<<

SCHREYER HONORS COLLEGE

DEPARTMENT OF POLITICAL SCIENCE

EQUIVALENCY FRAMES IN THE #METOO ERA: DO THEY MATTER?

JENNIFER CASTOE SPRING 2020

A thesis submitted in partial fulfillment of the requirements for baccalaureate degrees in Digital and Print and Political Science with honors in Political Science

Reviewed and approved* by the following:

Eric Plutzer Professor of Political Science and Sociology Thesis Supervisor

Michael Berkman Professor of Political Science Honors Adviser

* Electronic approvals are on file i

ABSTRACT

The #MeToo movement sparked a national conversation about sexual harassment and assault in the , including about which language should be used when engaging in this conversation. Women’s rights and sexual violence advocacy groups, before and especially since the birth of the #MeToo movement, have provided guidelines on recommended language use for sexual violence stories. This thesis includes a content analysis of the terms that are used in the coverage of ten events within the first year of the #MeToo movement by seven national media outlets, four of which lean liberal and three lean conservative, and labels those terms according to the advocacy groups’ recommendations. The seven media outlets altogether used recommended terms roughly 75% of the time, and the liberal media used recommended terms more often than the conservative media, suggesting that ideological bias could influence how reporters write about sexual harassment and assault. ii

TABLE OF CONTENTS

LIST OF FIGURES ...... iii

LIST OF TABLES ...... iv

ACKNOWLEDGEMENTS ...... v

Chapter 1 Introduction ...... 1

Chapter 2 Literature Review ...... 3

Types of Media Framing ...... 3 Effects of Media Framing ...... 4 What are advocacy groups saying? ...... 7 1. The victim of the sexual abuse (Ex: woman, accuser, victim) ...... 8 2. The perpetrator of the abuse (Ex: the chief executive, last name, abuser) ...... 9 3. The victim’s account (Ex: said, accused, came forward) ...... 10 4. The statement of the sexual abuse (Ex: accusation, claim, allegation) ...... 10 5. The sexual abuse (Ex: sexual harassment, misconduct, abuse) ...... 11 6. The action upon the victim (Ex: harassed, abused, raped) ...... 12 Evidence for these arguments ...... 14 What are journalism institutions saying? ...... 15

Chapter 3 Theory and Hypotheses ...... 17

Chapter 4 Methodology and Operationalization ...... 19

Operationalization for Each Category of Terms ...... 21 Operationalization of Term Labelling ...... 24

Chapter 5 ...... 26

Results and Analysis ...... 26

Removing the Subjective Labels ...... 28 Which categories performed better than others? ...... 31 Incorporating the Recommendations of Journalistic Institutions ...... 33 Breakdown by Publication ...... 36 Analyzing “Accuser” ...... 39

Chapter 6 Discussion ...... 44

Appendix A Resources for Content Analysis ...... 47

BIBLIOGRAPHY ...... 55 iii

LIST OF FIGURES

Figure 1. Percentage of Codable Explicitly/Probably Recommended Terms by Advocacy Organizations, by Ideology ...... 26

Figure 2. 95% Confidence Test Results for Percentages of Explicitly and Probably Not Recommended Terms by Advocacy Organizations, by Ideology ...... 28

Figure 3. Percentage of Codable Explicitly Recommended Terms by Advocacy Organizations, by Ideology ...... 29

Figure 4. 95% Confidence Test Results for Percentages of Explicitly Not Recommended Terms by Advocacy Organizations, by Ideology ...... 30

Figure 5. Percentage of Codable Explicitly/Probably Recommended Terms by Advocacy Organizations, by Category of Term ...... 31

Figure 6. Percentage of Codable Explicitly/Probably Recommended Terms by Advocacy Organizations and Journalistic Institutions, by Ideology ...... 34

Figure 7. Percentage of Codable Explicitly/Probably Recommended Terms by Advocacy Organizations, by Publication ...... 37

iv

LIST OF TABLES

Table 1. Explicitly Recommended and Not Recommended Terms by Advocacy Organizations………………………………………………………………………..12

Table 2. Percentage of All Explicitly/Probably Recommended Terms by Advocacy Organizations, by Ideology ...... 27

Table 3. Percentage of All Explicitly Recommended Terms by Advocacy Organizations, by Ideology ...... 30

Table 4. Percentage of All Explicitly/Probably Recommended Terms by Advocacy Organizations, by Category of Term ...... 32

Table 5 . Percentage of All Probably/Explicitly Recommended Terms for Only Advocacy Organizations and Both Advocacy Organizations and Journalistic Institutions, by Ideology…………………………………………………………….35

Table 6. Percentage of All Explicitly/Probably Recommended Terms by Advocacy Organizations, by Publication ...... 38

Table 7. Percentage of "Not Recommended" Terms if "Accuser" was Unanimously Labelled as "Not Recommended" or All Terms and Codable Terms, by Ideology ... 40

Table 8. Percentage of "Not Recommended" Terms if "Accuser" was Unanimously Labelled as "Not Recommended" for All Terms and Codable Terms, by Publication ...... 41

Table 9. Timeframe Within Which #MeToo Events Were Selected for Content Analysis ...... 47

Table 10. Print Articles and Cable Segments Chosen for Content Analysis ...... 47

v

ACKNOWLEDGEMENTS

First and foremost, I want to thank my mom and sister for supporting me throughout the thesis process and for bearing with me through my various stress management techniques. Thank you to my advisors, Dr. Eric Plutzer and Dr. Michael Berkman, for challenging me to perform at my highest capabilities, to my peers for encouraging me to explore this topic, and finally, to my for always lending an ear and understanding the demands of not only the thesis process, but everything happening in conjunction.

1

Chapter 1

Introduction

The #MeToo movement is profound in many ways. An entire population of women, it seems, found a voice. Corporate human resource departments changed how they respond to and work to prevent workplace sexual harassment. Abusers were and continue to be held accountable after years of predatory behavior without consequence. A national dialogue erupted over to what extent abusers should be held accountable. Another change, one perhaps not as obvious, was a shift in how we talk, and write, about sexual abuse and the implications of our word choice. Vox Constance Grady captured the core of this shift in her 2017 article. She writes:

As I spend more and more time writing about the sexual violence that undergirds

American culture, our vocabulary for this kind of violence has begun to seem profoundly

impoverished. I’ve started to feel that I am using a language that wants to make it as

difficult as possible to describe this particular kind of violence, that wants it to remain

unspeakable, in the shadows, unnamed. (Grady, 2017)

With the rise of sexual abuse stories in the media due to the #MeToo movement and frustrations such as Grady’s for the proper way to cover , several women’s rights and sexual violence advocacy groups have published media resource guides for how to report on sexual abuse, with the underlying assumption that the language used to describe events of sexual violence have distinct nuances and connotations that can affect how and how much readers believe the stories to be true. As a result, journalistic institutions have also weighed in on this 2 language debated. In 2018, the even added into its style guide, which is the industry standard, a preference for the term “sexual misconduct” over “sexual harassment” as

“sexual harassment” may be too mild a term for some of the acts described in many of the

#MeToo stories.

However, there are no research studies that examine whether word choice in stories of sexual abuse actually influences readers’ perception of the story. There isn’t even research about which terms the media uses in stories about sexual abuse and how often, aside from occasional anecdotal evidence. This thesis will aim to answer the latter question through a content analysis of ten high-profile abusers, hereafter called #MeToo events, across seven media outlets, four of which lean liberal and 3 of which lean conservative. Then, I will analyze four sexual violence advocacy centers’ language guidelines and place the terms found in the content analysis into the context of the recommendations from the advocacy groups. I will then divide the media into liberal and conservative categories and assess how each ideological media performs within the context of the recommendations and to what extend each adheres to these recommendations.

This will set the stage for future research to examine the effects of the terms found in this content analysis on how readers believe the events outlined in these stories and potentially justify and shed light on the #MeToo language debate.

3 Chapter 2

Literature Review

Types of Media Framing

Research on the effects of framing is expansive and still continues to be studied for more than a half century. Erving Goffman first introduced the concept of media framing in 1974. He defined framing as what allows people to “locate, perceive, identify and label… concrete occurrences” through a “schemata of interpretation” (Goffman, 1974, p.21). Robert Entman expanded Goffman’s definition to include selection and salience. He writes, “To frame is to select some aspects of a perceived reality and make them more salient in a communicating text, in such a way as to promote a particular problem definition, causal interpretation, moral evaluation, and/or treatment recommendation for the item described” (Entman, 1993, p.52). An increase in salience, or prominence, increases the likelihood that the receivers of the information will perceive meaning and store it in memory (Entman, 1993, p. 53).

Media framing can either be studied as the dependent or independent variable, called

“frame building” and “frame setting,” respectively (Scheufele, 1999, p.115-16). Frame building analyzes how frames become established in social discourse and how certain frames get adopted by members of the social elite and (Scheufele, 1999, p.115). Frame setting, rather, examines the effects of the chosen frames on the audiences that interact with it (Scheufele, 1999, p. 116). This thesis explores the concept of frame setting, though it will not go so far as to examine the effects of the frames identified in the content analysis. 4 Within frame setting, James Druckman (2001) introduces two types of frames: equivalency frames and emphasis frames, referred to as issue frames. Equivalency frames are defined as the usage of different, but logically equivalent words or phrases that cause people to change their preferences (Druckman, 2001, p.228). Issue frames, on the other hand, are defined as the emphasis of a subset of potentially relevant considerations that the speaker leads people to focus on when forming their opinions (Druckman, 2001, p.230). As women’s advocacy groups and journalistic institutions alike debate over the usage of specific terms over others rather than the emphasis of specific facts or angles, I will identify the equivalency frames in place for stories concerning sexual abuse.

Effects of Media Framing

One of the most widely-cited studies on framing effects is Kahneman and Tversky’s

Asian Disease problem in which they tested the effectiveness of an equivalency frame.

Kahneman and Tversky (1984) presented subjects of Stanford University and the University of

British Columbia with two solutions to the spread of an unusual Asian disease (n = 152 for lives saved, n = 155 for lives lost). The solutions were effectively identical except one was framed in terms of lives saved and the other in terms of lives lost. Despite the fact that the number of lives lost/saved were the same, participants flipped which solution they preferred – 72% preferring the lives saved option and only 22% preferring the lives lost option – based on how it was framed

(Kahneman & Tversky, 1984, p.343). While the Asian Disease study results suggest equivalency frames have an immense effect over people’s preferences and beliefs, Kahneman and Tversky

(1984) specifically examined Prospect Theory, which is a risk problem. In other words, there 5 was an element of risk added to the participants preference. This might explain why the results varied so much between the lives saved and lives lost frames. In order to know the true effects of equivalency frames, it would be valuable to conduct a study without the element of risk.

This question was addressed in Thomas Nelson and his colleagues’ (1997) study on

“Media Framing of a Civil Liberties Conflict and Its Effect on Tolerance” in which they presented 222 undergraduate student with two versions of a story on a local KKK rally, each using a different issue frame: “free speech” or “disruption of public order” (Nelson et al., 1997, p. 570). After reading one of the two news stories, the respondents answered two questions concerning their tolerance for the KKK as well. Those who read the story through the “free speech” frame were more tolerant of the KKK holding a rally in their community, suggesting media framing affects attitude, at least in the short-term (Nelson et al., 1997, p. 572).

Jennifer Merolla and her colleagues (2013) also tested the effects of issue framing, with respect to . They, however, went further to include equivalency frames into their study in a survey embedded in the 2007 Cooperative Congressional Elections Survey

(CCES), which was administered online to 2,188 respondents via YouGov (Merolla et al., 2013, p. 796). For the equivalency frame, Merolla and her colleagues tested the effects of placing the terms “illegal,” “undocumented,” or “unauthorized” before “immigrant” on public opinion. For example, Merolla and her colleagues asked respondents to what extent they agree or disagree on policies including legalization, the DREAM Act, and birthright citizenship (Merolla et al., 2013, p. 796). For each of the three questions, they varied the term that came before “immigrant.”

For the issue frame, they tested the effects of framing the story through an “amnesty” or a

“path to citizenship” lens (Merolla et al., 2013, p.796). They found, unlike in Kahneman and

Tversky’s Asian Disease Problem, that the effects of the equivalency frame had a small and 6 limited effect on public opinion, where the issue frame had a statistically significant effect

(Merolla et al., 2013, p.799). However, Merolla and her colleagues suggest this limited effect of the equivalency frame may have to do with the level of citizenship of the subjects in the story, a concept outlined in ’s “Semi-Citizenship in Democratic Politics” (Merolla et al.,

2013, p.791). Cohen explains that “undocumented persons” in the United States have neither autonomous rights - which include freedom of movement or the right to healthcare - nor relative rights - which include the right to vote and retain property (Cohen, 2009, p.67-72). Because of the low level of citizenship of unauthorized immigrants on both dimensions, “we might expect that attempts to change public opinion towards those immigrants may have limited effect, especially when compared to groups that rank low on only one dimension” (Merolla et al., 2013, p. 791).

While women are generally disadvantaged compared to men in the United States in terms of receiving equal pay and facing sexual harassment in the workplace, they do, as a population, have both autonomous and relative rights. Due to this citizenship status, readers may be more sympathetic to the subjects of the #MeToo stories, who are predominately American women, and the framing surrounding these stories may have a greater effect than in Merolla and her colleagues’ study.

The content analysis done for this thesis adds to the existing framing literature by examining equivalency frames that do not have an element of risk, such as in Kahneman and

Tversky’s Asian Disease Problem, and whose subjects generally have a higher or equal level of citizenship as the readership population, unlike in Merolla and her colleagues’ study on illegal immigration. 7 What are advocacy groups saying?

The National Sexual Violence Resource Center (NSVRC), the Women’s Media Center

(WMC), End Violence Against Women International (EVAWI), and Rape, Abuse, Incest

National Network (RAINN) have each published either a report or guidelines for how the media should report on sexual abuse, including what language not to use and why. The NSVRC, in a presentation by Kristen Houser, states that, “Historically, language used in coverage of [sexual abuse] has fueled victim distrust of the press and of the culture – this significantly impacts whether they tell others about what was done to them” (Houser, 2019, Slide 10). Similarly, End

Violence Against Women International believes “these recommendations for language use can improve our verbal and written communications as professionals in the field, helping us to provide information in ways that maximize our accuracy and clarity – and to avoid common tendencies that can create confusion, perpetuate misinformation, and contribute to a climate of doubt and victim blame” (Lonsway & Archambault, 2013, p. 2).

While these advocacy groups approach the topic from a standpoint of protecting the victim of sexual abuse, the underlying assumption is that the words used to describe acts of sexual abuse influence how much/if readers believe the stories, thus influencing the general culture around believing victims who choose to come forward. Together, these advocacy groups have identified six categories of terms and have provided recommendations for each. The following sections will outline these categories and their arguments for the recommended terms. 8 1. The victim of the sexual abuse (Ex: woman, accuser, victim)

End Violence Against Women International says the victim of the assault should be either named or referenced as a neutral identifier in relation to the perpetrator, such as

“girlfriend,” “neighbor,” etc., and the victim should choose whether or not to remain anonymous

(Lonsway & Archambault, 2013, p. 4). The National Sexual Violence Resource Center and End

Violence Against Women International recommend to avoid the term “accuser” as much as possible because “the word ‘accuser’ flips the power dynamic – referring to the victim as the

‘accuser’ turns the perpetrator into the victim of an accusation, and shifts the focus of the events onto the victim” (Houser, (n. d.), p. 2). However, the Women’s Media Center suggests there are some instances where “accuser” is acceptable. For example, the Women’s Media Center implies, although not explicitly, that “accuser” is only acceptable if the abuse is not yet proven (Ennis &

Wolfe, 2018, p. 11). End Violence Against Women International states that “’accuser’ … [is] not appropriate, unless [it] is used in a direct quote from another source and cited appropriately”

(Lonsway & Archambault, 2013, p. 4). For the terms “victim” and “survivor,” there is much less agreement among these advocacy organizations. The NSVRC does not explicitly make a case for or against “victim” or “survivor,” though it does overwhelmingly use “victim” in its examples for other categories, such as “the victim says” (Houser, (n. d.), p.1-2). EVAWI’s policy is to only use “victim” when in reference “to the context of the criminal justice system” and endorses the use of “survivor” (Lonsway & Archambault, 2013, p. 4). RAINN makes the distinction between

“victim” and “survivor” by defining “victim” in reference to “someone who has recently been affected by sexual violence, when discussing a particular crime, or when referring to aspects of the criminal justice system” and “survivor” in reference to “someone who has gone through the recovery process or when discussing the short- or long-term effects of sexual violence” (RAINN, 9 (n. d.)). Finally, the Women’s Media Center recommends “survivor” over “victim” because it is more “precise and empowering,” and “victim” should only be used if the person died (Ennis &

Wolfe, 2018, p. 14). Overwhelmingly, though, the general rule is to ask the victim what they prefer to be referred as (RAINN, (n. d.); Lonsway & Archambault, 2013).

2. The perpetrator of the abuse (Ex: the chief executive, last name, abuser)

The National Sexual Violence Resource Center says the abuser should be identified using

“people first language or non-charged words” such as “the father of four,” or “people who commit sexual offenses,” rather than heavily charged words such as “monster,” “predator,” etc. so as to avoid “an ‘us vs. them’ dynamic and make it difficult to recognize problematic sexual behavior in loved ones and associates” (Houser, (n. d.), p. 1). EVAWI makes the distinction on what to call the abuser based on whether it is referring to the criminal justice system. For example, within the criminal justice context, the term “perpetrator” will be used “only when a sexual assault conviction represents a final resolution of a case” (Lonsway & Archambault,

2013, p. 3). Outside of the criminal justice context, however, “the word ‘perpetrator’ will be used to refer in general terms to those who commit sexual offenses” (Lonsway & Archambault, 2013, p. 3). Beyond this term, EVAWI recommends to use the abusers name when referring to him/her

(Lonsway & Archambault, 2013, p. 3). The Women’s Media Center and RAINN do not state a preference for how the abuser is identified in the language guides being examined.

10 3. The victim’s account (Ex: said, accused, came forward)

The National Sexual Violence Resource Center believes that journalists should avoid implying responsibility and shame on the part of the victim, such as using the terms “confessed” or “admitted,” instead using neutral terms such as, “said,” “reported,” or “disclosed” as they are more objective (Houser, (n. d.), p. 2). Because “accuser” and “accused” have the same root word

– accuse – the NSVRC also recommends against using the word “accused” when describing how the victim came forward (Houser, (n. d.), p. 2). EVAWI only comments on the term “allege,” stating that as “almost all sexual assaults remain ‘unresolved’ by the legal system… it would be inappropriate to refer to all such reports (or even disclosures) of sexual assault as ‘alleged’”

(Lonsway & Archambault, 2013, p. 2). They go on to say that “this practice is not generally used for any type of crime other than sexual assault,” therefore it would be problematic to perpetuate the disbelief surrounding sexual assaults by continuing to use this term (Lonsway &

Archambault, 2013, p. 2). Neither RAINN nor WMC explicitly recommend which terms to use for this category, although WMC does endorse “alleged” when in reference to the events and the abuser if the abuse is not yet proven (Ennis & Wolfe, 2018, p. 10-11; RAINN (n. d.)).

4. The statement of the sexual abuse (Ex: accusation, claim, allegation)

End Violence Against Women International says the term “story” should be avoided due to the “connotation of skepticism conveyed,” and instead terms like “account,” “statement,” or

“description” should be used (Lonsway & Archambault, 2013, p. 5). The Women’s Media

Center notes that keeping the statement in the realm of hearsay – “accusation,” “allegation” – is appropriate if the abuse has yet to be proven (Ennis & Wolfe, 2018, p. 11). The National Sexual 11 Violence Recourse Center and RAINN do not note any preferences for this category in their guidelines (Houser, (n. d.); RAINN, (n. d.)).

5. The sexual abuse (Ex: sexual harassment, misconduct, abuse)

These terms should make clear that a crime occurred and should avoid using language that implies consent, such as “sex,” “relationship,” “sexual acts,” etc. because it “prevents the public from understanding that the act was one of violence and not a mutually consensual act”

(Houser, (n. d.), p. 1). All of the advocacy organizations examined for this thesis agree upon the general rule that journalists should avoid using language often used to describe consensual sex when portraying an act of sexual violence, including terms like “fondling,” “massaging,” or

“foreplay.” (Houser, (n. d.), p. 1; Lonsway & Archambault, 2013, p. 2; Ennis & Wolfe, 2018;

RAINN, (n. d.)). End Violence Against Women International recommends using objective language to describe the specific body parts and sexual acts involved instead of using a generic term; for example, saying “the defendant forced his penis into the victim’s vagina” instead of

“the defendant sexually assaulted the victim” (EVAWI, 2013, p. 3). RAINN and NSVRC even condemn the terms “date rape,” “acquaintance rape,” “legitimate rape,” and “forcible rape” because “these qualifiers and descriptors infer there are different kinds of rape, and none are as serious as ‘rape’” (Houser, (n. d.), p. 2; RAINN, (n. d.)). The NSVRC also recommends against using the term “nonconsensual sex” as it would be simpler and more accurate to say “rape”

(Houser, (n. d.), p.1). The NSVRC also recommends the media avoid labelling sexual abuse as a

“scandal” as this term could also be used in reference to a consensual affair or noncriminal events (Houser, (n. d.), p. 2). Finally, similar to “alleged” in reference to the disclosure of an 12 abuse, EVAWI says to avoid the term “alleged” because it instills doubt upon whether the event actually occurred (Lonsway & Archambault, 2013, p. 2). While “alleged” might be useful for journalists to avoid making a potentially false statement, the element of doubt can be avoided by simply attributing the statement to the victim using “said” (Fitts, 2013).

6. The action upon the victim (Ex: harassed, abused, raped)

Similar to other categories, the verb to describe the action upon the victim should include an element of force and should be careful to not portray the act as a consensual one (Houser, (n. d.), p. 1; Lonsway & Archambault, 2013, p. 2). For example, “had sex with victim” should be

“raped victim” because “had sex” leads the reader to believe there was no criminal behavior.

Other examples are using “strangled” instead of “choked,” “subjected to” instead of

“experienced,” and “molesting” instead of “fondling” (Lonsway & Archambault, 2013, p. 3, 5;

Houser, (n. d.), p. 2).

Table 1. Explicitly Recommended and Not Recommended Terms by Advocacy Organizations

National Sexual End Violence Women’s Media Rape, Abuse, Violence Resource Against Women Center (WMC) Incest National Center (NSVRC) International Network (RAINN) (EVAWI) Victim of the sexual Do: “Victim” Do: “Victim,” Do: “Survivor,” Do: “Victim” (if abuse “client,” “survivor,” “accuser” (when not recent), “survivor” Don’t: “Accuser” “neutral identifier,” yet proven) (if recovered) (“girlfriend,” “wife,” “daughter,” Don’t: “accuser” Don’t: N/A “neighbor”) when proven, “victim” (when still Don’t: “Alleged alive) victim,” “accuser” 13 Perpetrator of the Do: “People first Do: “Perpetrator” Do: N/A Do: N/A abuse language or non- when proven, “the charged words” name of the Don’t: N/A Don’t: N/A (Ex: “people who defendant” commit sexual offenses,” “man,” Don’t: N/A “father of four,” “good people,” “woman”

Don’t: “monsters,” “predators,” “evil” Victim’s account Do: N/A Do: “Account,” Do: “Allegation” Do: N/A “statement,” Don’t: N/A “description” Don’t: N/A Don’t: N/A

Don’t: “Story” Statement of the Do: “Reported,” Do: N/A Do: N/A Do: N/A abuse “said,” “disclosed,” “alleged” Don’t: N/A Don’t: N/A Don’t: N/A

Don’t: “Accused,” “admitted,” “confessed” Sexual abuse Do: “Rape,” Do: “language of Do: “Alleged rape,” Do: “Criminal act” “sexual assault,” the specific act” “Harassment” (use specific legal “forced using body parts description), penetration,” “force and element of Don’t: N/A “sexual violence” to perform,” “child forced coercion sexual abuse case,” Don’t: “Date rape,” “rape case,” “sexual Don’t: “alleged “acquaintance rape” abuse,” “sexually (act),” “sexual abusive intercourse,” “sex,” relationship,” “oral sex” “‘language of the charge,’” “unwanted sexual contact”

Don’t: “Sex,” “Intercourse,” “sex scandal,” “scandal,” “sexual relationship,” “nonconsensual sex,” “forcible 14 rape,” “child pornography” Action upon the Do: “Forced,” Do: “Subjected to,” Do: N/A Do: “Sexually victim pushed,” “held “strangled” assaulted,” down” Don’t: N/A “sexually abused” Don’t: “had (sex),” Don’t: “Engaged “fondling,” Don’t: N/A in,” “performing,” “massaging,” “fondle” “experienced,” “choked”

Evidence for these arguments

There is a lack of peer-reviewed research to suggest the validity of these arguments with respect to these specific terms or even other terms to describe events of sexual abuse. There is, however, some research to suggest that how women are framed in the media affects the behaviors of those who consume that media. For example, Silvia Galdi and her colleagues presented 141 men living in Northern Italy, both students and employed, one of three video clips

(Galdi et al., 2014). One video clip showed women being objectified in entertainment media, the second showed women in successful professional positions, and the third, which served as the control, was a nature documentary. Galdi and her colleagues coded for gender-harassing conduct by asking the men to participate in an online chat with a fictitious female partner after viewing one of the three clips. They gave the men the option to send pre-written jokes, which were presented in pairs – one sexually degrading toward women and one neutral (Galdi et al., 2014).

The men who viewed the objectifying media were more likely to send offensive jokes to the female partner in the online chat, thus behaving in a harassing manner. Then, to measure harassment proclivity, Galdi and her colleagues asked the men to rank which action they would take if they were in a successful position in which they would not face consequences for the 15 options chosen. The options were with respect to women in a subordinate position and had various degrees of objectification. Again, men exposed to objectifying media prior to ranking these hypothetical actions reported higher harassment proclivity than the men who watched the

Female Professionals and the control condition (Galdi et al., 2014).

Galdi and her colleagues’ study on the effects of objectifying media on men’s behaviors is relevant to this thesis because it suggests that how women are framed in the media has greater cultural consequences. Yet, it would be important to study these frames within the news media as this study was restricted to entertainment media. It also should be noted that the degree of objectification may vary between the two types of media in the sense that the news media may be more subtle in the ways it frames women and the sexual abuses they are subjected to. It would then be reasonable to expect that the effects of the framing would be subtle if the differences in framing is subtle to begin with.

What are journalism institutions saying?

Journalistic institutions are also weighing in on this debate, albeit to a lesser degree than the advocacy organizations. In response to the #MeToo movement and pressure from advocacy groups to change the discourse around sexual abuse, the Associated Press updated its Style

Guide in 2018. For example, the AP added a preference for the term “sexual misconduct” over

“sexual harassment,” because “’harassment’ is too mild to describe some of the activities that have been alleged in recent weeks,” according to AP vice president John Daniszewski (Perlman,

2018). He goes on to write that misconduct “encompasses a broader range of sexual misbehavior and does not run the risk of diminishing some of the alleged acts” (Perlman, 2018). The AP also 16 warned against using the words “victim” or “survivor” because they “label people, whereas the condition that affected them should bear the brunt” (Perlman, 2018). So, instead of writing

“sexual assault victims,” say “people who have been sexually assaulted.”

The Columbia Journalism Review published an article titled “The Right Way to Write

About Rape,” summarizing and adopting the recommendations of -based activist

Claudia Garcia-Rojas. Among the recommendations are a list of terms to avoid, such as

“alleged,” “admits,” “confesses,” “engaged in,” and “oral sex” for the same reasons listed by the advocacy groups already mentioned (Fitts, 2013).

The Poynter Institute, when writing about the new AP Style changes, added some recommendations of their own, going even further to condemn the use of the terms, “sexual misdeeds,” “ offenses,” “misbehavior,” “transgressions,” and “wrongdoing” because, similar to

“sexual harassment,” these synonyms fall short or seem stilted (Glover & Beard, 2017). They also recommended journalists stay away from the words, “scandal,” “sex,” “claimed,”

“proclaimed,” “reportedly,” and “shenanigans” (Glover & Beard, 2017). While these recommendations are not yet integrated into the AP Style Guide, the Poynter Institute acts as a resource for journalists, even providing courses for how to report in certain circumstances. Thus, it would be valuable to examine if the media adheres to these recommendations.

17 Chapter 3

Theory and Hypotheses

Frames are often analyzed on the basis of ideology. Liberals frame an issue one way, and conservatives frame it another. This can also be applied to media frames as some media outlets are loosely aligned with political parties in the United States. I use the term “loosely” because the media, mainstream or otherwise, rarely endorses one party over another. Yet, much of the reporting and especially the editorials, or commentary for the cable news media, tend to favor one party over the other, or at least one party’s policies over the other. Merolla and her colleagues (2013) explore ideological media framing and test the differences in frames between liberal, mainstream, and conservative media with respect to illegal immigration.1 They found the conservative media used the issue frame of “amnesty,” preferred by most Republicans, far more than the mainstream media. For example, and Washington Post’s mentions of amnesty were at 56 percent and 28 percent, respectively. Similarly, the

Post and New York Times mentions of amnesty were at 43 percent versus 23 percent respectively (Merolla, 2013, p. 792). This finding suggests that the mainstream, liberal, and conservative media use frames that align with the ideology of that outlet, thus using different frames that could have different effects on the people consuming that media.

1 Merolla and her colleagues labelled and Washington Post as mainstream media. While these newspapers are mainstream, they are comparatively more liberal than the conservative media tested – i.e. there is still a clear ideological difference. For this thesis, the New York Times and Washington Post will be labelled as liberal media as they tend to lean left and are significantly more liberal than the clearly conservative media.

18 Merolla and her colleagues’ theory is the basis of this thesis. I expect, because of the findings in their study, that liberal and conservative media will also use different frames for issues beyond illegal immigration, specifically on stories of sexual abuse. Victims of sexual violence are overwhelmingly women and girls – 82% of all juvenile victims and 90% of all adult rape victims are female (RAINN, NDL; Department of Justice, 2000). Because of this, sexual harassment and assault are often seen as a “women’s issues.” As there is a significant gender gap between the two major political parties in the United States, with women identifying as or lean

Democrat at 56 percent in 2017 versus just 44 percent of men, the issue of sexual violence has almost exclusively been adopted by the Democratic Party (Pew Research Center, 2018).

Additionally, it is no surprise that the sexual abuse advocacy organizations analyzed in this thesis would cater toward and support the Democratic Party. Therefore, my hypotheses for this thesis are as follows:

Hypothesis 1: The liberal media (the New York Times, , CNN, and

MSNBC) will use the equivalency frames recommended by these advocacy groups more

than the conservative media (the , the Washington Times, and ).

Also, simply because of the fact these advocacy groups have guidelines for how to report on sexual abuse, my second hypothesis is as follows:

Hypothesis 2: The media collectively will use not recommended equivalency frames more than it uses the recommended equivalency frames.

19 Chapter 4

Methodology and Operationalization

The methodology, like the theory, is based on Merolla and her colleagues’ study on media framing of illegal immigration. I decided to examine the same publications – the New

York Times, the New York Post, the Washington Post, the Washington Times, CNN, MSNBC, and Fox News. I agreed with Merolla and her colleagues’ decision to code these media outlets because they are national outlets with large readerships that have varying ideological tendencies.

The scope of my study, however, is smaller. Where Merolla and her colleagues analyzed five years of media coverage from 2007-2011, my timeframe spanned a month for each #MeToo event. I chose 10 events within the first year of the #MeToo movement in which well-known, powerful men were accused of sexual misconduct. These 10 events were, in chronological order, the accusations of sexual abuse against , Larry Nassar, , Roy

Moore, Louis C.K., , Al Franken, R. Kelly, Les Moonves, and Brett Kavanaugh.

While the events span the first year of the #MeToo movement2, I will only code articles within one month of each story breaking. For example, the New York Times broke the Harvey Weinstein sexual abuse story on October 5, 2017, so I coded the Harvey Weinstein stories between then and

November 5, 2017 for each of the seven media outlets. The timeline for each #MeToo event is located in the appendix on page 47 of this thesis.

2 Although Tarana Burke started the #MeToo hashtag in 2006, for the purposes of this thesis, I mark the beginning of the #MeToo movement as October 5, 2017, the day Harvey Weinstein was accused of sexual misconduct in the New York Times and New Yorker, due to its popularity and transition into a full-fledge movement. 20 Then, I randomly selected two stories within the first month of each #MeToo event breaking for each publication3 from the Nexis Uni database, resulting in a total of 122 stories.

Because some #MeToo events broke within the same week or even day, some publications combined these breaking 4stories into one about the #MeToo movement generally. I did not code these stories more than once even though they mentioned multiple events. This could mean that there are fewer stories for some publications, which contributes to a total number of stories smaller than 140 (7 media outlets x 10 #MeToo events x 2 stories). Also contributing to a smaller about of articles is that some publications did not write about some of the #MeToo events at all.

For example, MSNBC did not report at all on Larry Nassar, R. Kelly, or Les Moonves within the timeframe examined. Although I cannot know exactly why MSNBC did not report on these

#MeToo events, I can assume that because MSNBC is the cable affiliate of NBC Universal News

– where the other cable networks analyzed do not have broadcast news affiliates – they did not report on the breaking news about nonpolitical figures because they specialize in strictly political stories. Again, this is speculation, but it could be a reason as to why MSNBC has few stories than the other news outlets.

I excluded editorials and opinion articles, stories related to the perpetrator but that did not include language of the abuse, and stories published but not actually written by the publications analyzed, such as the Washington Times publishing an Associated Press articles. I excluded these because it is necessary that the language for the story be written by someone at that

3 Originally, I decided to select three articles within a month of each #MeToo event breaking for each publication, but after determining the results did not change significantly between 2 and 3 articles, I decided there was not enough value added for the amount of time and effort of coding the additional stories.

21 publication or chosen by a reporter or anchor, such as a quote or guest on air, as I am analyzing a difference in frames by ideology of the publication.

The following section will outline my coding operationalization for each of the six categories of terms identified by the advocacy organizations analyzed in this thesis.

Operationalization for Each Category of Terms

While coding the 122 stories, I followed some general coding rules that applied to multiple term categories. The first is that I only coded the headlines and body copy of the stories.

Photo captions and embedded links were not coded because they essentially repeated what was already included in the headline/story. The second is that quotes and attributions were coded just like the rest of the article. Even though the content of the quote and the terms used were not written or spoken by the reporter/publication, there was still the element of control to include the quote in the story. Third, I coded for both real and hypothetical scenarios as well as “not” terms.

For example, if a story said, “He did not rape the victim,” I still coded the term “rape” because they are using recommended language as opposed to saying, “He did not engage in nonconsensual sex with the victim,” which is not recommended. Finally, I only coded terms in reference to problematic sexual behavior so as to ensure my data was relevant to what the advocacy organizations made recommendations about. For example, every time “kissing” is coded in my dataset, it is in reference to nonconsensual behavior, such as in stories about Roy

Moore. On the contrary, this reference of kissing was not coded because consent was implied:

“Her only sexual experiences had been kissing boys her own age.” That way, when I label this term as not recommended, it is because it does not mention the presence of abuse that existed. As 22 these stories were all about the #MeToo movement, there were very few instances in which actual consensual sex or sexual activity was referenced. The following sections include the coding rules for each term category.

1. The victim of the sexual abuse (Ex: woman, accuser, Ms. Last name) – I coded terms only

in reference to the person/people coming forward with accusations. For example, the term

“women” may appear in an article 10 times, but only five may be in reference to women

who have come forward with accounts of sexual harassment or abuse. I also coded

pronouns in reference to the victim. However, I only coded the pronoun if it could be

logically replaced by the victim’s name or another identifier. For example, if the

sentence, “She said she was 18 when the abuse started,” appeared in an article, I only

counted the first pronoun because that was the instance in which the reporter had a choice

to use a different term. In some cases, another person came forward on behalf of the

victim, such as a parent of a minor. In these cases, I coded the person making the initial

accusation as the accuser. This does not include instances where close friend or

coworkers corroborated accusations. Overwhelmingly, though, the accuser was also the

victim.

2. The perpetrator of the abuse (Ex: the chief executive, Mr. last name, the abuser) –

President Trump is mentioned in many of the articles, but I only coded his name and any

references to him if the article mentioned the accusations of sexual harassment and abuse

against him specifically. For example, Brett Kavanaugh was often referred to as the

president’s nominee, but there would be no mention of Trump’s alleged abuse, so I did

not code his name in those instances. 23 3. The victim’s account (Ex: said, accused, came forward) – For this category, I coded only

the verbs portraying coming forward in which the accuser is the subject. For example, if

the perpetrator made a statement “accusing” the victim of lying, the term “accused”

would not be counted because it was not in reference to the accuser, and the advocacy

organizations focus on the credibility of the victim/accuser rather than the perpetrator.

Occasionally, a reporter will write, “The letter says” or “the article detailed,” in reference

to the accuser’s statement. In these instances, I coded “says” and “detailed” because the

statements originated from the accuser.

4. The statement of the sexual abuse (Ex: accusation, claim, allegation) – Similar to the

previous category, I coded only the statement of the sexual abuse by the accuser. For

example, if the alleged perpetrator made a “claim” that the victim was lying, it would not

be coded because it is not in reference to the statement of sexual abuse from the victim. If

there was a modifier before the noun, such as “sexual harassment claim,” I coded this as

one term instead of splitting “sexual harassment” and “claim” into separate categories.

5. The sexual abuse (Ex: sexual harassment, misconduct, abuse) – As outlined in the

general rules for operationalization above, I coded terms for this category only if they

were in reference to problematic sexual behavior. Therefore, “relationship” and “sex” are

included in this when referring to real or perceived acts of sexual abuse. When reporters

modified the sexual abuse with the term “alleged,” such as in “the alleged incident,” I

coded this as one term instead of separating “alleged” into the third category. I made this

decision based on End Violence Against Women International’s explicit condemnation of

inserting “alleged” before the act of sexual abuse. 24 6. The action upon the victim (Ex: harassed, abused, raped) – For this category, I coded

terms only in reference to verbs acted upon by the abuser to the victim. This included

physical, verbal, and emotional actions, such as retaliation, career pressure, and insults as

all of these are considered abusive actions.

Coding for theses six categories resulted in a data set with 190 row (one for each story and rows to calculate totals for groups of stories) and 1409 columns. Each column represented a term used in one category in addition to columns for category totals and three columns at the beginning of the dataset to identify the source, target, and headline of the article. Each cell recorded the number of times each term was mentioned in the story.

Operationalization of Term Labelling

After dividing terms into the categories above, I then applied the recommendations of each advocacy organization to each term. I analyzed each of the organizations’ guides on language use when referencing sexual violence and created a collective document to show which terms were unanimously recommended or not recommended and which were split. Based on these recommendations, I created three categories for labelling terms: explicitly recommended, explicitly not recommended, and not sure due to conflicting messages. These categories, however, did not account for all of the terms I coded. In fact, they only accounted for about 47% of the terms. Therefore, I made additional categories based on my subjective interpretation of how the advocacy organizations would likely label these terms based on their language guides.

These categories are: probably recommended, probably not recommended, and not sure as they could go either way. 25 For example, because the Women’s Media Center provided a caveat for “accuser” and the others collectively agreed that it should be avoided, “accuser” was labelled as “Not Sure c-

Conflicting Messages.” As for the subjective labels, an example is the term “complaint” categorized under the second category for the statement of the abuse. “Complaint” is not explicitly mentioned by any of the four advocacy organizations, but because “complaint” has a negative connotation and had the potential to downplay what the victim is “complaining” about,

I labelled this term as “Probably Not Recommended” as this is the likeliest categorization of this term after analyzing the advocacy organizations’ guidelines.

After executing this methodology, I had a final dataset consisting of 190 rows and 1409 columns. The keys columns were the counts of recommended or not recommended terms used in each story. The final data set is based on a total of 14,155 terms appearing in 122 stories, each coded on the six-point recommended to not recommended scale (including the “uncodable,”

“Not Sure” labels). To see the number of terms by category, see the left-hand column of table 4 on page 32.

26 Chapter 5

Results and Analysis

My first hypothesis anticipated that the liberal media – the New York Times, the

Washington Post, CNN, and MSNBC – would use the equivalency frames recommended by the

National Sexual Violence Resource Center, End Violence Against Women International, the

Women’s Media Center, and the Rape, Abuse and Incest National Network (RAINN) more often than the conservative media – the New York Post, the Washington Times, and Fox News. Figure

1 shows the percent of explicitly and probably recommended terms by the advocacy organizations, both by ideology and as a whole. The darker shading represents the explicitly recommended and not recommended terms, while the lighter shading represents the probably recommended and not recommended terms.

Figure 1. Percentage of Codable Explicitly/Probably Recommended Terms by Advocacy Organizations, by Ideology

Liberal Media

Conservative Media

All Media

0.0% 10.0% 20.0% 30.0% 40.0% 50.0% 60.0% 70.0% 80.0% 90.0% 100.0%

Probably Not Recommended Probably Recommended Not Recommended Recommended

Note: This chart displays only codable terms and excludes terms that advocacy organizations provided conflicting messages for or that could not be reliably coded as probably recommended or probably not recommended. Therefore, totals in the chart equal 100.

27 Table 2. Percentage of All Explicitly/Probably Recommended Terms by Advocacy Organizations, by Ideology

Not Sure - Not Sure - Probably Probably Not Not Recommended Conflicting Could go Recommended Recommended Recommended messages either way Liberal Media 35.4% 40.3% 7.8% 6.7% 5.9% 3.9% n = 9,814 Conservative Media 33.2% 38.5% 7.6% 8.5% 7.4% 4.9% n = 4,341 All Media 34.8% 39.7% 7.7% 7.2% 6.4% 4.2% n = 14,155

Figure 1 includes only the codable terms, meaning that the “Not Sure - Conflicting

Messages” and “Not Sure - could be labelled either as probably recommended or probably not recommended” labels are not included. After excluding these categories due to their lack of interpretation, I took the percentages out of the new total. Therefore, the totals in Figure 1 equal

100. The liberal media used not recommended terms 11.5% of the time compared to the conservative media who used not recommended terms 14.7% of the time, suggesting my first hypothesis is correct. After conducting a 95% confidence test for the codable terms shown in

Table 2 (liberal media: n=8,391, conservative media: n=3,646), the standard error for the liberal media was .7%, producing a range of 10.8% to 12.2%. The standard error for the conservative media was 1.1%, producing a range of 13.6% to 15.8%. These ranges are illustrated in Figure 2.

I performed the same 95% confidence test, shown in Figure 2 for the percentages shown in Table 1, which includes the “Not Sure” categories and shows the distribution of all the terms coded. The liberal media, whose n is now 9,814 when including the “Not Sure” category, used not recommended terms 9.8% of the time with a standard error of .06%, producing a range of

9.2% to 10.4%. The conservative media, with a new n of 4,341, used not recommended term 28 Figure 2. 95% Confidence Test Results for Percentages of Explicitly and Probably Not Recommended Terms by Advocacy Organizations, by Ideology

12.3% of the time with a standard error or 1.0%, producing a range of 11.3% to 13.3%. Again, these ranges to do not overlap. Therefore, I can say with confidence that Hypothesis 1 is correct.

My second hypothesis anticipated that the media would collectively use not recommended terms more often than using recommended terms. This hypothesis proved incorrect as the media as a whole used not recommended terms 12.4% of the time compared to

87.6% using recommended terms (of codable terms). When including the “Not Sure” categories, those percentages decreases to 10.6% and 74.5%, respectively. Still, the results for this hypothesis are significant.

Removing the Subjective Labels

The terms explicitly recommended or not recommended by the advocacy organizations only made up 46.7% of the total terms. Still, I feel it is valuable to analyze the distributions within only the explicitly recommended or not recommended terms as it eliminates any subjectivity. Figure 3 and corresponding Table 3 show the distribution of explicitly

29 Figure 3. Percentage of Codable Explicitly Recommended Terms by Advocacy Organizations, by Ideology

Liberal Media

Conservative Media

All Media

0.0% 10.0% 20.0% 30.0% 40.0% 50.0% 60.0% 70.0% 80.0% 90.0% 100.0% Not Recommended Recommended

Note: This chart displays only codable terms and excludes terms that advocacy organizations provided conflicting messages for. Therefore, totals in the chart equal 100.

recommended and not recommended. Similar to Figure 1 and Table1 2, the figure includes only codable terms, while the table incorporates the “Not Sure – Conflicting Messages” category.

Therefore, the totals in both tables equal 100.

Figure 3 displays similar trends as when incorporating probably recommended terms.

However, the differences between liberal and conservative media aren’t as large. So, I performed another 95% confidence test with only explicitly recommended and not recommended terms. For only codable terms, shown in Figure 3, the liberal media used not recommended terms 10% of the time with an n of 3,967. With a standard error of 0.9%, this produced a range of 9.1% to

10.9%.

The conservative media, on the other hand, used not recommended terms 12.8% of the time with an n of 1,654 and a standard error of 1.6%. This produced a range of 11.2% to 14.4%.

Still, even with eliminating probably recommended terms, the ranges of the liberal and conservative media do not overlap. However, the gap between the ranges is much smaller than 30 Table 3. Percentage of All Explicitly Recommended Terms by Advocacy Organizations, by Ideology

Recommended Not Sure - Conflicting Messages Not Recommended Liberal Media 75.2% 16.5% 8.3% n = 4,632 Conservative Media 72.7% 16.6% 10.7% n = 1,979 All Media 74.4% 16.5% 9.0% n = 6,610

when incorporating probably recommended terms, 0.3% and 1.6% respectively. So, when only analyzing explicitly recommended terms, I can say with 95% confidence that the conservative media uses not recommended terms more than the liberal media, but the conservative media uses not recommended terms at a higher rate when incorporating probably recommended terms.

When, performing the 95% confidence test on all explicitly recommended terms, organized in Table 3, the liberal media, which used not recommended terms 8.3% of the time, with an n of 4,633 and a standard error of 0.8%, produces a range of 7.5% to 9.1%. Meanwhile, the conservative media, which used not recommended terms 10.7% of the time, with an n of

1,979 and a standard error of 1.4%, produces a range of 9.3% to 12.1%. Again, these ranges shown in Figure 4 do not overlap, but they are closer to each other than when incorporating probably recommended terms.

Figure 4. 95% Confidence Test Results for Percentages of Explicitly Not Recommended Terms by Advocacy Organizations, by Ideology

31 Which categories performed better than others?

Though I did not provide a hypothesis for the performance of specific term categories, I feel it is valuable to analyze where the media generally performs well or poorly according to the advocacy organizations.

Figure 5. Percentage of Codable Explicitly/Probably Recommended Terms by Advocacy Organizations, by Category of Term

1 - The victim of the sexual abuse (Ex: woman, accuser, victim)

2 - The perpetrator of the abuse (Ex: chief executive, Last name, abuser)

3 - The victim’s account (Ex: said, accused, came forward)

4 - The statement of the sexual abuse (Ex: accusation, claim, allegation)

5 - The sexual abuse (Ex: sexual harassment, misconduct, abuse)

6 - The action upon the victim (Ex: harassed, abused, raped)

TOTAL FOR ALL MEDIA

0.0% 20.0% 40.0% 60.0% 80.0% 100.0% Probably Not Recommended Probably Recommended Not Recommended Recommended

Note: This chart displays only codable terms and excludes terms that advocacy organizations provided conflicting messages for or that could not be reliably coded as probably recommended or probably not recommended. Therefore, totals in the chart equal 100.

Figure 5 and corresponding Table 4 expand upon Figure 1 and Table 2’s “All Media” portion by breaking it down into category of term. Categories 1 and 2 (terms for the victim and the perpetrator of the sexual abuse) perform very well with close to 100% of all codable terms being either explicitly or probably recommended. Categories 4 and 5 (terms for the statement of the sexual abuse – “accusation,” “claim,” “allegation” – and the sexual abuse – “sexual harassment,” “misconduct,” “abuse”), however, perform quite poorly. For example, Category 4 32 (terms for the statement of the sexual abuse) performed the worst of all categories with 85.3% of the codable terms being not recommended. This could be because this category had the largest portion of uncodable terms at 65%. For example, “accusation” and “allegation” were coded as

“Not Sure due to Conflicting Messages.” These two terms alone, without a modifier in front such as “sexual assault allegation,” accounted for 51.6% of the terms for this category. When modifiers are included, this number jumps to 65%. Therefore, the codable n for Category 4 is

497 compared to its original n of 1421. This should be remembered when analyzing the performance of this category.

Table 4. Percentage of All Explicitly/Probably Recommended Terms by Advocacy Organizations, by Category of Term

Not Sure - Not sure - Probably Probably Not Not Recommended Conflicting Could go Recommended Recommended Recommended Messages either way 1 - the victim of the 25.3% 65.3% 5.6% 3.2% 0.6% 0.0% sexual abuse, n = 3,712 2 - the perpetrator of the 57.3% 37.0% 0.1% 3.4% 1.7% 0.4% sexual abuse, n = 5,249 3 - the victim's account, 37.0% 24.4% 2.9% 19.6% 4.0% 12.2% n = 1,741 4 - the statement of the 3.9% 1.2% 51.6% 13.4% 13.3% 16.6% sexual abuse, n = 1,421 5 - the sexual abuse, n = 12.8% 35.4% 6.0% 7.3% 30.4% 8.1% 1,361 6 - the action upon the 14.3% 49.9% 2.4% 13.7% 17.0% 2.7% victim, n = 671 TOTAL 34.8% 39.7% 7.7% 7.2% 6.4% 4.2% n = 14,155

Category 5 (terms for the sexual abuse) performs the worst when analyzing all terms, including the “Not Sure” categories, and second worst when analyzing only the codable terms.

When analyzing all terms, 38.5% of Category 5’s terms are not recommended compared to

29.9% of Category 4’s. The most significant criteria, according to the advocacy organizations, for how to label the sexual abuse was based on whether or not consensual language was being used to describe acts of sexual violence. It was unanimously agreed upon that consensual 33 language is not recommended as it normalizes sexual violence and blurs the line between consensual sexual acts and criminal sexual behavior. Figure 5 and Table 4 show that the media organizations analyzed collectively in this content analysis do a poor job of distinguishing being consensual acts and criminal acts in the language they use.

In fact, when focusing only on Categories 3-6, the percent of not recommended terms rises from 10.4% to 25.4% of all terms (shown in Table 3), and the percent of recommended terms falls from 74.5% to 43%. This suggests that aside from how the media, as a whole, labels the individuals on either end of the sexual abuse, it doesn’t do a great job of using recommended language around sexual violence, or at least as great a job as Figure 1 and Table 2 imply. Yet, the media still uses recommended terms more than not recommended terms. However, when removing Categories 1 and 2, the media does not use recommended terms a majority of the time when taking into account the uncodable or “Not Sure” labels.

Incorporating the Recommendations of Journalistic Institutions

As journalistic institutions, such as The Associated Press, the Columbia Journalism

Review, and the Poynter Institute, have weighed in on the debate over which terms are best to use in stories on sexual assault, I felt it was appropriate to incorporate their recommendations and analyze how much these recommendations influence the distribution of recommended and not recommended terms. Figure 6 shows the distribution of explicitly and probably recommended terms by both advocacy organizations and journalistic institutions by ideology. 34 Essentially, this table is the same as Figure 1, only Figure 6 includes the recommendations of journalistic institutions.

Table 5, then, compares Figures 1 and 6 to show the differences in the distribution of term labels between only advocacy organizations and both advocacy organizations and

Figure 6. Percentage of Codable Explicitly/Probably Recommended Terms by Advocacy Organizations and Journalistic Institutions, by Ideology

Percentage of Explicitly/Probably Recommended Terms by Advocacy Organizations and Journalistic Institutions, by Ideology

Liberal Media

Conservative Media

All Media

0.0% 10.0% 20.0% 30.0% 40.0% 50.0% 60.0% 70.0% 80.0% 90.0% 100.0%

Probably Not Recommended Probably Recommended Not Recommended Recommended

Note: This chart displays only codable terms and excludes terms that advocacy organizations provided conflicting messages for or that could not be reliably coded as probably recommended or probably not recommended. Therefore, totals in the chart equal 100.

journalistic institutions. The percentages including the recommendations of the journalistic institutions are shaded. The percentages of the “probably” labels each go down by anywhere between 0.1% and 2.0%. This is most likely due to the fact that there are more explicit recommendations, which leaves fewer instances in which I had to subjectively label terms based on the existing recommendations. The percentages of the “Not Sure” both increase by about a percentage point or less. This is probably due to the fact that there is a greater opportunity for conflicting messages with the more recommendations there are. Also, there is a greater chance 35 Table 5 . Percentage of All Probably/Explicitly Recommended Terms for Only Advocacy Organizations and Both Advocacy Organizations and Journalistic Institutions, by Ideology

Probably Not Sure - Not Sure - Could Probably Not Media, by Ideology Recommended Not Recommended Recommended Conflicting messages go either way Recommended Liberal Media (n = 9,814) Advocacy Organizations' Recommendations 35.4% 40.3% 7.8% 6.7% 5.9% 3.9% " including Journalistic Institutions 35.9% +0.5% 38.4% -1.9% 8.9% +1.1% 7.1% +0.4% 5.8% -0.1% 4.1% +0.2% Conservative Media (n = 4,341) Advocacy Organizations' Recommendations 33.2% 38.5% 7.6% 8.5% 7.4% 4.9% " including Journalistic Institutions 33.6% +0.4% 36.5% -2.0% 8.6% +1.0% 8.9% +0.4% 6.7% -0.7% 5.7% +0.8% All Media (n = 14,155) Advocacy Organizations' Recommendations 34.8% 39.7% 7.7% 7.2% 6.4% 4.2% " including Journalistic Institutions 35.2% +0.4% 37.8% -1.9% 8.8% +1.1% 7.6% +0.4% 6.0% -0.4% 4.6% +0.4%

that I wouldn’t be able to confidently label terms as probably recommended and probably not

recommended because there were more recommendations to consider.

Generally, when including the recommendations of journalistic institutions, there is a

higher proportion of explicitly not recommended terms. Interestingly, there is also a higher

proportion of explicitly recommended terms and lower proportions of probably

recommended/not recommended terms. One explanation for the higher proportion of probably

recommended terms among advocacy organizations and (slightly) lower proportions of explicitly

recommended and “Not Sure – Could go either way” is that journalistic institutions, namely the

Associated Press, preferred “sexual misconduct” over “sexual harassment” when speaking in

general terms because “sexual misconduct” covers more under its umbrella of behavior. This

caused “sexual misconduct” to move from “Probably Recommended” (because no advocacy

group explicitly recommended this term) to “Explicitly Recommended” and “sexual harassment”

to move from “Probably Recommended” to “Not Sure – Conflicting messages.” As for the

higher proportion of explicitly not recommended terms when incorporating journalistic

institutions, this is likely because of the Poynter Institute’s recommendation to stay away from

the terms “sexual misdeeds,” “sexual offenses,” “sexual misbehavior,” “sexual transgressions,” 36 and “sexual wrongdoing.” Instead, they suggested using the all-encompassing term “sexual misconduct.”

When considering the effect of incorporating journalistic institutions had on the distributions by ideology, the effect is almost exactly the same. The liberal media’s use of recommended terms dropped by 1.4% when including journalistic institution recommendations, while the conservative media’s dropped by 1.6%. Both the liberal and conservative media’s percent of not recommended terms increased by 0.1%. So, the liberal media slightly benefitted from incorporating the recommendations of journalistic institutions, but it is well within the margin of error.

Overall, these differences between the distribution of terms for only advocacy organizations and both advocacy organizations and journalistic institutions don’t ever vary more than 2%. Often, the differences are much smaller. Because these differences are not significant, I will here on out analyze data only including the recommendations of advocacy organizations.

Breakdown by Publication

In order to identify any potential outliers, I also broke the data down by publication. Figure 7 shows each publication in order from best performing to worst performing when analyzing the codable terms, including the subjective “Probably” labels. The New York

Times performed the best in this figure with the Washington Times performing the worst – i.e. using not recommended terms the most and recommended terms the least. Overall though, despite publication, the media’s use of recommended terms ranged from 83.4% - 90.1%, and the 37 Figure 7. Percentage of Codable Explicitly/Probably Recommended Terms by Advocacy Organizations, by Publication

New York Times

Washington Post

MSNBC

CNN

Fox News

New York Post

Washington Times

0.0% 10.0% 20.0% 30.0% 40.0% 50.0% 60.0% 70.0% 80.0% 90.0% 100.0%

Probably Not Recommended Probably Recommended Not Recommended Recommended Note: This chart displays only codable terms and excludes terms that advocacy organizations provided conflicting messages for or that could not be reliably coded as probably recommended or probably not recommended. Therefore, totals in the chart equal 100.

use of not recommended terms ranged from 9.9% to 16.6%. This is only a 6.7% difference between the best performing and the worst performing.

When looking at only the explicitly recommended or not recommended of the codable terms in Figure 7 (the portion of the bars shaded darker), the order of best performing changes.

Now, MSNBC has the largest portion of recommended terms at 46.4% of codable terms, followed by the Washington Post at 44%, Fox News at 43%, CNN at 38.6%, New York Post at

37.1%, Washington Times at 36.1%, and the New York Times at 36%. Interestingly, the

38 Table 6. Percentage of All Explicitly/Probably Recommended Terms by Advocacy Organizations, by Publication

Not Sure - Not Sure - Probably Probably Not Not Recommended Conflicting Could go either Recommended Recommended Recommended messages way New York Times 31.5% 47.3% 6.1% 6.4% 5.0% 3.6% n = 2,445 Washington Post 38.7% 38.8% 6.0% 6.2% 5.8% 4.5% n = 2,888 MSNBC 38.9% 35.3% 9.7% 6.5% 5.9% 3.8% n =2,400 CNN 31.8% 39.9% 10.0% 7.8% 6.8% 3.8% n = 2,085 Fox News 34.9% 35.0% 8.7% 10.3% 7.2% 3.9% n = 2,103 New York Post 32.4% 42.7% 6.3% 6.2% 7.3% 5.1% n = 1,404 Washington Times 30.8% 40.4% 6.9% 7.7% 7.1% 7.1% n = 830

New York Times went from best performing to worst performing in terms of its use of recommended terms when eliminating the probably recommended labels. Fox News also surpassed CNN and the New York Times when eliminating the probably recommended labels.

Fox News, however, also has the highest proportion of uncodable terms, shown in Table 5, at

19%, meaning that their proportions of recommended and not recommended terms shown in

Figure 7 are inflated compared to the other publications as all totals equal 100 and Fox News had the biggest gap to fill. When focusing only on use of explicitly not recommended terms, the order is generally the same as in Figure 7 with the exception of the Washington Post dropping from second to fifth in the ranking of best performing.

Now, if we include the uncodable terms shown in Table 6 – “Not Sure – Conflicting

Messages” and “Not Sure – Could Go Either Way” – the order of best performing in terms of the highest proportion of recommended terms is New York Times at 78.9%, Washington Post at 39 77.5%, New York Post at 75.1%, MSNBC at 74.1%, CNN at 71.6%, Washington Times at

71.2%, and Fox News at 69.9%. Generally, the liberal media had a higher proportion of recommended terms for all terms coded with the New York Post being an outlier and placing third while the other two conservative media – Washington Times and Fox News – placed last. In terms of not recommended terms for all terms coded, the divide was along strict ideological lines, with the New York Times performing the best with the least amount of not recommended terms at 8.6%, followed by MSNBC at 9.7%, Washington Post at 10.4%, CNN at 10.6%, Fox

News at 11.2%, New York Post at 12.4%, and Washington Times at 14.2%. Overall, though, there weren’t any significant outliers when dividing the data by publication.

Analyzing “Accuser”

The term “accuser” is one of the most debated terms within the discussion on sexual violence language use. It is also one of the most common ways for victims to be referred to in news articles. “Accuser” was the fifth most commonly used term within the category of how the victim is identified for a total of 137 times. The terms more common than “accuser” were

“woman,” “she/her/hers,” “Full name,” and “Ms. Last name.” Most sexual violence advocacy groups stand against the use of this term as it often places the abuser in the position of a victim.

Within the advocacy groups analyzed in this content analysis, the National Sexual Violence

Resource Center, End Violence Against Women, and RAINN all stand against this term. The

Women’s Media Center, however, approves of this term under the condition that the case hasn’t been proven and the story is still within the realm of hearsay. Because of the Women’s Media

Center’s caveat, “accuser” was label as “Not Sure – Conflicting Messages.” Due to the fact, 40 Table 7. Percentage of "Not Recommended" Terms if "Accuser" was Unanimously Labelled as "Not Recommended" or All Terms and Codable Terms, by Ideology

Not Recommended Not Recommended Media, by Ideology (of all terms) (of codable terms) Liberal Media (n = 9,814) Category for the victim of the sexual abuse (Ex: woman, accuser, victim) 0.0% 0.0% " if "accuser" was not recommended 3.4% +3.4% 3.6% +3.6% Total 3.9% 4.6% " if "accuser" was not recommended 4.9% +1.0% 5.6% +1.0% Conservative Media (n = 4,341) Category for the victim of the sexual abuse (Ex: woman, accuser, victim) 0.0% 0.0% " if "accuser" was not recommended 4.3% +4.3% 4.6% +4.6% Total 4.9% 5.8% " if "accuser" was not recommended 5.9% +1.0% 7.0% +1.2% All Media (n = 14,155) Category for the victim of the sexual abuse (Ex: woman, accuser, victim) 0.0% 0.0% " if "accuser" was not recommended 3.7% +3.7% 3.9% +3.9% Total 4.2% 5.0% " if "accuser" was not recommended 5.2% +1.0% 6.0% +1.0%

however, that “accuser” is so controversial, so commonly used, and only conditionally endorsed by one of the four advocacy groups, I feel it is useful to analyze the distribution of terms if

“accuser” was “Explicitly Not Recommended.”

Table 7 shows the media’s percentage of “Not Recommended” terms rose by 1 point for both all terms and codable term. Though not shown in the table, the percentage of “Not Sure –

Conflicting Messages” decreased by the same amount as “accuser” was relabeled from that category. While this does not initially seem like a significant amount, considering that the n for all media is over 14,000 and a single term, which happens to be a controversial one, influenced the distribution of terms by a percentage point is telling. The relabeling of “accuser” affected the liberal and conservative media the same, by 1 percent, for all terms and only slightly affected the conservative media more with a difference of 0.2% for codable terms. 41 Table 8. Percentage of "Not Recommended" Terms if "Accuser" was Unanimously Labelled as "Not Recommended" for All Terms and Codable Terms, by Publication

Not Recommended Not Recommended Media, by Publication (of all terms) (of codable terms) New York Times (n = 2,444) Category for the victim of the sexual abuse (Ex: woman, accuser, victim) 0.0% 0.0% " if "accuser" was not recommended 1.3% +1.3% 1.3% +1.3% Total 3.6% 4.1% " if "accuser" was not recommended 4.0% +0.4% 4.6% +0.5% Washington Post (n = 2,888) Category for the victim of the sexual abuse (Ex: woman, accuser, victim) 0.0% 0.0% " if "accuser" was not recommended 2.1% +2.1% 2.2% +2.2% Total 4.5% 5.2% " if "accuser" was not recommended 5.1% +0.6% 5.8% +0.6% MSNBC (n =2,400) Category for the victim of the sexual abuse (Ex: woman, accuser, victim) 0.0% 0.0% " if "accuser" was not recommended 5.8% +5.8% 6.0% +6.0% Total 3.8% 4.5% " if "accuser" was not recommended 5.1% +1.3% 6.0% +1.5% CNN (n = 2,085) Category for the victim of the sexual abuse (Ex: woman, accuser, victim) 0.0% 0.0% " if "accuser" was not recommended 6.3% +6.3% 6.7% +6.7% Total 3.8% 4.6% " if "accuser" was not recommended 5.3% +1.5% 6.4% +1.8% Fox News (n = 2,103) Category for the victim of the sexual abuse (Ex: woman, accuser, victim) 0.0% 0.0% " if "accuser" was not recommended 5.5% +5.5% 6.1% +6.1% Total 3.9% 4.9% " if "accuser" was not recommended 5.2% +1.3% 6.4% +1.5% New York Post (n = 1,404) Category for the victim of the sexual abuse (Ex: woman, accuser, victim) 0.0% 0.0% " if "accuser" was not recommended 3.3% +3.3% 3.5% +3.5% Total 5.1% 5.8% " if "accuser" was not recommended 6.0% +0.9% 6.8% +1.0% Washington Times (n = 830) Category for the victim of the sexual abuse (Ex: woman, accuser, victim) 0.0% 0.0% " if "accuser" was not recommended 3.3% +3.3% 3.4% +3.4% Total 7.1% 8.3% " if "accuser" was not recommended 7.8% +0.7% 9.1% +0.8%

42 Like Table 7, Table 8 shows the percentage of “Not Recommended” terms after relabeling “accuser” for both codable terms and all terms. Instead of breaking it down by ideology, however, Table 8 breaks the data down by publication. Table 7 showed that there wasn’t a significant difference in the liberal and conservatives use of “accuser.” Table 8, however, shows there is a difference between how often the print media and cable media use

“accuser.” The New York Times performed the best of all media with its total of “Not

Recommended” terms increasing by 0.4% and 0.5% for all terms and codable terms, respectively. Then, the Washington Post with 0.6% for all terms and codable terms, the

Washington Times with 0.7% and 0.8% for all terms and codable terms, and the New York Post with 0.9% and 1.0% for all terms and codable terms. So, within the print media, the conservative media performed slightly poorer than the liberal media. Within the cable networks, CNN performed the worst with an increase of 1.5% and 1.8% for all terms and codable terms, respectively. Interestingly, MSNBC and Fox News increased by the same amount at 1.3% and

1.5% for all terms and codable terms, respectively. One explanation for why the cable networks might perform worse than the print media with their use of “accuser” is that there is less of an editing process within the cable media, whereas every word is edited multiple times, at least in theory, for the print media. For example, in a New York Times article titled, “How We Describe

Sexual Assault,” Stephen Hiltner writes:

At The New York Times, articles are rigorously reported then dissected by desk editors

before publication. And major investigative stories are scrutinized to an even greater

degree. For those stories, not only do the desk’s editors probe the work, but masthead

editors — the most senior journalists at The Times — also weigh in, as do members of 43 our legal team. And especially careful consideration is given to the language used to

describe potentially criminal allegations. (Hiltner, 2017)

While this quote only speaks for the New York Times, print media generally have similar procedures. Another reason for the distinctions between print and cable could be that there is less control of what guests say on air, versus the print media’s control of using quotes within their stories.

44 Chapter 6

Discussion

The hypotheses of this thesis were that the liberal media would adhere to the recommendations of the four advocacy organizations more than the conservative media and that the media, as a whole, would use not recommended terms more than it used recommended terms.

After applying the advocacy organization’s recommendations to the content analysis, the first hypothesis, that liberal media would use recommended terms more than conservative media, was correct, at least for these particular media outlets. The second hypothesis, however, was not correct. Instead, the media, as a whole, used recommended terms far more than it used not recommended terms. These results held even after removing the subjective “probably” labels and analyzing only the explicitly recommended or not recommended terms. Although the difference between the liberal and conservatives use of not recommended terms was smaller than when including the “probably” labels, it still passed the 95% confidence test.

Beyond the hypotheses, other key takeaways were that the media, as a whole, performed quite well when referencing the victim and the abuser of the sexual violence, but quite poorly when referencing the statement of the sexual abuse (Ex: “accusation,” “claim,” “allegation,” etc.) and the sexual abuse (Ex: “sexual harassment,” “misconduct,” “abuse,” etc.). Also, incorporating the recommendations of journalistic institutions affects the conservative media more than the liberal media, but only ever so slightly. The journalistic institution recommendations overall did not affect the results in any significant way, suggesting that advocacy organizations are adhering to journalistic principles of accuracy and objectivity when issuing recommendations to the media. 45 When comparing the performance of each media outlet, the New York Times performs the best, according to advocacy organizations, and the Washington Times performs the worst, with the rest of the liberal-leaning outlets performing better than the rest of the conservative-leaning outlets. However, when eliminating the subjective “probably” labels, the New York Times performs the worst in terms of use of recommended terms, but remains best performing in terms of use of not recommended terms, implying that the New York Times has a higher proportion of terms labeled under a “probably” label.

Finally, relabeling “accuser” to “Not Recommended” affects the liberal and conservative media about the same, but increases the cable media’s use of not recommended terms more than the print media, suggesting that the less rigorous editing process or use of live guests, as opposed to the print media’s selection of source quotes, might affect the outlet’s performance of using recommended and not recommended terms.

One suggestion for future research would be to expand upon this content analysis to differentiate between quoted terms (or terms coming from someone not working for the media outlet, such as a live guest on cable news) and terms originating from that media outlet. Making this differentiation could provide more insight into the ideological bias of media outlets when reporting on sexual violence as it would eliminate or differentiate between people who represent a particular media outlet and those who simply appear in the media outlet. Although, an argument could be made for an ideological bias when choosing who to quote and invite on air.

Still, this would be a valuable addition to this content analysis.

I also suggest that further research on this topic includes the study of the effect of the terms found in this content analysis to either justify, nullify, or simply add to the debate on the usage of these terms in the media. This thesis sets the stage to answering the question of whether 46 these terms actually matter, how much, and what implications they have on the readers of these stories.

As for the implications of the results found in this content and data analysis, it should be remembered that the majority of the terms were subjectively labelled as they were not explicitly mentioned in the recommendations given by the advocacy organizations examined in this thesis.

Another person might have labelled these terms differently, thus altering the results of this study.

However, it should also be remembered that the results of the analysis including the subjectively labeled terms were on par with the results excluding these terms, suggesting the subjectivity in this thesis was as close to being objective as possible. 47 Appendix A

Resources for Content Analysis

Table 9. Timeframe Within Which #MeToo Events Were Selected for Content Analysis

Date of Initial Story Within #MeToo #MeToo Event Date range Movement Harvey Weinstein October 5, 2017 October 5, 2017 - November 5, 2017 Larry Nassar October 18, 2017 October 18, 2017 – December 6, 2017 Kevin Spacey October 29, 2017 October 29, 2017 – November 29, 2017 November 9, 2017 November 9, 2017 - December 9, 2017 Louis C. K. November 9, 2017 November 9, 2017 - December 9, 2017 Matt Lauer November 29, 2017 November 29, 2017 - December 29, 2017 Al Franken December 7, 2017 December 7, 2017 - January 7, 2018 R. Kelly May 10, 2018 May 10, 2018 - June 10, 2018 Les Moonves July 27, 2018 July 27, 2018 - August 27, 2018 Brett Kavanaugh September 16, 2018 September 16, 2018 - October 16, 2018

Table 10. Print Articles and Cable Segments Chosen for Content Analysis

Source Target Headline Date Byline

"Harvey Weinstein Paid Off Sexual Jodi Kantor and Megan New York Times Harvey Weinstein Harassment Accusers for Decades" 10/5/2017 Twohey "Refusing Weinstein’s Hush Money, Rose McGowan Calls Out New York Times Harvey Weinstein Hollywood" 10/28/2017 Susan Dominus

"The 'Me Too' Movement New York Times Larry Nassar Inevitably Spills into Sports" 10/19/2017 Juliet Macur

"Gymnast Says She New York Times Larry Nassar Was Molested by Team Doctor" 11/10/2017 Victor Mather

"Kevin Spacey Issues Apology to New York Times Kevin Spacey Actor After Sexual Accusation" 10/30/2017 Michael Paulson "Former TV Anchorwoman Accuses Kevin Spacey of Assaulting New York Times Kevin Spacey Her Son" 11/8/2017 Katherine Q. Seelye Richard Fausset, Jonathan "Sex Allegations Against Roy Martin and Campbell New York Times Roy Moore Moore Send Republicans Reeling" 11/9/2017 Robertson

"4 More Women Accuse Roy New York Times Roy Moore Moore of Misconduct" 11/15/2017 Alan Blinder 48 "Louis C. K. is accused by 5 women Melena Ryzik, Cara Buckley New York Times Louis C. K. of sexual misconduct" 11/9/2017 and Jodi Kantor "Louis C. K. responds to accusations: 'These stories are New York Times Louis C. K. true'" 11/10/2017 NO BYLINE Ellen Gabler, Jim Rutenberg, "NBC Fires Matt Lauer, the Face of Michael M. Grynbaum and New York Times Matt Lauer 'Today'" 11/19/2017 Rachel Abrams

"Lauer Offers an Apology, and a New York Times Matt Lauer Little Bit of a Denial" 12/1/2017 Michael M. Grynbaum "Al Franken Issues Apology After Accusations of Forcible Kissing and New York Times Al Franken Groping" 11/16/2017 Nicholas Fandos Sheryl Gay Stolberg, "Al Franken to Resign from Senate Yamiche Alcindor and New York Times Al Franken Amid Harassment Allegations" 12/7/2017 Nicholas Fandos

"R. Kelly faces a #MeToo reckoning New York Times R. Kelly as Time's Up backs a protest" 5/1/2018 Joe Coscarelli "What we know about R. Kelly's two-decade trail of sexual abuse New York Times R. Kelly allegations" 5/10/2018 Jacey Fortin "Les Moonves, CBS Chief, Faces Inquiry Over Misconduct New York Times Les Moonves Allegations" 7/27/2018 Edmund Lee "Les Moonves Stays as CBS C.E.O. While Its Board Plans an New York Times Les Moonves Investigation 7/30/2018 Edmund Lee "Kavanaugh's Nomination in Turmoil as Accuser Say He New York Times Brett Kavanaugh Assaulted Her Decades Ago" 9/16/2018 Sheryl Gay Stolberg

"The Women Who Have Accused New York Times Brett Kavanaugh Brett Kavanaugh" 9/26/2018 Christine Hauser "Weinstein in R-Rated shock Exposé reveals stars' & staff's claims of creepy come-ons, payoffs Lia Eustachewich and Bruce New York Post Harvey Weinstein by film mogul" 10/6/2017 Golding "DIRTY HARVEY Calling 'cut!' on toxic Wein Advisers & Mika bolt as New York Post Harvey Weinstein furor grows" 10/8/2017 Isabel Vincent "McKayla: #MeToo Olympic gymnast alleges abuse by team New York Post Larry Nassar doc" 10/19/2017 Yaron Steinbuch New York Post Larry Nassar "Gymnast: I trusted sick doc" 11/13/2017 Mark Moore New York Post Kevin Spacey "Actor: Spacey hit on me as a kid" 10/30/2017 Chris Perez "'Everyone knows' not cutting it New York Post Kevin Spacey anymore" 10/31/2017 Maureen Callahan "Sex claims roils race for 'bama New York Post Roy Moore Senate" 11/10/2017 Mark Moore

"Gop'ers tell Moore: Quit! New Kevin Sheehan, Max Jaeger New York Post Roy Moore accuser bares assault claim" 11/14/2017 and Ruth Brown 49 "It's C. K. and twisted Women accuse comedy star of pervy New York Post Louis C. K. misconduct" 11/10/2017 Ruth Brown

PERVANDO Louis admits: it's a;; Mara Siegler and Ruth New York Post Louis C. K. true' Fesses up to sick behavior 11/11/2017 Brown "Pervando Lauer outage is lust-see TV * 'Today' axes Matt amid sex Emily Smith and Danika New York Post Matt Lauer claims…" 11/30/2017 Fears "Pervando It wasn't news to NBC axes Complaints vs. Lauer 'all New York Post Matt Lauer brushed under the carpet" 12/1/2017 Emily Smith "Pervando Franken a big, fat 'creep' Sen. Groped and kissed me Mark Moore and Ruth New York Post Al Franken on my '06 tour" 11/17/2017 Brown "Pervando New accuser rips New York Post Al Franken Franken" 11/21/2017 Mark Moore

"Spotify's bad rap R. Kelly, New York Post R. Kelly XXXTentacion fight playlist ban" 5/11/2018 Nicolas Vega New York Post R. Kelly "Cut the R. Kelly" 5/16/2018 NO BYLINE "CBS boss 'got away with it for decades' Accusers slam TV honcho Moonves with sexual-misconduct New York Post Les Moonves claims" 7/28/2018 Ruth Brown

"Moonves can stay Will remain New York Post Les Moonves CBS boss amid sex harass probe" 7/31/2018 Alexandra Steigrad "Second accuser comes forward Brett exposed himself to me at New York Post Brett Kavanaugh Yale: woman" 9/24/2018 Chris Perez "JUDGMENT DAY Accusations at 11th hour Woman tells Brett-party Lia Eustachewich and Ruth New York Post Brett Kavanaugh 'rape' tale" 9/27/2018 Brown "3 women accuse Weinstein of Washington Post Harvey Weinstein rape" 10/11/2017 Paul Farhi and Elahe Izadi

"In Weinstein accusations, Washington Post Harvey Weinstein patterns of abuse span 30 years" 10/15/2017 Monica Hesse and Dan Zak "Maroney accuses doctor of sex Washington Post Larry Nassar abuse" 10/19/2017 Matt Bonesteel

"LGBT advocates denounce Washington Post Kevin Spacey Spacey's response to claim 10/31/2017 Amy B Wang and Elahe Izadi "In London, a raft of new Spacey Washington Post Kevin Spacey accusations" 11/17/2017 William Booth Stephanie McCrummen, "Moore accused of touching teen Beth Reinhard and Alice Washington Post Roy Moore girl" 11/10/2017 Crites

"Moore does not rule out that he Michael Scherer and David Washington Post Roy Moore dated teen girls" 11/12/2017 Weigel "Grappling not only with Louis C. K., but sexism and power in Washington Post Louis C. K. comedy" 11/18/2017 Elahe Izadi "The sexual harassment epidemic has been diagnosed. What's the Washington Post Louis C. K. cure?" 11/20/2017 Monica Hesse 50 "Matt Lauer fired; misconduct Washington Post Matt Lauer alleged" 11/30/2017 Paul Farhi "A cloudy morning after for Washington Post Matt Lauer 'Today'" 12/1/2017 Emily Yahr

"Al Franken accused of groping Paul Kane, Amy B Wang and Washington Post Al Franken woman in 2006" 11/17/2017 Lindsey Bever

"Democrats call for Franken to Elise Viebeck, Ed O'Keefe Washington Post Al Franken resign" 12/7/2017 and Karen Tumulty

"Woman accuses R. Kelly of giving Washington Post R. Kelly her a sexually transmitted disease" 4/16/2018 Geoff Edgers Washington Post R. Kelly "Behind the music, a lot of silence" 5/6/2018 Geoff Edgers

"CBS to probe sexual assault claims Washington Post Les Moonves against CEO Moonves" 7/28/2018 Steven Zeitchik

"The trick of enforcing Wall Washington Post Les Moonves Street's 'Weinstein clauses'" 8/12/2018 Jena McGregor Felicia Sonmez, Seung Min "Kavanaugh and accuser to testify Kim, Sean Sullivan and John Washington Post Brett Kavanaugh before the Senate" 9/18/2018 Wagner

Karoun Demirjian, Amy Washington Post Brett Kavanaugh "Second woman accuses nominee" 9/24/2018 Gardner and Seung Min Kim "TV news anchor accusses Harvey Weinstein of masturbating in front Washington Times Harvey Weinstein of her: Report" 7-Oct Andrew Blake

"Harvey Weinstein accused of Washington Times Harvey Weinstein raping three women: Report" 10/10/2017 Jessica Chasmar "Spacey's coming out as gay follows political sex scandal Washington Times Kevin Spacey playbook" 10/31/2017 Valerie Richardson "Kevin Spacey accused of dating 14-year-old, attempted rape: "He Washington Times Kevin Spacey is a pedophile'" 11/2/2017 Jessica Chasmar "Washington COP tries to oust Republican nominee Roy Moore in Washington Times Roy Moore Alabama Senate race" 11/9/2017 Stephen Dinan "Roy Moore accuser alters story, admits to adding to yearbook Alex Swoyer and Sally Washington Times Roy Moore signing" 12/8/2017 Persons

"Louis C.K. accused by five women Washington Times Louis C. K. of sexual misconduct" 11/9/2017 Jessica Chasmar " 'stunned' by Louis C.K. sex scandal, says friendship Washington Times Louis C. K. caused blindness" 11/14/2017 Douglas Ernst "NBC fired 'Today' show host Matt Lauer for inappropriate sexual Washington Times Matt Lauer behavior" 11/29/2017 Sally Persons

": Matt Lauer 'pinches Washington Times Matt Lauer me on the ass a lot'" 11/29/2017 Jessica Chasmar "Franken to face ethics investigation, outraged colleagues Washington Times Al Franken from both parties" 11/16/2017 Sally Persons 51 "Al Franken accuser Stephanie Kemplin 'appalled' by resignation Washington Times Al Franken speech" 12/7/2017 Jessica Chasmar "CBS exec Les Moonves investigated amid sexual Washington Times Les Moonves misconduct claims" 7/27/2018 Jessica Chasmar "Kavanaugh: 'This never Washington Times Brett Kavanaugh happened'" 9/17/2018 Stephen Dinan "Brett Kavanaugh denies new sexual assault accusation from Washington Times Brett Kavanaugh Deborah Remirez" 9/23/2018 Alex Swoyer , , "Harvey Weinstein Accused of Steve Moore, Dianne CNN Harvey Weinstein Sexual Harassment" 10/6/2017 Gallagher, Erin McLaughlin "Harvey Weinstein Now Facing Sex Assault Allegations; Paltrow and , , Jolie Say Weinstein Harassed , Emily CNN Harvey Weinstein Them" 10/10/2017 Jane Fox Isha Sesay, John "Trump's feud over condolence Vause, Jeff Zeleny, CNN Larry Nassar call…" 10/19/2017 , , , "Trump shifts focus to Boris Sanchez, David CNN Kevin Spacey Democrats…" 10/30/2017 McKenzie, Leyla Santiago

"Trump accuses Clinton and DNC Brooke Baldwin, Samantha CNN Kevin Spacey of real 'collusion'" 11/3/2017 Vinograd, Chloe Melas "Woman says Alabama Senator Candidate Roy Moore initiated , Alex Marquardt, sexual encounter when she was , , CNN Roy Moore 14, he was 32" 11/9/2017

Anderson Cooper, , , Amanda Carpenter, Jack "Roy Moore denies meeting Kingston, Pamela Brown, woman accusing him of sexually Richard Ben-Veniste, Mark CNN Roy Moore abusing her" 11/10/2017 Hertling, Gary Tuchman

Brooke Baldwin, Chloe "Comedian Louis C. K. Sexual Melas, MJ Lee, Brynn CNN Louis C. K. Misconduct Reports 'Are True'' 11/10/2017 Gingras, Kate Bennett

Poppy Harlow, , Christine Romans, Chad CNN Louis C. K. "Louis C. K. Admits to Misconduct" 11/13/2017 Myers,

Will Ripley, Jeff Zeleny, , , Philip Mudd, David Swerdlick, Rebecca Berg; ; CNN Matt Lauer "NBC Fires Matt Lauer" 11/29/2017 Hadas Gold "Matt Lauer's new job after being CNN Matt Lauer ousted" 11/30/2017 Ashley Banfield, 52

Jim Acosta, Phil Mattingly, , Wolf Blitzer, , Rebecca Berg, "Al Franken apologizes in wake of , David CNN Al Franken sexual misconduct allegations" 11/16/2017 Swerdlick, Elise Labott

"Democratic senators call on Dana Bash, MJ Lee; Wolf CNN Al Franken Franken to resign" 12/6/2017 Blitzer,

Manu Raju, , , Susan Hennessey, Rebecca Berg, Samantha " publishes Vinograd, , women's sexual harassment Dylan Byers, Paul allegations CBS Chairman & CEO Vercammen; Brynn Gingras; CNN Les Moonves Les Moonves" 7/27/2018 Evan Perez

Fredricka Whitfield, Sarah Westwood, Evan Perez, , David Swerdlick, Matt Lewis, Elise "CBS Board Knew of Sex Assault Labott, Kaylee Hartung, CNN Les Moonves Claims against Moonves" 8/4/2018 Martin Savidg , Jim Sciutto, "Brett Kavanaugh Accuser Says Sunlen Serfaty, Abby Phillip, She's Willing to Testify Before Ron Brownstein, Nia-Malika CNN Brett Kavanaugh Congress" 9/17/2018 Henderson, Abigail Tracy , Jessica Schneider, Boris Sanchez, Paul Callan, W. Kamau Bell, Patrick Snell, Nick Valencia, "Kavanaugh Accuser will testify on , James Gagliano, CNN Brett Kavanaugh Thursday" 9/23/2018 , Brian Stelter Daily with Chuck MSNBC Harvey Weinstein Todd 10/9/2017 , Beth Fouhy MSNBC Harvey Weinstein The Beat with Ari Melbur 10/10/2017 , Chris Matthews, Milissa Rehberger, Sophia Nelson, MSNBC Kevin Spacey Hardball with Chris Matthews 11/8/2017 Beth Fouhy

Chris Hayes, , MSNBC Kevin Spacey ALL IN with 11/10/2017 Carol Lee, Nick Akerman

Lawrence O'Donnell, Wendy The Last Word with Lawrence Sherman, Maria Teresa MSNBC Roy Moore O'Donnell 11/13/2017 Kumar Chris Matthews, James Peterson, Ken Dilanian, Jonathan Allen, Milissa MSNBC Roy Moore Hardball with Chris Matthews 11/9/2017 Rehberger;

Chuck Todd, Maria Teresa Meet the Press Daily with Chuck Kumar, , MSNBC Louis C. K. Todd 11/10/2017 MSNBC Louis C. K. Hardball with Chris Matthews 11/15/2017 Chris Matthews, 53 , Willie MSNBC Matt Lauer Discussion on Matt Lauer Firing 11/29/2017 Geist, Meet the Press Daily with Chuck MSNBC Matt Lauer Todd 11/30/2017 MSNBC Al Franken The Last Word with 11/16/2017 Ali Velshi Chris Matthews, Betsy Woodruff, Andrea Mitchell, Betsy Woodruff, Kasie Hunt MSNBC Al Franken Hardball with Chris Matthews 12/6/2017 and Page Hopkins

Shane Harris, Danielle Pletka, Cornell Belcher, John Kennedy, Chuck Rosenberg, Danielle Pletka and Shane MSNBC Brett Kavanaugh Meet the Press Daily 9/18/2018 Harris MSNBC Brett Kavanaugh The Show 9/24/2018 Rachel Maddow "Several Women Claim They Have Been Assaulted by Harvey , Kimberly Weinstein; Harvey Weinstein Faces Guilfoyle, Gillian Turner, Grownig Backlash From Kennedy, Richard Fowler Fox News Harvey Weinstein Hollywood" 10/10/2017 and Dan Springer "Multiple investigations of Harvey Fox News Harvey Weinstein Weinstein underway" 10/16/2017 Martha MacCallum , Lucas Tomlinson, , Jonathan Serrie, , Jonathan Hunt, Benjamin Hall, Mike Tobin, Caroline Shively, Adam Housley and Fox News Larry Nassar "Lost at sea…" 11/22/2017

"Former Trump campaign manager Martha MacCallum, Ed and assoiate face money Henry, David Bossie and Fox News Kevin Spacey laundering…" 10/30/2017

"Ex-Trump campaign chair and , , former campaign aide pleade not Mollie Hemingway and Fox News Kevin Spacey guilty…" 11/5/2017 Marie Harf

Bret Baier, Mike Emanuel, Jonathan Serrie, , , , Doug "Moore on the defense; Will of the McKelway, , Casey Fox News Roy Moore voters" 11/10/2017 Stegall and "Washington Post Reported a Jesse Watters, Juan Woman Accused Judge Roy Moore Williams, Kimberly Guilfoyle, Initiated Sexual Encounters When and Greg Fox News Roy Moore She was 14 years old"" 11/9/2017 Gutfeld "Louis C. K. Accused of Sexual Fox News Louis C. K. Misconduct" 11/9/2017

"Republican Roy Moore Accused Howard Kurtz, Mollie of; Republican Party Under Hemingway, Fox News Louis C. K. Pressure" 11/12/2017 and Erin McPike 54 "NBC News Fires Matt Lauer Over Fox News Matt Lauer Sexual Misconduct Allegations" 11/29/2017 Sean , Ed Henry

"More women come forward with Kimberly Guilfoyle, Dana allegations against Matt Lauer and Perino, Jesse Watters, Juan Fox News Matt Lauer former Today host issues apology 11/30/2017 Williams, "Melanie Morgan: Al Franken Fox News Al Franken Harassed Me" 11/16/2017 Laura Ingraham

"Senator Al Franken Faces Ethics Kimberly Guilfoyle, Dana Probe in Wake of sExual Perino, Juan Williams, Pete Fox News Al Franken Misconduct Allegations" 11/17/2017 Hegseth, Kennedy

Laura Ingraham, Mike Emanuel, Doug McKelway, Ellison Barber, Catherine Herridge, Mike Tobin, Juan "Despite Focus on , U.S. Intel Williams, Trace Gallagher, Fox News Les Moonves Eyes Threats" 7/27/2018 Daniel Hoffma Fox News Les Moonves "Media Furor Over Cohen Tape" 7/29/2018 Howard Kurtz, Guy Benson

"President had ordered declassification of Carter Fox News Brett Kavanaugh Page's FISA warrants" 9/17/2018 ,

"Interview with Sen. Lindsey , Gillian Turner, Fox News Brett Kavanaugh Graham" 9/23/2018 Juan Williams 55

BIBLIOGRAPHY

Cohen, E. (2009). Semi-Citizenship in Democratic Politics. Cambridge, MA; New York:

Cambridge University Press.

Department of Justice. (2000). Sexual Assault of Young Children as Reported by Law. Office of

Justice Programs, Buraeu of Justice Statistics.

Druckman, J. (2001). The Implications of Framing Effects for Citizen Competence . Political

Behavior, 225-256.

Ennis, E., & Wolfe, L. (2018). #MeToo: The Women's Media Center Report. The Women's

Media Center.

Entman, R. (1993). Framing: Toward Clarification of a Fractured Paradigm. Journal of

Communication, 51-58.

Fitts, A. S. (2013, July 18). The right way to write about rape. Retrieved from Columbia

Journalism Review:

https://archives.cjr.org/minority_reports/the_right_way_to_write_about_r.php

Gladi, S., Maass, A., & Cadinu, M. (2014). Objectifying Media: Their Effect on Gender Role

Norms and Sexual Harassment of Women. Psychology of Women Quarterly, 398-413.

Glover, A., & Beard, D. (2017, November 22). Is it assault or harassment? AP decides

‘misconduct’ is a more encompassing word choice. Retrieved from The Poynter Institute:

https://www.poynter.org/reporting-editing/2017/is-it-assault-or-harassment-ap-decides-

misconduct-is-a-more-encompassing-word-choice/ 56

Goffman, I. (1974). Frame analysis: an essay on the organization of experience. Cambridge,

MA: Press.

Grady, C. (2017, November 30). The Complicated, Inadequate Language of Sexual Violence.

Retrieved from Vox: https://www.vox.com/culture/2017/11/30/16644394/language-

sexual-violence

Hiltner, S. (2017, October 31). How We Describe Sexual Assault: Times Journalists and Lawyers

Respond. Retrieved from New York Times:

https://www.nytimes.com/2017/10/31/reader-center/sexual-assault-terminology.html

Houser, K. (2019). Words Matter: How language, context, and content can perpetuate or change

public bias about sexual violence. Retrieved from National Sexual Violence Resource

Center.

Houser, K. (n. d.). Words Matter: How language, context, and content can perpetuate or change

public bias about sexual violence. National Sexual Violence Recourse Center.

Kahneman, D., & Tverski, A. (1984). Choices, Values, and Frames. American Psychologist, 341-

350.

Lonsway, K., & Archambault, J. (2013, June). Suggested Guidelines on Language Use for

Sexual Assault. Retrieved from End Violence Against Women International:

https://www.evawintl.org/Library/Detail.aspx?ItemID=1219

Merolla, J., Ramakrishnan, S. K., & Haynes, C. (2013). “Illegal,” “Undocumented,” or

“Unauthorized”: Equivalency Frames, Issue Frames, and Public Opinion on Immigration.

American Political Science Association, 789-807.

Nelson, T., Clawson, R., & Oxley, Z. (1997). Media Framing of a Civil Liberties Conflict and Its

Effect on Tolerance. The American Political Science Review, 567-583. 57

Perlman, M. (2018, May 7). New AP Stylebook guidelines, influenced by #MeToo, hurricanes,

and online polls. Retrieved from Columbia Journalism Review:

https://www.cjr.org/language_corner/ap-stylebook.php

Pew Research Center. (2018, March 20). Wide Gender Gap, Growing Educational Divide in

Voters’ Party Identification. Retrieved from Pew Research Center: https://www.people-

press.org/2018/03/20/wide-gender-gap-growing-educational-divide-in-voters-party-

identification/

RAINN. (n.d.). Key Terms and Phrases. Retrieved from Rape, Abuse, Incest National Network:

https://www.rainn.org/articles/key-terms-and-phrases

Scheufele, D. (1999). Framing as a Theory of Media Effects. Journal of Communication, 103-

122.

58

ACADEMIC VITA

Education

Pennsylvania State University • Bachelor of Arts in Digital and Print Journalism • Bachelor of Arts in Political Science

Experience

CommAgency – Social Media Division 01/2020 – 05/2020 • Worked on behalf of CommAgency, a student-run media production agency, to create social media content for an off-campus creative student living space called co.space College of Communications - Research Assistant 05/2017 – 05/2020 • Traveled to South Africa, the first time for an embedded study abroad program and the second to collect qualitative data on the tabloid newspaper, the Daily Sun; Collected literature, designed and administered a questionnaire, transcribed interviews, and developed thematic categories; coauthoring two manuscripts for publication in academic journals. Teaching Assistant for Environmental Science Course 08/2019 – 12/2019 • Planned and conducted two weekly labs of 12 students each, graded 24 written journal assignments every two weeks, assisted students with any issues they had with the course all while teaching them about the threat of climate change Indiana Democratic Party - Communications Intern 05/2019 - 08/2019 • Created graphics for the Party's social media using Canva and Adobe Spark • Storyboarded, filmed, and edited approximately a dozen videos such as vlog material, candidate profiles, and more • Manned the front desk once a week, answering phones, and assisting people who reached out and helped run events such as fundraisers and dinners C-SPAN - Washington Journal Intern 09/2018 - 12/2018 • Promoted guests of the show on , greeted and assisted them with their C-SPAN experience, and helped with the production and promotion of the show; TweetDeck proficiency The Daily Collegian - Men's Volleyball Reporter 01/2017 - 05/2017 • Covered the Penn State Men's Volleyball team, averaging three articles per week abiding by AP-Style 59

• Conducted interviews, held press conferences, and coordinated with fellow reporters and editors on how to best cover the team