted-to posts in a given time period was lower than the degree of overlap between the top 10% of the engaging posts in the same time period. Focusing on the top 5% of engaging posts, the degree of overlap between commented-on posts and reacted-to posts ranged from about 19% to 58%, that is, between 42% and 81% of the most commented- on posts differed from the most reacted-to posts.
The tables also show the degree of overlap between the top shared posts and the top liked posts. Focusing on the top 10%, the degree of overlap ranged from about 43% to 67%.
Put another way, between 33% and 57% of the most shared posts differed from the most liked posts. For most of the organizations under study, the degree of overlap between the top
5% of shared posts and the top of 5% of liked posts in a given time period was lower than the degree of overlap between the top 10% of the engaging posts in the same time period. As for the top 5% of engaging posts: the degree of overlap between shared posts and liked posts ranged from about 34% to 57%, that is, between 43% and 66% of the most shared posts were not the same as the most liked posts.
The degree of overlap between the top shared posts and the top reacted-to posts in a given time period was higher than the degree of overlap between the top shared posts and the
121
top liked posts in the same time period for most of the organizations under study. The degree of overlap between the top 10% of shared posts and the top 10% of reacted-to posts ranged from about 45% and 76%. Put another way, between 24% and 55% of the most shared posts differed from the most reacted-to posts. For most of the examined organizations, the degree of overlap between the top 5% of shared posts and the top 5% of reacted-to posts in a given time period was lower than the degree of overlap between the top 10% of the engaging posts in the same time period. Focusing on the top 5% of engaging posts, the degree of overlap between shared posts and reacted-to posts ranged from about 37% to 82%, that is, between
18% and 63% of the most shared posts were not the same as the most reacted-to posts.
At the very top of different groups of engaging posts, the overlap may be even lower than it is among the top 5% of engaging posts, suggesting that content generating the highest volume of one mode of engagement may be particularly different from content generating the highest volume of another mode of engagement. I demonstrate this for two U.S. organizations and two Israeli organizations based on the actual content of posts. I present this content for the purpose of illustration—beyond correlations and percentages—but without delving at this point into specific content characteristics. The New York Times’ five most shared posts in the first half of 2019 differed from the five most commented-on posts in the same time period, and the latter posts differed from the five most liked posts. The Times’ most shared post and the most liked post was the same one—a post reporting that a patient had been cured of infection with H.I.V, the virus that causes AIDS (March 4, 2019). The following shared posts were about migrant children subject to conditions of overcrowding and filth at a Texas detention center (June 21, 2019); a U.N. report assessing that as many as one million plant and animal species across the globe were at risk of extinction (May 6, 2019); a son who testified in a drug trial of his father’s former partner (El Chapo’s trial) (January 3, 2019); and an Alabama woman who was shot and charged in her fetus’s death (June 27, 2019). In
122
addition to the post about the cured patient, the most liked posts focused on a Chicago woman who got 30 hotel rooms for homeless people during severe cold snap (February 2,
2019); banning single-use plastic bags for retail sales in New York State (March 28, 2019);
Tiger Woods who won the Masters Tournament, a major championship in professional golf
(April 14, 2019); and Jimmy Carter who became the longest-living U.S. president after reaching the age of 94 years and 172 days (March 22, 2019). The five most liked posts were also the most reacted-to posts (though their order was not identical).
However, the Times’ most commented-on posts featured other stories, which were mostly related to U.S. President Donald Trump. The most commented-on post was about
Donald Trump Jr., the president’s eldest son, who got a book deal and said he would offer his views on the “great achievements” of his father’s administration and his observations on “this significant time in history” (May 22, 2019). The next most commented-on posts were about
President Trump who wanted to give a prime time address on the federal shutdown and would travel to the southern border to make his case for a wall (January 7, 2019); President
Trump who sued two banks in a bid to prevent them from responding to congressional subpoenas (April 29, 2019); welcoming college football champions with fast-food buffet at the White House (January 15, 2019); and the mystery of Trump’s tan in the pale of winter
(February 2, 2019).
In a similar vein, the five most commented-on posts of The Daily Caller differed from its most shared, most liked, and most reacted-to posts. The most commented-on post contained a link to an opinion about Democrats’ initiative to lower the voting age to 16
(March 19, 2019). The next most commented-on posts mentioned prominent women. The posts were about a tweet against Trump by U.S. Representative Maxine Waters, a California
Democrat (“Trump is known for calling others disrespectful names, especially women. It’s your turn now. What name would you call him?”) (June 7, 2019); remarks by U.S.
123
Representative Alexandria Ocasio-Cortez, a New York Democrat, regarding a plan for tackling climate change (March 27, 2019); soccer player Megan Rapinoe who indicted she would not go to the White House if the women’s national team won the World Cup (June 26,
2019); and Ocasio-Cortez’s criticism of Trump’s State of the Union address (February 6,
2019).
The Daily Caller’s most shared posts differed from its most commented-on posts, as well as from the outlet’s most liked and reacted-to posts. Its most shared post featured a video about border walls and suggested that they were “an effective way to maintain border security” (January 10, 2019). The next most shared posts featured critical remarks by conservative commentator Candace Owens at a congressional hearing about hate crimes and the rise of white nationalism (April 9, 2019); a rundown of an incident involving a Native
American man and high school boys wearing “Make America Great Again” gear (January 23,
2019); an initiative of U.S. Senator Ted Cruz, a Texas Republican, to seize funds from convicted drug lord Joaquín “El Chapo” Guzmán to pay for a border wall with Mexico
(January 12, 2019); and a video about children in a Philadelphia Muslim center who were said to sing a song with violent lyrics (May 6, 2019).
However, The Daily Caller’s most liked post centered on Apollo 11 astronaut Buzz
Aldrin who saluted after being introduced by U.S. President Trump at the State of the Union address (February 5, 2019). The next most liked posts were about White House Press
Secretary Sarah Sanders who wondered after the arrest and indictment of Roger Stone, an aide and confident of Trump, whether the same standard would apply to people like Hillary
Clinton (January 25, 2019); Candace Owens’s response to U.S. Representative Ted Lieu
(Democrat of California) who played a clip where she made controversial comments about
Adolf Hitler (April 9, 2019); Trump who amid a government shutdown denied House
Speaker Nancy Pelosi and a congressional delegation the use of a military aircraft to visit
124
American troops in Afghanistan (January 17, 2019); and an off-duty policewoman in Brazil who shot a robber after he tried to attack her (March 27, 2019). Three of the five most liked posts were also among the five most reacted-to posts. The most reacted-to post reported that congressional Democrats who were already on their way to a military jet to fly to Europe were furious when they heard that Trump had cancelled their trip (January 17, 2019). Another highly reacted-to post reported that activists jumped the fence around Pelosi’s California house and set up a tent to protest immigration (January 14, 2019).
In Israel, the overlap at the very top of different engaging posts was also relatively low, showing a pattern that can transcend cultural/national contexts. The five most commented-on posts of Haaretz in the first half of 2019 included two posts that were also among the five most shared posts, and one post that was among the most liked posts and among the most reacted-to posts. Haaretz’s most engaging post (most commented-on, most shared, most liked, and most reacted-to post) featured critical opinion piece by children’s author Yehuda Atlas who was ashamed to live in a country where people were facing grave injustices and the prime minister was facing criminal allegations (March 28, 2019). The next most commented-on posts were about the Israeli prime minister’s elder son Yair Netanyahu who described some Israeli bereaved families as “mentally ill” because they wanted to have a joint Memorial Day ceremony with Palestinians bereaved families (May 7, 2019); Israelis who sought a haven in Portugal after feeling they could not live in Israel according to their desired lifestyle (May 13, 2019); Yair Netanyahu who left an open tab at a luxury hotel in
Brazil (March 11, 2019); and the top 10 pizzerias in the Northern Israeli city Haifa (May 11,
2019).
In addition to the opinion piece by the children’s author and the story about Yair
Netanyahu’s open tab, Haaretz’s most shared posts featured other pieces of content: an opinion piece suggesting that Israelis were paying too much and getting too little (e.g., a low
125
quality of public transportation) (February 7, 2019); an editorial calling to protect LGTBQ teenagers and not to appoint a “homophobe” politician education minister (April 18, 2019); and the Israeli prime minister’s wife Sara Netanyahu who reportedly accused journalists from a conservative website of being “leftists” and working for Netanyahu’s rival (April 8, 2019).
Haaretz’s most liked posts included two of the most shared posts (two opinion pieces), as well as other posts. The latter posts focused on actress Gal Gadot who expressed support for Jewish-Arab coexistence (March 10, 2019); actress Rotem Sela who criticized a
Jewish Israeli minister for delegitimizing Arab Israeli members of parliament (March 10,
2019); and a woman who killed her abusive husband and was expected to be released from prison after 18 years (June 19, 2019). Haaretz’s most reacted-to posts included four of the five liked posts, as well as a post about the death of Nechama Rivlin, the wife of Israel’s president (June 4, 2019).
The Israeli outlet Srugim (more conservative) showed a similar pattern regarding the overlap of engaging posts, suggesting that this pattern can transcend organizational contexts
(in addition to cultural/national contexts). Srugim’s five most commented-on posts in the second half of 2019 differed from the most liked and reacted-to posts, though one of the most commented-on post was also one of the five most shared posts. The most commented-on post centered on a former journalist (Igal Sarna) who indicated he would not live in Israel if
Netanyahu were reelected (April 4, 2019). The next most commented-on posts focused on a
TV host (Oded Ben-Ami) who suggested that god might be “a bit gay” (March 11, 2019); a rabbi who thought that voting for one of Israel’s right-wing parties was support for racism
(February 23, 2019); a Palestinian charged with murdering an Israeli woman (February 11,
2019); and a journalist (Yaron London) who thought Israel could have a different national anthem that Arab Israelis would identify with (May 19, 2019).
126
The outlet’s most shared posts included the above-mentioned story about the TV anchor and god, as well as content about other people or issues: hospital workers who were asked to remove flags of Israel so that people’s feeling would not be hurt (May 3, 2019); Bon
Jovi rock band that made its way to Israel (June 13, 2019); a rabbi’s son injured in a car accident (March 30, 2019); and demolishing buildings in an Israeli settlement (April 30,
2019). In comparison, Srugim’s most liked posts were different. They focused on the birth of a baby girl to a woman whose mother was stabbed to death by a Palestinian (March 16,
2019); a singer (Omer Adam) who chose not to perform in the Eurovision song contest because he observed the Jewish Shabbat (January 10, 2019); a TV host (Avri Gilad) responding to critics who suggested he was Islamophobe (April 23, 2019); Israeli culture minister who asked the European Broadcasting Union to allow a band to perform in the
Eurovision song contest without desecrating Shabbat (February 4, 2019); and the same TV host (Avri Gilad) who criticized the Israeli news media on a different occasion (April 10,
2019). The outlet’s most reacted-to posts included two of the most liked posts and three posts reporting death: a woman was killed in a car accident a month after getting married (January
13, 2019); a son died from his wounds two years after his father had been murdered in a shooting incident (March 31, 2019); and a grand rabbi who was a Holocaust survivor passed away (April 28, 2019).
Although the mentioned posts by U.S. and Israeli organizations illustrate patterns of overlap or lack thereof and give a sense of some of the content behind the numbers, they do not answer the question of larger systematic differences, of which content characteristics are associated with different modes of engagement. To answer this question, a quantitative content analysis was conducted.
127
Analysis of Commented-On, Shared, and Liked or Reacted-To Content
A quantitative content analysis of news organizations’ posts was conducted to examine relationships between content characteristics of these posts and modes of engagement the posts generated. A total number of 1,600 posts were analyzed, consisting of posts published on Facebook by the 16 organizations under study. To obtain this sample of posts, two steps were taken. First, constructed weeks were created to account for systematic variation due to day of week (Riffe, Aust, & Lacy, 1993). Based on the approach that for online news content at least two constructed weeks are needed in a six-month period (Hester
& Dougall, 2007), four constructed weeks were created—two for the last six months of 2018 and two for the first six months of 2019. The total number of posts published on the days of the constructed weeks is 12,191—8,969 posts by the U.S. organizations and 3,222 posts by the Israeli organizations. Second, systematic random sampling was used to select 800 U.S. posts (for eight organizations) and 800 Israeli posts (for eight organizations). These posts were analyzed by the author and three trained native speakers of English or Hebrew. Inter- coder reliability was examined for 150 of the U.S. posts (18.75%) and 150 of the Israeli posts
(18.75%). Only variables with Krippendorff’s alpha not lower than .700 were included in the analysis (See Appendix C for more information).
To analyze the coded posts, negative binominal regression analyses were performed.
Negative binominal regression is suitable for over-dispersed count data (when the conditional variance is greater than the conditional mean) and has been used for predicting sharing on social media (Valenzuela et al., 2017) and participation in an article’s comments section on a news site (Weber, 2014). As with Valenzuela and his colleagues, I report incidence-rate ratios
(IRRs), showing changes in incidence rate for a one unit increase in the predictor variable. To understand which content characteristics are associated with different modes of engagement
(RQ2c), bivariate associations between a given content characteristic and a given mode of
128
engagement are first examined, while controlling for the number of likes a Facebook page had at the time of posting. Later, multivariate models including the different content characteristics (predictors) are offered.
Older media logics: Topics and news values
This section focuses on the content characteristics of topics and news values, which can be seen as reflecting older media logics. In making news judgements, journalists prioritize content about certain topics, events, or people. For example, journalists may see public affairs—e.g., politics, economics, and international topics—as particularly important and may pay more attention to such affairs than audience members. Journalists also perceive stories that contain certain news factors, such as a conflict or surprise, as worthy of publication. In what follows, I show which topics and news values or factors are associated with different modes of engagement. The tables showing the associations focus on the U.S. sample, Israeli sample, and combined sample. If notable differences between newspapers and digital-native outlets were found, they are mentioned in the text.
The most prevalent topic in the sample combined across the U.S. and Israel (n =
1,600) is politics / government, accounting for more than a quarter (26.6%) of the posts. It is followed by arts / entertainment (12.9% of the posts), crime (6.8%), military / defense
(5.4%), human interest story (5.0%), society / welfare (4.4%), sports (4.3%), economy / business (3.5%), disasters / accidents (3.5%), health (3.3%), and food (3.3%). Topics with lower frequencies—topics that appeared in fewer than 50 posts—were not included in the analysis presented here (See Appendix D for detailed frequencies of topics, including frequencies in the U.S. and Israel).
Table 7 shows how each of the above-mentioned topics is related to each of the examined modes of engagement (commenting, sharing, liking, and reacting) after controlling for the number of likes a given Facebook page had at the time of posting. The table presents
129
Table 7
Relationships Between Topics and Modes of Engagement for News Organizations’ Posts
Topic Commenting Sharing Liking Reacting (IRR) (IRR) (IRR) (IRR)
Politics / Government General 2.264*** 2.123*** 1.652*** 1.766*** U.S. 2.340*** 1.557** 1.772*** 1.878*** Israel 2.224*** 3.544*** 1.520*** 1.648***
Military / Defense General 592** 1.050 1.295 1.256 U.S. .367* 1.341 ------Israel .630* 1.003 1.283 1.265
Economy / Business General .599* .659 .600* .529** U.S. .572* .655 .653 .564* Israel .657 .620 .423* .396*
Society / Welfare General .856 .501 .694 .810 U.S. .952 .852 .727 .878 Israel .719 1.311 .654 .720
Crime General .500*** .607** .360*** .561*** U.S. .608* .836 .352*** .590* Israel .406*** .454** .364*** .533***
Health General .722 2.354*** .798 .825 U.S. .978 4.074*** 1.105 1.149 Israel .497* 1.003 .536* .550*
Disasters / Accidents General .646 .905 .502** .918 U.S. 1.204 .679 .510 1.065 Israel .412*** 1.015 .492** .838
Arts / Entertainment General 1.188 .509*** 1.168 .976 U.S. .550*** .460*** .833 .670* Israel 1.721*** .539*** 1.416** 1.209
Sports General .492** .338*** 1.163 .825 U.S. .520* .400** 1.553 1.013 Israel .435** .220*** .556* .517*
Food General ------.535* .787 .529** U.S. ------.542 .911 .529* Israel .490* .516 .657 .526*
Human Interest General .744 1.351 1.617** 1.563* U.S. ------4.081** .670 .685 Israel .842 1.906** 2.070*** 1.978*
130
*** p < .001; ** p < .01; * p < .05; Note. Cells report incidence rate ratios (IRR) from negative binominal regressions. IRR > 1 indicates a positive effect while IRR < 1 indicates a negative effect. Bivariate associations were examined while controlling for the number of likes a Facebook page had at the time of posting. An empty cell indicates that no regression was performed due to a system’s error in computing the inverse LOG link function, the log likelihood, the gradient, or the Hessian matrix in the iterative process for at least one of the records. The general sample included 1,600 posts, and each of the subsamples (U.S. and Israel) included 800 posts.
131
incidence rate ratios (IRR) from negative binominal regressions. Whereas an IRR greater than one indicates a positive effect, an IRR smaller than one indicates a negative effect. For politics / government, all of the IRRs were greater than one (p < .001 or p < .01). Focusing on this topic in the general sample (U.S. posts + Israeli posts), the IRR was particularly high when commenting was the outcome variable: 2.264. This IRR suggests that posts where their main topic was politics / government were commented-on 126.4% more often than posts that their main topic was not politics / government. Posts about politics / government were shared
112.3% more often than other posts. In addition, the former posts were liked 65.2% more often than other posts. The positive associations between politics / government and modes of engagement were found both in the U.S. and Israel.
The table also shows positive associations between health and sharing. Focusing on the general sample, posts that their main topic was health were shared 135.4% more often than posts that their main topics was not health (p < .001). In the U.S. sample, posts about health were shared 307.4% more often than non-health posts (p < .001). Yet, in the Israeli sample, the association between health and sharing was not statistically significant.
H1 posited that politics / government would be positively associated with commenting, and health would be positively associated with sharing. Politics / government was indeed positively associated with commenting, though it was also positively associated with sharing and other modes of engagement. Health was positively associated with sharing, though not in Israel. Therefore, H1 is partially supported.
Other notable associations were also found. Specifically, human interest stories were positively associated with liking and reacting in the general sample and Israel. These stories are understood here as stories that focus on people’s experiences but do not clearly fall into one of the offered categories of topics (Tenenboim & Cohen, 2015)—e.g., a story about a person who survived against all odds, or a story about a person who won a lottery jackpot
132
after a stranger found and returned the winning ticket that he had left behind on a store counter. Focusing on the combined sample, human interest stories were liked 61.7% more often than stories that were not identified as human interest stories (p < .01). Among Israeli posts, the former stories were liked 107% more often than the latter stories. Yet, among U.S. posts, the examined association between human interest stories and liking was not statistically significant.
Negative associations were found between certain topics and modes of engagement.
In the general sample and among both U.S. and Israeli posts, posts about crime were commented-on and liked or reacted-to less often than non-crime posts, and posts focusing on sports or arts / entertainment were shared less often than posts that did not focus on these topics. Furthermore, posts about military / defense were commented-on less often than posts that did not center on military / defense, and posts focusing on economy / business were reacted-to less often than posts that did not focus on this topic.
For example, the IRR from the regression examining the relationship between crime and liking in the general sample was .360, suggesting that posts about crime were liked 64% less often than non-crime posts (p < .001). Focusing on the general sample, posts about sports were shared 66.2% less often than non-sports posts, and posts about arts / entertainment were shared 49.1% less often than posts where their main topic was not arts / entertainment.
However, in Israel the topic of arts / entertainment was positively associated with commenting and liking. Another analysis, which is not presented in the table, shows that this topic was positively associated with commenting among posts by digital-native outlets (IRR
= 1.916, p < .01). Yet, the examined association between arts / entertainment and commenting was not statistically significant among posts by newspapers.
Research has proposed that commenting plays a role in constructing a social or group identity through discussions of political issues (Boczkowski & Mitchelstein, 2012;
133
Tenenboim & Cohen, 2015). The present study suggests that sharing news messages about politics or reacting to them (by using a thumbs-up sign or another emoji) may also play a role in constructing such identity—at least in the U.S. and Israel. In these cultural/national contexts, the latitude for public expression on digital platforms seems considerable compared to other contexts (e.g., King, Pan, & Roberts, 2013; Klyueva, 2016), and political divisions are stark (Arian & Shamir, 2008; Iyengar et al., 2019; Bassan-Nygate &. Weiss, 2020).
As illustrated earlier, the most engaging political messages under study tended to focus on what top political actors, such as U.S. President Donald Trump and Israeli Prime
Minister Benjamin Netanyahu, said or did, and on reactions to what these actors said or did.
Another way to put it is that top political actors and issues related to them generated not only media attention, but also audience engagement (though it is also possible that at least some of the engagement is the result of a concerted activity by supporters or opposers, e.g., DW
News, 2019).
The findings regarding politics were also supported by the interviewees’ observations.
For instance, interviewees at Los Angeles Times indicated that stories related to Trump and
California tended to engage audience members. “Things like California is suing the Trump administration… those types of stories get a lot of comments,” said Jennifer of the newspaper’s audience engagement team. Yet, highly engaging political stories did not necessarily center on California. Jennifer shared a recent experience:
We’ve also found some success with writing national politics stories (not) focused on
California. I used to think that those stories might not do as well considering who our
audience is. But we actually wrote a story quite recently about Republican voters
from suburban Texas that are losing faith in the Republican party and Trump… might
not want to vote for him in 2020… It was not a California story, but there are a ton of
134
suburbs in Southern California, like in Orange County, that are extremely
Republican, so we got a lot of responses from people there.
Interviewees from other news organizations also indicated that political stories could engage audience members. Karen, a director at Vox, said: “our audience is very politics and policy focused, so a lot of our like less newsy pieces actually don’t do as well on Facebook because that’s not really what the audience is looking for all the time.” According to Uri, a social media editor at Haaretz, political content by the newspaper would not get a lot of likes, but could get many comments because people are agitated about it and would want to react.
David, a director at Haaretz, indicated that political content could also prompt comments attacking the newspaper for being leftist.
But prompting engagement with important political issues could be challenging. For example, Betty, a member of the audience development team at Slate, shared that the outlet had posted a story related to Special Counsel Robert Mueller’s hearing in U.S. Congress regarding Russian interference in the 2016 U.S. presidential election. Although the story was
“a really great piece, very timely, very solid reporting,” it did not perform well at first. But after changing the headline, the story did much better. As Betty suggested, the revised headline gave audience members the impression that there was something new or different about the story, and such an impression could help entice engagement:
We tried a different headline and finally that kind of cracked it, and then it started
picking up and people were reading it, and it was just an unexpected angle… I think
that got people interested because it signaled that it wasn’t just the same old same old
of this story that people have been hearing about for such a long time and it’s like
saturated the media.
Research has suggested that a distinct attribute of emailed news articles is their useful character (Boczkowski & Mitchelstein, 2012, 2013). Health stories can be seen as useful and
135
meaningful. Figure 7 shows a health related post from the examined sample (April 17, 2019).
The post by Vox contained a link to a piece on “8 things everybody should know about measles” and warned that unvaccinated people could easily get the virus. This post was highly shared by users.
Figure 7
A Highly Shared Facebook Post About Health
Although the examined Israeli posts did not yield a statistically significant
association between health stories and sharing, interviewees at Walla gave health-related stories as examples of engaging stories. One was a video about vaccinating children.
According to senior editor Yonatan, it generated engagement “at insane numbers – around
18,000 comments, shares, and likes, which are a lot.” Ariel also mentioned that content about a strike of medical assistants would be highly shared because this community could mobilize to spread the message.
136
The proportion of health-related content in the examined posts by U.S. and Israeli news organizations was not high to begin with, but in view of the performance of such posts in the U.S. (as presented earlier) and Sweden (Larsson, 2018) news organizations would do well to consider posting more health-related messages—not only during a global pandemic
(during which these lines are written).
Stories about crime on a news website were previously found to be more prevalent among heavily clicked items than among highly commented-on items. This finding suggested that users were more likely to read crime stories than to discuss them. They may have been curious to read about a crime, but did not necessarily feel they needed to convince people that committing a crime was wrong (Tenenboim & Cohen, 2015). This may be one explanation for the negative associations reported here between crime stories and modes of engagement on Facebook.
This study focuses on posts published by news organizations on their main Facebook pages, which can be seen as the organizations’ flagship spaces on Facebook. Yet, certain organization also had separate pages for specific sections in the news outlet not captured here. For example, USA TODAY had the Facebook pages USA TODAY Money and Tech,
USA TODAY Sports, and USA TODAY Life (“to talk all things entertainment”). Although the examined news organizations posted content from different sections of their outlet on their main Facebook page, the existence of more than one Facebook page may help explain why certain types of posts were less likely to entice engagement on the organizations’ main
Facebook page.
Lisa, an audience editor at USA TODAY, raised the possibility that users who were interested in sports got news from the organization’ accounts focusing on sports:
Sports content on our main account does not do well. We don’t really have an idea
why, but it does great on the sports platforms, and they’re pretty large… it’s almost
137
like that sports audience has migrated to the sports specific pages, and the general
news audience is not interested in sports. So every now and then we’ll share big
breaking news about a very well-known athlete, Serena Williams, LeBron James…
but we kind of scaled back on that.
As sports stories, content on arts / entertainment did not evoke a high volume of shares, but the latter content was positively associated with commenting and liking or reacting in Israel and with commenting among digital-native outlets. Engaging content about arts or entertainment focused, for example, on issues or people related to popular TV shows, and famous actors or singers. In some of the posts by Israeli outlets Yedioth Ahronoth and ynet, users were invited to tag other users who might be interested in the content.
Figure 8
A Highly Commented-On Facebook Post Related to Entertainment
Figure 8 shows a post reporting that American singer Lana Del Rey (or Elizabeth Grant) would visit Israel and perform in a music festival in the northern part of the country. Users
138
were invited to tag friends who would be happy to know about it. This demonstrates that comments can be used not only for expressing views or sharing information, but also for tagging users.
News values
News-related content has other characteristics in addition to topics. I now focus on the following news factors: conflict, surprise, prominent figure, reference to persons, impact, damage, success, and proximity (Galtung & Ruge, 1965; Harcup & O’Neill, 2017;
Tenenboim & Cohen, 2015; Weber, 2014; Ziegele et al., 2018a, 2018b). These factors are not mutually exclusive. Posts mentioning or referring to a conflict—a disagreement between parties, individuals, groups, or countries—accounted for 26.3% of the posts under study (420 of 1,600 posts). Posts containing an element of surprise or something unusual were 21% of the sample. A prominent figure, such as an actress or a president, was mentioned or referred to in 49.8% of the posts. Posts mentioning persons’ names or showing their faces accounted for 61.9% of the posts under study. Posts addressing an event or issue that had consequences for one or more social categories or groups (“impact”), such as nations, communities, and students, were 45.3% of the sample. Posts referring to a negative consequence of an event
(“damage”), such as a death, injury, illness, or material damage, accounted for 25.9% of the posts under study. Posts coded as referring to a positive consequence of an event (“success”), such as an improvement of the standards of living, breakthrough in health research, or raising money to help the elderly, were 4.3% of the sample. In addition, 70% of the posts in the U.S. sample (560 of 800) were coded as addressing an occurrence or occurrences in the U.S. only
(“proximity”), and 63.2 of the posts in the Israeli sample (506 of 800) were coded as addressing an occurrence or occurrences in Israel only (See Appendix D for detailed frequencies of news factors, including frequencies in the U.S. and Israel).
139
Table 8 shows how each of the above-mentioned news factors is related to each of the examined modes of engagement (commenting, sharing, liking, and reacting) after controlling for the number of likes a given Facebook page had at the time of posting. H2 posited that conflict would be positively associated with commenting, while surprise would be positively associated with sharing. Conflict was positively associated with commenting (p < .001), as well as with other modes of engagement. Focusing on conflict in the general sample, the IRR was particularly high in the regression for commenting: 2.348. This IRR suggests that posts that mentioned or referred to a conflict were commented-on 134.8% more often than non- conflict posts. The former posts were shared 126.1% more often than the latter posts. Surprise was positively associated with sharing (p < .001), as well as with other modes of engagement.
Focusing on surprise in the general sample, posts containing an element of surprise or something unusual were shared 190.9% more often than other posts. The former posts were commented-on 84.2% more often than the latter posts. Because conflict was positively associated with both commenting and sharing, and surprise was positively associated with both sharing and commenting, H2 is partially supported.
Table 8 also shows positive associations between other news factors and modes of engagement. Mentioning a prominent figure or referring to such a figure was positively associated with commenting, sharing, liking, and reacting. The IRRs were particularly high in the regressions for commenting and liking. Focusing on the general sample, posts mentioning a prominent figure or referring to such a figure were commented-on 105.8% more often than posts without a prominent figure. The former posts were liked 93% more often that the latter posts. References to persons (not necessarily prominent figures) yielded positive associations as well, except for sharing in the U.S. Furthermore, in each of the examined country’s posts addressing an occurrence in that country only were more likely to be engaged than other
140
posts. IRRs were particularly high in the U.S. sample. For example, posts addressing an occurrence in the U.S. only were shared 149.9% more often than other posts.
Table 8
Relationships Between News Factors and Modes of Engagement for Media Outlets’ Posts
News Factor Commenting Sharing Liking Reacting (IRR) (IRR) (IRR) (IRR)
Conflict General 2.348*** 2.261*** 1.607*** 1.828*** U.S. 2.553*** 1.413* 1.471** 1.689*** Israel 2.149*** 3.490*** 1.770*** 1.991***
Surprise General 1.842*** 2.909*** 1.391** 1.879*** U.S. 1.957*** 2.996*** 1.373* 2.047*** Israel 1.671*** 2.767*** 1.415** 1.659***
Prominent Figure General 2.058*** 1.487*** 1.930*** 1.867*** U.S. 2.051*** 1.316* 1.919*** 1.937*** Israel 2.077*** 1.699*** 1.933*** 1.791***
Reference to Persons General 1.998*** 1.456*** 1.840*** 1.891*** U.S. 1.864*** 1.095 1.620*** 1.726*** Israel 2.148*** 1.982*** 2.080*** 2.067***
Impact General 1.272** 1.577*** 1.102 1.193* U.S. 1.877*** 2.486*** 1.660*** 1.805*** Israel .857 .992 .721*** .770**
Damage General .894 1.786*** .742** 1.093 U.S. 1.072 2.508*** .803 1.240 Israel .712** 1.202 .679*** .946
Success General .396*** .625* 1.551* 1.190 U.S. .525* .529 1.712 1.228 Israel .278*** .728 1.399 1.157
Proximity General n/a n/a n/a n/a U.S. 2.399*** 2.499*** 2.211*** 2.465*** Israel 1.300* 1.733*** 1.240* 1.381** *** p < .001; ** p < .01; * p < .05; Note. Cells report incidence rate ratios (IRR) from negative binominal regressions. IRR > 1 indicates a positive effect while IRR < 1 indicates a negative effect. Bivariate associations were examined while controlling for the number of likes a Facebook page had at the time of posting. Proximity in the U.S. sample refers here to posts addressing an occurrence in the U.S. only, while proximity in the Israeli sample refers here to posts addressing an occurrence in Israel only. The general sample included 1,600 posts, and each of the subsamples (U.S. and Israel) included 800 posts.
141
The findings regarding impact, success, and damage were not as consistent as the above-mentioned findings. Posts addressing an event or issue that had consequences for one or more social categories or groups (“impact”) were positively associated with modes of engagement in the U.S. sample. But the associations between impact and engagement in the
Israeli sample were negative for liking and reacting (p < .001 or p < .01) and nonsignificant for commenting and sharing. Posts referring to success were less commented-on than success posts in the general sample and in both the U.S. and Israel (p < .001 or p < .05), though associations between success and other modes of engagement were not statistically significant in the U.S. and Israel. Posts referring to damage were more likely to be shared than non-damage posts in the general sample and in the U.S. The former posts were less likely to be liked than the latter posts in the general sample and in Israel. One type of damage is death, which was separately examined as well. Posts mentioning a death of at least one person were positively associated with reacting in the general sample (IRR = 1.501, p < .01) and in the U.S. sample (IRR = 1.762, p < .01), whereas the associations between death and commenting or sharing were similar to those found for damage.
In addition, some differences were found between newspapers’ posts (n = 1,081) and posts by digital-native outlets (n = 519). Impact was positively associated with commenting
(IRR = 1.389, p < .01) and with sharing (IRR = 1.839, p < .001) among posts by newspapers, but the relationships were not statistically significant among posts by digital-native outlets.
Success was negatively associated with commenting (IRR = .347, p < .001) and positively associated with liking (IRR = 1.606, p < .01) among posts by newspapers, but the relationships were not statistically significant among posts by digital-native outlets.
Furthermore, damage was positively associated with reacting (IRR = 1.390, p < .05) for digital-native outlets, whereas the relationship between damage and reacting was not statistically significant for newspapers.
142
Research has shown that controversial issues or references to a conflict can be linked to commenting on news websites (Boczkowski & Mitchelstein, 2012; Tenenboim & Cohen,
2015, but the findings here suggest that controversial issues or references to a conflict are not necessarily limited to prompting one mode of engagement on Facebook. Engaging content in the sample referred to issues such as abortion restrictions (in certain states in the U.S.), nominating to the U.S. Supreme Court a judge who faced accusations of sexual assault or misconduct (Brett Kavanaugh), election to the Israeli parliament, and an initiative to demolish the house of a suspected terrorist in the West Bank.
Interviewees also suggested that controversies could prompt user engagement.
For example, Nancy, a director at USA TODAY, said:
I think anything that’s controversial we’ve seen is definitely going to get higher
comments, so anything that has to do with politics—I shouldn’t say anything but a lot
of things that have to do with politics—the abortion debate… I can pretty much
guarantee that it’s a safe bet that we’re going to see lots of comments on things like
that.
According to Itamar, one of those who ran the Facebook page of the conservative newspaper Makor Rishon, the struggle between the political right and left is fierce in Israel.
“The (Israeli) people is divided. We’re not trying to divide, but when issues arise and there are attacks against the right-wing block… it arouses public interest,” he said. Posting controversial content can also be an opportunity to ask audience members what they think about an issue. As Uri, of the liberal newspaper Haaretz, put it:
If there’s something that is very controversial and I think that if I tell people to
express their opinion there’s a chance they will… I (will) do that. It just has to be
something controversial. I won’t tell people to express an opinion about… ‘There was
143
a murder of a woman in Nazareth (an Israeli city). What do you think?’ It’s clear that
I won’t do that.
In a similar vein, Dvir, of ynet’s new media team, said that team members sometimes encouraged discussions on Facebook by posing a question. For example, if the Israeli parliament rejects a bill related to LGTBQ rights, a post on ynet’s Facebook page can report this development and ask: “What do you think?” At Walla, a competing digital-native outlet,
Ariel shared that a week before his meeting with me an opinion by an Arab-Israeli member of parliament had been published on the outlet’s Facebook page. The lawmaker was critical of the Israeli government, and a response by a right-wing Jewish-Israeli lawmaker was expected to be published. “I know it will spark controversy, and I know it will work. And I have no problem with that,” Ariel said.
News organizations’ content on controversial issues can also elicit shares. “If the person sides with it, they’ll share it. If they don’t side with it, they’ll share it and say
‘Why?’”, according to Lisa, of USA TODAY. Ido, of Israel Hayom, further explained:
When someone says something that a lot of people agree with, and not just the
obvious… something that maybe they were not exactly able to word this way, or they
didn’t see what they were thinking in this way, they’d want to share it… because this
is what they’d want to have (on their timeline / wall).
Content containing a surprising element or something unusual can be shareable as well. For example, although most people may safely get from one place to another most of time, car accidents are not considered unusual in this study. But if an entire family is killed in a car accident, it is considered unusual in this study. If rockets are launched from the Gaza
Strip to southern Israel, it is not surprising or unusual. But if there is a rocket alert in another part of Israel, it can be considered surprising or unusual.
Figure 9 shows a shareable post that was coded as containing a surprising element:
144
U.S. actor Jussie Smollett was arrested on charges of filing a false police report weeks after claiming he was the victim of a racist and homophobic attack on a downtown street in
Chicago.
As indicated at the bottom of the post, it also generated “wow” faces from users. This
“wow” reaction is in line with the first words of Itai, a director at Yedioth Ahronoth, when
I asked him which type of content evokes shares. “Wow, wow content… something I’ve
never seen before,” he said. Such content can be about a beautiful lake or a strong human experience, Itai suggested. His examples and the post above point to the importance of additional news factors: reference to persons, and prominent figures.
Figure 9
A Facebook Post With a Surprising or Unusual Element
Prominent figures mentioned in the examined posts were politicians, artists, media actors, and others. Such figures were also mentioned by interviewees who spoke with me.
According to Eitan, of Haaretz, content related to Prime Minister Netanyahu and his wife, or content about certain Israeli women in politics (e.g., Miri Regev and Ayelet Shaked), was likely to generate interest and discourse. Eitan added that people wanted to hear what specific
145
writers (e.g. Yossi Verter, Raviv Druker, and Uri Misgav) had to say. For example, Raviv
Druker is a TV journalist who also writes for Haaretz. Whatever he writes usually performs well on Facebook, Eitan said. In a similar vein, Eyal, of Makor Rishon, indicated that a woman politician (Ayelet Shaked who was Israel’s minister of justice at the time) and a religious TV journalist (Amit Segal) were popular among readers. In the U.S., posts by The
Daily Caller about certain women in politics, such as House Speaker Nancy Pelosi and
Representative Alexandria Ocasio-Cortez, were also engaging. At The Dallas Morning News,
Linda shared that a highly engaging post the day before she met with me focused on actor
Matthew McConaughey who was appointed as a professor of practice at the Moody College of Communication at UT Austin. “The engagement on that was ridiculous. It was insane. So good,” Linda said.
According to Michael, of South Florida Sun-Sentinel, a celebrity name is one of the things that can help make a story engaging. As he put it:
The news stories that people really like to talk about will get more comments and
shares… it’ll be a celebrity name that pops up in the story, or it’ll just be something
that impacts everyone… that everyone can relate to. So those very relatable types of
stories, I think, get shared a lot more and retweeted a lot more.
Engaging stories can also focus on people who were not well known, such as people who survived an ordeal, overcame hardships, or broke a record. “We know that human stories are largely the fuel of Facebook,” said Eitan, of Haaretz. According to Itai, of Yedioth Ahronoth,
“people love heroes. It could be an Olympic champion; it could be someone overcoming a disability; it could be a soldier.” Itai further said that stories on such heroes were posted on
Yedioth Ahronoth’s Facebook page whenever it was possible, suggesting that they performed well.
146
Figure 10
An Engaging Facebook Post Focusing on a Woman Giving Birth
Stories about people’s ordeals and accomplishments were also posted by other news organizations. Figure 10 shows a post by Walla that tells the story of Shahar Levengrond-
Yehezkel who lost her sister in a terrorist attack and did not know how she would continue to live. But she did continue and gave birth to a baby boy. The post generated a relatively high volume of shares and reactions.
Whereas the topics and news factors analyzed in this section are seen as reflecting older media logics in that journalists have long made news judgements based on them, the following section centers on attributes that reflect newer media logics.
Newer media logics: Evaluations, emotions, and structural features
Although news topics and values play an important role in journalistic selection and construction of content, they may not be enough to draw people’s attention and entice their engagement in a media environment where content far exceeds what people can consume. In this environment, the boundaries between professional journalists and everyday citizens as well as between public and private are more blurred, giving rise to a more subjective or
147
emotional expression. Subjectivity and emotionality are not new in journalism, but they are particularly prominent and may be a vital force in enabling newer forms of engagement on digital platforms, such as social networking sites (Wahl-Jorgensen, 2016). Such sites can contain affective news streams (Papacharissi & Oliveira, 2012), which describe “how news is collaboratively constructed out of subjective experience, opinion, and emotion within an ambient news environment” (p. 279). In addition, news-related messages on these sites can be multimedia messages, including written texts, images, videos, and links. Multimedia features and affective aspects of messages are seen here as reflecting newer media logics. In what follows, I address these features and aspects and show their associations with modes of engagement. The tables showing the associations focus on the U.S. sample, Israeli sample, and combined sample. If notable differences between newspapers and digital-native outlets were found, they are mentioned in the text.
I first focus on evaluative aspects, including evaluative language, criticism, emotional language, and tone, in the Facebook posts under study. Posts using “evaluative language or… unattributed commentary beyond the fact of an occurrence or issue” (Lawrence et al., 2014, p. 10) accounted for 48.8% of the examined posts (781 of 1,600 posts). Posts containing criticism—“evaluative language that expresses disapproval of someone or something”
(Tenenboim, 2017, p. 3505–3506) —were 24% of the sample. Posts using emotional language, that is, posts explicitly referring to a human emotion (Ziegele et al., 2018b), accounted for 17.4% of the sample. An overall positive tone was found in 15.9% of the posts, compared to 14.3% negative and 51.9% neutral (17.8% of the posts were coded as “hard to determine”). Whereas Posts were coded as containing evaluative language if this language was unattributed, posts were coded as containing criticism or emotional language whether these were attributed or not.
148
Table 9 shows how each of the above-mentioned evaluative aspects is related to each of the examined modes of engagement (commenting, sharing, liking, and reacting) after controlling for the number of likes a given Facebook page had at the time of posting.
Consistent results across countries and modes of engagement were found regarding criticism:
Posts containing criticism were more likely to be commented-on, shared, and liked or reacted-to than posts without criticism in the general sample and in both the U.S. and Israel.
Table 9
Relationships Between Evaluative Aspects and Modes of Engagement for Media Outlets’ Posts
Evaluative Aspect Commenting Sharing Liking Reacting (IRR) (IRR) (IRR) (IRR)
Evaluative Language General .943 .846 1.190* 1.051 U.S. .728* .641** .953 .822 Israel 1.134 1.065 1.416*** 1.273*
Criticism General 2.629*** 2.840*** 1.564*** 1.823*** U.S. 3.145*** 1.964*** 1.530** 1.839*** Israel 2.072*** 4.067*** 1.600*** 1.802***
Emotional Language General 1.557*** 1.766*** 1.673*** 1.613*** U.S. 1.702** 1.105 1.329 1.296 Israel 1.446** 2.277*** 1.889*** 1.809***
Positive Tone General .756* .721* 1.780*** 1.235* U.S. .394*** .506** 1.567** .984 Israel 1.202 .967 1.978*** 1.470**
Negative Tone General 1.421** 1.812*** .839 1.101 U.S. 1.468 1.357 .773 1.116 Israel 1.380* 2.202*** .887 1.093 *** p < .001; ** p < .01; * p < .05; Note. Cells report incidence rate ratios (IRR) from negative binominal regressions. Bivariate associations were examined while controlling for the number of likes a Facebook page had at the time of posting. Posts were coded as containing evaluative language if this language was unattributed. Posts were coded as containing criticism or emotional language whether these were attributed or not. The general sample included 1,600 posts, and each of the subsamples (U.S. and Israel) included 800 posts.
For example, focusing on the general sample, posts containing criticism were shared
184% more often than posts without criticism, and the former posts were commented-on
149
162.9% more often than the latter posts. The IRR was the lowest in the regression for liking, though posts containing criticism were liked more often than posts without criticism.
Certain consistent patterns were also found regarding emotional language. First, posts explicitly referring to a human emotion were commented-on more often than posts that did not explicitly refer to a human emotion in the general sample and in both the U.S. and Israel.
In the general sample, the former posts were commented-on 55.7% more often than the latter posts. Second, posts containing emotional language were positively associated with each mode of engagement in the general sample and in Israel. In the Israeli sample, the IRR was particularly high in the regression for liking: Posts using emotional language were shared
127.7% more often than posts without emotional language.
Furthermore, a consistent pattern was found regarding a positive tone and liking:
Posts where their overall tone was identified as positive were liked more often than posts that their overall tone was not identified as positive in the general sample and in both the U.S. and
Israel. Posts that their overall tone was negative were commented-on and shared more often than posts that their overall tone was not identified as negative in the general sample and
Israel.
For evaluative language, most of the examined associations did not yield statistically significant results. However, in the U.S., posts containing an (unattributed) evaluative language were commented-on and shared less often than posts without this language (p < .01 or p < .05). In Israel, posts using an (unattributed) evaluative language were liked or reacted- to more often than posts without this language (p < .001 or p < .05).
In addition, some notable differences were found between newspapers’ posts (n =
1,081) and posts by digital-native outlets (n = 519). An overall negative tone was positively associated with commenting (IRR = 1.455, p < .05) and with sharing (IRR = 2.392, p < .001) among posts by newspapers, but the relationships were not statistically significant among
150
posts by digital-native outlets. An evaluative language was negatively associated with sharing
(IRR = .700, p < .05) among posts by newspapers, but the relationship was not statistically significant among posts by digital-native outlets. However, an emotional language was positively associated with sharing (IRR = 3.217, p < .001) among posts by digital-native outlets, but the relationship was not statistically significant among posts by newspapers.
The findings regarding criticism and tone suggest that expressions of disapproval can be engaging and a positive valence may contribute to likes. The examined posts contained various expressions of disapproval, such as “enemy of the people” (Trump’s words against the press), “completely inappropriate” (words by CNN legal analyst about comments made by U.S. Attorney General William Bar), “disgusting” and a “disgrace” (what a Florida county sheriff said about comments by Florida’s Democratic nominee for governor), “damn the man who gave you a position in government” (what a radio host said regarding the minister of culture in Israel), and “so much for free speech!” (a comment about the suspension of a person’s Twitter account). A positive tone was identified in posts containing words such as “a true patriot” (referring to decorated sniper Chris Kyle who was murdered in 2013), “simply amazing” (referring to a bridge in Vietnam), “a stunning Christmas gift” (paying off parents’ debt), “welcome to the world” (after the birth of a child), and “it’s perfect” (about a cookie).
However, the findings regarding the use of evaluative language suggest that the use of unattributed subjective expressions is not necessarily engaging in itself. It was even negatively associated with commenting and sharing among the examined U.S. posts. As indicated in the chapter on logics of media production in the U.S. and Israel, U.S. news organizations—at least mainstream media organizations—were generally more cautious about staff opining on their social media accounts. Indeed, according to the content analysis,
42.4% of the U.S. examined posts contained an (unattributed) evaluative language, compared
151
to 55.3% of the Israeli posts. Referring to such language and to the overall tone used by U.S. organizations, Jennifer, of Los Angeles Times, remarked:
I would say that most of the bigger, legacy newsroom social platforms are very
similar in terms of tone. I mean, LA Times, Washington Post, New York Times, The
Wall Street Journal, things like that. I think, specifically on the main social accounts,
there is a very unbiased, straight-to-the-point language.
Other interviewees offered similar accounts regarding the examined U.S. newspapers.
Linda, of The Dallas Morning News, indicated that posts on the main Facebook page of her newspaper typically included texts pulled from the stories without adding to them subjective expressions. Referring to the style of expression or presentation used by newspapers such as
The New York Times and Los Angeles Times, she said: “I didn’t see a ton of differences.”
Michael, of South Florida Sun-Sentinel, thought that his newspaper’s style of expression on
Facebook was similar to other newspapers but different from what TV stations did. Focusing on such stations in the South Florida market, he said:
We see a lot more kind of sensational posting using lots of exaggerated words and
lots of emojis and lots of graphics and visuals that we would not use. To us, it’s kind
of the difference between TV news and print news. It’s just really pronounced on
social media. So although the TV news social channels are bigger channels and get
more engagement than us, we also feel like they frequently cross journalistic lines
we’re not willing to cross… their conversations with readers are much more
sensational… shareable, as a result, but just have a tone that we don’t really choose
to take.
As shown in the chapter on logics of media production, social media editors in U.S. newspapers did think that being “a bit playful” on Facebook was OK. But digital-native outlets seemed to allow themselves more than that. For example, Betty, of Slate, said: “Slate
152
does try to have a clear distinctive voice and brings that to social media, and doesn’t really just try to be dry.” A qualitative look at posts by digital-native outlets in the examined sample was in line with this statement. Indeed, the quantitative content analysis suggests that 65.1% of the posts by such outlets contained unattributed evaluative language, compared to 41% of the posts by newspapers. That said, as discussed in the chapter on logics of media production, certain Israeli newspapers allowed themselves more than the examined U.S. newspapers.
The quantitative findings also suggest that referring to a human emotion in Facebook posts can be engaging. As an illustration, The Daily Caller posted the words “He’s mad,” along with a link to the story “Trump Really Wants Judge Jeanine Back On Fox News”
(March 17, 2019). The New York Times posted that “Times photographers look back at a terrifying and tragic time,” along with a link to snapshots from the first decade of the AIDS crisis (December 1, 2018). Yedioth Ahronoth posted about a model who experienced years of bulimia and “self-hatred” (May 22, 2019). And ynet posted the word “heartbreaking” regarding a baby who was buried among her parents after an accident (October 19, 2018).
These expressions demonstrate the importance of emotions on social networking sites. In what follows, I discuss posts that attempt to evoke anger, sadness, and anxiety.
Emotions
Using Berger and Milkman’s approach (2012), coders were presented with definitions of the examined emotions and were asked to code whether a given post evoked each emotion.
Posts evoking anger—“a strong feeling of annoyance, displeasure, or hostility” (Oxford
Dictionary, n.d. 1)—accounted for 21.2% of the sample (339 of 1,600 posts). Posts evoking sadness—“the condition or quality of being sad” Oxford Dictionary, n.d. 2)—were 18.6% of the sample. Posts evoking anxiety—“a feeling of worry, nervousness, or unease about something with an uncertain outcome” (Oxford Dictionary, n.d. 3)—accounted for 4.6% of the sample.
153
Table 10 shows how each emotion is related to each of the examined modes of engagement after controlling for the number of likes a given Facebook page had at the time of posting. H3 posited that anger would be positively associated with sharing and reacting.
Indeed, anger was positively associated with each of these modes of engagement. Focusing on the general sample, posts evoking anger were shared 174.3% more often than posts that did not evoke anger. The former posts were reacted-to 65.7% more often than the latter posts.
These patterns were found in both the U.S. and the Israeli samples. Therefore, H3 was supported. In addition, the analysis suggests that posts evoking anger were positively associated with commenting. Posts evoking anger in the general sample were commented-on
177.6% more often than posts that did not evoke anger.
Table 10
Relationships Between Emotions and Modes of Engagement for News Organizations’ Posts
Emotion Commenting Sharing Liking Reacting (IRR) (IRR) (IRR) (IRR)
Anger General 2.776*** 2.743*** 1.128 1.657*** U.S. 3.510*** 2.314*** 1.119 1.697*** Israel 2.020*** 3.313*** 1.140 1.607***
Sadness General 1.085 2.742*** .958 1.547*** U.S. 1.277 3.538*** 1.081 1.739*** Israel .872 1.936*** .809 1.313*
Anxiety General .926 2.290*** .561** .793 U.S. 1.331 3.739*** .833 1.335 Israel .612* 1.144 .352*** .384*** *** p < .001; ** p < .01; * p < .05; Note. Cells report incidence rate ratios (IRR) from negative binominal regressions. Bivariate associations were examined while controlling for the number of likes a Facebook page had at the time of posting. The reported emotions are the ones studied by Berger & Milkman (2012). Another emotion they studied—awe—is not presented here because the inter-coder reliability for this variable was relatively low. In addition, the frequency of awe in the sample was particularly low. The general sample included 1,600 posts, and each of the subsamples (U.S. and Israel) included 800 posts.
Sadness was positively associated with two modes of engagement—sharing and reacting—in the general sample and in both the U.S. and Israel. Focusing on the general
154
sample, posts evoking sadness were shared 174.2% more often than posts that did not evoke sadness. The former posts were reacted-to 54.7% more often than the latter posts. However, the findings regarding anxiety were less consistent. Anxiety was positively associated with sharing in the general sample and in the U.S., but not in Israel. It was also negatively associated with liking in the general sample and in Israel, but not in the U.S. Furthermore, anxiety was negatively associated with liking among posts by newspapers (IRR = .545, p <
.01), but the relationship was not statistically significant among posts by digital-native outlets.
Posts evoking anger were found for various issues, from whale hunting to disrespectful comments about women. Figure 11 shows a post by ynet with the words “The world is furious: after more than 30 years – Japan is renewing the whale hunting” and a photo of whales in bloody water (December 26, 2018). This post was highly engaging.
Figure 11
A Facebook Post Evoking Anger
A different example is from the U.S.: Slate posted a link to a story about Fox News host
Tucker Carlson who “says awful things about women in newly unearthed radio clips.” “The
155
Fox News host also has some things to say about statutory rape,” Slate added (March 10,
2019). The story itself reported that “highlights include the time Carlson seemed to give a nod to fantasizing about girls who were the same age as his 14-year-old daughter in boarding school. ‘If it weren’t my daughter I would love that scenario,’ he said” (Politi, 2019, para. 2).
Whether audience members opposed Carlson and his views or disapproved how Slate portrayed him, this content could trigger anger.
Posts evoking sadness were likely to address tragic events. For example, Yedioth
Ahronoth posted that an Israeli soldier was stabbed to death. The post included the words
“May his memory be blessed” (March 17, 2019). USA TODAY posted that Tina Turner’s firstborn son died of an apparent suicide (July 3, 2018). The post is shown in Figure 12.
Figure 12
A Facebook Post Evoking Sadness
Posts evoking anxiety addressed different issues. For instance, South Florida Sun-
Sentinel posted that U.S. mining sites dumped millions of gallons of toxic waste into drinking water sources (February 20, 2020), and Haaretz posted a link to text about identifying psychopaths around us (September 28, 2018).
156
The importance of emotionality on Facebook was also expressed in interviews with media professionals, particularly in Israel. According to Eitan, of Haaretz, certain issues or elements make people’s emotions “run wild,” and where emotions run wild – the engagement is greater. “We do want people to have some kind of a point of sympathy or anger or something like that… so that they’ll want to share or comment or like. Or sometimes there will even be motivations for action,” Eitan said. Itai, of Yedioth Ahronoth, suggested that
“emotional content, but something you don’t see every day” could be engaging. As assessed by Yair, of ynet, posts that generate a certain emotion or have the potential to move people or make them angry tend to perform better. According to Omer, of Walla, content tends to be highly shared when it can serve as a “campfire.” “It can be a collective grief; it can be a tragedy; it can be a collective joy,” he offered. Indeed, Moshe, of Maariv, was surprised to see how well a sad story had performed among users. The story, which appeared on Maariv’s
Facebook page, was about a former (Israeli) beauty queen who passed away. It suggested that different emotions may be linked to engagement.
Ariel, of Walla, further talked about the centrality of emotional elicitation in creating an item for social media. He wanted the user to love the item, to hate it, to burst out laughing, to cry, to be angry, to be moved, so that an action would be taken—“from a like, comment, share, to downloading the video and sending it to people on WhatsApp and so on.” Ariel explained that on Facebook “people operate with emotion and are there for that. This is not a municipal library. It won’t be a municipal library. This is how Facebook was built… It works
(based on) emotion. There I am too; this is also what usually motivates most of us.”
Although U.S. interviewees were generally less likely to explicitly mention emotionality, some interviewees did acknowledge the role it played. Mary, of Los Angeles
Times, said that stories with “an emotional sort of aspect” were “geared toward social media and getting that response and comments from users.” Lisa, of USA TODAY, referred to the
157
emojis Facebook users could use in reaction to posts. She also said that shares and comments were even more desired by her audience team:
We’re looking at the type of reaction. Are they angry? Are they sad? Are they happy?
…How does this make them feel? And we also focus on comments and shares more
than we focus on the reactions. A like is so easy to get, right? If someone takes the
time to actually comment, or decides it’s such a good piece of content that they want
to share it, that is a more meaningful social interaction, and that’s actually something
Facebook has to turn into.
Structural features
In an attempt to prompt engagement, news organizations also create multimedia messages. Structural features of digital messages are understood here as a video, a photo, a link, and a written text. A video was embedded into 6.1% of the examined posts. At least one photo (including as part of an embedded link) was identified in 84.6% of the posts. An embedded link to a webpage was found in 69.1% of the posts. Among the posts with an embedded link, 55.4% were to a news report, while the remaining posts were to an opinion piece, commentary, magazine feature, or other materials. In terms of written text, 68.3% of the examined posts (the general sample) contained a sentence or less than a sentence, compared to 24.3% with more than a sentence and not more than a paragraph, 4.9% with more than one paragraph, and 2.5% without a written text.
Table 11 shows the relationships between the examined features and the different modes of engagement after controlling for the number of likes a given Facebook page had at the time of posting. The presence of a video was positively associated with commenting and sharing in the general sample and in both the U.S. and Israel. Posts containing a video in the general sample were commented-on 183.5% more often than non-video posts. The former posts were shared 153.5% more often than the latter posts. In addition, the presence of a
158
video was positively associated with liking and reacting in the general sample and in Israel, but not in the U.S. A follow-up analysis suggests that the presence of a video was positively associated with sharing (IRR = 4.278, p < .001), liking (IRR = 1.946, p < .01), and reacting
(IRR = 1.902, p < .01) among posts by digital-native outlets, but these relationships were not statistically significant among posts by newspapers.
Table 11
Relationships Between Structural Features and Modes of Engagement for Media Outlets’ Posts
Structural Feature Commenting Sharing Liking Reacting (IRR) (IRR) (IRR) (IRR)
Video General 2.835*** 2.535*** 1.459* 1.404* U.S. 2.765*** 2.101* 1.197 1.138 Israel 2.845*** 2.869*** 1.687** 1.627*
Photo General .554*** .601** .864 .888 U.S. .650* 1.294 1.329 1.532* Israel .506*** .425*** .687** .680**
Embedded Link General .590*** .821 .708* .743* U.S. .428** .580* .917 .998 Israel .851 1.097 .605** .619**
Embedded Link to a General 2.069*** 2.252*** 1.638*** 2.178*** News Report U.S 2.096*** 2.467*** 1.773*** 2.422*** Israel 2.008*** 1.937*** 1.402** 1.778***
Length – More Than a General 1.574* 2.151** 1.861** 1.663** Paragraph U.S. 1.439 1.500 1.289 1.272 Israel 1.720* 3.092*** 2.628*** 2.188*** *** p < .001; ** p < .01; * p < .05; Note. Cells report incidence rate ratios (IRR) from negative binominal regressions. Bivariate associations were examined while controlling for the number of likes a Facebook page had at the time of posting. Embedded link to a news report focuses on the type of text to which a post linked – a news report rather than other types of texts (a magazine feature, a commentary or opinion piece, or other). Embedded link to a news report was examined among 1,108 posts in the general sample (those that included an embedded link), 746 posts in the U.S. subsample, and 362 in the Israeli subsample. The other variables were examined in the general sample that included 1,600 posts and in the subsamples (U.S. and Israel) that included 800 posts each.
The findings regarding the presence of at least one photo and an embedded link were less consistent. The presence of a photo was negatively associated with commenting in the general sample and in both the U.S. and Israel. It was also negatively associated with other
159
modes of engagement in Israel. However, these findings should be interpreted with caution because the vast majority of the examined posts included a photo (this seemed to be the standard in constructing a post). Posts that were not coded as containing a photo either included a video instead, or did not display a photo or video at the time of coding. The presence of an embedded link also yielded certain negative associations with modes of engagement. These findings should also be interpreted with caution because the vast majority of the examined pots in the U.S. included an embedded link (this seemed to be the standard in constructing a post). In Israel, certain news organizations were less likely to use this type of link.
A follow-up analysis suggests that among posts containing an embedded link, posts linking to a news report were positively associated with each of the modes of engagement under study in the general sample and in both the U.S. and Israeli. For example, posts with an embedded link to a news report were shared 125.2% more often than posts with an embedded link to another text or material.
In addition, posts with more than one written paragraph were more engaging than shorter posts or posts that did not contain a written text in the general sample and Israel. For example, in Israel the former posts were shared 209.2% more often than the latter posts.
Interviewees addressed multimedia aspects of posts their organizations produced. For example, Ashley, of Vox, said:
We do have such a robust video operation. Obviously other news organizations do
videos… we’re not the only people out here with videos, but it’s definitely part of our
strategy in that we have this wealth of really informative video content that we try to
share in the most effective, helpful ways.
According to Nancy, of USA TODAY, visual storytelling—through the use of videos, graphics, gifs, or other materials—was “part of the USA TODAY DNA” and “a really strong
160
piece of our social media presence.” She suggested that this might not be the priority of all news organizations. Linda, of The Dallas Morning News, shared that her organization had videographers who put together videos for social media platforms. Patricia, of Los Angeles
Times, said that her audience engagement team used videos, though it did not use as many live videos as it did before because Facebook’s algorithm deprioritized such videos. In Israel, the digital-native outlets ynet and Walla put more emphasis on videos than at least some of the examined newspapers.
As for written texts: Ashley, of Vox, acknowledged that long features to which posts linked did not always perform well. She explained:
Often the things that will take off will be like short news hits about whatever crazy
thing happened in the news today, or someone might write a take, an analysis on the
situation, that they might be short, quick news hits… that’s just the thing that people
are talking about, so it might catch fire. Whereas it could be a lot harder for us to find
our audience for an 8,000 word feature about whatever the topic is.
After separately analyzing and demonstrating relationships between different content characteristics and modes of engagement, it is also important to understand which content characteristics may be stronger predictors than others. Thus, in what follows I examine content characteristics in multivariate models.
Combining older and newer media logics
Messages by news organizations on social networking sites can reflect older as well as newer media logics rather than either older or newer media logics (See Chadwick, 2013 for “not only but also” type of thinking). In this final part of analyzing what in messages by news organizations is associated with different modes of engagement, components of older and newer media logics were included in the same models to examine which associations
161
exist when different components are considered and which components may be stronger predictors than others.
Negative binominal regression analyses were performed for commenting, sharing, liking, and reacting. The examined predictors were most of the content characteristics that were presented earlier (correlations among these predictors were < .700). Content characteristics were excluded if they belonged to one of the following categories: not yielding any statistically significant association when bivariate associations were examined (the topic society / welfare); appearing in almost all of the examined posts in the general sample (US and Israel) or in one of the countries, making it difficult to use the content characteristic as a predictor (e.g., the presence of at least one photo); being unique to one country (e.g., a focus on Israel only); addressing an aspect that is also represented by another variable (posts identified as a “human interest story” rather than as a business story, for example, tended to include the news factor “reference to persons”); and not generating a result due to a system’s error in computing the inverse LOG link function, the log likelihood, the gradient, or the
Hessian matrix in the iterative process for at least one of the records when bivariate associations were examined. This resulted in the inclusion of 24 content characteristics in the multivariate models.
Table 12 shows how the content characteristics under study are related to each mode of engagement after controlling for the number of likes a Facebook page had at the time of posting, as well as two additional variables: country (U.S. – yes / no) and organization
(newspaper – yes / no).
The table offers two key findings. First, certain content characteristics, such as a prominent figure and a surprise, were statistically significant predictors of each mode of engagement. Second, certain content characteristics were statistically significant predictors of one mode of engagement or more, but not all modes of engagement. For example, health
162
Table 12
Predicting Modes of Engagement With News Organizations’ Posts on Facebook: Multivariate Negative Binominal Models
Content Characteristic Commenting Sharing Liking Reacting (IRR) (IRR) (IRR) (IRR)
Topics Politics / Government 1.570*** 1.919*** 1.475** 1.598***
Military / Defense .772 1.565* 1.596** 1.586**
Economy / Business .868 .813 .857 .727
Crime .536*** .408*** .509*** .558***
Health 1.359 1.926** 1.344 1.223
Disasters / Accidents .817 .670 .822 1.081
Arts / Entertainment 1.363* .862 1.130 1.173
Sports .698 .470** .997 .864
News Conflict 1.208 .907 1.060 1.113 Values Surprise 1.983*** 2.632*** 1.664*** 2.039***
Prominent Figure 1.450** 1.339* 1.443** 1.445***
Reference to Persons 1.150 1.102 1.169 1.214*
Impact .809* 1.024 .981 .917
Damage 1.000 1.152 .877 .983
Success .662* 1.130 1.399 1.331
Evaluative Evaluative Language .944 .902 .900 .925 Aspects Criticism 1.423** 1.848*** 1.380** 1.285**
Emotional Language 1.299* 1.277* 1.428*** 1.348**
Tone (Positive) 1.251 1.791*** 2.050*** 1.737***
Emotions Anger 2.061*** 1.819*** 1.011 1.369**
Sadness 1.115 2.381*** 1.272* 1.683***
Anxiety 1.337 1.489* .776 .874
Structural Video 2.729*** 3.069*** 1.337* 1.299 Features Length (More Than a 1.010 1.294 1.332 1.189
Paragraph)
Control Number of Followers 1.000* 1.000** 1.000** 1.000**
Country (US) .502 3.051 .556 .735
Organization .602 .821 .514 .547 (Newspaper)
163
*** p < .001; ** p < .01; * p < .05; Note. Cells report incidence rate ratios (IRR) from negative binominal regressions for commenting, sharing, liking, and reacting. IRR > 1 indicates a positive effect while IRR < 1 indicates a negative effect. A mixed model negative binomial analysis has no direct analogue to R² (Coxe, West, & Aiken, 2009). The examined sample included 1,600 Facebook posts by U.S. and Israeli news organizations.
164
predicted sharing, but not other modes of engagement, and a positive tone predicted liking / reacting and sharing, but not commenting.
The following picture emerged about each mode of engagement. The characteristics positively associated with commenting were politics / government, arts / entertainment, surprise, prominent figure, criticism, emotional language, anger, and video. The characteristics negatively associated with commenting were crime, impact, and success.
Some of the strongest predictors of commenting were a video and anger. Posts containing a video were commented-on 172.9% more often than posts without it. Posts evoking anger were commented-on 106.1% more often than posts that were not identified as evoking anger.
The characteristics positively associated with sharing were politics / government, health, military / defense, surprise, prominent figure, criticism, emotional language, positive tone, anger, sadness, anxiety, video. The characteristics negatively associated with sharing were crime, and sports. Some of the strongest predictors of sharing were a video and surprise.
Posts containing a video were shared 206.9% more often than posts without a video. Posts with a surprising element or something unusual were shared 163.2% more often than posts without them.
The characteristics positively associated with liking were politics / government, military / defense, surprise, prominent figure, criticism, emotional language, positive tone, sadness, and video. However, crime was negatively associated with liking. One of the strongest predictors of liking was a positive tone. Posts with a positive tone were liked 105% more often than posts without a positive tone.
The characteristics positively associated with reacting were politics / government, military / defense, surprise, prominent figure, reference to persons, criticism, emotional language, positive tone, anger, and sadness. However, crime was negatively associated with reacting. One of the strongest predictors of reacting was a surprise. Posts containing a
165
surprising element or something unusual were reacted-to 103.9% more often than posts without them.
The findings help understand what in messages posted by the examined news organizations in triple-party news-spaces on a key platform can prompt user engagement.
Based on the findings, I suggest that according to logics of media usage underlying the triple- party news-spaces under study, particular news topics and factors, as well as particular affective aspects and structural features, can elicit different modes of engagement.
Summary
This chapter has investigated user engagement with messages by news organizations on Facebook. Specifically, it has examined to what extent and how news organizations’ posts on Facebook differ in the modes of engagement they generate (RQ2). I have shown the extent to which different engagement metrics correlate (RQ2a) and the degree of overlap among the most commented-on, shared, and liked/reacted posts (RQ2b). In addition, I have identified which content characteristics of the posts are associated with different modes of engagement
(RQ2c and H1-H3) and whether these characteristics differ across cultural/national and organizational contexts (RQ2d).
As explained earlier, different modes of user engagement can be regarded as different acts of civic engagement. Identifying which content characteristics prompt different acts of civic engagement is an important step in better understanding engaged citizenship (Dalton,
2016) in the digital media environment. Understanding what is commented-on, shared, or liked / reacted is also important because messages with greater engagement may be prioritized by the algorithm of a given social networking site and gain greater visibility. So even if many audience members do not take advantage of opportunities to comment, share, or like/react, those who do comment, share, or like/react can play a role in determining what would reach a larger audience.
166
Based on the analysis of news-related messages and the modes of engagement they generated, I propose three key principles of logics of media usage in triple-party news-spaces on Facebook. First, although highly commented-on posts, highly shared posts, and highly liked or reacted-to posts can overlap, they do not fully overlap. Each group of posts can differ from one another to a notable degree. As an illustration, focusing on the top 10% of commented-on posts and the top 10% of shared posts of a given news organization under study, about 40% to 60% of the most commented-on posts differed from the most shared posts. Second, certain content characteristics of posts are linked to each of the examined modes of engagement—commenting, sharing, or liking / reacting— in more than one cultural/national context. Third, certain content characteristics are linked to one or more modes of engagement, but not to all of them.
Related to the second and third principles, particular news topic and values (reflecting older media logics), as well as particular evaluative aspects, emotions’ triggers, and multimedia features (reflecting newer media logics) are associated with commenting, sharing, or liking/reacting. For example, posts that their main topic was politics were more likely to be commented-on, shared, and liked / reacted-to than posts that their main topic was not politics in the examined U.S. and Israeli contexts. Posts about health were more likely to be shared, particularly in the U.S. context, than non-health posts. But the former posts were not positively associated with other modes of engagement. One possible explanation is that news content deemed useful or meaningful is particularly likely to be circulated by users.
A reference to a prominent figure, such as a president or an actor, and a surprising element or something unusual, were positively associated with sharing, commenting, and liking / reacting among the examined posts by U.S. and Israeli organizations. Although conflict was positively associated with each mode of engagement when bivariate associations were examined, the associations were not statistically significant in the models with other
167
content characteristics. In the model for commenting, the association between conflict and commenting was marginally significant (p = .081). That said, stories about politics may involve controversial issues or conflicts. This may partly explain the results about conflict in the model with other predictors.
Criticism and emotional language were positively associated with commenting, sharing, and liking / reacting among the examined posts by U.S. and Israeli organizations.
Posts with an overall positive tone were more likely to be liked than posts without this tone in both the U.S. and Israel, but the former posts were not more likely to prompt comments.
Posts evoking anger were more likely to be commented-on, shared, and reacted-to than posts that were not identified as evoking anger. But anger was not associated with liking. Posts evoking sadness were more likely to be shared and liked / reacted-to than posts that were not identified as evoking sadness. However, sadness was not associated with commenting.
Anxiety was positively associated with sharing, but not with other modes of engagement. The findings about the role of anger are in line with previous studies (Amit-Danhi & Shifman,
2020; Berger & Milkman, 2012), though Berger and Milkman also found that stories by The
New York Times that evoked sadness were less likely to be emailed to other people. The findings offered here suggest that the dynamic of sharing sad news posts on Facebook may be different. Finally, posts containing videos were more likely to be commented-on, shared, and liked, demonstrating the power of multimedia messages.
Although future research would do well to further examine modes of user engagement on other platforms and in additional cultural/national and organizational contexts, this study offers a nuanced understanding of audience engagement and its enhancers and helps think about units of content as more engaging or less engaging than others, or as engaging in different ways (e.g., prompting shares but not comments). Democratic and business implications of the study are further discussed in the next chapter, which integrates insights
168
about production and usage of messages in triple-party news-spaces and offers news engagement logics.
169
CHAPTER 6
Discussion: News Engagement Logics in Triple-Party News-Spaces
In July 2020, U.S. lawmakers held a hearing on “online platforms and market power” with testimonies from Facebook’s chief executive, Mark Zuckerberg, and the leaders of
Amazon, Apple, and Google. Both Democrats and Republicans took aim at the four leaders and expressed concern about the tech companies’ dominance of the digital environment
(Sherman, 2020). “They have too much power,” said David Cicilline, Democrat of Rhode
Island and the chairman of the U.S. House of Representatives’ judiciary’s antitrust subcommittee (para. 9). “I share the concern that market dominance in the digital space is ripe for abuse,” stated Jim Sensenbrenner, Republican of Wisconsin and ranking member of the subcommittee (para. 10). Indeed, Facebook and other big tech companies (including companies that were not represented in the hearing) are powerful actors. They own platforms that attract a substantial number of people (e.g., Facebook, 2020a; Shaban, 2019) and create algorithms that can determine what people see on these platforms (Bell, 2016; Bell & Owen,
2017). These companies can control “even what format and type of journalism flourishes”
(Bell & Owen, 2017, p. 9). Facebook in particular has even been described as “swallowing” journalism (Bell, 2016).
With the tech companies’ power in mind, I set out to investigate the activity of news organizations and their audiences on non-proprietary platforms (Westlund & Ekström, 2018).
On these platforms, news organizations occupy what I term triple-party news-spaces: digital spaces that involve a news publisher, a platform owner, and users. The purpose of the dissertation was to identify and explicate the underlying logics of media production and usage in triple-party news-spaces and to examine them comparatively in different cultural/national and organizational contexts. On the production side, I drew on 28 interviews to examine how 14 news organizations in the U.S. and Israel produce messages for the spaces
170
under study. On the media usage side, I employed a content analysis of 1,600 messages and an analysis of engagement metrics for 157,962 messages to examine to what extent and how news organizations’ messages differ in the modes of engagement they generate: commenting versus sharing versus liking/reacting.
In this chapter, I review key findings and discuss the theoretical and practical contributions of the dissertation. In essence, I offer a framework of news engagement logics underlying triple-party news-spaces—a framework that addresses both media production and media usage in these spaces and contributes to a deeper understanding of journalism in the digital media environment. I also discuss business and democratic implications of the proposed logics, point to study limitations and offer avenues for future research, including the investigation of engagement and its outcomes in non-public triple-party news-spaces (e.g., a closed Facebook group). Such investigation can help further understand engagement with the news.
Key findings and contributions
Using the theoretical construct of media logics (Altheide &Snow, 1979; Chadwick,
2013; Klinger & Svensson, 2018), I suggest that underlying triple-party news-spaces are news engagement logics: logics by which news organizations act to evoke audience interaction with their content, and by which audience members actually interact with it. These include logics of media production and logics of media usage. I also suggest that logics can be shared across different cultural/national or organizational contexts.
Based on the interviews with media professionals in the U.S. and Israel, I propose that these logics of media production consist of five key dimensions: item selection, temporality considerations, hierarchy considerations, expression style, and approach toward user- generated content (UGC). Along each of these dimensions, news organizations combine
“old” and “new,” or older and newer norms, procedures, and processes, striving to strike a
171
balance between perceived journalistic imperatives or standards and perceived rules of the social media “game.” “Old” and “new” can either be applied in the same messages, that is coexist, or in separate messages, that is, message A reflecting an older logic is posted and afterwards message B reflecting a newer logic is posted. The use of the word “new” or
“newer” here does not necessarily suggest that something did not previously exist. Rather, this thing has been brought to the foreground in the era of digital media and fits affordances or constraints of social networking sites. In their selection of items, news organizations try to assess which stories or story components are important (older logic) and which ones would generate more interaction (newer logic). In determining the time and frequency of posting, news organizations seek to post about events in close proximity to their occurrence, but also attempt to post content at frequency or time that may help reach more users based on how the algorithm of a given platform is understood to be working (“old” and “new”). In displaying items, organizations signal the importance of certain content (“old”), though the hierarchy they offer is typically more flattened compared with the one in a news outlet (“new”). In wording or conveying messages, organizations try to use a style accessible to users—e.g., a more informal or conversational style—while using either a minimal degree or a higher degree of subjective expression (different combinations of “new” and “old”). Finally, organizations produce messages to be consumed by users (older logic), and solicit, highlight, or interact with user-generated content (newer logic).
Along three of the proposed dimensions—item selection, temporality considerations, and hierarchy considerations—considerable commonalities were found among news organizations in both the U.S. and Israel, but along the two other dimensions—expression style, and approach toward UGC—there were stark differences. In particular, the perceptions of interviewees from mainstream media in the U.S. reflected more caution about using unattributed subjective expression on social networking sites compared with interviewees
172
from Israel. The former thought they should avoid or minimize such expression in the presentation of news, particularly hard news. Trying to avoid the appearance of taking sides, especially on controversial issues, is in line with the norm of “objectivity” that took hold in
U.S. journalism over the late 19th and early 20th centuries. Possible explanations for the difference found between U.S. and Israeli organizations are the Israeli cultural characteristic of more direct talk and the tendency of Israeli journalists to be more interventionist in their role perception, that is, they are more likely than their U.S. colleagues to think it is one of their roles to promote certain agendas (Tenenboim-Weinblatt, 2014). With regard to their approach toward UGC, mass-market outlets (e.g., Yedioth Ahronoth) seemed more inclined than other outlets to post UGC possibly because they tried to come across as connected to
“the people.” Drawing on the Hierarchy of Influences (Shoemaker & Reese, 1996, 2014), the difference in the approach toward UGC can be explained at the level of news organizations, whereas the difference in the style of expression is mainly explained at the level of social systems.
The differences and similarities between countries and organizations pertain to the application of “old” and “new.” Old and new are relative terms. What can be seen as old today may have been regarded as new not so long ago, and what can be seen as new now may be considered old down the road. What is seen as new may also draw on something old. For example, selecting for publication stories that may interest audience members but are not necessarily deemed important is not new. However, trying to assess which stories or story components will maximize audience attention and evoke actual interaction on platforms not controlled by news organizations is relatively new. Immediacy, or reporting events in close proximity to their occurrence, was not always possible in a world of print newspapers but became common in a world that includes TV networks, cable channels, radio stations, and news websites. Immediacy has been described as a core value of online journalism and could
173
be considered new compared with print journalism. Immediacy can manifest in news organizations’ posting on both proprietary and non-proprietary platforms. But on the latter ones, news organizations also try to publish posts in a frequency that takes into consideration algorithms as they understand them (e.g., posting every 40 minutes on Facebook). In this context, posting about events in close proximity to their occurrence can be seen as reflecting older media logic, whereas trying to appease algorithms can be regarded as reflecting newer media logic.
Logics of media usage in triple-party news-spaces, as indicated with audience engagement metrics on Facebook, also combine “old” and “new.” As with production logics,
“old” and “new” can coexist in the same messages. Based on news value theory and literature on engagement enhancers, I identified five possible content-related components of media logics. Subsequently, I empirically examined the five components: news topics (e.g., politics and health), news factors (e.g., a prominent figure and a surprise), evaluative aspects (e.g., an overall tone and expressing criticism), triggers of emotions (e.g., evoking anger or sadness), and structural features (e.g., incorporating a video, or a photo). News topics, particularly those related to public affairs, and news factors are seen in this study as reflecting older media logics, as journalists have long made news judgements (e.g., what is newsworthy or what will be displayed more prominently) based on topics and the presence of certain factors
(e.g., a focus on a prominent person). Evaluative aspects, emotions’ triggers, and multimedia features are seen as reflecting newer media logics. Multimedia features are newer by definition. Subjectivity and emotionality are not new in journalism, but the boundaries between professional journalists and everyday citizens as well as between public and private are more blurred in the digital media environment, giving rise to a more subjective or emotional expression. Subjectivity and emotionality are particularly prominent and may be a vital force in enabling newer forms of engagement on digital platforms (Wahl-Jorgensen,
174
2016), such as social networking sites that can contain affective news streams (Papacharissi &
Oliveira, 2012). The dissertation shows that particular content characteristics reflecting older or newer logics are associated with commenting, sharing, or liking/reacting.
Based on a content analysis of news organizations’ messages and an analysis of engagement metrics, I propose three key principles of logics of media usage in triple-party news-spaces on Facebook. First, although highly commented-on posts, highly shared posts, and highly liked or reacted-to posts can overlap, they only partially overlap. Each group of posts can differ from one another to a notable degree. Second, certain content characteristics are linked to each of the examined modes of engagement—commenting, sharing, and liking/reacting—in more than one cultural/national context. Third, certain content characteristics are linked to one or more modes of engagement, but not to all of them.
Under news engagement logics (which address both the production and usage of messages), messages can hybridize different elements. Messages can focus on political topics while including multimedia features and being worded in a way that seeks to tap into user emotions (such as anger). Such messages can evoke comments, shares, and reactions.
Messages can focus on health issues and use a conversational style of expression, enticing shares, or they can focus on celebrities and use a positive tone, prompting likes. Under news engagement logics, messages are posted at frequency or time that takes into consideration the platform’s algorithm (e.g., posting every 30-40 minutes on Facebook), but some messages are posted as soon as possible if “breaking news” or a development deemed highly important occurs (even if it means that the last message on Facebook was posted five minutes earlier).
Many messages are posted without determining that message A would be more prominent than message B and that the latter would be more visible than message C and so forth, but some messages are signaled as more important (e.g., by using the label “BREAKING”).
175
Thus, news engagement logics reflect both perceived journalistic imperatives or standards and perceived rules of the social media “game.”
The comparative study of media logics in the U.S. and Israel points to the emergence of journalistic cultures on Facebook that share substantial parts in common. Put another way, substantial parts of journalistic cultures on Facebook can transcend national boundaries. Yet, although news organizations from more than one country operate on the same global platform, divergent cultural features of countries—at least the countries I examined—matter and are likely to play a role in shaping journalistic cultures on this platform, as I have demonstrated particularly with regard to the style of expression used by U.S. and Israeli organizations.
The identification and explication of news engagement logics in triple-party news- spaces in more than one country are an important theoretical contribution for five reasons.
First, rather than trying to understand media production and usage through the lens of either mass media logic or network media logic (Klinger & Svensson, 2015) or through the lens of either mass media logic or social media logic (van Dijck & Poell, 2013), I identify and explicate media logics that combine older and newer logics in a specific type of space—a digital space that involves a platform owner, news publisher, and users. Building on
Chadwick’s notion of a “hybrid media system” (2013) where older and newer media logics interact, I offer a nuanced understanding of the logics layered into each other in a triple-party news-space. Chadwick (2013) encouraged further investigation of combination of media logics in the context of political communication, and as I show, such an investigation can also help advance journalism studies. Second, the proposed news engagement logics advance audience research by addressing a key question posed by the participation paradigm
(Livingstone, 2013, p. 6): “How do people engage with, accede to, negotiate, or contest” media? My investigation was guided by this question, and the logics I identified reveal
176
patterns of audience interaction with news-related content. Third, the proposed logics add to news value theory (e.g., Galtung & Ruge, 1965; Harcup & O’Neill, 2017) and literature on engagement enhancers (e.g., Amit-Danhi & Shifman, 2020; Berger & Milkman, 2012) by showing which content characteristics are associated with different modes of engagement.
The proposed logics help think about content units as more engaging or less engaging than others, or as engaging in different ways (e.g., prompting shares but not comments). Fourth, the news engagement logics help better understand digital civic engagement and its drivers by providing a nuanced account of which content characteristics drive different acts of civic engagement. Fifth, the logics show which components of emerging journalistic cultures on
Facebook can transcend national boundaries and which components may be contingent upon divergent cultural features of countries.
The identification and explication of news engagement logics also offer a practical contribution. By showing which content characteristics prompt different modes of user interaction, the logics can help news organizations decide which content to prioritize for the spaces under study and how to present this content. Although media professionals I interviewed monitor which content performs well on social networking sites, they acknowledge that trial and error are an integral part of their work. As Lisa, an audience editor at USA TODAY, said: “No one has all the answers. We’re all experimenting.” The findings of this dissertation provide additional answers. For instance, the findings show that unattributed evaluative language in itself does not necessarily help news organizations to entice engagement with their messages, but messages focusing on health can be highly shared.
Business and democratic implications
The dissertation suggests that by including certain content characteristics in messages posted, news organizations may be able to enhance user engagement. User engagement in these spaces has both business and democratic implications. From a business
177
perspective, it may help increase the visibility of messages, potentially allowing them to reach more people. News publishers also hope that user engagement can help build their brand loyalty. User engagement in the distribution of content may also be regarded as a form of “free labor” (Jönsson & Örnebring, 2011), helping news organizations disseminate their content.
But engaging people “to the point where they feel compelled to pay for a digital subscription” (Batsell, 2015, p. 8) is a major challenge for news organizations. In August 2020, The New York Times Company reported that during the second quarter of 2020—a three-month period dominated by the coronavirus pandemic—it added 669,000 net new digital subscribers and brought in $185.5 million in revenue for digital subscriptions and ad compared with print revenue of $175.4 million (Tracy, 2020).
Yet, not every newspaper can report such numbers, and a time of global pandemic is, hopefully, an unusual time. Despite declines in print readership over time, the majority of revenue comes from the print side rather than the digital side of operation for major newspapers (Chyi & Tenenboim, 2017). Furthermore, when news organizations post content on digital platforms they do not own, they may help entice users to stay on these platforms
(where news-related content is also offered). Put differently, users may get the impression that they do not “need” to regularly visit news sites because they can get important or interesting content on social networking sites.
Although platforms non-proprietary to the news media attract a substantial number of people, news organizations would do well not to put all eggs in the basket of technology companies, that is, not to be fully dependent on platforms they do not control.
They would be wise to acknowledge there is still a room for both “old” and “new” and to avoid treating their proprietary products or platforms—e.g., newspapers and news websites— as doomed. That said, news organizations may also enhance user engagement with their
178
messages on non-proprietary platforms. The news engagement logics I identified offer guidance on how to do that. In addition, news organizations would do well to create opportunities for a deeper engagement—for example, in the form of a continuous conversation between news workers and audience members—in an attempt to bring audience members to the point where they pull out their credit cards to pay for something they feel is valuable to them. Some interviewees shared that the news organizations they worked for experimented with closed Facebook groups where conversations were held with audience members. Indeed, recent research has suggested that ongoing conversations between a journalist and audience members in a journalistic WhatsApp group—a non-public digital space—allowed a continuous co-construction of journalistic knowledge across the news- production process (including, but not limited to, generating source material for a story and evaluating information). Participants payed a monthly fee of 15 New Israeli Shekels (about 4
US Dollars) to be group members (Kligler-Vilenchik & Tenenboim, 2020). Yet, further research is needed on possible commercial benefits of dynamics in non-public digital spaces that do not require a payment to become a member.
An engagement with the news can also be seen as an important aspect of democratic life: Those who engage with the news may express their views or preferences about matters of public concern and even converse with other people. People’s political expression via digital channels can be meaningful in itself, but it may also drive or at least be linked to offline engagement in democratic life (e.g., participating in street marches)
(Boulianne, 2019).
The dissertation has shown that both in the U.S. and Israel news organizations’ messages focusing on politics / government were more engaging than messages that their main topic was not politics / government. This is a meaningful finding considering that
Facebook users can choose to engage with various messages that do not focus on politics,
179
such as personal messages by family and friends, or messages by celebrities they follow.
However, it is possible that at least some of the engagement with messages on political issues is not “authentic” in that users may be invited and even get paid by political actors or other actors (e.g., a marketing company) to engage with posts as part of a concerted effort to portray political actors in a positive or negative light (e.g., DW News, 2019).
Political expression (e.g., via commenting) by users of Facebook and other platforms can also be used for the spread of disinformation, misinformation, hateful speech, and toxicity. Thus, audience engagement does not always serve democratic life.
News organizations would do well to pay even more attention to user comments on Facebook and other platforms and to further interact with commenters as research has shown that it is possible to affect comment section norms by engaging directly with commenters (Stroud et al., 2015). Although some of the news organizations I studied indicated they did interact with commenters on Facebook—at least sometimes—by answering questions, interactions can be increased.
As mentioned earlier, conversations between news workers and audience members can also be held in non-public digital spaces (e.g., a closed Facebook group) where groups of people have opportunities to produce and share content among each other— yet not to a wide audience (see Tenenboim & Kligler-Vilenchik, 2020). Within such spaces, journalists and audience members may be able to get to know each other better, build relationships, and express themselves more freely. A possible way to create communities in non-public spaces is to pose screening questions to users in an attempt to better understand their intentions or interests. For example, USA TODAY created a closed Facebook group called “Across the Aisle, Across the Nation” where potential participants are asked what their thoughts are on the state of American politics and what political issues or causes they care the most about. The admins or group regulators also set rules for participation, such as no hate
180
speech or bullying, and respect everyone’s privacy (that is, what is shared in the group should stay in the group). The dynamics and their effects in this group and other groups in different cultural or national contexts are yet to be examined.
Limitations and ways forward
This dissertation is not without limitations. First, although media professionals
I interviewed addressed other non-proprietary platforms, the content analysis I conducted focused on Facebook. Messages on this highly dominant platform certainly deserve attention, but it is possible that some user engagement patterns are different on other platforms, such as Twitter. Thus, future research would do well to analyze content and different engagement metrics on other platforms as well. Second, although the dissertation offers a comparative perspective beyond the typically nation-centered studies of the press, it focuses on newspapers and digital-native outlets in two countries. By comparing the U.S. and
Israel I gained important insights about the likely role of cultural features in shaping logics underlying triple-party news-spaces. But future research would do well to examine logics of media production and usage in additional cultural/national and organizational contexts to understand what transcends these contexts and what may be contingent upon them. Third, I analyzed which messages prompt more user comments than other messages, but I did not examine the content of comments. Future research can include a content analysis of messages posted by news organizations and user comments that are attached to them to identify which types of messages generate different types of user comments. Analyzing both editorial content and user comments attached to it is not entirely new (e.g., Ziegele et al., 2018a), but it may help further understand user engagement in different triple-party news-spaces. Fourth, I talked with producers of messages and not with users of messages. Talking with the latter can help better understand their preferences and experiences.
I stated early on that the problem I was confronting in this study was to best
181
understand the ways in which news organizations and their audiences behaved on digital platforms non-proprietary to the news media. I learned a great deal about these ways, as well as about the challenges of appeasing algorithms and maximizing user interaction. As news organizations try to meet these challenges, it is worth bearing in mind the importance of quality journalism, which is one of crucial bulwarks that can protect democracy by providing reliable information about the workings of government and other institutions or people of power. Quality journalism does not have to be “swallowed” by Facebook and other technology companies. It can exist in or be assisted by different spaces, including triple-party news-spaces. But for triple-party news-spaces to realize their potential, each party needs to do more. News publishers would do well to engage more with users (e.g., by conversing with them in different stages of the news production) in addition to trying to entice user engagement by posting content. Platform owners need to further support journalism and its ability to reach people and build relationships with them. Allowing news publishers to label content as “breaking news” and giving this content more visibility is one way to do that. A recent partnership between Facebook and The New York Times to develop an augmented reality reporting project is another possible way (Fischer, 2020). Sharing with news publishers insights about issues they are struggling with (e.g., temporality issues) may also be helpful. Users would do well to engage with the news and news workers rather than leaving engagement to those who spread toxicity or disinformation. Researchers can also do more to engage with each of these parties and to try to promote shared benefits
(such as enhancing knowledge). It is not too late to experiment and to improve.
182
Appendices
Appendix A
Interview Guide
What is your position in the organization, and how long have you been in this position?
On which social media platforms does your organization have at least one account? (Facebook, Twitter, Instagram, etc.)
Does the organization have more than one account on these platforms? How many? Why?
What is the purpose – or what are the purposes – of your organization's activity on these platforms?
Does your organization have guidelines for using social media? If so – could you talk about them, and give examples of guidelines?
I have several questions about the social media team in your organization: How many members does the team include? What is the hierarchy within the team? How does the team operate in relation to other teams/divisions/sections within the organization? Who does the team report to? To what extent – if at all – does the team collaborate with people in the news division/desk?
About how many social media posts per day does the team typically publish? Are there minimal number and maximal numbers? What is the estimated number of daily posts the team publishes on each of the social media platforms?
What considerations are taken into account in determining the volume and frequency of posting?
What considerations are taken into account in deciding when to post content?
The social media team doesn't typically post every unit of content that appears in your newspaper and news website. How do you select what to post, and what not to post?
Are certain types of content created only for the social media pages/accounts of the organization, and do not appear on the organization's website? If so – what types of content?
Does the social media team use live videos? On what occasions? For what purposes? Is the use of live videos beneficial as you expected, or not?
Does the social media team share on the organization's pages/accounts posts by other users, such as politicians, celebrities, and ordinary citizens? If so – what types of posts / by what types of users? What are the considerations for selecting these posts? How frequent is this practice of sharing?
183
What are the differences and similarities between producing and writing posts for the social media pages / accounts of the organization, and producing and writing units of content for the organization's news site / newspaper?
I would be happy to know more about the writing of the organization's posts on social media. What writing styles do you use? Are there styles you more commonly use? How formal or informal is the language you use? How likely are you to express opinions in posts, even if the links they contain are to news reports? Does an expression of opinion require someone's approval within the organization? How likely are you to criticize people in power in the posts? How likely are you to use humor, irony, or sarcasm? Are there expressions you often use in posts? What are they? What considerations are taken into account in the wording of posts? Are they similar or different compared with the considerations for wording content for the news website / newspaper?
To what extent – if at all – are the social media pages / accounts of the organization used for soliciting or promoting user-generated content? Could you give examples?
To what extent – if at all – does the social media team interact with users in the comment sections – replying to comments, engaging in conversations, liking people's comments?
How often do you – or the social media team – check the number of likes, shares, and comments per post? What do you do with this information? Is a certain type of engagement more desired than other types? Are certain types of content more likely to generate a certain type of engagement, but not other types of engagement?
Could you share some of the successes and disappointments you or the social media team have had with the use of social media (for the organization)?
Could you share some concerns or dilemmas that you or the team have when you use social media?
How – if at all – are the ways your organization use social media different from the ways competing organizations use them?
For Israelis: Based on what you have seen, how – if at all – are the ways Israeli news organizations use social media different from the ways news organizations in other countries (such as the US) use them?
How – it at all – do you hope to improve your activity on social media, or what do you hope to accomplish that you haven't accomplished yet?
Is there something else you want to share or add?
Many thanks for your time and help!
184
Appendix B
News Posts – Codebook
This codebook is designed to help analyzing content that was posted on Facebook by news organizations. Please provide the requested information for each post separately. You are to address only the post’s content and not the comments that may be attached to it or the content of links it may contain. Make sure to read carefully the posts, the instructions about the requested information, as well as the explanations and examples that are supposed to help you in the coding process. Thank you for your cooperation and assistance.
A. Topic
The main topic of the post
1) Politics/government 2) Military/defense 3) Economy/business 4) Society/welfare* 5) Crime 6) Health 7) Education 8) Science/technology 9) Disasters/accidents 10) Arts/entertainment 11) Sports 12) Weather 13) Food 14) Lifestyle/Leisure (other than food) 15) Human interest story** 16) Other ______(please specify) 17) Hard to determine
*For example: Content focusing on society/welfare may address inequality, poverty, or union protests. **For example: A human interest story may be about a person who survived against all odds, or the tallest person in the world.
B. Conflict
The post reflects disagreement between parties, individuals, groups, or countries
0) No 1) Yes
C. Surprise
The post contains an element of surprise / something unusual
185
0) No 1) Yes
The coding should be yes if, for example, the post reports that a political candidate who was trailing in the polls is now leading, or it addresses a terrorist attack in Oslo – a place that is perceived as peaceful, or it mentions a rude statement by a person who is considered polite.
D. Reference to persons
The post mentions a person’s name or shows a person’s face
0) No 1) Yes
E. Prominent figure
The post mentions a prominent figure, such as an actress or a president
0) No 1) Yes
If the coding is no, please skip to V.
F. Type of prominent figure (I)
The prominent figure mentioned in the post is a/an
1) Political actor (e.g., a president, or a senator) 2) Military actor (e.g., a general, or a chief of general staff) 3) Artist (e.g., a movie actor, a singer) 4) Media actor (e.g., a commentator, or a TV anchor) 5) Sportsperson (e.g., a football player, or an athlete) 6) Fashion model 7) Other ______(please specify)
If there is more than one prominent figure, code the first figure that is mentioned. Then code the second figure below. If only one figure is mentioned, please skip to H.
G. Type of prominent figure (II)
The second prominent figure mentioned in the post is a/an
1) Political actor (e.g., a president, or a senator) 2) Military actor (e.g., a general, or a chief of general staff) 3) Artist (e.g., a movie actor, a singer) 4) Media actor (e.g., a commentator, or a TV anchor) 5) Sportsperson (e.g., a football player, or an athlete) 6) Fashion model 7) Other ______(please specify)
186
H. Damage
The post refers to a negative consequence of an event, such as a death, injury, illness, or material damage
0) No 1) Yes
I. Death
The post mentions a death of at least one person
0) No 1) Yes
J. Injury
The post mentions an injury of at least one person
0) No 1) Yes
K. Violence
The post mentions violence, whether it is a military attack or a fight between teenagers
0) No 1) Yes
L. Success
The post refers to a positive consequence of an event, such as an improvement of the standards of living, breakthrough in health research, or raising money to help the elderly
0) No 1) Yes
M. Proximity to the U.S.
The post addresses an occurrence or occurrences in
1) The United States of America only 2) A country or countries other than the United States of America 3) The United States of America and at least one other country 4) Other option ______(please specify)
If the coding is 1 or 3, please skip to O. If the coding is 2 or 4, please address N as well.
N. Relation to the U.S.
The post addresses an occurrence related to the United States of America
187
0) No 1) Yes 2) Hard to determine
The coding should be yes if, for example, the post reports that an American soldier was killed in Afghanistan, or an American actress received an award in France.
O. Proximity to Israel
The post addresses an occurrence or occurrences in
1) Israel only 2) A country or countries other than Israel 3) Israel and at least one other country 4) Other option ______(please specify)
If the coding is 1 or 3, please skip to Q. If the coding is 2 or 4, please address P as well.
P. Relation to Israel
The post addresses an occurrence related to Israel
0) No 1) Yes 2) Hard to determine
The coding should be yes if, for example, the post reports that the Israeli prime minister is visiting Australia, or an Israeli actress received an award in France.
Q. Affected country
The country mostly affected by the event/issue described in the post
1) The United States of America 2) Israel 3) Another country 4) Hard to determine
R. Impact
The post addresses an event/issue that has consequences for one or more social categories or groups (such as nations, communities, workers, and students)
0) No 1) Yes
S. Frequency
The post addresses an event/issue that has a determinable beginning and end
188
0) No 1) Yes
Emotions
T. Anger
Anger is defined as a strong feeling of annoyance, displeasure, or hostility. The post evokes anger among users
0) No 1) Yes 2) Hard to determine
U. Anxiety
Anxiety is defined as a feeling of worry, nervousness, or unease about something with an uncertain outcome. The post evokes anxiety among users
0) No 1) Yes 2) Hard to determine
V. Awe
Awe is defined as a feeling of reverential respect mixed with fear or wonder. The post inspires awe among users
0) No 1) Yes 2) Hard to determine
W. Sadness
Sadness is defined as the condition or quality of being sad. The post evokes sadness among users
0) No 1) Yes 2) Hard to determine
Subjectivity
X. Evaluative language
The post uses evaluative language or offers unattributed commentary beyond the facts of an occurrence or issue
189
0) No 1) Yes
Examples of posts that should be coded as 1 (yes): “Just when we thought the president couldn't go lower, he made this statement about immigrants”; “The parliament has just passed an outrageous bill”; “The woman was brutally murdered.”
Y. Main purpose
The primary purpose of the post
1) Conveying information 2) Seeking information 3) Conveying opinion 4) Other _____ (please specify)
A post conveys an opinion if it uses evaluative language or offers unattributed commentary beyond the facts of an occurrence or issue.
Z. Humor
Regardless of what else the writer of the post is doing (seeking information, stating an opinion or conveying information), is the writer of the post trying to be funny?
0) No 1) Yes
AA. Emotional language (I)
The post explicitly refers to a human emotion
0) No 1) Yes
Examples of posts that should be coded as 1 (yes): “They were sad to hear the news”; “We were shocked at the president’s statement.”
AB. Emotional language (II)
The post contains an unattributed emotional expression (in other words: the writer of the post expresses his / her emotion or an emotion of people he / she represents)
0) No 1) Yes
Examples of posts that should be coded as 1: “Former secretary of state dies after car crash. Our condolences”; “New Zealand’s prime minister gives birth to baby girl. Congratulations!”; “We were shocked at the president's statement.”
AC. Emoji
190
The post contains an emoji, such as or
0) No 1) Yes
AD. Criticism (I)
The post contains evaluative language that expresses disapproval of someone or something
2) No 3) Yes
AE. Criticism (II)
The post contains unattributed evaluative language that expresses disapproval of someone or something (in other words: the criticism is expressed by the writer of the post or those he / she represents, and is not attributed to a source)
0) No 1) Yes
Examples of posts that should be coded as yes: “The prime minister’s behavior was disgraceful”; “We think this is not the best way to address the problem.”
Tone
AF. Overall tone
The overall tone of the post
1) Negative 2) Positive 3) Neutral 4) Hard to determine
Pronouns
AG. Use of “we”
The post uses the pronoun “we” in an unattributed way (“we” is said on behalf of the news organization and is not attributed to a source)
0) No 1) Yes
AH. Use of “you”
The post uses the pronoun “you” in an unattributed way (it is not attributed to a source)
0) No
191
1) Yes
Engagement
AI. Commenting
The post includes a call/invitation to express views or ask questions
0) No 1) Yes
Examples of posts that should be coded as 1 (yes): “Leave your questions in the comments”; “Read the column and share your thoughts.”
AJ. Sharing
The post includes a call/invitation to share content
0) No 1) Yes
Examples of posts that should be coded as 1 (yes): “Share it with your friends”; “Retweet if you agree.”
AK. Liking
The post includes a call/invitation to “like” content or to use another reaction (such as a sad emoji)
0) No 1) Yes
Examples of posts that should be coded as 1 (yes): “Like and Share with your friends”; “Like if you agree, press if you disagree.”
AL. Tagging
The post includes a call/invitation to tag other users
0) No 1) Yes
AM. Watching or clicking
The post includes a call/invitation to read, watch, or click on content
0) No 1) Yes
AN. Engaging in other ways
192
The post includes a call/invitation to engage in other ways – e.g., submitting photos/videos, or taking a quiz
0) No 1) Yes
AO. Another request
The post includes a different request (not a request to comment, take a quiz, etc.)
0) No 1) Yes
Examples of posts that should be coded as 1 (yes): “You may lose your driver’s license even if your passenger is the one throwing eggs. Please tell friends who need to hear this”, “Contaminated food was also found in the city. Watch out”
Other characteristics
AP. Question
The post includes a question posed by the writer of the post (rather than by a source quoted in the post)
0) No 1) Yes
AQ. Shared post
The post was originally produced by a user other than the examined news organization. In other words, the examined news organization shared/retweeted a post by another user
0) No 1) Yes
If the coding is 1 (yes), please address AR as well. If not, please skip to AS.
AR. User of the original post (I)
Please copy the username of the user that produced the post______
AS. Shared post with addition
The post contains a post by another user, as well as other content. In other words, the news organization shared a post by another user and added something to it (e.g., a statement)
0) No 1) Yes
193
If the coding is 1 (yes), please address AT as well. If not, please skip to AU.
AT. User of the original post (II)
Please copy the username of the user that produced the shared post______
AU. Breaking news
The post contains the words "breaking news" / "breaking" / a similar word, or an image/logo with these words
0) No 1) Yes
AV. Length
In terms of written text, the post contains
1) A sentence or less than a sentence 2) More than a sentence and not more than a paragraph 3) More than one paragraph 4) No written text
A paragraph is regarded here as a distinct division of written matter that begins on a new line. A long piece of writing that consists of sentences that are not separated from one another should be coded as “more than a sentence and not more than a paragraph.” A post's date and words that appear above the date (at the top of a post), such as the name of the Facebook user or “new photos have been added to the photo album” should not be coded here as a written text. If words appear in a photo, they should not be coded here as a written text.
Photos, videos, and links
AW. Use of photos
The post contains
1) No photos 2) One or two photos 3) Three or more photos
Logos and figures should not be coded as photos. The small head photo that appears at the top of a Facebook post, adjacent to the name of the account holder, should also not be coded here as a photo.
AX. Video
The post contains at least one video player
0) No 1) Yes
194
AY. Live video
The post contains a video that was live-streamed
0) No 1) Yes
The content of the post can help identify such a video. For example: “Nick Confessore, New York Times investigative reporter, is live answering your question about the article.”
AZ. Embedded link (I)
The post contains an embedded link
0) No 1) Yes
If the coding is 0 (no), please skip to BD.
BA. Embedded link (II)
The embedded link is to the website of the examined news organization
0) No 1) Yes
If the coding is 0 (no), please address BB as well. If the coding is 1 (yes), please skip to BC.
BB. Embedded link (III)
If the embedded link is not to the site of the examined news organization, please indicate to which site/content it is ______
BC. Embedded link (IV)
The embedded link is to a
1) News report 2) Commentary / opinion piece 3) Magazine feature 4) Other ______(please specify)
BD. URL link (I)
The post includes at least one URL link (which is not embedded)
0) No 1) Yes
195
If the coding is 1 (yes), please address BE as well. If the coding is 0 (no), you are done coding.
BE. URL link (II)
The URL link is to the website of the examined news organization
0) No 1) Yes
If the coding is 0 (no), please address BF as well.
BF. URL link (III)
If the URL link is not to the site of the examined news organization, please indicate to which site/content it is ______
BG. URL link (IV)
The URL link is to a
1) News report 2) Commentary / opinion piece 3) Magazine feature 4) Other ______(please specify)
196
Appendix C
Inter-coder reliability was examined for 150 of the U.S. posts (18.75%) and 150 of the Israeli posts (18.75%) under study. The inter-coder reliability was examined using Krippendorff’s alpha, and variables with Krippendorff's alpha not lower than .700 were included in the analysis. However, variables with Krippendorff's alpha lower than .700 were disqualified, that is, they were not included in the analysis presented in the findings section.
Krippendorff’s alpha for the examined variables in the U.S. sample and the Israeli sample
(calculations were conducted separately) is shown here. Another measure of inter-coder reliability—Cohen’s kappa—is presented for comparison.
Table 13
Inter-Coder Reliability for the Examined Variables
U.S. Sample (n = 150) Israeli Sample (n = 150) Krippendorff's Cohen’s Krippendorff's Cohen’s alpha kappa alpha kappa Topic .823 .823 .888 .888
News Values Conflict .712 .711 .749 .750
Surprise .714 .715 .739 .740
Prominent .973 .973 .933 .933 Figure
Reference to .800 .799 .840 .840 Persons
Impact .772 .772 .834 .833
Damage .763 .764 .802 .802
Death .902 .902 .892 .892
Injury .797 .797 1.000 1.000
Violence .953 .953 .748 .747
Success .868 .868 .814 .813
Proximity .759 .760 .915 .915
197
U.S. Sample (n = 150) Israeli Sample (n = 150) Krippendorff's Cohen’s Krippendorff's Cohen’s alpha kappa alpha kappa Frequency* .654* .654 .587* .601
Evaluative Evaluative .903 .903 .801 .800 Aspects Language
Criticism .878 .878 .867 .867
Emotional .736 .735 .803 .802 Language
Tone .835 .835 .828 .828
Humor* -.027* -.021 .902 .902
Emotions Anger .854 .853 .713 .714
Sadness .926 .926 .868 .867
Anxiety .827 .826 .892 .892
Awe* .664* .664 .794 .793
Structural Video .902 .902 1.000 1.000 Features Photo .893 .892 .908 .908
Embedded .957 .956 .933 .933 Link
Embedded .708 .708 .852 .851 Link – Type of Content
Length .916 .915 .841 .841
* The variable was not included in the analysis due to low inter-coder reliability.
198
Appendix D
Table 14
Frequencies of Content Characteristics Among Posts by Media Outlets in the U.S. and Israel
General Sample U.S. Sample Israeli Sample (n = 1,600) (n = 800) (n = 800) % % % Topic Politics / 26.6 30.6 22.5 Government
Military / 5.4 1.5 9.4 Defense
Economy / 3.5 5.5 1.5 Business
Society / 4.4 5.3 3.6 Welfare
Crime 6.8 6.4 7.2
Health 3.3 3.1 3.5
Education 1.3 2.0 .6
Science / 3.1 3.3 2.9 Technology
Disasters / 3.5 2.3 4.8 Accidents
Arts / 12.9 10.9 15.0 Entertainment
Sports 4.3 5.6 3.0
Weather 1.6 1.5 1.8
Food 3.3 3.6 2.9
Lifestyle / 2.9 2.3 3.5 Leisure
Human interest 5.0 2.9 7.1
Other 9.4 9.6 9.1
199
General Sample U.S. Sample Israeli Sample (n = 1,600) (n = 800) (n = 800) % % % Hard do 2.7 3.8 1.6 Determine
News Values Conflict 26.3 27.8 24.8
Surprise 21.0 24.8 17.3
Prominent 49.8 52.5 47.1 Figure
Reference to 61.9 62.5 61.3 Persons
Impact 45.3 54.3 36.4
Damage 25.9 26.0 25.8
Success 4.3 4.3 4.4
Proximity n/a 70.0 63.2
Evaluative Evaluative 48.8 42.4 55.3 Aspects Language
Criticism 24.0 26.1 21.9
Emotional 17.4 13.0 21.9 Language
Positive Tone 15.9 13.6 18.3
Negative Tone 14.3 12.8 15.9
Neutral Tone 51.9 61.3 42.6
Hard to 17.8 12.4 23.3 Determine
Emotions Anger 21.2 24.9 17.5
Sadness 18.6 21.1 16.1
Anxiety 4.6 4.1 5.0
Structural Video 6.1 5.9 6.4 Features
200
General Sample U.S. Sample Israeli Sample (n = 1,600) (n = 800) (n = 800) % % % Photo 84.6 83.8 85.5
Embedded Link 69.1 93.1 45.1
Embedded Link 55.4 56.4 53.3 to a News Report*
Length – More 4.9 5.5 4.4 Than a Paragraph
Length – More 24.3 21.0 27.6 Than a Sentence and Not More Than a Paragraph
Length – A 68.3 73.5 63.0 Sentence or Less
Length – No 2.5 .0 5.0 Written Text * Embedded link to a news report was examined among 1,108 posts in the general sample (those that included an embedded link), 746 posts in the U.S. subsample, and 362 in the Israeli subsample.
201
References
Ad Fontes Media. (2020). All News Sources. Retrieved from /https://www.adfontesmedia.com/rankings-by-individual-news-source Alexa. (2020). The top sites on the web. Retrieved from https://www.alexa.com/topsites Allan, S., & Thorsen, E. (2009) Citizen journalism: Global perspectives. New York: Peter Lang. Altheide, D. L. (2014). Media edge: Media logic and social reality. New York: Peter Lang. Altheide, D. L., & Snow, R. P. (1979). Media logic. Beverly Hills, CA: Sage. Amit-Danhi, E. R. & Shifman, L. (2020). Off the charts: User engagement enhancers in election infographics. Information, Communication & Society. doi:10.1080/1369118X.2020.1761858 Anderson, C. W. (2013). Rebuilding the news: Metropolitan journalism in the digital age. Philadelphia: Temple University Press. Arian, A., & Shamir, M. (2008). A decade later, the world had changed, the cleavage structure remained: Israel 1996–2006. Party Politics, 14(6), 685–705. doi:10.1177/1354068808093406 Avraham, D. (2013, February 18). Data from the Internet research board for January 2013: Walla the leading Israeli site – also in video. Walla (in Hebrew). Retrieved from https://b.walla.co.il/item/2617731 Barber, B. (1984). Strong democracy: Participatory politics for a new age. Berkley CA: University of California Press. Bassan-Nygate, L., & Weiss, C. M. (2020). It’s us or them: Partisan polarization in Israel and beyond. APSA MENA Politics Newsletter, Spring. The Middle East and North Africa Politics Section of the American Political Science Association. https://static1.squarespace.com/static/5bda793b0dbda3f980201462/t/5ed5271d 00577d3c5b10d27c/1591027491231/BW_MENA.pdf Batsell, J. (2015). Engaged journalism: Connecting with digitally empowered news audiences. New York: Columbia University Press. Bell, E. (2016). Facebook is eating the world. Columbia Journalism Review. Retrieved from http://www.cjr.org/analysis/facebook_and_media.php Bell, E., & Owen, T. (2017). The platform press: How Silicon Valley reengineered journalism. New York: Tow Center for Digital Journalism. Retrieved from https://academiccommons.columbia.edu/catalog/ac :15dv41ns27
202
Bennett, W. L. (2007). Civic learning in changing democracies: Challenges for citizenship and civic education. In P. Dahlgren (Ed.), Young citizens and new media: Learning democratic engagement (pp. 59–77). New York: Routledge. Bennett, W. L. (2008). Changing citizenship in the digital age. In W. L. Bennett (Ed.), Civic life online (pp. 1–24). Cambridge, MA: MIT Press. Bennett, W. L., Wells, C., & Freelon, D. (2011). Communicating civic engagement: Contrasting models of citizenship in the youth web sphere. Journal of Communication, 61(5), 835–856. doi:10.1111/j.1460-2466.2011.01588.x Benzaquen, I. (2017, June 15). How many visitors do the news websites in Israel have? The Seventh Eye (in Hebrew). Retrieved from https://www.the7eye.org.il/250845 Berger, J., & Milkman, K. L. (2012). What makes online content viral? Journal of Marketing Research, 49(2), 192–205. doi:10.1509/jmr.10.0353 Boczkowski, P. J. (2004). Digitizing the news: Innovation in online newspapers. Cambridge: MIT Press. Boczkowski, P. J. (2010). News at work: Imitation in an age of information abundance. Chicago, IL: University of Chicago. Boczkowski, P. J., Mitchelstein, E., & Walter, M. (2011). Convergence across divergence: Understanding the gap in the online news choices of journalists and consumers in Western Europe and Latin America. Communication Research, 38(3), 376–396. doi:10.1177/0093650210384989 Boczkowski, P.J., & Mitchelstein, E. (2012). How users take advantage of different forms of interactivity on online news sites: Clicking, e-mailing, and commenting. Human Communication Research, 38(1), 1–22. doi:10.1111/j.1468-2958.2011.01418.x Boczkowski, P.J., & Mitchelstein, E. (2013). The news gap: When the information preferences of the media and the public diverge. Cambridge, Mass.: MIT Press. Boczkowski, P.J., & Peer, L. (2011). The choice gap: The divergent online news preferences of journalists and consumers. Journal of Communication, 61(5), 857–876. doi:10.1111/j.1460-2466.2011.01582.x Bromwich, J. E., & Haag, M. (2018, January 12). Facebook is changing. What does that mean for your news feed? The New York Times. Retrieved from https://www.nytimes.com/2018/01/12/technology/facebook-news-feed-changes.html Boulianne, S. (2019). Revolution in the making? Social media effects across the globe. Information, Communication & Society, 22(1), 39–54.
203
doi:10.1080/1369118X.2017.1353641 Bowman, W., & Willis, C. (2003). We media: How audiences are shaping the future of news and information. The Media Center. Retrieved from http://www.hypergene.net/wemedia/download/we_media.pdf Bullard, S. B. (2015). Editors use social media mostly to post story links. Newspaper Research Journal, 36(2), 170–183. doi:10.1177/0739532915587288 Burscher, B., Odijk, D., Vliegenthart, R., de Rijke, M., & de Vreese, C. H. (2014). Teaching the computer to code frames in news: Comparing two supervised machine learning approaches to frame analysis. Communication Methods and Measures, 8(3), 190–206. doi:10.1080/19312458.2014.937527 Cameron, A.C., & Trivedi, P. K. (2013). Regression analysis of count data. New York: Cambridge University Press. Chadwick, A. (2013). The hybrid media system: Politics and power. Oxford, UK: Oxford University Press. Chyi, H. I. (2009). Information surplus in the digital age: Impact and implication. In Z. Papacharissi (Ed.), Journalism and citizenship: New agendas in communication (pp. 47–60). New York: Taylor & Francis. Chyi, H. I., & Tenenboim, O. (2017). Reality check: Multiplatform readership in the United States, 2007–2015. Journalism Practice, 11(7), 798–819. doi:10.1080/17512786.2016.1208056 Collier, D. (1993). The comparative method. In A. W. Finifter (Ed.), Political science: The state of the discipline II. Washington, D.C.: American Political Science Association. Couldry, N. (2008). Mediatization or mediation? Alternative understandings of the emergent space of digital storytelling. New Media & Society, 10(3), 373–391. doi:10.1177/1461444808089414 CrowdTangle Team. (2020). CrowdTangle. Menlo Park, California: Facebook. Coxe, S., West, S. G., & Aiken, L. S. (2009). The analysis of count data: A gentle introduction to Poisson regression and its alternatives. Journal of Personality Assessment, 91(2), 121–136. doi:10.1080/00223890802634175 Dahlgren, P. (2005). The Internet, public spheres, and political communication: Dispersion and deliberation. Political Communication, 22(2), 147–162. doi:10.1080/10584600590933160 Dalton, R. J. (2016). The good citizen: How a younger generation is reshaping American politics (second edition).Washington, DC: CQ Press.
204
Delli Carpini, M. X., Cook, F. L., & Jacobs, L. R. (2004). Public deliberations, discursive participation and citizen engagement: A review of the empirical literature. Annual Review of Political Science, 7(1), 315–344. doi:10.1146/ annurev.polisci.7.121003.091630 Dewey, J. (1939). Creative democracy: The task before us. Retrieved from http://www.faculty.fairfield.edu/faculty/hodgfinalguaranteeson/Courses/progress/Dew ey.pdf Domingo, D., Quandt, T., Heinonen, A., Paulussen, S., Singer J. B., & Vujnovic, M. (2008). Participatory journalism practices in the media and beyond: An international comparative study of initiatives in online newspapers. Journalism Practice, 2(3), 326–342. doi:10.1080/17512780802281065 Dvir-Gvirsman, S. (2019). I like what I see: studying the influence of popularity cues on attention allocation and news selection. Information, Communication & Society, 22(2), 286–305. doi:10.1080/1369118X.2017.1379550 DW News. (2019, December 19). German political parties paid for Instagram, Facebook likes. Retrieved from https://www.dw.com/en/german-political-parties-paid- for-instagram-facebook-likes/a-51733728 Economist Intelligence Unit. (2020). Democracy index 2019: A year of democratic setbacks and popular protest. Retrieved from http://www.eiu.com/Handlers/WhitepaperHandler.ashx?fi=Democracy-Index- 2019.pdf&mode=wp&campaignid=democracyindex2019 Edgar, D. (2014). L.A. Times updates newsroom ethics guidelines. Los Angeles Times. Retrieved from https://www.latimes.com/local/readers-rep/la-rr-la-times-updates- newsroom-ethics-guidelines-20140618-story.html Editor & Publisher. (2018). Newspaper Databook 2018. Fountain Valley, CA: Editor & Publisher. Ellis, J. (2015, September 16). What happened after 7 news sites got rid of user comments. Nieman Lab. Retrieved from http://www.niemanlab.org/2015/09/what- happened-after-7-news-sites-got-rid-of-reader-comments Esser, F. (2013). Mediatization as a challenge: Media logic versus political logic. In H. Kriesi, S. Lavanex, F. Esser, J. Matthes, M. Bühlmann, & D. Bochsler (Eds.), Democracy in the age of globalization and mediatization (pp. 155–176). Basingstoke, UK: Palgrave Macmillan. Evans, S. K., Pearce, K. E., Vitak, J., & Treem, J. W. (2017). Explicating affordances: A
205
conceptual framework for understanding affordances in communication research. Journal of Computer-Mediated Communication, 1(1), 35–52. doi:10.1111/jcc4.12180 Eveland, W. P., & Thomson, T. (2006). Is it talking, thinking, or both? A lagged dependent variable model of discussion effects on political knowledge. Journal of Communication, 56(3), 523–542. doi:10.1111/j.1460-2466.2006.00299.x Facebook. (2020a). Press releases. Retrieved from https://investor.fb.com/investor-news/default.aspx Facebook. (2020b). What’s the difference between a profile, page and group on Facebook? Retrieved from https://www.facebook.com/help/337881706729661 Fandos, N., Huang, J., Kaplan, T., & Rogers, K. (2019, February 5). Transcript: Trump’s State of the Union, annotated. The New York Times. Retrieved from https://www.nytimes.com/interactive/2019/02/05/us/politics/trump-state-of-union- speech-transcript.html Fischer, S. (2020, September 2016). NYT, Facebook launch multi-year augmented reality reporting project. Axios. Retrieved from https://www.axios.com/new-york-times-facebook-augmented-reality-0973fa9e-ffcc- 4a68-91ec-ffb77d757745.html Fishman, M. (1980). Manufacturing the news. Austin: University of Texas Press. Galtung, J., & Ruge, M. (1965). The structure of foreign news: The presentation of the Congo, Cuba and Cyprus Crises in four Norwegian Newspapers. Journal of International Peace Research, 2(1), 64–90. doi:10.1177/002234336500200104 Gans, H. J. (1979). Deciding what’s news: A study of CBS Evening News, NBC Nightly News, Newsweek and Time. New York: Vintage. Gil de Zúñiga, H., Molyneux, L., & Zheng, P. (2014). Social media, political expression, and political participation: Panel analysis of lagged and concurrent relationships. Journal of Communication, 64(4), 612–634. doi:10.1111/jcom.12103 Ha, L., Xu, Y., Yang, C., Wang, F., Yang, L., Abuljadail, M., Hu, X., Jiang, W., & Gabay, I. (2018). Decline in news content engagement or news medium engagement? A longitudinal analysis of news engagement since the rise of social and mobile media 2009–2012. Journalism: Theory, Practice & Criticism, 19(5) 718–739. doi:10.1177/1464884916667654 Haaretz. (2016, December 15). “Amos Shocken, are you Zionist?”: Questions and answers with Haaretz publisher. Retrieved from https://www.haaretz.co.il/20years/1.3152618 Haaretz. (August 12, 2018). For the attention of Facebook user Benjamin Netanyahu: In a
206
democratic country, different people want all sort of things. [Facebook status update]. Retrieved from https://www.facebook.com/haaretz/posts/10156578534577520 Hagey, K. (2020, June 10). Fox News host Tucker Carlson leaves the Daily Caller. The Wall Street Journal. Retrieved from https://www.wsj.com/articles/fox-news-host-tucker- carlson-leaves-the-daily-caller-11591804588 Hågvar, Y. B. (2019). News media’s rhetoric on Facebook. Journalism Practice, 13(7), 853– 872. doi:10.1080/17512786.2019.1577163 Hallin, D. C., & Mancini, P. (2004). Comparing media systems: Three models of media and politics. New York: Cambridge University Press. Hanitzsch, T. (2007). Deconstructing journalism culture: Toward a universal theory. Communication Theory, 17(4), 367–385. doi:10.1111/j.1468-2885.2007.00303.x Hanitzsch, T., Hanusch, F., Ramaprasad, J., & De Beer, A. S. (Eds.). (2019). Worlds of journalism: Journalistic cultures around the globe. New York: Columbia University Press. Harcup, T., & O’Neill, D. (2001). What is news? Galtung and Ruge revisited. Journalism Studies, 2(2), 261–280. doi:10.1080/14616700118449 Harcup, T., & O’Neill, D. (2017). What is news? News values revisited (again). Journalism Studies, 18(12), 1470–1488. doi:10.1080/1461670X.2016.1150193 Heiss, R., Schmuck, D., & Matthes, J. (2019). What drives interaction in political actors’ Facebook posts? Profile and content predictors of user engagement and political actors’ reactions. Information, Communication & Society, 22(10), 1497–1513. doi:10.1080/1369118X.2018.1445273 Hermida, A. (2016). Social media and the news. In T. Witschge, C.W. Anderson, D. Domingo, & A. Hermida (Eds.), The Sage Handbook of Digital Journalism (pp. 81– 94). London: Sage. Hermida, A., Fletcher, F., Korell, D., & Logan, D. (2012). Share, Like, Recommend: Decoding the social media news consumer. Journalism Studies, 13(5–6), 815–824. doi:10.1080/1461670X.2012.664430 Hermida, A., Lewis, S. C., & Zamith, R. (2014). Sourcing the Arab Spring: A case study of Andy Carvin’s sources on Twitter during the Tunisian and Egyptian revolutions. Journal of Computer-Mediated Communication, 19(3), 479–499. doi:10.1111/jcc4.12074 Hermida, A., & Mellado, C. (2020). Dimensions of social media logics: Mapping forms of
207
journalistic norms and practices on Twitter and Instagram. Digital Journalism. doi:10.1080/21670811.2020.1805779 Hester, J. B., & Dougall, E. (2007). The efficiency of constructed week sampling for content analysis of online news. Journalism & Mass Communication Quarterly, 84(4), 811–824. doi:10.1177/107769900708400410 Hille, S., & Bakker, P. (2013). I like news. Searching for the 'Holy Grail' of social media: The use of Facebook by Dutch news media and their audiences. European Journal of Communication, 28(6), 663–680. doi:10.1177/0267323113497435 International Symposium on Online Journalism (ISOJ) (2015). Transcript for the session “The art and science of audience development: How news organizations navigate a world where they don’t control the distribution channels.” Austin, TX: The University of Texas at Austin. Retrieved from https://isoj.org/wp- content/uploads/2016/12/Day1.Panel3_.pdf Israel Hayom. (2017, November 27). The editor of Israel Hayom: “We are a right-wing newspaper, but do not work for Netanyahu.” Retrieved from https://www.israelhayom.co.il/article/517969 Iyengar, S., Lelkes, Y., Levendusky, M., Malhotra, N., & Westwood, S. J. (2019). The origins and consequences of affective polarization in the United States. Annual Review of Political Science, 22(1), 129–146. doi:10.1146/annurev-polisci-051117- 073034 Jamieson, K. H. (2017, March 6). Jamieson offers new name for fake news: ‘Viral Deception’ or VD. Annenberg Public Policy Center. Retrieved from https://www.annenbergpublicpolicycenter.org/on-cnn-jamieson-offers-new-name-for- fake-news-viral-deception-or-v-d/ Jönsson, A. M., & Örnebring, H. (2011). User-generated content and the news: Empowerment of citizens or interactive illusion? Journalism Practice, 5(2), 127–144. doi:10.1080/17512786.2010.501155 Karnowski, V., Leiner, D. J., Kümpel, A. S., & Leonhard, L. (2020). Worth to share? How content characteristics and article competitiveness influence news sharing on social network sites. Journalism & Mass Communication Quarterly. doi:10.1177/1077699020940340 Katriel, T. (1986). Talking straight: Dugri speech in Israeli Sabra culture. Cambridge, UK: Cambridge University Press.
208
Kelle, U. (2007). The development of categories: Different approaches in grounded theory. In A. Bryant, & K. Charmaz (Eds.), The Sage handbook of grounded theory (pp. 191– 213). London: Sage. Kennedy, T. (2020, February 25). Syracuse.com to end reader comments on website beginning Thursday. Syracuse.com. Retrieved from https://www.syracuse.com/news/2020/02/syracusecom-to-end-reader-comments-on- website-beginning-thursday.html King, G., Pan, J., & Roberts, M. E. (2013). How censorship in China allows government criticism but silences collective expression. The American Political Science Review, 107(2), 326–343. doi:10.1017/S0003055413000014 Kligler-Vilenchik, N., & Tenenboim, O. (2020). Sustained journalist–audience reciprocity in a meso news-space: The case of a journalistic WhatsApp group. New Media & Society, 22(2), 264–282. doi:10.1177/1461444819856917 Klinger, U., & Svensson, J. (2015). The emergence of network media logic in political communication: A theoretical approach. New Media & Society, 17(8), 1241–1257. doi:10.1177/1461444814522952 Klinger, U., & Svensson, J. (2016). Network media logic: Some conceptual consideration. In A. Bruns, G. Enli, E. Skogerbö, A. Larsson & C. Christensen (Eds.), Routledge companion to social media and politics (pp. 23–38). New York: Routledge. Klinger, U., & Svensson, J. (2018). The end of media logics? On algorithms and agency. New Media & Society. doi:10.1177/1461444818779750 Klyueva, A. (2016). Taming online political engagement in Russia: Disempowered publics, empowered state, and challenges of the fully functioning society. International Journal of Communication, 10, 4661–4680. doi:1932–8036/20160005 Krotz, F. (2009). Mediatization: A concept with which to grasp media and social change. In K. Lundby (Ed.), Mediatization: Concept, changes, consequences (pp. 21–40). New York: Peter Lang. Krug, S. (2016). Reactions now available globally. Facebook. Retrieved from https://about.fb.com/news/2016/02/reactions-now-available-globally/ Ksiazek, T. B. (2016). Commenting on the news: Explaining the degree and quality of user comments on news websites. Journalism Studies, 19(5), 650–673. doi:10.1080/1461670X.2016.1209977 Ksiazek, T. B., Peer, L., & Lessard, K. (2016). User engagement with online news:
209
Conceptualizing interactivity and exploring the relationship between online news videos and user comments. New Media & Society, 18(3), 502–520. doi:10.1177/1461444814545073 Larsson, A. O. (2018). “I shared the news today, oh boy”: News provision and interaction on Facebook. Journalism Studies, 19(1), 43–61. doi:10.1080/1461670X.2016.1154797 Lasorsa, D. L., Lewis, S. C., & Holton, A. E. (2012). Normalizing Twitter: Journalism practice in an emerging communication space. Journalism Studies, 13(1), 19–36. doi:10.1080/1461670X.2011.571825 Lawrence, R. G., Molyneux, L., Coddington, M., & Holton, A. (2014). Tweeting conventions: Political journalists’ use of Twitter to cover the 2012 presidential campaign. Journalism Studies, 15(6), 789–806. doi:10.1080/1461670X.2013.836378 Lawrence, R. G., Radcliffe, D., & Schmidt, T. R. (2018). Practicing engagement: Participatory journalism in the Web 2.0 era. Journalism Practice, 12(10), 1220–1240. doi:10.1080/17512786.2017.1391712 Lewis, S. C., Holton A. E., & Coddington, M. (2014). Reciprocal journalism: A concept of mutual exchange between journalists and audiences. Journalism Practice, 8(2), 229– 241. doi:10.1080/17512786.2013.859840 Lewis, S. C., & Molyneaux, L. (2018). A decade of research on social media and journalism: Assumptions, blind spots, and a way forward. Media and Communication, 6(4), 11– 23. doi:10.17645/mac.v6i4.1562 Lewis, S., & Westlund, O. (2015). Actors, actants, audiences, and activities in cross-media news work: A matrix and a research agenda. Digital journalism, 3(1), 19–37. doi:10.1080/21670811.2014.927986 Lewis, S. C., Zamith, R., & Hermida, A. (2013). Content analysis in an era of big data: A hybrid approach to computational and manual methods. Journal of Broadcasting & Electronic Media, 57(1), 34–52. doi:10.1080/08838151.2012.761702 Lis, J. (2019, November 17). Netanyahu at ‘emergency' rally of Likud party faithful: Arab Joint List wants to destroy the country. Haaretz. Retrieved from https://www.haaretz.com/israel-news/elections/.premium-netanyahu-at-emergency- party-rally-arab-joint-list-wants-to-destroy-the-country-1.8136337 Lischka, J. A. (2018). Logics in social media news making: How social media editors marry the Facebook logic with journalistic standards. Journalism: Theory, Practice & Criticism. doi:10.1177/1464884918788472 Livingstone, S. (2013) The participation paradigm in audience research. Communication
210
Review, 16(1–2), 21–30. doi:10.1080/10714421.2013.757174 Livingstone, S., & Markham, T. (2008). The contribution of media consumption to civic participation. British journal of sociology, 59(2), 351–371. doi:10.1111/j.1468-4446.2008.00197.x Loosen, W., & Schmidt, J. H. (2016). Multi-method approaches. In T. Witschge, C. W. Anderson, D. Domingo, & A. Hermida (Eds.), The Sage Handbook of Digital Journalism (pp. 562–575). London: Sage. Lundby, K. (2009). Media logic: Looking for social interaction. In K. Lundby (Ed.), Mediatization: Concept, Changes, Consequences (pp. 101–119). New York: Peter Lang. Makor Rishon. (2019, February 5). A strong right needs a strong newspaper (an ad). Retrieved from https://mashrokit.info/makor/yamin/?ref=twitter Manosevitch, I. (2011). User generated content in the Israeli online journalism landscape. Israeli Affairs, 17(3), 422–444. doi:10.1080/13537121.2011.584670 Manosevitch, I., & Tenenboim, O. (2017). The multifaceted role of user-generated content: An analytical framework. Digital Journalism, 5(6), 731–752. doi:10.1080/21670811.2016.1189840 Media Bias / Fact Check. (2020). About. Retrieved from /https://mediabiasfactcheck.com/about Meraz, S., & Papacharissi, Z. (2013). Networked gatekeeping and networked framing on #Egypt. The International Journal of Press/Politics, 18(2), 138–166. doi:10.1177/1940161212474472 Mitchell, A., Gottfried, J., Barthel, M., & E. Shearer. (2016, July 7). The modern news consumer: News attitudes and practices in the digital era. Pew Research Center. Retrieved from http://www.journalism.org/2016/07/07/social-engagement/ Molyneux, L. (2015). What journalists retweet: Opinion, humor, and brand development on Twitter. Journalism: Theory, Practice & Criticism, 16(7), 920–935. doi:10.1177/1464884914550135 Montgomery, S. (2016, August 17). Beyond comments: Finding better ways to connect with you. NPR. Retrieved from http://www.npr.org/sections/npr- extra/2016/08/17/490208179/beyond-comments-finding-better-ways-to-connect-with- you Morris, W. (2019). The meaning of the scene: When Pelosi clapped at Trump. The New York
211
Times. Retrieved from https://www.nytimes.com/2019/02/20/magazine/pelosi-trump- and-the-spectacle-of-red-meat-politics.html Mourão, R. R., & Molyneux, L. (2020). Tweeting outside the lines: Normalization and fragmentation as political reporters break from the mainstream. Journalism Practice. doi:10.1080/17512786.2020.1771753 Muddiman, A., & Stroud, N. J. (2017). News values, cognitive biases, and partisan incivility in comment sections. Journal of Communication, 67(4), 586–609. doi:10.1111/jcom.12312. Nave, N. N., Shifman, L.,Tenenboim-Weinblatt, K. (2018). Talking it personally: Features of successful political posts on Facebook. Social Media + Society, 4(3), 1–12. doi:10.1177/2056305118784771 Neiger, M., & Tenenboim-Weinblatt, K. (2016). Understanding journalism through a nuanced deconstruction of temporal layers in news narratives. Journal of Communication, 66(1), 139–160. doi:10.1111/jcom.12202 Netzer, Y., Tenenboim-Weinblatt, K., & Shifman, L. (2014). The construction of participation in news websites: A five-dimensional model. Journalism Studies, 15(5), 619–631. doi:10.1080/1461670X.2014.895527 Newman, N. (2011). Mainstream media and the distribution of news in the age of social discovery. Oxford, UK: Reuters Institute for the Study of Journalism, University of Oxford. Retrieved from http://reutersinstitute.politics.ox.ac.uk/sites/default/files/2017- 11/Mainstream%20media%20and%20the%20distribution%20of%20news%20in%20t he%20age%20of%20social%20discovery.pdf Newman, N., Dutton W. H., & Blank, G. (2012). Social media in the changing ecology of news: The fourth and fifth estates in Britain. International Journal of Internet Science, 7(1), 6–22. Newman, N., Fletcher, R., Schulz, A., Andı, S., Nielsen, R. K. (2020). Digital news report 2020. Oxford, UK: Reuters Institute for the Study of Journalism, University of Oxford. https://reutersinstitute.politics.ox.ac.uk/sites/default/files/2020- 06/DNR_2020_FINAL.pdf Olmstead, K., Mitchell, A., & Rosenstiel, T. (2011, May 9). Navigating news online: Where people go, how they get there and what lures them away. Pew Research Center. Retrieved from http://www.journalism.org/2011/05/09/navigating-news-online/ Oremus, W. (2017, March 5). Twitter’s new order. Slate. Retrieved from
212
http://www.slate.com/articles/technology/cover_story/2017/03/twitter_s_timeline_alg orithm_and_its_effect_on_us_explained.html Oxford Dictionary. (n.d. 1). Anger. Retrieved from https://www.lexico.com/en/definition/anger Oxford Dictionary. (n.d. 2). Sadness. Retrieved from https://www.lexico.com/en/definition/sadness Oxford Dictionary. (n.d. 3). Anxiety. Retrieved from https://www.lexico.com/en/definition/anxiety Oxford Learner's Dictionary. (n.d.). News. Retrieved from https://www.oxfordlearnersdictionaries.com/us/definition/english/news Papacharissi, Z., & Oliveira, M. (2012). Affective news and networked publics: The rhythms of news storytelling on #Egypt. Journal of Communication, 62(2), 266–282. doi:10.1111/j.1460-2466.2012.01630.x Peri, Y. (2012). The impact of national security on the development of media systems: The case of Israel. In D. C. Hallin & P. Mancini (Ed.), Comparing media systems beyond the Western world (pp. 11–25). New York: Cambridge University Press. Perrin, A., & Anderson, M. (2019). Share of U.S. adults using social media, including Facebook, is mostly unchanged since 2018. Pew Research Center. Retrieved from https://www.pewresearch.org/fact-tank/2019/04/10/share-of-u-s-adults-using- social-media-including-facebook-is-mostly-unchanged-since-2018/ Pew Research Center. (2019, July 9). Newspapers fact sheet. Retrieved from https://www.journalism.org/fact-sheet/newspapers/ Politi, D. (2019, March 10). Tucker Carlson says awful things about women in newly unearthed radio clips. Slate. Retrieved from https://slate.com/news-and- politics/2019/03/tucker-carlson-bubba-love- sponge.html?fbclid=IwAR2aR9i3MDGG9DRxv8VREywN9UJQzoFpfwFDM1utb3N VxnMfbKsjegsnARM Quandt, T. (2018). Dark participation. Media and Communication, 6(4), 36–48. doi:10.17645/mac.v6i4.1519 Reichertz, J. (2007). Abduction: The logic of discovery of grounded theory. In A. Bryant, & K. Charmaz (Eds.), The Sage handbook of grounded theory (pp. 214–228). London: Sage. Rhyu, J. (2020a, April 6). Facebook breaking news label expands. Facebook. Retrieved from https://www.facebook.com/journalismproject/facebook-breaking-news-label-expands
213
Rhyu, J. (2020b, April 6). Enabling publishers to label breaking news on Facebook. Facebook. Retrieved from https://www.facebook.com/journalismproject/facebook- breaking-news-label Riffe, D., Aust, C. F., & Lacy, S. R. (1993). The effectiveness of random, consecutive day and constructed week sampling in newspaper content analysis. Journalism & Mass Communication Quarterly, 70(1), 133–139. doi:10.1177/107769909307000115 Rogers, K. (2018). Trump says a ‘big operation’ kept wife, Melania, from G-7. The New York Times. https://www.nytimes.com/2018/06/08/us/politics/trump-melania- surgery.html Ryfe, D. (2012). Can journalism survive? An inside look at American newsrooms. New York: Polity. Rudat, A., Buder, J., & Hesse, F. W. (2014). Audience design in Twitter: Retweeting behavior between informational value and followers’ interests. Computers in Human Behavior, 35, 132–139. doi:10.1016/j.chb.2014.03.006 Russell, A. (2016). Networked journalism. In T. Witschge, C. W. Anderson, D. Domingo, & A. Hermida (Eds.), The Sage Handbook of Digital Journalism (pp. 149–163). London: Sage Publications. Schudson, M. (2001). The objectivity norm in American journalism. Journalism: Theory, Practice & Criticism, 2(2), 149–170. doi:10.1177/146488490100200201 Schudson, M. (2000). The domain of journalism studies around the globe. Journalism: Theory, Practice & Criticism, 1(1), 55–59. doi:10.1177/146488490000100110 Schultz, T., Fielding, K., Newton, F., & Louis, W. (2018). The effect of images on community engagement with sustainable stormwater management: The role of integral disgust and sadness. Journal of Environmental Psychology, 59, 26–35. doi:10.1016/j.jenvp.2018.08.003 Shaban, H. (2019, February 7). Twitter reveals its daily active user numbers for the first time. The Washington Post. Retrieved from https://www.washingtonpost.com/technology/2019/02/07/twitter-reveals-its-daily- active-user-numbers-first-time/ Shah, D. V. (2016). Conversation is the soul of democracy: Expression effects, communication mediation, and digital media. Communication and the Public, 1(1), 12–18. doi:10.1177/2057047316628310 Sheafer, T., & Wolfsfeld, G. (2009). Party systems and oppositional voices in the news
214
media: A study of the contest over political waves in the United States and Israel. International Journal of Press/Politics, 14(2), 146–165. doi:10.1177/1940161209333089 Shearer, E., & Matsa, K. E. (2018, September 10). News use across social media platforms 2018. Pew Research Center. Retrieved from https://www.journalism.org/2018/09/10/news-use-across-social-media-platforms- 2018/ Sherman, A. (2020, July 30). U.S. lawmakers agree Big Tech has too much power, but what to do about it remains a mystery. CNBC. Retrieved from https://www.cnbc.com/2020/07/30/us-lawmakers-agree-big-tech-has-too-much- power-remedies-unclear.html Shoemaker, P. J., & Cohen, A. A. (2006). News around the world: Content, practitioners, and the public. New York: Routledge. Shoemaker, P. J., & Reese, S. D. (1996). Mediating the message: Theories of influences on mass media content (2nd ed.). New York: Longman. Shoemaker, P. J., & Reese, S. D. (2014). Mediating the message in the 21st century. New York: Routledge. Shoemaker, P. J., & Vos, T. P. (2009). Gatekeeping theory. New York: Routledge. Silverman, C. (2009, November 19). Profiles in courage: Social media editors at big media outlets. MediaShift. Retrieved from http://mediashift.org/2009/11/profiles-in-courage- social-media-editors-at-big-media-outlets323/ Singer, J. B. (2005). The political j-blogger: ‘Normalizing’ a new media form to fit old norms and practices. Journalism: Theory, Practice & Criticism, 6(2), 173–198. doi:10.1177/1464884905051009 Singer, J. B., Domingo, D., Heinonen, A., Hermida, A., Paulussen, S., Quandt, T., Reich, Z., & Vujnovic, M. (2011). Participatory Journalism: Guarding Open Gates at Online Newspapers. Oxford: Wiley-Blackwell. Srugim. (2020). Srugim – the home site of the religious sector. Srugim.co.il. StatCounter. (2020). Social media stats Israel. Retrieved from https://gs.statcounter.com/social-media-stats/all/israel Strömbäck, J., & Esser, F. (2014). Mediatization of politics: Transforming democracies and reshaping politics. In K. Lundby (Ed.), Mediatization of Communication (pp. 374– 404). Berlin: De Gruyter Mouton. Stroud, N. J. (2017). Attention as a valuable resource. Political Communication, 34(3), 479–
215
489. doi:10.1080/10584609.2017.1330077 Stroud, N. J., & Muddiman, A. (2019). Social media engagement with strategy- and issue- framed political news. Journal of Communication, 69(5), 443–466. doi:10.1093/joc/jqz029 Stroud, N. J., Scacco, J. M., & Curry, A. L. (2016). The presence and use of interactive features on news websites. Digital Journalism, 4(3), 339–358. doi:10.1080/21670811.2015.1042982 Stroud, N. J., Scacco, J. M., Muddiman, A., & Curry, A. L. (2015). Changing deliberative norms on news organizations’ Facebook sites. Journal of Computer-Mediated Communication, 20(2), 188–203. doi:10.1111/jcc4.12104 Stroud, N. J., Van Duyn, E., & Peacock, C. (2016). News commenters and news comment readers. Report by the Center for Media Engagement (formerly the Engaging News Project). Austin, TX: The University of Texas at Austin. Retrieved from https://mediaengagement.org/wp-content/uploads/2016/03/ENP- News-Commenters-and-Comment-Readers1.pdf Strübing, J. (2007). Research as pragmatic problem-solving: The pragmatist roots of empirically-grounded theorizing. In A. Bryant, & K. Charmaz (Eds.), The Sage handbook of grounded theory (pp. 580–602). London: Sage. Taylor, G. (2014). Scarcity of attention for a medium of abundance. In M. Graham & W. H. Dutton (Eds.). Society and the Internet: How networks of information and communication are changing our lives. Oxford: Oxford University Press. Tenenboim, O. (2017). Reporting war in 140 characters: How journalists used Twitter during the 2014 Gaza-Israel conflict. International Journal of Communication, 11, 3497– 3518. doi:1932–8036/20170005 Tenenboim, O., & Cohen, A. A. (2015). What prompts users to click and comment: A longitudinal study of online news. Journalism: Theory, Practice & Criticism, 16(2), 198–217. doi:10.1177/1464884913513996 Tenenboim, O., & Kligler-Vilenchik, N. (2020). The meso news-space: Engaging with the news between the public and private domains. Digital Journalism, 8(5), 576–585. doi:10.1080/21670811.2020.1745657 Tenenboim-Weinblatt, K. (2014). Producing protest news: An inquiry into journalists’ narratives. The International Journal of Press/Politics, 19(4), 410–429. doi:10.1177/1940161214540941 Tenenboim-Weinblatt, K., Hanitzsch, T., & Nagar, R. (2016). Beyond peace journalism:
216
Reclassifying conflict narratives in the Israeli news media. Journal of Peace Research, 53(2), 151–165. doi:10.1177/0022343315609091 Tenenboim-Weinblatt, K., & Neiger, M. (2018). Temporal affordances in the news. Journalism: Theory, Practice & Criticism, 19(1), 37–55. doi:10.1177/1464884916689152 Tewksbury, D. (2003). What do Americans really want to know? Tracking the behavior of news readers. Journal of Communication, 53(4), 694–710. doi:10.1111/j.1460- 2466.2003.tb02918. The New York Times. (October 13, 2017). Social media guidelines for the newsroom. Retrieved from https://www.nytimes.com/editorial-standards/social-media-guidelines.html The New York Times. (August 13, 2018). We’re live with Stephen Colbert, Emmy- and Peabody Award-winning comedian, writer, producer and television host, in conversation with Sopan Deb, culture reporter for the New York Times. Get ready for laughs, and if you have any questions, leave them in the comments and Deb will ask some. [Facebook status update]. Retrieved from https://www.facebook.com/nytimes/posts/2072350509483710 Tracy, M. (2020, August 5). Digital revenue exceeds print for 1st time for New York Times Company. The New York Times. Retrieved from https://www.nytimes.com/2020/08/05/business/media/nyt-earnings-q2.html Tracy, S. J. (2013). Qualitative research methods: Collecting evidence, crafting analysis, communicating impact. Malden, MA: Wiley-Blackwell. Trilling, D., Tolochko , P., & Burscher, B. (2017). From newsworthiness to shareworthiness: How to predict news sharing based on article characteristics. Journalism & Mass Communication Quarterly, 94(1), 38–60. doi:10.1177/1077699016654682 Tuchman, G. (1972). Objectivity as strategic ritual: An examination of newsmen’s notions of objectivity. The American Journal of Sociology, 77(4), 660–679. doi:10.1086/225193 Tuchman, G. (1978). Making news: A study in the construction of reality. New York: The Free Press. Usher, N. (2014). Making news at the New York Times. Ann Arbor: University of Michigan Press. Valenzuela, S., Piña, M., & Ramírez, J. (2017). Behavioral effects of framing on social media users: How conflict, economic, human interest, and morality frames drive news sharing. Journal of Communication, 67(5), 803–826. doi:10.1111/jcom.12325
217
van Dijck, J., & Poell, T. (2013). Understanding social media logic. Media and Communication, 1(1), 2–14. doi:10.12924/mac2013.01010002 Verthaim, D. (2018, July 31). TGI survey – first half of 2018: Israel Hayom and Yedioth Ahronoth neck and neck on the weekend; The Marker plunges. Walla (in Hebrew). Retrieved from https://b.walla.co.il/item/3177392 Vox Staff. (2020, April 8). Ethics and guidelines at Vox.com. Retrieved from https://www.vox.com/2018/12/7/18113237/ethics-and-guidelines-at-vox-com Vujnovic, M., Singer, J. B. Paulussen, S., Heinonen, A., Reich, Z., Quandt, T., Hermida, A., & Domingo, D. (2010). Exploring the political-economic factors of participatory journalism: Views of online journalists in 10 countries. Journalism Practice, 4(3), 285–296. doi:10.1080/17512781003640588 Wahl-Jorgensen, K. (2016). Emotion and journalism. In T. Witschge, C.W. Anderson, D. Domingo, & A. Hermida (Eds.), The Sage Handbook of Digital Journalism (pp. 128– 143). London: Sage. Weber, P. (2014). Discussions in the comments section: Factors influencing participation and interactivity in online newspapers’ reader comments. New Media & Society, 16(6), 941–957. doi:10.1177/1461444813495165 Webster, J. G. (2014). The marketplace of attention: How audiences take shape in a digital age. Cambridge, MA: The MIT Press. Welbers, K., & Opgenhaffen, M. (2019). Presenting news on social media: Media logic in the communication style of newspapers on Facebook. Digital Journalism, 7(1), 45–62. doi:10.1080/21670811.2018.1493939 Website IQ. (2020). Website traffic statistics. Retrieved from https://www.websiteiq.com/ Wendelin, M., Engelmann, I., & Neubarth, J. (2017). User rankings and journalistic news selection: Comparing news values and topics. Journalism Studies, 18(2), 135–153. doi:10.1080/1461670X.2015.1040892 Westlund, O., & Ekström, M. (2018). News and participation through and beyond proprietary platforms in an age of social media. Media and Communication, 6(4), 1–10. doi:10.17645/mac.v6i4.1775 White, D. M. (1950). The ‘gate keeper’: A case study in the selection of news. Journalism Quarterly, 27(4), 383–396. Witschge, T., Anderson, C. W., Domingo, D., & Hermida, A. (2019). Dealing with the mess (we made): Unravelling hybridity, normativity, and complexity in journalism studies.
218
Journalism: Theory, Practice & Criticism, 20(5), 651–659. doi:10.1177/1464884918760669 Wu, L. (2016). Did you get the buzz? Are digital native media becoming mainstream? ISOJ Journal, 6(1). Retrieved from https://isojjournal.wordpress.com/2016/04/14/did-you- get-the-buzz-are-digital-native-media-becoming-mainstream/ Zago, G. D. S. & Bastos, M. T. (2013). Visibility of news items on Twitter and Facebook: Comparative analysis of the most replicated news in Europe and the Americas. Brazilian Journalism Research, 9(1), 114–131. doi:10.25200/BJR.v9n1.2013.568 Zelizer, B. (2018). Epilogue: Timing the study of news temporality. Journalism: Theory, Practice & Criticism, 19(1), 111–121. doi:10.1177/1464884916688964 Ziegele, M., Weber, M., Quiring, O., &, Breiner, T. (2018a). The dynamics of online news discussions: Effects of news articles and reader comments on users’ involvement, willingness to participate, and the civility of their contributions. Information, Communication & Society, 21(10), 1419–1435. doi:10.1080/1369118X.2017.1324505 Ziegele, M., Quiring, O., Esau, K., & Friess, D. (2018b). Linking news value theory with online deliberation: How news factors and illustration factors in news articles affect the deliberative quality of user discussions in SNS’ comment sections. Communication Research. doi:10.1177/0093650218797884
219