Introduction

Dating violence (DV) is a social and public health problem affecting adolescents in several parts on the world, with negative consequences on their education, health and social and interpersonal lives (Taquette and Maia Monteiro 2019). The importance of this study lies in the fact that adolescence is a stage of life in which the construction of identity evolves rapidly and develo** gender roles may be reproduced in adult relationships. Indeed, DV is considered a precursor to intimate partner violence in adults (Exner-Cortens et al. 2017). In our study, we rely on the following definitions of "adolescence" and "DV". Adolescence refers to a stage in life which ranges from 10 to 19 years old according to the World Health Organization ([WHO] 2021). DV includes many specific forms of violence, but psychological violence is the main form of DV reported by adolescents, followed by physical and sexual violence (Taylor et al. 2017), as well as cyber violence and control (Wright 2020).

A considerable number of studies conducted with adolescents have revealed that attitudes towards DV and traditional gender stereotypes are important precursors to this social problem (McNaughton et al. 2016; Ontiveros et al. 2020). However, a limited number of studies have explored the sociocultural norms underlying the attitudes and beliefs that encourage violence in romantic relationships. Furthermore, in most of the studies on attitudes conducted with Latin American adolescents, the participants were not living in their country of origin (DuPont-Reyes et al. 2015; Ontiveros et al. 2020). This lack of knowledge about what is or is not DV from the perspective of the Latin American adolescents living in southern countries is a considerable shortcoming, given that this population is exposed to high levels of violence and numerous social and gender inequalities that may influence their experiences and perceptions of DV (Fries et al. 2013). Indeed, Chile is a develo** country that, since the 1990s, has made significant progress in several fields (e.g., economic, educational, social). And while Chile remains in first place on the Human Development Index (HDI) for Latin America and the Caribbean, and is ranked 40 out of 189 countries overall, it does not seem to be advancing in gender equality (United Nations Development Program [UNDP], 2020).

In Chile, physical and psychological DV affects about 10% of adolescents (n = 2.781) (Instituto Nacional de Juventud [INJUV] 2012). Despite this sizable rate, there have been limited scientific studies on the subject, with research tending to focus on the violence experienced by adults (Bacigalupe 2000) or young adults (Fries et al. 2013). Therefore, we still know very little about what exactly Chilean adolescents consider to be violence and about the contextual elements that influence their perceptions.

The increase in Chilean adolescents in such at-risk behaviors as schoolyard violence, street gang membership, the consumption and trafficking of drugs and alcohol (Tijmes 2012), low self-esteem, high levels of sadness, and isolation at home (Florenzano et al. 2009), all make these adolescents a particularly vulnerable population group. Despite this complex portrait, Chile has no public policy aimed at preventing DV or hel** adolescents struggling with violence (Servicio Nacional de la Mujer [SERNAM] 2012; Ministerio de Salud [MINSAL] 2013).

The current law regarding intrafamily violence promulgated on September 22, 2005, states in article number 5:

Intrafamily violence. Any mistreatment that affects the life or physical or mental integrity of the person who has or has had the status of spouse of the offender or a relationship of coexistence with him is constitutive of intrafamily violence; or a relative of the offender, or of the offender’s spouse our current partner, by consanguinity or by affinity in an entirely straight line or in a collateral line up to and including the third degree. There will also be intrafamily violence when the conduct referred to in the preceding paragraph occurs between the parents of a common child, or affects a minor, an elderly or disabled person who is under the care of or depends on any of the members of the family group. (Law 20.066. Ministerio de Justicia (2005), our translation).

This practically excludes adolescents, the majority of whom are not married, are not living together, and do not have children (INJUV 2012).

As indicated, a major gap in the literature of adolescent DV is the lack of documentation for adolescents living in southern countries. Accordingly, the relevance of this study is that it is contributing to the diversification of this field of research. Indeed, the results of our study can help to balance the narrative distinctions in DV studies from the northern and southern hemisphere.

This paper attempts to answer the question: What is DV for Chilean adolescents? It presents the results of a qualitative study that used social representations (SRs) as its theoretical framework (Wagner 2020) and a multi-methodological approach. In addition to expanding our knowledge of DV from the adolescents’ point of view, it also aims to facilitate the development of an appropriate socio-political response to this social and public health problem. Our results thus help to formulate recommendations for DV prevention and provide guidance for future research.

Influence of gender and social-cultural norms on DV representations in Latin American

The results of studies on gender roles and DV are inconclusive. On the one hand, the study conducted by Fries et al. (2013) shows that girls use physical violence in romantic relationships as much as boys or even more. On the other hand, however, when sexual violence and severe physical violence are measured, women and girls are the primary victims (Lehrer et al. 2010). The literature highlights the gender differences in DV, boys and girls having different perceptions of these issues (Lehrner and Allen 2018; Taylor et al. 2017). Indeed, DV is socially unacceptable when perpetrated by boys and is perceived as more severe. On the contrary, certain contexts make violence by girls more socially acceptable (e.g., a legitimate response to potentially sexualized disrespect) (Lehrner and Allen 2018) and portray it as less harmful. Finally, emotions associated with DV are also seen to be modulated by gender. Fear is reported to be the most common feeling provoked by physical or sexual violence, specifically by girls (Ross 2012). It should be noted however that gender norms reinforce boys’ tendency to under-report the fear experienced in the context of DV (Hamby and Turner 2013), which might explain the difference.

Furthermore, a relationship between gender, socio-economic conditions, and DV was revealed in Spriggs et al. (2009), who concluded that victimization varied according to gender when the relationship between the family’s socio-economic status and school is considered. The study shows that the influence of an underprivileged environment on DV differs according to gender: boys felt more vulnerable exercising their traditional masculinity than did girls exercising traditional femininity. That being said, studies on the relationship between DV, socio-economic status, and educational, family, and environmental contexts are limited (e.g. Foshee et al. 2015).

To our knowledge, no study has looked at the SRs of DV in adolescents. The literature focuses mainly on young adults (from 18 to 30 years old). Three studies are listed: first, a study conducted by Saldivia (2011) with Chilean youth showed, for example, that hits were defined as the essential attribute of DV in its physical form. Content analysis revealed four other semantic groups associated with DV: “behavior” (hit, kill, aggress, etc.); “feelings” (insecurity, mistrust, fear, etc.); “characteristics” (frequent, mean, physical, etc.) and “consequences” (damage, death, etc.). In addition, jealousy and infidelity were the main explanations given for the violence by participants.

Second, gender differences in SRs of DV were revealed in a study conducted by Campbell et al. (1992). The results showed that the image of the violence used by women was associated with “expressive aggression” and was seen to be the result of a loss of self-control in response to a stressful situation. On the other hand, violence used by men was associated with being “instrumental” and was seen to either control or hurt the other person, or in response to a threat to their self-esteem or public image. Men justified aggression more than women did, seeing it as an expression of their masculinity. Women experienced emotions such as sadness, panic, fear, and loss of balance. The emotions most attributed to men were anger, indignation, and fury.

Third, a study conducted by Lelaurian et al. (2018) showed that young adults’ discourse on DV is structured by normative injunctions, such as: “the inadmissibility of violence, specifically when directed towards a woman” or, “the unacceptable nature of flirting when you are in relationship.” (Lelaurin et al. 2018, p. 641). This study showed a link between attitudes about violence and gender; three factors were identified. First, the endorsement of ambivalent sexism, comprising two types: (1) “hostile sexism” reflecting openly negative attitudes and stereotypes towards a sex, and (2) “benevolent sexism,” representing ideas that may seem subjectively positive with respect to gender stereotypes, but which are in fact detrimental to people and gender equality more broadly. Second, intimate partner violence myths as mediators of the effect of gender on perceived severity and justification; and thirdly, the importance of context, where norms and values shaped by society (e.g., male domination, patriarchal society, education) give meaning to violent behavior.

Though cultural factors influence the emergence of DV (Viejo et al. 2018), only a limited number of studies have examined their role (McNaughton et al. 2016). The few studies that have done so in a Latin American population suggest that tolerance towards violence can be explained by machismo, which comprises gender stereotypes and beliefs supporting masculine violence towards women (Haglund et al. 2019; Terrazas-Carrillo and Sabina 2019). In addition, machismo has been described in the literature as alternativity “a largely negative construct,” “a neutral concept,” and “a source of pride and honor” (Arciniega et al., 2008, p. 19). Indeed, some traditional aspects of Latin American culture can be considered as factors that protect against DV. This would include the importance of family, caballerismo (i.e., it is men’s role to protect women), and marianismo, which refers to the symbolic influence of the religious figure of Mary on feminine identity, representing an ideal based on values of purity, sacrifice, maternity, responsibility, and care for others (Terrazas-Carrillo and Sabina 2019). Thus, endorsement of DV attitudes is likely to decline in men when they embrace the ideology of marianismo. Arciniega et al. (2008), conclude that traditional constructs—machismo and marianismo—are both risk and protective factors in DV. In fact, they describe both positive and negative aspects of traditional Latin femininity and masculinity based on various dimensions, be they symbolic, economic, educational, cultural, or gender.

Theoretical framework

Gender is a social construction opposing femininity and masculinity that is reinforced by sexual stereotypes (e.g., women/weakness vs. man/strength). Indeed “Adolescents’ beliefs about gender roles and their expectations of how [girls and boys] behave and interact may influence their own behavior in dating relationships and their engagement in [DV].” (Haglund et al., 2019, p. 1040). In fact, SRs “of gender and gender schemas are foundational for discrimination against women” (Coron, 2019, p. 828). Therefore, SRs are a particularly relevant framework for studying DV. It was first elaborated by Moscovici in the 1960s (cited in Wagner 2020). SRs are common sense knowledge, whose specificity lies in the social character of the processes that produce them. They are conceived as a form of knowledge socially elaborated and shared, which interprets reality and guides our practices and individual choices and positions (Wagner 2020). This theory allows for a better understanding of why the same word or phenomenon does not have the same meaning for everyone. SRs can be both individual and collective, being determined by a person’s history, emotions, and the sociological and ideological system in which that person develops (Coron 2019).

Different approaches have been developed to analyze SRs. The present study draws its inspiration from work conducted by Abric (Rateau et al. 2012; Wagner 2020), who developed a structural approach dealing specifically with the content of SRs by focusing on their common elements. The analysis conducted through this approach aims, on the one hand, to identify the central core of an SR, its fixed base; and, on the other hand, to shed light on its peripheral elements, those that are less consensual, more flexible, heterogeneous, determined by the changes in SRs as well as by the people’s characteristics and daily contexts. We were, therefore, interested in drawing out and shedding light on the most common elements composing the SRs of DV in Chilean adolescents so as to identify the group’s more stable conceptual foundations. We also studied the distribution of these representations based on the different social subgroups (gender and social class) to which they belonged to deepen our understanding of the influence of social context and individual characteristics on these adolescents' representations. This theory is quite useful in understanding the transformations in people's conceptions of a social object, in this case DV (see Fig. 1).

Fig. 1
figure 1

Schematic concepts of the study

Full size image

Methodology

The qualitative research design used here was an exploratory-descriptive strategy (Denzin and Lincoln 2005) aimed at understanding a single topic, namely adolescents’ SRs of DV. It should be noted that exploratory is not used here to imply that absolutely no research has been conducted on this issue. However, the studies addressing adolescents from southern countries are limited. Given that it is recommended to use a multimethodological approach in SR studies (Rateau et al. 2012), this study was carried out with two data collection methods, namely free association and focus groups, which facilitated the triangulation of each data set. Based on the research question: What is DV for Chilean adolescents? The following goals were developed: (1) Describe the contents of SRs of DV among Chilean adolescents, distinguishing between central and peripheral elements; and (2) Identify, where applicable, the differences and similarities in the SRs of DV among Chilean adolescents, according to gender and socio-economic status.

Recruitment strategy and participants

We used a non-random sample composed of volunteer participants. Eleven schools were contacted; of these, four public schools and three private schools agreed to participate. Supported by administrators from these seven schools, meetings of approximately 20 min were organized with adolescents from 42 classrooms. According to the schools’ records, an average of 300 adolescents (aged 14 to 18; grades 9 and 10) from private schools and 720 adolescents from public schools attended these meetings.

To ensure informed consent, the researcher explained in simple terms to these adolescents why the research was being conducted, its objectives, methodology and ethical guidelines, what would be done with the results, and the benefits of their participation, i.e., the advancement of knowledge about DV. They were invited to participate in the two data collection phases, one based on free association and the other on focus groups. Their questions about the research were also answered at this time. Finally, a consent form was given to them so that they could discuss the research goals with their parents and return the signed forms within approximately two weeks. Among the 1,020 adolescents present at these meetings, 142 volunteered to participate in the free association phase (68 from public schools and 74 from private schools) and of these, 48 also volunteered to participate in the focus groups (25 from public schools and 23 from private schools). The socio-demographic characteristics of our sample are shown in Table 1.

Table 1 Participants’ socio-demographic characteristics
Full size table

Even though most of the respondents described themselves as being middle class (87.4%), the socio-economic differences were more evident in their perception of their neighborhood. Indeed, 14 participants from public schools stated that they lived in a “dangerous” or “very dangerous” neighborhood, whereas no participants from private schools did. This is in line with the high degree of social segregation in Chilean schools (Torche 2005). The more privileged participants go to private schools whereas the less privileged go to public schools. Since the perception of the social class and socio-economic status differed according to the participants, we used the notion of the type of school to indicate differences noted in the focus groups.

Data collection methods

Data collection was conducted between February and May 2014 in three urban cities in the Province of Concepción in Chile: San Pedro de la Paz, Concepción and Hualpén (see map, Fig. 2), following the approval of the concerned research ethics committee. The schools that agreed to participate are located in these cities.

Fig. 2
figure 2

Province of Concepción map. Source: Authors illustration. 2020. Population of three cities in the Province of Concepción. CENSO 2017 Statistics National Institute, Chile

Full size image

During the data collection of the free association method (see Online Appendix 1), the participants had to write five words that came to mind after reading the expression “dating violence.” The answers could be written in the form of verbs, nouns, adjectives, adverbs, etc. This method gave us access to the semantic dimension of the object being studied (Dany et al. 2015). This method along with a brief questionnaire about the participants’ socio-demographic characteristics was completed in a 30-min-long session during normal class time.

The focus groups were held outside of class time. So that the participants would be more at ease discussing the subject, single-sex groups were chosen. According to Morgan (1997), the number of groups required to reach empirical saturation ranges from 3 to 5, with 6 to 8 participants per group. In the present study, 12 groups (same number of groups of girls and boys, as well as of public and private schools) were composed with a mean of 4 participants per group. The group meetings lasted from 90 to 120 min and were led by the head researcher who was accompanied by a male or female research assistant depending on each group’s make-up. Based on the literature review and the results of the free association method, an interview guide was developed to examine the following themes: the components of a healthy romantic relationship; the attitudes and conduct considered to constitute violent behavior; and the similarities and differences regarding DV according to gender, stage in life (adolescents vs. adults), and social class.

The aim of the research was not to investigate and examine the experiences of the participants. Some participants in the focus groups did however give examples or testimonies of personal DV experiences or experiences of people close to them. As planned, each participant was given information about resources that they could access at health centers if needed.

Data analysis

The analysis of the data obtained through free association took place in two steps based on the work of Dany et al. (2015). First, semantic groups and dimensions were created. These dimensions were then assembled into higher level meaning categories. During the first phase, the 706 words (terms or expressions) that were collected were translated from Spanish into French to be codified with the NVivo 10 software. A manual process of lemmatization helped us to reduce, under the same term, words that were written in different forms, such as the singular, plural, participle, adjective, etc. This process allowed us to reduce inflectional forms and, sometimes, forms derived from a word to a common base form that, unlike stemming, considers the context and converts the word to its meaningful base form, referred to as a lemma (Balakrishnan and Lloyd-Yemoh 2014). For example, words were grouped semantically according to their morphological (e.g., suffer, suffering) or synonymic base (e.g., relationship breakup, separation of the couple). This made it possible to calculate the frequency with which each semantic dimension was mentioned. Forty semantic groups were ultimately identified.

Following a thematic approach in the second phase, the semantic groups were placed in 15 higher-level categories. Categorizing the semantic groups allowed us, based on the thematic analysis of the focus group content, to distinguish central from peripheral elements in the SRs of DV and to identify where the study participants’ content agreed or disagreed. It is worth noting that some semantic groups could be placed in more than one category. For example, the semantic group “fear” can represent both an emotion and the consequence of endured violence depending on the context.

The semantic dimensions were categorized using inter-judge agreement to maximize the objectivity of the data analysis process. Indeed, during the content analysis of the first and second methods, several debriefing meetings between members of the thesis committee (composed by the first and third authors and two professor-researchers) were held to maximize the credibility of the analysis. This is a strategy commonly suggested to enhance the trustworthiness of a study by reducing the risks related to researcher bias (Patton 2002). Overall, the committee members shared similar viewpoints and enriched the analysis with suggestions here and there.

The focus group data were analyzed using thematic analysis (Denzin and Lincoln 2005). This analysis strategy allows researchers to organize emerging themes from unstructured material and interpret them in the context in which they were constructed. The analysis was conducted in four steps. The collected data were first recorded and then were transcribed verbatim in the original language of the collected data (Spanish) to preserve as much of the interviewees’ idiomatic expressions as possible and to ensure research credibility. Preliminary readings were then carried out several times so that we could familiarize ourselves with the material and establish a list of statements that would serve as a foundation for the coding. In the third step, NVivo 10 software was used to categorize and classify the content into successive groups of statements based on their semantic proximity meanings. In the fourth and final step, the data interpretation documented the adolescents’ most common SRs of DV as well as those which were more influenced by context and some of the adolescents’ characteristics.

In accordance with the criteria proposed by Lincoln and Guba (1985) for qualitative research, namely dependability, credibility, transferability, and confirmability, several procedures were put in place to ensure the trustworthiness of this study. Dependability corresponds to the results’ degree of stability over time, thereby showing that the findings are consistent and could be repeated. The data coding process and the detailed explanation of the results analysis method made it possible to understand the process by allowing other researchers to arrive at substantially the same results when performing the study again. The large sample size for a qualitative study, the detailed description of the characteristics of the respondents and the recruitment methods, and the use of a logbook allows for the transferability of the results to populations with similar characteristics. The multi-method approach, the triangulation of the data, the inter-rater verification of the coding, and the fact of having interviewed students from different types of schools all helped to increase the credibility of the research. The use of a well-defined theoretical framework made it possible to better understand what social representations are and to present the results, which were always supported by quotes from the participants met in the recorded discussion groups. Likewise, the researchers’ use of reflexivity to avoid their biases and their interests helped to ensure the study’s confirmability.

Results

In this article, the results of the two data collection methods are triangulated. The data from the free association method are first presented in Table 2 and then in Fig. 3 in the form of both semantic groups and higher-level categories. Each of the categories is then illustrated by comments from focus groups, which allows us to place these answers in the discursive context of the participants. It is worth noting that the participants’ names were replaced by fictitious names to ensure their anonymity.\

Table 2 Semantic groups of words associated with DV
Full size table
Fig. 3
figure 3

Semantic categories

Full size image

Semantic groups

A total of 706 words or expressions were associated with the term “dating violence.” Table 2 presents the semantic groups that were most frequently named by the participants. The first column indicates the identified semantic dimension, the second presents the words mentioned in each group, and the third details the number of times these words were employed. This table shows that the figurative image most evoked by the participants was hits. The physical nature of DV was thus deeply rooted in the studied SRs.

Semantic higher-level categories

The semantic categories were grouped according to their semantic proximity. Given the limited space of this article, we only present, in Fig. 3 and in subsequent results, 6 of the 15 higher-level categories that were identified. These were retained because the first five categories were the most frequently mentioned by the participants, and the sixth one occupied a larger place in the focus groups. The categories are presented in decreasing order in Fig. 3 and the differences according to the diversification criteria (gender and school type) are highlighted. Extracts from the focus groups are then presented to further develop our analysis of the discursive and social context in which SRs of DV are constructed.

As is shown in Fig. 3, we note that the “emotions and feelings” category was the most common in the adolescents’ SRs in the free association method. Girls were more likely than boys to associate DV with emotions such as fear and pain. This gendered difference also emerged from the thematic content analysis of the focus groups where the girls mentioned having felt fear in their couple relationship whereas the boys did not.

I would see changes [in behavior] and then he would start to get angry, and I would start to feel afraid that he was going to hit me or bully me … You shouldn’t be bullied by your boyfriend (Margarita, 16 years old, public school).

As regards the second semantic category, “source of conflict,” jealousy was the most frequently mentioned source of conflict and reason for using violence. The results of the free association method (Fig. 3) were coherent with the statements of the adolescents in the focus group as shown by Valeria, “Why do arguments happen? It’s because of jealousy…” (16 years old, public school). The analysis associating gender and type of school pointed to some notable differences. Boys from public schools were more likely to associate jealousy and arguments with DV than were the other three groups. The focus group participants agreed that boys were more jealous than girls were, as can be seen in the following extract:

[...] bitch, several times “maraca” « [Chilean expression to designate prostitute], words that [a man] generally [uses] when a man [...] I don’t know if I can generalize, but I have been in situations where men say that to women when they are jealous, angry (Rodrigo, 16 years old, private school).

In addition, the vast majority consider that "social media" generate a great deal of jealousy in couples and ultimately conflict. This is shown by Adolfo, “Facebook causes a lot of problems. For example, a guy writes a comment like, “Oh! You’re so beautiful!” It's like a comment, a little flirt, and then you get jealous” (16 years old, public school).

The third and fourth semantic categories involved “acts of physical violence” and “acts of psychological and verbal violence” respectively. As seen in Fig. 3, the girls referred to words associated with these types of violence more often than the boys did. However, an analysis of the adolescents’ discourse in the focus groups pointed to some interesting nuances. For example, the adolescents considered that the meaning or impact of hits depended on whether they were given by a boy or girl. Most of the participants considered that the physical violence used by boys was more serious, because of their greater strength, as shown by Joaquin, “We don’t have the same strength… A girl’s body is more delicate. If you hit a man in the chest, it doesn’t hurt him. A woman, yes. I mean, yeah, it might hurt a man, but for a woman, it hurts a lot more” (18 years old, public school). For some participants, this strength would place them in a position of superiority over girls as illustrated by Gianella, “Men think they are better than women… The man believes he's better because he’s stronger [than women]” (16 years old, public school).

Another example from the focus groups, is that a girl’s violent behavior seemed to be associated with the use of verbal violence, whereas that of the boys was associated with physical violence. Participants explained the differences by gender stereotypes: women are more delicate and have better verbal skills than men do, whereas men are physically stronger. This idea is shown in Nina’s words: “Women have more verbal skills, more vocabulary. I realized that men, on the other hand, they use punches more often” (15 years old, private school).

The importance given to physical strength as a central element in defining violence, the association of violence with the male stereotype and the low impact of physical aggression of girls would explain, according to some participants, the possible social tolerance of violence by girls. “It’s the result of the society we’re in. Even [in] films, soap operas, you see women slap. So, it's like something social… It seems normal, but it's also violence because it's a hit” (Valeria, 16 years old, public school).

The results of the present study thus shed light on a certain paradox: while the participants noted that both genders can exercise violence, they also noted a difference in the way they used it, as shown by Maria Carolina: “I think there is less difference… because, in both cases, they [men and women] treat each other poorly, […] hit, […] say bad things to each other” (15 years old, public school).

The fifth semantic category considers the “consequences” of DV. These refer to the physical, psychological, and relationship aspects. For example, the semantic group injuries refer to both physical and psychological scars. The semantic groups loveless and break-up, in contrast, refer to consequences at the relationship level. In general, consequences are mentioned more by girls than boys (Fig. 3). Some participants in the focus groups attributed more importance to psychological consequences than physical, as can be seen in the following extract:

Psychological violence [hurts more] than physical violence because physical violence can leave a bruise or something like that. But psychological violence, keeps coming back to you, it doesn’t go away. […] Sometimes a word hurts more than a hit (Ivan, 15 years old, private school).

Despite the fact that, in the free association method, the words associated with the “social phenomena” category were only mentioned by a few participants (Fig. 3), they nonetheless occupied a larger place in the focus groups. Machismo, which promotes the dominant image of an aggressive man, was given as an explanation for DV. Moreover, the adolescents we met explained the girls’ violence by changes in Chilean society, which now attaches greater importance to the rights of women, who in turn are becoming freer to stand up to men. Participants also talked about changes in mentality and the influence of religion as illustrated by the following excerpts:

… it used to be […] that women let themselves be dominated because it was such a macho society, because women were raised to always let themselves be dominated by men. Now women demand respect... and this leads to a lot of violence too... Often, on the woman's side ... (Flavia, 16 years old, private school).

People's mentalities have changed a lot ... before, women were always degraded [and] men were always the important [ones] because they are in the image of God according to the Bible… many religions do not take Mary as [an important figure] [some people] use religion as a rationale for abuse… (Camelia, 15 years old, private school).

The focus groups also helped to situate the SRs in a complex social context. The participants talked about the socio-historical changes occurring in gender relationships in Chile and the presence of traditional practices related to male/female hierarchy. These changes are perceived differently by girls and boys. While the former see them as an opportunity to overcome machismo, the latter react to them in a contradictory manner, displaying perplexity and uneasiness.

I think [violence against women] will be reduced because [now] there is an assumption that [men and women] are more equal, [and women] are no longer inferior... (Leonor, 16 years old, public school).

Before... the man was... the leader of the relationship. Now, I think the roles have completely changed and the woman is [the boss]. Now, instead, the man is the one who has to follow the woman's lead, he’s the lap-dog now (Bastian, 16 years old, private school).

Discussion

Drawing inspiration from Abric's structural approach to SRs (Wagner 2020), the analysis of semantic dimensions led us to observe that the semantic group hits was the most often mentioned as regards DV perception, as likewise noted by Saldivia (2011). Consequently, the physical dimension of violence would seem to be the most profoundly rooted in the participants’ SRs. That being said, when the semantic groups were organized into higher-level categories and triangulated with the results of the focus groups, the importance given to gender stereotypes could also be seen in the participants’ answers (see Fig. 4).

Fig. 4
figure 4

Central core and peripheral elements of DV

Full size image

The focus group showed that girls’ behavior was associated with verbal abuse, which girls practiced more than boys, the latter being more associated with physical abuse. The participants explained these differences by gender stereotypes: women are more delicate and have more verbal skills, men are physically stronger. In the free association method, the girls referred to "physical acts" more often than boys did, but the analysis of the data collected through the focus group method revealed new avenues for interpreting gender differences. First, the interpretation of the term hits differed depending on whether the act was committed by a boy or girl, given that violent female acts were trivialized by participants and seen as less dangerous because of their lesser physical strength. The society would seem to be more tolerant of violence by girls towards boys. Our results are consistent with other studies asserting that gender is central to understanding the perception of DV (Hamby and Jackson 2010; Lehrner and Allen 2018). Although girls generally qualify violence as being unacceptable, in some cases they justify and approve of it when it is used as a means of self-defense or revenge (Haglund et al. 2019). Second, some participants in the focus groups attributed more importance to psychological consequences than physical. In relation to this, the most surprising finding is that the adolescents identified emotional abuse and psychological aggression as violence even though they probably have not been educated about these forms of violence, at least not through government prevention programs.

Gender is also present when we analyze the emotions that adolescents associate with DV. The two methods allowed us to identify fear as a determinant in the SRs of DV. This study showed especially that, for the girls, the free association method revealed the latent dimensions that structured the semantic dimension due to its projective character (Dany et al. 2015). The position of victim was thus associated with DV. In groups however, the girls adopted a discourse where they were powerful, a discourse probably reinforced by group dynamics (nobody wants to be considered a victim). In our opinion, this is not a contradiction in the girls' discourses, since they were obtained by two different methods. On the contrary, the combination of these methods would seem to give rise to complementary results that reveal the coexistence of two social influences. On the one hand, the girls feel vulnerable because they belong to the female gender which they consider to be more exposed and fragile; as such, they show a continuity with the position of adult women in abusive relationships. On the other hand, they want to project an image of modern, independent women who will not accept to be abused. This symbolizes a generational gap and embodies the dynamic character of gender as it is influenced by changes in Chilean society.

These results are consistent with those of other studies conducted with adolescents. For example, Ross (2012) noted that girls experience more fear and sadness than boys do and that the former associate this emotion with the feeling that they have less control and power than the latter. It should also be noted that, while male violence is associated with greater physical strength (Lehrner and Allen 2018), social power, and a gender socialization that reinforces this behavior, gender stereotypes likewise increase the tendency of boys to underreport the fear they can feel in a DV context (Hamby and Turner 2013). Campbell et al.’s (1992) model, which distinguishes between “expressive aggression” (resulting from a loss of self-control in women) and “instrumental aggression” (used by men to control their partner), cannot be easily applied to the participants’ SRs examined in the present study. As regards emotions, girls readily mentioned anger and hate to justify violence, whereas boys mentioned fear and pain, and did so in smaller numbers. Given that this topic has not been widely studied in adolescents, future research should take a more in-depth look at gender stereotypes, emotions, and DV.

The discourses of the participants were characterized by ambivalence about condemning, minimizing, and justifying violence. The principle of SR conditionality (Flament 1994, cited in Lelaurian et al. 2018) furthers our understanding of a context described as "ambivalent sexism" (Ramiro-Sánchez et al. 2018). The importance of taking socio-cultural and gender aspects into account so as to better understand what violence and violent practices mean for adolescents is emphasized by several authors (Lehrner and Allen 2018; Taylor et al. 2017). Adolescents are develo** their identity and learning about romantic relationships, and the social context influences these processes. That is why it is essential to develop preventive strategies intended for the context in which adolescents and their romantic relationships develop.

Examining how SRs of DV are constructed in a given socio-cultural context would also seem to be quite relevant if we consider, like Jodelet (2012), that culture can be used as a symbolic and material framework for the production of behaviors and as an element for structuring relationships in society. During the focus group meetings, the “machismo culture” was identified by the participants as the main factor explaining male violence against women. The category of machismo has often been used in different studies of Latin American populations to understand male violence (Haglund et al. 2019; Terrazas-Carrillo and Sabina 2019). According to the participants in this study, the machismo ideology was quite present in earlier decades (i.e., the generations of their parents and grandparents). That said, the cultural changes characterizing modern Chile have had a mitigating effect on male dominance. The participants considered that machismo still exists to a lesser extent and that it reinforces the image of men as the main aggressors. Contrary to the results of the study conducted by Arciniega et al. (2008), who argue that “there are both positive and negative aspects of traditional masculinity” (p.31), the participants in the present study attributed a negative connotation to machismo. For the boys who participated in the focus groups, socio-cultural changes and the evolution of women meant a loss of power for them, explaining the need to "adjust" and change their practices. Future studies should examine which resources lead to a change in the perception by adolescents of the traditional gender hierarchy and what strategies boys use to adapt to an apparently unfavorable context.

Marianismo, moreover, is not considered to be a protective factor among participants, as other studies in Latin American populations have shown (Arciniega et al. 2008; Terrazas-Carrillo and Sabina 2019). On the contrary, this figure symbolizes—especially in the girls’ discourse—the subordination of women in a context of male dominance and is associated with older times. Indeed, our results point to possible generational differences in the way gender identities and strategies influence how the girls face violence. This result is in kee** with the idea of DuPont-Reyes et al. (2015) who underline the importance of a heterogeneous view of the Latin American population. Future studies should explore these generational changes in greater depth.

Social thought functions in such a way that it integrates new elements encountered in people’s physical and social surroundings. Cultural changes identified by the participants concerning gender relationships were thus integrated into traditional SRs. The concomitance of social thought rooted, on the one hand, in the machismo of the past and, on the other hand, in neo-machismo (Montecino 2003)—which explains the new forms of machismo in response to the loss of masculine privileges—highlights the mixed characteristics of “Latin American culture.” The opinions of the adolescents we interviewed showed the continuities and breaks existing between the society of the past and the one in which they now live, a society where traditions and modernity coexist. This cultural characteristic of develo** countries like Chile—but also of other southern countries—should be taken into account in future studies to better understand complex problems with socio-cultural roots such as DV.

This study explored the impact of socio-economic conditions on SRs of DV. Even though most of the respondents described themselves as being middle class (87.4%), the socio-economic differences in their perception of their neighborhood are consistent with the high degree of social segregation in Chilean schools (Torche 2005). The main discrepancy in the data regarding the types of schools emerged in the free association method. It concerned the answers in which the theme of jealousy was associated with DV. In our results, the boys from public schools were more likely to make this association than were the other three groups. Since a limited number of studies have looked at aspects related to social class as a factor associated with DV (e.g. Spriggs et al. 2021). To date, no Chilean studies have been published on the effects of the pandemic on intimate or domestic violence. It is, therefore, impossible to know whether adolescents' SRs of DV have changed with this pandemic. Studies in other countries specifically target adult women who live with their abuser (Agüero 2021). Therefore, knowledge about the increase in violence, the effects of confinement in the romantic relationships of adolescents who do not live together, and the impact of increased social media use in DV on violence is limited.

Despite these limitations, the present study proposes original findings that contribute to our understanding of adolescents’ SRs of DV, that is, how the SRs influence their reality and guide their practices, choices, and positions. Recognizing our limited knowledge of SRs of DV in a Latin American context, but also, in the context of North American countries, our study clearly serves as a point of comparison for future studies that aim to explore representations of both gender and DV.

Conclusions

In the present article, the Chilean adolescents’ SRs of DV were explored through SR theory using a multi-method approach comprising free association and focus groups. Several implications for prevention emerge from our analysis. First, SRs of DV are built in a context of ambivalent sexism; promoting actions for a culture of gender equality is therefore critical. In addition, this study shows that SRs of DV are dynamically co-constructed, interacting with gender identities, and that gender has an influence over the perception of DV, associated emotions, forms of violence used by girls and boys, etc. Consequently, DV prevention strategies must consider gender and cultural and socio-environmental factors that influence the behaviors of adolescents and their experiences of DV. It is not enough to change DV favorable attitudes, as most prevention programs propose. We must change the ground in which these attitudes take root, which implies large-scale awareness-raising strategies and several levels of intervention (individual, family, community, social) and multi-component programs (e.g., sexuality, drug use prevention). Because DV is a gender-based problem (Viejo et al. 2018), these DV prevention strategies should develop “gender-specific” content. It could, for example, discuss positive masculinities, communication skills, expression of emotions, and conflict resolution with boys; and with girls, talk about empowerment, the meaning of being respected, and the social tolerance of girls' violence.

Second, while school is a privileged place for prevention, a few workshops are not enough; the curriculum must include DV prevention and the promotion of healthy romantic relationships. Based on differences between participants from public and private schools and recognizing the high degree of social segregation in Chilean schools, we believe that this factor should also be taken into account when designing prevention programs in schools. Adolescents can be reached in various settings (Goodmark 2018), so prevention campaigns must be multi-site: school, sports activities, social media, religious groups, etc. In addition, peers, parents, teachers, but also health professionals, police, and other actors (coaches, religious leaders, etc.) should be the object of awareness-raising and prevention actions. Third, changing the current intrafamily violence law (20.066) is critical. The underlying traditional conception of the romantic relationship anchored in the married couple renders DV in unmarried couples invisible. In 2020, the Chilean law on femicides (21.212, art 390, promulgated on March 2, 2020 (Ministry of Justice (2020)), was modified and now includes crimes committed by an intimate partner even without cohabitation. Though progress has been made, we are still without effective prevention or punishment to combat effectively. Finally, because DV is a social and public health problem, intersectoral collaboration is essential (Goodmark 2018). Several sectors such as education, health, leisure, media, and community, as well as the ministry for women and gender equality (currently SERNAMEG, formerly SERNAM) must work in synergy to provide adolescents with relevant, suitable services that take into account what DV means for them.