“But she was drunk”: Sexual violence and Blaming the victim “Mas ela estava bêbada”: Violência Sexual e Culpabilização da vítima “Pero ella estaba ebria”: violencia sexual y Culpabilización de la víctima

Various studies have been demonstrating the importance of patriarchal values in the structuring of society, especially with regard to sexual violence against women, pointing out that women themselves are still blamed for the violence that they suffer. The work presented here falls within this context and analyzes how the social representations of university students about women victims of sexual violence are structured (Study 1; N = 288) and how information about the victim's skin color can affect the judgment that is made about women victims of sexual violence (study 2; N = 99). Together, the results indicate that the woman was blamed for the violence she suffered and this becomes more evident in the situation in which the woman was Black.


RESUMEN
Varios estudios han demostrado la importancia de los valores patriarcales en la estructuración de la sociedad, especialmente en lo que respecta a la violencia sexual contra la mujer, señalando que aún se le culpa por la violencia que ha sufrido. El trabajo que aquí se presenta se enmarca en este contexto y analiza cómo se estructuran las representaciones sociales de los estudiantes universitarios sobre las mujeres víctimas de violencia sexual (Estudio 1; N = 288) y cómo la información sobre el color de piel de la víctima puede afectar el juicio que se hace sobre las mujeres víctimas de violencia sexual (estudio 2; N = 99). En conjunto, los resultados indican que se culpó a la mujer por la violencia que sufrió y esto es más evidente en la situación en la que la mujer era negra.
ccording to data from the Atlas da Violência (Atlas of Violence) published in 2019, sexual violence cases in Brazil broke the 2018 record of 66,041 cases, with 81.8% of the victims being women (Cerqueira, 2018). A recent event took place in December 2020, when a state congressman from Partido Cidadania (Citizen's Party) touched the breasts of a colleague, a state congresswoman from the PSOL party, A Psicología, Conocimiento y Sociedad -12(1), 84-109 (mayo-julio 2022) -Trabajos originales ISSN: 1688-7026 86 during a session of the Legislative Assembly of the State of São Paulo. If this type of episode takes place in a public agency, whose mission is precisely to legislate, among other things, about violence against women, we can easily imagine the situation in more private places.
Sexual violence is defined as any act or attempted sexual act, comments, attacks, or any other forms of harassment given its non-consensual character (WHO, 2017).
According to Bandeira (2014), in the context of violence against women, sexual violence is considered gender violence, having its origin in the imbalance of power and the division of roles between men and women, and in the domination of women's bodies. The concept of gender, on the other hand, was proposed to overcome the biological determinism related to the use of the term sex, or sexual differentiation, and to highlight the social construction of the identities of men and women, deconstructing the sexual division of roles based on social norms that designate what the supposed roles of men and women are.
Social norms are patterns of influence that the individual builds in interaction with other people or acquires from groups of which they are a part (Sherif, 1966). Thus, it is thought that social norms help people understand themselves in social situations in which they are involved, especially when they are young or uncertain (Cialdini & Trost, 1998). Given that, historically, women's social representations have been marked by exclusion and violence (Estramiana & Ruiz, 2006), it is clear that, despite some advances regarding the division of gender roles, attributions based on the idea of patriarchy are currently being culturally incorporated and reinforced in our society.
Several studies have shown how much patriarchal values still structure society, especially with regard to sexual violence against women (Hirt et al., 2017;Silva & de Oliveira Gomes, 2018), as happens for example, in situations where women themselves are blamed for being sexually assaulted. The study developed by Hockett, Smith, Klausing and Saucier (2016), for example, demonstrated that the blaming of the Psicología, Conocimiento y Sociedad -12(1), 84-109 (mayo-julio 2022) -Trabajos originales ISSN: 1688-7026 87 victim of sexual violence tends to increase as the victim deviates from the stereotype considered as being a "real" victim. The true victim would be the one who had not consumed alcohol, who was raped by a stranger in a quiet public space, who fought against the aggressor and who suffered visible physical and emotional injuries. Thus, victims who do not fit this stereotype are seen as possibly more guilty for what happened to them.
Thus, it is clear that the victim blaming comes from the acceptance and replication of concepts that normalize sexual violence against women, so that society itself raises questions that test the morality of the raped woman, such as: "What was she wearing?
What was she doing at this place, at this time? Was she drunk?", taking the focus off the aggressor and, in a way, legitimizing the victim's guilt.
Over the past 30 years, new approaches have been developed by some groups of feminists and Black activists, with the aim of analyzing the relationship between the gender and skin color categories. Most research adopts the intersectionality approach (Crenshaw, 1994), which was proposed with the aim of analyzing Black women's concomitant experiences with sexism and racism in the US. The central idea behind this approach is that all individuals occupy different positions in different hierarchical systems. Thus, this matrix of domination should be studied taking into consideration the multiple hierarchical systems that are interconnected, such as gender, race/ethnicity, and social class (Crenshaw, 1994).
A survey commissioned in 2016 by the Brazilian Public Security Forum (FBSP) to Datafolha identified that 33.3% of the Brazilian population believes that in cases of rape the victims themselves are guilty. In the same direction, the results of the survey "Sexual Violence -Perceptions and behaviors about sexual violence in Brazil", carried out by the Patrícia Galvão Institute in 2016, indicated that, when asked about the relation between alcohol consumption and sexual violence, 89% of the participants stated that being drunk is not a reason for a man to abuse a woman, but 20% think that having sex with an alcoholic colleague or girlfriend is not sexual violence.
Together, these surveys show that: a) violence against women is currently considered a serious social and public health problem and b) women are still blamed for the violence they suffer. The work presented here falls within this context, unfortunately so frequent in Brazil. Social representations are defined as knowledge built and shared socially (Moscovici & Lage, 1976) and that functions as a system of interpretation of reality that organizes the relations of individuals with the environment, influencing their behavior (Abric, 1994). Thus, analyzing the way in which the social representations of women victims of sexual violence are constituted is understood as having great importance for studies of this theme.
In view of the above, this work aims to investigate what are the social representations of university students about sexual violence against women. To this end, two studies were developed with quantitative and qualitative approaches that aimed to analyze, in Study 1, how the social representations of university students about women victims of sexual violence are structured and, in Study 2, to investigate how information about the skin color of the victim can affect the judgment of women who are victims of sexual violence.

Study 1
Discussing the issue of violence against women in the university context is relevant because, depending on the major area, these participants will be the future professionals who are likely to deal with these victims. Thus, we chose to have as participants students of the following major areas: physiotherapy, pharmacy, nursing, and social work. We used the Theory of Social Representations to support this study because they would have the function of determining the field of values or ideas shared by the groups, as well as enabling the understanding of the meanings that people attribute to the world and to themselves (Almeida, Santos, & Trindade, 2000). Thus, considering that Social Representations are responsible for assigning meaning to the real world and for guiding behaviors (Jodelet, 2001), this study sought to analyze how the social representations of university students about women victims of sexual violence are structured.

Participants
There were 288 university students from a public institution in the city of João Pessoa, PB, of which 149 were men, 138 women, and 1 did not report their gender. Participant ages ranged from 18 to 33 years, with a mean of 20.5 years of age (SD = 3.36).

Procedure and instrument
Initially, participants were instructed to read a story in which a situation of sexual violence against a young woman was reported. From the data previously presented in the research published by the Atlas of Violence and Patrícia Galvão Institute, the choice of this scenario was due to the fact that in the addressed context this type of violence is, in a way, more naturalized. The participants answered the instrument individually, administered collectively in the classroom. The story presented is described below: Eduardo had just started at the university and decided to go to a freshmen hazing event at a student fraternity, promoted by the upperclassmen from his department. When he gets there, he realizes that most of his classmates have drunk too much and are quite "happy". Eduardo pays attention to a girl who is quite drunk and is talking to an apparently sober young man. Eduardo notices that the boy is speaking into her ear and pointing out the way to one of the bathrooms in the house. Eduardo realizes that she resists entering, but ends up giving in and accompanies the boy, staggering. Minutes later, Eduardo sees the young man coming out of the bathroom alone, in a hurry, closing his pants, and looking from side to side somewhat suspiciously.
Following this scenario, participants answered a question about blaming the victim, adapted from the study by Katz, Merrilees, Hoxmeier and Motisi (2017): "If the girl hadn't been drinking she probably could have avoided this situation." The participants' task was to indicate the point on the scale that was closest to their opinion, ranging from 1 = totally disagree to 7 = totally agree. Soon after, the participants were asked to justify their position on the previous question. Finally, the participants answered a questionnaire containing sociodemographic data, with questions related to gender and age.

Data analysis
Sociodemographic data were analyzed with SPSS software, version 22, using frequency analyses. Discursive data were processed through textual analysis using the IRAMUTEQ (Interface de R pour les Analyses Multidimensionnelles de Textes et de Questionnaires) software, which is a program that allows the analysis of textual data through statistical and graphic resources (Ratinaud, 2009).
The data were analyzed using Descending Hierarchical Classification (DHC), a method that classifies small segments of text according to the vocabularies present in it and then breaks them down into their reduced forms, organizing them into Discursive Classes (Camargo & Justo, 2013). We emphasize that the variables used for our analyses were participant age and gender, as well as the attributed guilt measured using the Likert scale that varied from 1 = strongly disagree to 7 = strongly agree, so that the greater the assent, the greater the blaming.

Results and discussion
The DHC retained 80.19% of the total corpus, which was divided into two subcorporas composed of six discursive Classes. The first subcorpora agglomerated Classes 3 and 2, being called "Women and the ingestion of alcohol". The second, called "Women: victim or guilty?", aggregated Classes 5, 4, 1, and 6 ( Figure 1), which subsequently were subdivided into Classes 5 and 4 and, contrary to Classes 1 and 6. Class 1 was the most significant, corresponding to 20% of the corpus, being called "No is no". Class 4 represented 17.6% of the corpus and was called "Both have a share of the blame".
Class 2 expressed 16.4% of the corpus and was called "She was not guilty, but she could have avoided it". Class 5 also expressed 16.4% of the corpus and was called "She is guilty and could have avoided it". Class 3 expressed 15.3% of the corpus and was called "Nothing justifies it". Finally, Class 6 represented 14.1% of the corpus, called "No victim is guilty".

Figure1. Dendrogram of the Descending Hierarchical Classification
Initially, following the corpus partition logic, in the first subcorpora partition, "Women and the ingestion of alcohol", we have a general idea that the victim is not guilty for what happened, although Classes 3 and 2 are opposed due to the victim's relation to drinking. That is, while Class 3 subtly passes on the idea that drinking facilitated the occurrence of a crime and it would have been better to avoid it, Class 2 transmits the idea that, even though the woman ingested alcohol, that does not justify the crime committed against her and she should not be held responsible for the violence suffered.
Specifically, the analysis of Class 3, called "Nothing justifies it", presents statements that emphasize the idea that the fact that the woman ingested alcohol does not justify the action of the man, which could have happened even if she was sober. Despite this, it should not occur in either circumstance, since it is one of the main forms of violence against women typified by law. In this Class there was a low level of victim blaming, since the most representative variable was the level of attributed guilt, with level 1 standing out on a Likert scale from 1 to 7. Examples: "The situation has nothing to do with the girl being drunk or not, the problem in that situation is the boy who abused her. He could have abused her either way, drunk or not." (Male -21 years old, blaming: 1) "The girl has nothing to do with experiencing this situation, whether she drinks or not, it does not justify such an act". (Female -23 years old, blaming: 1) In Class 2, called "She was not guilty, but she could have avoided it", we have statements very close to those of the previous class, differentiating themselves by the fact that, while indicating that the woman was not guilty, it also states that she could have avoided experiencing that situation if she had not ingested alcohol. In this way, there is blaming of the victim, but in an implicit way since it presents an argument that goes in the direction of control and restriction, that is, that she could have avoided drinking, as this allows a risky situation for other scenarios. Examples: "The boy is to blame for the incident, but her level of drunkenness has contributed to making her an easier target." (Male -22 years old, blaming: 2) "She was not to blame for being in this situation, but if she was aware of what she was doing, she could have asked for help before." (Female -20 years old, blaming: 2) The general idea in the statements gathered in the opposite subcorpora, called "Women: victim or guilty?", is the analysis of the situation of blaming the victim herself.
In a subsequent subdivision, there is the opposition from Classes 5 and 4, which condense statements aimed at blaming the victim in an explicit and strong way, while Classes 1 and 6 defend that the guilt is never the victim's, but the aggressor's.
Class 5, as in the other Classes, also emphasized the issue of alcohol ingestion, being called "She was guilty and could have avoided it". However, here the statements more directly portray the woman's guilt, emphasizing the fact that she could have avoided it if she had been sober or if she had kept an awareness of reality. Also noteworthy is the high rate of blaming the victim, since the variable associated with this class was the level of attributed guilt, with level 5 standing out on a Likert scale from 1 to 7 (little blame and a lot of blame ). Examples: "If she were sober I believe she would have realized that the boy wanted to hurt her".
(Female -18 years old, blaming: 5) "In this situation, yes (she is guilty) because if the girl were sober she would know how to make her own decisions". (Male -23 years old, blaming: 7) Analyzing the statements of Class 4, "The man is at fault", it is clear that the blame is meant for the man, emphasizing that he took advantage of the woman's situation. It is important to point out that, despite the statements blaming the man, it is clear that there was a high level of blaming the woman. When analyzing the variables associated with this Class, the prevalence of responses from male participants was noted, as well as the level of guilt attributed, highlighting level 6, with the scale being evaluated in an interval from 1 to 7 (little blame and a lot of blame). Thus, it is understood that the participants indicated a high level of blame for the woman, but in their statements this blaming appeared implicitly, justifying this with the fact that she should not have ingested alcohol. Examples: "I believe the drink left her vulnerable, but she was not to blame for the abuse, he is to blame." (Male -18 years old, blaming: 6) "In my view, the boy took advantage of the drunk girl. If she were not drunk, she would never end up in a situation like this". (Male -22 years old, blaming: 7) At the other pole, in Class 1, called "No is no", the statements were based on the idea that, regardless of the situation the woman is in, drunk or not, the right to not want to have intimate relations with someone must be respected. They also portray the fact that the man took advantage of the situation, in which the woman was under the influence of alcohol, to violate her. In addition, they present the importance of other people's intervention when they perceive the situation, in order to avoid violence.

Examples of statements:
"The fact that she is drunk does not give the boy the right to do what he wants with her" (Male -19 years old, blaming: 1) "If the girl in question was in the mood to drink, she did not mean that people could take the liberty and take advantage of the situation". (Female -18 years old, blaming:

1)
Finally, in Class 6, "No victim is guilty", the statements present the idea that the victim is never guilty for experiencing a situation of violence, and has the right to enjoy their leisure without being harassed. Furthermore, the blame of the abusive man who can bring himself to rape a woman, regardless of her condition, was emphasized. In this Class there was a low rate of victim blaming, with the level of guilt of 2 standing out as an associated variable, on a Likert scale from 1 to 7 (little blame and a lot of blame), as well as the female gender. Examples: "The aggressor is the one to blame, not the victim, with her being at any stage of consciousness." (Female -19 years old, blaming: 1) "Being drunk or not, she would never need or deserve to go through such a situation." (Female -19 years old, blaming: 2)

Study 2
Several studies show that the phenomenon of blaming women for the violence they suffer is more frequent when they are Black than when they are White (Donovan, 2007;Lewis, Mendenhall, Harwood, & Browne Huntt, 2016). According to the Atlas of Violence (Cerqueira, 2018), in 2016, 22,918 cases of rape were registered by the Unified Health System (Sistema Único de Saúde -SUS), in which 34% of the victims were White and more than 54% were Black. Here it is important to note that, historically, in Brazil, Black women have always been at the bottom of the social pyramid, either due to lower wages or due to being seen as a sexual object. In fact, Gilberto Freyre, in his book Casa Grande e Senzala, had already pointed out the abuses practiced by White owners against Black women, who served both as work and sexual objects for the owners and their children (Freyre, 2000).
In this study we seek to investigate how information about the victim's skin color can affect the judgment that is made about women victims of sexual violence. For that, we used the theory of the central core (or nucleus) proposed by Abric (1994). This theory proposes that all social representation is constituted as an organized and structured set of information, beliefs, opinions, and attitudes, composed of two subsystems: the central and the peripheral (Abric, 1994). The choice of this theory is due to the fact that it makes it possible to compare social representations between two different groups or from the same group at two different times. Thus, it can be considered that the representations about that particular object are different if the compositions of the central cores are different (Sá, 1996).
Thus, there would be a central system, formed by the central core of representation, composed of stable elements and with a normative and functional nature (Sá, 1996), and constituting the essential element of representation (Abric, 1994). A peripheral system, constituted by more flexible elements, allows the simpler incorporation of the changes that occur in that context, to the group's representations (Abric, 1994). The central nucleus has the function of generating the basic meaning of representation, determining the organization of all elements, and is relatively immune to change (Sá, 1996). The peripheral system is responsible for making the connection between concrete reality and the central system (Abric, 1994). In addition, it would be more sensitive to changes, aiming at adaptations to protect the elements of the central nucleus (Sá, 1996).

Participants
Participants included 99 university students from a public institution in the city of João Pessoa, PB, of which 50 were men, 46 women, and 3 did not report their gender.
Participant ages ranged from 18 to 28 years, with a mean of 20.21 years of age (SD = 2.18). The sample consisted of students from the following major areas: physiotherapy, pharmacy, nursing, and physical education.

Instruments and Procedures
The scenario used in this study was the same as in Study 1, being changed only by using the two experimental conditions where the girl's skin color varied between White and Black. After describing the scenario, participants were asked to look at the photo of the girl involved in the situation and freely write down five characteristics that society would attribute to her. Finally, the participants answered a questionnaire containing sociodemographic data, composed of questions related to gender and age.
Participants answered the instrument individually, administered collectively in a classroom, and were randomly assigned to one of the two experimental conditions.

Data analysis
Sociodemographic data were analyzed with SPSS software, version 22, using frequency analyses. While the discursive data were processed through textual analysis using the IRAMUTEQ. The corpus was constructed from the transcription of the texts written by the participants and divided into separate files depending on the girl's skin color (Black or White). For the analysis of the structure of social representations, Prototypical Analyses were performed, which are based on the calculation of the frequency and order of the evoked words, generating a table with four quadrants divided into Central Nucleus, First Periphery, Second Periphery, and Contrast Zone.
The structures of the social representations studied will be presented by means of tables, in which elements are organized according to frequency and average order of evocations, the frequency (F) represents the number of times a given word was mentioned, and the mean recall order (MRO) indicates the position in which the term evoked was hierarchized by the participant.

Condition 1: White girl victim of sexual violence
The words that make up the central core of the social representations of the White girl were those that had the most frequency and were promptly evoked. These words are presented in the upper left quadrant (Table 1). Here the word most often evoked in the first place was "beautiful", followed by the evocations "blonde" and "cheerful". These elements of the central nucleus were considered by the participants as the most significant in representing the White victim of sexual violence, giving real meaning to the representations. According to Abric (1994), these elements can be considered as consensual, socially shared and stable, and are the ones that most resist changes.
In the second quadrant is the first periphery, located in the upper right corner. Here are the words mentioned with high frequency and average order of evocation, that is, they were mentioned a lot, but were not promptly evoked. In this quadrant the evocations "friendly", "drunk", and "donkey" appear. The "friendly" evocation, although not in the central nucleus, seems to reinforce the idea in it. The words "drunk" and "donkey" refer to the negative characteristics attributed to the girl.
The third quadrant, called the contrast zone and located on the lower-left, is composed of elements with low evocation frequency, but which were readily mentioned.
Evocations such as "young", "easy", "smart", "smiling", "careless", and "blameful" appear here. While in the fourth quadrant, called the second periphery, are the words that had a low frequency of evocation and were mentioned later (in a second or third moment). The evocations that mostly stand out are "innocent", "rich", "bitch", "cool", and "vulgar".
Together, the evocations expressed in the condition of the White woman victim of sexual violence, present in the central core, express positive shared characteristics related to the girl. The peripheral and contrast-zone elements present a certain ambivalence, since they either coincide with the central core, reinforcing positive characteristics, or they denote negative attributes. This ambivalence can be a product of the main function of the peripheral elements, which is to protect the central core.
Thus, at the same time that it points to common sense and "blames" the victim of sexual violence, since she is White, this victim has her blame mitigated.

Condition 2: Black girl victim of sexual violence
On analyzing the characteristics that compose the central core of the social representations evoked for the Black girl victim of sexual violence (Table 2), the most evoked word was "black", followed by "pretty", "drunk", "poor", and "irresponsible". In the second quadrant, or first periphery, the words "easy" and "blameful" converge. It is seen that, compared to the condition of the White girl, here the central core was composed of more negative than positive evocations, tending to discredit the victim, and the words of the first periphery reinforce these social representations. In the quadrant referring to the contrast zone, evocations such as "dark-skinned", "young", and "vulnerable" appear. In relation to the fourth quadrant, several contradictory evocations appear, some good characteristics such as "nice/friendly" and "curlyhaired", and some negative characteristics such as "irresponsible" and "slut". The second axis, on the other hand, presents a divergence in relation to blaming the victim. The classes' statements expressed that both the victim and the aggressor were guilty, also using the victim's alcohol consumption as a justification, and that, in most cases, men take advantage of situations like these. In the opposite direction, some classes of this axis presented statements in defense of the victim, emphasizing that in no situation of violence is the victim guilty.
In view of these results, it is clear that most of the representational content encompasses blaming the victim, corroborating what has been found in the literature (Acosta, Gomes, Oliveira, Marques, & Fonseca, 2018). These results are also in line with studies that indicate that there is a shared representation that women attending parties are seen as more accessible and more sexually available (Gunby, Carline, Bellis, & Beynon, 2012). Given that social representations are formed by sets of knowledge that are socially valued and shared (Jodelet, 2001), it is considered that the way students represent sexual violence against women may be related to the meaning of positioning themselves against this practice.
The results of study 2 indicated that in the condition in which the victim was White, more positive evocations appeared, such as "beautiful" and "blonde". In the condition that the victim was Black, more negative evocations appeared, such as "irresponsible" and "guilty". These results corroborate those found by Batista, Leite, Torres and Camino (2014), whose participants stated that society attributes a greater number of positive adjectives to Whites, as happy and friendly, than to Blacks. They are assigned more negative adjectives, such as stupid and dishonest.
It is relevant to highlight that the "blameful" evocation is present in both conditions, however, in the condition in which the girl was White, it appeared in the contrast zone, demonstrating that it had a low frequency and a low evocation by the participants. In the condition that the victim was Black, that word appeared in the first periphery, which had a high frequency and average number of evocations.
It is thus relevant to note that the "blameful" evocation is present in both conditions, however, when the girl was White, it appeared in the contrast zone, thus demonstrating that it had a low frequency and a low evocation by the participants. While in the condition in which the girl was Black, this word appeared in the first periphery, indicating that it had a high frequency and average amount of evocation. This result shows that in the condition of the Black victim there was a more explicit blaming.
Together, the results of these studies help understand how social representations that are shared by society about women victims of sexual violence are organized. We can consider that social representations differ according to the victim's skin color, since, according to the central nucleus theory, two representations might be judged as distinct if their central nuclei are composed of different evocations (Sá, 1996).

Final considerations
From these results, it is concluded that in social representations about sexual violence against women, the idea that the woman victim of sexual violence is perceived as coresponsible for the violence suffered is perpetuated. It is also understood that the fact that women ingest alcohol might serve as a basis to justify that the violence suffered is deserved and, consequently, their being blamed. It is also concluded that the representations of the victims vary depending on their skin color, observing a representation with a negative connotation for the Black victim. Thus, it is clear that the analysis of these phenomena is not only based on gender, but also on skin color and the situation the victim is in, showing that the treatment received by women depends on many variables, ranging from the color of their skin, their social class, and the situation in which the violence occurred.
In view of this, we highlight the relevance of knowing and analyzing how the social representations of future professionals who will deal directly with these victims are structured, since social representations are manifestations of how social thinking of social groups is organized (Abric, 1994). Thus, these results can provide theoretical support that assists in the construction and development of actions to confront and prevent sexual violence against women. This also highlights the need to train these future professionals based on more effective qualification, including discussions on this topic, as well as a better structure of care services for these victims, providing integrated and quality care. We point out as a limitation the fact that the sample includes the participation of university students with an average age of 20 years and is restricted to a single socio-cultural context. Possibly, the results observed would vary if a sample of the general population were used, with a higher average age and various levels of education. Thus, the importance of expanding the sample is pointed out as a future direction, with a view toward carrying out intercultural analyses.

Editor/a de sección
La editora de sección de este artículo fue Pilar Bacci.