On-line program for the prevention of dating violence: mixed descriptive cross-sectional and post-test pre-test study

Abstract

Introduction

Dating violence (DV) is a public health problem for which intervention is necessary. The beginning of adulthood is where more relationships begin, and where it is required to prevent DV. Although various prevention programs have been developed in high schools, no effective programs have been developed at universities. Aim: To identify the prevalence and associated factors of DV and to evaluate the effectiveness of a training program to prevent DV in college students.

Methods

Mixed, descriptive, cross-sectional pre-test post-test study. The study population included students at the University of Castilla-La Mancha. Data collection was conducted through an online questionnaire that included the following validated scales: the Conflict in Adolescent Dating Relationship Inventory (CADRI), the Cyber Dating Abuse Questionnaire (CDAQ), the Scale of Attitudes toward Violence (EAV), and the Violence in Intimate Relationships (IRV) scale. The intervention consisted of an online training program through training pills. The Ethics Committee approved the study.

Results

A total of 354 students participated in the descriptive phase. A total of 85.6% perpetrated and 85.1% were victims of some type of violence, with psychological violence being the most common form of violence. 26 students participated in the pre-test post-test phase; there were differences, although not statistically significant, between the scores on the attitudes and knowledge scales. There were correlations between the scores obtained from perpetrating violence and the scale of attitudes toward violence and the knowledge scale. Among the factors associated with DV, having a history of violence and living with a partner were significant in increasing the risk of suffering and perpetrating any type of violence.

Conclusions

DV is a major problem today, with a high prevalence of young people involved. Owing to the high prevalence of violence, the lack of training and knowledge highlights the importance of preventive training programs to reduce dating violence.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12889-025-24265-z.

Introduction

During adolescence, the first couple relationships begin to be established, with the first violent relationships being installed around the age of 15 [1]. The incidence of dating violence (DV) increases as age increases, with the highest DV peak occurring in the 15- and 30-year-old age groups [2].

DV is a set of abusive behaviors, both physical and psychological, exerted by one of the members of a couple relationship against the other [3]. These violent acts can include control, threats, emotional manipulation, physical or verbal attacks, and isolation [4]. Unlike other forms of violence, it occurs in the context of affective relationships that have not yet formalized commitment and can affect both adolescents and young adults [5]. In addition, it can also be produced through electronic media and the internet while talking about cyber DV or online DV [6].

A systematic review [7] reported that the instrument most widely used to measure the prevalence of DV was the Conflict in Adolescent Dating Relationship Inventory (CADRI). The prevalence of psychological violence is between 5.6% and 95.5%, that of physical violence is between 0.8% and 32.9%, that of sexual violence is between 2.4% and 41.0%, and that of electronic DV or cyber DV is between 0.6% and 48.0%. These figures may vary depending on gender [7, 8] and sexual orientation, with young people belonging to sexual minorities being a risk group [9]. This prevalence is high in the university stage since, according to one study [10], one-fifth of college students experience DV, and half of the students testify to DV.

Due to the high prevalence of DV, in both its face-to-face and digital forms, it is considered a phenomenon that is increasingly recognized for its impact on the physical, psychological and social health of young people. Among the consequences of suffering from DV, we find low self-esteem, depression, anxiety, suicidal thoughts, mental problems, difficulties in emotional self-regulation, drug use, risky sexual behaviors, and deterioration of academic performance [11, 12]. Experiences of early violence in partner relationships may predict the continuity of the violent pattern in future relationships, either as victims or aggressors, generating a cycle that is difficult to break [13]. In the specific case of cyber-violence, the negative effects include social isolation, deterioration of digital well-being, and impairment of self-esteem, even when the violent acts are minimized or normalized by those involved [14]. In addition, it poses a risk to the development, well-being, and the creation of future healthy relationships [15].

There are also risk factors for DV, they could be individual, family, and sociocultural factors that increase the likelihood of involvement in DV relationships, either as a victim or as an aggressor. At the individual level, having experienced violence in childhood, having low self-esteem, poor conflict resolution skills, and substance use are consistently associated with higher levels of perpetration and victimization [16]. In addition, socio-cultural factors such as maintaining traditional gender roles, living in a hetero-patriarchal culture, the ideals of romantic love, socioeconomic status, a history of violence in the immediate environment and the normalization of violence in the media and social networks reinforce permissive attitudes towards abusive behavior, especially among adolescents [6, 11, 17]. In the family setting, exposure to violent dynamics in the home, such as interparental abuse or authoritarian parenting styles, has also been linked to the reproduction of violent behaviors in subsequent romantic relationships [18, 19]. The acceptance of violent behaviors is a mediating factor in most of these relationships, as well as between interparental violence and DV [20–22]. Distorted perceptions contribute to minimizing the impact of emotional and psychological violence, facilitating the reproduction of abusive patterns in relationships and continuing the cycle of DV [23, 24].

Understanding the consequences and, especially, the risk factors associated with the acceptance of violence is essential to designing effective prevention programs. The minimization or invisibilization of violent behaviors hinders the perception of risk and delays help-seeking, which increases negative psychological consequences such as anxiety, isolation and revictimization [25]. Therefore, preventive programs should focus on challenging permissive attitudes towards violence, promoting skills to identify abusive behaviors and fostering relationships based on respect, equality and autonomy.

Given the magnitude of the problem, interventions have been carried out for adolescents in the United States and America [26]. These programs include “Safe Dates” [27], “Dating Matters“ [28], “Date Smart“ [29] and “Lights4Violence“ [30], all of which are carried out with adolescents.

In Spain, we also find several programs conducted in schools and institutes, such as the PREVIO program [31], which focuses interventions on risk factors and protective factors to prevent DV. The DARSI program [32] focuses on modifying the perception of love (traditionally associated with jealousy, control, and dependence) and tolerance toward DV. The program “PRO-Move Healthy Relationships“ [33] focused on hostile sexism, myths of romantic love, and gender-based violence. The Date-e Adolescence program [34] significantly modified the myths of romantic love, improved self-esteem and favored anger regulation. We also found initiatives as “Amor 2.0” and “Violencia Zero” [35, 36] carried out by social entities, but without evaluation of their effectiveness.

To our knowledge in the Spanish context, there are no DV prevention programs for university students or young adults, nor are any programs conducted online or through technology in Spain. Studies conducted in the university population worldwide obtain satisfactory results by increasing knowledge and attitudes toward DV but are not effective when trying to obtain responses/actions from observers [37]. Training and awareness about DV are fundamental tools for its prevention, especially in the university stage, when more relationships start and when there is a greater risk of suffering and exercising DV [1, 38].

Due to its magnitude, high prevalence, and the widespread social normalization of violent behaviors, dating violence constitutes a global public health issue that urgently requires comprehensive intervention [39, 40]. Relationships are especially important in adolescence and early adulthood and affect emotional and psychological development, making it necessary to create prevention and promotion programs for healthy relationships [41]. As mentioned above, training programs are effective tools for prevention, and no programs of this type have been found in Spanish universities thus far. Given that the university stage is when adolescents start their first violent relationships, training young people about DV is important. In addition, online learning has proven to be highly effective, positively affecting student motivation [42].

Therefore, the main objective of this study was to evaluate the effectiveness (based on knowledge and acceptance of violent behaviors) of a training program on DV in university students. As secondary objectives, the prevalence and types of DV (online and offline) and associated factors were measured to better understand the characteristics of the university sample studied.

Methods

Design

A mixed study was conducted: a cross-sectional descriptive study and a longitudinal and prospective pre-test post-test among students at the University of Castilla‒La Mancha (UCLM) from March 2023 to May 2024.

The descriptive cross-sectional study was carried out following the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) [43]. The pretest and posttest studies were conducted following the Transparent Reporting of Evaluations with Nonrandomized Designs (TREND) guidelines [44].

Participants

The reference population was the UCLM students enrolled in the 2023/2024 courses, consisting of a total of 27,631 students. The participants were contacted through institutional email by convenience, which was disseminated through social networks.

For the descriptive cross-sectional phase, a random sample of 348 individuals was necessary to estimate, with a confidence level of 95% and a precision of +/- 5% units, a population percentage that is expected to be approximately 68%, taking as a reference the study carried out by Tarriño-Concejero [12]. A loss to follow-up rate of 5% was estimated, and the POISSON approach was used.

For the pre-test post-test phase, with an alpha risk of 0.05 and a statistical power greater than 0.8 in a bilateral contrast, 24 subjects were required to detect a difference equal to or greater than 0.06 units as statistically significant. Concerning the results obtained by K. Peterson [45], the Gender Violence Scale was used to measure the acceptance of violence. Standard deviations of 0.4 and 0.5 were estimated for the first and second measures, respectively. It is calculated assuming a correlation between the two measures of 1. A loss to follow-up rate of 10% was estimated.

Instruments

An online questionnaire composed of an ad hoc questionnaire was administered to collect the sociodemographic variables: age, sex, sexual orientation, marital status (single, in an undefined relationship, in a serious relationship or married, the first two options being understood as low attachment relationships and the other two as high attachment relationships), cohabitation with a partner, socioeconomic status, social class, toxic habits, history of abuse in the family environment, history of DV, knowledge of the UCLM protocol on violence and harassment, and training on intimate partner violence. And the following validated scales:

• Conflict in Adolescent Dating Relationship Inventory (CADRI): validated for Spanish by Fernández Fuertes et al. (2006) [46] of the original scale of Wolfe et al. (2001) [47]. This questionnaire consists of 35 items (25 items on violence and 10 on positive behaviors in relationships that are not addressed in this study) rated on a Likert scale with four response options ranging from 1, ‘never’, to 4, ‘always’ that evaluate violent behavior toward the partner. It examines five dimensions of violence: relational, threats, verbal-emotional, and physical violence. Initially designed for use in adolescents, it was also validated in a Spanish university population (19–25 years old, n = 976) showing a good fit of the confirmatory factorial model and test-retest reliability and concordance with external observations [48]. The reliability (Cronbach’s alpha) is 0.86. It has also been widely used in studies of teenage and dating violence and is considered one of the most reliable for measuring psychological, physical and verbal perpetration and victimization [49, 50].

• Cyberg Dating Abuse Questionnaire (CDAQ): Developed by E. Borrajo et al. (2015) [14] in Spain, this scale comprises 20 items and uses a 6-point Likert-type scale ranging from 1 never to 6 always or more than 20 times. In addition, the scale is composed of two factors, control and direct aggression, with parallel items for victimization and perpetration of violence. With a Cronbach’s alpha (α) of 0.73 for the direct assault perpetration scale, α = 0.84 for direct assault victimization, α = 0.81 for control perpetration and α = 0.87 for control victimization. In addition, is highlighted in recent reviews [51] as one of the most recommended instruments for assessing digital violence in young people, including cyber aggression and control, and several cross-cultural validations [52, 53] have been carried out, showing acceptable fit in Confirmatory Factor Analysis, convergent/discriminate validity, and overall reliability α > 0.87.

• Scale of Attitudes toward Violence (EAV): Validated in the Spanish context by J. Ortuño-Sierra et al. (2021) [54] and developed by MB Vizcarra Larrañaga and AM Poo Figueroa (2010) [55]. It is composed of 10 items with a 5-point Likert-type scale ranging from 1 (totally disagree) to 5 (totally agree); the higher the score is, the greater the acceptance or positive attitudes toward the use of violence, with an α value greater than 0.90.

• Violence in Intimate Relationships (IRV): Developed by the MA. Rodrigues Dixe and da Silva Amaro de Oliveira Fabião (2013) [56]. It consists of 47 true or false response items; the higher the score is, the greater the degree of knowledge of violence. The scale addresses multiple dimensions, including physical, psychological, sexual, economic, and normative aspects related to violence. It has been used with university populations and professionals in training (nursing, psychology, education), with the purpose of identifying knowledge gaps and guiding prevention and training programs. It has shown evidence of internal consistency of scores in previous studies, with alpha values greater than 0.90 [57], and content validity reviewed by experts in health and gender violence.

Procedure

The cross-sectional descriptive phase consisted of the administration of an online questionnaire, which included different tests. Before the administration, the participants read the information sheet and signed the informed consent form. At the end of the questionnaire, the participants were asked if they wanted to participate in a pre-test post-test study to prevent and reduce DV. Therefore, the participants in the pre-test post-test phase were recruited from those in the descriptive cross-sectional study who expressed a desire to participate in the intervention aimed at preventing and reducing DV by providing an email to participate in the online program, all participants who showed interest were invited to participate. The only exclusion criterion for participation was to be over 30 years of age.

The questionnaire of the cross-sectional descriptive phase was taken as the questionnaire of the pre-test phase for the participants who showed their desire to participate in the DV prevention program. The pre-test questionnaire included sociodemographic variables, the CADRI, the CDAQ, the EAV and the IRV, while the post-test questionnaire only included the EAV and the IRV.

The intervention consisted of an online training program comprising educational short videos, each between 5 and 10 min long, which participants viewed independently over 5 weeks. The online videos gave participants the advantage of completing the program whenever and wherever they wanted, without being forced to adhere to a schedule or go to a certain location to complete the program. Once a week, a link to the new video was sent individually to each of them, and they were reminded to watch the previous videos. The videos were available to participants continuously for the duration of the program. Once all the videos were sent and two weeks before the post-test evaluation, a reminder was sent with the aim that participants could finish watching all the videos.

The theme and information contained in the videos was as follows: definition of DV, types of violence and how to identify them, who is affected by DV, consequences of suffering and exercising violence, potential causes and factors that favor DV, normalization of violence and abusive behavior, acceptance of DV, and finally, characteristics of a healthy relationship.

At the beginning of the visualization of the videos (pre-test) and after the visualization (post-test), the participants answered the survey, which measured the acceptance of violence and knowledge about DVs.

Statistical analysis

For the statistical analysis, the SPSS version 29 program, licensed by the University of Castilla-La Mancha, will be used. The qualitative variables were measured as counts and percentages. The quantitative variables are expressed as arithmetic means (m) and standard deviations (SD).

Inferential analyses were conducted to examine the relationships between the independent and dependent variables:

For qualitative variables, a comparison of the proportions of the categorical variables was performed via chi-square tests for contingency tables. The chi-square statistic with Yates correction was used, and when the expected frequency was ≤ 5, Fisher’s exact test was applied. For the qualitative variables, the goodness of fit to a normal distribution was determined (using the Shapiro‒Wilk test), and the homogeneity of the variances was verified (using the Levene test). The data did not fit a normal distribution; therefore, nonparametric tests (Wilcoxon) were applied.

Spearman correlation was performed for the scores of the different scales used. In addition, to control for the influence of sex, a partial correlation was performed.

The different dimensions of DV and victimization of the CADRI questionnaire and the CDAQ, as well as the total scores of both dimensions, were dichotomized to categorize the groups (0 = no and 1 = yes). A zero-tolerance criterion was applied. Specifically, any item in which the participant indicated that a violent act occurred at least once (i.e., selecting “rarely”, “sometimes” or “often”) was considered a positive case of perpetration or victimization. Students were classified as perpetrators or victims of DV if they marked one or more elements of perpetration/victimization. The same method was applied for each DV perpetration and victimization subscale, as in previous studies.

Logistic regression was also performed to identify the variables associated with violence. Odds ratios (ORs) were calculated with their respective 95% confidence intervals (95% CIs). All hypothesis tests were bilateral. In all the statistical tests, those whose confidence level was 95% (p <.05) were considered statistically significant.

To determine if there were significant differences in the pretest and posttest phases, the means obtained from the EAV and IRV were compared through a nonparametric test of the Wilcoxon signed test for each scale.

Ethical considerations

The present study was approved by the Clinical Research Ethics Committee of the Health Area of Talavera de la Reina (Toledo) with code 01/2021. All participants read the information sheet before answering the questionnaires and gave their written consent to participate in the study, they could not proceed to answer the online questionnaires without signing the informed consent form.

Results

Sociodemographic variables of the participants

A total of 354 university students participated in the cross-sectional study, of whom 18.6% were men and 81.1% were women. The mean age was 21.80 years (SD ± 4.715).

Concerning tobacco consumption, 90% of the participants were nonsmokers, with an average consumption of 7.86 cigarettes per day among smokers. Table 1 show the sociodemographic characteristics of the participants.

Table 1.

Sociodemographic variables of the participants

		Sex			p-value
		Men	Women	Total
		N (%)	N (%)	N (%)
Sexual orientation	Straight	56 (84.8%)	236 (81.9%)	292 (82.5%)	0.091
	Homosexual	4 (6.1%)	6 (2.1%)	10 (2.8%)
	Bisexual	6 (9.1%)	46 (16.0%)	52 (14.7%)
Marital status	Single	27 (40.9%)	103 (35.8%)	130 (36.7%)	0.834
	In an undefined relationship	6 (9.1%)	25 (8.7%)	31 (8.8%)
	In a serious relationship	31 (47%)	153 (53.1%)	184 (52%)
	Married	2 (3.0%)	7 (2.4%)	9 (2.5%)
Living as a couple	Yes	8 (12.1%)	26 (9.0%)	34 (9.6%)	0.645
	No	58 (87.8%)	262 (90.9%)	320 (90.4%)
Social class	Class I	5 (7.6%)	27 (9.4%)	32 (9.0%)	0.306
	Class II	12 (18.2%)	30 (10.4%)	42 (11.9%)
	Class III	20 (30.3%)	67 (23.3%)	87 (24.6%)
	Class IV	10 (15.2%)	49 (17.0%)	59 (16.7%)
	Class V	12 (18.2%)	78 (27.1%)	90 (25.4%)
	Class VI	7 (10.6%)	37 (12.8%)	44 (12.4%)
Socioeconomic level	High	3 (4.5%)	8 (2.8%)	11 (3.1%)	0.700
	Medium	53 (80.3%)	241 (83.7%)	294 (83.1%)
	Low	10 (15.2%)	39 (13.5%)	49 (13.8%)
History of family violence	No	57 (86.4%)	251 (87.2%)	308 (87%)	0.863
	Yes	9 (13.6%)	37 (12.8%)	46 (13%)
DV background	No	60 (90.9%)	221 (76.7%)	281 (79.4%)	0.010
	Yes	6 (9.1%)	67 (23.3%)	73 (20.6%)
DV training	No	49 (74.2%)	187 (64.9%)	236 (66.7%)	0.148
	Yes	17 (25.8%)	101 (35.1%)	118 (33.3%)
Alcohol consumption	Never	12 (18.2%)	42 (14.6%)	54 (15.3%)	0.929
	Rarely	19 (28.8%)	86 (29.9%)	105 (29.7%)
	Occasional	32 (48.5%)	148 (51.4%)	180 (50.8%)
	Habitual	3 (4.5%)	11 (3.8%)	14 (4.0%)
	Daily	0 (0%)	1 (0.3%)	1 (0.3%)

Violence results

Among the total participants, 85.6% perpetrated and 85.1% were victims of some type of violence, the most common being psychological violence. The mean score for the CADRI perpetration subscale was 30.09 points (SD ± 4.89), and that for the victimization subscale was 34.74 points (SD ± 9.82). Table 2 shows the prevalence according to the type of DV by sex.

Table 2.

Types of violence exercised and suffered by sex

Type of VN		Sex			p-value
		Men	Women	Total
		N (%)	N (%)	N (%)
V Physical Perpetrated	No	62 (93.9%)	275 (95.5%)	337 (95.2%)	0.596-
	Yes	4 (6.1%)	13 (4.5%)	17 (4.8%)
V Suffered Physical	No	59 (89.4%)	265 (92%)	324 (91.5%)	0.491
	Yes	7 (10.6%)	23 (8%)	30 (8.5%)
Threats Perpetrated	No	62 (93.9%)	263 (91.3%)	325 (91.8%)	0.484
	Yes	4 (6.1%)	25 (8.7%)	29 (8.2%)
Threats Suffered	No	56 (84.8%)	237 (82.3%)	293 (82.8%)	0.620
	Yes	10 (15.2%)	51 (17.7%)	61 (17.2%)
V Sexual Perpetrated	No	45 (68.2%)	253 (87.8%)	298 (84.2%)	< 0.001
	Yes	21 (31.8%)	35 (12.2%)	56 (15.8%)
V Sexual Suffered	No	44 (66.7%)	201 (69.8%)	245 (69.2%)	0.620
	Yes	22 (33.3%)	87 (30.2%)	109 (30.8%)
V Realational Perpetrated	No	59 (89.4%)	276 (95.8%)	335 (94.6%)	0.036
	Yes	7 (10.6%)	12 (4.2%)	19 (5.4%)
V Relational Suffered	No	48 (72.7%)	231 (80.2%)	279 (78.8%)	0.180
	Yes	18 (27.3%)	57 (19.8%)	75 (21.2%)
V Perpetrated Psychological	No	11 (16.7%)	47 (16.3%)	58 (16.4%)	0.945
	Yes	55 (83.3%)	241 (83.7%)	296 (83.6%)
V Suffered Psychological	No	10 (15.2%)	47 (16.3%)	57 (16.1%)	0.816
	Yes	56 (84.8%)	241 (83.7%)	297 (83.9%)
Control Perpetrated	No	43 (65.2%)	189 (65.6%)	232 (65.5%)	0.942
	Yes	23 (34.8%)	99 (34.4%)	122 (34.5%)
Control Suffered	No	38 (57.6%)	188 (65.3%)	226 (63.8%)	0.240
	Yes	28 (42.4%)	100 (34.7%)	128 (36.2%)
Direct Aggression Perpetrated	No	61 (92.4%)	275 (95.5%)	336 (94.9%)	0.307
	Yes	5 (7.6%)	13 (4.5%)	18 (5.1%)
Direct Aggression Suffered	No	54 (81.8%)	243 (84.4%)	297 (83.9%)	0.610
	Yes	12 (18.2%)	45 (15.6%)	57 (16.1%)
V Electronic Perpetration	No	42 (63.6%)	187 (64.9%)	229 (64.7%)	0.843
	Yes	24 (36.4%)	101 (35.1%)	125 (35.3%)
V Suffered Electronics	No	37 (56.1%)	182 (63.2%)	219 (61.9%)	0.282
	Yes	29 (43.9%)	106 (36.8%)	135 (38.1%)

Correlations between DVs

Table 3 shows the correlations between the total DV scores. Cyber DV, acceptance of violence and knowledge scores. The partial correlations adjusted for sex are also shown.

Table 3.

Simple and partial correlations between perpetrating and suffering from DV, cyberdv, acceptance of and knowledge about DV

Bilateral Spearman correlation	DV Perpetrated	DV Suffered	Cyber DV Perpetrated	Cyber DV Suffered	Acceptance toward DV	DV knowledge
DV Perpetrated	-	0.772**	0.452**	0.392**	0.145**	−0.142**
DV Suffered		-	0.366**	0.556**	0.104*	−0.077
CyberDV Perpetrated			-	0.566**	0.136*	−0.113*
CyberDV Suffered				-	0.010	−0.054
Attitudes toward DV					-	−0.043
DV knowledge						-
Partial Correlation Mediated by Sex	DV Perpetrated	DV Suffered	CyberDV Perpetrated	CyberDV Suffered	Acceptance toward DV	DV knowledge
DV Perpetrated	-	0.548**	0.502**	0.314**	0.125*	−0.167*
DV Suffered		-	0.220**	0.659**	0.089	−0.035
CyberDV Perpetrated			-	0.377**	0.099	−0.178**
CyberDV Suffered				-	0.078	−0.047
DV attitudes					-	−0.086
DV knowledge						-

Factors associated with exercising and suffering DV

The sociodemographic variables associated with perpetrating DV were having a history of DV (OR = 2.732, p =.041) and living with a partner (OR = 6.339, p =.072). The DV of victimization was also associated with a history of DV (OR = 2,878 p =.031) and living with a partner (OR = 6.659, p =.065). The perpetration of CyberDV was associated with having a history of DV (OR = 1.820, p =.025) and victimization of CyberDV with being a man (OR = 1.649, p =.85), having a history of DV (OR = 4.397, p <.001), living with a partner (OR = 1.962, p =.089) and being single or in a relationship with a low level of attachment (OR = 1.547, p =.072). Table 4 shows the results of the logistic regressions conducted for each type of violence.

Table 4.

Logistic regressions for the associations between sociodemographic data and types of violence

Violence Perpetrated

DV

Cyber DV

Physics

Threats

Sexual

Social

Psychological

Control

Direct Aggression

OR (95%CI)

p-value

OR (95%CI)

p-value

OR (95%CI)

p-value

OR (95%CI)

p-value

OR (95%CI)

p-value

OR (95%CI)

p-value

OR (95%CI)

p-value

Background of VN No Yes

Reference 3.72

0.009

Reference 3.042

0.009

-

Reference 3.821

0.009

Reference 2.589

0.036

Reference 1.681

0.054

Reference 7.055

< 0.001

Living as a couple No Yes

-

Reference 3.600

0.009

Reference 2.953

0.010

Reference 0.000

0.998

Reference 3.513

0.092

Reference 1.838

0.096

Social class I II III IV V VI

-

Reference 4.430 3.635

0.057 0.102

-

-

-

Marital status Single Relationship

-

-

2.589 Reference

0.087

Sex Female Man

Reference: 3.536

< 0.001

Reference 3.753

0.012

3.043 Reference

0.064

Sexual orientation Bisexual Straight Homosexual

Reference: 1.735 0.000

0.275

Suffered Violence

DV

Cyber DV

Physical suffered

Threats Suffered

Sexual Suffered

Social Suffered

Psychological Suffered

Control

Direct Aggression

OR (95%CI)

p-value

OR (95%CI)

p-value

OR (95%CI)

p-value

OR (95%CI)

p-value

OR (95%CI)

p-value

OR (95%CI)

p-value

OR (95%CI)

p-value

Background of VN No Yes

Reference 5.353

< 0.001

Reference 8.320

< 0.001

Reference 4.624

< 0.001

Reference 4.882

< 0.001

Reference 3.813

0.009

Reference 3.589

< 0.001

Reference 8.241

< 0.001

Living as a couple No Yes

-

Reference 3.558

0.012

Reference 2.753

0.014

-

Reference 7.156

0.056

Social class I II III IV V VI

-

Reference 2.773 2.407

0.034 0.069

-

-

-

VN training Yes No

-

Reference 2.288

0.024

-

Marital status Single Relationship

3.122 Reference

0.001

2.210 Reference

0.003

3.494 Reference

< 0.001

1.964 Reference

0.034

Sex Female Man

-

Reference 1.979

0.048

1.715 Reference

0.062

Socioeconomic level High Medium Low

Reference 5.014 5.163

0.025 0.044

Acceptance toward violence and knowledge scale

A mean score of 11.20 (SD ± 5.53) was obtained for the pre-test phase on the scale of acceptance of violence. The mean score for men was 10.86 (SD ± 5.12), and that for women was 11.28 (SD ± 5.63) (p =.629). For the knowledge scale, a mean score of 42.28 (SD ± 2.94) was obtained; men obtained a mean score of 41.65 (SD ± 3.69), and women obtained a mean score of 42.43 (SD ± 2.72) (p = 178) (Appendix 1).

Results of the pretest and posttest interventions

26 participants (7.34%) completed the pretest-posttest study. The differences between the scores on the scale of attitudes toward and knowledge of violence can be seen in Appendix 2. The score on the scale of acceptance of violence decreases, although it is not significant. Similarly, the knowledge score also increased, although the difference was not significant.

Disscusion

This study identified the prevalence of DV and its associated factors and measured the effectiveness of the DV prevention program through online training pills viewed over 5 weeks in university students. The high prevalence found agrees with the results found in other similar studies [7] and in others that have analyzed the data according to zero tolerance [48, 58].

Among the factors associated with experiencing DV, living with a partner and the antecedents of DV in past relationships are the main factors. This may be due to the acceptance of DV, which, as we have observed in the results, has a positive relationship with perpetration or the cycle of violence. According to another study [59], violence in the peer group is a predictor of suffering and exercising DV in the future.

Importantly, 37% of the participants in the present study had not received training in DV and were not willing to participate in the intervention; 66% of the participants had not received any type of training, and 15% despite having prior training, were willing to participate. This, as in other studies [60], supposes a need to involve and raise awareness to a greater extent among the young people to whom the intervention is directed.

Concerning the differences by sex, significant differences were found in the perpetration of sexual violence, where women perpetrated 12.2% and men perpetrated 31.8%, and relational or social violence, where women perpetrated 4.2% and men perpetrated 10.6%, which is also reflected in the logistic regressions performed, men perpetrate sexual violence 3.5 times more than women and social violence 3.7 times more than women. In terms of online violence or cyberviolence, women suffer 1.7 times more violence than men, but they exert more direct aggression, 3 times more than men. This could be explained through the relationship between structural sexism and hostile sexism with sexual violence, since its relationship with gender has been proven in different studies, with a higher prevalence in the male population [19, 20].

These differences align with the international literature on interpersonal and digital violence, in university samples, men consistently report higher perpetration of sexual coercion and other forms of physical and verbal aggression, whereas women tend to engage more in indirect or psychological aggression (such as rumors or social exclusion) [61, 62]. In addition, in digital contexts, it has been documented that men tend to perpetrate direct aggression through social networks, but women are more frequently involved in cyber monitoring or control practices, as well as in more subtle forms of indirect aggression [63]. Other studies indicate that women are more vulnerable as victims of online violence, especially sexual harassment or cyber-stalking, and experience greater psychological impacts [64]. Consequently, our results reflect complex gender dynamics: sexual and social violence remains predominantly male, while in the digital realm we find duality, with women both suffering more violence and often exhibiting different forms of aggression, more direct in some cases. This dual profile underscores the need to design differential interventions, sensitive to the type of violence and gender, to address these disparities in aggression and victimization.

The results of the pre-test and post-test phases indicate that the intervention could be effective for some participants, reducing the acceptance of violence in relationships and increasing the knowledge of the participants about DV. The fact that the differences in the mean scores were not significant may be due to different reasons. Although, number of participants had exceeded the sample size initially calculated, it would be necessary to conduct future studies with larger sample sizes. This could also be due to the duration of the intervention, since previous studies with similar samples involved interventions with longer durations [65, 66].

It may also be since the median violence acceptance score was at the minimum of acceptance in both the pre-test and post-test phases, as well as the scores on the knowledge scales, which were high in both cases. While this may generate a floor effect, previous studies have pointed out that a minimal baseline score may limit the detection of subsequent changes. In similar contexts, some interventions designed to modify attitudes toward violence have reported null results on pre-post measures [67–70], attributable precisely to very low baseline scores that leave little room for improvement. Therefore, in our case it is plausible that pretest levels may have restricted the potential for change, even in the face of effective interventions, suggesting that future studies should consider this phenomenon and consider samples with sufficient variability in baseline attitudes.

Learning through online videos has shown benefits for learning, increased effectiveness, reduced effort, and increased attractiveness to students. However, to achieve these benefits, videos must be easy to view, short, relevant, and interesting, and students must have the necessary resources to view them [71]. These characteristics were considered for the design of the videos in the intervention of this study.

Other studies, such as those carried out by Coker A. et al. [72, 73] found that different types of populations demonstrate the effectiveness of bystander programs in reducing the acceptance of DV. Furthermore, the study by Savasuk-Lxton R. et al. [74] examines the relationship between education in healthy relationships and the acceptance of intimate partner violence and beliefs about traditional gender roles, obtaining positive results after the intervention and managing to reduce both variables.

In line with education in healthy relationships and knowledge about DV, another study conducted with nurses [75] reported negative correlations between the scores obtained on the knowledge scale and the scores obtained on the DV scale. These data coincide with those obtained in the present study, implying that the greater the level of knowledge of violence is, the lower the probability of suffering and perpetrating DV. In addition, the increase in knowledge about DV also generates an increase in interventions in these people when they observe a case of DV [76, 77]. This support can be key to addressing situations of violence since DV victims often seek help informally and mainly in their circle of peers [26, 78].

According to a systematic review of DV prevention programs, 80% of the programs analyzed included education on healthy relationships, followed by the promotion of equitable attitudes according to gender with 40% agreement, with the same level of agreement for the rest of the contents and varying according to the type of program and the audience it was aimed at [71]. In addition to the studies analyzed in this review, only 50% achieved significant results after the interventions were evaluated. Other interventions to prevent violence in young people through digital media have sought to increase skills such as conflict resolution, emotion regulation, empathy, self-efficacy, knowledge, and social skills [79].

Limitations

First, although the sample was representative, the convenience sampling of participants in the descriptive phase may have affected the results of the intervention. The low participation rate could have biased the results toward those students with a preexisting interest in issues of violence, potentially excluding those who do not consider the issue relevant or do not recognize signs of violence in their relationships.

Second, the online format of the intervention may have influenced its effectiveness since the lack of interaction with the participants could have reduced the impact of the training content and reflection after viewing the videos. In addition, the virtual environment may have affected the attention and level of commitment of the participants, who could have experienced distractions or technical difficulties that limited their participation in the study. Moreover, any specific technical mechanisms were not implemented to verify that all participants viewed the video in its entirety. Although it was clearly stated that it was necessary to view the entire content before continuing, it cannot be guaranteed with complete certainty that all participants did so.

Finally, as the questionnaires are self-reported, it is possible that there was bias in the responses. Moreover, the use of an online quantitative methodology may not adequately capture nuances or relevant contextual factors in the relationships of the students.

One strength is the size of the sample in a descriptive cross-sectional study. Similarly, another key aspect of this study was the experimental design, which allowed the evaluation of changes in knowledge and attitudes. In addition, this is the first study to evaluate the effectiveness of a training program for university students in Spain. Finally, as a strength within the intervention, we find that the program has been conducted online, adapting the training to the needs of the students and facilitating their participation.

Conclusion

The high prevalence rates of DV in university students, especially of psychological violence understood as verbal and emotional violence and of cyber-DV through control actions, make DV a public health problem and necessitate the creation of more programs for prevention in this age group.

Prevention programs through online videos seem to be effective in improving knowledge and reducing the acceptance of violent behaviors in the young university population since improvements were observed after the implementation of the program. However, future studies with larger sample sizes are necessary to obtain more solid conclusions.

In addition, there were significant correlations between the perpetration of DV, the acceptance of violent behaviors, and the level of knowledge, which indicates that DV can be decreased if we increase knowledge and decrease the acceptance of violent behaviors through awareness of violence.

It is necessary to continue researching how to make DV prevention programs more attractive to the young population to increase their effectiveness. Furthermore, this study was the first to evaluate the effectiveness of a training program on DV in the university population in Spain.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

DV

Dating Violence

CADRI

Conflict in Adolescent Dating Relationship Inventory

UCLM

University of Castilla‒La Mancha

STROBE

Strengthening the Reporting of Observational Studies in Epidemiology

TREND

Transparent Reporting of Evaluations with Nonrandomized Designs

CDAQ

Cyberg Dating Abuse Questionnaire

EAV

Scale of Attitudes toward Violence

IRV

Violence in Intimate Relationships

m

Arithmetic means

SD

Standard deviations

ORs

Odds ratios

95% CIs

95% confidence intervals

Author contributions

Conceptualization, B.C., C.T., and L.A.; methodology, B.C. and C.T.; validation, M.G. and C.H.; investigation, B.C.; resources, C.T.; writing—original draft preparation, B.C.; writing—review and editing, R.C., C.T., M.G., C.H., and L.A.; visualization, B.C.; supervision, C.T.; project administration, C.T. All authors have read and agreed to the published version of the manuscript.

Funding

This research was funded by a grant from the European Regional Development Fund (ERDF) [European Regional Development Fund (ERDF), (DOCM 01/27/2021)].

Esperanza Barroso-Corroto is supported by a grant (SBPLY/23/180502/000002) from the Junta de Comunidades de Castilla-La Mancha (Spain) and cofinanced by the European Social Fund Plus (ESF + 2021–2027) Program.

Data availability

The minimum data set is published within the body of the article. More data supporting this study’s findings are available from the corresponding author upon reasonable request.

Declarations

Ethics approval and consent to participate

This study was approved by the Ethics Committee of Clinical Research of the Health Area of Talavera de la Reina (Toledo) with code 01/2021. All participants read the information sheet before answering the questionnaires and gave their written consent to participate in the study, they could not proceed to answer the online questionnaires without signing the informed consent form.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.

Clinical trial number

Not applicable.