Domestic Violence against Women in the Brazilian Media: Study of Social Representations

Abstract This study examined the content published by the newspaper Folha de São Paulo regarding domestic violence before and after Law 11340/06, commonly known as Maria da Penha Law, came into force. A retrospective and comparative documentary research analyzed 3408 news reports published between 1994 and 2018. Divided into two corpora, ‘12 years before’ and ‘12 years after’ the Law, the material was analyzed using the IRaMuTeQ software and Descending Hierarchical Classification. The first corpus included news reports on cases involving celebrities and little about ordinary people. It also covered the feminist struggle to reduce domestic violence, focusing on specialized police stations and shelters. The second corpus included reports on the achievements generated by the Law and its challenges, pointing out the need to regard the law as more than a punitive instrument, addressing its preventive and care spheres. Problematizing how the media discloses such law is paramount, since the content divulged affects the construction of social representations.

As an ever-present phenomenon in human life, violence is considered one of the main causes of death worldwide (Cerqueira et al., 2019;Dahlberg & Krug, 2007).Specialized literature presents specific concepts regarding violence, such as violence against women, domestic violence and intrafamily violence, which overlap despite their particular aspects regarding the relationship with the male perpetrator and the victim's gender (Krenkel, 2014).
Violence against women defines victimization perpetrated against someone for being female and the relationship between the victim and the male perpetrator is broader, encompassing family members and strangers (Teles & Melo, 2017).Domestic violence occurs when the male perpetrator has an affective relationship with the victim, and can even be a family acquaintance.Intrafamily violence, in turn, specifically requires a family bond between victim and the male perpetrator.The latter two target only women, but children, older adults and adolescents (Saffioti, 2002).Our study, however, focuses on domestic violence against women perpetrated by a spouse or ex-spouse, group with the highest rate of authorship (Bueno et al., 2021).
According to Law 11340/06, domestic violence against women is any gender-based act or omission that causes death, injury, or suffering (physical, sexual or psychological), as well as moral or material damage.Discussed since the 1970s by feminists and researchers who sought to unveil the naturalization of household violence (Blay, 2003;Pinto & Gago, 2012), the phenomenon gained greater notoriety only in 2002, when Brazil signed the Convention on the Elimination of All Forms of Discrimination Against Women (CEDAW), the first international treaty aimed at the human rights of women (Lima & Peterke, 2010;Pimentel, 1979).
Based on this mechanism, Maria da Penha Maia Fernandes addressed an international complaint and Brazil created legislation to prevent and eradicate domestic violence against women (Azeredo & Weingartner, 2015).According to the latest Atlas of Violence, however, 3,737 women were murdered in 2019, a 7.3% reduction compared to 2018, whereas records of Violent Deaths from Undetermined Causes (MVCI) grew 35.2% from 2018 to 2019.Note that 33.3% of deaths in 2019 occurred inside the victim's home, an important indicator for the number of femicides (Cerqueira et al., 2021).
In western societies, this phenomenon is influenced by a patriarchal culture that shapes individuals according to gender stereotypes.Historically, this system has restricted participation in private and public spheres to man, contributing to the oppression and subordination of women (Saffioti, 1999).Rather than biological, this oppression is culturally established as a form of colonization (Sant'Anna & Penso, 2017).Hence, the Social Representations Theory can contribute to understanding this phenomenon, as social representations are culturally constructed and anchored in shared knowledge.As part of an individual scope under which the construction of femininities occurs, social representations can influence social practices (Amâncio, 2001;Silva, 2013).
A form of socially elaborated knowledge shared within groups, social representations fulfil practical objectives and contribute to the constructed reality of a certain group.Opposite scientific knowledge, this type of knowledge can also be called 'common sense' knowledge (Jodelet, 2001).Representations encompass common knowledge, establishing individuals as producers of meanings (Camargo, Scholösser, & Giacomozzi. 2018) who create their social reality.
According to Moscovici (1978), the communication process is fundamental for building thought and social representations.As such, the media, as an important mass communication tool and source of information, is therefore of interest to research on social representations (Castro, Aguiar, Berri, & Camargo, 2016), since it not only influences the construction of social representations, but also people's stance on issues (Camargo, 2003).
Widely used since the 1990s, the term media is relatively recent in Brazilian research and lacks a consensual definition in Communication Studies.Until 2004, however, media referred to an extension or natural consequence of a set of means of communication.Specialized publications sometimes used the term as synonymous with journalism, the mainstream press, the press, and the means and vehicle of communication (Guazina, 2007).
Traditionally studied separately from culture and society, previous research on the media focused only on the effects that their messages have on the subject and institutions.However, this distinction can no longer be considered in a contemporary society so completely permeated by the media.We must therefore comprehend the changes in character, function and structure that the media's ubiquity brings about in social institutions and cultural processes (Hjarvard, 2008).
Until the 1970s, Communication Studies argued that certain media messages had behavioral, limited, and short-term effects.By this decade research started viewing the media as a potential knowledge producer, tool for reading the world and responsible for influencing the public agenda.Accordingly, it came to be seen as an autonomous producer of knowledge about reality and meanings (Guazina, 2007;Camargo & Bousfield, 2011).
Mass media communication (Moscovici, 1981) conveys ideas and information according to values upheld by the diverse, dispersed groups that compose society.It thus informs the different points of view and determines common interests, simultaneously creating and disintegrating social solidarities (Beltrão & Quirino, 1986).Everyday mass communication bombards groups with information to change, create or crystallize their attitudes and opinions (Alexandre, 2001).But each group or sector of the population appropriates this information according to their social status and schooling level (Moscovici, 1978).
Mass communication also aims to relay the greatest possible number of messages to the largest number of people by newspapers, magazines, radio, cinema, television, and posters, among others (Kientz, 1973).Its recipients are anonymous individuals inserted in social groups that influence their attitudes, opinions, reaction and interpretation of the message received.Thus, it is the group and cultural context to which the individual belongs that determine their possibilities of modifying a given behavior as a result of the message received (Furtado, 2009).
For Moscovici (2003), communication plays a key role in transforming social representations by reciprocal influences between subjects.Information broadcasting starts by the science being disclosed to the media and then disseminated to social groups, which interpret the message and comment with other groups and among themselves.This process is called a rumor, and allows to identify the exact moment when a social representation is created.Media messages are usually very complex and abundant or simple and elliptical.Thus, recipients perform information reduction or inference, that is, they focus on the peculiar or strange elements in a message.This same process occurs in objectifying social representations (Clémence, Green, & Courvoisier, 2011).
Research on social communication shows that the media's functions depend on the existing means.Once responsible for disseminating news by transmitting the events, newspapers currently contend with social media already informing people of said events and must therefore assume a new role, that of carefully divulging the particularities of the facts (Maletzke, 1976).Such an ever-changing media environment produces ever-changing social groups.As the media and new technology provides social groups with a vast network of sensory stimuli and triggers, subjects are afforded new ways of perceiving, feeling, intuiting and thinking (Mitozo, Massuchin, & Carvalho, 2017).Moscovici (1978) argued that human beings are not passive machines that record messages and react to stimuli; rather, we all have imagination and the will to interpret the world to which we belong.When representing something, the subject is not simply unfolding, repeating or reproducing it, but recreating, editing and modifying it (Moscovici, 1978).Moreover, social representations fulfill the function of knowing, as it provides knowledge so one can assimilate and interact with values and beliefs; of identity, allowing individuals to feel like they belong to a group and define group identity; of orientation, acting as a guide for social action and practices; and the function of justifying behaviors and positions.Thus, representations uphold and justify social differentiation, allowing groups to stereotype, discriminate against, and distance themselves from the Other (Abric, 1998).
The media appears therefore to cultivate an important link with the groups.By influencing the construction of social representations, it concomitantly provides knowledge, strengthening group identity, controlling practices, and justifying stances on a given topic (Abric, 1998).Consequently, the media can also reproduce prejudiced discourses regarding domestic violence, or contribute with messages that strengthen women's groups.Hence the importance of identifying the arguments conveyed by news articles on domestic violence (Camargo, 2003).
A study by Miranda, Santos, and Anjos (2017) on feminicide in Tocantins investigated online media news published in 2016.Using the terms femicide, crime of passion, murder of women, family violence, domestic violence, violence against women and death of women, the authors found that 15 media outlets published debates, initiatives and research on violence against women.Interestingly, publications on homicide by partners or ex-partners never used the term femicide.The authors also reflected on the media's contribution to the cultural change necessary to combat this type of crime (Miranda, Santos, & Anjos, 2017).
A study on violence against women in the state of Pará found that most news reports described the action simply as physical violence (78.10%).As for the perpetrator, most were acquaintances (19.61%), spouses (19.44%), and ex-partners (13.84%) (Augusto, Lima, Sena, Silva, Gomes, & Santos, 2015).This shows a media interest in relaying news that involve physical violence, and consequently that most of these violent acts are committed by men who have had or have a marital relationship with the victim.
Thus, this study examined the content published by the newspaper Folha de S. Paulo regarding domestic violence against women before and after Law 11340/06 (Maria da Penha) came into force.Its importance is justified by how the media influences the construction of social representations and social practices.

Procedures
For documentary search, we selected the Folha de S. Paulo newspaper, a media outlet of great national visibility with publications since 1921 available online (Vieira, 2018).For unlimited access to the news reports, the researcher made a digital subscription.Documentary search was performed using 'domestic violence' as a descriptor.Inclusion criterion consisted of news discussing domestic violence against women in marital relationships published from 1994 to 2018, so that we could analyze reports from 12 years prior and 12 years after the creation of Law 11340/06.Reader's comments, image of the day and photography sections1 were excluded.

Data analysis
After organizing the selected news reports into two corpora for comparison, the obtained data were analyzed using the IRaMuTeQ software, applying lexical analysis of text segments which later underwent Descending Hierarchical Classification (DHC).DHC creates classes of text segments (TS) containing similar vocabulary similar (Camargo & Justo, 2013).

Results
Documentary search performed on the Folha de S. Paulo newspaper online platform for news reports published between 1994 and 2018, using the descriptor 'domestic violence' obtained 3408 news.After reading the reports and applying the inclusion and exclusion criteria, we selected 848 publications for analysis.The year 2016 had the highest number of news published (78 reports), followed by 2017 (63 reports) and 2010 (62 reports) (Figure 1).The lowest number was found for 1997, with only one report.In 2006, year that Law 11340/06 was created, 25 news were published on the topic.Class 1, separated from classes 4 and 5 in a second analysis, contained more general issues related to the phenomenon of domestic violence in Brazil.Classes 4 and 5 gathered reports about research on the topic and news about the delegacia da mulher (police station for women), respectively.Figure 2 shows the DHC classes with their respective words, frequencies and chi square values.After manually selecting the first 20 characteristic words of each class, their significant association was verified by the chi-squared test.A small number of news reported cases of ordinary Brazilians, making an articulation between fantasy and real life: Domestic violence against women in the Brazilian media When I see him, I remember the things he did to me, when I begged him not to hit me.But he didn't listen to me, she said.It was 2 months ago that Neila left home after a fight with her husband and never went back (rep. 732, sec. Daily life, year 2003).
This class also addressed cases of celebrities, but in less instances than class 2: "In 2001, irritated by phone calls he received saying he was being betrayed, Corrales punched his wife, Maria.Worse, she was pregnant with the couple's third child" (rep.700, sec.Sports, year 2004).
Class 1 consisted of 184 TS and was associated with the year 2000 and the 'More' section.Labeled 'problematizations,' it included discussions about domestic violence and related aspects such as religion, as shown by the following excerpt: "Until the Constitution of 88, the man was the head of conjugal society in Brazil.Finally, there is the influence of Catholicism in Latin cultures.Our religion sometimes describes the woman as a dangerous element and sometimes as the virgin" (rep.818, sec.Illustrated, year 1998).
Another related factor is how this phenomenon is naturalized by culture: Against this background, the fact that a man hits a woman is seen as natural.Olivia recalls that it was not until the 1970s that women began to denounce this domestic aggression, launching campaigns such as 'those who love do not kill' (rep.760, sec.Daily life, year 2001).In 2001, after 18 years of the practice of the crime, in an unprecedented decision, the Inter-American Commission condemned the Brazilian State for negligence and omission in relation to domestic violence, recommending that the State, among other measures, continue and intensify the reform process in order to break with the tolerance and discriminatory treatment of the state regarding domestic violence against women in Brazil (rep. 740, sec. Opinion, year 2002).Finally, class 5 consisted of 368 TS and was associated with 1998 and the Daily life, Ribeirão and Campinas sections.It gathered content related to institutions that assist women in situations of domestic violence, and was thus labeled 'secondary network': Promises of behavior change, financial dependence and resistance to putting people with whom she has lived for more than 10 years behind bars are some of the reasons that help to understand why the battered woman still hesitates to go to a police station.A survey commissioned by the IBGE for the Secretariat, 15 years after the first Police Station for Women was implemented in the country, the majority of victims of intrafamily violence continue to resist pressing charges against their partner.Even among those that choose to report them, many end up dropping the complaint before the process is concluded (rep.795, sec.Daily life, year 1999).A total of 9 women and 29 children are now staying in a shelter in São Paulo, the address of which is kept secret by the police.These women are in hiding because they have been threatened with death by their husbands or companions.They were insulted, beaten and raped.They decided to run away from home with their children.The hiding place is actually a shelter maintained by the Public Security Department (rep.845, sec.Daily life, year 1994).
As observed, this first corpus focuses mainly on celebrities and TV characters, giving little attention to everyday women in situations of violence and incipiently citing resources created to serve these women.Some news reports feature problematizations regarding this topic, focusing on police stations and shelters as resources for protection given the lack of a specific legislation.Moreover, they showed concern about the fact that some women withdraw the complaint or do not file it.
The corpus 'Reports 12 years after Law 11340/06' consisted of 672 news reports from January 2006 to December 2018 divided into 7548 text segments (TS), of which 87.35% were retained in the analysis.This corpus was divided into two sub-corpora, generating four classes.Classes 2 and 3 covered cases of famous people that were widely publicized in the media and the daily lives of people suffering domestic violence.
Classes 1 and 4 gathered elements related to legislation, studies, and discussions about domestic violence in Brazil.Figure 3 shows the DHC classes with their respective words, frequency and chi-square values, which were also selected as in Figure 2. Class 2 consisted of 125 TS and was associated with the Daily life section.It was labeled 'ordinary people' and gathered news about ordinary people who committed or suffered domestic violence: "A man was arrested in the city of Itaguaí RJ on suspicion of having the woman killed with 2 shots last Wednesday, the 24 th " (rep.523, sec.Daily life, year 2010).It also feature some US cases: "American Davon Crawford, 33, who this Thursday 5 th killed his wife, sister-in-law and 3 children in a duplex in Cleveland, Ohio, committed suicide when the police tried to arrest him, official sources reported" (rep. 578, sec. World, year 2009).
Class 1, labeled 'problematization and studies,' consisted of 174 TS and was associated with the Columns and Blogs and Social Entrepreneur sections.It included theoretical discussions and research data on domestic violence: "Athenas.Luis Felipe Miranda, president of Avon in Brazil, stated that just the fact of talking about, disseminating and reflecting on the research data is already extremely important.This is not a topic to punish men, but to provide knowledge" (rep.440, sec.Social Entrepreneur, year 2011); "In 70 percent of cases of domestic violence against women, the aggressor is the partner or the spouse" (rep.301, sec.Social Entrepreneur, year 2013).
Finally, class 4 consisted of 165 TS and was associated with the Daily life section and the month of August, when Law 11340/06 was implemented.Aptly labeled 'legislation,' it included news addressing issues related to laws that protect women, pointing out challenges and achievements: In the case of Marcia, the now ex-husband left after the ban.He disappeared; he was intimidated when he learned he could be arrested if he failed to comply with the measure.It's not always like that.Enforcing the protective measures is one of the main obstacles of the Maria da Penha Law (rep.195, sec.Daily life, year 2016).A woman who, in order not to suffer domestic violence, separates from her husband or partner can receive financial assistance of at least R$622 per month for one year.This is what a bill approved yesterday by the Senate Economic Affairs Committee establishes (rep.351, sec.Daily life, year 2012).
Comparative analysis shows that before Law 11340/06, the media outlet mostly focused on cases involving national and international celebrities, given how little this topic was debated in the Brazilian context.After the law came into force, we see a significant increase in news reporting on cases of ordinary people, not only celebrities, of Brazilians and foreigners, as it became important to show the reason why the Law was created, anchoring it in the cases of daily life in our country.Mention of institutions geared towards supporting victims indicated possibilities of enforcing the Maria da Penha Law in institutional practices.

Discussion
Considering that media outlets influence the formation of social representations (Moscovici, 1978), this study investigated the content published by the Brazilian newspaper Folha de S. Paulo regarding domestic violence against women.First, we highlight the difference in the number of news reports before (175 reports) and after (672 reports) implementation of Law 11340/06.Created to prevent and eradicate domestic violence, its enactment may have given greater visibility to a topic until then naturalized and trivialized (Pinto & Gago, 2012).
For Castro, Seixas, Neca, and Bettencourt (2018), the media is an important tool for legitimizing new laws.Despite the common belief that domestic violence is a private issue and therefore state intervention is not necessary, when faced with media news about the legislation groups begin their own debates on the topic, guiding their communications according to the values circulated by the media (Castro, Seixas, Neca, & Bettencourt, 2018), thus leading to concrete actions forseen by the Law (Bertoldo & Castro, 2019).
Class 2 'cases of celebrity,' however, showed that despite this naturalization, the celebrity cases disclosed already caused repercussions and debates in society.As with O. J. Simpson, a former US football player and actor, who in 1994 was accused of stabbing his ex-wife Nicole Brown and her friend Ronald Goldman, and acquitted in 1995 after a long trial that received great media attention (Souza & Andrade, 2019).This media coverage was essential for social groups, since the media disseminates information seeking to change, create, and crystallize their attitudes and opinions (Alexandre, 2001).However, each social group will appropriate information according to their social status and schooling level (Moscovici, 1978).
Similarly, class 3 'real and fantasy characters' featured the social repercussions of violent characters in soap operas.Blay (2003) states that television programs usually dramatize passionate crimes such as domestic violence, providing a double message: they criminalize the perpetrator while romanticizing this type of crime.Cases involving famous athletes, for example, were followed by people as entertainment (Souza & Andrade, 2019).This same class included reports about ordinary people, but in fewer numbers.This lack of of reporting on domestic violence in Brazil stems from the pervasive belief that marital conflict is something private and natural (Saffioti, 1999).For many years this same belief ruled out the need for state intervention, resulting in unpunished assault and death of women and no preventive measures (Pinto & Gago, 2012).
Class 1 'problematizations' highlights the power and influence of religious dogmas and attitudes that tend to consolidate patriarchy (Cordovil, 2016).Interested in reinforcing the idea of the sacred family and eternal love (Saffioti, 2004), religion pressures many women to remain in the cycle of violence and maintain their family (Santi, Nakano, & Santiere, 2010).This class also confirms the trivialization, naturalization and legitimization of domestic violence against women due to gendered social representations (Saffioti, 1999;Teles & Melo, 2017).Rather than biological differences, these differences are based on culturally constructed stereotypes socially imposed by the patriarchy which validates the relations of domination and violence (Santos, Oliveira, Viegas, Ramos, Policarpo, & Silveira, 2018;Sant'Anna & Penso, 2017).Such social representations also reinforce the sexual asymmetry, favoring male sexuality, justifying the supposed need and instinctive biological impulse of women, thus supporting sexual violence in marital relationships (Santos et al., 2018).
However, this class also reaffirms the history of feminist struggles started in the 1970s against the naturalization of domestic violence (Blay, 2003;Pinto & Gago, 2012), which was only consolidated in 1985 when the first specialized police station for women was created in São Paulo (Pasinato & Santos, 2008).In 2002, Brazil signed the first international treaty aimed at supporting the human rights of women, called CEDAW (Lima & Peterke, 2010), through which Maria da Penha Maia Fernandes, a woman from Ceará, made an international complaint regarding the lack of punishment for her ex-husband.This resulted in the approval of Law 11.340/2006, known as the Maria da Penha Law (Azeredo & Weingartner, 2015).
These achievements also draw heavily on scientific studies on domestic violence against women to prove the need for changes in legislation and the social representations regarding this issue.Class 4 'statistics' covers these studies, showing that despite little broad debates on the topic, people were already looking to make changes.These news reports are important because they present scientific data in a more simplified manner to reach different social groups (Moscovici, 1978).In other words, it is by the rumor process that complex or simple scientific information about domestic violence against women disseminated by the media is assimilated by the individual through information reduction or inference, thus creating a social representation (Clémence, Green, & Courvoisier, 2011).
Finally, class 5 'secondary network' covered the issues faced by specialized police stations for women prior to Law 11340/06, including reports on the fact that some women withdraw their complaint or do not make it, hoping that their partner would change (Frugoli, Miskolci, Signorelli, & Pereira, 2019) or due to financial dependence (Rolim & Falcke, 2017;Souza & Resende, 2018).Despite specific legislation created to prevent and eradicate domestic violence, this is a recurrent issue (Souza & Resende, 2018).
Besides hope of behavioral change and financial issues, external pressures exerted by the family and society for women to preserve and maintain "the sanctity of marriage" also contributes to women withdrawing or not filling complaints (Saffioti, 2004;Santi et al., 2010).Moreover, these women often suffer death threats from the perpetrator (Saffioti, 2004) and feel ashamed to share what happened, as they are sometimes stigmatized and blamed by professionals (Souza & Rezende, 2018).
This class also covered the secondary network, that is, institutions established to assist women in situations of violence (Grossi, Tavares, & Oliveira, 2008).Such dissemination was viewed as important because the specialized Police Station for Women created in 1985 was still little known (Pasinato & Santos, 2008;Saffioti, 2002).The news also cited shelter houses, safe and confidential places where women in situations of violence and at imminent risk of death are temporarily housed to ensure their physical and psychological integrity (Krenkel & Moré, 2017).
They provide women with a safe space to reflect on their experiences and strategies to resume their lives (Krenkel & Moré, 2015).Accordingly, this represents information dissemination, since the media can produce knowledge about reality (Camargo & Bousfield, 2011;Guazina, 2007).
We observed some similarities with the corpus 12 years after Law 11340/06.Chief among them the fact that even after the legislation, most news focused on reporting cases involving famous people, as seen in class 3 'famous people' (20.95%).Class 2 contained reports on ordinary people (17.01%), in Brazil and abroad.In other words, despite the greater visibility brought by the law to cases in Brazil, these reports were still in the minority, mirroing the invisibility of Brazilian women under domestic violence (Bálsamo, 2019).The most recent data on femicide, however, points out that 3,737 women were murdered in 2019 (Cerqueira et al., 2021) with little media repercussion.This stems from the fact that the media produces representations that determine interest on certain identities to the detriment of others, upholding systems of power and inequalities (Howarth, 2016).
Class 1 'problematization and studies' was similar to class 4 (Statistics) of the first corpus, but with a greater number of news reports due to the visibility generated by the legislation.Currently, many studies are conducted on domestic violence against women, showing the academic advances achieved (Bálsamo, 2019;Frugoli et al., 2019).However, according to the last issue of the Visible and Invisible study, 536 women were physically assaulted every hour in 2018 (Brazilian Forum on Public Security, 2019).In other words, the state must verify the applicability of these results in rescuing women in situations of violence and in stoping perpetrators.
Moreover, given its function of disclosing particularities with a careful and intensive interpretation of facts (Maletzke, 1976) to guide different perspectives and determine the common interests of individuals (Beltrão & Quirino, 1986), the media plays a key role in scientific dissemination.This is how science becomes part of everyday life, leaving laboratories and entering more simplified conversations and reading (Moscovici, 2001).
Finally, class 4 'legislation' covered the obstacles faced by women when invoking the law.Protective measures are not enough to prevent further aggression or even femicide.Studies emphasize that flaws in the application of protective measures go unnoticed due to difficulties in conducting their inspection (Cruz & Rosa, 2017;Mendes, Bitu, & Nóbrega, 2017).This shows that we currently lack tools to mitigate flaws in the Maria da Penha Law (11340/06) and to ensure its effectiveness (Campos, 2017;Somenzari, 2017).
But this class also addresses some improvements achieved by the legislation in meeting the needs of women (Santos & Medeiros, 2017).Nonetheless, despite also aiming to prevent domestic violence against women, the law often ends up being seen only as punitive (Souza, 2016).This may explain why the number of femicides continues to increase in Brazil (Cerqueira et al., 2019).A paradigmatic turn is needed in the Maria da Penha Law, one that favors prevention and assistance, treating women with dignity and humanity (Campos, 2017).
We must also change the social representations regarding Law 11340/06, pushing for it to be thought of and remembered mainly as a preventive law.And the media is fundamental in this process, since it reaches a great number of people (Kientz, 1973).Despite being part of a social group that influences their attitudes and opinions, and their individual reactions and interpretions of the message received, subjects can still change social representations and behavior (Furtado, 2009).

Final considerations
Our study found that differences between the publications of a traditional Brazilian media (newspaper Folha de S. Paulo) regarding violence against women before and after the Maria da Penha Law came into force, with the number of news reports on the topic increasing substantially after the law's creation.The textual corpus found was significantly larger 12 years after its implementation, indicating that greater visibility was afforded to domestic violence against women after international demands for changes regarding the prevention and punishment of this crime in Brazil.News reports published 12 years before Law 11340/06 focused mainly on celebrities and TV characters, revealing how little household violence was discussed in the country.Given the widespread social representation at the time that marital strife was something private, the Brazilian popular saying 'you don't take sides in a fight between husband and wife' applies here.Even after Law 11340/06 was created, the media often focused on famous people, corroborating the idea that the "bodies of ordinary Brazilian women" are invisible.
Moreover, reports from 12 years before the Law often featured religious arguments that defended more normative social representations, consolidating patriarchal practices.Interestingly, they also reaffirmed the history and struggles of the feminist movements to denaturalize domestic violence.Both corpora publicized studies on the topic, simplifying the information to reach different groups, thus generating debates and new social representations.
When discussing punishment or help to victims, news reports before the Law often mentioned the secondary networks available at the time, such as specialized police stations and shelter houses, making note of the concerning fact that women usually withdrew the complaint or did not make it.News published after Law 11340/06 came into force, in turn, focused on disclosing the achievements and challenges to be overcome.
Our study point to the need for a new look at the Maria da Penha law (11340/06), since it is still viewed and used for punitive, rather than preventive or assistance purposes.Consequently, support from the media is necessary for future changes in social representations regarding this legislation.An important limitations of this study was its sample size, since we selected only one media vehicle for analysis.Discussing the political stance of newspapers, Taschner (1992) states that Folha de S. Paulo, founded by a group of journalists in 1921, positioned itself against the rather conservative Estado de São Paulo newspaper by adopting a more liberal viewpoint.Despite controversies on this aspect, we must consider the above conditions in our analysis.Thus, further studies are necessary on this theme so the field can contribute to social changes.

Figure 1
Figure 1 Number of news reports by publication year.
50 different newspaper sections published news coverage on the topic, showing the variety produced by Folha de S. Paulo, most frequently Daily life (332 reports), Illustrated (113 reports), World (102 reports), Celebrities (55 reports), Columns and Blogs (31 reports), Opinions (31 reports), BBC Brasil (24 reports) and Power (20 reports).Other sections had a frequency of less than 10 news reports.For comparing data before and after Law 11340/06, we organized the content into two corpora: Reports 12 years before Law 11340/06 and Reports 12 years after Law 11340/06.Both underwent lexical analysis.The corpus 'Reports 12 years before Law 11340/06' consisted of 175 news reports from January 1994 to December 2005 divided into 1646 text segments (TS), of which 91.68% were retained in the analysis.DHC divided the corpus into two sub-corpora, generating five classes.Classes 2 and 3 presented elements referring to cases of domestic violence involving celebrities and other famous people that were widely publicized in the media.

Figure 2
Figure 2 Dendrogram of the corpus Reports 12 years before Law 11340/06.
Class 4 consisted of 242 TS and was associated with the years 2002 and 2005 and the Folha online section.It was labeled 'statistics' and gathered content related to scientific research data: A survey recently released by the WHO, an agency linked to the UN, highlighted that 27 percent of women living in the city of São Paulo and 34 percent of those living in the Zona da Mata of Pernambuco have been victims of domestic violence (rep.675, sec.Opinion, year 2005).

Figure 3
Dendrogram of the corpus Reports 12 years after Law 11340/06.These classes were labeled based on the associated TS content.Class 3 consisted of 123 TS and was associated with the Illustrated, Celebrities and Sports sections.Labeled 'famous,' it gathered cases of domestic violence committed or suffered by a personality, as shown by the excerpt: "The singer Rihanna and her ex-boyfriend Chris Brown, who was sentenced to 5 years of probation for assaulting her in 2009, returned to collaborate musically and rekindled the rumors of a reunion" (rep.405, sec.Illustrated, year 2012).