Violence Against Women

Violence Against Women http://vaw.sagepub.com/

Fact or Fiction? Gender Issues Related to Newspaper Reports of Intimate Partner Homicide Jessica A. Wozniak and Kathy A. McCloskey Violence Against Women 2010 16: 934 DOI: 10.1177/1077801210375977 The online version of this article can be found at: http://vaw.sagepub.com/content/16/8/934

Published by: http://www.sagepublications.com

Additional services and information for Violence Against Women can be found at: Email Alerts: http://vaw.sagepub.com/cgi/alerts Subscriptions: http://vaw.sagepub.com/subscriptions Reprints: http://www.sagepub.com/journalsReprints.nav Permissions: http://www.sagepub.com/journalsPermissions.nav Citations: http://vaw.sagepub.com/content/16/8/934.refs.html

Downloaded from vaw.sagepub.com at Tehran University on December 2, 2010

Article

Fact or Fiction? Gender Issues Related to Newspaper Reports of Intimate Partner Homicide

Violence Against Women 16(8) 934–952 © The Author(s) 2010 Reprints and permission: http://www. sagepub.com/journalsPermissions.nav DOI: 10.1177/1077801210375977 http://vaw.sagepub.com

Jessica A. Wozniak1 and Kathy A. McCloskey1

Abstract The present study compared newspaper articles reporting female-perpetrated intimate partner violence (IPV) homicides with those reporting male-perpetrated IPV homicides. Domestic violence was not mentioned in 72% of the articles. Female-perpetrated IPV homicide, although relatively rare, was not portrayed as more newsworthy nor did it receive more coverage when compared to male-perpetrated IPV homicide. When analyzing explanations for homicides given in newspaper reports, as well as descriptors used to describe the perpetrator, there was no support found for gender bias in favor of male perpetrators. Female victims were more likely to be portrayed as innocent when compared to males. Keywords domestic homicide, intimate partner violence, newspaper reporting

Introduction: Intimate Partner Violence-Related Newspaper Coverage As a public health problem, domestic violence is but a subset of overall violent crime in the United States and constitutes about 33% of all violent crime recorded by police nationwide (Durose et al., 2005). However, in some jurisdictions domestic violence may

1

University of Hartford, West Hartford, CT

Corresponding Author: Jessica A. Wozniak, Graduate Institute of Professional Psychology, 44B Nanel Drive, Glastonbury, CT 06033 Email: [email protected]


Wozniak and McCloskey

935

constitute the majority of calls to the police and represent the most common type of felony arrest (Durose et al., 2005; McManus & Dorfman, 2003). In terms of severe violence resulting in death, about 22% of all homicides are domestic murders (Durose et al., 2005) and, once again, some jurisdictions show higher proportions than this. The most common form of domestic violence is intimate partner violence (IPV), or that which occurs between past or present adult intimates. Recent national statistics show that males represent 83% of all spousal murderers and 75% of those who killed a boyfriend or girlfriend (Durose et al., 2005), suggesting that IPV is largely a male-perpetrated problem. However, cultural depictions of IPV, including representations by the news media, may or may not accurately reflect the true shape or tenor of this public health problem. Studies conducted by McManus and Dorfman (2003) and Sorenson, Manz, and Berk (1998) found that IPV homicides are less likely to be covered in newspapers than homicides committed by strangers, out of proportion to their actual occurrence within the local population. Furthermore, when IPV homicides are reported, they are covered very differently than other homicides because they are often less visible in their placement in the newspaper and are less likely to have humanistic story frames that are sometimes added to news reports to provide depth and context (Bullock & Cubert, 2002; Sorenson et al., 1998; Taylor & Sorenson, 2002). In addition, unlike reports concerning other types of homicide, IPV homicide articles seldom provide background issues or use a “hook,” referring to the first few sentences in an article used to draw in the reader; also, they are often unemotional in tone (Bullock & Cubert, 2002; Sorenson et al., 1998). Citizens, therefore, may not read the article because it does not catch their attention or play on their emotions. When reporting IPV homicides, stories are often episodic and focus only on that particular incident rather than the broader context of IPV and rarely include expert opinion (Delaware Coalition Against Domestic Violence [DCADV], 2006; Taylor & Sorenson, 2002). As noted above, an IPV homicide is often portrayed as a single event and not reported as the endpoint in a pattern of abuse that has occurred over time (Bullock & Cubert, 2002; Websdale, 1998). By presenting stories of violence against women as separate and discrete events, newspapers portray to the public that IPV incidents against women are isolated events, thus denying the widespread nature of IPV in our culture (Carll, 1999; DCADV, 2006; Rhode Island Coalition Against Domestic Violence [RICADV], 2000; Websdale & Alvarez, 1997). Bullock and Cubert (2002) found that out of 230 newspaper articles on domestic violence homicides, less than a quarter actually labeled the incident as domestic violence and only 10% of the articles framed the homicide within the larger context of IPV. Furthermore, almost half (48%) of the articles suggested some type of excuse for IPV perpetrator behavior, and 17% included victim-blaming language. Similarly, a study published by the RICADV in 2000 showed that less than 20% of relevant newspaper articles concerning IPV homicides made clear reference to the specific type of violence involved, while some reporters labeled such homicides as a “family tragedy” or framed IPV homicides as unpredictable or as the result of a tragic “love story” even with evidence of prior IPV incidents in hand. In 2006, the DCADV released a report that showed out of 30 newspaper articles concerning IPV homicides only 10% referred to domestic violence, about a quarter


936

Violence Against Women 16(8)

of the articles attempted to establish the presence/absence of abuse before the homicide, another 20% of the articles used positive perpetrator descriptors (e.g., “nice,” “hard-working,” “normal,” “upbeat,” “well-liked”), and slightly more than half (52%) of the articles described the murder as a “shock” or unpredictable. In a study by Taylor and Sorenson (2002), only one out of 33 IPV homicide articles addressed the broader contextual issues surrounding IPV, yet even that one article focused on women’s violence against men. Unfortunately, this is all too common. One important difference between male-perpetrated IPV homicides and femaleperpetrated IPV homicides is that when women kill their male partners it is often in selfdefense within a context of ongoing victimization, or in response to a physical provocation from their male partners (a “victim precipitated conflict;” Felson & Messner, 1998). More than a decade ago, Snell and Morton (1994) noted that women incarcerated for homicide were more likely to have killed an intimate (50%) within the context of ongoing victimization than men (16%). In addition, male IPV homicide victims were more likely to have a violent history than female victims. Conversely, female homicide IPV perpetrators were less likely to have a violent history than other female homicide perpetrators, suggesting that the former may need greater provocation to commit murder (Felson & Messner, 1998). However, information about female IPV homicide perpetrators may be depicted in the media very differently than the criminal justice and public health care systems show. For instance, Bond-Maupin (1998) has shown that female offenders appearing on television shows do not fit the actual offender profile based on statistics nationwide. Unfortunately, when an IPV homicide perpetrator is female, it has been shown that newspapers often emphasize the nature of the crime in the headline and usually do not provide context (e.g., active resistance and/or self-defense) (Jewkes, 2004). If context is included, it often appears later in the article or as a follow-up weeks or months later (Carll, 1999). When the IPV homicide perpetrator is a male, however, the headlines may provide exculpatory reasons for the murder (e.g., “Man Kills Cheating Wife,” Carll, 1999). For male IPV homicide perpetrators, an altered mental state may also be used to explain why the homicide occurred, including obsession with the victim, uncontrollable responses to the victim’s actions (e.g., suspected infidelity), and other rationales as to why the man was not in command of his actions during the killing, thus obscuring the man’s responsibility (Carll, 1999; Meyers, 1994, 1997). By portraying the male IPV homicide perpetrator in this manner, news stories may lead the public to falsely believe that the female victim was the one in control of the situation (Jewkes, 2004). They may also promote the belief that women cause their own victimization, thereby reducing the male IPV homicide perpetrator’s responsibility for his lethal actions (Meyers, 1994, 1997). When prior IPV victimization of the female is portrayed in an article, studies have found that it is weakened by the use of terms such as “alleged,” which can give readers the impression that this information may not be as accurate or as important as that obtained from the male perpetrator (Consalvo, 1998; Jewkes, 2004). Once again, blame is not placed on the individual who actually committed the crime. Reporters choose a point of view based on the sources available to them, such as police reports or court documents, the testimony of the perpetrator, the lawyers, or witnesses, and interviews with surviving



937

friends and relatives of the victim and/or perpetrator, and it is believed that these sources are chosen because of ease of access and availability (Consalvo, 1998). When only “dry” sources are chosen based on availability, such as police reports or court documents, the human face of IPV is not given and the crime may be dehumanized when information from people who knew the victim and/or perpetrator is omitted (Consalvo, 1998). Conversely, reporting only forensic facts with little or no context can easily obscure the overarching social meaning of the crime (Websdale, 1998; Websdale & Alvarez, 1997). In addition, studies have shown that domestic violence experts are rarely seen as a valuable source when IPV homicides are reported even though they could easily help provide the social context behind these crimes (Bullock & Cubert, 2002; DCADV, 2006). Not all studies show the problematic patterns outlined above. Notably, McManus and Dorfman (2003) recently examined IPV-related articles appearing in the San Jose Mercury News and the Los Angeles Times, two California newspapers well known for their high journalistic quality. While these authors found that compared to other types of violent crime, IPV did not receive the same attention proportionate to the frequency of actual arrests and was treated less frequently within each article as an issue topic, they also found that IPV articles are usually not framed to blame the victim or make excuses for the perpetrator. It should be noted, however, that IPV articles nevertheless provided such frames more often than articles concerning other types of violent crime. Thus the story, the social context, the violent acts themselves, and the responsibility for the crime can all be portrayed in many different ways and from vastly different viewpoints depending on the perspectives and sources that reporters use, illustrating the way journalists influence how IPV is presented to society at large (Bullock & Cubert, 2002). Vagueness in an article can also conceal the social context within which the crime occurred, obscure responsibility for the homicide through positive self-presentation/ negative other-presentation by the perpetrator, and fail to educate the public about the problem of IPV (Meyers, 1997). Because the public relies on the media for information concerning such social problems, inaccurate or incomplete news coverage of IPV homicides can lead to false views concerning IPV problems as well as the implied solutions to such problems. As we discuss next, misrepresentation in news reports may lead to a distorted reality and imprecise representation of what is seen as right and wrong concerning violence in general (Coleman & Thorson, 2002; Gilliam & Iyengar, 2000) and IPV in particular (McEvoy, 1996).

Why Is It Important to Examine Media Reports? It has been known for some time that news stories influence public perceptions of social reality, reflect public interest, and play a large role in how people understand certain societal problems, especially crime (e.g., Curran, Gurevitch, & Woollacott, 1979), and IPV is no exception. It seems clear that mass media, specifically television news and newspaper reports, are a primary source of information about crime and violence and thus play a major role in shaping societal views of morality (e.g., McEvoy, 1996). A connection between crime reporting and attitudes about crime has been demonstrated, illustrating the


938


influence that the media have on public opinion and perceptions (Gilliam & Iyengar, 2000, 2005; Holbert, Shah, & Kwak, 2004). There are numerous reasons why this is so. More than a quarter of a century ago, Sheley and Ashkins (1981) found that the relative distribution of crimes covered by the media does not match distributions found in official justice system statistics, often resulting in a public view of crime that is more similar to the image presented by the media than to that actually found in the community. Such a mismatch continues today; as noted previously, a recent California study examined newspaper coverage of different types of violent crime and found that coverage did not match incidence (McManus & Dorfman, 2003). Part of the reason for this mismatch may be that certain crimes receive more newspaper coverage than others; crimes against people, especially homicide, appear to receive the most attention. News reporters often use police wire services as sources of information, yet these services only alert reporters to sensational offenses or those in which the police or reporters have expressed a particular interest; this often leads to crime being represented through a distorted lens by the media and presented inaccurately to the public (e.g., Sheley & Ashkins, 1981). Furthermore, television and newspapers often compete with each other for viewership and readership. This can lead to a search for marketable news stories (e.g., homicide) that are often guided by factors such as dramatization and simplification, both of which can obscure the true meaning of an event and fail to provide context (Chibnall, 1977; Sheley & Ashkins, 1981). In addition, crime stories are often chosen if they can be presented in a way that displays action and drama, thereby further distorting the actual context surrounding the crime (Chibnall, 1977). This is sometimes referred to as part of the “frame” used to structure news stories; for better or worse, such frames have been shown to influence public beliefs about crime (Gilliam & Iyengar, 2000, 2005). Websdale and Alverez (1997) discuss the tendency for newspaper journalists to structure crime reports solely around the forensic details of each case and report only on the immediate situation in which the crime took place. However, as Websdale (1998) has noted, this journalistic approach often obscures the contextual factors surrounding IPV homicides: This approach to the reporting of violent crime usually focuses on individual offenders and virtually never on structural factors. . . . While we may learn demographic details of the perpetrator and victim . . . we learn very little of the political nature of the offense or the gendered pattern of killing in tragedies such as domestic homicidesuicides. For example, although many homicide-suicides are preceded by a history of woman battering, we rarely learn this from the press accounts. . . . The press rarely tells readers that the typical perpetrator of homicide-suicides are men [yet] such simple details inform the public about the gendered edge to these types of killings and could therefore serve as a powerful educational tool regarding violence against women. (pp. 205-206) Reporters, then, often miss the opportunity to educate the general public about the problem, even though there is evidence that more accurate understanding can be achieved by including larger contextual factors. For instance, Coleman and Thorson (2002) found



939

that when newspaper reports about general violent crime were written in such a way to include patterned context, risk factors, policy issues, and prevention strategies, shifts in reader attribution of responsibility occurred; attribution focus shifted from solely on the individual to societal influences on that individual. Studies on overall homicide reporting in newspapers have also shown that victim and perpetrator characteristics are related to newspaper coverage. One analysis of the Los Angeles Times concluded that if the suspect was Latino or if the suspect had less than a high school education, homicides received less coverage than homicides committed by individuals of other ethnicities or with a higher education (Sorenson et al., 1998). In addition, when a victim was female, the story concerning homicide was two times more likely to be reported than when the victim was male even though, overall, men are murdered in larger numbers in our society. In contrast, Gilliam and Iyengar (2000) examined the effects of news reporting on the viewing public and found that a common “script” used in televised news broadcasts includes two core messages, namely, that crime is usually violent and perpetrators are usually non-White. These messages are also inaccurate given the fact that most crime is nonviolent and the largest number of perpetrators are White. Consequently, there is ample evidence of both over- and undercoverage of crime incidence that varies with the demographic characteristics of both victims and perpetrators. It seems clear that individuals in our culture look to the news media to provide an accurate portrayal of social issues, particularly crime (Curran et al., 1979; Gilliam & Iyengar, 2000, 2005). However, due to the need for television and newspaper reporters to find marketable stories while dealing with intense deadline pressures, sensational crimes of an episodic and decontextualized nature are often more likely to make print. As noted above, available crime stories may also be overdramatized and adhere to a particular news frame, thus distorting the reality of the actual crime (Chibnall, 1977; Gilliam & Iyengar, 2000, 2005; Sheley & Ashkins, 1981). In addition, both race and gender play a crucial role in which stories are chosen for coverage. The tendency for journalists to report individual case facts without including the overarching context leading up to the crime because of deadlines, pressures to deliver, and resource allocation also can obscure important information (Websdale, 1998; Websdale & Alvarez, 1997). These factors can lead to widespread and inaccurate influence on opinions about crime that do not match reality because people often perceive societal problems based on what journalists choose to report (Hollbert et al., 2004; Sheley & Ashkins, 1981). Unfortunately, journalistic choices often stem from viewership and readership patterns (i.e., what seems to sell) as well as limits on space for coverage, budget issues, and journalistic deadlines, rather than solely the need to provide accurate, impartial, and contextualized reporting. Because the public gets much of its knowledge concerning IPV from the news media, it is important to understand what media sources such as newspapers currently report. The role the media play in forming public conceptions or misconceptions about IPV leads to a need for more in-depth examination of what the media actually present to the public. If the media do not portray IPV as important, then society at large will probably not view it as important. There will likely be little motivation for societal change if the severity of violence against women is not portrayed accurately and completely. In addition, perpetrators and/or potential


940


perpetrators will not learn to alter their behavior or take responsibility for their actions without consistent cultural messages outlining the unacceptability of IPV. Political actions concerning IPV prevention and intervention may also be affected by such reporting because of the devaluing of the crime and the victim. The results of the present study can perhaps be used to educate media professionals and improve the media representation of IPV homicide as well as improve public knowledge about IPV in general.

Rationale for the Present Study The present effort is wider in scope than previous studies in that a large number of newspapers and geographic regions were sampled. The present study also provides a more in-depth analysis of how IPV homicide is portrayed in initial newspaper reports throughout the United States than has yet to be compiled in one effort alone. In addition, the present study differs from previous research because it analyzes only the first article reported about the homicide to investigate the initial representation the public receives in regard to IPV homicide. Such initial depictions are assumed to best represent the effects of deadlines, the information available to reporters immediately after a homicide occurs, and other journalistic pressures. In addition, follow-up articles are only counted for 2 weeks after the first report to both decrease the number of confounding variables (such as courtroom dramatics or problems with the judicial process) that may lead to varied reporting as well as to eliminate thematic stories. Our focus was the first article about the IPV homicide to appear in print. It is expected that initial newspaper articles concerning IPV homicides will show many of the problematic patterns discussed in the research literature. The first and most important hypothesis is that very few, if any, of the newspapers sampled will provide discussions of IPV within a broader context. Because journalists often portray IPV murders as isolated and unpredictable incidents rather than resulting from a pattern of abuse, it is likely that the majority of the articles will lack discussions of domestic violence within a broader context. To test this hypothesis, each article will be examined for whether or not (a) terms specifically referring to IPV were used, (b) the context of community intervention was included, and (c) statistics, patterns of violence, and quotes from IPV experts were included. In addition, the type of sources quoted, if any, will be noted. The second hypothesis centers on the characteristics of the perpetrator and victim. Based on the demographic differences in victimology outlined above, it is expected that non-White, uneducated, or poor victims and perpetrators will be represented more often in the articles sampled than victims or perpetrators who are White, educated, or middle to upper class. A set of tertiary hypotheses centers on the effects of gender. First, it is expected that initial articles concerning female perpetrators of IPV homicide will receive more coverage than those involving male perpetrators due to the relative rarity and sensationalism associated with such homicides. More coverage is defined here as (a) longer word counts in the initial article and (b) more numerous follow-up articles within a 2-week period. Second, it is expected that more explanations/motivations within each article will be given for the behavior of male perpetrators compared to female perpetrators. This expectation is based



941

on findings reviewed in the literature that suggest male IPV homicide perpetrators often receive exculpatory coverage. Third, it is expected that, regardless of whether she is perpetrator or victim, the female in the situation will be portrayed more negatively than the male.

Method Procedure IPV newspaper articles database. The current database consists of 100 IPV homiciderelated newspaper articles that were selected from a large, previously existing database that had been sampled randomly from newspaper reports across the United States by a colleague of the second author. Seventy articles reporting male-perpetrated IPV homicides and 30 articles reporting female-perpetrated IPV homicides were randomly selected from within this larger database. This 70/30 ratio approximated the gender representation of the IPV homicide perpetrator population as reflected in criminal justice and public health statistics. This sampling approach was used to avoid sampling bias by closely approximating what is found in the homicide population of interest (Hirsch & Riegelman, 1992). All resultant data were then analyzed using a chi-square (c2) computerized statistical module (SPSS) that allowed experimenter manipulation of the expected frequencies. Thus all Chisquare statistics reported below were conducted with expected frequencies that were commensurate with the 70/30 selection ratio (George & Mallery, 1999). Articles were published between 2000 and 2003, with 23 articles published in 2000, 21 in 2001, 34 in 2002, and 22 in 2003. Lexis-Nexis Academia On-Line was then used to confirm that the articles from the original database were the first reported on that specific homicide. If they were not, the original article was substituted into the sample. Of the 100 articles collected, all referred to heterosexual couples, 50 of whom were married, 28 were boyfriend/girlfriend, 14 were estranged, 6 were divorced, and 2 were separated. The 100 original reported articles were then coded on a rating form developed specifically for the present research effort. Two raters, who were blind to the hypotheses, were trained on the use of this rating form by using three articles that were not included in the sample (rating forms are available from the first author on request). Interrater reliability was determined to be 80% using these three articles. Each rater then coded all 100 articles independently. Any discrepancies between raters were discussed and recoded once consensus was reached, following the procedure of Wahl, Hanrahan, Karl, Lasher, and Swaye (2007). Lexis-Nexis Academia On-Line was also used to identify any follow-up articles for each of the 100 articles in the study sample. Keywords were used to search for these articles, and articles were counted if they were published within a 2-week period after the murder occurred. It should be noted that articles that dealt only with the judicial disposition of the case (e.g., arraignment, sentencing) were not included A simple count of the number of follow-up articles was then obtained to see what types of IPV homicides received more coverage within a 2-week period. Ratings of newspaper articles. As noted, a rating form was used to code all 100 articles in the sample. The rating form included seven sections based on content domains that past


942


researchers found to be important in analyzing reports of IPV homicide. The first section consisted of identifying information about the article and the type of IPV that was reported. The second and third sections focused on victim and perpetrator characteristics, respectively, as well as the way in which the victim and perpetrator were depicted. In other words, the focus of these sections was on the overall “take home” messages that the article conveyed to the average reader regarding the victim and perpetrator. Raters had to select between two opposite descriptors. The descriptive terms used here were selected based on past research that suggests both victims and perpetrators are often described using such words. Raters based their choice on specific words used or examples given that implied these characteristics. In some circumstances, neither of the descriptors was used to describe the individual, or the characterization could not be determined from the article. In some instances, the individual was described in both ways in the article, based on different sources or opinions. The fourth section examined specific elements given within each article, including whether or not specific terms referring to IPV were used as well as whether the broader social issue of domestic violence was addressed. The fifth section identified sources used by the journalist, and the sixth section included any motivations reported for the crime. The seventh section asked who was held responsible for the IPV homicide. It should be noted that this particular section did not target legal responsibility, but rather how the perpetrator/victim was portrayed in the article by the reporter. This portrayal could be based on source statements or word choice of the reporter. Raters determined whether the victim or the perpetrator was considered responsible for the victim’s death, the victim was held responsible for his/her own death, the perpetrator was held responsible for the death of the victim, or the perpetrator and the victim were portrayed as both at fault for the death of the victim. Finally, space was included on the rating form for any other information that may have been relevant to understanding the crime but was not included in the seven previous coding sections. (The rater coding instruction manual is also available from the first author on request.)

Results IPV Within Context Seventy-two percent of the articles did not even mention the term domestic violence or any derivative thereof. None of the articles discussed the IPV homicide in terms of community intervention efforts. Only one of the 100 articles reported domestic violence statistics and only one referred to domestic violence as a broader social issue. Thus the overwhelming majority of articles lacked discussion of domestic violence within a broader context, supporting the first major hypothesis. Nevertheless, results also indicated that when the perpetrator was female, it was more likely that reports of past violence by male victims were included by reporters, c2 (1, N = 100) = 4.05, p = .044. Of the 30 articles concerning female IPV homicide perpetrators,



943

23.3% reported past violence by male victims, whereas only 8.6% of the 70 articles concerning male IPV homicide perpetrators reported past violence by female victims. The overwhelming majority of articles (98 out of 100) lacked opinions by any experts or officials other than justice system entities such as police officers or court officials. Results show that when opinions of justice system workers were included, such opinions were more likely to appear in articles concerning female perpetrators (19 out of 30, or 63.3%) than when the perpetrator was male (8 out of 70, or 11.4%; c2 (1, N = 100) = 28.70, p < .0001). Interviews of family members, friends, or neighbors were poorly represented as sources of information (14 out of 100). However, such information was more likely to appear in articles concerning female perpetrators (8 out of 30, or 26.7%) than when the perpetrator was male (6 out of 70, or 8.6%), c2 (1, N = 100) = 5.71, p = .017. Consistent with the original hypothesis that few, if any, of the articles in this sample would include context, it was found that about three quarters of all articles failed to even mention IPV (or use derivative terms thereof). All of the articles failed to discuss the homicide within the context of community interventions and 99% did not provide IPV statistics or provide a discussion of this social problem as background. Furthermore, 87% of the articles lacked reports of any violence prior to the homicide, 98% of the articles lacked quotes from any expert source with the exception of justice system workers, and 86% of the articles lacked quotes from family members, friends, or neighbors. These results all support the first major research hypothesis, namely, that articles would largely fail to provide different sources of contextual information for the homicides. On the other hand, when articles did include context, there appeared to be gender differences. Specifically, past violent behavior by IPV homicide victims was reported more often when the victim was male than when the victim was female. Likewise, quotes from justice system workers, family members, friends, or neighbors, were included more often when the perpetrator was female and the victim male. It seemed that when contextual factors were included, they could be used to help the reader make sense of the relative rarity of female-perpetrated IPV homicide.

Ethnicity, Occupation, Education, and Income With regard to victims, the lack of demographic information was evident. Only two of the 100 articles reported the victim’s education level and none noted income. The occupation of the victim was reported in only 18 of the articles. Ethnicity was rarely noted as well, with only five of the 100 articles making a reference to the victim’s ethnic background. Statistical analyses concerning ethnicity, educational level, or socioeconomic status could not be performed because of the small number of articles specifically giving information on these variables. Similar to the results for victims, very few articles presented demographic information concerning the perpetrators. Eighteen articles reported the occupation of the perpetrator, one article mentioned the education of the perpetrator, and none noted income. Ethnicity was reported in only 6 of the 100 articles. Once again, statistical analyses concerning any differences between the above demographic variables and the amount of coverage could not be performed because of the small number of articles that provided this information.


944


Because only a handful of articles reported ethnicity, educational level, or occupation for either the perpetrator or victim and none reported income, the present sample did not allow an examination of perpetrator/victim “worthiness,” the second hypothesis of this study. Although other researchers have found that ethnicity and class variables affect the amount and type of news coverage given to a particular crime, this effect could not be tested in the present study.

The Effects of Gender Extent of coverage. Results indicated that the mean number of words used in the articles was 290 (SD = 192.9). The mean number of follow-up articles within a 2-week period was 2.20 (SD = 2.5). Independent t tests were used to determine any significant gender differences in length of the articles (male perpetrator M = 299.76, SD = 203.13; female perpetrator M = 267.20, SD = 167.44) or number of follow-up reports between articles (male perpetrator M = 2.29, SD = 2.52; female perpetrator M = 2.07, SD = 2.42). Results revealed no significant gender differences for either article length, t(98) = .77, p = .44, or number of follow-up reports, t(98) = .40, p = .69. Therefore, the third major hypothesis—that female perpetrated homicides would receive more newspaper space within a 2-week period than those involving male perpetrators—was not supported. Explanations/motivations for homicide. Perpetrator motivations or reasons for the homicide could not be found in slightly more than half (55 out of 100) of the articles. In the remaining 45 articles where at least one motivation/reason was included, the most common reason given was that an argument between the perpetrator and victim led to the murder. Proportionately, reporting that an argument preceded the homicide tended to be more likely when the perpetrator was female (12 out of 30 articles, or 40%) than when the perpetrator was male (15 out of 70, or 21.4%), but the chi-square analysis did not reach significance for this difference, c2 (1, N = 100) = 3.68, p = .055. Chi-square results revealed there were also no significant gender differences for any of the other motivations identified within the articles, and thus no support for gender bias was found within these variables.

Article Depictions Perpetrator depiction. Because of small numbers or zeros in many of the categorical cells shown in Table 1, chi-square analysis could be performed on only a few of the perpetrator descriptors. (Please refer to the table for explanation on how cells were collapsed for a 2 × 2 analysis.) For those that could be tested, it was found that descriptors such as warmth/ coldness, stability, or friendliness showed no significant differences for perpetrator gender (see Table 1). In addition, no perpetrators were explicitly referred to as victims, regardless of gender. Results did indicate, however, that articles were more likely to portray the perpetrator as blameworthy (65 out of 70, or 92.9%) when the perpetrator was male than when the perpetrator was female (16 out of 30, or 53.3%), c2 (1, N = 100) = 21.32, p < .0001. It should be noted that the perpetrator was not described as strictly innocent in any of the articles, regardless of gender.


945

Wozniak and McCloskey Table 1. Depiction of Articles Concerning IPV Homicide Perpetrator Interpersonal Warmth

Male perpetrator N = 70 Female perpetrator N = 30

Neither

Warm/Loving

Cold/Abusive/ Violent

Both

47 (67.1%)

3 (4.3%)

18 (25.7%)

2 (2.9%)

23 (76.7%)

1 (3.3%)

5 (16.7%)

1 (3.3%)

Neither Male perpetrator N = 70 Female perpetrator N = 30

0 (0.0%)

6 (8.6%)

0 (0.0%)

26 (86.7%)

1 (3.3%)

2 (6.7%)

1 (3.3%) Both

2 (2.9%)

2 (2.9%)

27 (90.0%)

0 (0.0%)

2 (6.7%)

1 (3.3%)

Blameworthiness Blameworthya Innocent

Both

1 (1.4%)

65 (92.9%)

0 (0.0%)

4 (5.7%)

6 (20.0%)

16 (53.3%)

0 (0.0%)

8 (26.7%)

Referred to as victim

Labeling Referred to as perpetrator

Both

70 (100%)

0 (0.0%)

0 (0.0%)

0 (0.0%)

30 (100%)

0 (0.0%)

0 (0.0%)

0 (0.0%)

Neither Male Perpetrator N = 70 Female perpetrator N = 30

Friendliness Friendly

2 (2.9%)


Unfriendly

64 (91.4%)


Both

64 (91.4%)


Stability Unstable

Stable

Responsibility Victim held Perpetrator responsible held responsibleb

Both

1 (1.4%)

0 (0.0%)

69 (98.6%)

0 (0.0%)

10 (33.3%)

0 (0.0%)

13 (43.3%)

7 (23.3%)

a. c2 (1, N = 100) = 21.32, p < .0001; male perpetrators depicted as blameworthy more often than female perpetrators (male blameworthiness = 65 out of 70 versus neither, innocent, or both = 6 out of 70; female blameworthiness = 16 out of 30 versus neither, innocent, or both = 14 out of 30). b. c2 (1, N = 100) = 43.41, p < .0001; male perpetrators held responsible more often than female perpetrators (male perpetrator responsibility = 69 out of 70 versus neither, victim responsible, or both = 1 out of 70; female perpetrator responsibility = 13 out of 30 versus neither, victim responsible, or both = 17 out of 30).


946


Table 2. Depiction of Articles Concerning IPV Homicide Victim

Interpersonal Warmth

Neither Warm/Loving Female victim N = 70 Male victim N = 30

60 (85.7%) 19 (63.3%)

Female victim N = 70 Male victim N = 30

70 (100%) 30 (100%)

Female victim N = 70 Male victim N = 30

60 (85.7%) 26 (86.7%)

Neither

Neither

Neither

Cold/Abusive/ Violent

Both

10 (14.3%) 0 (0.0%) 2 (6.7%) 6 (20.0%) Stability Stable Unstable

0 (0.0%) 3 (10.03%)

0 (0.0%) 0 (0.0%) 0 (0.0%) 0 (0.0%) Friendliness Stable Unstable

0 (0.0%) 0 (0.0%)

0 (0.0%) 10 (14.3%) 0 (0.0%) 4 (13.3%) Blameworthiness Blameworthy Innocenta

0 (0.0%) 0 (0.0%)

Both

Both

Both

Female victim N = 70 3 (4.3%) 0 (0.0%) 65 (92.9%) 2 (2.97%) Male victim N = 30 12 (40.0%) 0 (0.0%) 12 (40.0%) 6 (20.0%) Labeling Neither Referred to Referred to as as Victim Perpetrator Both Female victim N = 70 Male victim N = 30

58 (82.9%) 23 (76.7%)

Female victim N = 70 Male Victim N = 30

44 (62.9%) 20 (66.7%)

12 (17.1%) 0 (0.0%) 0 (0.0%) 7 (23.3%) 0 (0.0%) 0 (0.0%) Responsibility Caused Did Not Cause Bothb Perpetrator’s Perpetrator’s Neither Actions Actions 6 (8.6%) 3 (10.0%)

14.(20.0%) 0 (0.0%)

6.(8.6%) 7 (23.3%)

a. c2 (1, N = 100) = 33.13, p < .0001; female victims were portrayed as innocent more often than male victims (female victim innocence = 65 out of 70 versus neither, blameworthy, or both = 5 out of 70; male victim innocence = 12 out of 30 versus neither, blameworthy, or both = 18 out of 30). b. c2 (1, N = 100) = 4.05, p = .044; male victims were portrayed as both causing and not causing perpetrator’s actions more often than female victims (male victim portrayed as both causing and not causing perpetrator’s actions = 7 out of 30 versus neither, victim caused perpetrator’s actions, or victim did not cause perpetrator’s actions = 23 out of 30; female victim portrayed as both causing and not causing perpetrator’s actions = 6 out of 70 versus neither, victim caused perpetrator’s actions, or victim did not cause perpetrator’s actions = 64 out of 70).

More of the articles held the perpetrator directly accountable for the victim’s death (69 out of 70; 98.6%) when the homicide perpetrator was male than when the perpetrator was female (13 out of 30; 43.3%). This was a statistically significant difference, c2 (1, N = 100) = 43.41, p < .0001. Another 10 out of 30 (33.3%) female perpetrator articles did not hold either perpetrator or victim responsible, and both the female perpetrator and male



947

victim were portrayed as at fault in seven out of 30 (23.3%). These results are in direct contrast with the original hypothesis that depictions of accountability would show a gender bias in favor of male perpetrators. Victim depiction. Table 2 shows that once again, because of small numbers or zeros in many of the categorical cells, chi-square statistics could be performed on only a few of the victim descriptors. Overall, articles were neutral in the words used to describe victim stability, friendliness, or labeling as victim/perpetrator. In addition, the majority of articles depicted victims as neither warm/loving nor cold/abusive (60 out of 70, or 85.7% for female victims, and 19 out of 30, or 63.3% for male victims).However, female victims were depicted as warm/loving 14.3% of the time (10 out of 70 articles), whereas male victims were described as warm/loving only 6.7% of the time (2 out of 30 articles). Chi-square results indicated that female victims were more likely to be depicted as innocent in 65 out of 70 articles (92.9%) compared to male victims (12 out of 30 articles, or 40%), c2 (1, N = 100) = 33.13, p < .0001. No articles depicted the victim as totally blameworthy regardless of gender. Victim responsibility was not referenced at all in the majority of articles (44 out of 70, or 62.9% for female victims, and 20 out of 30, or 66.7% for male victims). However, when both the victim and perpetrator were held responsible for the homicide, victims tended to be male (7 out of 30 articles, or 23.3%) compared to female (6 out of 70 articles, or 8.6%). Again, these results are in contrast to the original hypothesis of gender bias in favor of males.

Discussion Explanation of Results Contextual factors. As expected, the overwhelming majority of these initial articles lacked discussion of IPV within a broader context and did not include the concept of domestic violence in general. Based on previous research, this finding is not surprising. With regard to IPV homicide reports, newspapers often do not provide the context of the crime (e.g., as a widespread social phenomenon based on gender) and, if context is included, it often appears as a thematic follow-up many weeks or months later (Carll, 1999). In the few articles that did provide some type of context, results showed that, when the perpetrator was female, it was more likely that reports of past violence by the male victim were included than when the perpetrator was male and the victim was female. This finding is both consistent with and contradictory of past research findings. In terms of forensic research, it has been known for some time that men killed by their female partners are more likely than not to have a criminal history, including prior arrests for domestic violence (e.g., Felson & Messner, 1998). Some studies have found, however, that media reports often do not accurately reflect the common criminal history of male IPV homicide victims as well as other contextual issues (e.g., Taylor & Sorenson, 2002). The good news from the results obtained here is that when context was provided, it accurately reflected what is known from the research literature. The reporters who included this information are to be commended for helping educate readers about context in this instance, especially in light of the pressures many journalists face relative to deadlines.


948


The finding that the articles covering female perpetrators were also more likely to include opinions of the justice system, family, and friends was unexpected. This finding may indicate that journalists attempted to provide a more contextualized frame in articles concerning female IPV perpetrators to increase understanding of the relative rarity of this type of crime and the events leading up to it. On the other hand, this result contradicts what others have found when examining contextual factors provided in other media accounts. It is hoped that this contradiction may have occurred because of improved reporting by journalists in recent years. Ethnicity, occupation, education, and income. Analyses indicated that minimal demographic information was given about victims in the text of each article. This led to the inability to test whether or not our sample also exhibited similar demographic effects found in previous media studies (e.g., Sorenson et al., 1998). As noted previously, a lack of demographic information dehumanizes the problem. The limited information provided is not surprising given other research that suggests articles about IPV homicides are less likely to have humanistic frames and to have fewer details than other types of crime stories (Bullock & Cubert, 2002; Sorenson et al., 1998; Taylor & Sorenson, 2002).

Gender Coverage. There were no significant differences in word count between male- and female-perpetrated homicides, indicating that these initial stories were given the same amount of space regardless of gender of the perpetrator. Thus there was no empirical support for the expectation that female-perpetrated homicides were viewed as more sensational and newsworthy by journalists. Instead, both male- and female-perpetrated homicides received similar amounts of coverage in contrast to past research (e.g., Taylor & Sorenson, 2002). It is possible that reporters are no longer overfocusing on female-perpetrated homicides and instead are reporting both types of crimes with equal coverage. Explanations for homicides. Results indicate that, contrary to findings elsewhere, gender was not significantly associated overall with the number or type of explanations for the perpetrator’s actions although they did trend toward significance. These results provide some support that IPV homicide is being reported in an unbiased manner although trends in the expected direction weaken this assumption. It appears that journalists may no longer be justifying male perpetrator behaviors as strongly as before, at least in initial reports of IPV homicides within the present sample. Article depictions. Overall, few articles provided positive or negative depictions of the perpetrator or victim, suggesting that neutral reporting was the norm for this sample. This type of reporting is often considered unbiased and is therefore commended. However, neutrality and vagueness can conceal the context in which the crime occurred and therefore the social context of IPV homicide may remain obscured. Even though the amount of information was minimal in the majority of articles, when journalists did provide context, it appears they did so in an unbiased manner, contrary to one of the original hypotheses suggesting that exculpatory information would be more likely to appear in articles concerning male perpetrators. Instead, results suggest that the



949

majority of negative descriptors were used to describe the person at fault in the situation regardless of gender. It should also be noted that articles appeared to place more blame on males regardless of whether they were the victim or the perpetrator, implying that journalists may be quite accurate in portraying to the public that females are often acting in selfdefense as forensic research has illustrated and therefore should not be taken as evidence of gender bias against males. Contrary to the hypothesis that female victims would be portrayed more negatively, it was found that female victims were seen as innocent 92.9% of the time. Female victims were also more likely to be portrayed as warm and loving, were never described as cold or abusive, and were not seen as the cause of perpetrators’ actions when compared to male victims. These results suggest that blaming the female victim did not occur in the majority of articles. This also suggests that journalists, even under pressure to get the first report into print, are making strides in their reports of IPV homicide when it comes to accurately portraying circumstances to the public.

Limitations and Suggestions for Future Research One of the major strengths of the present effort was also a weakness: only the first article to appear in print was examined in-depth. By choosing only the first article, we were able to capture all the pressures and possibly negative influences with which journalists must deal to get their articles into print quickly. However, the analysis of only the first article concerning the homicide limited our results. For example, follow-up articles were not examined to see if, with more time, journalists may have accessed other information to provide greater context. It is left to others to provide a longitudinal look at follow-up articles because the focus of this effort was on the most pressured situation that journalists face—getting the initial article in before deadline in competition with other journalists. Another possible limitation was the coding method. Domestic violence is a controversial issue and gender issues are emphasized. In our study, both raters were female. Rater gender may have influenced the coding of articles although checks and comparisons were in place to help control such influence. Nevertheless, an increase in the number of raters and varying the gender of raters may have increased the reliability and validity of the coding of the articles. A larger sample size may also have helped increase the power of the study. In addition, instead of using newspaper articles obtained via Lexis-Nexis Academia, perhaps Internet news sources provided by newspaper conglomerates could have been included to allow coders to view pictures associated with the articles that were unavailable otherwise. Analyzing pictures and other graphics could have provided more data with respect to ethnicity and race as well as the individual’s portrayal in the picture.

Conclusion It is crucial for journalists to portray IPV homicides in an unbiased and accurate manner within initial newspaper reports in order for the community to understand the severity and


950


extent of the problem. Findings from the present study pertaining to how females are portrayed are promising. The fact that female victims were not being blamed for their own deaths suggests that journalists are attempting to help the public accurately perceive IPV violence. Hopefully, this is evidence of a trend toward a more consistent and unbiased portrayal of IPV homicide within the media that has the potential to eventually change society’s perceptions as well. However, results also suggest that there is room for improvement. By not referencing domestic violence in a larger context within initial articles, reporters are doing a disservice to victims of IPV homicide. Facts about and examples of domestic violence and its prevalence are crucial when reporting these crimes because homicide of an intimate partner is the ultimate form of domestic violence.

Acknowledgments The authors would like to thank the reviewers for their valuable feedback.

Declaration of Conflicting Interests The authors declared no potential conflicts of interest with respect to the authorship and/or publication of this article.

Funding The authors received no financial support for the research and/or authorship of this article.

References Bond-Maupin, L. (1998). “That wasn’t even me they showed”: Women as criminals on America’s Most Wanted. Violence Against Women, 4, 30-44. Bullock, C., & Cubert, J. (2002). Coverage of domestic violence fatalities by newspapers in Washington State. Journal of Interpersonal Violence, 17, 475-499. Carll, E. (1999). Violence in our lives: Impact on workplace, home, and community. Boston: Allyn & Bacon. Chibnall, S. (1977). Law-and-order news. London: Tavistock. Coleman, R., & Thorson, E. (2002). The effects of news stories that put crime and violence in context: Testing the public health model of reporting. Journal of Health Communication, 7, 401-425. Consalvo, M. (1998). “3 shot dead in courthouse:” Examining news coverage of domestic violence and mail-order brides. Women’s Studies in Communication, 21, 188-211. Curran, J., Gurevitch, M., & Woollacott, J. (Eds.). (1979). Mass communication and society. Beverly Hills, CA: Sage. Delaware Coalition Against Domestic Violence. (2006). Media study: Key findings. In Domestic violence media guide (pp. 23-29). Wilmington, DE: Author. Retrieved May 27, 2008, from http://www.dcadv.org/04resources/DCADV%20Media%20Guide.pdf Durose, M. R., Harlow, C. W., Langan, P. A., Motivans, M., Rantala, R. R., & Smith, E. L. (2005). Family violence statistics including statistics on strangers and acquaintances [NCJ 207846]. Washington, DC: U.S. Department of Justice, Bureau of Justice Statistics.



951

Felson, R., & Messner, S. (1998). Disentangling the effect of gender and intimacy on victim precipitation in homicide. Criminology, 36, 405-423. George, D., & Mallery, P. (1999). SPSS for Windows: Step by step. Boston: Allyn & Bacon. Gilliam, F., & Iyengar, S. (2000). Prime suspects: The influence of local television news on the viewing public. American Journal of Political Science, 44, 560-573. Gilliam, F., & Iyengar, S. (2005). Super-predators or victims of societal neglect? Framing effects in juvenile crime coverage. In K. Callaghan & F. Schnell (Eds.), The framing of American politics (pp. 148-166). Pittsburgh, PA: University of Pittsburgh Press. Hirsch, R. P., & Riegelman, R. K. (1992). Statistical first aid: Interpretation of health research data. Cambridge, MA: Blackwell Science. Holbert, R. L., Shah, D. V., & Kwak, N. (2004). Fear, authority, and justice: Crime-related TV viewing and endorsements of capital punishment and gun ownership. Journalism & Mass Communication Quarterly, 81, 343-363. Jewkes, Y. (2004). Media and crime. Thousand Oaks, CA: Sage. McEvoy, K. (1996). Newspapers and crime: Narrative and the construction of identity. In J. Morrison & C. Bell (Eds.), Tall stories? Reading law and literature (pp.179-200). Dartmouth, UK: Aldershot. McManus, J., & Dorfman, L. (2003). Distracted by drama: How California newspapers portray intimate partner violence (Issue 13). Berkeley, CA: Berkeley Media Studies Group. Retrieved May 25, 2008, from http://www.ncdsv.org/images/DistractedbyDramaHowCA Newspapers.pdf Meyers, M. (1994). News of battering. Journal of Communication, 44, 47-63. Meyers, M. (1997). News coverage of violence against women: Engendering blame. London: Sage. Rhode Island Coalition Against Domestic Violence (2000). Domestic violence: A handbook for journalists. Warwick, RI: Author. Sheley, J., & Ashkins, C. (1981). Crime, crime news, and crime views. Public Opinion Quarterly, 45, 492-506. Sorenson, S., Manz, J., & Berk, R. (1998). News media coverage and the epidemiology of homicide. American Journal of Public Health, 88, 1510-1514. Snell, T. L., & Morton, D. C. (1994). Women in prison. Washington, DC: Bureau of Justice Statistics. Taylor, C., & Sorenson, S. (2002). The nature of newspaper coverage of homicide. Injury Prevention, 8, 121-128. Wahl, O., Hanrahan, E., Karl, K., Lasher, E., & Swaye, J. (2007). The depiction of mental illnesses in children’s television programs. Journal of Community Psychology, 35, 121-133. Websdale, N. (1998). Rural woman battering and the justice system. Thousand Oaks, CA: Sage. Websdale, N., & Alvarez, A. (1997). Forensic journalism as patriarchal ideology: The newspaper construction of homicide-suicide. In F. Y. Bailey & D. C. Hale (Eds.), Popular culture, crime and justice (pp. 123-141). Belmont, CA: Wadsworth.


952


Bios Jessica A. Wozniak, PsyD (University of Hartford) is an adjunct professor at the University of Hartford Graduate Institute of Professional Psychology in Hartford, CT. She is also a Project Coordinator for the Family Advocacy Center at Baystate Children’s Hospital in Springfield, MA. Her areas of interest and specialization include psychological assessment, trauma, and abuse. Kathy A. McCloskey, PhD (Columbia Pacific), PsyD (Wright State University), Diplomate of the American Board of Professional Psychology in Clinical Psychology (ABPP), is an associate professor at the University of Hartford Graduate Institute of Professional Psychology in Hartford, CT. Her specialties include domestic violence, trauma, forensic populations, and the training of doctoral-level clinical psychologists.
