Patriarchy v. Functional Truth:
Assessing the Feminist Critique of Intimate Violence Reporting
Director, Grade the News
Stanford, CA 94305-2050
Tel. (650) 725-7092
[log in to unmask]
Director, Berkeley Media Studies Group
Submitted to the Media and Family Competition, MC & S Division
Please address all correspondence to the first author.
This research was supported by The California Wellness Foundation as part
of its Violence Prevention Initiative.
Patriarchy v. Functional Truth:
Assessing the Feminist Critique of Intimate Violence Reporting
Journalism assumes reporters are able to pursue "functional truth"—an
account of issues and events reliably describing social reality. Critical
feminist scholars, however, contend that journalists working in
male-dominated corporations are constrained by a culture of patriarchal
values. The present study is the first in the U.S. to test this critique as
it applies to reporting the vast social pathology of intimate partner
violence. Contrary to that critique, newspapers very rarely blamed female
battering victims or mitigated suspect blame. However, intimate violence
was covered much less often and with less depth than other violence.
Patriarchy v. Functional Truth:
Assessing the Feminist Critique of Intimate Violence Reporting
In their distillation of the essence of journalism at the beginning of the
21st century, Kovach and Rosenstiel posit the pursuit of truth as the
This is what journalism is after_a practical or functional form of truth.
It is not truth in the absolute or philosophical sense. It is not the truth
of a chemical equation. But journalism can_and must_pursue truth in a sense
by which we can operate day to day. 
Criticial feminist scholars aren't buying it, especially when it comes to
reporting about issues particularly affecting women, such as violence
against intimate partners. Some argue that journalists do not_and others
that they cannot_pursue the truth because they are caught up in a culture
of patriarchal values that reflect, reinforce and legitimize viewpoints
that privilege men over women. In News Coverage of Violence Against
Women: Engendering Blame, Meyers wrote:
News scholars … note that the values, norms and conventions that shape the
news support the status quo by representing the interests of a white,
middle- and upper-class, male elite.
The result when applied to news of violence against women, she asserted,
"positions the female victim as deviant and deserving of condemnation if
she in any way appears to have disregarded or flouted socially approved
gender roles and expectations" [italics in the original].
The present study tests the core assumption of journalism that reporters
and editors can discern and provide "functional truth," against the
competing idea that journalists are unable, unwilling, or too unaware to
provide an account of reality free of patriarchal values. The field of
contest is coverage of the most common type of violence against
women_intimate partner violence. It's a vast public health problem that in
the U.S. is the most frequent cause of non-fatal injury to women and
kills an average of three women every day. Given that news is a primary
resource citizens use to make sense of important issues of the day, if
journalists are warping coverage of intimate violence as feminist
scholars charge, efforts to develop social solutions to battering may be
stymied. How this topic is portrayed matters both as a public health issue
and as a challenge to the fundamental professional standard of journalism.
Functional truth: By the early 1980s, Hackett observed, the journalistic
paradigm of objectivity was spent: "No longer can we simply assume the
possibility of unbiased communication, of objective and detached reporting
on an allegedly external social and political world." Schudson
substituted "mature subjectivity" for the unreachable standard of
objectivity. He argued for reporting,
firmly grounded in the process of screening tips, assembling and weighing
evidence, fitting facts, and attempting to disconfirm the resulting story.
In the end, these exercises yield a degree of 'moral certainty' about the
convergence of facts into a truthful report.
Similarly, Ettema urged a "practical wisdom," an effort to justify fact
claims with as much evidence as practicable. Allan described three
streams of feminist research about news, the most traditional of which
holds that both male and female reporters can be sensitized to sexist
stereotypes and learn to exclude or counter them.
For the purposes of this study, "functional truth" means two things,
reporting that: 1) avoids or challenges sexist stereotypes about intimate
violence against women; and 2) treats intimate violence no differently than
other crimes of violence.
Patriarchal biases in reporting about violence against women: Most studies
of news accounts of violence against women concern rape rather than the far
broader problem of violence between intimate partners. Benedict, for
example, reported that a rape victim was likely to be blamed not only if
she stepped outside the role of housewife, but if no weapon were used, she
knew the assailant, was young, or pretty. Quantitative studies of
sexist stereotyping of violence reporting are scarce, contradictory and
dated. A 1984 study of sexual assault reporting in three Toronto newspapers
found that almost half of these stories carried attributions of fault
directed at the female victim. Only one-in-six articles blamed the male
perpetrator. The only other study examined assignment of responsibility
across all types of crime in 11 Canadian dailies in the early 80s. It
identified victim responsibility in about one article in four. But male
victims were more likely to be blamed than female for all crimes of
When narrowed to analyses of intimate partner violence, the literature is
more recent, but scant. It also turns away from quantitative measures to
discourse analysis of individual stories or stories selected purposively,
and it employs feminist critiques. To the existing scholarship we've
added observations about press coverage from intimate violence experts who
work in the counties where we conducted our study. Three primary
First, critics accuse the press of failing to take intimate violence as
seriously as other violent crimes. For example, Santa Clara County
(California) Deputy District Attorney Rolanda Pierre-Dixon who has
specialized in domestic violence cases for 20 years said, "You don't see
informative [stories] about domestic violence year-long; [reporters] only
react to homicides." Billie Weiss, executive director of the Injury and
Violence Prevention Program in the Los Angeles County Department of Health
Services, complained that even intimate homicides don't necessarily rate
coverage in the Los Angeles press.
Second, scholars charge that violence against women is the only crime in
which the victim is blamed. If a female victim of violence is a child or
elderly, she may be described as a true victim, noted Meyers. But in
between those ages "chances are she will be represented as somehow
responsible for her own suffering because she was on drugs, drunk, not
properly cautious, stupid, engaged in questionable activities, or involved
in work or exhibiting behavior outside the traditional role of women."
Third, more than in any other crime, reporters mitigate or obscure male
perpetrators' guilt. Finn described how some men escape blame: White,
middle-class men who batter their wives are "constructed in state and media
discourses as the victims of provocation or personal stress, more deserving
of mercy and compassion than condemnation and constraint." Lamb and
Keon found newspaper reports of domestic violence often used the passive
voice. "The words 'they were beaten' or 'they were subjected to abuse' are
ambiguous and direct the focus of the reader away from the male
perpetrator. They also make the victim the linguistic center of the
sentence." Thus the very language used to describe intimate partner
violence shields the perpetrator from responsibility.
The literature suggests three research questions highlighting a conflict
between "functional truth" and patriarchal values and stereotypes in news:
• RQ1: Is intimate violence covered less frequently, less representatively,
and with less depth than other kinds of crime?
• RQ2: How often does intimate violence reporting blame female victims for
the violence inflicted on them?
• RQ3: How often does intimate violence reporting mitigate or deflect blame
from male suspected perpetrators?
We chose to answer these questions by analyzing how violence is reported
at a regionally and a nationally prominent newspaper, respectively the San
Jose Mercury News and the Los Angeles Times. In addition to their
influence on journalistic practice, these papers were selected because they
are located in counties with domestic violence death review teams, which
facilitated our analysis. We concentrated on just two newspapers because in
order to detect even rare negative stereotypes we would have to analyze
every article describing intimate violence over an extended time—in this
case one year. Also, to discover whether those stereotypes were challenged
when they appeared would require a secondary qualitative analysis. Finally,
to determine whether such violence was reported differently than other
kinds of violence, we needed to analyze a second simultaneous sample from
the same papers.
Using computer databases we identified more than 5,200 articles about
criminal violence occurring in the U.S. and published in 2000. We
printed and analyzed every story about intimate partner violence (488
articles). We also gathered a random 1-in-9 sample of all other
violence stories. That sample, comprising 529 articles, was stratified
chronologically over the 12 months.
Answering the Research Questions. We answered RQ1 about inferior coverage
of intimate violence with three measures. First we determined whether
intimate violence gets press attention commensurate with its frequency as a
felony arrest. We calculated the number of intimate violence stories
per felony arrest for that crime in Santa Clara County (the primary
circulation area of the Mercury News) and in Los Angeles County (the
primary circulation area of the Times) and compared those ratios with the
number of stories per felony arrest for other violence. We included all
stories about these violent incidents that occurred in the newspapers' home
counties as well as any thematic stories originated by the papers' own staffs.
Prior research has established that crime reporting over-represents
homicide. So we asked not just whether intimate violence is portrayed
as more fatal than it is, but whether coverage is more murder-oriented than
reporting on other violence. To compare the "murder-centricity" of intimate
with other violence, we calculated the ratio of stories about intimate
murders to all articles about violence between intimates and compared it to
the ratio of such homicides to all felony arrests for intimate
violence. We calculated similar ratios for the other violence sample
using non-intimate violence crime data.
Finally, we examined whether intimate violence is covered as often as an
issue as other kinds of violence. Researchers have found thematic
reporting—stories that look at issues or patterns of events rather than
focusing on a particular episode or incident—is often substantially more
helpful to readers as a resource for making sense of their environment.
Episodic reporting, on the other hand, may lose sight of the forest for the
trees. So we counted the percentage of all stories that were thematic for
both intimate and other violence samples.
To answer RQs 2 and 3—blaming the victim and deflecting blame from the
perpetrator—we analyzed frames. A frame is a theme or sometimes just a cue
that activates a scenario in the minds of some readers. Frames are
powerful because they aid certain interpretations and hinder others_often
without the reader's awareness. Frames create tracks for a train of thought.
In both the intimate violence and other violence samples, we looked for
causal content frames—including necessary, sufficient or contributory
causes. For the intimate violence articles alone, we conducted a second,
more qualitative analysis. The first author re-read the stories in which we
found causal frames to determine whether they explicitly or implicitly
assigned responsibility. He also examined the gender of the victim and
whether the frame was challenged or contradicted—which might lessen its
Most frame analyses avoid implied frames because they are less obvious to
coders and thus less likely to meet criteria for reliability. But we
included them because the criticisms we tested came from discourse analyses
that would be sensitive to implication; we couldn't assess them fairly
using only manifest indicators. We also think the decisions—conscious and
otherwise—that reporters and editors make to include certain information
and exclude other elements sometimes reveal a subtle causal logic. News is
a terse and "objective" genre in which extraneous content is stripped away.
Why, for example, include a comment about a woman's remaining with a
violent man in a story of her battering if journalists don't judge it
relevant to the violence?
Counting implied frames also increases the likelihood of accepting the
presence of a frame that's not really there in order to avoid missing one
that is. In other words, our results are more likely to exaggerate than
understate the presence of victim blame and suspect mitigation. Despite
these tradeoffs, this research design allows us to generalize beyond a few
instances without sacrificing nuance.
Re RQ1: Intimate partner violence coverage was covered less frequently,
less representatively, and with less depth than other kinds of crime.
Overall coverage of intimate violence was much less frequent. In 2000,
the Mercury News reported 17 thematic stories about intimate violence plus
38 episodic articles about such violence in Santa Clara County for a total
of 55. There were 2,450 arrests for this crime in the county in 2000,
so the story per arrest ratio is .0224. Our sample of stories about other
kinds of violence in the Mercury News contained 24 thematic stories plus 35
episodic stories describing such violence in the county, for a total of 59.
Since each of the other violence stories represents 9 articles in the
paper, we estimate that the paper reported 531 stories about other violence
that occurred in Santa Clara County. There were 2,952 arrests for other
types of violence, which yields a story per arrest ratio of .1799. That's 8
times larger than the ratio for intimate violence stories. Even if you
account for the largest margin of error around the ratio for other
violence, the gap would only shrink to 6.4 times as many other violence as
intimate violence stories.
The Times staff wrote 21 thematic stories about intimate violence in 2000
plus 82 episodics describing incidents in the LA County for a total of 103.
There were 14,706 intimate violence arrests in the county, so the ratio
of story per arrest is .007. For the other violence sample, the Times
produced 105 thematics and 116 episodics for a sample of 221. Multiplied by
9, the total comes to 1,989 stories. There were 25,734 other violence
felony arrests in the county, so the story per arrest ratio is .0773.
That's 11 times larger than the intimate violence ratio. The maximum margin
of error would only reduce the ratio to 10.2 times as many other as
intimate violence stories.
Intimate violence coverage was even more murder-centric than reporting on
other violence. While 63% of all Mercury News stories about intimate
violence in Santa Clara County described murders, the ratio of such murders
per felony arrest for intimate violence was just 2.4 for every 1000.
Thus for intimate violence, the proportion of murder stories to all
intimate violence articles exaggerates the ratio of murders to felony
arrests by .63/.0024 or 263 times. The same analysis for other violence
stories shows that 37% of all Mercury News articles concerned homicide.
County records show 24 murders for other violence compared to 2,952 arrests
for a ratio of 9.5 homicides per thousand felony arrests. So the
proportion of murder stories to all other violence articles overstates the
ratio of murders to felony arrests by .37/.0095 or 39 times. Therefore,
coverage of IPV emphasizes murder 263/39, or almost 7 times more than
coverage of other violence. Even at the greatest margin of error for the
other violence sample, coverage of IPV is about 5 times more murder-centric.
Similarly at the Times, the 62% of all LA County intimate violence stories
concerning murder compares to a ratio of 3.3 murders per thousand felony
arrests for intimate violence. Dividing .62 by .0033, we see that
murders between intimates were exaggerated over violence short of murder
by a factor of 188. For other violence, 61% of all stories concerned
murder. There were 951 murders not involving intimate partners in the
county in 2000 compared to 25,734 felony arrests for such non-intimate
violence, a rate of .037. So the portrayal of murder exaggerates its
incidence by .61/.037, about 17-fold. Thus, intimate violence coverage is
188/17 or 11 times more murder-oriented than other violence reporting. Even
at the greatest margin of error for the other violence sample, intimate
violence coverage is 10 times more murder-centric.
Intimate violence is less likely to be covered as an issue. When we counted
all violence articles in the papers, including wire service stories, only
one-in-eight intimate violence stories rated treatment as an issue, trend
or theme rather than as a simple description of a particular violent
episode. In contrast, a third of all stories about other violence was
Re RQ2: Intimate violence reporting rarely blames a female victim. To
answer RQ2, we examined intimate violence stories for 11 frames found in
the literature or suggested by battered women's advocates. To establish
whether victims were also blamed in coverage of other kinds of violence, we
also created a set of 9 parallel frames for that sample. Table 1 shows the
frames previous research suggested would be most common and the percentage
of intimate violence stories in which the frame appeared.
[Table 1 goes about here]
After conducting the secondary qualitative analysis, the pattern of blaming
a female victim for intimate violence that we expected from prior research
grew fainter still. Almost half of the victims blamed turned out to be
male. We also noticed that female victim-blame statements were sometimes
contested in the same article, presumably attenuating the strength of the
Take the frame of the victim being unfaithful, for example. Of 11 stories
in which it was coded, this frame was explicit in just 4. Contrary to
previous research, the frame was applied almost as often_in 5 of the 11
articles_to a male victim of a female assailant as to the more usual case
of a woman victim of a man's assault. When the context surrounding the
frame is considered, women victims were blamed for violence because they
saw other men in about 1 story in 100 and usually by implication.
The most frequent particular frame was staying with a violent
partner. Although this frame appeared in 24 stories, it was explicit in
only 7. In 3 of the 7 explicit uses, the frame was challenged. For example,
in a Times column in which a student blames her girlfriends who continue to
date abusive boys, the columnist then quotes the teacher responding: "I
won't accept that anyone likes getting beat up". Rather than becoming a
reinforcement of a stereotype, it became an opportunity to undermine its
This frame, in implied form, was more common than any frame blaming the
victim for sexual reasons. According to Mercury News reporter Michelle
Guido, who consulted on this research, blaming the victim for staying has a
commonsense ring to it. "People have a hard time sympathizing with victims
of domestic violence. Their visceral reaction is 'get out!'" But in
reality, the frame is often naïve. Women are at greatest risk when they
attempt to leave and just after leaving an abusive relationship.
Further, the frame assumes the battered woman and any children have a safe
place to go and the resources to live independently. It also takes for
granted that she has the self-confidence to get out. Physical abuse is
almost always accompanied by belittling psychological abuse.
Overall we turned up little victim blame directed at women. Even if we
aggregate such frames and add up all of the articles in which blame was
explicit, uncontested and directed against female victims, they constitute
fewer than 4% of the 488 analyzed.
In coverage of other kinds of crime, victims are blamed even less. In our
parallel analysis of stories describing other kinds of violent crime, we
looked for victim blame frames such as victims entering dangerous places,
flaunting wealth, failing to heed warning signs of danger, becoming
impaired through drinking or drug use, failing to cooperate with police,
appearing vulnerable, etc. Only one registered in more than 2% of the
articles sampled. The "victim provoked the violence with a physical
attack" frame appeared in 8.5% of stories. The relative frequency of this
frame, however, may not indicate negative stereotyping in the same way it
might in the intimate violence sample. We would expect this situation to be
more common in violence other than between intimates because of the absence
of the strength differential between men and women in the typical intimate
relationship. Though a man who fights back when attached by another man is
considered more justified in American culture than one who retaliates
against his wife or girlfriend, this frame does not question a violent
response in general.
In conclusion, our analysis does not support the implication in the
literature that blaming women for domestic violence is common; at these
papers it was rare. On the other hand, victim-blame in other kinds of
violence reporting was almost non-existent.
Re RQ3: Intimate violence reporting rarely deflects responsibility from the
batterer. First we examined the finding in the literature that the
suspect's identity was often obscured and the focus shifted to the victim.
Second, we looked for frames that let the perpetrator off the hook, either
partly or completely. Last, we tested the claim that the couple is blamed
for the husband's abuse. To be fair to journalists, we also looked for
frames that held the suspect's feet to the fire_by attributing the violence
to the suspect's jealous, controlling behavior, for instance, or by
rejecting blame-shifting, e.g., "there's no excuse for domestic violence."
Batterers' identities are not obscured. We measured whether suspects were
identified, and how much vis-a-vis victims. We looked for identifiers such
as name, age, residence, occupation, education level, achievements, etc.
and computed one total for the suspect and another for the victim. Far from
being obscured, we found that reporters consistently named intimate
violence suspects and identified them more completely (averaging 3.17
attributes) than victims (averaging 1.93). The practice of reporting more
fully on suspects than their victims was not unique to intimate violence,
however. Suspects were better identified in reports of other kinds of
violence as well.
Perpetrators are rarely excused. Here we looked for frames mitigating the
perpetrator's responsibility. Table 2 summarizes our findings.
[Table 2 goes about here]
Again, when we looked more closely at the context of these frames, we saw
they were often challenged. Take the "perpetrator snapped" frame, for
example. It was usually explicit (in 15 articles of 21). But in all but 3
of the explicit stories, the frame was contested. In addition, 3 articles
contained the reverse of the frame. An example of this opposite frame was
found in a Mercury News book review: "Murray makes it clear that abuse is
not a natural reaction; it's deliberate. 'His hand doesn't just jump out of
his pocket and slap her in the face. They don't act out of anger, but out
of a need for dominance and control.'"
The most common mitigation frame, that both partners bear some
responsibility, was usually implied (31 of 47 articles). In the 16 stories
with explicit uses, the preponderance of blame was levied on men in 11, on
women in 2 and on both parties equally in 3. Blame usually took the shape
of criminal charges. For example, although San Diego Padres outfielder Al
Martin was described as having "exchanged punches" with a woman who
claimed to be his wife, only he was charged with assault. In the two
cases where women were blamed, police were investigating them for murders.
If we sum all explicit uses of these four mitigation frames in which the
batterer is male, we find the expectation of the research literature met in
only 14 articles of 488, about 3%.
In addition to mitigation frames, we also saw outright innocent pleas. As
trial coverage comprised a large part of both papers' reporting of intimate
violence, this frame was prominent_appearing in 120 stories, or 24.6% of
all such articles. Most were connected to a trial. Whether the invocation
of this frame represents an attempt to deflect blame from intimate violence
suspects depends on whether it occurs more often in these stories than in
reporting of other types of violent crime. As we shall see, it doesn't.
Batterers are blamed. We found several counter-stereotypical frames to be
about as common as suspect-mitigation themes. The two most frequent were:
"The suspect acted violently at least partly because s/he was battered
previously" (22 of 488 stories, or 4.5%); and "The suspect may have acted
at least partly out of a desire to control/dominate/bully/intimidate a
partner" (53 stories or 10.9%). The controlling behavior frame was most
often explicit—described as a cause of the violence in 40 of 53 stories—and
overwhelmingly directed toward male perpetrators.
Mitigating factors are rarer still in coverage of other kinds of crime. Our
analysis of stories describing other kinds of criminal violence turned up
four parallel frames of interest. The possible mitigation frame of
"impairment" due to alcohol or drug consumption showed up in 4.2% of
articles—almost as often in other violence stories as in those describing
intimate violence. That squares with research findings showing that
alcohol use accompanies both domestic and other kinds of
violence. However, the "snapped" or "acting out of character" frame
was very rare, appearing in 1.5% of other violence stories. That's much
less than in intimate violence stories even when sampling error is
accounted for. The difference suggests that the stereotype uniquely
attaches to intimate violence. But the frequent challenge of this frame
demonstrates that journalists are not merely transcribing it verbatim from
The "bullying/controlling" frame appears in 4% of the other violence
stories—less than half as often in other violence stories as in intimate
violence articles. But one would expect this frame to be less frequent
given the wider range of criminal situations in the other violence sample.
The "innocence" claim appears slightly more frequently in other
violence stories than in intimate violence articles, even when margin of
error is considered. Were reporters using the frame to shield batterers
we'd expect it to be more common in intimate violence coverage.
In conclusion, we find little support for the idea that male suspects are
often excused or their responsibility diminished in intimate violence
stories published in the newspapers studied.
Two out of three falls in this contest between the critical feminist
critique of patriarchal values in news content and the journalism norm of
functional truth go to the latter. Patriarchal stereotypes can be weeded
out or challenged by journalists, even those working for large,
publicly-traded corporate news media whose boards of directors are
dominated by elite white males. In combating these stereotypes,
functional truth triumphed over patriarchy. This is a narrow conclusion,
however. We found only that the cultures of these two newsrooms allowed
reporters to avoid disseminating a particular set of patriarchal cliches.
In fact, the failure to take intimate violence as seriously as other
violent felonies, suggests that the locus of patriarchal influence may have
shifted from identifiable stereotypes in stories to harder-to-detect news
selection strategies. A feminist scholar, Nancy Berns, interpreted our data
Intimate violence is still not readily viewed as crime. More likely
journalists may view it as a 'woman's issue' along with topics such as
health and child care. These types of issues are covered less than crime.
A content analysis, of course, can't tell us why intimate partner violence
receives inferior coverage, and other explanations are plausible. But
whatever the reason, paying scant attention to so common a crime as
intimate violence reinforces the notion that domestic violence is a private
rather than public problem. As van Zoonen observed: "The definition of the
issue as a matter of the private sphere prevent[s] its recognition as a
social problem and [leaves] the women affected without means to talk about
and fight against it."
From a public health point of view, the finding that journalists can—and
did—avoid sexist stereotypes is hopeful. But the relative lack of thematic
coverage of intimate violence indicates that it's rarely the topic of
enterprise reporting—stories journalists originate to answer the public's
broad questions about current issues and events_or policy/government
reporting, which is almost always issue-oriented. Enterprise journalism
allows journalists to investigate and explain the news rather than react to
events on deadline. The preponderance of episodic reporting signals a
passive, low-cost approach to journalism that yields few resources the
public can use to make sense of such crimes and construct solutions.
As importantly, if the only newsworthy intimate violence is homicide,
journalists are exposing just the tip of an enormous and threatening social
pathology. Intimate partner violence ends in murder less frequently than
other types of violent crime. Yet in many American cities, police make
more arrests and answer more 911 calls about domestic violence than any
other kind of violent offense. Beyond the toll on millions of victims'
physical and mental health, intimate violence also correlates with child
abuse. And because children are often present when spouses are abused,
young ones become 7 to 15 times more likely than non-exposed children to
abuse partners when they become adults. Treating intimate violence as
less worthy of coverage than other kinds of felony violence constitutes a
serious failure on the part of journalists to provide a representative
picture of the world around us.
While the theoretical finding that large male-dominated corporations can
avoid transmitting at least some sexist stereotypes may generalize to
similar news media, the descriptive finding that they did may not. Our data
are not drawn from typical newspapers. Columbia Journalism Review named the
Times as the nation's fourth best newspaper and the Mercury News eighth
best. Further, with their domestic violence death review committees,
Santa Clara and Los Angeles counties may be more progressive in coping with
intimate violence than other parts of the nation. Research covering a wider
sample of media—and across a number of issues—is needed.
 We thank Michelle Guido, a police reporter at the San Jose Mercury
News, who consulted on this study, and Visiting Assistant Professor Nancy
Berns of Drake University who helped interpret our findings. We thank
research assistants Elena O. Lingas, MPH, Jennifer Carlat, Saleena Gupte,
MPH and Karen White. We are also grateful to the Mercury News for pro bono
access to their electronic news archive. This research was funded in part
by The California Wellness Foundation.
 Bill Kovach and Tom Rosenstiel, The Elements of Journalism: What
Newspeople Should Know and the Public Should Expect, (New York: Crown,
2001) p. 42.
 Lana Rakow and Kimberlie Kranich, for example, argued that television
news discourse constitutes a "masculine narrative form," in which women are
not subjects but 'signs' representing femininity or sex, and rarely
portrayed in instrumental roles as experts, attorneys, police officers,
etc. ("Women as Sign in Television News," Journal of Communication 41(1),
1991, 8-23). Others, such as British scholar Paula Skidmore, depicted "a
macho culture of newsgathering_aggressive and domineering, but also one of
male camaraderie and 'bonding'_which excludes women" ("Gender and the
Agenda: News Reporting of Child Sexual Abuse," in Cynthia Carter, Gill
Branston and Stuart Allan, ed. News, Gender and Power (New York: Routledge,
1998) 204-218. Liesbet van Zoonen described a linear model of news
production in which male senders process information introducing gendered
(and other) distortions creating a message containing stereotypes that
socializes news consumers thereby inculcating sexist attitudes, in Feminist
Media Studies London: Sage, 1994).
 Marian Meyers, News Coverage of Violence Against Women: Engendering
Blame (Thousand Oaks, CA: Sage, 1997), 19.
 Meyers, News Coverage of Violence Against Women: Engendering Blame, 24.
 Meyers defines patriarchy as "the systematic institutionalization of
women's inequality within social, political, economic, and cultural
structures," in News Coverage of Violence Against Women: Engendering Blame, 3.
 Joshua R.. Vest, Tegan K. Catlin, John J. Chen and Ross C. Brownson,
"Multistate Analysis of Factors Associated with Intimate Partner Violence,"
American Journal of Preventive Medicine 22(3) (2002) 156-164.
 Callie Marie Rennison, "Special Report, Intimate Partner Violence and
Age of Victim, 1993-99," (U.S. Bureau of Justice Statistics, Washington,
DC, October 2001).
 Doris A. Graber, Processing the News: How People Tame the Information
Tide (White Plains, NY: Longman, 1984).
 For parsimony we will use intimate violence as a synonym of intimate
 Robert A. Hackett, "Decline of a Paradigm? Bias and Objectivity in
News Media Studies" Critical Studies in Mass Communication, 1(3) (1984), 253.
 Michael Schudson, Discovering the News (New York: Basic Books,
 James S. Ettema, "Journalism in the 'Post-Factual Age'" Critical
Studies in Mass Communication, 4(1) (1987): 82-86.
 Stuart Allan, "(En)Gendering the Truth: Politics of News Discourse"
in News, Gender and Power, 121-137. The other two types of feminist
critiques would require greater change in the newsroom: the first holds
that "only women are justified in speaking for women as a social group,"
thus newsrooms should be balanced by gender and masculine values
counterposed by the feminine in all reports; the second jettisons the
notion of journalistic objectivity altogether for "its perceived complicity
in legitimizing patriarchal hegemony," p. 122.
 Meyers, News Coverage of Violence Against Women: Engendering Blame.
 Helen Benedict, Virgin or Vamp: How the Press Covers Sex Crimes (New
York: Oxford University Press, 1992).
 Sophia E. Voumvakis and Richard V. Ericson, News Accounts of Attacks
on Women: A Comparison of Three Toronto Newspapers (Toronto: Centre of
Criminology, University of Toronto, 1984).
 Gabriel Weimann and Thomas Gabor, "Placing the Blame for Crime in
Press Reports," Deviant Behavior 8, (1987): 283-297.
 Meyers writes that "only within the past 15 years have researchers
combined feminist theory about violence against women with studies of news
coverage," in News Coverage of Violence Against Women: Engendering Blame, 28.
 Interview with first author, 4/16/2001
 Personal communication to the first author, 9/26/2002.
 Meyers, News Coverage of Violence Against Women: Engendering Blame, 61.
 Geraldine Finn, "Taking Gender into Account in the 'Theatre of
Terror': Violence, Media and the Maintenance of Male Dominance," Canadian
Journal of Women and the Law, 3(2), (1989-1990), 381, cited in Meyers, News
Coverage of Violence Against Women.
 Sharon Lamb and Susan Keon, "Blaming the Perpetrator: Language that
Distorts Reality in Newspaper Articles on Men Battering Women," Psychology
of Women Quarterly, 19 (1995), 211.
 Our funding source limited us to California newspapers. However, as
the largest—and arguably most media-genic—state, California powerfully
influences the national culture.
 After experimenting and cross checking with microfilmed copies of
newspapers, we settled on a combination of keywords and topics: "subject
(crime) or subject (violent crime) or subject (justice) or subject
(domestic violence) or police or violen* or batter* or gun* or prison*."
The asterisk includes any word beginning with the letters supplied,
regardless of ending.
 We defined such violence as "a deliberate physical attack on a
person, including violence to self and self-defense, or a written or verbal
threat of bodily harm, or stalking, harassing, or otherwise menacing an
individual's physical person." We chose this definition because it
parallels the one used by the National Violence Against Women Survey
sponsored by the U.S. Justice Department and the Centers for Disease
Control and Prevention.
 We chose 2000 because more recent state crime statistics were not
 Conforming to the U.S. Justice Department's National Violence Against
Women Survey classification, we defined intimate partner violence as that
taking place between current or former romantic/sexual partners, including
those on an initial date.
 That assured us equal numbers of fat Sunday editions and thin Monday
papers and diminished the distortion of a major violent event occurring at
one point in time. Researchers have found such constructed week designs
provide the best picture of reporting for a given sample size. See Daniel
Riffe, Charles F. Aust and Stephen R. Lacy, "The Effectiveness of Random,
Consecutive Day and Constructed Week Sampling in Newspaper Content
Analysis," Journalism Quarterly, 70(1) (1993): 133-139.
 We used arrest data rather than counts of incidents because
California measures intimate partner violence only by arrests. In a
category called spousal abuse, the state includes physical violence between
people who are married, formerly married, cohabiting or had a child
together. It is slightly narrower than our definition of intimate partner
violence in that it does not include all dating partners.
 Lori Dorfman and Vincent Schiraldi, Off Balance: Youth, Race and
Crime in the News (Washington DC: Youth Law Center, 2001); Doris A. Graber,
Crime News and the Public (New York: Praeger, 1980); Sanford Sherizen,
"Social Creation of Crime News: All the News Fitted to Print" in Charles
Winick, ed. Deviance and Mass Media (Beverly Hills, CA: Sage, 1978).
 While arrests constitute the best available surrogate for incidents,
they substantially underestimate incidents of intimate violence because
most violence between intimates isn't reported to police. Therefore the
actual ratio of homicides to less extreme incidents of spousal abuse is
smaller than official figures indicate. See Rosemary Chalk and Patricia A.
King, ed. Violence in Families: Assessing Prevention and Treatment
Programs (Washington, DC: National Academy Press, 1998).
 Shanto Iyengar, Is Anyone Responsible? How Television Frames
Political Issues (Chicago: University of Chicago Press, 1991). Also see
Graber, Processing the News: How people Tame the Information Tide and Dolf
Zillman and Hans-Bernd Brosius, Exemplification in Communication: The
Influence of Case Reports on the Perception of Issues (Mahwah, NJ: Lawrence
 Stephen D. Reese, "Prologue—Framing Public Life: A Bridging Model
for Media Research," in Stephen D. Reese, Oscar H. Gandy, Jr. and August E.
Grant, ed. Framing Public Life (Mahwah, NJ: Lawrence Earlbaum, 2001): 7-31.
 Explicit frames causally link a prior behavior or motivation with a
real or threatened act of violence. B happened because of A. Implicit
frames mention a prior behavior, but don't expressly connect it to the
violence: B happened; earlier A occurred.
 A second researcher independently coded 20% of the stories in both
samples. For manifest or obvious measures, such as whether the suspect was
identified by name, we accepted those with chance-corrected reliability
statistics of .8 or better using Scott's pi. For latent measures that
required coder judgment, such as the presence or absence of a frame, we
accepted those with pi's of .7 or better. See Daniel Riffe, Stephen Lacy
and Frederick G. Fico, Analyzing Media Messages: Using Quantitative Content
Analysis in Research (Mahwah, NJ: Lawrence Earlbaum, 1998). To correct for
the conservative bias in pi when one response category is chosen
disproportionately, we calculated the correction for chance using the
normal polynomial distribution when 80% or more of the coding decisions
fell in a single category. See W. James Potter and Deborah
Levine-Donnerstein, "Rethinking Validity and Reliability in Content
Analysis," Journal of Applied Communication Research, 27 (1999): 258-284.
 But they will do so equally for both story samples and thus won't
affect comparisons between intimate violence reporting and coverage of
other kinds of crime.
 Data come from the California Bureau of Criminal Information and
Analysis website (http://justice.hdcdojnet.state.ca.us/cjsc_stats).
 All margins of error for the other violence sample are calculated at
the standard 95% confidence level. Note that the IPV sample has no margin
of error since it contains every story in the population.
 Data come from the California Bureau of Criminal Information and
 Data come from the Santa Clara County Death Review Committee Chair,
Rolanda Pierre-Dixon, 7/18/2002 and the California Bureau of Criminal
Information and Analysis website.
 California Bureau of Criminal Information and Analysis website.
 Data come from the California Bureau of Criminal Information and
Analysis website and an interview with Sung Yu, research analyst at the
Injury and Violence Prevention Program, Los Angeles County Department of
health Services, 10/02/02.
 The difference is significant at p<.001; chi-square statistic for a
test of independence = 77.6 with df=1 and n=1,017.
 The frequency of male victims here raises the question of whether
the newspapers disproportionately reported cases in which women attacked
men. The gender ratio in the papers, however, mirrored state statistics,
which showed 83.5% of those arrested for spousal abuse were men. Data from
Marie K. Herbert, ed. Report on Arrests for Domestic Violence in
California, 1998 (Sacramento: State of California, Office of the Attorney
General, Criminal Justice Statistics Center Report Series 1 (3) August, 1999).
 "Sandy Banks: At-Risk Girls, Mentor Get Together and Get Real," Los
Angeles Times, 30 July 2000, sec. E, p. 1.
 Rennison, "Intimate Partner Violence and Age of victim, 1993-99.
 Patricia Tjaden and Nancy Thoennes, Extent, Nature, and Consequences
of Intimate Partner Violence (Washington, DC: U.S. Department of Justice,
Office of Justice Programs, 2000).
 We are grateful to Nancy Bern, who has conducted research about
similar negative stereotypes in magazines, for the suggestion to consider
the aggregate impact of blame-shifting frames lest our break-out of many
specific frames dissipate the force of these themes.
 The margin of error around the largest of these point estimates is
plus or minus 1.2 percentage points.
 In forthcoming research we will show that more sophisticated
reporting_which recognizes environmental causes rather than assuming crime
arises only from the disposition of the criminal_is even scarcer in
intimate violence coverage than in reporting of other types of violent crime.
 The level of suspect identification, which sums 9 attributes, was
marginally reliable in the other violence sample: Intercoder agreement,
within 1 attribute, was 85%; Scott's pi was .7.
 "Dating's Ugly Little Secret, Psychotherapist Examines Abusive Teen
Relationships," San Jose Mercury News, 10 October 2000, sec. D, p.5.
 "Spring Training Daily Report; Martin Faces Rough Road," Los Angeles
Times, 23 March 2000, sec. D, p.3
 The intimate violence percentage falls within the margin of error of
the OV estimate.
 James J. Collins and Pamela M. Messerschmidt, "Epidemiology of
Alcohol-Related Violence," Alcohol Health & Research World 17, 2 (1993):
 Plus or minus 1 percentage point in this case because of the large
sample and small standard deviation.
 The margin of error around this estimate—plus or minus 1.7 percentage
points—would not come close to closing the gap.
 This frame was marginally reliable in the other violence sample;
coders agreed 88% of the time, but Scott's pi was .66.
 The margin of error here is plus or minus 4 percentage points, not
quite enough to bridge the gap from the estimate of 29.3% of other violence
stories with an innocent claim to the 24.6% of intimate violence stories
containing this claim.
 The Los Angeles Times is owned by the Tribune Company whose CEO when
this was written was John W. Madigan. Of the 15 members of the Tribune
board of directors 12 were male; 11 members were white non-Hispanics. The
San Jose Mercury News is owned by Knight Ridder whose CEO was P. Anthony
Ridder. Of the 11 members of the Knight-Ridder board of directors, 8 are
male; 9 members are white non-Hispanics. Both Mr. Madigan and Mr. Ridder
are white. Knight Ridder is the second largest publisher of newspapers in
the U.S., by daily circulation and the Tribune Company is the eleventh
largest, but also owns other media properties including television
stations, and the Chicago Cubs baseball team. Information from corporate
websites: http://www.tribune.com/about/bios/board_index.html and
http://www.kri.com/ on 12/12/2002.
 Personal communication to first author, 10/29/2002.
 From a business perspective neglecting intimate violence makes sense.
The lower ratio of homicides to lesser felonies for intimate violence makes
it a less dramatic story than other types of violence. Also, assaults by
strangers may excite more fear than attacks within personal relationships.
Like drama, fear compels attention, making these crimes more commercially
valuable than intimate violence. Since news media revenues are based on the
size of the audience they attract, profit-minded publishers may foster a
newsroom culture that privileges extreme or random violence over lesser or
more predictable attacks. For more on how economics affects news selection
see John H. McManus, Market-Driven Journalism: Let the Citizen Beware?
(Thousand Oaks, CA: Sage, 1994).
 Liesbet van Zoonen, Feminist Media Studies (London: Sage, 1994), 40.
 Stories about specific crimes fill space very inexpensively: The
who, what, where, when and how_and sometimes the why_are all available at
one place and time, or in a press release. It's one-stop reporting.
 Doris A. Graber, Crime News and the Public (NY: Praeger, 1980).
 In Santa Clara County in 2000, 45% of the felony arrests were for
spousal abuse, but only 18% of the homicides. In Los Angeles County, 36% of
the felony arrests concerned spousal abuse, but only 5% of the homicides.
Non-intimate violence accounted for a greater ratio of homicides per felony
arrest. Data from the Criminal Justice Statistic Center of the California
Department of Justice, available at
 In California, police agencies received almost 200,000 calls for help
in domestic violence situations and made more than 51,200 felony arrests in
2000, according to the California Department of Justice
(http://caag.state.ca.us/cvpc/fs_dv_in_ca.html). Domestic violence was the
number one violent felony arrest in the state in 1997, according to the
Mercury News, "The Face of Abuse," 9 September 1999, sec. A, p.1.
 Chalk and King, ed. Violence in Families: Assessing Prevention and
 J. Kaufman and E. Zigler, "Do Abused Children Become Abusive
Parents?" American Journal of Orthopsychiatry 57(2), (1987): 186-192, cited
in Chalk and King, Violence in Families: Assessing Prevention and Treatment
 See Kovach and Rosenstiel, The Elements of Journalism, esp. ch. 9,
"Make the News Comprehensive and Proportional."
 "America's Best Newspapers," Columbia Journalism Review, Nov/Dec,
Table 1: Frequency of Victim-Blame Frames
Table 2: Frequency of Suspect-Mitigation Frames