Killings in Context: an Analysis of the News Framing of Femicide

Abstract

While attacks against members of the LGBT community are increasingly covered as hate crimes and are widely viewed as a form of repression, attacks on women are almost never covered as violations of human rights. We propose that until violence against women is recognized as a form of repression and a threat to the physical security of women, we cannot expect much to be done to prevent it. We posit that policies aimed at preventing violence against women are unlikely to come about unless this abstract concept is connected, through a connection frame, to concrete crimes against women. We conducted a framing analysis of news coverage of all confirmed femicides in Massachusetts in 2013 and find that while journalists have the potential to draw these connection frames, they seldom put these killings in the context of violations of women’s rights. Full Text available: http://rdcu.be/Hxhb

Full text

1
This is a post-peer-review, pre-copyedit version of an article published in Human Rights Review. The final
authenticated version is available online at: https://doi.org/10.1007/s12142-018-0497-3
Killings in Context: An analysis of the news framing of femicide
Camelia Bouzerdan and Jenifer Whitten-Woodring*
Department of Political Science
Global Studies Doctoral Program
University of Massachusetts Lowell
Dugan Hall, 203
Lowell MA 01854
Corresponding Author
Jenifer Whitten-Woodring
Email: jenifer_whittenwoodring@uml.edu
Phone: 858-829-3142
*Names are listed in alphabetical order; authorship is equal

2
Killings in Context: An analysis of the news framing of femicide
Introduction
On October 6, 1998 in Laramie Wyoming, Matthew Shepard, age 21, a blond, white
student, was robbed, beaten, tied to a fence and left to die. According to news reports, Shepard
was killed because he was gay.
1
His killing was covered as a hate crime. The case drew national
news coverage, fueled public outrage and eventually led to the passage of the Matthew Shepard
and James Byrd Hate Crimes Prevention Act of 2009, which expanded hate crime law to include
crimes motivated by the victim’s sexual orientation or gender.
On February 23, 1997 in Costa Mesa California, Sunny Sudweeks, age 26, a blond, white
student, was found dead in her apartment by her boyfriend. She had been raped and strangled.
According to news reports, robbery was ruled out as a motive, and there was no evidence that
drugs were involved. Though the news accounts did not mention it, the only motivation for the
crime appeared to be the sexual attack. In other words, Sudweeks was killed primarily because
she was a woman. Her death was covered as a mystery. The case drew limited attention and
remained unsolved until 2017 when DNA evidence pointed to a suspect who had previously been
convicted on domestic violence charges.
This tale of two killings points to a disparity in the way that violence against women is
covered compared to violence against other marginalized groups. While attacks against members
of the LGBT community are increasingly covered as hate crimes and are widely viewed as a
form of repression, we propose attacks on women are almost never covered as violations of
human rights. Previous studies have found that journalists tend to cover killings of women as
1
It was later revealed that drugs may also have played a role in Shepard’s killing (for more see
Jimenez 2013).

3
isolated crimes rather than part of a larger problem of violence against women (Radford and
Russell 1992, Humphries 2009, Lundman 2003). Building on these studies, we propose that until
violence against women is recognized as a form of repression and a threat to the physical
security of women, we cannot expect much to be done to prevent it. In this study, we predict that
violence against women will generally be covered and, in effect, dismissed as isolated crimes
unless there are extenuating circumstances. We posit that policies aimed at preventing violence
against women are unlikely to come about unless this abstract concept is connected to concrete
crimes against women. Journalists clearly have the potential to draw these connection frames
2
when they cover killings of women, but we find that they seldom put these killings in the context
of violations of women’s rights.
We begin by reviewing the literature relevant to news coverage of violence against
women, including research on human rights, political communication, criminal justice and media
studies. We develop our theory through an inductive comparison of the news coverage of the
Matthew Shepard and Sunny Sudweeks cases. We then test our hypotheses with a deductive
framing analysis of the news coverage of all 15 women killed by men in Massachusetts in 2013.
3
Through this in-depth analysis of a small number of cases, we find that violence against women
is almost never treated as a human rights problem. Instead, with one important exception, these
2
Below and Whitten-Woodring (2008) introduced the term connection frames as the practice
of connecting abstract concepts to concrete events.
3
We selected Massachusetts because we needed a manageable sample so that we could take
an in-depth look at all cases for a given year and assess the effects on state policy. We chose
2013 because we wanted a year that was recent, but distant enough to have played out in the
news media (and following the passage of the Matthew Shepard and James Byrd Hate Crimes
Prevention Act, which expanded hate crimes to include gender-based crimes), yet also a year
for which crime statistics were readily available. 2013 was the most recent year for which
comprehensive data were available.

4
killings were treated as unconnected crimes or dismissed as individual or family matters. We
conclude with a discussion of the implications of our findings.
Violence Against Women in the News
Violence against women is a global problem that is present in highly developed countries
as well as developing countries (Hudson, Ballif-Spanville, Caprioli and Emmett 2012; Richards
and Haglund 2015). The United Nations Declaration on the Elimination of Violence against
Women (1993, Article 1) defines it as, “any act of gender-based violence that results in, or is
likely to result in, physical, sexual or psychological harm or suffering to women, including
threats of such acts, coercion or arbitrary deprivation of liberty, whether occurring in public or in
private life.” Thus, violence against women constitutes a violation of women’s human rights.
Yet, more than two decades after this declaration, women around the world continue to suffer
from violence ranging from verbal abuse to murder. At the global level, one in three women
experience violence, usually from an intimate partner (UN Women 2017). Though brutality
against women can come at the hands of strangers, all too often it comes from those who are
closest to them. “All over our world, those most likely to physically injure women are men they
love or have loved: husbands, boyfriends, ex-husbands, ex-boyfriends, fathers, brothers,”
(Hudson et al. 2012: 21). Indeed Petrosky et al. (2017) found that more than half of the women
murdered in the United States were killed by intimate partners. Comparatively Bridges et al.
(2008) found that only about 5% of male homicide victims were killed by current or former
partners.
According to a Violence Policy Center (2015) study, more than 1600 women were
murdered by men in the United States in 2013. Most of these killings can be characterized as

5
femicides.
4
Following the World Health Organization (2012: 1), we conceptualize femicide as
the “intentional murder of women because they are women.” Though, more expansive
definitions include any killings of women or girls, for the purposes of our study we prefer this
narrow definition. Femicide can be divided into intimate and non-intimate femicide, depending
on the relationship between the killer and victim, as well as the motive of the crime. Intimate
femicides are killings committed by those in close relationships with the victim, especially
current or former husbands or boyfriends (WHO 2012). Non-intimate femicides are killings
committed by perpetrators who have little or no knowledge of the victim and are often motivated
by anger against women and/or sexual aggression (WHO 2012). Although femicide is a leading
cause of death, studies of femicide are scarce and the concept of femicide seldom appears in
academic literature to describe the killing of women.
Recent studies suggest that media coverage of violence against women could potentially
shape public knowledge and opinion about the issue in ways that could lead to political action
(Clark 2012; Haydari and Kara 2015; Easteala et al. 2014; Wood 2009). Yet, these studies find
that the opportunity to bring about policy change is often missed because media seldom cover
femicide as femicide or violence against women. In keeping with these studies, we propose
words matter when it comes to covering femicide and violence against women.
News media typically cover the killing of women as separate incidents rather than as
parts of a broader context. This presentation of femicide suggests that the killing was motivated
by isolated circumstances rather than as part of a pattern of violence against women. York
(2011:66) argued that this type of coverage normalizes violence against women and discourages
4
Radford and Russell (1992) introduced the concept of femicide, which was then adopted by
Della Giustina (2010), Humphries (2009), and others.

6
efforts to prevent it. Similarly, Meyers (1997) argued that by reporting with a sense of urgency
and excitement, the media fail to put the crime in context. In other words, the focus on shocking
details draws attention away from human rights and women’s rights. Moreover, Heinze and
Freedman (2009) identified the gap between the degree of media coverage and the gravity of
human rights implications. Although it is difficult to quantify the opportunity cost of this failure
to focus on human rights violations, we posit that news media could narrow the gap by explicitly
incorporating the human rights violation angle when covering femicide. The argument for doing
so is that it might slow down and eventually stop the normalization of violence against women.
In short, if incidences of male aggression continue to be portrayed as isolated crimes, violence
against women will persist and policy change is unlikely (Wood 1994).
Several studies across countries and over time have found that news media can make a
difference in human rights in general and women’s rights in particular. Clark (2012) identified a
positive relationship between newspaper circulation and physical integrity rights (the rights to be
free from political imprisonment, torture, murder and disappearance), and Whitten-Woodring
(2016) found that the interaction of media freedom and internet access is associated with
improved physical security for women. Yet, news media often fail to cover violations of human
rights (Brandle 2016, forthcoming; Powers 2016). Because of the pivotal role that the media can
play in promoting human rights, Haydari and Kara (2015) recommend that journalism schools
include human rights coverage in their curriculum. In particular, they suggest that future
journalists have the opportunity while in school to be part of social projects outside the
university in order to build strong skills to promote social justice and cultural dialogue (Haydari
and Kara 2015). Yet, Reilly (forthcoming) finds that US journalism degree programs seldom
offer training in the reporting on human rights.

7
A common critique of news media stems from the influence that sources have on news
coverage. According to the Indexing theory, rather than holding government accountable through
independent watchdog reporting and analysis, journalists tend to index the opinions and
comments of official sources (Bennett 1990). Not surprisingly then, reporters’ reliance on
sources strongly influences how media cover the killing of women. After conducting in-depth
interviews with local journalists in a metropolitan area, York (2011) concluded that reporters’
reliance on police spokespersons has made the police perspective the foundation of news stories.
Moreover, by endorsing the police perspective as the legitimate source in most every case, this
dependence is amplified. Another crucial factor in shaping news coverage of femicide is the
legislation in the state where the crime takes place. In discussing the reasons why violence
against women is not portrayed as hate crime, Young (1999) posited that when hate crime laws
were introduced, these laws were not concerned with the protection of the victim, but rather with
a political argument or the punishing of the offender. Yet, others have argued that when news
media give too much attention to hate crimes, a moral panic will ensue, as with the killing of
Matthew Shepherd (Colomb and Damphousse 2004).
In contrast, we hypothesize that when journalists emphasize the hate crime angle and
draw attention to a particular crime as an example of the repression of human rights, this will
draw public attention to the underlying social problem. In effect by connecting a specific crime
to an abstract human right, journalists are drawing a connection frame (Below and Whitten-
Woodring 2008). This is an extension of framing. According to Entman (1993: 52), by
highlighting some aspects of a story over another, journalists are constantly framing their stories:
Framing essentially involves selection and salience. 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

8
interpretation, moral evaluation, and/or treatment recommendation for the item
described.
Below and Whitten-Woodring (2008: 1) developed the idea of the connection frame in the
context of climate change and posited that news media can help to boost such abstract issues on
the policy agenda by “providing a ‘connection frame’ linking the abstract issue of impending
climate change to the concrete issue of severe weather events.” Here we build on this idea and
argue that if journalists connect individual acts of femicide to the relatively abstract problem of
violence and against women and repression of women’s rights, there will be more public
pressure for the development of policies and law enforcement practices aimed at curbing
violence against women.
Additionally, crime news coverage is culturally specific, meaning the manner in which
the story is reported will correspond with social, political, and economic trends in society
(Meyers 1997). In patriarchal societies where the violence against women is more accelerated,
the socioeconomic environment is crucial in determining how women experience domestic
violence (Bhattacharjee and Banda 2016). Wood (2009) proposed that institutions promote
gendered violence by normalizing violence against women and identified the role of language
used in communication as a key factor. Additionally, Easteal, Judd and Holland (2015: 111),
found that media consistently frame violence against women with a theme of “mutuality of
responsibility” and that this “victim-blaming” detracts from the underlying social causes of the
violence. Although there are many reasons why gendered violence persists, Tarshis (1992)
claimed that increased violence against women is a response to the women’s movement. These
outcomes are most visible in democratic societies where freedom of expression is allowed. Yet,
Winston and Pollock (2016) argued that while freedom of expression is necessary, it is not

9
sufficient because it has to be complemented by the freedom of press. When these conditions are
met, the news media’s capacity for reaching mass audiences increases.
Lastly, journalistic assessment of the newsworthiness of a particular murder is a function
of certain particularities. Several studies indicate that journalists consider gender and race when
they decide whether to cover a murder, or how many times a murder will be covered
(Gruenewald, Pizarro and Chermak 2009; Lundman 2003; Lundman, Douglass and Hanson
2004). Sorenson and Peterson (1998) found that homicides with victims from wealthier
neighborhoods received more coverage. In addition, victims from black or hispanic communities,
or victims with less education, typically receive less coverage (Sorenson and Peterson 1998).
In summary, previous findings indicate that (1) when news media cover femicide, the
story is generally outlined as an isolated incident, not as a part of a bigger problem of violence
against women, (2) the reporters rely mainly on the police and the legislation to formulate the
story, (3) the crafting of the story is culturally specific, and (4) the newsworthiness of a story is
decided based on the actors involved, in particular their race, gender, education, and economic
status.
A Tale of Two Killings: A theory of how news coverage might make a difference
We build on previous research by first comparing news coverage of two killings: the
Matthew Shepard case, which propelled violence against gays into the national spotlight and
eventually led to policy change, and the Sunny Sudweeks case, which, though clearly a case of
violence against women, garnered minimal news coverage at the local level and did not prompt
any changes in policy or practice. We use this comparison to derive our theory and inductively
identify the frames for our analysis. We then test our hypotheses with a deductive paragraph-by-

10
paragraph framing analysis of the news coverage of the 15 confirmed killings of women by men
in Massachusetts in 2013.
We selected the Matthew Shepard case because it is one of only a few gendered violence
cases where intense news coverage arguably led to reform. We then searched for a most similar
case involving a femicide and found the Sunny Sudweeks case. Both were students in their 20s,
white and blond, and both were killed in the late 1990s. Though it is true that a homicide in
Wyoming is likely to garner more news coverage than one in a heavily populated area like
Orange County, California, the purpose of this comparison is not to compare the quantity of the
coverage but the quality of the coverage. We seek to identify how news coverage, in particular
news framing, contributed to the disparity of public and political attention to these two cases. To
analyze the news coverage we used Nexus Lexis to capture all available articles about the Sunny
Sudweeks case from the Orange County Register from February 24, 1997 to March 3 2017 (14
articles total), which included stories about the killing, the investigation, reactions of family,
friends and community members, memorial services and finally more than two decades later, the
identification of the suspected killer.
5
We also used Nexus Lexis to capture all available articles
from the Wyoming Tribune Journal about Matthew Shepard from October 10, 1998 to October
14, 1998 (25 articles total), which included stories about the crime and Shepard’s death, the
investigation, reactions of family, friends, community members, activists and political leaders,
vigils and memorial services. Both newspapers are well-respected regional publications. We
begin with a summary of the news coverage of each crime.
5
Many news articles were analyzed as part of this study. To conserve space, we include only
the references to articles that are quoted directly in our reference section. References to all of
the articles used in our analyses are included in the online appendix at
http://faculty.uml.edu/Jenifer_whittenwoodring/.

11
Sunny Sudweeks’ boyfriend found her body in their apartment when he returned home
from his shift as a cab driver early in the morning on February 23, 1997. She had been raped and
strangled in her bed. Investigators quickly eliminated Sudweeks’ boyfriend as a suspect and
determined that robbery was not the motive. From the beginning, the news articles in the Orange
County Register focused on the procedures of the police investigation and the human-interest
angle, in particular the reactions of friends and family members and their memories of
Sudweeks’ striking appearance, her penchant for vintage clothing and her fondness for children.
It was not until the one year anniversary of her death that a news article implied Sudweeks’
killing was a case of gendered violence, “More likely it was a stalker—or an acquaintance—who
sexually assaulted and killed her,” (Rhoads 1998: B01). Yet, Sudweeks’ death was never framed
as a case of violence against women, a hate crime or a violation of human rights. In fact, the
terms “violence against women” and “domestic violence” appeared only once each in the
coverage of Sudweeks’ murder, not in connection to her death, but respectively in reference to a
fundraiser held by Sudweeks’ boyfriend in her honor and to a previous conviction of her
suspected killer (Rhoades 1998; Fausto 2017). In the 20 years it took to solve the crime,
Sudweeks’ family and friends struggled to keep her story in the news, including offering a
reward for information leading to an arrest. In the article marking the first anniversary of
Sudweeks’ slaying, her mother noted that, “I have to market my daughter’s death,” (Rhoades
1998).
In contrast to Sudweeks, the killing of Matthew Shepard was admittedly more dramatic.
She was killed in her bed and her death was relatively immediate. Shepard was lured to a bar to
meet his killers (who allegedly pretended to be homosexual), then driven to a remote area where
he was robbed, severely beaten, tied to a fence and left to die. Shepard clung to life for a week

12
before he succumbed to his injuries. From the beginning the Wyoming Tribune Eagle’s news
coverage framed the attack on Shepard as violence against gays, a hate crime, and evidence of
the need for policies to combat repression of homosexuals. This depiction was initiated by
Shepard’s friends and gay rights activists and confirmed by police.
6
Even before Shepard’s
death, the case was connected to the recent failure of Wyoming state lawmakers to pass an anti-
hate crime law that would have increased penalties in cases where victims were targeted based
on a number of characteristics including sexual orientation (but not gender). The suspects were
quickly identified, arrested and arraigned, so there was less reporting on the procedures of
investigation. Instead, journalists focused on how the University of Wyoming community rallied
in support of Shepard and against hate crime and an online petition that urged Wyoming
Governor Jim Geringer and state lawmakers to pass hate crime legislation. By the time Matthew
Shepard died from his injuries, reporters had interviewed the governor and several legislators
about the case and failed hate crime legislation. Although news coverage also included human
interest frames with reactions from family and friends of Shepard, the reporting in the Wyoming
Eagle Tribune often framed the case as violence against gays, a hate crime and a violation of
human rights.
To analyze and compare the use of language in the Sudweeks and Shepard cases and
identify frames to code for in our primary analysis, we identified the most frequently used words
in each case. Figure 1 depicts the use of language with a word cloud for each case. The most
frequently used words are larger and more central in the word clouds. For the Sudweeks case, the
most frequently used word was “police” followed by “school” and “crime.” The frequency of the
6
Recently there has been some debate as to the motives behind Shepard’s killing. Jiminez
(2013) proposed that it was drugs rather than sexual orientation that prompted the attack, but
these claims have been widely criticized.

13
words “police” and “crime” is indicative of a procedural frame focusing on the police procedures
and investigation. The frequency of “school” reflects the human-interest frame of Sudweeks as a
student. Although “rape” is somewhat prominent in the Sudweeks coverage, violence against
women is not. In contrast, “Hate,” is front and center in the Shepard case, indicating a morality/hate
crime frame, followed by “crimes,” “community,” “people,” and “university,” which reflect more
of a human-interest frame. In the Shepard case, references his being “gay” are somewhat frequent,
suggesting a frame of violence against gays.
We analyzed the news coverage of these cases with an open view to identify possible
frames as well as some the typical news frames identified in previous studies, especially
responsibility, human interest, and morality (Smetko and Valkenburg 2000). In addition to these
standard frames, we found a great deal of what we characterized as procedural frames, frames
focusing on factual non-dramatic coverage of the investigation, court proceedings, or basic
description of the crime. In the Shepard case, there were also frames connecting the crime to the
problem of violence against gays and the need for policies to prevent hate crimes. We also noticed
some morality frames had more of a human rights angle than the standard religious angle,
emphasizing the immorality of anti-gay violence and violence as a form of repression.
FIGURE 1 ABOUT HERE
Based on the coverage of the Sudweeks case, we expect that femicide will be framed as
an isolated crime, with an emphasis on human interest and police and court proceedings. Based
on the coverage of the Shepard case, we theorize that news coverage of femicide is unlikely to
bring about policy change, unless it is framed as violence against women, a hate crime and/or a
violation of women’s rights. Moreover, we posit more generally that coverage of femicide and

14
sexual aggression as hate crimes and human rights violations could help to bring about a shift in
public opinion by pressuring law makers to make policy changes to mitigate violence against
women. Yet, we do not expect these changes in news coverage because news media not only
shape public opinion, but are also in part driven by public opinion—especially in developed
democracies with media systems that are relatively free from direct government control, like the
United States. News media in these environments are dependent on advertising revenue and
therefore vulnerable to commercial pressures. For these reasons, news media are unlikely to
cover stories in a way that is out of sync with public attitudes. We propose that for the most part
the public in the US view violence against women as an abstract problem that is more likely to
pertain to others than oneself. Most readers would prefer to read news articles that reassure them
that cases of seemingly random violence against women are isolated instances that happen to
women who find themselves in the wrong place at the wrong time and that cases of domestic
violence are tragic, but private family matters.
Building on these proposals, we derive the following hypotheses:
H1: Stories about femicide will be dominated by human interest and procedural frames
rather than morality/human rights or morality/hate crime frames
H2: Stories about femicide with morality/human rights frames will be more likely to bring
about policy change
H3: Stories about femicide with connection frames linking the crime to violence against
women will be more likely to bring about policy change

15
Research Design and Data
To test our hypotheses about media coverage of violence against women, we needed a
sample that was representative and manageable. To that end, we chose to focus on all available
cases in the state of Massachusetts in 2013. As mentioned earlier, at the time of this writing,
2013 was the most recent year for which complete crime data were available for Massachusetts.
In part, we chose to focus on Massachusetts because it was a convenient case for us. Beyond
convenience, we posit that Massachusetts is a rigorous test for our hypotheses, given its
relatively high median household income of $70,628 in 2015 compared to the national average
($55,775), and the percentage of college educated adults, 54% in 2010, compared to the national
39% (US Department of Education 2012). We would expect news media catering to a wealthier,
more educated audience to be more likely to cover femicide as violence against women.
Moreover, we opted to use the Boston Globe as our news source because it is renowned
nationally for its watchdog coverage, especially regarding human rights. (Take for example the
paper’s investigation of the Catholic Church’s cover-up sex abuses by clergy.) If ever a
newspaper in a state was inclined to cover femicide as a violation of human rights, it would be
the Boston Globe in Massachusetts. We focused on newspaper coverage rather than television
news reports because again, we expected newspaper coverage to be a tougher test since
newspapers tend to be more in-depth and more informative in their coverage (Druckman 2006).
According to FBI reports, in 2013 Massachusetts had 29 killings of women that were
classified as murder, homicide, manslaughter or murder-suicide. We collected all of the stories
published in the Boston Globe mentioning these cases using the archives available through
Proquest. Seven of the cases were not covered by the Boston Globe, but were covered by other
local news sources, so we used alternative sources for these cases. Since femicide was our focus,

16
we eliminated all cases that were not femicides (where the victims were targeted for reasons
other than gender, such as drugs or robbery or the Boston Marathon bombings), as well as all
unsolved cases because the motives were unclear.
7
This left 15 cases, two non-intimate
femicides and 13 intimate (domestic violence) femicides. Table 1 provides details about these
cases.
TABLE 1 ABOUT HERE
To analyze the coverage of these cases, we coded all news articles mentioning the victim,
identifying the dominate frame of each paragraph. Following Entman (2010), we chose to code
at the paragraph level because reporters will often use more than one frame in a given story and
because it helps to account for stories that mention the victim, but focus primarily on a different
issue. For example, a story mentioning trials involving insanity pleas was coded “not applicable”
for paragraphs that did not reference the femicide victim or killer. We coded for the following
frames:
8
• Connection Frame linking crime to violence against women problem/policy
• Connection Frame linking crime to other problem/policy (for example, mental
health)
• Morality featuring a moral prescription related to religion or ethics
• Morality featuring a moral prescription related to human rights
• Morality featuring a moral prescription related to hate crime
7
We would have included unsolved cases involving sexual assault because such violence would
indicate that sexual aggression was a motivating factor; however, there was no indication that
sexual assault was involved in the any of the 2013 unsolved killings. We do concede that we are
probably underestimating the number of femicides.
8
A more detailed description of these frames is available in the online appendix
http://faculty.uml.edu/Jenifer_whittenwoodring/.

17
• Human Interest putting a human face on the killer (or killer’s
friends/family/community)
• Human Interest putting a human face on the Victim (or victim’s
friends/family/community)
• Responsibility of the killer
• Responsibility of the victim
• Responsibility of some other entity (usually police/court/government)
• Procedural (basic procedures of police/court or basic crime description)
• Not Applicable
Two coders coded a total of 4184 paragraphs. To check inter-coder reliability, both coders coded
a subset of randomly selected stories. The Krippendorff’s alpha coefficient was .816, within the
acceptable range.
To get a better idea of the language used for different frames and different cases, we also
generated word clouds in Nvivo. We compared the frames used in the two non-intimate
femicides to those used in intimate femicides. Because the majority of the stories on intimate
femicide that appeared in the Boston Globe (80%) were from the same case—that of Jennifer
Martel, who was engaged to the son of a renowned Red Sox broadcaster—we analyzed this case
in greater detail and found it to be an important exception to the typical coverage of 2013
femicides in the Boston Globe.
Findings
Framing of Femicide

18
The results of the framing analysis are presented in Table 2. As expected, the dominate
frames found in the paragraphs of news coverage of femicide were human interest (35%) and
procedural frames (32%). Less than 1% of the paragraphs featured morality/human rights frames
and only 2.5% featured connection frames linking the crime to violence against women. There
were no morality/hate crime frames. Therefore, we find support for H1: much like the coverage
of Sunny Sudweeks, human interest and procedural frames dominated the coverage femicides in
general.
This was especially true of the two cases of non-intimate femicide in 2013. The killings
of Amy Lord and Colleen Ritzer were both covered extensively in the Boston Globe. Yet, our
framing analysis shows that their murders were covered as isolated crimes rather than violence
against women or hate crimes, even though both were clearly targeted primarily because they
were women.
9
Regarding other frames, overall, about 5% of the paragraphs featured responsibility
frames, with less than 1% attributing responsibility to the victim (generally because of a failure
to press charges), about 2% attributing responsibility to the killer and about 3% attributing
responsibility to police (for the failure to catch the killer earlier), the courts, or government
(mostly for failure to address mental health problems). Religious morality frames made up about
1% of the frames, and these were generally in articles covering the victim’s funeral. Less than
1% of the frames linked the crime to problems other than violence against women (for example,
mental health issues and government failure to regulate “sober” homes). About 22% of the
9
A more detailed analysis of the coverage of these non-intimate femicides is available in the
online appendix.

19
paragraphs in articles mentioning one of the 2013 femicides were coded as “not applicable.”
Usually these were articles focusing on lawyers or judges involved in the court proceedings.
Interestingly, intimate femicides (with and without inclusion of the Jennifer Martel case)
had a statistically significant higher percentage of paragraphs with violence against women
connection frames (4.4%) than non-intimate femicides (.3%). The Jennifer Martel case had a
comparatively high percentage (5.8%) of violence against women connection frames, but almost
no (.1%) morality/human rights frames. At 4.1%, intimate femicides (excluding Martel) had
significantly higher percentage of morality/human rights frames than non-intimate femicides
(which had only .3%). The language used in these violence against women connection frames
and morality human rights frames is depicted in word clouds in Figure 2. Here we see that the
most frequent words in the violence against women connection frames include “domestic
violence,” “victims” and “women.” The most frequent words in the morality/human rights
frames are “violence,” “defendant” and “death.” Because there were so few incidences of these
frames, it is not possible to test H2 and H3, our hypotheses that news coverage with these frames
will be more likely to lead to policy change. Yet, there is some evidence that connection frames
linking the Jennifer Martel case to violence against women helped bring about the passage of a
new law aimed at protecting domestic violence victims in Massachusetts.
The Jennifer Martel Case
We identified 13 cases of confirmed intimate femicide, or domestic violence deaths, three
of which were not covered in the Boston Globe. Of the 121 articles that appeared in the Boston
Globe about these cases, 95 were about Jennifer Martel. The Martel case garnered substantial
news coverage in part because her fiancé/killer was the son of a prominent Red Sox broadcaster
and in part because just one day before her death, her fiancé had been charged with assaulting

20
her and was then released on his own recognizance. Unlike the other intimate femicide cases in
Massachusetts in 2013, the Martel case was vigorously covered. It was also used to shine a
spotlight on the problems with both policies and practices in addressing domestic violence. We
analyzed all 176 references to domestic violence in the Boston Globe’s coverage of the Martel
case and found 6 references to problems with the application of the dangerousness law, which
allows a prosecutor to move to detain a defendant if she believes he represents a substantial
threat—an option that was not used to protect Martel. We also found 27 references to the need
for better policies and practices to prevent domestic violence. Of the 95 articles, 63 were focused
specifically on the Martel case. The remaining 32 articles mentioned the case in connection to
political developments and the call for changes in domestic violence policy that came about after
Martel’s killing (for this reason, 36% of the paragraphs in stories mentioning Martel were coded
“not applicable” in their framing. These mentions do provide evidence that journalists were
connecting Martel’s killing to the larger and more abstract problem of violence against women.
In short, both the articles focusing on Martel’s death and the ones that mentioned it only in
passing were drawing connection frames between the specific case and violence against women.
Additionally, the use of language in the Martel coverage indicates that the main focus of the
coverage was domestic violence and how her killing could have been prevented. The word cloud
on the left in Figure 3 (which was created using only the 63 articles that focused on the Martel
case) indicates that “domestic violence,” “daughter” and “family” were the most frequently used
words in the news articles, followed by “murder,” and “court.” As indicated in Figure 2, the term
“domestic violence” was the most frequent term in connection frames. There is some indication
that this focus made a difference.

21
As one women’s rights advocate put it, “The murder of Jennifer Martel seems to be the
straw that broke the proverbial camel’s back, motivating the speaker and the Legislature to act in
a way that begins to address the systemic issues, and not simply apply a Band-Aid,” (Troop
2013, quoted in Abraham 2014: A1). In fact, the Massachusetts Legislature did pass a bill in
August 2014 that increased penalties for repeat offenders and required employers to provide up
to 15 days of leave to victims and their family members. Interestingly the bill, which was signed
into law by then Governor Deval Patrick, also called for police to shield domestic violence cases
from the publically available police logs until the arraignment of a suspect. Supporters of the
measure claimed it would protect victims’ privacy and encourage them to report domestic
violence, but the executive director of the Massachusetts Newspaper Publishers Association,
Robert Ambrogi, argued that it could also shield those in power. “If a teacher, if a coach, if a
guidance counselor—any number of public officials—is arrested under these circumstances,
that’s something the public has a right to know,” (Ambrogi quoted in Scharfenberg 2014: B1).
FIGURE 3 ABOUT HERE
Thus, news coverage was different in the Martel case because it focused on her killing as
an example of violence against women. The articles detailed Martel’s killer’s long history of
domestic violence and provided evidence that Martel’s death could have been prevented if
prosecutors had used the dangerousness law to detain her killer. In contrast to the argument of
Meyers (1997), in the Martel case, journalists’ focus on the shocking details, including the
offender’s long history of abuse, helped to bring attention to the issue of domestic violence. This
reporting likely contributed to the pressure that prompted lawmakers to develop and pass new
policies aimed at mitigating domestic violence; however, the inclusion of the measure that
prohibited police from including information about domestic violence cases in their logs was

22
viewed by local media as an effort to stifle news coverage of domestic violence allegations
following the 2006 coverage of the arrests of a local sports celebrity and his wife on domestic
violence charges (that were later dropped). Regardless of the measure limiting media access to
information, the Martel case provides limited support for H2, the hypothesis that news coverage
with connection frames linking femicide to violence against women will be more likely to bring
about political change. H3, the hypothesis that femicide coverage with morality/human rights
frames will bring about policy change, was not supported.
News coverage of the other 12 intimate femicides in 2013 was minimal at best, and in
some cases, non-existent in the Boston Globe. The three cases that were not covered the Boston
Globe (but were covered in local newspapers and websites) were murder-suicides. This lack of
coverage is not surprising because news organizations often avoid covering suicides due to
concerns that such coverage could prompt others to commit suicide. Yet, murder-suicides are a
common type of domestic violence and violence against women. The Boston Globe did include
one article on the murder-suicide of Shirley Ju and three articles on the murder of Mei Kum
Jones, whose husband also killed their twin infant sons before killing himself. For the most part
though, these cases were covered as individual incidents rather than part of a broader pattern of
violence against women. Reporters typically used human interest frames (36% of paragraphs),
focusing on personal or family tragedies, and procedural frames (31% of paragraphs), focusing
on police investigations and court proceedings. Yet, they did at times use morality/human rights
frames (about 4% of paragraphs) and connection frames to link these crimes to violence against
women (also about 4% of paragraphs). Generally though, reporters quoted official sources, such
as police and attorneys, with some reactions from family, friends and neighbors. These foci are
evident in the word cloud presented on the right in Figure 3, which was generated using news

23
coverage all intimate femicides (including the murder-suicides) except Martel. The most
frequently used word is “Police” followed by “murdering,” “court” and “family.” Thus, as with
non-intimate femicide, intimate femicide more often than not was covered as isolated crime, with
dominate frames of human interest and police and court procedures.
Conclusion
Since the passage of the Matthew Shepard and James Byrd Hate Crimes Prevention Act
of 2009, hate crimes in the US are defined as crimes in which the offenders are motivated by a
bias against the victim’s religion, national origin, race, color, sexual orientation, gender identity,
disability or gender. Yet femicides, meaning the killings of women because of their gender, are
seldom covered as hate crimes or violations of human rights. In 2013 in Massachusetts, Amy
Lord was killed by a stranger who was motivated by what one investigator described as a
“complete disdain” for women. Colleen Ritzer was raped, violated with a stick and stabbed to
death by a student who wrote a note expressing his hate. Still neither of these crimes was covered
as a hate crime by the Boston Globe, nor were they prosecuted as such. In contrast to Colomb
and Damphousse (2004), who argued that too much media attention to hate crimes can lead to a
moral panic, we propose that media failure to cover violence against women—especially non-
intimate femicides—as hate crimes is part of the reason why there are not better policies in place
to prevent femicide. That said, it is not fair to put all of the blame on media or law enforcement.
The failure to address the problem of femicide is a structural problem deeply rooted in society
(Hudson et al. 2012).
Nevertheless, our findings indicate that covering the killings of women as femicides and
violations of human rights could help to bring about policy changes aimed at preventing violence

24
against women by increasing the perception of government responsibility.
10
If journalists had
focused on the fact that both Amy Lord and Colleen Ritzer were killed primarily because they
were women, they could have drawn attention to violence against women. Advocates for
women’s rights and human rights could encourage journalists to cover non-intimate femicides as
hate crimes, just as gay rights activists did in the case of Matthew Shepard. Similarly, our
findings point to potential benefits of training journalists about the importance of connecting
domestic violence deaths to the larger societal problem and covering femicides in general as
violations of human rights.
In 2013, most cases of domestic violence killings in Massachusetts were covered briefly
at best and were often treated as isolated incidents rather than part of a broader pattern of
violence against women. The one exception was the Jennifer Martel case. Her murder garnered
front page news coverage, drew attention (through journalists’ use of connection frames) to the
larger problem of domestic violence and led to the passage of state legislation aimed at
protecting victims and prevent domestic violence. Perhaps this was because her killer was the
son of a local celebrity or because her death could potentially have been prevented had her killer
not been released just hours before he killed her. The lesson learned is that news coverage of
domestic violence can help, not only to raise public awareness of the problem, but also to bring
about policy changes. Ironically though, the legislation passed in the wake of Martel’s murder
included a measure that limits journalists’ access to information by requiring that domestic
violence cases be kept out of the publically available police logs until and if the suspect is
arraigned.
10
This is in keeping with the findings of Pollock, Richardella, Jahr, Morgan and Cook
(forthcoming) that hate crimes are associated with increased perception of government
responsibility regarding violence against women.

25
We acknowledge that this study is merely a starting point. We cannot be sure that what
held true for Massachusetts in 2013 will hold true for other states and other years, but we do find
indications here that news coverage of femicide matters and that reports framing violence against
women as just that can make a difference. Our study points to the need for more research in this
area. In particular, it would be interesting to study whether there are differences depending on
the gender of the reporter covering a femicide.

26
Tables and Figures
Table 1: Femicides in Massachusetts in 2013
Name/Date
Motive
Most frequent
word(s)
Covered in the Boston Globe
Rebecca Felteau
3/10/13
Domestic Violence/
Intimate Femicide
police
Anita Clark
5/4/13
Domestic Violence/
Intimate Femicide
police
Celestine Walker
5/12/13
Domestic Violence/
Intimate Femicide
court=police=
prosecutors
Melissa Hardy
6/23/13
Domestic Violence/
Intimate Femicide
sober
Tyshianna Atkins
7/4/13
Domestic Violence/
Intimate Femicide
domestic=violence
Amy Lord
7/23/13
Robbery/
Non-Intimate
Femicide
police
Sherry Leigh Bradley
10/1/13
Domestic Violence/
Intimate Femicide
police
Jennifer Martel
10/15/13
Domestic Violence/
Intimate Femicide
domestic=violence
Colleen Ritzer
10/22/13
Rape, Robbery/
Non-Intimate
Femicide
court
Shirley Ju
11/3/13
Domestic Violence/
Intimate Femicide
Murder-Suicide
violence=domestic
Mei Kum Jones
11/18/13
Domestic Violence/
Intimate Femicide
Murder-Suicide
family
Colleen Butler
11/30/13
Domestic Violence/
Intimate Femicide
police
Not Covered in the Boston
Globe
Julie Treadwell
2/8/13
Domestic Violence/
Intimate Femicide
Murder-Suicide
shot
Glomerys Martinez
4/22/13
Domestic Violence/
Intimate Femicide
Murder-Suicide
police
11Anonymous
7/14/13
Domestic Violence/
Intimate Femicide
Murder-Suicide
police
11
name not disclosed at the victim’s family request

27
Table 2: Framing of Femicides in Massachusetts in 2013
Frame Type
Femicide
Type
Connection
Morality
Human Interest
Responsibility
Procedural
N/A
Total N
Not VAW
VAW
Religious/
Ethics
Human
Rights
Killer
Victim
Other
Killer
Victim
Non-
Intimate
20
.8%
6
.3%
50
2%
8
.3%
471
20%
434
18%
68
3%
58
2%
0
0%
916
38%
373
16%
2404
Intimate
9
.5%
98
5.5%
9
.5%
17
1%
331
18.6%
235
13.2%
76
4.3%
28
1.6%
9
.5%
417
23.4%
551
31%
1780
Intimate
w/o Martel
6
1.5%
17
4.4%
3
.8%
16
4.1%
46
11.8%
96
24.6%
9
2.3%
21
5.4%
0
0%
120
30.7%
57
14.6%
391
Martel only
3
.2%
81
5.8%
6
.4%
1
.1%
285
20.5%
139
10%
67
4.8%
7
.5%
9
.7%
297
21.4%
494
35.6%
1389
Total N
29
104
59
25
802
669
144
86
9
1333
924
4184
%
.7%
2.5%
1.4%
.6%
19.2%
16%
3.4%
2.1%
.2%
31.9%
22.1%
100%

28
Figure 1: A Tale of Two Killings
Sunny Sudweeks: Violence against women as isolated crime Matthew Shepard: Violence against gays as hate crime

29
Figure 2: Language Used in Frames linking Femicide to Violence Against Women and Human Rights
Connection Frame to Violence Against Women Morality/Human Rights Frame

30
Figure 3: Coverage of Intimate Femicide
Jennifer Martel Coverage of Intimate Femicide Excluding Martel

31
References
Abraham, Yvonne. "A Speaker Galvanized by the Deadly Cycle of Abuse." Boston Globe Mar 27
2014. ProQuest. Web. 31 May 2017: A1.
Below, Amy and Jenifer Whitten-Woodring, “Global Warming Heats Up: Climate Change, Connection
Frames and the US Policy Agenda.” American Political Science Association Annual Meeting,
Boston, August 2008.
Bennett, W. Lance (1990) Toward a theory of press-state relations. Journal of Communication. 40 (2):
103-125.
Bhattacharjee, A., & Banda, S. (2016). Domestic violence against women: The socioeconomic
scenario. Indian Journal of Health & Wellbeing, 7(7), 765-768.
Boston Globe Editorial. 2013. "A Preventable Attack?" Boston Globe. Jul 30 2013. ProQuest. Web. 31
May 2017: A10.
Brandle, S. M. (2016). Tele vision news a nd human rights in the US & UK: The violations will not be
televised. Routeledge: New York.
Brandle, S.M. (forthcoming). “Media Coverage of Human Rights in the US and UK: The violations still
won’t be televised (or Published)”
Bridges, F.S., Tatum, K. M., & Kunselman, J.C. (2008). Domestic violence statutes and rates of
intimate partner and family homicide: A research note. Criminal Justice Policy Review, 19(1),
117-130
Clark, R. (2012). Bringing the media in: Newspaper readership and human rights. Sociological Inquiry,
82(4), 532-556. doi:10.1111/j.1475-682X.2012.00417.x

32
Colomb, W., & Damphousse, K. (2004). EXAMINATION OF NEWSPAPER COVERAGE OF HATE CRIMES:
A moral panic perspective. American Journal of Criminal Justice, 28(2), 147-163.
Della Giustina, J.,. (2010). Why women are beaten and killed: Sociological predictors of femicide.
Druckman, J. N. “Media Matter: How Newspapers and Television News Cover Campaigns and Influence
Voters.” Political Communication, 22(4), 463-481.
Easteal, P., Judd, K., & Holland, K. (2015). Enduring themes and silences in media portrayals of
violence against women. Women's Studies International Forum, 48, 103-113.
doi:10.1016/j.wsif.2014.10.015
Entman, R. M. (1993). “Framing: Toward Clarification of Fractured Paradigm.” Journal of
Communication 43(4), 51-58.
Entman, Robert M. (2010). “Media Framing Biases and Political Power: Explaining Slant in News of
Campaign 2008.” Journalism 11(4), 389-408.
Fausto, Alma (2017). “Police ID suspect: Former Santa Ana resident is accused of killing OCC student
in ’97.” Orange County Register. March 3, 2017: Cover.
Gruenewald, J., Pizarro, J., & Chermak, S. M. (2009). Race, gender, and the newsworthiness of
homicide incidents. Journal of Criminal Justice, 37(3), 262-272.
doi:10.1016/j.jcrimjus.2009.04.006
Haydari, N., & Kara, M. (2015). A collaborative media production project on human rights: Bridging
everyday and media education. Innovations in Education & Teaching International, 52(4), 383-
392. doi:10.1080/14703297.2013.861759
Heinze, E., & Freedman, R. (2010). Public awareness of human rights: Distortions in the mass media.
International Journal of Human Rights, 14(4), 491-523. doi:10.1080/13642980802645804
Hudson,Valerie M.,,. (2012). Sex and world peace. New York: Columbia University Press.

33
Humphries, D. (2009). Women, violence, and the media. Hanover: Northeastern University Press.
Investigation Discovery (2017). See No Evil: Watching Amy Lord. Season 02, Episode 01.
Lundman, R. J. (2003). The newsworthiness and selection bias in news about murder: Comparative
and relative effects of novelty and race and gender typifications on newspaper coverage of
homicide. Sociological Forum, 18(3), 357.
Lundman, R. J., Douglass, O. M., & Hanson, J. M. (2004). News about Murder in an African American
Newspaper: Effects of relative frequency and race and gender typifications. Sociological
Quarterly, 45(2), 249-272.
McCabe, Kathy. "Speliotis on School Safety Panel." Boston GlobeJan 26 2014. ProQuest. Web. 31 May
2017.
Petrosky E, Blair JM, Betz CJ, Fowler KA, Jack SP, Lyons BH. Racial and Ethnic Differences in
Homicides of Adult Women and the Role of Intimate Partner Violence — United States, 2003–
2014. MMWR Morb Mortal Wkly Rep 2017;66:741–746. DOI:
http://dx.doi.org/10.15585/mmwr.mm6628a1.
Pollock, Richardella, Jahr, Morgan and Cook (forthcoming) “Nationwide Newspaper Coverage of Rape
and Rape Culture on College Campuses: Testing Community Structure Theory.”
Powers M. (2016) “A new era of human rights news? Contrasting two paradigms of human rights
news-making.” Journal of Human Rights 15(3): 314-329.
Radford, Jill,, Russell,Diana E.H.,. (1992). Femicide: The politics of woman killing. New York; Toronto;
New York: Twayne ; Maxwell Macmillan Canada ; Maxwell Macmillan International.
Reilly, Janet E. forthcoming. “Training Human Rights Reporters: Human rights education in U.S.
journalism degree programs.”

34
Rhoades, Brady (1998). “A mother’s torment continues over daughter’s unsolved murder; CRIME: A
year after Sunny Sudweeks’ slaying, police say there are still no suspects.” Orange Country
Register. February 23, 1998: B01
Richards, David L., Haglund, Jillienne,,. (2015). Violence against women and the law.
Scharfenberg, David. (2014) “Abuse measure wins approval.” Boston Globe, August 1, 2014: B1.
Proquest May 30, 2017.
Shackelford, T. K. (2001). Cohabitation, marriage, and murder: Woman-killing by male romantic
partners. Aggressive Behavior, 27(4), 284-291.
Sorenson, S. B., & Peterson, J. G. (1998). News media coverage and the epidemiology of homicide.
American Journal of Public Health, 88(10), 1510-1514.
Tarshis, L. (1992). The war on women. (cover story). Scholastic Update, 124(14), 14.
UN Women. (2017). Facts and figures: Ending violence against women. Retrieved from
http://www.unwomen.org/en/what-we-do/ending-violence-against-women/facts-and-figures
United Nations. (2017). Declaration on the elimination of violence against women. Retrieved from
http://www.un.org/documents/ga/res/48/a48r104.htm
US Department of Education. (2012). “New State-by-State College Attainment Numbers Show
Progress Toward 2020 Goal.” Available at: https://www.ed.gov/news/press-releases/new-state-
state-college-attainment-numbers-show-progress-toward-2020-goal. Accessed on July 14, 2017.
Violence Policy Center. (2015). When men murder women: An analysis of 2013 homicide data.
Retrieved from http://www.vpc.org/studies/wmmw2015.pdf
Whitten-Woodring, J. (2016). “News about Her: The Effects of Media Freedom and Internet Access on
Women’s Rights.” Journal of Human Rights, 15(3): 383-407.

35
Winston, M. E., & Pollock, J. C. (2016). Human rights in the news: Balancing new media participation
with the authority of journalists and human rights professionals. Journal of Human Rights, 15(3),
307-313. doi:10.1080/14754835.2016.1153413
Wood, J. T.,. (1999). Gendered lives: Communication, gender, and culture. Belmont, CA: Wadsworth
Pub.
World Health Organization. (2012). Understanding and addressing violence against women. Retrieved
from http://apps.who.int/iris/bitstream/10665/77421/1/WHO_RHR_12.38_eng.pdf
Young, C. (1999). Gender war crimes. Reason, 30(8), 55.