RESPONSES TO BLACK AND WHITE NEWS ANCHORS: EFFECTS OF TWEET TYPES, AND MODERATION BY HUMANITARIANISM/EGALITARIANISM MOTIVATIONS By Linda R. White A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of Information & Media – Doctor of Philosophy 2023 ABSTRACT Increasingly network affiliate stations are hiring reporters and anchors of color, with the belief that racially diverse newsrooms provide coverage that satisfies the diverse news needs of the many Americans who rely on local television news. Also increasingly, network affiliate stations encourage their on-air professionals to use social media, to post on Facebook, Instagram, and Twitter. This dissertation first asks from a structural racism point of view, how African Americans/Blacks and Whites response to African American/Black and White news anchors and their stories. It then asks how audiences respond when these anchors are said to not tweet, tweet a neutral news statement, or to tweet their own opinion about controversial issues. The tweet impact is theorized to result from the degree to which the two types of tweets depart from norm violations for professional journalists—who tweet the news, but generally eschew tweeting their own opinions. Results show Whites are less positive about African American/Black than White anchors, and that there is negativity about opinionated tweets. Further, the results show complicated relationships among anchor race, audience race, type of tweet, age, and political affiliation. Keywords: ANOVA, humanitarianism-egalitarianism, expectancy violation theory, novelty, TV news anchors Copyright by LINDA R. WHITE 2023 This dissertation is dedicated to my parents Rev. Dr. Edward O. and Jo Ellen White, my daughter Ashley, my brother David and his wife JoAnn, their children Paige, Zachary, and Taylor, my extended family, my ancestors, and my Heavenly Father, Jesus Christ. iv ACKNOWLEDGEMENTS I’d like to thank my Heavenly Father for sending me to Godly parents. They taught me the value of faith, family, knowing the word of GOD, and of a great education. That education began in their home but continued in the Akron Public School System (Ohio), Ohio University for my undergraduate degree in Broadcast Journalism (Athens, Ohio), Ball State University for my master’s in journalism (Muncie, Indiana) and finally here in the Journalism Department at Michigan State University for my PhD in Information & Media (East Lansing, Michigan). I’d like to thank my brother David and his wife JoAnn, who became surrogate parents to me upon the passing of both our father (2001) and mother (2017) and encouraged me, listened to my complaints, and most importantly prayed for me during this process. I want to acknowledge my daughter Ashley and her partner Brodreck for their encouragement as well as my extended family and friends in Ohio, Indiana, Alabama, and Tennessee for their encouragement and prayers. To my Michigan State family, friends, and cohort who I shared dinners with, adult beverages with, and who helped me through this program, just know you are friends for life! I’m so proud of all of us. Special shout outs to Bob Gould, Mike Castelluci, Teanna Barnes, Miriam Bingham, Erin Bowling, Asya Lawrence, Chloe Porforino, Marvin Pride for their help with my stimuli. And lastly to my advisor and dissertation committee co-chair Dr. Esther Thorson, dissertation co-chair Dr. Saleem Alhabash, dissertation committee member and interim CAS Dean, Dr. Teresa Mastin, and dissertation committee member and journalism grad program director, Dr. Rachel Mourao, as well as I&M PhD program director, Dr. Pat Huddleston, v Journalism department chair Dr. Tim Vos and all the wonderful professors in our program, thank you for all that you do! It’s been a wonderful ride! vi PREFACE After spending 25 years working as an anchor/reporter in the Midwest and the Southeast, I returned to higher education to obtain my masters and PhD degrees in 2016. This research project is part of a larger research umbrella plan to study racial diversity in television news. For my masters I conducted a content analysis of television news stories to learn the impact of BIPOC news directors on diverse staff, stories, and covering communities of color and their television ratings. This project takes a look at audience attitudes of African American/Black and White anchors. My hope is that my scholarship contributes to improving the diversity in television newsrooms, the television news product, and how newsrooms serve their diverse audiences. vii TABLE OF CONTENTS LIST OF ABBREVIATIONS .......................................................................................... ix Chapter 1 Introduction ..................................................................................................... Chapter 2 Racial Stereotypes and Structural Racism ...................................................... Chapter 3 Expectancy Violations Theory and Journalists’ Tweets ................................. Chapter 4 Journalistic Norms and Practices and Expectations About Them................... Chapter 5 Overview of the Theoretical Perspectives and Hypotheses ............................ Chapter 6 Experimental Design and Methodology.......................................................... Chapter 7 Results ............................................................................................................. Chapter 8 Discussion, Limitations, and Conclusion ........................................................ BIBLIOGRAPHY ............................................................................................................ 1 7 18 29 39 42 51 81 88 APPENDIX A: Pretest I Tweets ...................................................................................... 100 APPENDIX B: Pretest II Tweet Pairs ............................................................................. 102 APPENDIX C: News Consumption Scale....................................................................... 103 APPENDIX D: Social Media Use by Journalists ............................................................ 104 APPENDIX E: Story Evaluation Scale............................................................................ 105 viii LIST OF ABBREVIATIONS AMP ATB Affect Misattribution Procedure Attitude toward Blacks BIPOC Black, Indigenous, and People of Color CRT EF EMS EV EVT HE IAT IMS Critical Race Theory Executive Function External Motivations Expectancy Violation Expectancy Violation Theory Humanitarianism-Egalitarianism Implicit Association Test Internal Motivations LGBTQ Lesbian, Gay, Bisexual, Transgender, Queer/Questioning MCPR Motivations to Control Prejudice Reactions PE SJ Protestant Ethic Social justice ix Chapter 1 Introduction In 2020, millions in the U.S. took to the streets to peacefully protest the killing of a Black man in Minneapolis, Minnesota – George Floyd, and the hundreds more who came before him (Kishi & Jones, 2020). Millions more marched in solidarity around the world. The U.S. had seen plenty of racial injustice protests dating back to the 1960s. Like the 1960s, the 2020 protests included large numbers of White faces. Some of these new faces of protestors begin the deep dive discovering and understanding the systemic racial injustices not just in law enforcement but in education, health, and housing among many others. This experimental research project seeks to investigate similar biases associated with television news anchors and a news user’s choice of a news anchor in addition to the role social media and prejudice play in that choice. Social justice movements like Black Lives Matter, #MeToo, and Stop Asian Hate, have influenced every industry in the U.S. including broadcast news organizations. There have been calls to diversify the workforce. Reporter positions have been created to cover how news impacts Black, Indigenous, and People of Color (BIPOC) communities in TV news (Tameez, 2022). But television news continues to have a diversity problem (Papper, 2022). Various entities have called for BIPOC parity and/or equal treatment of BIPOC in news coverage (ASNE, 2016; Butler, 2012). One of the earliest was the 1947 Commission on the Freedom of the Press report (Press, 1947). Created through the University of Chicago, this commission sought to answer the question was the freedom of the press in danger. They concluded yes and proceeded to offer solutions. One solution addressed accurate images of Chinese and Black people. They warned against hate speech and feeding stereotypes, concluding that understanding one’s culture can lead to greater respect for it and them. The most famous group to call for the creation of jobs to level the economic playing field for Blacks was the National Advisory Commission on Civil 1 Disorders, AKA the Kerner Commission (Disorders & Commission, 1968). It was formed to investigate the 1967 race riots. The commission concluded the riots were due to inequality in banking, housing, education and were caused by White racism (George, 2018). The commission’s 426-page document about race inequality in the U.S. included a long list of recommendations for a variety of industries including news – both print and broadcast. The commission concluded the journalism profession had done a disservice in not hiring, training, and promoting Blacks and should do a better job of including them in newsrooms. Without minority voices, race riot coverage was flawed and inaccurate. For example, it was reported that there was Black on White crime, instead it was majority Black on Black crime. While the report was largely ignored by the federal government when it was released (George, 2018), media companies took it seriously and attempted to diversify its staff. But Black, Indigenous, and People of Color reporters struggled to cover more than just race related and/or gender related stories and had to prove they were just as good as their White counterparts to cover breaking news and more serious topics (Byerly & Wilson II, 2009). Today minorities and women lead newscasts on both network and local news, but there are still newsrooms where diverse voices are missing (Tameez, 2022). Journalists and responses to racial occurrences in the U.S. Today’s journalists aren’t just gathering, producing, and writing the news. They must also maintain a social media presence to interact with news users. News companies want their journalists to be accessible on social media platforms (Weaver et al., 2019). In my 25-year career as a television anchor/reporter, social media use began as an experiment. It was used to promote stories and develop sources. In addition to doing that, now journalists are expected to interact with audiences. My last media employer now tracks social media engagement in newsrooms. Most of these are pleasant interactions, but there are some negative circumstances that have cost 2 journalists their jobs. Meteorologist Rhonda Lee was fired at a Mississippi station for defending her hair on her station’s Facebook page (Smith, 2012). Pittsburg anchor Wendy Bell lost her job when she spoke negatively about the alleged suspects in a quadruple homicide on her Facebook page (Post, 2016). In 2022, Felicia Somnez lost her job at the Washington Post for what the paper called disparaging comments about her colleagues and the workplace culture on Twitter (Darcy, 2022). Both Lee and Somnez maintain they were defending themselves. So, in this day and age, do we have an expectation of how news anchors should behave on social media? Obviously, employers do, but what about viewers? What can anchors/reporters say online? What shouldn’t they say? Additionally, how would social justice issues in social media posts impact the credibility of the news anchor and/or the content? Do the public hold TV news journalists to different standards based on race when they talk about opinions? We know that journalistic roles are ever changing, especially with the advent of social media platforms. In September 2022, RTDNA issued social media guidelines to be boldly objective and scrutinize social media activity in order to regain trust from local communities (How to Improve Trust in Local Journalism, 2022). Their suggestions were to not allow social media to be a place to debate objectivity and to avoid personal opinions on controversial topics. This dissertation explores first the role of race of on-air anchors in TV news. It then explores how the kind of social media posts the anchors are reported to have made influence attitudes toward the Black and White anchors. Expectancy violation theory (EVT) is used to guide expectations about how news vs news opinion posts affect attitudes toward Black and White TV journalists. EVT explains when these types of posts have an impact on anchor and story attitudes. Personal opinion vs. news-related posts are used to create possible expectancy expectations. Burgoon (1978) suggested that how we utilize and perceive spatial relationships in 3 communication is directly related to our expectations and how those expectations are realized or broken. In other words, our interpretation of the significance of spatial connections is influenced by our past knowledge and experience, and when those expectations are not realized, the communication suffers. Overall, Burgoon and Jones (1976) argue that spatial relationships are crucial in communication and that knowing how expectations and violations of those expectations influence our perception of those relationships is critical to good communication. The next sections will uncover literature impacting this dissertation. Chapter 2 addresses racial stereotypes and structural racism. Scholars have investigated how and why people remember stereotypes about people in various social categories and express discrimination in negative behaviors. You’ll read about their conclusions, how some measure stereotypes, as well as how critical race theory (CRT) and intersectionality impact this study about television news anchors. These frameworks explain how society views female television news anchors and the attitudinal challenges Black women bear in relation to their daily news assignments. Chapter 3 reviews expectancy violation theory (EVT) and humanitarianism- egalitarianism (HE) attitudes. EVT focuses on how a person will respond to an unexpected violation of social media action by a news anchor. This literature addresses how will the unexpected event impact the likeability of the news anchor and the stories she reads. The HE scale measuring racial attitudes, places a priority on the wellbeing and equality of all people regardless of their social category. It is a belief that everyone is equal and has access to opportunities and resources. How HE impacts what a news user thinks will be addressed in this chapter. I’ll include details on the Katz and Hass (1988) research investigating kindness and 4 ambivalence without directly asking how a survey participant felt about a person of a different race and ethnicity. Chapter 4 examines journalism norms and practices overall and in relation to how journalists navigate social media. We’ve seen a change in how television news companies require journalists to be present on social media platforms in the last 15 years. They face challenges in this growing medium. Minoritized journalists have an extra burden in expectations of what to cover both when the industry added them to staffs in the early 1970s but also today. Journalists navigate these concerns while including social media posts when the public isn’t always appreciative and welcoming. You’ll find a summary of the theoretical perspectives, hypotheses, and research model in Chapter 5. Here I explain what I expect to happen. You’ll also view a detailed model of the research hypotheses and research question. Chapter 6 elaborates on the experimental design and methodology. The independent variables of anchor race and expectancy violation and the moderator variable humanitarianism- egalitarianism are further explained. You’ll see photos of the anchors and learn about the stories in the experiment. The dependent measures, control variables, and additional survey scales are provided. Additionally, efforts to avoid confounds in the videotaping of the anchors is detailed along with the experimental procedure and data analysis plan. Chapter 7 discusses the analysis of two tweet pretests and the main study data. Two pretests were conducted to determine the best tweets to provoke an expectancy violation. Factor and reliability analysis determined the factor loading of the various scales used. ANOVAs, univariate ANCOVAs, bivariate correlation, pre-planned post hoc pairwise comparisons, and moderation (PROCESS Model 1 and 2) were used to analyze the data. 5 The implications of the findings are found in Chapter 8 and their connection to novelty in television. Novelty is described by scholars as the unexpected or new revelation. News professionals may find this information helpful in how social media posts could be used to attract younger audiences and BIPOC to linear television. At the conclusion of this dissertation, you’ll understand how this research adds to the body of knowledge regarding the usage of expectancy violation theory scale and humanitarianism-egalitarianism attitude scale in television news anchor evaluation. To my knowledge, they have not been used in this manner and on this topic within empirical research. 6 Chapter 2 Racial Stereotypes and Structural Racism “Challenging power structures from the inside, working the cracks within the system, however, requires learning to speak multiple languages of power convincingly.” ~ Patricia Hill Collins, On Intellectual Activism Newborns enter the world as a blank slate. The world is their oyster is an oft quoted phrase, meaning, there’s an infinite amount of knowledge to learn or not learn. This research delves into the psychological mechanisms of prejudice, stereotypes, bias, and racism supporting the hypotheses and research questions proposed by this dissertation. These triggers develop into negative behavior and evaluation toward marginalized people. Scholars are divided about how and why these triggers develop. The following is what the relevant literature says about prejudice, stereotypes, bias, and racism and their impact on the television news industry. Stereotypes There are many reasons why people use stereotypes – to feel superior, to continue inequalities, and/or to provide a status quo paradigm – ‘Why change things now?’ to list a few (Bodenhausen & Macrae, 2013). Stereotypes have been defined as generalized judgments, which are either negative or positive, on every member of a social or cultural group (Allport, 1958; Jones et al., 2013; Karlins et al., 1969). Bodenhausen and Macrae (2013) implied a lack of self- esteem led people to stereotype others. They believed hierarchical control played a role in stereotyping. In this process, structures in the brain called “superordinate goal states” oversee “subordinate components” to control stereotyping due to the demands of societal pressures (p. 20). The former is higher order functioning of the brain, the latter, lower order. This higher and lower order functioning of the brain provides the mechanism to recall memories, to form an opinion and to act on it. Lower order operations are effortless and easily guide a person 7 throughout the day like walking in the crosswalk when the crossing light is green. This is what Bargh and Chartrand (1999) called automaticity. Bodenhausen and Macrae (2013) used the example of how to treat a new neighbor. Superordinate goals of treating others equally, fairly, and with kindness, overruled any lower-level order or subordinate goal to recall stereotypes of the neighbor, leading to negative treatment of the new neighbor based on negative stereotypes. Bargh and Chartrand (1999) described a conscious (i.e., controlled) versus automatic mental process. Controlled processing of stereotypes and stereotypic portrayals requires forethought, adding an element of cognitive control. In other words, an individual’s conscious and controlled stereotypic thinking and expression is the outcome of elaborate thinking that requires redistribution of mental energy. Automatic processing deals with making decisions without much forethought. In the case of automatic processing, stereotyping occurs with minimal cognitive effort and with lack of control for the emergence of stereotypic thoughts and expressions (Bargh & Chartrand, 1999; Devine & Sharp, 2009). Bargh and Chartrand (1999) believed the conscious forethought operated at a snail’s pace and ran out of steam quickly. The automatic process is fast paced. On the other hand, Thagard (2019) described stereotypes as a combination of “concepts, beliefs, and emotions” controlled by one’s mental energy (p 109). In this case, when one encounters a member of a particular social or cultural group, automatically, the brain retrieves previously stored attributes of that group and applies them to perceiving the individual based on stereotypic thoughts. The author also said, when the power structure is lopsided, one has more power over another, stereotypes provide a semblance of structure (Hall, 1997). Gorham (1999) said stereotypes for good or bad help us understand the world in which we navigate. The author operationalized misconceptions about different racial groups according to social categories. In essence, these myths could be half-truths. Blaine (2007) said in thinking about stereotypes, 8 misconceptions develop by relying on biased data in our stored memory causing potentially impaired actions, reactions, or inactions towards others. The top three social categorizations are age, race, and sex (Stolier & Freeman, 2016). The age category is based on the number of years a person is alive (Scherbov & Sanderson, 2016). Scholars commonly think of sex as anatomical and chromosomal differences (Johnson & Repta, 2012). In the United States, race does not divide by country, language, or ancestry, but by physical traits (Lucal, 1996). Class, educational rank, and religion are other examples of social categories. Social class can be defined by socioeconomic status. It has been divided up into five categories: lower class, working class, middle-class, upper middle class, and upper class (Bird, 2017). In their investigation of community college versus university graduates, Anisef et al. (1992) uncovered, community college graduates do not attain higher status careers than their university counterparts. In analyzing a survey of college graduates, the researchers used fifteen variables in a multiple regression with current occupation as the dependent variable. The social category of religion plays a role in life satisfaction as Yaden et al. (2022) uncovered in their meta-analysis of more than 1300 journal articles about religion/spirituality and life satisfaction. Coding for 16 variables and five dimensions of religiosity, they concluded there was a significant positive relationship between religion/spirituality and life satisfaction. Social categories play a role when people experience stereotyping and negative behaviors. Jim Crow laws in the south, separate but equal educational systems throughout the United States and the Japanese internment camps are a few examples of how prejudice influenced various laws in U.S. and state governments (Allport, 1958). Jones (1972) tells of an incidence of prejudiced by individual(s) on two men returning to a U.S. airport after vacationing in Jamaica. Custom agents searched their luggage and their bodies for illegal contraband 9 presumably drugs. Jones (1972) said when customs agents found nothing, the men were later x- rayed. When still nothing was found, they were released. In a verbal exchange between the men and a customs agent, the author said when the men asked why they were targeted, the agent asked them what was their nationality and their age? They responded Hispanic and mid -twenties. The agent intimated those were the reasons why they were targeted (Jones, 1972). How we view the social roles of others can impact how we treat them or our expectations of them, individually or as a group. I believe this will impact my study as the survey taker accesses the conscious forethought by controlling their stereotypes but will run out of steam and be unable to control it later in the survey. Measuring Stereotypes Scholars have attempted to measure stereotypes, prejudice, and explicit and implicit bias. Princeton researchers tested stereotypes with its students three times at pivotal moments in history after World War I, World War II, and during the Vietnam War (Devine & Elliot, 1995; Gilbert, 1951; Karlins et al., 1969). Devine and Elliot (1995) recreated the Princeton trilogy scholarship by updating the stereotype phrases within the scale. They observed high- and low- prejudiced respondents were knowledgeable about the updated stereotypes about Blacks specifically, but low-prejudiced respondents had personal beliefs that were opposite, whereas high-prejudice individuals had personal beliefs that reverberated the stereotypical beliefs. Madon et al. (2001) also revisited the Princeton trilogy with updated stereotype phrases. They concluded marginalized groups faced different stereotypes and may be exposed to them more often than previous generations. Some scholars believe racism is alive and well. Racism is a result of racial prejudice and the power that one group exerts over another because of perceived, known and unknown, 10 difference (Jones, 1972). Jones et al. (2013) said racism is a systematic interaction including biases such as prejudice and discrimination, which are perpetuated by societal and cultural factors. In other words, racism is more than just an individual occurrence, it is profoundly rooted in a society’s social and cultural systems, and it is maintained by both conscious and unconscious prejudices. Because of this coordinated interaction, individuals and groups may get uneven treatment and opportunities based on their race or ethnicity. Arhin and Thyer (2004) approached racial prejudice through the lens of respondent learning or classical conditioning, operant learning, and learning via observation. According to the authors respondent learning can result in automatic reactions to certain environmental cues. Reactions can be both positive and negative which can alter the development of certain behaviors. In other words, when an individual is repeatedly exposed to events, they may build a conditioned response based on the emotions connected with it. When an individual has learned to associate a specific trigger with a specific response, this process is frequently automatic and not under conscious control. According to Arhin and Thyer (2004), the process through which an individual’s behavior is influenced by the consequences it causes in their environment is referred to as operant learning. This suggests that the individual’s conduct gets stronger or is weakened as a result of the good or negative effects. A reward that increases the likelihood of a behavior is positive reinforcement. Negative reinforcement removes unpleasant feedback, increasing the chance of the behavior being repeated. Punishment on the other hand, involves either the introduction of a negative consequence or the removal of a positive stimulus, both of which reduce the chance of the conduct being repeated. This type of learning has the potential to shape future behavior in a particular environment. Arhin and Thyer (2004)’s observational learning involves witnessing the 11 actions of others. Observed actions can be linked to behavior, influencing the development of social conduct. CRT & Intersectionality Some legal scholars have explained societies treatment of African American women with intersectionality. This argument is important in how news users and society as a whole view anchors who are women of color. Intersectionality has been used as an analytical tool to highlight the inequalities in society’s long-held institutions in its treatment of BIPOC (Crenshaw, 1990). As a legal scholar, Crenshaw (1990) intended to explain how Black women were double discriminated against because they were women and of color. Ultimately, she pointed to intersectionality’s usefulness in multiple disciplines defining structural intersectionality and political intersectionality. The former in its treatment of women who are domestic violence and rape victims. The overlapping endorsement of feminist and antiracist policies, which, according to her interpretation, basically opposed one another and deepened the marginalization of women of color. Essentially, it’s the inability to recognize patriarchy in one and race in the other. Collins and Bilge (2020) suggested a half dozen strategies in intersectionality analysis: relationality, power, social inequality, context, complexity, and social justice. Social inequality analysis provides acknowledgement that some interactions with others can lead to unequal institutional treatment. Within domains of power, a scholar will find an overlap of identities such as race, class, gender, and sexuality among others. Relationality examines interconnections, not differences. For social context, a researcher would combine social inequality, relationality, and power relations. Either individually or when used together, these analyses provide complexity to how a person is treated based on their various social and racial identities. The social justice lens is often used by those battling systemic institutional racism. 12 Intersectionality analysis has been used by researchers to explain its impact in other disciplines and regions of work. Rigoni (2012) analyzed European women of color in media using intersectionality. She found more women of color are leaders and managers within ethnic media in comparison to legacy media. European women of color often battle inequalities in mainstream media more than in ethnic media. The author found they experience segregation in divergent ways. In explaining the premise of horizontal segregation, the author said work tasks were assigned based on gender. Another example of segregation was described as vertical in nature. Vertical segregation is the equivalent of the glass ceiling. Women who value their roles in the traditional family sense, tend to avoid managerial roles and if they are in those positions, they are childless. The author also found other women hoping to avoid unwanted sexual advances or keeping the nature of their personal lives out of the boardroom and their workplace, may miss out on managerial opportunities. Additionally, Rigoni (2012) uncovered that European women of color see their journalistic roles as activists for both women and people of color. This suggests when BIPOC aren’t making a way for themselves, the hegemonic globalization remains. In essence, within the Global North, South, and developing nations, the elites have the power, developing countries grow into a similar power structure, and inequality increases in media and among the non-elites (Artz, 2003). Crenshaw (2018) originally wrote about intersectionality in 1989. Since then, it’s been cited more than 25,600 times. While some scholars may not directly engage in intersectionality scholarship, it can be seen indirectly. Blaine (2007) examined gender stereotypes and sexism, Klein and Shiffman (2009) analyzed race, gender, and lack of LGBTQ characters in animated cartoons, and Ryan and Mapaye (2010) conducted a content analysis of network news determining an uptick in diversity among reporters. There is no doubt intersectionality is an 13 effective analysis tool. It puts inequality in our various institutions, both physical and social, on full view. Structural racism as outlined here leads to Hypothesis 1. Hypothesis 1: Participants will express more favorable evaluations of the anchor in terms of (a) attractiveness, (b) trustworthiness, and (c) expertise, as well as (d) the news stories they read when the anchor is White than African American/Black. Main Effect of Anchor Race Much effort has been expended looking for measures of racism. McConahay et al (1981) developed the Modern Racism scale. They compared antiquated (old fashion) racist ideology and phrases with 1970s (new) racist ideology and phrases. Their 2 x 2 experiment changed the race of the experimenter and the type of scale (old fashioned vs new). They found that the older racist statements were impacted by the race of the experimenter because participants found them offensive, while the more modern racist statements paired with a Black or White experimenter weren’t impacted. According to their results, the modern racist statements weren’t seen as offensive by participants. They concluded Whites recognized old fashioned racism but didn’t see their other negative behaviors like personal judgements, their strongly held convictions, and their intentional or unintentional behaviors toward others based on racial stereotypes against Blacks as prejudice. In assessing the racists attitudes of college students, Brigham (1993) validated the attitude toward Blacks (ATB) scale when Reverend Jesse Jackson was a candidate in the 1988 presidential election. Their scale correlated with direct assessments of prejudice made in anonymous, low-pressure environments. There have been many efforts to measure racism in a way that bypasses conscious experiences of racism. The Implicit Association Test (IAT ) is a thought experiment that 14 measures implicit (unconscious) attitudes (Greenwald et al., 1998). The idea is that the test reveals negative attitudes towards a person or object for participants who don’t want to reveal them. The authors conducted three computerized experiments using pleasant and unpleasant to describe things and people. The first involved flower and insects. The second was Japanese and Korean names and the third was Black and White names. Following the computer test, participants took paper/pencil tests on race related attitudes and beliefs. The authors concluded their experiments led to an effective measure of prejudice. Others have tested the effectiveness of the IAT. In their study, Amodio and Devine (2006) developed a new IAT to examine the ability to measure implicit bias through a dual categorization implicit association test (IAT) measuring prejudice and stereotyping in three different studies. Study 1 utilized IAT by asking participants to link pleasant or unpleasant words to either Black or White faces. To stereotype IAT, a new research design asked participants to match mental or physical description words to Black or White faces. Studies 2 and 3 were designed to replicate Study 1 and to test the hypothesis that implicit evaluation and stereotyping predict discriminatory behaviors. The study results of Amodio and Devine (2006) indicated implicit evaluation and implicit stereotyping were separate constructs and that implicit evaluation predicted physical forms of discriminatory behaviors while implicit stereotyping predicted mental forms of discriminatory behaviors. In three experiments, Greenwald et al. (1998) tested discrimination in flowers versus insects, Korean names versus Japanese full names and Korean names versus shortened Japanese names, and names associated with Black and White families. They too found the IAT was effective in uncovering hidden prejudice. Bernstein et al. (2010) also used IAT in their research to determine bias before and after the election of President Barack Obama in 2008. Their results showed decreased implicit bias. They also 15 discovered explicit attitudes didn’t change because initial attitudes were already favorable. Sternadori (2017) also discovered a decrease in implicit bias when they tested empathetic news stories. On the other hand, one of the first studies to introduce an alternative to IAT was Payne et al. (2005). They defined affect misattribution procedure (AMP) as the tendency for individuals to misattribute their emotional reactions to a stimulus, such as a picture or word, to a different source than the actual cause of the emotion. They argued the AMP, in addition to its research results, is simpler to administer and takes less time. They also believed their test showed better validity and reliability in explicit bias compared to IAT. Their sixth experiment tested attitude bias. Black and White participants viewed 24 photos, half of them were of African American men and the other half were White men. Participants also took an explicit bias test. The authors hypothesized Whites would prefer the photos of White people over the African American ones. They also hypothesized explicit bias would be present in those not interested in controlling bias. Their results discovered those who don’t care what others think about them regarding prejudice were linked to implicit and explicit measures. Those who wanted to control racial bias, didn’t show a connection between implicit and explicit tests. Payne et al. (2005) believed the AMP was better than the IAT at exposing bias. Other attitude scales have been deemed an alternative to the IAT and AMP measures in detecting prejudice. Plant and Devine (1998) tested the validity of their internal (IMS) and external (EMS) motivations to withhold prejudice scale against the ATB and motivation to control prejudice reactions (MCPR) scales. Their multi-part study scale items tested a person’s unease with inner prejudices and what others would think about their prejudices. Plant and 16 Devine (1998) believed their scale went a step further than the MCPR in testing both internal and external motivations and evaluating them separately or together. Ito et al. (2015) tested Plant and Devine (1998)’s internal and external motivations of prejudice in compilation with Dunton and Fazio (1997)’s motivation to control prejudice response and Attitudes Toward Blacks scales (Brigham, 1993) when studying how executive function (EF) impacts results. They found internal motivations weren’t impacted by EF. Their results and Plant and Devine (1998)’s study provide a prejudice measure without the complexity of photos and cultural stimuli of the IAT and AMP. 17 Chapter 3 Expectancy Violation Theory and Journalists’ Tweets "In recognizing the humanity of our fellow beings, we pay ourselves the highest tribute." –Thurgood Marshall, in Furman v. Georgia In this chapter an overview of EVT is provided including its development and its application to expectations about social media posting by professional TV journalists. Journalists have a desire for attention from audiences in their use of social media environments (Wilhelm et al., 2021). It has become essential to create interactions between journalists and audiences, and of course this also leads to the greater opportunity to monetize a news product. The journalist and news consumer relationship evolution continues in light of a new level of visibility in the digital age from positive and negative digital interactions (Wilhelm et al., 2021). This study proposes to use expectancy violation theory (EVT) as a lens to view what happens in the relationship between news anchors and their audience following a social justice referenced tweet. It is a useful framework for understanding how people interpret and respond to unexpected behaviors in communication, and it has proven to be a useful tool for researchers studying various communication environments, including social media and television news. EVT began as an anthropological term to explain the physical and emotional space in personal relationships known as proxemics (Burgoon, 2015). Burgoon (2015) explained spatial distances can be determined in a variety of ways due to age, culture, and personality. Burgoon and Jones (1976) described expectancy violation as an “invisible, dynamic, and transportable space” managed by an individual (p 131). It involves two participants: the initiator, who is held accountable for the act and the respondent (Burgoon, 2015; Burgoon & Jones, 1976). The suggestion is that the impact of what happens in these relationships can be measured in what is 18 now called expectancy violations (Burgoon, 1993). In other words, our society has norms for various people – family, celebrities, doctors, neighbors, politicians, and more. We hold them to a certain flame that can burn bright, hot, or be snuffed out. These are various reactions to a negative or positive new behavior. We like it, dislike it, embrace it, endure it, or demand accountability when the behavior crosses the line. Burgoon (2015) said when expectations are not satisfied, they are referred to as “expectancy violations”, and when they are, they are referred to as “expectancy confirmations” (p 3). Key concepts include expectations, communicator reward valence, arousal‐distraction, the interpretation–evaluation appraisal process, and violation valence (Burgoon, 2015). According to Burgoon (2015), there can be four different outcomes because expectancy confirmations and violations can be either positive or negative. The four outcomes are a positive confirmation equals a likely outcome; a negative confirmation equals a likely outcome but committed by a negative persona; a positive persona who approaches nearer than anticipated is a positive violation; and a negative violation can be defined as a negative persona who approaches nearer than anticipated (Burgoon, 2015). Contrary to the long-accepted use of EVT, Burgoon (2015) said newer research assumes positive violations can result in outcomes that are more favorable than positive confirmations, whereas negative violations can have an opposite effect and produce outcomes that are more unfavorable than confirmations. While EVT began as a theory defining personal space violations, scholars have expanded its use into a variety of communicative areas including close relationships (Afifi & Metts, 1998), customer service (Houston III et al., 2018), gaming (Zhou et al., 2023), parasocial relationships with celebrities (Walther-Martin, 2015), phubbing (Kadylak, 2020), AI journalism (Waddell, 2018), social media interaction (Bevan et al., 2014; Grinberg et al., 2017; Lee et al., 2020), 19 stereotypical behavior (Bettencourt et al., 1997), and verbal communication (Bennett et al., 2020; Kim, 2014; Yuan & Lu, 2020). Afifi and Metts (1998) created a typology of EVT from three studies in personal relationships. They defined EVT as an unexpected behavior that comes with three significant violations incurring a range of emotion (violation valence), power (violation expectedness and importance), and impact on insecure relational outcomes (violations and uncertainty). Their results indicated positive violations outweighed negative ones. Zhou et al. (2023) wanted to determine how fast unhappy gamers would abandon gaming software or game apps. To determine how long players would endure competitions within the game, they adapted expectancy violation items in three ways: reward, achievement, and competition. They discovered players whose expectations aren’t met, meaning the rules aren’t clear, will quickly abandon the game. Walther-Martin (2015) expanded EVT into parasocial relationships when they conducted an experiment using a comedic celebrity. In their 2 (African American v White source) x 2 (political ideology) between-subject experimental design, it was discovered political ideology moderated source evaluation. Overall, evaluators rated the African American comedian higher than the White comedian. The author concluded expectancy violation can be experienced in multimedia messaging even when there is no personal relationship with the source and violations do not always have to be negative. But what can be seen as negative is phubbing or monitoring your cell phone during face-to-face interactions. Kadylak (2020) surveyed senior citizens about cell phone etiquette and phubbing using EVT. As one might expect, the survey participants who found this behavior abhorrent admitted they had negative feelings about the practice which led to symptoms of depression and loneliness. News users also expressed negative feelings about AI 20 authorship. Waddell (2018) conducted two experiments utilizing M-Turk participants in scholarship regarding EVT. Participants read either articles by a person or artificial intelligence. Waddell (2018) discovered that news written by AI was perceived to be less credible. Several scholars have used EVT to study social media interactions. Bevan et al. (2014) studied the business of unfriending on Facebook and found the person on the receiving end of being unfriended, is not particularly fond of the practice. The authors adapted Afifi and Metts (1998)’s EVT scale for a Qualtrics survey. Bevan et al. (2014) uncovered it’s a higher expectancy violation when the person doing the unfriending was considered a close friend. Grinberg et al. (2017) also studied Facebook but looked at reactions to Facebook posts or statements. They wanted to know what the expectations for likes and comments were. They discovered when people responded in the comments of a post, and what type of post it was (celebrating success) was more important than liking the post. Lee et al. (2020) studied how politicians use Twitter to connect with their constituents. While they did not test EVT, they used the theory as a guide in their research. Two studies of fictitious politician tweets either personal or scandalous determined participants judged politicians positively even after the scandalous tweet. The authors concluded Twitter users saw the politicians as being more authentic. Their outcome variable was Twitter as the medium. If conducted today, any of these social media studies might have different outcomes based on changes in how people use Facebook and after Elon Musk’s purchase of Twitter. In their scholarship regarding EVT and stereotypical behavior, Bettencourt et al. (1997) conducted three job applicant experiments. Two experiments used job applicant scenarios, one from the football team, the other from the speech team. The third experiment tested job applicants and their resumes. EVT was determined to be a mediator in how participants 21 evaluated either the football player or speech team member. Those who violated expectations were rated more positively. Their research produced “expectancy-violating”, expectancy- inconsistent, and “expectancy-consistent” behaviors (p. 272). They explained that violating behaviors and consistent behaviors were on opposite ends of the spectrum with inconsistent ones between the two. Additionally, they concluded evaluation may depend on whether the evaluated is a known persona or unknown persona. If unknown, severe responses to these scenarios will be the result of expectancy violations but if known, severe responses develop when norms are violated. In the area of verbal communication, scholars have used EVT to study emotional responses, effective crisis communication messages, and climate change messaging. In their research the emotional responses, Bennett et al. (2020) adapted Afifi and Metts (1998)’s EVT scale to study three message types of violation expectedness, violation valence, and violation importance, testing the experience of hurt, surprise, and anger. They discovered the message matters and can determine how people feel after receiving the message. Additionally, participants experienced surprise when the violation was unexpected, however anger and hurt weren’t connected to violation expectedness. They further revealed a negative connection between hurt and violation valence and a positive connection to violation importance. They concluded negative messages from known personas can lead to an erosion of the relationship. Kim (2014) studied the relationship between a corporation and its consumers in using EVT. To prevent that relationship from declining, they discovered companies should be honest with their consumers during and after a crisis. They conducted both survey and experimental phases in their scholarship. The first phase looked at the pre-crisis expectations, and the second phase dealt with post-crisis messages. The author discovered participant’s evaluation of the crisis 22 communication, depended on whether participants liked the company or not prior to the crisis. If they liked the company, they were less critical and if they didn’t like the company, they were more critical, which confirmed EVT. Yuan and Lu (2020) also confirmed EVT in their investigation of climate change messaging. They conducted an online experiment on aggressive messaging. In their 2 (communication style: more vs. less aggressive) × 2 (target of address: deniers vs. we Americans) between-subjects factorial design, they modified a news article to be more or less aggressive. They found the more the participant denied there was a climate problem, the more the participant saw the message as being aggressive. They also discovered the results were moderated by political ideology. While the majority of EVT scholarship uses the EVT scale directly or adapted it, Houston III et al. (2018) instead used the theory to guide their research on customer service and interactions between a Black or White salesperson and a Black or White customer. They hypothesized and discovered White customers have a higher expectation of positive sales service than Black customers because of the past experiences of Black customers and poor customer service due to their race. The role of expectancy violations related to journalist’s tweets There have also been a number of research projects covering EVT and journalists for which this scholarship has particular interest. The following section will dive deeper into these projects. For TV news anchors, we know they are judged more harshly than actors, singers, and sports media personalities (Cohen, 2010). They discovered TV hosts were held to a higher standard in the categories of social and trust. Women were impacted negatively by all four expectancy violations that the author studied. Two of those violations dealt with societal morality. They were major moral (infractions), using drugs or evading taxes, and minor moral 23 (infractions) which included violations like shoplifting and romantically cheating. The author concluded these violations have the potential to damage the long-term relationship with the celebrity persona. In Lee’s (2015) research, audiences reacted negatively when journalists interacted with them in social media. In that online experiment of a fictitious journalist on Facebook, audiences liked the interaction, but it was seen as unprofessional. They also determined that negative comments could impact the journalist negatively. But that is not the case for everyone. Young social media users were not fazed by negative comments. In fact, they appreciated the honesty on the social media platform stud ied (Cohen, 2010; Houston et al., 2020; Lee, 2015). Credibility was also studied by Masullo et al. (2021). They conducted two experiments about incivility in online comment threads in online news stories. They used two survey services. Study 1 utilized M-Turk and study 2 utilized Dynata. Participants were exposed to news comment threads through a mock news website. After reading comments, survey participants responded to 17 descriptors such as credible, believable, attractive, and fair to name a few. Masullo et al. (2021) uncovered hostile or uncivil thread comments led news users to not only view the story negatively, but the news company negatively as well. Gong and Eppler (2022) explored anchor delivery mistakes using EVT. They discovered viewers have an expectation for mistake free news anchors. They uncovered a maximum of three mistakes that was acceptable to viewers. This research informed by these scholars predicts a negative reaction to a social justice tweet, positing that social media messages seen as negative could impact viewers negatively lessening the credibility of the anchor and the content by the survey participant. 24 Humanitarianism-Egalitarianism Attitudes Many psychologists consider humanitarianism a primary attitude (Ferguson, 1944; Fraser & Murakami, 2022). While there are variations in the definition of humanitarianism, according to Barnett (2005) an often agreed upon definition is “the impartial, independent, and neutral provision of relief to those in immediate danger of harm” (p. 724). International relief aid organizations use this definition as a guiding principle to help those in need in war-torn communities and to those escaping these conditions (Barnett, 2005). There are also many definitions of egalitarianism. It often refers to relationships among men and women (Beere et al., 1984; King & King, 1986). However, egalitarianism references a fundamental idea that all humans are in some ways equal (Sheehy-Skeffington & Thomsen, 2020; Sigman & Lindberg, 2019). Both humanitarianism and egalitarianism are related to ways that humans see themselves in relationship to others different from themselves, and therefore are concepts importantly related to racial perceptions (Gaither et al., 2020; Gallagher, 2015; Kteily et al., 2017). A humanitarianism-egalitarianism (HE) attitude scale was developed by Katz and Hass (1988). They set out to test two contrasting scales that would measure the divisive attitudes and behaviors that arise between individuals of different races due to their uncertain or conflicting emotions, beliefs, and opinions towards one another. At its essence, this belief asserts that all individuals should have an equal opportunity to thrive and succeed in our world. It acknowledges that disparities in resources and opportunities can create various obstacles for certain marginalized groups or individuals, resulting in systemic disadvantages and inequalities. The goal of advocating for equal access to these opportunities and resources is to level the playing field and establish a more fair and just society. 25 To further elaborate on this attitude, it ensures equal access to opportunities such as education, employment, healthcare, housing, and legal rights. It means that every person, regardless of their background or situation, should have an equal chance to pursue education, gain knowledge, secure employment with a living wage, receive adequate healthcare services, access safe and affordable housing, and enjoy equal protection under the law. The basis of this scale also recognizes the importance of addressing historical and systemic injustices that have contributed to ongoing inequality. It acknowledges that certain marginalized communities have faced historical disadvantages and discrimination that continue to affect their access to opportunities and resources today. Achieving equality, therefore, necessitates addressing and rectifying these systemic issues to create a more inclusive and just society. Promoting equal access to opportunities and resources goes beyond merely ensuring an even distribution of goods and services. It also involves addressing the social, economic, and political structures that perpetuate inequality. This may involve implementing policies and initiatives that challenge systemic discrimination, foster inclusivity, and provide targeted support to historically marginalized and disadvantaged groups. The inherent questions within the HE scale acknowledges the value and dignity of every individual and strives for a world where everyone can flourish and realize their full potential, regardless of their background or circumstances. Other scholars have used this scale to measure parent/child volunteerism (White, 2021), attitudes towards protecting endangered species (Harnish et al., 2023), attitudes towards refugees (Fraser & Murakami, 2022), the decision to attend graduate school for social work, (Osteen et al., 2023), and understanding the impact of an HE intervention on the IAT (Röhner & Lai, 2021). 26 Humanitarianism and egalitarianism should be predictive of how people perceive African American and White anchors. Those with high HE attitudes would be expected to be far more accepting of racial differences in professional on-air anchors. Thus, an important approach would be to measure the presence of the two attitudinal structures and use them to predict responses to African American and White anchors. The 10-item HE scale (Katz & Hass, 1988) includes the following: One should be kind to all people, one should find ways to help others less fortunate than oneself, a person should be concerned about the well-being of others, there should be equality for everyone—because we are all human beings, those who are unable to provide for their basic needs should be helped by others, a good society is one in which people feel responsible for one another, everyone should have an equal chance and an equal say in most things, acting to protect the rights and interests of other members of the community is a major obligation for all persons, in dealing with criminals the courts should recognize that many are victims of circumstances, and prosperous nations have a moral obligation to share some of their wealth with poor nations. This study predicts those who don’t have those same beliefs will not like the African American anchor as much as they like the White anchor. HE will moderate the anchor and story evaluation. Application of the concepts of humanitarianism and egalitarianism leads to the following hypotheses: Hypothesis 2: HE will moderate the effect of anchor race on perceived anchor (a) attractiveness, (b) trustworthiness, (c) expertise, and (b) story evaluation, such that those high on the HE will show less negativity toward African American/Black anchors and the stories they deliver than those low 27 on the HE scale, while HE will not lead to differences in evaluating the White anchors and their stories. Hypothesis 3: Participants will express more favorable evaluation of the anchor in terms of perceived (a) attractiveness, (b) trustworthiness, and (c) expertise, as well as more favorable (d) news story evaluations upon exposure to the news story with no tweet (control), followed by neutral tweet (low expectancy violation) and social justice tweet (high expectancy violation), respectively. Hypothesis 4: There will be a significant interaction between anchor race and expectancy violation on (a) perceived attractiveness, (b) perceived trustworthiness, (c) perceived expertise, and (d) story evaluations, such that participants will express more favorable evaluations toward anchor and stories by a White anchor than an African American/Black anchor when she tweets about a social justice issue (expectancy violation condition) compared to neutral and no tweet, while they will express less favorable anchor and story evaluations upon exposure to stories by an African American/Black anchor following a social justice tweet, compared to neutral and no tweets conditions. 28 Chapter 4 Journalistic Norms and Practices, and Expectations About Them “The way to right wrongs is to turn the light of truth upon them.” ~ Ida B. Wells, Investigative Journalist Journalism is a key component of communication that has the potential to shape and grow knowledge and opinion but also to tear down and destroy self-esteem and self-worth. When it is constructed objectively, and that is not always the case, it educates, enlightens, entertains, informs, and more (Deuze & Witschge, 2018; Hanitzsch & Vos, 2016; Raeijmaekers & Maeseele, 2017). The nearly 400-year-old institution is the practice of informing the public about events – locally, nationally, and internationally (Schudson, 2012). The basic principles of journalism remain constant, objectively reporting, perhaps playing a watchdog role, however journalists themselves and media companies are constantly evolving, shaped by a complex coaction of historical, cultural, social, and technological factors. These factors may include the advent of social media platforms, adapting new revenue models, changing public attitudes about news, all while navigating journalism’s norms and values. With today’s deadlines, journalists often find themselves at a crossroads between what they say they do, should do, want to do, and what they actually do. That is the practice of journalism as described by Hanitzsch and Vos (2017), narrated (what journalists say they do), normative (what journalists say they should do), cognitive (what journalists want to do), and practiced (what journalists actually do). Norms, formats, and routines are what constitute a shared ideology among journalists such as a commitment to truth, accuracy, fairness, impartiality, independence, and accountability (Deuze, 2019). Additionally, Deuze (2019) said journalists see themselves as individuals working for media companies, not media company employees, meaning that journalists have values that include public service (seeing themselves 29 as watchdogs), autonomy (free and independent work), immediacy, and a sense of ethics. Deuze (2019) also argued these norms and values are synonymous with daily rituals and tasks. This chapter explores journalistic norms, a journalist practitioners role using social media, and what behavior is expected of them in the realm of social media. Institution An institution has three characteristics: authority over some sector of society; structured habits and routines, and the capacity to expand over time. As an institution, journalism continues to be known as the Fourth Estate, playing a watchdog role in our society over other institutions (Wahl-Jorgensen & Hanitzsch, 2009). As part of a democratic society, that watchdog role holds other institutions accountable. But news producers also provide a forum for debate, helping to shape the way people understand and interact with the world around them (Steensen & Westlund, 2020). However, it’s important to note that journalism’s role in a democratic society is complex and multifaceted. It’s also subject to continuous debate and discussion among scholars, policymakers, and the public (Zelizer, 2004). Certainly, journalism can have a positive impact on democracy, but it can also be subject to bias, manipulation, and censorship. These negative impacts can be influenced by a range of factors including business pressures, political whims, and social and cultural norms. Scholars and professionals vary in their interpretation of journalism as an institution (Zelizer, 2004). Some speak of a crisis for print, low advertising revenue, and the decrease in trust of media (Nerone, 2013; Steensen & Westlund, 2020; Wahl-Jorgensen & Hanitzsch, 2019). Others disagree – and believe journalism is adjusting to a different business model while maintaining a public service and watchdog role (Deuze & Witschge, 2018; Hanitzsch & Vos, 2016; Schudson, 2012). 30 There is also debate about what type of institution journalism is. Is it social, cultural, or political? Wahl-Jorgensen and Hanitzsch (2009) argued that journalism comprises all three institutional types. Hanitzsch and Vos (2017) argued journalism is not a political institution but a social and discursive one, meaning discursive work by and between journalists around the world, is consistently re-shaping the institution. There are challenges to journalism as an institution. Wahl-Jorgensen and Hanitzsch (2019) argued that the digital era, social media, and the decline of trust in journalism are prominent challenges of contemporary journalism. Zelizer (2004) described the institution as powerful enough to shape opinion and control the information flow to viewers and readers. Given that few media companies mostly controlled by White men own a majority of news producing television stations here in the United States, I came to realize the importance of interpreting journalism institutions through the lens of societal and political power and the imbalance of power that comes with it for journalists today and in the future (Ehrlich, 2019; Macker, 2020). Discourse The discursive struggle is palpable for the journalism institution regarding its meaning and role in society (Hanitzsch & Vos, 2016). Hanitzsch and Vos (2016) said the internal discourse within journalism is that practitioners have the power to control the journalism narrative. The internal discourse for the journalist is that she juggles multiple hats as a witness to history, storyteller, with a main focus of objectivity. The discourse isn’t just internal as mentioned above but journalism also shapes public discourse (Peters & Carlson, 2019). It’s shaped in what communities and stories are and are not covered. Daily, newsroom gatekeepers decide what’s news especially when issues aren’t 31 apparent or forthright (e.g., breaking). This application of discourse extends to the public who may also decide what is covered by pressing the “like” button on a viral video. Identity/roles Peters and Carlson (2019) described theories of journalism and democracy as usually normative which tackle democratic norms that includes its use as a communication tool for other institutions like government to convey events, ideas, and decisions to residents, voters, and constituents. This symbiotic relationship can also be adversarial due to journalism’s role in our society as a watchdog of public and government institutions (Deuze, 2005; Wahl-Jorgensen & Hanitzsch, 2009). In that watchdog role, journalists give a voice to the voiceless, hopefully providing multiple sides of an issue to readers, listeners, and viewers. Those receiving these messages learn about injustices, unfair practices, inequality in their communities, and more, thus contributing to the democratic deliberative process within the public sphere. Journalist uses of Social Media Noted challenges within broadcast journalism include credibility, layoffs, and successfully deploying an ever-evolving profitable business model in the era of new technologies that include mobile internet and social media (Chan, 2019). Smart phones with internet capability help journalists receive real time information from consumers about news events. New technologies and the internet that have made jobs like audio and video editing obsolete and the requirement for multitasking and multiplatform skills have changed the way journalists do their jobs in the last 20 years (Weaver et al., 2019). That assistance from the public can also hinder journalists whose owners seek to monetize news. It can be challenging to media companies when information about a news-related event is already in a social environment/public forum, to draw consumers to watch or read their news daily for monetization. 32 The websites that allow internet users to interact with one another in real time are commonly referred to as social media (Aichner et al., 2021; Van Looy, 2022). Several people connected to AOL, an early internet email service, and Matisse, an online media environment, take credit for creating and popularizing the term (Bercovici, 2010). In Aichner et al. (2021)’s systematic review of 25 years of social media scholarship of how the name social media has evolved, they discovered scholars have also used the terms virtual communities and social networks in the past while describing the same phenomena. Social media uses Web 2.0, a technical platform to allow internet users to interact (Van Looy, 2022). Aichner et al. (2021) said social media became the commonly used term after 2010. Social media websites include blogs (e.g., Medium, WordPress), social networking sites (e.g., Facebook, Twitter) and virtual worlds (e.g., Minecraft, Second Life) to name a few (10 Best Virtual Worlds for Adults, 2022; Haan, 2023; Watermeyer, 2012). With the popularity of social media networks, broadcast journalism had to find a way to engage viewers who had migrated to these platforms back to their television channels whether on TV or their channel’s website. What started as an experiment for many of us working during this timeframe became mandatory with company social media policies (Hedman, 2020; Kempton & Connolly Ahern, 2021; Santana & Hopp, 2016). Social media is one of those technologies that has changed how journalists approach stories daily. Social media is an accepted toolkit in the way news is gathered, sourced, and distributed to audiences (Bruns & Nuernbergk, 2019; Molyneux & Mourão, 2019; Neuberger et al., 2019; Zubiaga, 2019). The norms and routines around social media are continuously evolving. What began as an experiment, testing social media platforms to promote stories, perhaps find sources has become a routine (Dutta & Gangopadhyay, 2019; Santana & Hopp, 2016). Now broadcast news practitioners openly engage with their audience. About a third of 33 journalists surveyed by Weaver and Willnat (2016) post comments on their work-related social networking sites and another third reply to comments. In its survey of journalists, they found evidence that social media has had a positive impact on journalists to interact with news consumers, get information to them quickly, and market themselves and their stories. While journalists divide their time between providing and interpreting news events, the majority choose to inform the public about news (Bodrunova et al., 2018; Schudson, 2001). In their interviews of journalists from newspaper, broadcast television, and online only media companies, Twitter and Facebook were used most for breaking news, real-time reporting, and to distribute videos about a news story (Neuberger et al., 2019). Vázquez-Herrero et al. (2022) analyzed the use of TikTok by 19 news media and programs from around the world with verified profiles. They discovered the news media is using TikTok to inform and brand their product in order to attract TikTok audiences to their product. Journalists also share their opinions on social media. Cozma and Chen (2013) found journalists have a tendency to share both in social media and as long as it’s not controversial, Johnson (2020) found news users don’t care. In their experiment, Johnson (2020) discovered that news users neither agree nor disagree when reporters share thoughts from their personal lives on their news social media accounts. Houston et al. (2020) discovered non- opinionated shares by news producers in social media were perceived as more informative, accurate, and trustworthy. However, opinionated posts were viewed as more engaging and emotional especially by twenty-somethings who have grown up in the social media age. Additionally, employers may require news professionals to make their presence known on social media platforms (Hedman, 2020; Lysak et al., 2012). Before this researcher left the industry, reporters were required to post to social media about their story before leaving the newsroom for their first interview. This interaction on social media, a formula for some, has been 34 determined to increase an account’s following and popularity (Kozman & Cozma, 2021). Determined to monetize social media, some media companies keep track of social media engagement in newsrooms (Kempton & Connolly Ahern, 2021). In their interviews with journalists, Kempton and Connolly Ahern (2021) discovered organizations also have workplace manuals for social media coverage and public interaction. During the 2020 racial reckoning many journalists dealt with objectivity head -on as they covered the movement and debated internally how to externally express their feelings about what they were covering. In their ethnographic study about journalists who covered the movement, Harlow (2022) found of the 28 journalists interviewed, seven posted a square on the Instagram page in support of #Blackout Tuesday, an action to protest perceived racism and police brutality. Four were White. Most of the journalists in this study thought protesting crossed the line but a social media post for BLM expressed empathy instead of bias. There is, however, a negative side to social media interaction between TV news professionals and their audience. Weaver et al. (2019) said new job duties increased job unhappiness. Journalists reported social media threatened accuracy and journalism integrity (Weaver et al., 2019). Additionally, anchors and reporters have been harassed on various platforms, especially young women (Lewis et al., 2020). They concluded negative interactions may dissuade broadcast journalists from interacting with the public. There are those who discard harassment and have become desensitized to it, just like covering negative news stories day in and day out. What minoritized journalists believe they should cover The Kerner Commission encouraged news companies to hire more people of color in their newsrooms in all positions but specifically as managers and reporters. When BIPOC were 35 first hired they were pigeonholed into covering communities of color and their issues (Gans, 2004). Some journalists of color have struggled with their identity versus journalistic norms in how they cover stories. Pritchard and Stonbely (2007) discovered in their content analysis of the Milwaukee Journal Sentinel, minority journalists were more likely than their White counterparts to cover BIPOC issues. Nishikawa et al. (2009) found some journalists of color will argue against only covering issues in their communities and felt being an accurate and balanced journalist was more important. They thought of themselves as journalists first, and a person of color was secondary when working. Others in their research told of having to use a subtle form of advocacy, clearing up misinformation in the newsroom about marginalized communities. Still others spoke of embracing advocacy because of the historical reference of being left out of the conversation and of coverage. As a former television anchor/reporter, I relate to the latter. I embraced covering communities of color, the related issues, and providing positive images of people of color for the viewing audience. Meyers and Gayle (2015) interviewed African American women from television and print media who echoed that same sentiment. The subjects not only used strategies to provide positive Black role models to contradict any stereotypical coverage but when it came to sources, they also advised them on their appearance. I nodded my head in understanding. I discovered as did others, these practices can help the anchor/reporter with their sources for future stories and confirm a reputation for having “your sources back.” This refers to making sure they look presentable during the interview. In every detail of every day as an anchor/reporter, BIPOC are crafting positive images of themselves, building relationships, and producing quality work – which means conquering stereotypes of competence and intelligence while providing evidence, they have a right to work in the newsroom (Brigham, 1993; Cha & Roberts, 2019). 36 In this era of racial reckoning, more BIPOC anchors/reporters understand their important place recording history and how these moments are covered and from whose perspective (Walker, 2022). Yes, news norms are still key – watchdog, inform, educate, but Walker (2022) uncovered African American journalists who wanted to be connected to their communities and their race was primary, while being a journalist was secondary. They also discovered journalists were conflicted covering matters of race. Reporters obviously supported Black Lives Matter. Some of their White counterparts openly supported BLM. But under the watchful eye of community members, Black reporters believed they were held to a different standard. They consciously kept objectivity in the forefront in how they’re seen covering stories impacting race. But what about what people expect about social media posting? News user reactions can vary widely depending on the individual's experiences and biases. Possible reactions could be agreement and support, disagreement and criticism, sharing and liking, engagement and dialogue, or harassment and trolling. Thus far I have hypothesized that there will be more negative responses to African American anchors, and that there will be more negative responses to anchors who tweet about personal opinions about race. And that race and expectancy violations about tweets will create more negativity for African American than for White anchors. However, I also suggest that high HE will diminish the negativity for African American anchors. Will it also reduce the negative effects of personal tweet perceived violations on attitudes toward African American anchors? In other words, is there an interaction between high and low HE and the effects of the tweet conditions? I approach this as a research question: Research Question 1: How does HE moderate the effect of the interaction between anchor race and tweet condition on (a) perceived attractiveness, (b) perceived trustworthiness, (c) perceived expertise, and (d) story evaluation. 37 Combining Racism, Expectancy Violations, and Humanitarian/Egalitarian Motives Theory that combines HE with expectancy violation negativity effects and applied to situations where racism attitudes may occur is shown in Figure 1, which is developed in the next chapter. In this model HE moderates the effects of race of the anchor/reporter and the effect of perceived violation of expectations about appropriate tweets. This model and the hypotheses derived from it will be tested in the experiment described in the method section. 38 Chapter 5 Overview of the Theoretical Perspectives and Hypotheses In this chapter the theoretical perspective developed is summarized and integrated with all the hypotheses and the research question. First, based on racism and structural racism theory: H1: Participants will express more favorable evaluations of the anchor in terms of (a) attractiveness, (b) trustworthiness, and (c) expertise, as well as (d) the news stories they read when the anchor is White than African American/Black. H2: HE will moderate the effect of anchor race on perceived anchor (a) attractiveness, (b) trustworthiness, (c) expertise, and (b) story evaluation, such that those high on the HE will show less negativity toward African American/Black anchors and the stories they deliver than those low on the HE scale, while HE will not lead to differences in evaluating the White anchors and their stories. Third, I expect that tweets that violate expectations from viewers about their appropriateness (i.e., personal racial opinion vs tweets about the news versus no tweets) people will have more negative views of those anchors who do personal racial tweets. H3: Participants will express more favorable evaluation of the anchor in terms of perceived (a) attractiveness, (b) trustworthiness, and (c) expertise, as well as more favorable (d) news story evaluations upon exposure to the news story with no tweet (control), followed by neutral tweet (low expectancy violation) and social justice tweet (high expectancy violation), respectively. 39 Fourth, there will be an interaction between negativity toward anchors and stories such that the negativity toward anchors who tweet racial opinions will be greater for African American anchors than for White anchors. H4: There will be a significant interaction between anchor race and expectancy violation on (a) perceived attractiveness, (b) perceived trustworthiness, (c) perceived expertise, and (d) story evaluations, such that participants will express more favorable evaluations toward anchor and stories by a White anchor than an African American/Black anchor when she tweets about a social justice issue (expectancy violation condition) compared to neutral and no tweet, while they will express less favorable anchor and story evaluations upon exposure to stories by an African American/Black anchor following a social justice tweet, compared to neutral and no tweets conditions. Finally, I ask whether HE will moderate the interaction between anchor race and the effects of tweet expectancy violations: RQ1: How does HE moderate the effect of the interaction between anchor race and tweet condition on (a) perceived attractiveness, (b) perceived trustworthiness, (c) perceived expertise, and (d) story evaluation. Shown below is a model that brings together all of the theoretical expectations. As can be seen, race (African American) is expected to negatively impact anchor and story evaluation. Second, when a personal racial tweet violates a subject’s expectations, it makes their attitude toward anchor and story more negative. This effect is even stronger when race is African American than when it is White. HE moderates the effect of negative racial bias toward the African American and White anchors by reducing it for African American anchors. It also 40 possible that HE mediates the effect of expectancy violations by the personal tweets by reducing the negativity effect of the violation. Figure 1. Theoretical Model 41 ace eporter Eval.Story Eval.Expect. Viol.HE a a a a a Chapter 6 Experimental Design and Methodology This study employs a 2 (Anchor Race: African American v White) x 3 (Expectancy Violation: social justice vs. neutral vs. no tweet) x 2 (story topic: education test scores vs. garbage strike) mixed factorial design with repeated measures on topic repetition. This design investigates the effects of these independent variables on anchor and story evaluations. Participants are exposed to two video stories in random order about a pending garbage strike (neutral) and educational test scores (race-related) delivered by either an African American or White anchor. Story topic is a repeated measure, not an independent variable. Participants will be exposed to the same two stories in one of six experimental conditions: African American anchor with SJ tweet, African American anchor with neutral tweet, African American anchor with no tweet, White anchor with SJ tweet, White anchor with neutral tweet, and White anchor with no tweet. Based on the G*power, 300 participants will be needed. A Michigan State University Information and Media program dissertation research grant funded payment for survey takers at the market rate for the length of the survey. Stimuli Figure 2 Anchor Photos 42 Figure 2 (cont’d) Anchor #1 (view anchor #1 video here) Anchor #2 (view anchor #2 video here) Anchor #3 (view anchor #3 video here) Anchor #4 (view anchor #4 video here) Conditions: BL Anchor 1: Story 1 BL Anchor 1: Story 2 BL Anchor 2: Story 1 BL Anchor 2: Story 2 WH anchor 1: Story 1 WH anchor 1: Story 2 WH Anchor 2: Story 1 WH Anchor 2: Story 2 Each of the anchor/story combinations will be paired with a neutral, racial, or no tweet. Therefore, there will be 24 total conditions. Independent & Moderator Variables Anchor Race. Participants will be randomly assigned to view two video stories delivered by either a Black or White anchor. A TV news anchor employs characteristics, credibility, parasocial relationships, and style to be the face of a television station, reading the news at an established day and time (Hill, 2007; Meltzer, 2010). Based on this, the manipulation of anchor race will be based on physical and facial attributes of the anchor. 43 Expectancy Violation. The social media microblogging website, Twitter has been around since 2006 (Greenhow & Gleason, 2012). The authors declared it a new literacy practice. A form of microblogging at its peak in 2022, Twitter had nearly 370 million worldwide users but today there are 336 million (eMarketer, 2022; Number of monetizable daily active Twitter users in the US, 2022). Some journalists are required to promote their stories on twitter using a short text-based tweet that can include links to URLs, videos, or other graphics. Journalists also use it to look for sources and connect with viewers (Hedman, 2020; Houston et al., 2020). The conceptualization of expectancy violation theory begins with an intrusion of space, visible or invisible, physical, or emotional (Burgoon & Jones, 1976). The violation can evoke either negative or positive responses (Burgoon, 2015). To manipulate expectancy violation, participants will be randomly assigned to view the two video stories following a tweet by the anchor either discussing a social justice issue in an opinionated way (expectancy violation) or presenting news-relevant information (neutral). A control condition will also be included, where participants will not see any tweet preceding the video stories. Adapted expectancy violation, expectancy violation valence were used for this research Afifi and Metts (1998). Their Cronbach reliability coefficients were .71 and .94 respectively (p 376). Story Topic. Carlson (2016) said “journalism studies explain a news story as a constructed account shaped by and through a complex array of professional, organizational, technological, political, economic, and cultural factors” (p 351). With that in mind, participants in the current study will be exposed to two stories (order counterbalanced), where one story will discuss a neutral topic and the other one will include a reference to race. Moderator Variable Humanitarianism-egalitarianism Scale Evaluation 44 The humanitarianism-egalitarianism (HE) attitude scale was developed by Katz and Hass (1988). They set out to test two contrasting scales that would measure the divisive attitudes and behaviors that arise between individuals of different races due to their uncertain or conflicting emotions, beliefs, and opinions towards one another. In Study 1 they tested and retested more than 40 anti-Black and pro-Black statements. Ten remaining items were combined with the 11- item Protestant Ethic (PE) scale, revised and retested, finding an adequate internal consistency. In Study 2, students were primed with humanitarianism statements, PE statements, or control items. They discovered two scales, one with a social justice component, friendlier towards Blacks, and one based on independence and endurance, a pulling yourself up by your bootstraps mentality that was critical of Blacks. The former is the humanitarianism-egalitarianism scale providing items that indicate compassion for humankind not for any particular race of people. Participants will use a 7-point Likert scale (strongly disagree to strongly agree). All item-test correlations for the HE Scale were above .50. Coefficient alpha for the 10 items was .84 (p. 897). The higher the score, the more extreme the attitude. Dependent Measures The current study will include two sets of variables that will be measured upon exposure to each news story. The first will entail evaluation of the anchor credibility, while the second will entail evaluating the news story. All items will be measured on a seven-point Likert-type scale anchored by “Strongly Disagree” to “Strongly Agree” (1-7 scale). Upon satisfactory validity and reliability indices, items will be averaged per news story. For regression analyses, the two-story variables will be averaged into a single variable. News Anchor Credibility 45 To assess the perceived credibility of the anchor, the study will measure five elements of perceived attractiveness (5 items), trustworthiness (5 items), and expertise (5 items). All items will be adopted from Ohanian (1990). Each subscale had a reliability coefficient of 0.8 or higher. Factor loadings ranged from .548-.748 for attractiveness, .524-.696 for trustworthiness, and .556- .702 for expertise (p 47). News Story Evaluation The study will use four items to measure news story evaluation. These are attitudinal measures assessing whether the story is interesting, engaging, important, and enjoyable (Kohring & Matthes, 2007). Kohring and Matthes (2007) indicated the factor loading should be higher than .6, and the indicator reliability should be ≥ .4. They said the goodness of fit (x2/df = 1.633) provided evidence of validity. Additional items were used from Hinnant et al. (2023) story evaluation scales. Participants were asked to evaluate the story as relevant, trusted and told the entire story. A 7-point Likert scale was used. A higher score on these items indicated a higher perceived relevance of the story. Control Variables Political ideology Participants will be asked a series of pre-stimulus questions to determine their political ideology including party affiliation by choosing one of the following options: Strong Republican, Republican, Strong Democrat, Democrat, Independent leaning Republican, Independent leaning Democrat, and Independent. To reduce the number of groups examined for this analysis, this variable will be reduced to a three-level variable categorizing participants into Republican, Democrat, or Independent. Manipulation Checks 46 To check the manipulation of anchor race, participants will indicate whether the anchor who delivered the story was Black or White at the end of the study. As for checking the manipulation of the tweet conditions, I will use Afifi and Metts (1998) scales for violation expectedness and violation valence. For violation expectedness, participants will rate four items using seven-point Likert-type scales (anchored by “Strongly Disagree” and “Strongly Agree” (1- 7 scale) to evaluate the anchor’s performance as (1) completed expected (reverse-coded), not at all expected, surprised me a great deal, surprised me only very slightly (reverse-coded). With regard to violation valence, the anchor’s behavior will be evaluated using four items, rated on seven-point Likert-type scales (anchored by “Strongly Disagree” to “Strongly Agree”) as to whether the anchor’s behavior was (1) a very positive performance, (2) a behavior I liked a lot, (3) a behavior that I did not like at all (reverse-coded), (4) I’d like to see much more of the anchor’s work. Additional survey scales Additional survey scales were added to gauge daily news consumption, media literacy, and social media literacy required of journalists. A daily news consumption questionnaire was adapted from a previous unpublished study (see Appendix C). Vraga et al. (2015)’s multidimensional scale on media literacy was added to the survey to track participants knowledge of and attitudes towards news media. In this divisive era, news consumers tend to choose media producers who align with their political leanings even if news consumed isn’t factual (Barnidge et al., 2020). This scale paired with political leanings may provide additional incite into how participants evaluated the news anchors. Finally, a social media use scale was created to measure how well participants understood how journalists use social media for sources, story ideas, and to promote their news product (Appendix D). 47 Participants were prompted to divulge if fourteen statements were not at all acceptable (1) to perfectly acceptable (7). Sample Description Using G*power software to estimate the sample size needed for this experiment, the following parameters were used ηp 2 = .04 (Mastro et al., 2012), effect size of f = .2294, α err prob = .05, power = .95. This analysis yielded a minimum of 300 participants. Due to potential missing values and outliers highly prevalent in online samples, this study will oversample by 35%, thus the final sample size recruited for the study will be 405 participants. Figure 3. G*Power Results Survey companies Prolific (pre-tests) and Qualtrics (main study) were used to engage a convenience sample of participants for this experiment. An equal number of Democrats, Independents, and Republicans were recruited. Only African Americans/Blacks and Whites were recruited. An equal number of men and women and age range (18-34, 35-45, 55+) were recruited. Stimuli/Procedures Two White and two African American/Black college students in the Michigan State University Journalism program (one of whom was a graduate) were used as anchors. The 48 students all wore their hair in a bun. They completed their makeup together before the video shoot, and it was similar. They wore white blouses with a dark colored jacket. Each anchor read the same two stories. One story was about a potential garbage strike. The other was about education test scores and included how test scores for people of color were lower than their White counterparts. Each story was a composite from Associated Press and a nationwide news story about the topic. The setting was Cincinnati, Ohio for its size and diversity. The July 2022 census data indicated the city’s White population was 48.6% and African American/Black population was 40.3% (Bureau, 2022). The stories were written by the researcher who had 25 years of broadcast news experience as an anchor/reporter. The researcher consulted with a Michigan State University professor of practice for final edits. The education story was 124 words and approximately 45 seconds in length. The video was of teachers and children/teens in school settings. The garbage strike story was 122 words and approximately 40 seconds in length. The video was of a garbage truck driver, attaching garbage cans to a truck and emptying them into a garbage truck. The video for both stories was from CBS Newspath, to which the Journalism School is a subscriber. A sixth student helped with technology and recording. It took three hours to record the four anchors reading two stories. Each anchor was recorded approximately four times in order to make sure both the researcher and anchor were comfortable with the product. Experimental Procedure The experiment was conducted online via Prolific for the pretest and Qualtrics for the main study. Both are survey companies. At the initial stage of the survey, participants answered questions about their race, gender, and political ideology to ensure representative quotas for sampling were met. Upon answering quota-relevant questions, participants answered questions 49 related to their news consumption behavior and knowledge of news media norms. Afterwards, participants were exposed to two story blocks (random order) in their respective experimental conditions: Black anchor/no tweet, Black anchor/neutral tweet, Black anchor/social justice tweet, White anchor/no tweet, White anchor/neutral tweet, and White anchor/social justice tweet. In each story block, participants were first exposed to the Twitter profile of the anchor, followed by a tweet that is either social justice oriented or neutral (participants in the no tweet condition were not exposed to the profile nor the tweet. See Appendix B). Participants then watched a video story read by an African American/Black or White anchor (depending on their random assignment for the anchor race conditions). Following each video story, participants answered questions from scales regarding violation expectedness, violation valence, anchor credibility evaluation, and story evaluations. Upon exposure to both story blocks, participants answered the anchor race manipulation check question. The final leg of the survey included demographic information: education level, sexual orientation, and income level. Data Analysis Data for an anchor’s credibility rating (attractiveness, expertise, and trustworthiness) and story evaluations were submitted 2 (race) x 3 (tweet conditions) by 2 (story topic) ANOVA with repeated measures on the last factor to examine the main effects of race and expectancy violation, and the interaction between them. With regard to examining the moderating effect of humanitarianism and egalitarianism, Hayes’ P OCESS Model 2 (Hayes, 2017) was used for the following DVs: perceived anchor attractiveness, expertise, and trustworthiness; and story evaluations. Given that the HE scale is a continuous one, I examined Johnson-Neyman’s regions of significance for potential curvilinear effects of HE on the relationship between anchor race, tweet conditions, and the DVs. The results can be found in the next chapter. 50 Chapter 7 Results Pretest I A pretest was conducted to select tweets that correspond with the two tweet conditions of neutral (news only) and social justice focus. The social justice tweet represented the opinion of the news anchor. The pretest consisted of 10 neutral (news only) tweets and 10 social justice tweets (see Appendix A). The neutral tweets focused only on promoting a news story, while the social justice or opinionated tweets included the anchor’s opinion and advocacy for social justice issues. Additionally, the social justice tweet included the #BlackLivesMatter hashtag, while the neutral ones did not include that hashtag. Participants evaluated all 20 tweets (presented in random order) using the following scales: violation expectedness, violation valence and attribution for the violation (Afifi & Metts, 1998) as well as an adaptation of a news story evaluation scale (Lang et al., 2003). Participants (N = 38) were recruited through Prolific, an online research platform. Participants’ age averaged 33.86 (SD = 10.85). Six in 10 participants (63%) identified as female, while 34.2% identified as male, and 2.6% self-identified as transgender. Half of the sample identified as African American/Black, 47.4% identified as White/Caucasian, and 2.6% did not answer the question. One third of the participants reported having completed a bachelor’s degree (34.2%), the next highest group reported having their high school diploma/GED (28.9%), followed by some college but no degree (18.4%). The median household income was $25,000 to $49,999 per annum. In terms of political party affiliation, 13.1% identified as Republican or leaning Republican, 68.4% as Democrat or leaning Democrat, and 18.4% as Independent. 51 Results of the pretest showed that the neutral and social justice tweets did not differ significantly on the following measures. To remedy this, I identified two pairs of tweets that were largest in difference on the three variables of interest: offends me; seems inappropriate for a TV anchor to tweet (reversed); and includes the anchor’s opinion. Based on that, the tweets of interest were reworked to make sure the manipulation was stronger and clearer to participants (see Appendix B). They were pretested again (see Pretest II) with another sample of 40 participants, using Prolific survey service. Pretest II Participants (N = 40) were recruited through Prolific, an online research platform. Participants age was 37.45 (SD = 13.79) on average. Men and women were evenly split within the sample at 19 each (47.5%) while 2.5% self-identified as transgender, gender variant, or non- conforming. In terms of racial background, the sample was intentionally split between African American/Black and Caucasian. Nearly one third reported having a bachelor’s degree (30%), the next highest group reporting having some college, but no degree (22.5%), followed by a high school diploma/GED (17.5%). The median household income was $25,000-49,999 per annum. For party affiliation, the largest group 62% self-identified as democrat (strong, leaning), 22.5% as republican (strong, leaning), and 12.5% as independent. 52 Table 1. Mean Differences of Experimental Stimuli used in the Main Experiment Interesting Important The kind of message I’d expect from TV anchors & reporters T1 News M(SD) 4.38 (1.78) 5.18 (1.82) T1 Opinion M(SD) 4.43 (1.89) 4.78 (2.044) 5.30 (1.522) 3.50 (2.05) Offends me 2.03 (1.42) 2.90 (1.89) Seems inappropriate for a TV anchor to tweet 2.58 (1.95) 3.98 (2.18) Is an example of good use of Twitter by a journalist 5.00 (1.59) 3.30 (2.12) Is something I’d like to see in my Twitter feed 4.03 (2.00) 3.63 (2.08) Includes the anchor’s opinion 2.17 (1.62) 6.13 (1.47) Mixes news with opinion Expectancy Violation (higher value indicates greater violation of expectancy) Expectancy Violation Valence (higher value indicates more positive valence) 2.20 (1.40) 6.03 (1.54) 3.6 (1.10) 4.33 (1.18) 4.68 (1.31) 3.88 (1.79) T2 News M(SD) 4.55 (1.85) 5.18 (1.88) T2 Opinion M(SD) 4.55 (1.85) 4.93 (2.09) 5.60 (1.22) 3.55 (1.87) 1.63 (1.15) 2.65 (2.01) 2.63 (1.96) 4.47 (1.89) 5.33 (1.35) 3.15 (2.01) 4.28 (1.85) 3.90 (2.11) 2.42 (1.88) 6.40 (.98) 2.45 (1.87) 6.23 (1.29) 3.88 (1.05) 4.26 (1.18) 4.96 (1.32) 4.05 (1.83) Paired Samples t- test t(39) = <.001, ns t(39) = 1.136, ns t(39) = 6.07, p < .001, Cohen’s d = .96 t(39) = -3.03, p < .05, Cohen’s d = -.15 t(39) = -4.25, p < .001, Cohen’s d = -1.01 t(39) = 6.20, p < .001, Cohen’s d = .98 t(39) = 1.25, ns t(39) = -10.51, p < .001, Cohen’s d = -2.14 t(39) = -9.45, p < .001, Cohen’s d = -1.49 t(39) = -1.57), ns t(39) = -3.03, p < .05, Cohen’s d = .48 Paired Samples t- test t(39) = -.21, ns t(39) = 1.79, ns t(39) = 4.26, p < .001, Cohen’s d = .67 t(39) = -2.9, p < .05, Cohen’s d = -.46 t(39) = -3.08, p < .05, Cohen’s d = -.81 t(39) = 4.77, p <.001, Cohen’s d = .75 t(39) = 1.29, ns t(39) = -9.04, p < .001, Cohen’s d = -1.43 t(39) = -10.24, p < .001, Cohen’s d = -1.62 t(39) = -3.5, p < .001, Cohen’s d = -.552 t(39) = -3.49, p = .001, Cohen’s d = .47 53 As seen in Table 1, participants were not different in evaluating the two news and two opinion tweets (See Appendix B) in terms of how interesting they perceived them, their perceived importance, and whether they would like to see more tweets like these from journalists. These three variables provide great insight about the elimination of confounding factors in the tweets. However, results showed that the neutral and social justice tweets differed significantly in terms of expectancy violation. It also shows that participants, while recognizing the opinionated tweets as opinionated, they did not mind seeing them on Twitter. Past research showed that news audience members are interested in opinion-based use of social media. Even though the second tweet pair showed no difference in the expectancy violation, the significant differences in all other measures provide enough justification for selecting the two tweet pairs for the main study. Additionally, in the main study, participants saw the tweets followed by the TV report/news, which might enhance the perceptions of expectancy violation. Main Study Demographics After data were cleaned and 160 incompletes eliminated, the final sample size was 673. Participants were recruited through Qualtrics Panels, an online survey research platform using a quota-based sampling technique in terms of gender (males vs. females), race (White vs. African American/Black), political affiliation (Democrats, Independents, and Republicans) as well as age range (18-34, 35-54, 55+). Participants’ age was 48.04 (SD = 17.98) on average. More than half of the participants identified as female (52.3%), while 46.1% identified as male, and 1.5% self - identified as non-binary. The sample was about evenly split between African American/Black (48.9%) and White (51.1%). One quarter of the participants reported having completed some college, but no degree (24.7%), the next highest group reported having their high school 54 diploma/GED (20.8%), followed by a bachelor’s degree (17.7%), graduate degree (12.5%), Associates or technical degree (11.7%), and some high school or less (1.6%). The median household income was $50,000-$74,999 per annum. In terms of political party affiliation, 18.9% identified as Republican, 54.5% as Democrat, and 26.6% as Independent. Main Study Results Factor and Reliability Analyses As seen in Table 2, all measures were submitted to factor and reliability analyses. The humanitarianism egalitarianism scale showed acceptable internal consistency with factor loadings ranging between .474 and .820, which explained 54.20% of the variance. The measure also showed acceptable reliability (Cronbach’s α = .90). For all DV measures, the factor and reliability analyses were run for each of the two messages separately (education scores vs. garbage strike stories). The latent factors, as shown in Table XX, explained more than two-thirds of the variance. Table 2. Items, Factor, and Reliability Analysis for all Study Measures Variable Humanitarianism Egalitarianism (HE) 5.42 Eigenvalue % Variance Exp. 54.196 Factor Loadings Cronbach’s α M (SD) Expectancy Violation The anchor’s performance was… Was (not) completely expected (at all) Surprised me slightly (a great deal) Expectancy Violation Valence .474 - .820 .90 54.07 (11.38) Edu Scores Story Garbage Strike Story 1.33 Eigenvalue % Variance Exp. 66.28 Factor Loadings Cronbach’s α M (SD) .814 .49 7.82 (3.02) 1.41 70.36 .839 .58 7.83 (3.15) Eigenvalue % Variance Exp. 75.86 3.03 3.05 76.15 55 Table 2 (cont’d) The anchor’s performance was… made me feel worse (better) about how she relates to viewers made me feel she does (not) really care about viewers did not like (liked) her performance would rather (never again) see more stories from this anchor News Anchor Attractiveness The anchor was… attractive classy beautiful elegant sexy News Anchor Trustworthiness The anchor was… dependable honest reliable sincere trustworthy News Anchor Expertise The anchor was… experienced knowledgeable qualified competent unbiased News Story Evaluation The news story was… informative comprehensible interesting engaging believable important enjoyable fair Factor Loadings Cronbach’s α M (SD) .860 - .874 .89 20.08 (5.96) .852 - .894 .90 20.49 (5.971) 3.69 73.73 .777 - .897 .91 22.93 (7.46) 4.16 83.11 .889 - .929 .95 25.81 (6.77) 3.87 77.37 .797 - .911 .92 26.90 (6.62) 7.35 66.823 .745 - .867 .95 57.75 (13.748) 3.63 Eigenvalue % Variance Exp. 72.50 Factor Loadings Cronbach’s α M (SD) .776 - .879 .90 21.67 (7.49) 4.14 Eigenvalue % Variance Exp. 82.79 Factor Loadings Cronbach’s α M (SD) .897 - .918 .95 25.74 (6.67) 3.87 Eigenvalue % Variance Exp. 77.34 Factor Loadings Cronbach’s α M (SD) .785 - .912 .92 26.52 (44.07) 7.52 Eigenvalue % Variance Exp. 68.37 Factor Loadings Cronbach’s α M (SD) .750 - .877 .95 57.23 (13.93) 56 Table 2 (cont’d) biased accurate truthful tells the entire story Manipulation Checks To check the manipulation of the tweet factor, data for expectancy violation (two item) were submitted to a 3 (tweet condition) x 2 (message repetition) ANOVA with repeated measures on the last factor. The main effect of expectancy violation manipulation on the expectancy violation measure was not significant, F(2, 667) = .62, ns. Additionally, the main effect of message repetition was also not significant, F(1, 667) = .01, ns, nor was the interaction between EVT manipulation and repetition, F(2, 667) = 1.94, ns. To further examine the effect of the manipulation on perceived expectancy violation, data for expectancy violation valence was submitted to a 3 (EVT) x 2 (repetition) ANOVA with repeated measures on the last factor. The main effect of EVT manipulation approached significance, F(2, 670) = 2.77, p = .06, η2 p = .01. Participants in the social justice tweet condition expressed more negative expectancy violation (M = 4.93, SE = .09), followed by those in the no tweet (M = 5.17, SE = .09) and neutral news tweet (M = 5.20, SE = .09), respectively. Planned post-hoc pairwise comparisons showed the no tweet and neutral news tweet conditions as well as the social justice tweet and no tweet did not differ significantly (p>.05), yet the difference between the social justice tweet and neutral news tweet approached significance (p = .09). The main effect of message repetition approached significance, F(1, 670) = 3.83, p = .051, η2 p = .01 and the interaction between EVT and message repetition also approached significance, F(2, 670) = 2.87, p = .057, η2 p = .01. Based on this, the EVT manipulation did not successfully alter participants’ perceived expectancy violation based on the type of tweet they saw. 57 Given that the differences in expectancy violation among the three tweet conditions were not significant, while the difference between the social justice and neutral tweet conditions approached significance, I conducted additional analyses to check whether the perceived expectancy violation and expectancy violation valence were different as a function of participants’ age and race. Data for perceived expectancy violation (two items) were submitted to a 3 (tweet condition: no tweet vs. neutral tweet vs. social justice tweet) x 2 (participant race: White vs. Black) x 2 (participant age: younger vs. older) x 2 (message repetition) repeated measures ANOVA with repeated measures on the last factor. The effect of the three-way interaction among participants’ age, participants’ race, and the tweet condition on perceived expectancy violation was significant F(2, 583) = 3.06, p < .05, η2 p = .01. As shown in Figure 2, younger White participants expressed greater violation of expectancy than older ones, and those assigned to the neutral tweet condition expressed the greatest expectancy violation. Younger Black participants expressed higher expectancy violation than older Blacks, but this difference was only apparent in the no tweet and social justice tweet conditions. Additionally, there was a main effect of participants’ age on expectancy violation, F(1, 583) = 17.58, p < .001, η2 p = .03. Younger participants, in general, perceived the anchor’s performance to be more violating of their expectancy (M = 4.11, SD = 1.59) than older participants (M = 3.60, SD = 1.48). The main effect of participants’ race on expectancy violation was also significant, F(1, 583) = 6.71, p < .05, η2 p = .01, Black participants perceived the anchor’s performance to be more violating of their expectancy (M = 4.01, SD = 1.63) compared to White participants (M = 3.71, SD = 1.47). 58 White Participants No Tweet Neutral Tweet Social Justice Tweet Younger Older Black Participants 5 4.5 4 3.5 3 2.5 2 1.5 1 0.5 0 5 4.5 4 3.5 3 2.5 2 1.5 1 0.5 0 No Tweet Neutral Tweet Social Justice Tweet Younger Older Figure 4. Mean differences in Expectancy Violation, by tweet condition, participant race, and participant age To check the manipulation of anchor race, participants were asked to indicate the race of the anchor delivering the tweet and/or news story to them during the experiment, where they could choose from the following racial categories: White, Black/African American, American Indian or Alaska Native, Asian, Native Hawaiian or Pacific Islander, Other, and “Can’t 59 remember.” Data for the race manipulation check were submitted to a Chi-Square analysis, χ2(4), 202.01, p < .001, Cramer’s V = .58. Of all participants, only one individual indicated that the anchor was other than White or African American/Black, two participants indicated “Other”, and seven participants said they could not remember the race of the anchor. Among the rest of the sample, 70.27% of participants assigned to view a White anchor correctly remembered the anchor’s race as White, while 27.03% identified the White anchor as African American/Black, 0.68% selected “Other” and 2.03% indicated they could not remember. On the other hand, among participants assigned to view stories by African American/Black anchors, 84.21% of them correctly identified the anchor’s race as African American/Black, 15.13% thought the anchor was White, 0.33% thought the anchor was American Indian or Alaska Native, 0.33% said they could not remember. Based on this, the race manipulation was successful. Additionally, data for expectancy violation valence were submitted to a 3 (tweet condition) x 2 (participant race) x 2 (participant age) x 2 (message repetition) repeated measures ANOVA with repeated measures on the last factor. The main effect of tweet condition was not significant, F(2, 583) = 2.49, p = .08, η2 p = .01. Planned post-hoc pairwise comparisons showed the difference between social justice and neutral tweet conditions approached significance (p = .09). Taken together, the tweet condition manipulation check findings showed the measure of perceived expectancy violation, i.e., whether participants perceived the anchor’s performance violated the expectancy of how a journalist should act was not sensitive to the tweet condition, yet these differences were affected by participants’ age and race. On the other hand, findings for the valence of expectancy violation, i.e., the perceived pleasantness/unpleasantness of the expectancy violation of the anchor’s performance was different when comparing the neutral and 60 social justice tweet conditions. The effect of tweet condition was not sensitive to participants’ age and race. Based on this, there is not enough evidence to show a straight-forward effect of the manipulation, despite its interaction with age and race of participants. Therefore, the reporting of the results will focus on comparing the three tweet conditions in isolation of potential violation of expectancy. H1: Participants will express more favorable evaluations of the anchor in terms of (a) attractiveness, (b) trustworthiness, and (c) expertise, as well as (d) the news stories they read when the anchor is White than African American. Main Effect of Anchor Race Hypothesis 1 posited that participants will express more favorable evaluation of the anchor in terms of attractiveness, trustworthiness, and expertise (H1a-c) and story (H1d) when the anchor is White than when she is African American. To test this hypothesis, data for perceived anchor attractiveness, trustworthiness, and expertise, as well as news story evaluation were submitted to identical parallel 2 (anchor race) x 2 (message repetition) ANCOVA with repeated measures on the last factor. In these analyses, participant gender, party affiliation, race, and age were added as covariates. Here, I will report each set of analyses separately. For each DV, two sets of analyses are reported. First, I report the analyses for the entire sample of participants. The second set of analyses only include participants in the control EVT manipulation condition (no tweet) to guard against any potential effect of the EVT manipulation on anchor evaluation. H1a: Perceived Attractiveness. For perceived anchor attractiveness, the main effect of anchor race was significant, F(1, 591) = 38.59, p < .001, η2 p = .06. Even though this effect was significant, it was in the opposite direction from my hypothesis. Participants perceived the Black 61 anchor (M = 4.83, SE = .08) to be more attractive than the White anchor (M = 4.16, SE = .08); H1a was not supported. Neither the main effect of message repetition nor its interaction with the anchor race as well as the control variables were significant (ps > .05). Of the four covariates, the effects of participant race, F(1, 591) = 6.17, p = .01, η2 p = .01, and age, F(1, 591) = 32.24, p < .001, η2 p = .05, were significant. Bivariate correlation analyses showed that younger participants perceived the anchors to be more attractive than older participants, r(673) = -.245, p < .001. For participant race, Black participants perceived the anchors to be more attractive than White participants, r(673) = -.186, p < .001. Regarding the sub-sample of the no tweet condition, a similar pattern of results emerged. The main effect of anchor race was significant, F(1, 201) = 8.24, p = .005, η2 p = .04. Participants assigned to view the story by a Black anchor (M = 4.82, SE = .13) evaluated the anchor to be more attractive than those assigned to the White anchor condition (M = 4.27, SE = .14). The effects of message repetition and its interaction with the covariates were not significant. Of the four covariates, the effects of participant race, F(1, 201) = 4.49, p = .035, η2 p = .02, and age, F(1, 201) = 6.55, p = .011, η2 p = .03, were significant. H1b: Perceived Trustworthiness. The main effect of anchor race on perceived trustworthiness was significant, F(1, 591) = 4.09, p < .04, η2 p = .01. This was similar to the perceived anchor attractiveness whereas this was significant, it was in the opposite direction from my hypothesis. Participants also perceived the Black/African American anchor (M = 5.30, SE = .07) to be more trustworthy than the White anchor (M = 5.01, SE = .07); H1b was not supported. Neither the main effect of message repetition nor its interaction with the anchor race 62 as well as the control variables were significant (ps > .05). Of the four covariates, the effect of participant race approached significance, F(1, 591) = 3.41, p = .07, η2 p = .01. With regard to the sub-sample of the no tweet condition, the main effect of anchor race was not significant, F(1, 201) = .17, ns. Participants assigned to view the stories by either the Black/African American or White anchors found them to be nearly similar in attraction (M = 5.32, SE = .12), (M = 5.25, SE = .12) respectively. The effects of message repetition and its interaction with the covariates were not significant. Of the four covariates, the effects of participant race was significant, F(1, 201) = 4.34, p = .04, η2 p = .02. H1c: Perceived Expertise. The main effect of anchor race on perceived expertise was significant, F(1, 591) = 4.18, p < .04, η2 p = .01. This was similar to the perceived anchor attractiveness and trustworthiness whereas this was significant, it was in the opposite direction from my hypothesis. Participants also perceived the Black/African American anchor (M = 5.49, SE = .07) to be more of an expert than the White anchor (M = 5.29, SE = .07); H1c was not supported. Neither the main effect of message repetition nor its interaction with the anchor race as well as the control variables were significant (ps > .05). With regard to the sub-sample of the no tweet condition for anchor expertise, the main effect of anchor race was not significant, F(1, 201) = .03, ns. Participants assigned to view the stories by either the Black/African American or White anchors found them to be nearly similar in expertise (M = 5.54, SE = .12), (M = 5.51, SE = .12) respectively. The effects of message repetition and its interaction with the covariates were not significant. H1d: News Story Evaluation. The main effect of anchor race on news story evaluation was significant, F(1, 591) = 4.64, p < .03, η2 p = .01. This too was in the opposite direction of my hypothesis. Participants enjoyed the Black anchor (M = 5.36, SE = .07) more than the White 63 anchor (M = 5.16, SE = .07); H1d was not supported. Neither the main effect of message repetition nor its interaction with the anchor race as well as the control variables were significant (ps > .05), however participant race approached significance, F(1, 591) = 3.30, p = .07, η2 p = .01. With regard to the sub-sample of the no tweet condition for story evaluation, the main effect of anchor race was not significant, F(1, 201) = .03, ns. Participants assigned to view the stories by either the Black/African American or White anchors found them to be nearly similar in attraction (M = 5.54, SE = .12), (M = 5.51, SE = .12) respectively. The effects of message repetition and its interaction with the covariates were not significant. H2: HE will moderate the effect of anchor race on perceived anchor (a) attractiveness, (b) trustworthiness, (c) expertise, and (b) story evaluation, such that those high on the HE will show less negativity toward Black anchors and the stories they deliver than those low on the HE scale, while HE will not lead to differences in evaluating the White anchors and their stories. Humanitarianism Egalitarianism Moderating the Effect of Anchor Race on Anchor and Story Evaluations In Hypothesis 2, I posited that humanitarianism-egalitarianism would moderate the effect of anchor race on anchor evaluations in terms of attractiveness, trustworthiness, and expertise (H2a-c) and news story evaluation (H2d). Data for each of the four DVs were each submitted to moderation effect using PROCESS macros for SPSS (Model 1) with 10,000 bootstrap samples with anchor race as an IV, HE as a moderator, perceived attractiveness, trustworthiness, expertise, and news story evaluations as DVs (ran models separately for each DV), and participant age, race, gender, and political party affiliation as covariates. The following reports the findings for each of the DVs. 64 Table 3. Moderating Effect of Humanitarianism/Egalitarianism (HE) on the Effect of Anchor Race on Perceived Anchor Attractiveness (constant) Anchor Race HE Race x HE Gender Age Participant Race Party Affiliation Coeff. 5.27 .75 .46 -.02 -.11 -.02 -.34 -.02 SE .78 .52 .07 .09 .05 .003 .11 .02 t 6.79 1.44 6.80 -.17 -2.06 -7.53 -3.20 -1.12 p < .001 .15 < .001 .87 .04 < .001 .001 .26 CILL, UL 3.74, 6.79 -.27, 1.77 .33, .59 -.20, .17 -.21, -.005 -.03, -.02 -.55, -.13 -.06, .02 R = .50, R2 = .25, MSE = 1.49 F(7, 589) = 28.23, p < .001 H2a: Perceived Attractiveness. The moderation model explained 25% of the variance in perceived anchor attractiveness, R = .50, R2 = .25, MSE = 1.49, F(7, 589) = 28.23, p < .001. The main effect of anchor race on attractiveness was not significant, Coefficient = .75, SE = .52, t = 1.44, p = .15, CILL,UL = -.27, 1.77. The effect of HE on attractiveness was significant, Coefficient = .46, SE = .07, t = 6.80, p < .001, CILL,UL = .33, .59; the higher participants’ HE was, the more attractive they perceived the anchor, regardless of her race. The effect of the interaction between race and HE was not significant, Coefficient = -.02, SE = .09, t = -.17, p = .87, CILL,UL = -.20, .17, thus H2a was not supported. As shown in Table 3, participants’ gender, age, and race were negative predictors of perceived attractiveness. Table 4. Moderating Effect of Humanitarianism/Egalitarianism (HE) on the Effect of Anchor Race on Perceived Anchor Trust (constant) Anchor Race HE Race x HE Gender Coeff. 4.24 .54 .56 -.06 -.01 t 6.10 1.17 9.31 -.76 -.25 p < .001 .24 <.001 .45 .81 CILL, UL 2.87, 5.60 -.37, 1.45 .44, .68 -.23, .10 -.10, -.08 SE .69 .46 .06 .08 .05 65 Table 4 (cont’d) Age Participant Race Party Affiliation -.004 -.27 -.05 .003 .10 .02 -1.34 -2.78 -2.71 .18 <.001 <.05 -.01, -.002 -.46, -.08 -.08, .01 R = .47, R2 = .22, MSE = 1.19 F(7, 589) = 24.35, p < .001 H2b: Perceived Trustworthiness. The moderation model explained 22% of the variance in perceived anchor trustworthiness, R = .47, R2 = .22, MSE = 1.19, F(7, 589) = 24.35, p < .001. The main effect of anchor race on trust was not significant, Coefficient = .54, SE = .46, t = 1.17, p = .24, CILL,UL = -.37, 1.45. The effect of HE on trust was significant, Coefficient = .56, SE = .06, t = 9.31, p < .001, CILL,UL = .44, .68; the higher participants’ HE was, the more they trusted the anchor, regardless of her race. The effect of the interaction between race and HE was not significant, Coefficient = -.06, SE = .08, t = -.76, p = .45, CILL,UL = -.23, .10, thus H2b was not supported. As shown in Table 4, participants’ race and party affiliation were negative predictors of perceived trustworthiness. Table 5. Moderating Effect of Humanitarianism/Egalitarianism (HE) on the Effect of Anchor Race on Perceived Anchor Expertise (constant) Anchor Race HE Race x HE Gender Age Participant Race Party Affiliation Coeff. 4.02 .49 .59 -.05 -.001 -.003 -.24 -.05 SE .66 .44 .06 .08 .04 .003 .09 .02 t 6.17 1.12 10.44 -.69 -.004 -1.07 -2.65 -3.21 p < .001 .27 <.001 .49 1.00 .28 <.05 <.001 CILL, UL 2.76, 5.33 -.37, 1.35 .48, .71 -.21, .10 -.09, -.09 -.01, -.002 -.42, -.06 -.08, .02 R = .52, R2 = .27, MSE = 1.06 F(7, 589) = 31.00, p < .001 H2c: Perceived Expertise. The moderation model explained 27% of the variance in perceived anchor expertise, R = .52, R2 = .27, MSE = 1.06, F(7, 589) = 31.00, p < .001. The main 66 effect of anchor race on expertise was not significant, Coefficient = .49, SE = .44, t = 1.12, p = .27, CILL,UL = -.37, 1.35. The effect of HE on expertise was significant, Coefficient = .59, SE = .06, t = 10.44, p < .001, CILL,UL = .48, .71; the higher participants’ HE was, the more they thought the anchor was an expert, regardless of her race. The effect of the interaction between race and HE was not significant, Coefficient = -.05, SE = .08, t = -.69, p = .49, CILL,UL = -.21, .10, thus H2c was not supported. As shown in Table 5, participants’ race and party affiliation were negative predictors of perceived trustworthiness. Table 6. Moderating Effect of Humanitarianism/Egalitarianism (HE) on the Effect of Anchor Race on Perceived News Story Evaluation Coeff. 4.28 .07 .56 -.56 -.03 -.01 -.26 -.02 (constant) Anchor Race HE Race x HE Gender Age Participant Race Party Affiliation CILL, UL 3.07, 5.48 -.73, .88 .46, .67 -.21, .17 -.11, -.05 -.01, -.002 -.42, -.09 -.05, .01 p < .001 .86 <.001 .77 .46 < .05 <.05 .11 t 6.96 .18 10.51 -.29 -.74 -2.88 -3.01 -1.62 SE .61 .41 .05 .07 .04 .002 .09 .02 R = .55, R2 = .30, MSE = .93 F(7, 589) = 35.88, p < .001 H2d: News Story Evaluation. The moderation model explained 30% of the variance in news story evaluation, R = .55, R2 = .30, MSE = .93, F(7, 589) = 35.88, p < .001. The main effect of anchor race on news story evaluation was not significant, Coefficient = .07, SE = .41, t = .18, p = .86, CILL,UL = -.73, .88. The effect of HE on news story evaluation was significant, Coefficient = .56, SE = .05, t = 10.51, p < .001, CILL,UL = .46, .67; the higher participants’ HE was, the higher they evaluated the news story. The effect of the interaction between race and HE was not significant, Coefficient = -.65, SE = .07, t = -.29, p = .77, CILL,UL = -.21, .17, thus H2d 67 was not supported. As shown in Table 6, participant gender and race were negative predictors of news story evaluation. H3: Participants will express more favorable evaluation of the anchor in terms of perceived (a) attractiveness, (b) trustworthiness, and (c) expertise, as well as more favorable (d) news story evaluations upon exposure to the news story with no tweet (control), followed by neutral tweet (low expectancy violation) and social justice tweet (high expectancy violation), respectively. Main Effect of Tweet Condition Hypothesis 3 posited the tweet condition would affect both anchor evaluations (H3a-c) and news story evaluation (H3d), where more favorable evaluations of the anchors and stories would be expressed by participants assigned to the news story with no tweet (control), followed by the neutral or news only tweet and the social justice tweet, respectively. To test this hypothesis, data for the perceived anchor attractiveness, trustworthiness, expertise, and news story evaluation were submitted to identical parallel 3 (tweet conditions) x 2 (message repetition) ANOVA with repeated measures on the last factor. In these analyses, participant gender, party affiliation, race, and age were added as covariates. Here, I will report each set of analyses separately. H3a: Perceived Attractiveness. The main effect of tweet condition on perceived attractiveness was not significant, F(2, 590) = .89, ns. Even though participants rated the anchor as least attractive when she tweeted a social justice message (M = 4.39, SD = 1.51) compared to when the tweet only included news (neutral tweet) (M = 4.55, SD = 1.42) and the control (no tweet) condition (M = 4.56, SD = 1.54), respectively, yet these differences were not significant, 68 thus H3a was not supported. Of the four covariates, the effects of participant race, F(1, 590) = 6.23, p = .01, η2 p = .01, and age, F(1, 590) = 28.75, p < .001, η2 p = .05, were significant. Bivariate correlation analyses showed that younger participants perceived the anchors to be more attractive than older participants (r (673) = -.245, p < .001). For participant race, Black participants perceived the anchor to be more attractive than White participants, (r (673) = -.186, p < .001). s s e n e v i t c a r t t A d e v i e c r e P n a e M 5.5 5 4.5 4 3.5 3 Figure 5. Perceived attractiveness White Participants Black Particpants White Anchor condition Black Anchor condition Effect of the Interaction between Anchor Race and Participant Race on Perceived Attractiveness (ns) I conducted additional analyses to compare Black and White participants. Data for perceived anchor attractiveness (composite score for the two messages) were submitted to a univariate ANCOVA with anchor race, tweet condition, and participant race as IVs and gender, age, and party affiliation as covariates. Black participants (M = 4.75, SD = 1.41) perceived the African American/Black anchors to be more attractive than White participants (M = 4.26, SD = 1.36), F(1, 582) = 6.06, p = .014, η2 p = .01. Additionally, the interaction between anchor race and 69 participant race approached significant, F(1, 582) = 3.19, p = .074, η2 p = .01. White participants assigned to the White anchor condition perceived the anchor to be less attractive (M = 4.02, SD = 1.39) than Black participants (M = 4.32, SD = 1.47). On the other hand (See Figure 3), White participants (M = 4.50, SD = 1.29) perceived the Black anchor to be less attractive than Black participants (M = 5.16, SD = 1.21). H3b: Perceived Trustworthiness. The main effect of tweet condition was not significant, F(2, 590) = 1.84, ns. Participants assigned to social justice tweet conditions evaluated the anchor as least trustworthy (M = 5.05, SD = 1.38), compared to those assigned to the neutral tweet(M = 5.25, SD = 1.30) and no tweet(M = 5.28, SD = 1.30) conditions. However, these differences were not significant, thus H3b was not supported. Of the four covariates, the effects of participant race approached significance, F(2, 590) = 3.39, p = .07, η2 p = .01. Univariate ANCOVA results for the effect of anchor race, tweet condition, and participant race on perceived trustworthiness showed that White participants (M = 5.12 , SD = 1.24) perceived the anchors to be less trustworthy than Black participants (M = 5.28, SD = 1.22), F(1, 582) = 3.27, p = .07, η2 p = .01. The interaction between anchor race and participant race was significant F(1, 582) = 16.81, p < .001, η2 p = .03. White participants (M = 5.22, SD = 1.17) perceived the White anchor to be more trustworthy than Black participants (M = 4.96, SD = 1.19). On the other hand, White participants (M = 5.03, SD = 1.30) perceived the Black anchor to be less trustworthy than Black participants (M = 5.59, SD = 1.17). 70 i s s e n h t r o w t s u r T d e v i e c r e P n a e M 5.7 5.6 5.5 5.4 5.3 5.2 5.1 5 4.9 4.8 4.7 4.6 Perceived Trustworthiness White Participants 1 Black Participants 2 White Anchor condition Black Anchor condition Figure 6. Effect of the Interaction between Anchor Race and Participant Race on Perceived Trustworthiness The interaction between tweet condition and participant race approached but was not significant, F(2, 582) = 2.50, p = .08, η2 p = .01. Among participants assigned to the no tweet condition, White participants (M = 5.11, SD = 1.31) perceived the anchors to be less trustworthy than Black participants (M = 5.45, SD = 1.05). Among participants to the neutral tweet condition (news only), White participants (M = 5.34, SD = 1.11), perceived the White anchors to be more trustworthy than Black participants (M = 5.16, SD = 1.35). Finally, among participants assigned to the social justice tweet condition, White participants (M = 4.91, SD = 1.26) perceived the White anchors to be less trustworthy than Black participants (M = 5.21, SD = 1.25). H3c: Perceived Expertise. For perceived expertise of the anchor, the main effect of tweet condition was significant, F(2, 590) = 5.49, p < .004, η2 p = .02. Participants assigned to the social justice tweet condition evaluated the anchor to less of an expert (M = 5.15, SD = 1.36), compared to those assigned to the neutral tweet (M = 5.47, SD = 1.25) and no tweet conditions 71 (M = 5.53, SD = 1.27) (see Figure 5). Planned post hoc pairwise comparisons showed that the no tweet and news tweet condition were not significantly different from one another, however, the no tweet and social justice tweet conditions differed significantly (p = .007), and the news tweet and the social justice tweet conditions also were significantly different (p = .02). H3c was supported. None of the covariates were significant (ps > .05). Univariate ANCOVA analyses for the effect of anchor race, tweet condition, and participant race on perceived expertise, showed that the interaction between anchor race and participant race was significant F(1, 582) = 9.81, p = .002, η2 p = .02. White participants (M = 5.38, SD = 1.15) perceived the White anchor to be more of an expert than Black participants (M = 5.18, SD = 1.20). On the other hand, White participants (M = 5.29, SD = 1.24) perceived the Black anchor to be less of an expert than Black participants (M = 5.70, SD = 1.15). Perceived Expertise e s i t r e p x E d e v i e c r e P n a e M 5.8 5.7 5.6 5.5 5.4 5.3 5.2 5.1 5 4.9 Figure 7. White Participants Black Participants White Anchor condition Black Anchor condition Effect of the Interaction between Anchor Race and Participant Race on Perceived Expertise 72 H3d: News Story Evaluation. For news story evaluation, the main effect of tweet condition was significant, F(2, 590) = 4.26, p < .02, η2 p = .01. Participants assigned to the social justice tweet condition (M = 5.06, SD = 1.28) less favorably than those assigned to the neutral tweet (M = 5.38, SD = 1.21) and no tweet (M = 5.33, SD = 1.20) conditions. For news story evaluation, Planned Post hoc Pairwise comparisons showed that the no tweet and news tweet conditions were not significantly different from one another, however, the difference between the no tweet and social justice tweet conditions approached significance (p = .057), and the news tweet and the social justice tweet conditions were significantly different (p = .019). H3d was supported. None of the covariates were significant, however participant race approached significance, F(2, 590) = 2.26, p = .07, η2 p = .01. Bivariate correlation analyses showed that Black participants evaluated the stories more favorably than White participants (r (673) = -.084, p = .029). Univariate ANCOVA analyses for the effect of anchor race, tweet condition, and participant race on news story evaluation, showed that the main effect of participant race approached significance F(1, 582) = 3.18, p = .075, η2 p = .01. White participants (M = 5.16, SD = 1.14) evaluated the news stories less favorably than Black participants (M = 5.36, SD = 1.15). The interaction between anchor race and participant race was significant F(1, 582) = 11.20, p < .001, η2 p = .02. White participants (M = 5.21, SD = 1.10) assigned to view stories read by the White anchor evaluated the stories more favorably than their Black participants (M = 5.11, SD = 1.14). On the other hand, White participants (M = 5.12, SD = 1.18) assigned to view stories read by Black anchors evaluated the stories less favorably than Black participants (M = 5.62, SD = 1.10). 73 News Story Evaluation White Participants Black Participants White Anchor condition Black Anchor condition n o i t a u l a v E y r o t S s w e N n a e M 5.7 5.6 5.5 5.4 5.3 5.2 5.1 5 4.9 4.8 Figure 8. Effect of the Interaction between Anchor Race and Participant Race on News Story Evaluation H4: There will be a significant interaction between anchor race and expectancy violation on (a) perceived attractiveness, (b) perceived trustworthiness, (c) perceived expertise, and (d) story evaluations, such that participants will express more favorable evaluations toward anchor and stories by a White anchor than an African American anchor when she tweets about a social justice issue (expectancy violation condition) compared to neutral and no tweet, while they will express less favorable anchor and story evaluations upon exposure to stories by an African American anchor following a social justice tweet, compared to neutral and no tweets conditions. Interaction Between Anchor Race and Tweet Condition Hypothesis 4 posited that the effect of the interaction between anchor race and tweet condition on anchor and story evaluations would be significant. To test this hypothesis, data for perceived anchor attractiveness, trustworthiness, expertise, and news story evaluation were 74 submitted to identical parallel 2 (anchor race) x 3 (tweet condition) x 2 (message) repeated measures ANOVA on the last factor. In these analyses, participant gender, party affiliation, race, and age were added as covariates. Here, I will report each set of analyses separately. H4a: Perceived Attractiveness. For perceived anchor attractiveness, the effect of the interaction between anchor race and tweet condition was not significant, F(2, 587) = .66, ns; H4a was not supported. While the interaction effect between anchor race and expectancy violation wasn’t significant, the anchor race alone was significant, as noted under tests of H1. F(2, 587) = 38.98, p < .001, η2 p = .06 as were the covariates participant race F(2, 587) = 6.11, p < .01, η2 p = .01. and age F(2, 587) = 32.59, p < .001, η2 p = .05. H4b: Perceived Trustworthiness. For perceived anchor trust, the effect of the interaction between anchor race and expectancy violation was not significant, F(2, 587) = .65, ns; H4b was not supported. While the interaction effect between anchor race and expectancy violation wasn’t significant, the anchor race alone was significant, F(2, 587) = 4.26, p < .05, η2 p = .01 as noted under tests of H1. The covariate participant race approached significance, F(2, 587) = 3.36, p = .07, η2 p = .01. directionality of effect? H4c: Perceived Expertise. For perceived expertise of the anchor, the effect of the interaction between anchor race and expectancy violation was not significant, F(2, 587) = .96, ns; H4c was not supported. While the interaction effect between anchor race and expectancy violation wasn’t significant, the anchor race F(2, 587) = 4.44, p < .05, η2 p = .01 and expectancy violation F(2, 587) = 5.60, p < .05, η2 p = .02 separately were significant. H4d: News Story Evaluation. For news story evaluation, the interaction between anchor race and expectancy violation was not significant, F(2, 587) = .67, ns; H4d was not supported. While the interaction effect between anchor race and expectancy violation wasn’t significant, the 75 anchor race F(2, 587) = 4.93, p < .05, η2 p = .01 and expectancy violation F(2, 587) = 4.35, p < .05, η2 p = .02 were significant as main effects, as seen in prior analyses. Participant race approached significance, F(2, 587) = 3.18, p = .08, η2 p = .01. RQ1: How does HE moderate the effect of the interaction between anchor race and tweet condition on (a) perceived attractiveness, (b) perceived trustworthiness, (c) perceived expertise, and (d) story evaluation. Moderating Effect of HE on the Interaction Between Anchor Race and Tweet Condition The study’s research question asked about how HE would moderate the effects of anchor race and tweet condition on anchor evaluations (RQ1a-c) and news story evaluation (RQ1d). To test this moderation effect, data for each DV were submitted to a two-moderator model using PROCESS macros for SPSS (Model 2) with 10,000 bootstrap samples, where anchor race was an IV, tweet condition was a first moderator, HE was a second moderator, and participants’ age, gender, race, and political affiliation as covariates. Given that tweet condition was a three- category variable, this variable was recoded into two main variables, where the first one compared the neutral tweet condition to the two other conditions, and the second one anchored the comparison to social justice tweet condition against the two the other conditions. RQ1a: Perceived Attractiveness. Anchor race, tweet condition, and HE (including covariates) explained 25% of the variance in perceived anchor attractiveness. Neither the main effects of race and tweet condition nor the interaction between them were significant (see Table X). HE was positively associated with perceived attractiveness, Coefficient = .45, SE = .07, t = 6.69, p < .001, CILL,UL = .32, .59. Additionally, gender (males > females), age (younger > older), and race (Blacks > Whites) were significant covariates. 76 Table 7. Moderating Effect of Humanitarianism/Egalitarianism (HE) on the Effect of Anchor Race and Tweet Condition on Perceived Anchor Attractiveness (constant) Anchor Race (AR) Neutral Tweet Condition (NTC) Social Justice Tweet Condition (SJTC) AR x NTC AR x SJTC HE Race x HE Gender Age Participant Race Party Affiliation Coeff. 2.98 .65 .002 -.23 -.04 .21 .45 -.008 -.11 -.02 .34 -.02 SE .44 .54 .17 .18 .24 .25 .07 .09 .05 .003 .11 .02 t 6.78 1.20 .009 -1.31 -.15 .87 6.69 -.09 -2.01 -7.54 3.15 -1.16 p < .001 .23 .99 .19 .88 .39 < .001 .93 < .05 < .001 < .05 .25 CILL, UL 2.12, 3.84 -.41, 1.71 -.34, .34 -.58, .12 -.51, .44 -.27, .70 .32, .59 -.19, .17 -.21, -.002 -.03, -.02 .13, .55 -.06, .02 R = .50, R2 = .25, MSE = 1.50 F(11, 585) = 18.12, p < .001 RQ1b: Perceived Trustworthiness. Anchor race, tweet condition, and HE (including covariates) explained 23% of the variance in perceived anchor trust. Neither the main effects of race and tweet condition nor the interaction between were significant. HE was positively associated with perceived trust, Coefficient = .56, SE = .06, t = 9.16, p < .001, CILL,UL = .44, .67. Additionally participant race (Blacks > Whites) and party affiliation (Republicans > Democrats) were significant covariates. Table 8. Moderating Effect of Humanitarianism/Egalitarianism (HE) on the Effect of Anchor Race and Tweet Condition on Perceived Anchor Trust (constant) Anchor Race (AR) Neutral Tweet Condition (NTC) Coeff. 2.48 .46 .004 SE .39 .48 .15 t 6.34 .95 .02 p < .001 .34 .98 CILL, UL 1.72, 3.26 -.49, 1.41 -.30, .31 77 Table 8 (cont’d) Social Justice Tweet Condition (SJTC) AR x NTC AR x SJTC HE Race x HE Gender Age Participant Race Party Affiliation -.25 -.05 .18 .56 -.06 -.01 -.003 .26 -.05 .16 .22 .22 .06 .08 .05 .003 .10 .02 -1.60 -.21 .80 9.16 -.66 -.19 -1.38 2.73 -2.72 .11 .84 .42 < .001 .51 .85 .17 < .05 < .05 -.56, .06 -.47, .38 -.26, .61 .44, .67 -.22, .11 -.10, .08 -.01, .002 .07, .45 -.08, -.01 R = .48, R2 = .23, MSE = 1.19 F(11, 585) = 15,82, p < .001 RQ1c: Perceived Expertise. Anchor race, tweet condition, and HE (including covariates) explained 28% of the variance in perceived anchor expertise. The main effect of anchor race was not significant. While the neutral tweet condition was not significant, the social justice tweet was significant, Coefficient = -.42, SE = .15, t = -2.82, p < .05, CILL,UL = -.71, -.13. Participants assigned to the social justice tweet condition expressed higher perceived expertise toward the anchors compared to those assigned to the two other conditions. HE was positively associated with perceived expertise, Coefficient = .58, SE = .06, t = 10.26, p < .001, CILL,UL = .47, .69. Additionally participant race (Blacks > Whites) and party affiliation (Republicans > Democrats) were significant covariates. Table 9. Moderating Effect of Humanitarianism/Egalitarianism (HE) on the Effect of Anchor Race and Tweet Condition on Perceived Anchor Expertise (constant) Anchor Race (AR) Neutral Tweet Condition (NTC) Social Justice Tweet Condition (SJTC) Coeff. 2.58 .35 -.06 -.42 SE .37 .45 .14 .15 78 t 7.01 .76 -.43 p < .001 .45 .67 CILL, UL 1.86, 3.30 -.54, 1.24 -.35, .22 -2.82 < .05 -.71, -.13 Table 9 (cont’d) AR x NTC AR x SJTC HE Race x HE Gender Age Participant Race Party Affiliation -.04 .21 .58 -.04 .004 -.003 .23 -.05 .20 .21 .06 .08 .04 .003 .09 .02 -.19 1.03 10.26 -.55 .08 -1.11 2.59 -3.18 .85 .30 < .001 .58 .93 .27 < .05 < .05 -.36, .44 -.19, .62 .47, .69 -.20, .11 -.08, .09 -.008, .002 .06, .41 -.08, -.01 R = .53, R2 = .28, MSE = 1.05 F(11, 585) = 21.04, p < .001 RQ1d: News Story Evaluation. Anchor race, tweet condition, and HE (including covariates) explained 31% of the variance in news story evaluation. The main effects of anchor race was not significant. While the neutral tweet condition was not significant the social justice tweet was significant, Coefficient = -.29, SE = .14, t = -2.07, p < .05, CILL,UL = -.56, -.01, indicating that those assigned to the social justice tweet condition evaluated the stories less favorable than those assigned to the two other conditions. HE was positively associated with news story evaluation, Coefficient = .55, SE = .05, t = 10.39, p < .001, CILL,UL = .45, .66. Additionally participant age, (younger > older) and party race (Blacks > Whites) were significant covariates. Table 10. Moderating Effect of Humanitarianism/Egalitarianism (HE) on the Effect of Anchor Race and Tweet Condition on Perceived Anchor News Story Evaluation Coeff. 2.58 .04 CILL, UL 1.90, 3.26 -.79, .87 p < .001 .93 t 7.46 .09 SE .35 .43 (constant) Anchor Race (AR) Neutral Tweet Condition (NTC) Social Justice Tweet Condition (SJTC) AR x NTC AR x SJTC .83 .41 -.15, .38 -2.07 -.79 .71 < .05 -.56, -.01 .43 .48 -.53, .23 -.24, .52 .11 -.29 -.15 .14 .14 .14 .19 .19 79 Table 10 (cont’d) HE Race x HE Gender Age Participant Race Party Affiliation .55 .03 -.02 -.007 .25 -.02 .05 .07 .04 .002 .08 .02 10.39 .40 -.57 -2.94 2.93 -1.62 < .001 .69 .57 < .05 < .05 .11 .45, .66 -.11, .17 -.10, .06 -.01, -.002 .08, .41 -.05, .005 R = .56, R2 = .31, MSE = .92 F(11, 585) = 23.99, p < .001 RQ1d: News Story Evaluation. Anchor race, tweet condition, and HE (including covariates) explained 31% of the variance in news story evaluation. The main effects of anchor race was not significant. While the neutral tweet condition was not significant the social justice tweet was significant, Coefficient = -.29, SE = .14, t = -2.07, p < .05, CILL,UL = -.56, -.01, indicating that those assigned to the social justice tweet condition evaluated the stories less favorable than those assigned to the two other conditions. HE was positively associated with news story evaluation, Coefficient = .55, SE = .05, t = 10.39, p < .001, CILL,UL = .45, .66. Additionally participant age, (younger > older) and party race (Blacks > Whites) were significant covariates. 80 Chapter 8 Discussion, Limitations, and Conclusion The first focus in this study is to examine the current state of attitudes of African Americans/Blacks and Whites toward African American/Black and White female news professionals. Research on racial attitudes and news personalities has long shown negative attitudes by Whites, consistent with structural racism (Dixon, 2007, 2017; Entman, 1992). In the late 1960s, the Kerner Commission acknowledged structural racism throughout U.S. institutions like banking, education, healthcare, and housing (Disorders & Commission, 1968). More than 60 years later during the 2020 racial reckoning, activists and allies made it known that structural racism hadn’t disappeared (Kishi & Jones, 2020). Entman (1992) said modern racism included the denial of systemic racism. A second focus is to ask how African Americans/Blacks and Whites react to an important norm in the news business (Deuze & Witschge, 2018; Hanitzsch & Vos, 2016; Raeijmaekers & Maeseele, 2017), that is reporters tell stories, but they do not share their personal opinions about the news (Deuze, 2019; Schudson, 2001, 2012). This has long been the norm in Western news but has become even more important as the social media environment came into being and became so crucial for financial success of news companies. News professionals are commonly required to post on Facebook, Instagram, Twitter, etc. (Dutta & Gangopadhyay, 2019; Santana & Hopp, 2016). Does the objectivity norm apply to these social media environments? Here, tweets were created to be neutral or opinionated, or in a control condition, were not present. Effects of the tweets were theorized in terms of expectation violation theory and used as a guide as others have used it (Lee et al., 2020). The theory suggests that when audience members perceive a 81 behavior to be different from what they expect, and especially when the behavior is considered inappropriate to a norm, there is negative attitude response. Finally, the study aimed to examine the interaction between tweet types (and presumed degree of norm violation). Would the negative attitudes of White audiences of African American/Black news professionals be increased by the opinionated tweet? While these hypotheses and the research question about the interaction of tweets and race were straight- forwardly derived from theory, as is often the case, the results were different, interesting, and important. Race and response to African American/Black and White anchors The results showed that anchor race, participant race, and types of tweets matter to news viewers. The different measures of attitude toward the African American/Black and White anchors showed variations in how they were affected by race. It’s evident, viewers like to see people like themselves on television news. When they see more people like themselves, they’re more accepting of the entire newscast crew. It’s also important to connect with newscast anchors and reporters. Viewers like to know their local celebrities have the same highs and lows personally as they do. That may mean talking about staying up all night with a sick child or enjoying a new hobby like riding bicycles with new friends around the city. Implications of these findings are that structural racism is alive and well, at least in this case. If looking at the main effects, one might believe African American/Black anchors will be well liked by a majority White viewing population but delving deeper into the results, when looking at just the White participants, they trusted, and believed the White anchor was more of an expert, and evaluated stories read by White anchors more favorably. This latter finding is supportive of structural racism operating in the television news audience. 82 Effects of tweets Opinionated, neutral, and no-tweet conditions did not show the expected values of expectation violation that were expected. This was because expectation violation values of the tweets varied extensively as a function of audience race, anchor race, age, and political affiliation. This research confirms Bennett et al. (2020), that message matters. While the SJ tweet was liked the least, the under 45 age group liked the anchor more. Ultimately, TV news personnel should be cautious when tweeting or sharing their opinion on social media. Tweeting, “I dislike ketchup on my hot dog” could spark an animated debate but may not impact a loyal viewers habit of watching your channel. However, tweeting about something controversial like “Black Lives Matter,” “Gun ights,” “ eproductive ights/Pro-Life” issues may gain and/or lose viewers depending on where they fall on the political spectrum. The caveat is this norm may only apply to older viewers. Younger viewers appear to appreciate personal opinions from news professionals, for good or for bad. It’s plausible that because younger viewers have grown up with social media, they’re less concerned with the foibles news people make or for that matter anyone makes in the social media universe, no matter the platform. Additionally, it's plausible that they experienced a positive violation (Burgoon, 2015). This scholarship confirms that violations aren’t always negative (Walther-Martin, 2015). This shows that expectation violation processes are important to the response to African American/Black and White anchors and the stories they tell. But just what stimuli lead to these violations vary and are not correlated with the tweet manipulations here. Further, the expected negative effect of violations did not occur. Instead, the more surprised people were (more expectation violation) by the tweets, the more positivity they expressed about the anchors and their stories. This may be an effect of novelty. It has been described as something new and/or 83 unexpected, a surprise if you will (Rodero & Cores-Sarría, 2023). Online news sites use algorithms to detect novelty by screening what online users like to read, view, and read and try to provide similar options to their online users. News websites keep track of what their users do and try to provide viewers and readers with updated information (Broussard & Boss, 2018; Gaughan & Smeaton, 2005; Kumari et al., 2021). Scholars continue to look for new ways to use novelty detection to get news stories in the social media feeds of news users (Gaughan & Smeaton, 2005). In the case of my research, I suspect the unexpected tweet caused a novelty reaction, leading participants to like the anchors more. The relationship between novelty and expectation violation should be examined more closely in follow-up studies. It may be that the relationship is of considerable relevance to stereotyped responses by African Americans/Blacks and Whites to the race of news professionals. There can be positive expectation violations as well as negative ones. Here, the positive ones were operant in influencing attitudes. Interaction of tweets, audience race, and anchor race There were significant differences between the interaction of tweets, participant race and anchor race that align with the hypotheses that White participants would like the White anchors more than the Black anchors. It is not surprising that White participants would be offended by the social justice tweet. It is surprising that younger Black participants experienced more expectancy violation due to the social justice tweet. These results suggest African Americans/Blacks and Whites differ in how they understand journalism norms and have expectations of how journalists should behave in terms of tweeting their opinions or objective representations of the news. Their trust of news anchors to be fair and balanced has evolved from the time period when in the early days of TV news, they didn’t see people like themselves on television news (Mills, 2004). 84 Importance of the study The results showed that even in 2023, there are important differences in how African American/Black and White audience members deal with African American/Black and White news professionals. Overall: 1. Audience members like African American/Black anchors, but White audiences trust White anchors more and evaluate their stories more positively. 2. As I look at the results again, I think this isn’t true. There’s no difference for Whites. It’s African American/Blacks that like African American/Blacks more than Whites. 3. Tweet conditions make a difference. In the no tweet and social justice tweet conditions, White participants perceived the anchors to be less trustworthy. In the neutral tweet condition, White participants perceived the anchors to be more trustworthy. This difference can also be seen by age of the participant. Younger participants experienced a greater expectancy violation. For younger Whites the shock was due to the neutral tweet. For younger Blacks it was due to the no tweet and social justice tweet conditions. Race of anchors and tweet type do not interact. However, expectancy violation relates in a significant way with three main IVs but relates in the opposite direction to what was hypothesized. The violations may have led to a novelty result (Rodero & Cores-Sarría, 2023). “EV” is an important variable. It is not working like hypothesized, but it is a boost. This boost does not turn off viewers but turns them on. It could be a “looky loo” effect, whereas the younger viewer is waiting for a shoe to drop, or they want to know what else the anchor may say that they either agree or disagree with. It also appears, the age of the news user does matter in how they respond to the tweet of an anchor. 85 Higher expectation violation scores were associated with higher favorability of the news anchors and news stories. That’s novelty. Expectancy violation helps, not hurts anchors. Surprisingly, they’re positive. The less they expect the tweets to occur, the more positive they are about the “whole” thing. When looking at tweet condition, race of the anchor, and looking at interaction of race of the participant, where those tweet conditions, didn’t align with expectancy violation there were expectancy violations going on. Further research is needed. By the way, just a pure tweet isn’t going to tell you about expectancy violation, one will have to look at the interaction of participant race, anchor race, the attitude of the participant, and the type of tweet. Limitations As with all experiments, there are possible shortcomings that need to be explored in follow-up studies. First, there was no pretesting of attitude toward the African American/Black and White anchors. This will be remedied by conducting a post-test before the paper is prepared for publication. Although, the sample found them to be equally attractive and for the control condition participants, they observed the anchors to be similar in expertise. It should be noted, however, that the varied responses of the African American/Black and White audience members to the anchors suggests that there were not prior attitudinal differences to the anchors that were producing those responses. Also, examination of the stimulus anchors does intuitively suggest there were no differences in how “attractive” they were. The tweets were pretested for expectancy violation, but they did not perform the same in the main study. In the pretest, the social justice tweet was flagged by participants as being inappropriate, offensive, and mixing opinion with news. The hope was that these pretested tweets would produce hypothesized results. Unfortunately, that was not the case. 86 It is clear from the results that expectancy violation is not straight-forwardly related to the norm of objectivity of news professionals. For some of the respondents the opinionated tweet was positive, not universally rejected. It should not be assumed that journalistic norms like refraining from sharing one’s own views about the news are what audience members believe or are influenced by. The effects of the tweet types here were not due to expectancy violation, even though expectancy violation did affect the attitudes toward the anchors. New theorizing about opinionated vs non-opinionated tweets or social media posts is needed, especially as posting is such a major activity of journalists. Further, there is need for more study of how social media behavior norms of journalists are shared or not by audience members. Conclusion The study provides important findings about news audiences, news professional race, and journalistic tweeting. The humanitarianism-egalitarianism scale provides insights to the person who cares for others and is empathetic. Some might describe it as a “Pollyanna” view of the world. Despite the non-significance of the HE variable, when participants had a high HE score, they found the anchor attractive, trusted her, found her to be an expert, and evaluated her story more favorably than those with a low HE score. This research supports some of the guidelines issued by media companies and RTDNA to news employees (How to Improve Trust in Local Journalism, 2022). The no tweet condition or a neutral news tweet – just report the facts - are favored. And even though those under 45 had higher expectancy violations, they were shocked by the tweets, they appreciated the candor. Some of the response to tweets could be attributed to the current state of twitter by its current owner who has randomly changed policies to the ire of its users. 87 Between African American/Black participants and White participants, African American/Black participants trusted the anchors more, no matter the race. They perhaps are a loyal audience worth courting. That courting may mean adding BIPOC anchors and targeting social media posts towards the under 45 crowd. This may be the way to attract this age group back to linear news. These results also indicate it’s okay to tweet, just don’t include your opinion for those over 45. While the social justice tweet may have been liked the least, the younger participants appreciated the candor. And just as important, the study raises intriguing next questions in this important area of journalistic research. These include the use of social media to attract news users to linear television and/or online, and not forgetting that attracting loyal BIPOC viewers could be just as important as keeping loyal White viewers. The fine line could be the perfect social media post that keeps them engaged and watching what’s up next. 88 BIBLIOGRAPHY 10 Best Virtual Worlds for Adults. (2022). Sensorium. Retrieved February 26 from https://sensoriumxr.com/articles/best-virtual-worlds-for-adults Allport, G. W. (1958). The nature of prejudice (Abridg ed.) [Non-fiction]. Doubleday. https://search.ebscohost.com/login.aspx?direct=true&db=cat09242a&AN=msuc.b213611 74&site=eds-live Afifi, W. A., & Metts, S. (1998). Characteristics and consequences of expectation violations in close relationships. Journal of Social and Personal Relationships, 15(3), 365-392. Aichner, T., Grünfelder, M., Maurer, O., & Jegeni, D. (2021). Twenty-five years of social media: a review of social media applications and definitions from 1994 to 2019. Cyberpsychology, behavior, and social networking, 24(4), 215-222. Altemeyer, B. (1983). Right-wing authoritarianism. Univ. of Manitoba Press. Anderson, C. W., Bell, E., & Shirky, C. (2015). Post-Industrial Journalism: adapting to the Present. Geopolitics, history & international relations, 7(2). Anisef, P., Ashbury, F. D., & Turrittin, A. H. (1992). Differential effects of university and community college education on occupational status attainment in Ontario. Canadian Journal of Sociology/Cahiers canadiens de sociologie, 69-84. Arhin, A., & Thyer, B. A. (2004). The causes of racial prejudice: A behavior analytic perspective. The psychology of racial prejudice and discrimination, 1, 1-19. Artz, L. (2003). Globalization, media hegemony, and social class. The globalization of corporate media hegemony, 3-31. ASNE. (2016). American Society of Newspaper Editors. asne.org/diversity Barnett, M. (2005). Humanitarianism transformed. Perspectives on politics, 3(4), 723-740. Barnidge, M., Gunther, A. C., Kim, J., Hong, Y., Perryman, M., Tay, S. K., & Knisely, S. (2020). Politically motivated selective exposure and perceived media bias. Communication Research, 47(1), 82-103. Bargh, J. A., & Chartrand, T. L. (1999). The unbearable automaticity of being. American psychologist, 54(7), 462. Beere, C. A., King, D. W., Beere, D. B., & King, L. A. (1984). The Sex-Role Egalitarianism Scale: A measure of attitudes toward equality between the sexes. Sex Roles, 10, 563-576. 89 Bélair-Gagnon, V., & Usher, N. (2021). Journalism Research that Matters. Oxford University Press. Bennett, L. K., Scruggs, X., & Woods, J. M. (2020). Surprise, hurt, and anger as emotional responses to expectancy violations following feedback messages. Communication Research Reports, 37(1-2), 22-33. Bercovici, J. (2010, December 9th, 2010). Who Coined 'Social Media'? Web Pioneers Compete for Credit. Forbes. https://www.forbes.com/sites/jeffbercovici/2010/12/09/who-coined- social-media-web-pioneers-compete-for-credit/?sh=12f1b6dd51d5 Bettencourt, B. A., Dill, K. E., Greathouse, S. A., Charlton, K., & Mulholland, A. (1997). Evaluations of ingroup and outgroup members: The role of category-based expectancy violation. Journal of experimental social psychology, 33(3), 244-275. Bevan, J. L., Ang, P.-C., & Fearns, J. B. (2014). Being unfriended on Facebook: An application of expectancy violation theory. Computers in Human Behavior, 33, 171-178. Bird, R., Frank Newport. (2017). What Determines How Americans Perceive Their Social Class? Gallup. https://news.gallup.com/opinion/polling-matters/204497/determines-americans- perceive-social-class.aspx Blaine, B. (2007). Understanding gender stereotypes and sexism. Understanding the Psychology of Diversity, 101-121. Bodenhausen, G. V., & Macrae, C. N. (2013). Stereotype activation and inhibition. In Stereotype activation and inhibition (pp. 9-60). Psychology Press. Bodrunova, S. S., Litvinenko, A. A., & Blekanov, I. S. (2018). Please follow us: media roles in Twitter discussions in the United States, Germany, France, and Russia. Journalism Practice, 12(2), 177-203. Brewer, M. B. (2007). The social psychology of intergroup relations: Social categorization, ingroup bias, and outgroup prejudice. Bruns, A., & Nuernbergk, C. (2019). Political journalists and their social media audiences: New power relations. Media and Communication, 7(1), 198-212. Bureau, U. S. C. (2022). QuickFacts Cincinnati city, Ohio. Retrieved August 7, 2023 from https://www.census.gov/quickfacts/fact/table/cincinnaticityohio/PST045222 Burgoon, J. K. (2015). Expectancy violations theory. The international encyclopedia of interpersonal communication, 1-9. Burgoon, J. K. (1993). Interpersonal expectations, expectancy violations, and emotional communication. Journal of language and social psychology, 12(1-2), 30-48. 90 Burgoon, J. K. (1978). A communication model of personal space violations: Explication and an initial test. Human communication research, 4(2), 129-142. Burgoon, J. K., & Jones, S. B. (1976). Toward a theory of personal space expectations and their violations. Human communication research, 2(2), 131-146. Butler, B. (2012). 2012 NABJ Diversity Census. http://www.nabj.org/news/103235/NABJ- Releases-2012-Television-Newsroom-Management-and-Network-Diversity-Census.htm Byerly, C. M., & Wilson II, C. C. (2009). Journalism as Kerner turns 40: Its multiculticultural problems and possibilities. The Howard journal of communications, 20(3), 209-221. Carlson, M. (2016). Metajournalistic Discourse and the Meanings of Journalism: Definitional Control, Boundary Work, and Legitimation [Report]. Communication Theory, 26(4), 349. https://doi.org/10.1111/comt.12088 Carlson, M., Robinson, S., Lewis, S. C., & Berkowitz, D. A. (2018). Journalism studies and its core commitments: The making of a communication field. Journal of Communication, 68(1), 6-25. Chan, J. M. (2019). From networked commercialism to networked authoritarianism: The biggest challenge to journalism. Journalism, 20(1), 64-68. Cohen, E. L. (2010). Expectancy violations in relationships with friends and media figures. Communication Research Reports, 27(2), 97-111. Coleman, R. (2011). Color blind: Race and the ethical reasoning of Blacks on journalism dilemmas. Journalism & Mass Communication Quarterly, 88(2), 337-351. Collins, P. H., & Bilge, S. (2020). Intersectionality. John Wiley & Sons. Cozma, R., & Chen, K.-J. (2013). What's in a tweet? Foreign correspondents' use of social media. Journalism Practice, 7(1), 33-46. Crenshaw, K. (2018). Demarginalizing the intersection of race and sex: A Black feminist critique of antidiscrimination doctrine, feminist theory, and antiracist politics [1989]. In Feminist legal theory (pp. 57-80). Routledge. Crenshaw, K. (1990). Mapping the margins: Intersectionality, identity politics, and violence against women of color. Stan. L. Rev., 43, 1241. Darcy, O. (2022, June 10, 2022). The Washington Post fires reporter Felicia Sonmez after a week of feuding publicly with her colleagues. CNN. Retrieved April 15 from https://www.cnn.com/2022/06/09/media/felicia-sonmez-washington-post/index.html Deuze, M. (2019). What journalism is (not). Social media+ society, 5(3), 2056305119857202. 91 Deuze, M. (2005). What is journalism? Professional identity and ideology of journalists reconsidered. Journalism, 6(4), 442-464. Deuze, M., & Witschge, T. (2018). Beyond journalism: Theorizing the transformation of journalism. Journalism, 19(2), 165-181. Devine, P. G., & Sharp, L. B. (2009). Automaticity and control in stereotyping and prejudice. Disorders, U. S. N. A. C. o. C., & Commission, U. S. K. (1968). Report of the national advisory commission on civil disorders. US Government Printing Office. Dunton, B. C., & Fazio, R. H. (1997). An individual difference measure of motivation to control prejudiced reactions. Personality and social psychology bulletin, 23(3), 316-326. Dutta, S., & Gangopadhyay, S. (2019). Digital journalism: Theorizing on present times. Media Watch, 10(3), 713-722. Ehrlich, A. (2019). Who Owns Your Local News? Sinclair Broadcasting Invests in 'The Medford Experiment''. Jefferson Public Radio, Southern Oregon University. Retrieved June 14th, 2022 from https://www.ijpr.org/business-and-labor/2019-07-16/who-owns-your-local- news-sinclair-broadcasting-invests-in-the-medford-experiment eMarketer, I. I. (2022). Number of Twitter users worldwide from 2019 to 2024. Retrieved March 15th, 2023 from https://www.statista.com/statistics/303681/twitter-users-worldwide/ Ferguson, L. W. (1944). Socio-psychological correlates of the primary attitude scales: I. Religionism; II. Humanitarianism. The Journal of Social Psychology, 19(1), 81-98. Fraser, N. A., & Murakami, G. (2022). The role of humanitarianism in shaping public attitudes toward refugees. Political Psychology, 43(2), 255-275. Gaither, S. E., Perlin, J. D., & Doan, S. N. (2020). Race, gender, and the development of cross- race egalitarianism. Frontiers in Psychology, 11, 1525. Gallagher, C. A. (2015). Color-blind egalitarianism as the new racial norm. Theories of race and ethnicity: Contemporary debates and perspectives, 40-56. George, A. (2018, March 1, 2018). The 1968 Kerner Commission Got It Right, But Nobody Listened. Smithsonian Magazine. https://www.smithsonianmag.com/smithsonian- institution/1968-kerner-commission-got-it-right-nobody-listened-180968318/ Gong, Z. H., & Eppler, J. (2022). Exploring the impact of delivery mistakes, gender, and empathic concern on source and message credibility. Journalism Practice, 16(8), 1653- 1672. 92 Gottfried, J., Amy Mitchell, Mark Jurkowitz, Jacob Liedke. (2022, June 14, 2022). Many journalists say social media helps at work, but most decry its impact on journalism. Pew Research Center. Retrieved July 5, 2023 from https://www.pewresearch.org/journalism/2022/06/14/many-journalists-say-social-media- helps-at-work-but-most-decry-its-impact-on-journalism/ Greenhow, C., & Gleason, B. (2012). Twitteracy: Tweeting as a new literacy practice. The Educational Forum, Grinberg, N., Kalyanaraman, S., Adamic, L. A., & Naaman, M. (2017). Understanding feedback expectations on Facebook. Proceedings of the 2017 ACM conference on computer supported cooperative work and social computing. Haan, K., Main, Kelly. (2023). Best Blogging Platforms (February 2023). Forbes Advisor. https://www.forbes.com/advisor/business/software/best-blogging-platforms/ Hall, S. (1997). The spectacle of the other. Representation: Cultural representations and signifying practices, 7. Hanitzsch, T., & Vos, T. P. (2016). Journalism beyond democracy: A new look into journalistic roles in political and everyday life. In: SAGE Publications. Hanitzsch, T., & Vos, T. P. (2017). Journalistic roles and the struggle over institutional identity: The discursive constitution of journalism. Communication Theory, 27(2), 115-135. Harlow, S. (2022). Journalism’s Change Agents: Black Lives Matter,# BlackoutTuesday, and a Shift Toward Activist Doxa. Journalism & Mass Communication Quarterly, 99(3), 742- 762. Harnish, R. J., Nataraajan, R., Tarka, P., & Slack, F. J. (2023). Attitudes toward protecting endangered species: The impact of perceived physical attractiveness of animals and political ideology. Psychology & Marketing, 40(1), 73-88. Hayes, A. F. (2017). Introduction to mediation, moderation, and conditional process analysis: A regression-based approach. Guilford publications. Hedman, U. (2020). Making the most of Twitter: How technological affordances influence Swedish journalists’ self-branding. Journalism, 21(5), 670-687. Hill, A. (2007). Restyling factual TV: Audiences and news, documentary and reality genres. Routledge. Hinnant, A., Hu, S., Hong, Y., & Young, R. (2023). Contested Certainty and Credibility: The Effect of Personal Stories and Scientific Evidence in User Comments on News Story Evaluation and Relevance. Science communication, 45(1), 65-94. 93 Houston, J. B., McKinney, M. S., Thorson, E., Hawthorne, J., Wolfgang, J. D., & Swasy, A. (2020). The twitterization of journalism: User perceptions of news tweets. Journalism, 21(5), 614-632. Houston III, L., Grandey, A. A., & Sawyer, K. (2018). Who cares if “service with a smile” is authentic? An expectancy-based model of customer race and differential service reactions. Organizational Behavior and Human Decision Processes, 144, 85-96. How to Improve Trust in Local Journalism. (2022). Magid. Johnson, J. L., & Repta, R. (2012). Sex and gender: Beyond the binaries [Chapter]. SAGE Publications Inc. https://doi.org/10.4135/9781452230610.n2 Johnson, K. A. (2020). I got a new puppy! The impact of personal, opinion, and objective tweets on a journalist’s and a news organization’s perceived credibility. Journalism Practice, 14(1), 48-66. Jones, J. M. (1972). Prejudice and racism [Bibliographies Non-fiction]. Addison-Wesley Pub. Co. https://search.ebscohost.com/login.aspx?direct=true&db=cat09242a&AN=msuc.b142550 17&site=eds-live Kadylak, T. (2020). An investigation of perceived family phubbing expectancy violations and well-being among US older adults. Mobile Media & Communication, 8(2), 247-267. Katz, I., & Hass, R. G. (1988). Racial ambivalence and American value conflict: Correlational and priming studies of dual cognitive structures. Journal of personality and social psychology, 55(6), 893-905. https://doi.org/https://doi.org/10.1037/0022-3514.55.6.893 Kempton, S. D., & Connolly Ahern, C. (2021). Mastering Metrics: The Impact of Social Media on the Routines and Values of Broadcast Journalists. Electronic News, 15(3-4), 109-122. Kim, S. (2014). The role of prior expectancies and relational satisfaction in crisis. Journalism & Mass Communication Quarterly, 91(1), 139-158. King, L. A., & King, D. W. (1986). Validity of the sex-role egalitarianism scale: Discriminating egalitarianism from feminism. Sex Roles, 15, 207-214. Kishi, R., & Jones, S. (2020). Demonstrations and Political Violence in America (US Crisis Monitor, Issue. https://acleddata.com/acleddatanew/wp- content/uploads/2020/09/ACLED_USDataReview_Sum2020_SeptWebPDF_HiRes.pdf Klein, H., & Shiffman, K. S. (2009). Underrepresentation and symbolic annihilation of socially disenfranchised groups (“out groups”) in animated cartoons. The Howard journal of communications, 20(1), 55-72. 94 Kohring, M., & Matthes, J. (2007). Trust in news media: Development and validation of a multidimensional scale. Communication Research, 34(2), 231-252. Kozman, C., & Cozma, R. (2021). Lebanese television on Twitter: A study of uses and (dis) engagement. Journalism Practice, 15(4), 508-525. Kteily, N. S., Sheehy-Skeffington, J., & Ho, A. K. (2017). Hierarchy in the eye of the beholder:(Anti-) egalitarianism shapes perceived levels of social inequality. Journal of personality and social psychology, 112(1), 136. Lang, A., Potter, D., & Grabe, M. E. (2003). Making news memorable: Applying theory to the production of local television news. Journal of Broadcasting & Electronic Media, 47(1), 113-123. Lee, E.-J., Lee, H.-Y., & Choi, S. (2020). Is the message the medium? How politicians’ Twitter blunders affect perceived authenticity of Twitter communication. Computers in Human Behavior, 104, 106188. Lee, J. (2015). The double-edged sword: The effects of journalists' social media activities on audience perceptions of journalists and their news products. Journal of Computer- Mediated Communication, 20(3), 312-329. Lewis, S. C., Zamith, R., & Coddington, M. (2020). Online harassment and its implications for the journalist–audience relationship. Digital Journalism, 8(8), 1047-1067. Lucal, B. (1996). Oppression and privilege: Toward a relational conceptualization of race. Teaching Sociology, 245-255. Lysak, S., Cremedas, M., & Wolf, J. (2012). Facebook and Twitter in the newsroom: How and why local television news is getting social with viewers? Electronic News, 6(4), 187-207. Macker, N. (2020, January 1st, 2020). New Coverage of TV Station Owners. MorningStar. https://www.morningstar.com/articles/961093/new-coverage-of-tv-station-owners Mastro, D., Seate, A. A., Blecha, E., & Gallegos, M. (2012). The wide world of sports reporting: The influence of gender-and race-based expectations on evaluations of sports reporters. Journalism & Mass Communication Quarterly, 89(3), 458-474. Masullo, G. M., Tenenboim, O., & Lu, S. (2021). “Toxic atmosphere effect”: Uncivil online comments cue negative audience perceptions of news outlet credibility. Journalism, 14648849211064001. McConahay, J. B., Hardee, B. B., & Batts, V. (1981). Has racism declined in America? It depends on who is asking and what is asked. Journal of conflict resolution, 25(4), 563- 579. 95 Meltzer, K. (2010). TV news anchors and journalistic tradition: How journalists adapt to technology. Peter Lang. Mills, K. (2004). Changing channels: The civil rights case that transformed television. Univ. Press of Mississippi. Molyneux, L., & Mourão, . . (2019). Political journalists’ normalization of Twitter: Interaction and new affordances. Journalism studies, 20(2), 248-266. Nerone, J. (2013). Why Journalism History Matters to Journalism Studies [Article]. American Journalism, 30(1), 15-28. https://doi.org/10.1080/08821127.2013.767693 Neuberger, C., Nuernbergk, C., & Langenohl, S. (2019). Journalism as Multichannel Communication: A newsroom survey on the multiple uses of social media. Journalism studies, 20(9), 1260-1280. Number of monetizable daily active Twitter users in the US. (2022). statista. Retrieved April 15, 2023 from https://www.statista.com/statistics/970911/monetizable-daily-active-twitter- users-in-the-united-states/ Ohanian, R. (1990). Construction and validation of a scale to measure celebrity endorsers' perceived expertise, trustworthiness, and attractiveness. Journal of advertising, 19(3), 39- 52. Osteen, P., Morris, R., Castillo, J., & Baffour, T. D. (2023). Professional Values, Gatekeeping, and Motivations for Seeking the MSW Degree. Journal of Social Work Education, 59(1), 66-77. Papper, B. a. K. H. (2022). TV, Radio news profits rise, but short of pre-COVID levels. RTDNA. Retrieved June 3rd, 2022 from https://www.rtdna.org/article/tv_radio_news_profits_rise_but_short_of_pre_covid_levels Peters, C., & Carlson, M. (2019). Conceptualizing change in journalism studies: Why change at all? In (Vol. 20, pp. 637-641): SAGE Publications Sage UK: London, England. Plant, E. A., & Devine, P. G. (1998). Internal and external motivation to respond without prejudice. Journal of personality and social psychology, 75(3), 811. Post, T. W. (2016). Ex-TV anchor sues broadcaster for discrimination after she was fired for comments on black violence. https://nationalpost.com/news/world/ex-tv-anchor-sues- broadcaster-for-discrimination-after-she-was-fired-for-comments-on-black-violence Press, C. o. F. o. t. (1947). A free and responsible press: A general report on mass communication: Newspapers, radio, motion pictures, magazines, and books. University of Chicago Press. 96 Raeijmaekers, D., & Maeseele, P. (2017). In objectivity we trust? Pluralism, consensus, and ideology in journalism studies. Journalism, 18(6), 647-663. Rigoni, I. (2012). Intersectionality and mediated cultural production in a globalized post-colonial world. Ethnic and Racial Studies, 35(5), 834-849. yan, K. M., & Mapaye, J. C. (2010). Beyond “Anchorman” a comparative analysis of race, gender, and correspondent roles in network news. Electronic News, 4(2), 97-117. Santana, A. D., & Hopp, T. (2016). Tapping into a new stream of (personal) data: Assessing journalists’ different use of social media. Journalism & Mass Communication Quarterly, 93(2), 383-408. Scherbov, S., & Sanderson, W. C. (2016). New approaches to the conceptualization and measurement of age and aging. Journal of aging and health, 28(7), 1159-1177. Schudson, M. (2012). The sociology of news (ed.). In: New York: WW Norton & Company. Schudson, M. (2001). The objectivity norm in American journalism. Journalism, 2(2), 149-170. Sheehy-Skeffington, J., & Thomsen, L. (2020). Egalitarianism: psychological and socio- ecological foundations. Current opinion in psychology, 32, 146-152. Sigman, R., & Lindberg, S. I. (2019). Democracy for all: conceptualizing and measuring egalitarian democracy. Political Science Research and Methods, 7(3), 595-612. Smith, J. (2012, December 12, 2012). Meterologist fired after defending her 'ethnic' hair. MSNBC. Retrieved April 15, 2023 from https://www.msnbc.com/melissa-harris- perry/meteorologist-fired-after-defending-her-ethn-msna16859 Steensen, S., & Westlund, O. (2020). What is Digital Journalism Studies? Routledge. http://ezproxy.msu.edu/login?url=https://search.ebscohost.com/login.aspx?direct=true&A uthType=ip,uid,cookie&db=edsnor&AN=edsnor.10642.9280&site=eds-live Stolier, R. M., & Freeman, J. B. (2016). Chapter 7 - The Neuroscience of Social Vision. In J. R. Absher & J. Cloutier (Eds.), Neuroimaging Personality, Social Cognition, and Character (pp. 139-157). Academic Press. https://doi.org/https://doi.org/10.1016/B978-0-12- 800935-2.00007-5 Tameez, H. (2022, March 8, 2022). Ameircan Journalism's "Racial Recloning" Still Has Lots of Reckoning to Do. Nieman Lab. https://www.niemanlab.org/2022/03/american- journalisms-racial-reckoning-still-has-lots-of-reckoning-to-do/ Times, N. Y. (2017, October 13, 2017). The Times Issues Social Media Guidelines for the Newsroom. Retrieved July 5, 2023 from https://www.nytimes.com/2017/10/13/reader- center/social-media-guidelines.html 97 Van Looy, A. (2022). Definitions, social media types, and tools. In Social Media Management: Using Social Media as a Business Instrument (pp. 21-50). Springer. Vraga, E., Tully, M., Kotcher, J. E., Smithson, A.-B., & Broeckelman-Post, M. (2015). A Multi- Dimensional Approach to Measuring News Media Literacy. Journal of media literacy education, 7(3), 41-53. Vázquez-Herrero, J., Negreira-Rey, M.-C., & López-García, X. (2022). Let’s dance the news! How the news media are adapting to the logic of TikTok. Journalism, 23(8), 1717-1735. Waddell, T. F. (2018). A robot wrote this? How perceived machine authorship affects news credibility. Digital Journalism, 6(2), 236-255. Wahl-Jorgensen, K., & Hanitzsch, T. (2009). Introduction: On why and how we should do journalism studies. In The handbook of journalism studies (pp. 23-36). Routledge. Wahl-Jorgensen, K., & Hanitzsch, T. (2019). Journalism studies: Developments, challenges, and future directions. In The handbook of journalism studies (pp. 3-20). Routledge. Walther-Martin, W. (2015). Media-generated expectancy violations: A study of political humor, race, and source perceptions. Western Journal of Communication, 79(4), 492-507. Watermeyer, R. (2012). Social networking sites. Weaver, D. H., & Willnat, L. (2016). Changes in US journalism: How do journalists think about social media? Journalism Practice, 10(7), 844-855. Weaver, D. H., Willnat, L., & Wilhoit, G. C. (2019). The American journalist in the digital age: Another look at US news people. Journalism & Mass Communication Quarterly, 96(1), 101-130. White, E. S. (2021). Parent values, civic participation, and children’s volunteering. Children and Youth Services Review, 127, 106115. Wilhelm, C., Stehle, H., & Detel, H. (2021). Digital visibility and the role of mutual interaction expectations: Reframing the journalist–audience relationship through the lens of interpersonal communication. New Media & Society, 23(5), 1004-1021. Yaden, D. B., Batz-Barbarich, C. L., Ng, V., Vaziri, H., Gladstone, J. N., Pawelski, J. O., & Tay, L. (2022). A meta-analysis of religion/spirituality and life satisfaction. Journal of Happiness Studies, 23(8), 4147-4163. Yuan, S., & Lu, H. (2020). “It’s global warming, stupid”: Aggressive communication styles and political ideology in science blog debates about climate change. Journalism & Mass Communication Quarterly, 97(4), 1003-1025. 98 Zelizer, B. (2004). Taking journalism seriously: News and the academy. Sage Publications. Zhou, F., Lin, Y., Mou, J., Cohen, J., & Chen, S. (2023). Understanding the dark side of gamified interactions on short-form video platforms: Through a lens of expectations violations theory. Technological Forecasting and Social Change, 186, 122150. Zubiaga, A. (2019). Mining social media for newsgathering: A review. Online Social Networks and Media, 13, 100049. 99 APPENDIX A: Pretest I Tweets TWEET: 1) Just like Philando Castile, the shooting of a Cincinnati Black man by police was streamed live on Facebook. We’ll show you the video and have reaction at 11. 2) Three years after the 2020 George Floyd protests, how activists are fighting for racial equality. The story at 11. 3) People from all over Ohio gathered peacefully to protest the latest shooting of a Black man, Jayland Walker, in Akron, Ohio. The story at 11. 4) Study concludes that more Black physicians could improve life expectancy of Black patients. The story at 11. 5) A Black man in Cincinnati was killed by police. The shooting was streamed live on Facebook. The story at 11. 6) Climate activists address air pollution in their urban communities. We’ll show you how they’re fighting for clean air at 11. 7) Police officer shooting on Broadway. Black man killed in traffic stop. Live report coming up at 11. 8) New study projects hate crimes against Blacks and other racial minorities are set to spike during the 2024 presidential election. The story at 11. 9) Ebony Williams, a Black mother, talks about protecting her son from harassment by a Cincinnati police officer. The story at 11. 10) Black Lives Matter protests in Cincinnati today started peacefully and ended with arrests. Story at 11. TWEET with OPINION: 1) Just like Philando Castile, the shooting of a Cincinnati Black man by police was streamed live on Facebook. These are scary times. I’m afraid for my male family members/African American friends. #BlackLivesMatter The story at 11. 2) We said Black Lives Matter. We never said only Black Lives Matter. We know all lives matter. It’s more than just a movement. #BlackLivesMatter The story at 11. 100 3) People of all races/ages and backgrounds gathered peacefully to protest the latest shooting of a Black man, #JaylandWalker in Akron, Ohio. Our voices matter. Solidarity Matters #BlackLivesMatter The story at 11. 4) Study concludes more Black physicians could improve life expectancy of Black patients. Institutional racism is alive and well #BlackLivesMatter The story at 11. 5) So sad to see yet another Black man killed on our streets. The story at 11 #BlackLivesMatter 6) Climate activists address air pollution in their urban communities. Unfortunately, I’ve seen this as a problem in many Black communities. See how they’re fighting for their rights to clean air. #BlackLivesMatter 7) Police officer shooting on Broadway. Black man killed in traffic stop. Why is this still happening? Live report coming up at 11. #BlackLivesMatter 8) New study projects hate crimes against Blacks and other racial minorities are set to spike during the 2024 presidential election. Some of us aren’t surprised. The story at 11. #BlackLivesMatter 9) Ebony Williams, a Black mother, went to great length to protect her son from harassment by a police officer. These are scary times. The story at 11. #BlackLivesMatter 10) Black Lives Matter protests in Cincinnati today started peacefully and ended with arrests. It’s upsetting only Black people were arrested. Story at 11. #BlackLivesMatter 101 APPENDIX B: Pretest II Tweet Pairs Pretest II Tweet Pairs News News Opinion Tweets 8 & 18 New study projects hate crimes against Blacks and other racial minorities are set to spike during the 2024 presidential election. The story at 11. Tweets 10 & 20 Black Lives Matter protests in Cincinnati today started peacefully and ended with arrests. Story at 11. New study projects hate crimes against Blacks and other racial minorities are set to spike during the 2024 presidential election. In my opinion, this shows an extremely disturbing and sadly not so surprising trend. Even though the perpetrators of hate crimes are responsible in the public’s eye, there is no doubt that words by certain politicians have led to this upsurge in hate crimes that wreak havoc in our communities. Not acceptable. The story at 11. #BlackLivesMatter Black Lives Matter protests in Cincinnati today started peacefully and ended with arrests. I am so upset and devastated. This is not right. It should not be this way. People are fed up. It is scary that such acts of denying people their basic human rights continues in 2023. This is not acceptable. There is a need for accountability and more transparency in how police departments handle social justice protests in this country. Story at 11. #BlackLivesMatter 102 APPENDIX C: News Consumption Scale Cable Television News (e.g., CNN, Fox, MSNBC) Network Television News (e.g., ABC – David Muir, CBS – Noral O’Donnell, NBC – Lester Hold) Local Television News in your area (ABC, CBS, Fox, NBC, Independent) National Newspaper – Print Copy (e.g., New York Times, Wall Street Journal, Washington Post) Local Newspaper – Print Copy (e.g., Lansing State Journal, Detroit Free Press, Columbus Dispatch, Cincinnati Inquirer) Online website of National Newspaper (e.g., New York Times, Wall Street Journal, Washington Post) Online website of Local Newspaper (e.g., Lansing State Journal, Detroit Free Press, Columbus Dispatch, Cincinnati Inquirer) Other Online News Sources (e.g., Breitbart, Huffington Post, NewsOne) News on Facebook News on Instagram News on Reddit News on TikTok News on Twitter News on YouTube News on Other Social Media Platforms 103 APPENDIX D: Social Media use by journalists Promote Stories Connect with their audience Find sources for stories Identify stories to cover Gather accurate information for stories Build trust in news they produce Express partisan opinions Promote their own political views Endorse their favorite candidates Make offensive comments Give their opinion about a story Take a personal stance about a political issue Take a personal stance about a social issue 104 Informative Comprehensible Interesting Engaging Believable Important Enjoyable Fair Biased Accurate Truthful Tells the entire story APPENDIX E: Story evaluation scale 105