Dating Victimization Among Chilean University Students: Gender Roles and Christian Spirituality.

This correlational-multivariate, cross-sectional quantitative study differentially determined by sex the prevalence of dating victimization by violence, the self-perception of victimization, and the attempt to ask for help to end a problematic relationship, as well as analyzing the explanatory weight of the attitudes to gender roles and Christian spirituality on these variables. The study sample, non-probabilistic and by convenience, was comprised of 759 Chilean university students. 63.9% were women and the average age was 20.5 years (SD=1.69). Men report suffering more physical violence and violence by coercion, and women have more tools to perceive their situation of victimization and to ask for help. Transcendent attitudes are a protective factor, stereotypes a risk factor, and the influence of religion is paradoxical. However, the explanatory power of these variables is low. In conclusion, dating violence is a problem present in the study sample, with the experience of victimization being different for men and women.

1. Introduction

Dating violence (DV), defined as “any type of intentional aggression by one partner against another” (Rubio-Garay et al., 2017, p. 135), is a public health issue and a violation of human rights (Santoro et al., 2018). Internationally, the systematic review performed by Rubio-Garay et al. (2017), based on 113 publications, shows a great variability in the numbers on the prevalence of the victimization. In Chile, the numbers vary between 41% and 85% for psychological violence, and between 16.1% and 37.9% for physical violence (Lehrer et al., 2009; Saldivia & Vizcarra, 2012; Vivanco et al., 2015; Vizcarra & Póo, 2011).

These differences among studies denote a lack of consensus among researchers in conceptualizing and measuring DV, which promotes limitations and controversies on our knowledge of and approach to the phenomenon (F. Rubio-Garay et al., 2015).

Most of the research on DV frames the violence in the physical, psychological-emotional, and sexual categories. Compared with this, the scientific community has noted deficiencies in the traditional grouping, particularly in the psychological dimension, as it pools dissimilar behaviors (Aizpitarte Gorrotxategi & Rojas-Solís, 2019; F. J. Rodríguez et al., 2017).

Some authors recommend moving beyond the classic proposal by offering alternative categories such as (1) coercion, which groups behaviors that pressure someone by force of will; (2) humiliation, which refers to criticisms leveled at the persons self-esteem and respect, as behaviors like refusal of support; (3) and detachment, which groups behaviors of indifference and rudeness towards the partner and their feelings (F. J. Rodríguez et al., 2017). It is added to this difficulty that young people perceive some of these aggressions as normal practices and relative to the context or inherent to conflict resolution. Thus, the works that take the perception about suffering from DV as a reference do not reflect the true extent of the problem. Numerous studies have contributed empirical evidence about young people who do not consider themselves abused despite experiencing violence (Cortés-Ayala et al., 2014; López-Cepero et al., 2015; Rodríguez-Franco et al., 2012).

This dissonance between reality and perception is relevant to the search for help: labeling the violent experience is the starting point to end the situation (López-Cepero et al., 2015). Vizcarra and Póo (2011) determined in their study that more than half of the victimized participants never told anyone. In the study by Lehrer et al. (2009), one third of women and 42.7% of men kept the secret. Those that speak out turn to a friend and to an institution as a last resort, but they never did to the police (Lehrer et al., 2009; Vizcarra & Póo, 2011).

The literature indicates that preconceived notions of masculinity and femininity affect the understanding and perception of DV and the search for help. Female violence is perceived as less harmful and justified. The man as victim is inconceivable due to his position of power and the expectations of his physical and emotional infallibility. When he is recognized as a victim, he is ridiculed, so his search for help is impeded in order to maintain his male identity. As a result, the gender hierarchy has been favored in studies on DV: the tendency to assign the woman the role of victim and the man as the aggressor (Aizpitarte Gorrotxategi & Rojas-Solís, 2019; Ferrer-Pérez & Bosch-Fiol, 2019; Moreno et al., 2016; Pereda & Tamarit, 2019; R. Rodríguez et al., 2018; Rojas-Solís et al., 2019; Santoro et al., 2018; Savage et al., 2016; Scarduzio et al., 2016; Walker et al., 2018; Wilchek-Aviad et al., 2019).

Archer’s metaanalysis by 2000 drives the study of gender symmetry. This approach questions acts of aggression as being linked decisively to one sex and admits the existence of bidirectional violence, without detracting from the evidence on the higher rates of injury suffered by women. Currently, studies report the reciprocity of aggression in young couples. Some studies report that men admit to suffering a greater number of abusive behaviors than women, tolerate more female violence, and ask less for help. Studies with samples of Chilean young people also report this, although there are also studies in which women are more victimized, or there are no differences between sexes (Courtain & Glowacz, 2018; Lehrer et al., 2009; López-Cepero et al., 2015; R. Rodríguez et al., 2018; Rubio-Garay et al., 2017; Saldivia & Vizcarra, 2012; Vivanco et al., 2015; Vizcarra & Póo, 2011). The results of the systematic review by Rubio-Garay et al. (2017) suggest a greater prevalence in the psychological and sexual victimization in women, although the women report victimizing their partners more psychologically.

From an ecological point of view, DV is the result of the interaction of risk factors on different levels: individual, close relations, community, and sociocultural. Based on these, explanatory models have been developed, founded on different theoretical approaches. Among them, the perspective of gender stands out, widely endorsed by the scientific community: it lends great weight to the patriarchal mechanisms that define the role of men and women in society, and defends gender as a central category of analysis in the explanation of violence (Ferrer-Pérez & Bosch-Fiol, 2019; Moreno et al., 2016; Santoro et al., 2018; Savage et al., 2016; Wilchek-Aviad et al., 2019). Other authors strongly question this conceptualization, considering it a biased approach that does not encompass the complexity of the phenomenon (Archer, 2000; Arnoso et al., 2017; Pereda & Tamarit, 2019). Based on this heated debate, our interest focuses on the attitudes of young people regarding gender roles (GR) and their explanatory weight in DV.

Stereotypical attitudes -an unequal attribute of functions by sex in the public and private spheres, the conflict between the competences assigned to these spheres, and the segregation of functions in the couple- are identified among the factors of greatest relevance to explain violence. However, this variable is accompanied by a list of precipitating, facilitating, and modulating risk factors. Authors recommend analyzing it as a moderating variable of DV, or they report that it only explains DV weakly. Others determine that the stereotypical attitudes are related not only to DV in itself, but also to other related variables (Arnoso et al., 2017; Berkel et al., 2004; Bringas-Molleda et al., 2017; García-Cueto et al., 2015; Ramiro-Sánchez et al., 2018; F. Rubio-Garay et al., 2015). Conversely, young people with egalitarian attitudes/transcendent attitudes -assessment of the behavior transcending gender- perceive abusive behaviors more easily and tolerate violence less (Baber & Tucker, 2006; Berkel et al., 2004; Bringas-Molleda et al., 2017).

GR are part of cultural baggage and conveyed through their different forms of expression, such as religion, justifying or legitimizing direct or structural violence (Galtung, 2016). Roeser et al. (2008) describe religious identity as the personal identification with a collective or group, characterized by a particular religious tradition and defined by a common worldview. In Chile, Christian spirituality is strong, because most of the population identifies with Catholic and evangelical groups. According to the traditional archetypes of femininity, these groups espouse values and moral rules on the family, impacting on the beliefs and expectations of gender and dating relationships, as well as promoting the tolerance of violence (Alcaino, 2017; Berkel et al., 2004; Betancourt & Cartes, 2019; Jankowski et al., 2018; King & Boyatzis, 2015; Nelson, 2009).

However, it has been estimated that the role of religion can be paradoxical. It also favors moral codes and expectations in favor of happiness, resilience, care of others, and respectful and decent treatment. Studies indicate that young people with less religious identification refer to suffering more types of abuse, and those who have greater intrinsic religiosity are less likely to be victims (Berkel et al., 2004; Fernández-Ríos et al., 2018; Fernández-Ríos et al., 2015; King & Boyatzis, 2015; Tussey & Tyler, 2019). For their part, Berkel et al. (2004) establish spirituality as a protective factor for the tolerance of DV, but they also concluded that intrinsic and extrinsic religiosity were not related to this variable.

Betancourt and Cartes (2019) conducted a qualitative study with young Christian Chileans. At discourse level, they identify the Christian religion and the values it promotes as a protective factor. However, they consider that one way to prevent violence is to fulfill the GR valued in the Bible. They also recognize the church as a network of support against violence, while at the same time they admit they would not ask their community for help out of shame and fear of being judged as bad Christians.

According to the information reviewed, we hypothesize that men and women experience similar levels of violence, highlighting psychological violence as the most frequent; low levels of perception of victimization in general, although with men having to suffer a greater number of violent behaviors to be selfidentified as victims; finally, men and women preference for friends and relatives instead of specialized services when seeking help, although female sex is a determining variable to decide seeking help.

Additionally, we expect that the stereotypical attitudes to GR and spirituality will correlate positively, and that in turn these variables will be defined as predictor variables of the three types of behaviors indicated, although with low explanatory weight.

Finally, the aim of this study is to differentially determine by sex, in a sample of Chilean university students, the prevalence of victimization by DV, the selfperception of victimization, and the intention to seek help to end a problematic relationship, as well as to analyze whether the attitudes to GR and Christian spirituality contribute to explaining these variables.

2. Method

2.1 Participants

This correlationalmultivariate crosssectional quantitative study had an initial nonprobabilistic sample by convenience of 1,080 Chilean university students, a sufficient size for a representative sample of the countrys student population -1,194,311 for 2019 (Ministerio de Educación, 2020). However, as it is a sample by convenience, it is not representative. Participants who stated they had been in a dating relationship for at least one month were selected: 759 students from the central (27.1%, n=206) and southern Chile (72.8%, n=553). 63.9% (n=484) were women. The average age was of 20.5 years (minimum of 18 years, maximum of 25). Most participants are studying education sciences, social sciences, and humanities (45.1%, n=342). 60% (n=367) identify with Christian religions, and 23.6% (n=179) with no religion. Analysis of a priori power, using the G-Power program to determine the minimum sample size necessary in each analysis to be performed in this study (Faul et al., 2009), showed that this number of participants was sufficient.

In addition, two groups were formed through twostage clustering based on the scores in the factors of the Dating Violence Questionnaire-R (DVQ-R) (victimization by DV): medium level of victimization (MLV) (n= 160; coercion, mean= 1.22, DT=1.33; detachment, mean= 1.44, DT=1.60; humiliation, mean= .76, DT= .95; physical, mean= .94, DT= .32; sexual, mean= .29, DT= .69); high level of victimization (HLV) (n=594; coercion, mean= 4.61, DT=2.84; detachment, mean= 5.64, DT=3.44; humiliation, mean= 3.41, DT=2.80; physical, mean= 1.06, DT=1.69; sexual, mean= 2.17, DT=2.64).

2.2 Instruments

Sociodemographic questionnaire. Using a brief questionnaire created ad hoc, personal information was collected, including sex, age, origin, religion, and so forth.

Dating Violence Questionnaire-Revised (DVQ-R). This instrument (F. J. Rodríguez et al., 2017) assesses victimization in dating relationships in both men and women. It has 20 items on a Likert-type scale from 0 (Never) to 4 (Almost always), grouped into 5 factors: coercion, detachment, humiliation, physical and sexual violence. This instrument has been shown to be reliable for a Chilean university population (Pérez & Rodríguez-Díaz, 2017). The scale has an adequate internal consistency (McDonalds coefficient omega for ordinal data and multidimensional instruments): coercion, .74; detachment, .81; humiliation, .83; physical violence, .90; sexual violence, .90; overall, .92.

Items Perception of abuse and asking for help. The DVQ-R includes a section with questions with dichotomous reply options (No=0 and Yes=1) that assess the perception of abuse, as “Have you felt abused?”, and on seeking help, such as “If you needed to ask for help to break up with your partner, (1) would you ask friends?”, (2) “would you ask relatives?”, (3) “would you ask professors, advisers at your school?”, and (4) “would you request specialized resources?”.

Gender Role Attitudes Scale (GRAS). This assesses attitudes to female and male GR. It consists of 17 items in a Likert format with 5 response alternatives, 1 (Strongly agree) to 4 (Strongly disagree), grouped into two correlated factors: Stereotypical attitudes - SA - (12 items) and Transcendent Attitudes - TA - (5 items). High scores are indicative of Stereotypical Attitudes and high Transcendent Attitudes respectively. This instrument has presented evidence of internal and external validity with a sample of Chilean university students (Pérez et al., 2020). In the study sample, internal consistency for the factor of Stereotypical Attitudes was excellent (McDonald’s Omega, .92) and good for the factor of Transcendent Attitudes (McDonald’s Omega, .83).

Religious Orientation Scale Items. The dimensions of religiosity are measured through Likert-type response questions (Fernández-Ríos et al., 2018; Fernández-Ríos et al., 2015): (1) Religious Identification (RI), “To what extent would you say you are religious?”, with response levels between 1 (Not at all) to 9 (Very much); (2) Extrinsic religiosity (ER) refers to participation in religious events and services, and it is assessed by the question, “Other than special occasions (weddings, funerals, baptisms), how often do you attend religious services?”, with response levels from 1 (Never) to 5 (Once a day); (3) and Intrinsic Religiosity (IR) on the experience and feeling that a connection to the sacred offers, which it is measured through the question, “Do you consider that your religious beliefs affect your happiness?”, with a response range between 1 (Not at all) and 9 (Very much). The same scoring model was used for the three variables by recoding the response values to 5 levels in RI and IR. The internal consistency in the study sample was excellent (McDonald’s Omega, .92).

2.3 Procedure

Voluntary and confidential participation was requested of the young people by signing an informed consent approved by the Science Ethics Committee at the Universidad de La Frontera. Then they were given the questionnaires, which were answered during school hours in an average of 30 minutes.

2.4 Plan of analysis

Descriptive and frequency analyses were used. Kolmogorov-Smirnov and Levenes tests show the violation of normality and homoscedasticity assumptions, which is not an impediment to the selection of techniques when the sample sizes are large. For the comparison of means, Welchs parametric t-test was chosen, suitable when the sizes and variations of the sample are uneven between the groups, and Cohens d corrected for uneven samples. The bivariate correlations were studied using Pearsons coefficient. As predictive techniques, linear regression is chosen when the dependent variable is quantitative and the logistic when it is dichotomous (Delacre et al., 2017; Fagerland, 2012). Finally, a two-phase clustering analysis was used. The Statistical Package for the Social Sciences (SPSS) v. 23, FACTOR 10.8.04 and JASP were used.

3. Results

3.1 Prevalence of victimizacion in dating relationships

94.2% (n=258) of men and 87.8% (n=425) of women stated having suffered at least one type of violent behavior. The most common types of violence are detachment, coercion, and humiliation (Table 1). The men reported greater victimization by coercion -Men (n=274). Mean 2.53; Women (n=484). Mean=1.62; t(559.6)=5.393; p<.001; d = .41-, and physical violence -Men (n=274). Mean=.49; Women (n=484). Mean=.20; t(418.2)=3.756; p<.001, d = .30-.

Table 1

Descriptive information on the sample in study variables

		Total			Men		Women
		n	%	n	%	n	%
Have	Coercion (n = 758)	522	68.9	218	41.8	304	58.2
suffered at	Detachment (n=758)	532	70.2	206	38.7	326	61.3
least one	Humillation (n=758)	442	58.2	175	39.6	267	60.4
violent	Physical (n = 758)	125	16.5	73	58.4	52	41.6
Behavior...	Sexual (n = 758)	220	29	76	34.5	143	65.3
Perceive themselves as abused (n = 754)		84	11.1	23	27.4	61	72.6
Would	Friends (n=731)	633	86.6	225	35.5	408	64.5
seek help	Family (n=769)	454	63.2	135	29.7	319	70.3
from...	Professors/Advisors (n=669)	96	14.3	26	27.1	70	72.9
	Specialized resources (n=678)	182	26.8	33	18.1	149	81.9
		Total			Men		Women
		Mean	SD	Mean	SD	Mean	SD
V. by Coercion (N=758)		1.94	2.24	2.52	2.22	1.62	2.19
V. by Detachment (N=758)		2.34	2.73	2.46	2.71	2.27	2.75
V. by Humiliation (N=758)		1.33	1.88	1.43	1.71	1.27	1.97
Physical V. (N=758)		.30	.92	.48	1.12	.19	.76
Total V. (N=758)		6.69	6.98	7.50	6.27	6.13	7.31
Sexual V. (N=758)		.69	1.56	.59	1.26	.75	1.71
Stereotypical A. (N=754)		21.16	7.51	25.83	7.53	18.56	6.10
Transcendent A. (N=755)		22.95	2.75	22.19	3.00	23.38	2.50
Religious Identification (N=726)		2.62	1.25	2.56	1.31	2.66	1.22
Intrinsic Religiosity (N=724)		2.44	1.46	2.45	1.51	2.44	1.44
Extrinsic Religiosity (N=738)		1.07	1.06	1.18	1.13	1.01	1.01

and negatively with the Transcendent (RI, n = 653, r = −.124, p = .001; IR, n = 663, r = −.171, p < .000; ER, n=663, r = −.112, p < .000).

The men scored higher in Stereotypical Attitudes - Men (n=271). Mean=25.83; Women (n=483). Mean= 18.54; t(470)= 13.6; p<.001; d=1.06-, and the women in Transcendent Attitudes -Men (n=272). Mean= 22.2; Women (n=483). Mean= 23.4; t(483.1)= -5.503; p<.001; d=-0.42-. No differences by sex were found in the variables on religion. These correlate positively with Stereotypical Attitudes (RI, n = 652, r = .191, p < .000; IR, n = 653, r = .247, p < .000; ER, n = 650, r = .212, p < .000)

Using linear regression analysis, the predictive value of these variables on the different types of victimization was analyzed, stratifying by sex.

For women, high stereotypical attitudes and low religious identification predict victimization by coercion; high stereotypical attitudes and low intrinsic religiosity predict victimization by detachment; and a lower score in transcendent attitudes predicts victimization by humiliation and physical violence. For men, high scores in stereotypical attitudes predict suffering violence by coercion, and lower scores in transcendent attitudes, sexual violence. The model on violence by detachment shows that this variable is predicted by high intrinsic religiosity and low extrinsic religiosity (see Table 2).

Table 2

Linear Regression Models. Prediction of victimization in DV stratified by sex

	Beta	t	Sig.	Model			Collinearity
				R2	F	p	T	DW	VIF
Coercion
Woman
SA	.200	4.233	<.001	.046	10.839	<.001	.951	.870	1.05
RI	-.135	-2.849	.005
Man
SA	.140	2.242	.026	.020	5.026	.026
Detachment
Woman
SA	.149	3.078	.002	.025	5.830	.003	.926	.732	1.08
IR	-.110	-2.261	.024
Humiliation
Woman
TA	-113	-2.410	.106	.013	5.809	<.001
Man
IR	.305	3.411	<.001	.044	5.818	.003	.476	1.143	2.10
ER	-.222	-2.482	<.001
Physical
Woman
TA	-.185	-3.994	<.001	.034	15.95	<.001
Sexual
Man
TA	-.146	-2.344	.020	.021	5.476	.020

Note. SA= Stereotypical Attitudes; TA= Transcendent Attitudes; RI= Religious Identification; IR= Intrinsic Religiosity; ER= Extrinsic Religiosity; T= Tolerance; DW= Durbin-Watson; VIF= Variance Inflation Factor

3.2 Self-perception of victimization in dating relationships

11.1% (n=84) of participants felt abused. They stated having experienced a mean of 9.14 (SD=4.27) different types of violent behaviors. The remaining 88.9% (n = 670) reported having experienced a mean of 4 (SD=3.03) types of violent behaviors (see Table 1). The difference in score on DV suffered between the two groups was statistically significant in all cases (Table 3).

Table 3

Difference of means. Number of violent behaviors suffered by type of DV

	NPA∗ (n = 670)	PA∗∗ (n = 84)	Welch’s T	gl	p	d
	M(SD)	MD(SD)
V. by Coercion	1.64(1.83)	4.39(3.47)	-7.140	88.8	<.000	-.99
V. by Detachment	2.01(2.34)	4.90(3.93)	-6.588	90.51	<.000	-.89
V. by Humiliation	1.04(1.38)	3.44(3.16)	-6.85	87.02	<.000	-.98
Physical V.	.16(.53)	1.35(2.03)	-5.33	84.44	<.000	-.80
Sexual V.	.51(1.13)	2.06(2.97)	-4.70	86.03	<.000	-.69

Note. *Do not perceive themselves as abused (NPA); ** Perceive themselves as abused (PA).

95.3% (n=566) of the participants classified with MLV, and 65% (n=104) of those classified with HLV did not feel abused. There was no significant difference between men and women regarding the self-perception of victimization, but the men who did not feel abused had a higher score on the overall DVQ-R scale -Men (n = 251), Mean=6.85; Women (n = 423), Mean=4.60; t(495.97) = 5.561; p < .001, d=.45-.

After eliminating the participants who had never suffered violent behavior (n=49, 7.2%), for each group (MLV and HLV) we performed logistic regression models to determine the effects of the attitudes to GR and the variables of religion on the perception of victimization. Sex and total score on the DVQ-R are control variables.

The participants with MLV: low scores in religious identification and high in overall victimization increase the likelihood of perceiving themselves as abused. The model explains 10.7% of the variance, although religious identification explains 7.4% itself. The participants with HLV: being a woman and high overall scores in victimization increases this likelihood. This model explains 36% of the variance, although sex by itself explains 31% (see Table 4).

Table 4

Binary logistic regression model. Self-perception as victim of DV and asking for help to end a problematic relationship

						Model
	B	S.E.	Wald	p	Exp(B)	R2	χ2	p
Perception of Abuse
MLV
RI	-.905	.209	8.875	.003	.405	.107	16.78	<.000
OV	.478	.209	5.232	.022	1.613
HLV
Sex	1.060	.411	6.644	.010	2.887	.36	45.56	<.000
OV	1.372	.325	17.822	<.000	3.994
Asking for Help
General
Sex	.567	.219	6.692	.010	1.172	.082	6.383	.012
Friends
TA	.567	.219	6.692	.010	1.172	.082	6.383	.012
Family
Sex	1.251	.401	9.732	.002	3.492	.243	28.37	<.000
SA	-.534	.206	6.706	.010	.586
Resources
Sex	1.143	.442	6.688	.010	3.135	.079	7.23	.007

Note. MLV= Medium Level of Victimization; HLV= High Level of Victimization; RI= Religious Identification; OV: Overall Victimization; TA= Transcendent Attitudes; SA= Stereotypical Attitudes.

3.3 Intention of asking for help to end a problematic relationship

Selecting the group of participants with HLV, logistic regression models were created, considering the intention to ask for help in general as a dependent variable, and for each group consulted on the DVQ-R -the model about professors for help does not converge- (descriptive data in Table 1). As covariables, stereotypical and transcendent attitudes, the variables on religion, sex and the overall score on the DVQ-R were considered.

Being a woman increases the likelihood of asking any group for help: the higher the score in Transcendent Attitudes, the greater the likelihood that the participants will turn to a friend for help. Being a woman and having low scores in Stereotypical Attitudes increases the likelihood of turning to family. This model explains 24.3% of the variance, with 18.8% being attributable to sex. Finally, being a woman is the only variable that increases the likelihood of going to specialized resources (See Table 4).

4. Discussion

In this work, we endeavor to shed light on the complexity of DV by studying victimization by DV in a sample of Chilean university students, considering a broad range of types of violence and adopting the perspective of gender symmetry. We examine the prevalence of the victimization differentiated by sex, self-perception, and intention to seek help to end a problematic relationship, and analyzed the explanatory role of sex, attitudes to GR and identification with Christian groups on these variables.

The results obtained expose DV as a problem present in the sample. This is consistent with the national and international scientific literature, highlighting high numbers for violence by detachment, coercion, and humiliation. DV affects both male and female participants, with the former reporting greater violence by coercion and physical violence (Courtain & Glowacz, 2018; Lehrer et al., 2009; R. Rodríguez et al., 2018; Rubio-Garay et al., 2017; Saldivia & Vizcarra, 2012; Vivanco et al., 2015; Vizcarra & Póo, 2011).

The literature has responded to these incongruities between the sexes on the basis of societies settling into traditional GR. The greater male victimization is attributed to a greater tolerance of female violence (Scarduzio et al., 2016; Walker et al., 2018). Chilean adolescents identify female DV as reaffirmation of woman and as a tool to demand respect (Sanhueza & Geneviève, 2018). For their part, Courtain and Glowacz (2018) state that men tolerate any type of violence more, and that female physical and sexual violence as well as male psychological violence are the most tolerated. However, the evidence suggests men and women have similar underlying motivations and factors for violence (Walker et al., 2018). This points to the need to overcome the gender hierarchy and allows for bidirectional and reciprocal violence dynamics (Archer, 2000; Arnoso et al., 2017; Courtain & Glowacz, 2018).

These results contrast with the low levels of self-perception as victims. This dissonance is indicative of a deep normalization of violent behaviors, especially the more subtle ones (Cortés-Ayala et al., 2014; López-Cepero et al., 2015; Rodríguez-Franco et al., 2012). The literature refers to a tendency to not perceive violence due to a lack of serious injuries (Moreno et al., 2016). However, those participants who perceive themselves as abused suffer any type of violence to a greater extent. Self-perception as a victim does not depend on the type of violence experienced -more or less serious-, but rather on the amount: the participants must experience an average of at least nine different types of violent behaviors to perceive themselves as abused. This number is repeated in similar studies (López-Cepero et al., 2015; Rodríguez-Franco et al., 2012). It is confirmed in this study sample that friends are the ones preferred when seeking help, with the last being specialized resources. This points to young people not knowing about complaint procedures and the distrust of institutions (Lehrer et al., 2009; López-Cepero et al., 2015; Vizcarra & Póo, 2011).

Finally, we analyzed the explanatory role of sex, attitudes to GR, and religion, considering the significant relation between the last variables. This result supports the role of religion as part of the symbolic sphere of culture that promotes and conveys traditional gender-based rules of behavior. The patterns of influence on victimization differ between sexes. This result is to be expected given that GR establish different rules of behavior for men and women, which is why they have a different meaning for each group (Berkel et al., 2004; Betancourt & Cartes, 2019; Galtung, 2016; Jankowski et al., 2018; King & Boyatzis, 2015; Nelson, 2009).

Women experience pressure to behave a certain way against their will, such as violence by coercion and emotional punishment; violence by detachment, when they accept the patriarchal rules about gender, stereotypical attitudes, and religion does not occupy an important part in their lives (religious identity and intrinsic religiosity). Stereotypical attitudes are definitely outlined as a risk factor in women (Ferrer-Pérez & Bosch-Fiol, 2019; Moreno et al., 2016; Santoro et al., 2018; Savage et al., 2016; Wilchek-Aviad et al., 2019). Religion has emerged as a protective factor, although this study is not sufficient to explain whether this is because it promotes values about respect, dignity and the care of others, or that it teaches women to behave in a relationship according to these traditional roles: men do not need to resort to violence to modify their behavior (Berkel et al., 2004; Betancourt & Cartes, 2019; Fernández-Ríos et al., 2018; Fernández-Ríos et al., 2015; Tussey & Tyler, 2019).

In addition, a greater distancing from traditional GR -transcendent attitudes- favor that women suffering less direct violence from their partner through humiliations and physical aggressions: transcendent attitudes are a useful tool to reduce tolerance of violence that is more easily labeled as such (Baber & Tucker, 2006; Bringas-Molleda et al., 2017; Galtung, 2016).

It is paradoxical that men with Stereotypical Attitudes experience greater coercion, since their gender mandate does not allow them to tolerate female violence. In addition, those who frequently attend religious acts -extrinsic religiosity-, and therefore commonly interact with people who promote and reinforce traditional values, suffer greater violence by humiliation, although greater religious identification is a protective factor. We distinguished two possible explanations: (1) men do not draw attention to the woman as the aggressor, and interpret their aggressions as of little importance; (2) to reveal themselves means to accept their role as victim, so their pattern of masculinity is disrupted. Consequently, they end up tolerating more violence. Finally, an egalitarian view of gender protects them from sexual violence, whereas none of these variables influences violence by detachment or physical violence (Arnoso et al., 2017; Courtain & Glowacz, 2018; Moreno et al., 2016; Santoro et al., 2018; Savage et al., 2016; Scarduzio et al., 2016).

Attitudes to GR are discarded as predictor variables of the perception of victimization. On the other hand, those who identify more with Christian spirituality do not perceive the violent behaviors they have experienced as abuse. This result is consistent with studies that attribute greater tolerance of violence to religion, although this tolerance is not unconditional, because the risky influence of religion disappears when victimization levels are elevated (Berkel et al., 2004; Jankowski et al., 2018; King & Boyatzis, 2015; Nelson, 2009). Conversely, the weight of the variable sex for the group with HLV (31% of the variance) is very worrying: although men suffer the same levels of victimization, being a woman is determinant in self-defining as a victim of abuse. The literature indicates that they tolerate any type of DV more because being a man and a victim disrupts the pattern of masculinity. This leads the man to: (1) minimize female violence, and (2) struggle to reconcile victimization and gender role, evading the (private and public) recognition of the violent experience to avoid humiliation and the judgment of his fellow men (Courtain & Glowacz, 2018; Rojas-Solís et al., 2019; Savage et al., 2016; Walker et al., 2018).

On the other hand, being a woman increases the likelihood that the participants classified with HLV will decide to seek help in general, from family and specialized resources. Again, man as victim does not fit in the pattern of masculinity, which entails little social support. The institutional assistance mechanisms are even oriented towards women -paradoxically reinforcing their traditional view as victims. Consequently, men prefer to keep the secret to avoid being ridiculed. However, the literature indicates that the feelings of shame, fear, and putting their credibility in doubt also paralyzes women (Arnoso et al., 2017; López-Cepero et al., 2015; F. J. Rodríguez et al., 2017; R. Rodríguez et al., 2018; Rojas-Solís et al., 2019; Vizcarra & Póo, 2011; Walker et al., 2018).

Despite what has been discussed in the previous paragraphs, we cannot avoid the influence of these variables being restricted to certain types of violence and groups, and that these models, with a few exceptions, explain a smaller percentage of the variance. This leads us to question the weight afforded by the scientific community to patriarchal mechanisms to approach an ecological view of DV (Archer, 2000; Arnoso et al., 2017; Ferrer-Pérez Bosch-Fiol, 2019; Pereda & Tamarit, 2019; Rubio-Garay et al., 2017; Santoro et al., 2018; Savage et al., 2016; Wilchek-Aviad et al., 2019). However, the literature indicates that the effect of patriarchy also operates from other variables not measured in this study (Berkel et al., 2004; García-Cueto et al., 2015). The systematic review conducted by Ramiro-Sánchez et al. (2018) establishes that sexist attitudes are related to a series of variables, such as positive attitudes to violence, risky sexual behaviors or emotional dependency, which (F. Rubio-Garay et al., 2015) identifies as precipitating, facilitating or modulating variables of the aggression suffered in DV. Consequently, we consider the analysis of the indirect and mediating effect of attitudes to GR on DV as well as the effect of patriarchal mechanisms from other related variables a future line of enquiry.

Finally, we conclude that this work is a theoretical contribution to the understanding of DV in Chilean society from gender symmetry, highlighting the impact that traditional GR have not only on woman, but also on male victimization, mens perception of violence, and seeking help. We consider that this is an element to orient DV prevention programs in schools and to encourage institutions to also consider men as possible victims.

Limitations here include the lack of representativeness of the sample, of information on the temporality of the dating relationships, and on the sexual orientation of the participants, as well as the intrinsic limitations to the self-reported measurement systems and the effects of social desirability, which have not been controlled.