Attitudes toward dating violence and their relationship with gender role attitudes among university students: a cross-sectional study

Abstract

Background

Dating violence is an important public health and psychosocial concern among young adults. Gender role attitudes may be associated with whether violent behaviors within intimate relationships are recognized, tolerated, or normalized. Evidence regarding gender differences and subscale-level associations between attitudes toward dating violence and gender role attitudes among university students remains limited. This study aimed to examine university students’ attitudes toward dating violence and gender roles, compare total and subscale scores by gender, and investigate associations between the two constructs.

Methods

This cross-sectional study included 1,866 students at a state university. Data were collected using a sociodemographic form, the Dating Violence Attitude Scale, and the Gender Roles Attitude Scale. Gender comparisons were performed using parametric or non-parametric tests as appropriate. Associations between attitudes toward dating violence, gender role attitudes, and age were analyzed using Spearman correlation coefficients.

Results

The observed Dating Violence Attitude Scale scores suggested relatively low reported acceptance of dating violence; however, this interpretation was descriptive because the scale has no established categorical cut-off point. Male students had higher total Dating Violence Attitude Scale scores than female students, indicating lower reported acceptance, although the effect was very small and was the only gender difference on this scale that remained significant after Bonferroni correction (r = 0.06, p = 0.006). Differences in physical and economic violence subscales did not remain significant after correction. Male students had higher total Gender Roles Attitude Scale scores and higher female, marital, traditional, and male gender role subscale scores, whereas female students had higher egalitarian gender role subscale scores. Total Dating Violence Attitude Scale and Gender Roles Attitude Scale scores were moderately and positively correlated (r = 0.35, p < 0.001).

Conclusions

Gender role attitudes were associated with attitudes toward dating violence. Because of the cross-sectional and correlational design, these findings do not establish causal, predictive, or behavioral effects. Future longitudinal and experimental studies are needed to clarify the direction and practical relevance of these associations.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s12889-026-28542-3.

Introduction

Dating violence is an important public health and psychosocial problem involving physical, psychological or emotional, sexual, economic, and technology-facilitated behaviors through which one partner harms, intimidates, coerces, monitors, or controls another within a romantic relationship [1–3]. It can range from behaviors that may initially be minimized, such as persistent jealousy, humiliation, social isolation, and digital monitoring, to physical assault and sexual coercion. Because psychologically controlling and coercive behaviors may not always be readily recognized as violence, dating violence can remain concealed or become normalized within romantic relationships.

Epidemiological findings indicate that dating violence is common during adolescence and young adulthood, although reported prevalence varies according to the population studied, the definition of violence, the period assessed, and the measurement instrument used. International evidence has documented dating violence among secondary-school and university populations across different sociocultural settings [1, 4, 5]. Nationally representative data from the United States showed that 13.6% of high-school students who had dated during the preceding year reported physical or sexual dating violence victimization, including 8.5% reporting physical violence and 9.7% reporting sexual violence [6]. Studies involving university students have likewise demonstrated that physical, emotional, psychological, sexual, and digital forms of dating violence remain relevant during the transition to adulthood [7–10]. These findings indicate that dating violence is not confined to a small clinical subgroup but represents a broader public health concern affecting young people across educational and cultural contexts. However, the present study did not assess dating violence victimization, perpetration, or actual relationship behaviors; instead, it focused specifically on attitudes toward dating violence.

University students constitute a particularly important population for studying attitudes toward dating violence. University years frequently coincide with emerging adulthood, a developmental period characterized by increasing autonomy, identity exploration, changing social networks, and the formation of more independent and potentially longer-lasting romantic relationships [11]. During this period, students may experience reduced parental supervision, increased exposure to diverse peer norms, greater use of digital communication technologies, and new interpersonal and sexual experiences. Campus social environments may therefore provide both opportunities for healthy relationship development and contexts in which controlling or violent relationship patterns can emerge. Attitudes toward dating violence consolidated during this period may influence how young adults recognize, interpret, tolerate, or reject harmful relationship behaviors.

Dating violence during adolescence and young adulthood has been associated with a wide range of adverse outcomes. Victimization may contribute to depression, anxiety, posttraumatic stress symptoms, reduced self-esteem, suicidal ideation, substance use, impaired academic functioning, and deterioration in social and interpersonal relationships [4, 5, 12]. Psychological abuse, including persistent criticism, humiliation, threats, and isolation, may cause substantial harm even in the absence of physical injury [3, 13]. Sexual coercion may be particularly traumatic and may increase the likelihood of posttraumatic stress and other long-term psychological difficulties [14]. Dating violence experiences may also affect trust, attachment, conflict-resolution patterns, and expectations regarding future romantic relationships [15]. Furthermore, attitudes that tolerate or normalize violence may increase vulnerability to subsequent victimization or perpetration and may reduce the likelihood that harmful behaviors are recognized or disclosed [5, 16]. Nevertheless, attitudes toward dating violence should be distinguished from actual experiences of victimization or perpetration, because attitudinal measures do not directly assess behavior.

Gender role attitudes provide an important theoretical framework for understanding how attitudes toward dating violence may be shaped. Gender roles refer to socially constructed expectations concerning the behaviors, responsibilities, authority, and relational positions considered appropriate for women and men. Traditional gender beliefs may portray men as dominant, controlling, decisive, or sexually assertive and women as passive, dependent, compliant, or responsible for preserving the relationship [17]. Such beliefs may encourage controlling behavior to be interpreted as protection, jealousy as an indication of affection, or compliance as an expected component of femininity. They may also reduce recognition of psychological, economic, or sexual coercion as violence.

Gender-related socialization may therefore influence not only the occurrence of violent behavior but also how individuals evaluate its acceptability and seriousness. Rigid or patriarchal gender beliefs, hostile or benevolent sexism, and norms legitimizing male authority have been associated with greater tolerance or justification of dating and intimate partner violence [9, 18–21]. Conversely, more egalitarian gender role attitudes have generally been associated with lower acceptance of violent behavior within romantic relationships [7, 22]. However, these relationships are unlikely to be uniform. Gender norms may relate differently to physical, emotional, economic, sexual, and digitally mediated forms of violence, and particular behaviors may carry different meanings across gender and sociocultural contexts.

Previous studies have reported gender differences in both dating violence attitudes and gender role attitudes, but the findings have not been entirely consistent. Women have frequently been found to hold more egalitarian gender role attitudes and to report lower acceptance of dating violence than men [7, 22]. Nevertheless, recent research suggests that gender differences may vary according to the form of violence being assessed, the measurement instrument, and the cultural and relational context [8–10, 21]. Some studies have identified gender-specific differences in physical or sexual violence, whereas others have found smaller or inconsistent differences in psychological, emotional, relational, or economic violence. These findings suggest that analyses limited to scale totals may conceal meaningful domain-specific patterns.

Research conducted in Türkiye has demonstrated an association between gender role attitudes and attitudes toward dating violence among university students [7, 18, 22]. However, the available evidence remains limited in several respects. Many studies have focused primarily on total scale scores, have examined a restricted range of violence-related dimensions, or have interpreted statistically significant gender differences without considering their effect magnitude. In addition, the relationship between specific dimensions of dating violence attitudes and individual dimensions of gender role attitudes has received comparatively limited attention. Given that egalitarian, marital, female, male, and traditional gender role beliefs represent conceptually distinct domains, their relationships with physical, emotional, economic, sexual, and general violence attitudes may not be equivalent. Therefore, even single-center studies may contribute meaningful evidence when they include large student samples, examine both total and subscale scores, and interpret statistically significant findings together with effect-size estimates.

Examining these relationships is scientifically relevant because it may clarify whether the association between gender role beliefs and dating violence attitudes is broad and uniform or concentrated in particular dimensions. It is also socially relevant because attitudes influence whether problematic behaviors are recognized, challenged, disclosed, or normalized within young adults’ relationships. A more differentiated understanding may help future research identify which beliefs and forms of violence require closer assessment and may provide an empirical basis for the subsequent development and evaluation of culturally appropriate university-based prevention strategies. In this respect, university settings are important public health contexts because they provide opportunities for relationship education, awareness programs, and violence prevention initiatives targeting young adults. Nevertheless, because attitudinal associations do not establish behavioral prediction or causality, such findings should be interpreted as an initial basis for prospective and experimental research.

The primary aim of this study was to examine the association between attitudes toward dating violence and gender role attitudes among university students. Secondary aims were to compare dating violence attitude scores and gender role attitude scores by gender and to examine subscale-level associations between the two constructs.

Methods

Study design

This study was designed as a cross-sectional, questionnaire-based observational study examining university students’ attitudes toward dating violence and gender roles. The manuscript was prepared and reported in accordance with the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) Statement. The completed STROBE checklist is provided as Supplementary Material 1.

Setting and participants

The study was conducted at Karabük University, a state university in Türkiye. The target population comprised undergraduate students enrolled at the university during the predefined data collection period. Data collection started immediately after ethical approval was obtained and continued for two months, from 30 September 2025 to 30 November 2025. At the time of data collection, Karabük University had approximately 52,000 students.

Students were eligible to participate if they: (1) were registered as undergraduate students at Karabük University during the study period; (2) were able to understand and complete the questionnaires; and (3) voluntarily agreed to participate by providing electronic informed consent.

Students were excluded if they: (1) were unable to understand or complete the online questionnaires; (2) declined to provide informed consent; or (3) submitted questionnaires with incomplete responses for the variables required for analysis.

Participation was voluntary, and no financial or academic incentive was provided.

Data collection procedure

Data were collected using an anonymous online self-report questionnaire created through Google Forms. The survey link was distributed to students through university-wide institutional communication channels during the predefined data collection period.

Before accessing the questionnaire, potential participants were provided with information regarding the study aims and procedures, the voluntary nature of participation, the confidentiality of their responses, and their right to discontinue participation at any time without penalty. Only students who provided electronic informed consent were permitted to proceed to the questionnaire.

A total of 2,112 students accessed the online form. Of these, 55 students did not provide electronic informed consent and were therefore not allowed to proceed to the questionnaire. Among the remaining 2,057 students who provided consent, 191 forms were excluded because of incomplete responses in variables required for analysis. The final analytical sample therefore consisted of 1,866 students with complete questionnaire data. Because the survey link was distributed through WhatsApp and student communication networks, the exact number of students who received or viewed the invitation could not be determined. Therefore, the response rate could not be calculated.

The questionnaire package included a sociodemographic data form prepared by the researchers, the Dating Violence Attitude Scale, and the Gender Roles Attitude Scale. Participants completed the questionnaires without direct interaction with the research team. No names, student identification numbers, Internet Protocol addresses, or other directly identifying information were collected. No separate instrument assessing socially desirable responding or impression management was administered.

Study variables and instruments

The primary study variables were the total and subscale scores of the Dating Violence Attitude Scale and the Gender Roles Attitude Scale. The study assessed self-reported attitudes only and did not measure dating violence victimization, perpetration, or actual behaviors within romantic relationships. Age and gender were included as sociodemographic variables. Gender was analyzed according to the categories recorded in the original study questionnaire.

Dating Violence Attitude Scale

The Dating Violence Attitude Scale was developed by Terzioğlu et al. in 2016 to assess young people’s attitudes toward violence in dating relationships [23]. The scale consists of 28 items rated on a 5-point Likert scale ranging from 1 (“strongly disagree”) to 5 (“strongly agree”). The possible total score ranges from 28 to 140. The general violence, economic violence, and physical violence subscales each range from 5 to 25; the emotional violence subscale ranges from 6 to 30; and the sexual violence subscale ranges from 7 to 35. Of the 28 items, 23 are reverse coded. Higher total and subscale scores indicate lower approval or acceptance of dating violence. The scale does not have an established diagnostic or categorical cut-off point. Therefore, expressions regarding lower or higher approval of dating violence were interpreted descriptively according to the observed score distributions and were not used to classify participants into predefined categories.

In the original validation study, the Cronbach’s alpha coefficient was 0.91 for the total scale and ranged from 0.72 to 0.85 across the subscales [23]. In the present study, Cronbach’s alpha was 0.91 for the total scale. Psychometric assessment in the current sample was limited to internal consistency; factor structure, measurement invariance, and other forms of validity were not independently reassessed.

Gender Roles Attitude Scale

The Gender Roles Attitude Scale was developed by Zeyneloğlu to assess university students’ attitudes toward gender roles [24, 25]. The scale consists of 38 items rated on a 5-point Likert scale ranging from 1 (“strongly disagree”) to 5 (“strongly agree”).

It includes five subscales: egalitarian gender role, female gender role, marital gender role, traditional gender role, and male gender role. The male gender role subscale contains 6 items, whereas each of the remaining subscales contains 8 items. The total score ranges from 38 to 190. Subscale scores range from 6 to 30 for the male gender role subscale and from 8 to 40 for the other subscales. All items were scored according to the original scoring instructions of the scale, including reverse coding where required. After scoring, higher total and subscale scores indicate more egalitarian gender role attitudes, whereas lower scores indicate more traditional gender role attitudes. Therefore, higher scores on the traditional gender role subscale indicate lower endorsement of traditional gender role beliefs, not greater support for traditional roles. The scoring and reverse-coding procedures were rechecked before analysis. The Gender Roles Attitude Scale does not have an established diagnostic or categorical cut-off point; therefore, score interpretations were made descriptively and comparatively rather than by assigning participants to predefined attitude categories.

The original psychometric evaluation reported a Cronbach’s alpha coefficient of 0.92 for the total scale [25]. In the present study, Cronbach’s alpha was 0.75. Psychometric assessment in the current sample was limited to internal consistency; factor structure, measurement invariance, and other validity properties were not independently reassessed.

Sample size

The minimum required sample size was estimated a priori using G*Power version 3.1.9.7 (Heinrich-Heine-Universität Düsseldorf, Düsseldorf, Germany) [26]. The calculation was based on a previous study examining the relationship between gender role attitudes and attitudes toward dating violence among university students [22].

Assuming a medium correlation effect size of r = 0.30, a two-sided alpha level of 0.05, and statistical power of 80%, the minimum required sample size was calculated as 84 participants. The survey remained available throughout the predefined data collection period, and recruitment was not discontinued when the minimum required sample size was reached. All eligible respondents with complete data were included in the analysis. Because the final sample size substantially exceeded the minimum required sample size, emphasis was placed on effect sizes and confidence intervals rather than statistical significance alone.

Statistical analysis

Descriptive statistics were calculated for all variables. The distribution of continuous variables was assessed using the Kolmogorov–Smirnov test together with visual inspection of histograms and Q–Q plots and evaluation of descriptive distributional characteristics. Because the scale total and subscale scores deviated from normality and some subscales showed marked clustering or ceiling effects, these variables are presented as median and interquartile range (IQR) and were compared between male and female students using the Mann–Whitney U test. Age showed an approximately normal distribution and is therefore presented as mean ± standard deviation (SD) and was compared between groups using Student’s independent-samples t-test. For the independent-samples t-test, the t value, degrees of freedom, p-value, and the number of participants included in the analysis were reported.

For comparisons performed using the Mann–Whitney U test, Mann–Whitney U and standardized Z values were reported together with p-values and effect sizes. Effect sizes were calculated from the standardized test statistic using the formula r = |Z|/√N, where Z represents the standardized Mann–Whitney U test statistic and N represents the total number of participants included in the comparison. Effect sizes are reported with 95% confidence intervals. Values of approximately 0.10, 0.30, and 0.50 were interpreted as small, moderate, and large effects, respectively. Given the large sample size, p-values were interpreted together with effect-size estimates to distinguish statistically detectable differences from differences of potential practical relevance.

Relationships among the Dating Violence Attitude Scale, Gender Roles Attitude Scale, their respective subscales, and age were examined using Spearman’s rank correlation coefficient because the scale-score distributions were non-normal and the relationships were evaluated using ranked observations. Spearman correlation coefficients were reported with p-values, 95% confidence intervals, and the number of participants included in each analysis. Correlation coefficients of approximately 0.10, 0.30, and 0.50 were interpreted as small, moderate, and large associations, respectively.

To account for multiple comparisons, Bonferroni-adjusted significance thresholds were applied separately to each prespecified family of analyses. A threshold of p < 0.0083 was used for the six gender comparisons involving the Dating Violence Attitude Scale total and subscale scores and separately for the six comparisons involving the Gender Roles Attitude Scale total and subscale scores. A threshold of p < 0.0167 was used for the three principal correlation analyses among the two total scale scores and age. A threshold of p < 0.002 was used for the 25 correlations between the Dating Violence Attitude Scale and Gender Roles Attitude Scale subscales. All statistical tests were two-sided. After recalculation, statistical significance decisions were checked according to the relevant Bonferroni-adjusted thresholds, and consistency among the Abstract, Results section, tables, and supplementary materials was reviewed.

Statistical analyses were conducted using IBM SPSS Statistics for Mac, version 26.0 (IBM Corp., Armonk, NY, USA).

Ethical considerations

Ethical approval was obtained from the Non-Interventional Clinical Research Ethics Committee of Karabük University on 30 September 2025 (decision number: 2025/2468). The study was conducted in accordance with the principles of the Declaration of Helsinki and applicable institutional requirements.

Participation was entirely voluntary. All potential participants received electronic information regarding the purpose and procedures of the study, the voluntary nature of participation, their right to withdraw at any time without penalty, and the measures used to protect confidentiality. Electronic informed consent was obtained before participants accessed the study questionnaires.

The survey was anonymous, and no personally identifying information was collected. Participant confidentiality was maintained throughout data collection, analysis, and reporting.

Results

A total of 1,866 students with complete questionnaire data were included in the analysis, comprising 886 female and 980 male students. Participants were aged 18–24 years. Mean age was slightly lower among female students than male students (20.31 ± 2.16 vs. 20.54 ± 2.53 years; t = − 2.12, df = 1864, p = 0.034). However, the magnitude of this difference was very small.

The comparison of Dating Violence Attitude Scale total and subscale scores is presented in Table 1, including possible score ranges, Mann–Whitney U values, standardized Z values, p-values, effect sizes, and 95% confidence intervals. After application of the Bonferroni-adjusted significance threshold of p < 0.0083, only the total scale score differed significantly between male and female students. Male students had higher total scores, indicating lower reported acceptance of dating violence; however, the effect size was very small (r = 0.06, 95% CI 0.01–0.11).

Table 1.

Comparison of Dating Violence Attitude Scale and subscale scores between male and female students

Variable	Possible score range	Male (n = 980)	Female (n = 886)	U	Z	p-value	Effect size (r)	95% CI
Dating Violence Attitude Scale	28–140	57.0 (50.5–64.0)	52.5 (48.0–59.0)	402,200	−2.75	0.006*	0.06	0.01 to 0.11
General Violence	5–25	9.0 (9.0–10.0)	9.0 (9.0–9.0)	411,600	−1.94	0.053	0.04	−0.01 to 0.09
Physical Violence	5–25	7.0 (5.0–9.0)	6.0 (5.0–9.0)	403,800	−2.61	0.009	0.06	0.01 to 0.11
Emotional Violence	6–30	11.0 (8.0–15.0)	10.0 (8.0–14.0)	416,900	−1.48	0.140	0.03	−0.02 to 0.08
Economic Violence	5–25	10.0 (8.0–13.0)	9.0 (8.0–11.0)	404,200	−2.58	0.010	0.06	0.01 to 0.11
Sexual Violence	7–35	19.0 (16.0–19.0)	19.0 (19.0–19.0)	425,400	−0.75	0.452	0.02	−0.03 to 0.07

Values are presented as median (interquartile range) because the scale-score distributions deviated from normality and some subscales showed marked clustering or ceiling effects. Comparisons were performed using the Mann–Whitney U test. Mann–Whitney U and standardized Z values are reported for each comparison. The total number of participants included in each comparison was 1,866. The possible total score range of the Dating Violence Attitude Scale is 28–140. The possible subscale score ranges are 5–25 for general violence, physical violence, and economic violence; 6–30 for emotional violence; and 7–35 for sexual violence. Higher total and subscale scores indicate lower approval or acceptance of dating violence. A Bonferroni-adjusted significance threshold of p < 0.0083 was applied for six comparisons

*Statistically significant after Bonferroni correction. Effect sizes were calculated as r = |Z|/√N, where N represents the total number of participants included in the comparison. Effect sizes are presented with 95% confidence intervals. Values of approximately 0.10, 0.30, and 0.50 indicate small, moderate, and large effects, respectively

The nominal differences in the physical and economic violence subscales did not remain statistically significant after correction for multiple comparisons. No significant gender differences were observed in the general, emotional, or sexual violence subscales. Effect sizes across the subscale comparisons were negligible or very small.

The comparison of Gender Roles Attitude Scale total and subscale scores is presented in Table 2, including Mann–Whitney U and standardized Z values. All six comparisons remained statistically significant after Bonferroni correction. Male students had higher total Gender Roles Attitude Scale scores, whereas female students had higher egalitarian gender role scores. Because higher scores indicate more egalitarian attitudes after scoring according to the original scale instructions, higher scores on the female, marital, traditional, and male gender role subscales should be interpreted as lower endorsement of traditional or unequal gender role beliefs within those domains. Thus, the higher traditional gender role subscale score among male students does not indicate greater support for traditional roles; rather, it reflects lower endorsement of traditional gender role statements after reverse coding. The largest gender difference was observed for the traditional gender role subscale (r = 0.60), followed by the egalitarian gender role subscale (r = 0.55). The total scale and marital gender role differences were moderate to large, the female gender role difference was moderate, and the male gender role difference was small.

Table 2.

Comparison of Gender Roles Attitude Scale and subscale scores between male and female students

Variable	Possible score range	Male (n = 980)	Female (n = 886)	U	Z	p-value	Effect size (r)	95% CI
Gender Roles Attitude Scale	38–190	112.0 (101.0–122.0)	98.0 (91.0–105.0)	198,200	−20.30	< 0.001*	0.47	0.43 to 0.50
Egalitarian Gender Role	8–40	35.0 (32.0–37.0)	39.0 (37.0–40.0)	158,000	−23.76	< 0.001*	0.55	0.52 to 0.58
Female Gender Role	8–40	24.5 (21.0–27.0)	22.0 (20.0–24.0)	278,500	−13.39	< 0.001*	0.31	0.27 to 0.35
Marital Gender Role	8–40	15.0 (12.0–18.0)	12.0 (12.0–13.0)	218,200	−18.57	< 0.001*	0.43	0.39 to 0.47
Traditional Gender Role	8–40	24.0 (20.75–29.0)	15.0 (12.0–20.0)	132,900	−25.92	< 0.001*	0.60	0.57 to 0.63
Male Gender Role	6–30	11.0 (8.0–16.0)	9.0 (6.0–12.0)	338,700	−8.21	< 0.001*	0.19	0.15 to 0.23

Values are presented as median (interquartile range) because the scale-score distributions deviated from normality. Comparisons were performed using the Mann–Whitney U test. Mann–Whitney U and standardized Z values are reported for each comparison. The total number of participants included in each comparison was 1,866. The possible total score range of the Gender Roles Attitude Scale is 38–190. The possible subscale score ranges are 8–40 for the egalitarian, female, marital, and traditional gender role subscales and 6–30 for the male gender role subscale. After scoring according to the original scale instructions, higher total and subscale scores indicate more egalitarian gender role attitudes. A Bonferroni-adjusted significance threshold of p < 0.0083 was applied for six comparisons

*Statistically significant after Bonferroni correction. Effect sizes were calculated as r = |Z|/√N, where N represents the total number of participants included in the comparison. Effect sizes are presented with 95% confidence intervals. Values of approximately 0.10, 0.30, and 0.50 indicate small, moderate, and large effects, respectively

Dating Violence Attitude Scale and Gender Roles Attitude Scale total scores showed a statistically significant moderate positive correlation (r = 0.35, 95% CI 0.31 to 0.39, p < 0.001), as illustrated in Fig. 1. Because higher scores on both scales indicate lower acceptance of dating violence and more egalitarian gender role attitudes, respectively, this finding indicates that more egalitarian gender role attitudes were associated with lower reported acceptance of dating violence.

Fig. 1.

Scatter plot showing the correlation between Gender Roles Attitude Scale and Dating Violence Attitude Scale scores

Age was weakly correlated with Dating Violence Attitude Scale scores (r = − 0.05, 95% CI − 0.10 to − 0.01, p = 0.031) and Gender Roles Attitude Scale scores (r = 0.10, 95% CI 0.05 to 0.14, p < 0.001). However, after applying the Bonferroni-adjusted significance threshold of p < 0.0167, only the correlation between age and Gender Roles Attitude Scale scores remained statistically significant. Both correlations were very small in magnitude and should be interpreted cautiously.

The correlations between the Dating Violence Attitude Scale and Gender Roles Attitude Scale subscales are presented in Table 3, including Spearman’s rank correlation coefficients, 95% confidence intervals, p-values, and the number of participants included in the analyses. All 25 cross-subscale correlations remained statistically significant after application of the Bonferroni-adjusted threshold of p < 0.002. The associations were positive and ranged from small to moderate in magnitude (r = 0.16–0.39).

Table 3.

Correlations between Dating Violence Attitude Scale and Gender Roles Attitude Scale subscales

Dating Violence Attitude Scale subscale	Egalitarian gender role	Female gender role	Marital gender role	Traditional gender role	Male gender role
General violence	r = 0.24, 95% CI 0.20 to 0.28, p < 0.001	r = 0.20, 95% CI 0.16 to 0.24, p < 0.001	r = 0.18, 95% CI 0.14 to 0.22, p < 0.001	r = 0.27, 95% CI 0.23 to 0.31, p < 0.001	r = 0.16, 95% CI 0.12 to 0.20, p < 0.001
Physical violence	r = 0.29, 95% CI 0.25 to 0.33, p < 0.001	r = 0.25, 95% CI 0.21 to 0.29, p < 0.001	r = 0.22, 95% CI 0.18 to 0.26, p < 0.001	r = 0.34, 95% CI 0.30 to 0.38, p < 0.001	r = 0.21, 95% CI 0.17 to 0.25, p < 0.001
Emotional violence	r = 0.31, 95% CI 0.27 to 0.35, p < 0.001	r = 0.27, 95% CI 0.23 to 0.31, p < 0.001	r = 0.25, 95% CI 0.21 to 0.29, p < 0.001	r = 0.36, 95% CI 0.32 to 0.40, p < 0.001	r = 0.23, 95% CI 0.19 to 0.27, p < 0.001
Economic violence	r = 0.26, 95% CI 0.22 to 0.30, p < 0.001	r = 0.24, 95% CI 0.20 to 0.28, p < 0.001	r = 0.21, 95% CI 0.17 to 0.25, p < 0.001	r = 0.32, 95% CI 0.28 to 0.36, p < 0.001	r = 0.20, 95% CI 0.16 to 0.24, p < 0.001
Sexual violence	r = 0.35, 95% CI 0.31 to 0.39, p < 0.001	r = 0.29, 95% CI 0.25 to 0.33, p < 0.001	r = 0.28, 95% CI 0.24 to 0.32, p < 0.001	r = 0.39, 95% CI 0.35 to 0.43, p < 0.001	r = 0.25, 95% CI 0.21 to 0.29, p < 0.001

Spearman’s rank correlation coefficients are presented as r values with 95% confidence intervals. The number of participants included in each analysis was 1,866. A Bonferroni-adjusted significance threshold of p < 0.002 was applied for 25 comparisons. All correlations remained statistically significant after Bonferroni correction.

The strongest association was observed between sexual violence and traditional gender role scores (r = 0.39), followed by emotional violence and traditional gender role scores (r = 0.36) and sexual violence and egalitarian gender role scores (r = 0.35). The weakest association was between general violence and male gender role scores (r = 0.16).

Discussion

In this cross-sectional study of 1,866 university students, attitudes toward dating violence were generally non-supportive and were significantly associated with gender role attitudes. The principal finding was a moderate positive correlation between the total Dating Violence Attitude Scale and Gender Roles Attitude Scale scores, indicating that more egalitarian gender role attitudes were associated with lower reported acceptance of dating violence. The subscale-level correlations were consistently positive but varied from small to moderate in magnitude, demonstrating that the relationship was present across multiple dimensions but was not uniform. Gender differences followed a more complex pattern. Although male students had higher total Dating Violence Attitude Scale scores than female students, this difference was very small and was the only comparison on this scale that remained significant after correction for multiple testing. In contrast, gender differences across the Gender Roles Attitude Scale were more pronounced and ranged from small to large in magnitude.

The association between more egalitarian gender role attitudes and lower acceptance of dating violence is consistent with previous research conducted in Türkiye and other sociocultural settings. Kök et al. reported that university students with more egalitarian gender role attitudes were less supportive of dating violence [22]. Similarly, Yıldırım-Hamurcu and Terzioğlu found that egalitarian gender role attitudes were associated with lower support for dating violence among Turkish university students [7]. Research examining ambivalent sexism and traditional gender norms has likewise suggested that beliefs legitimizing male authority, female compliance, and rigid relationship roles may increase the tolerance or normalization of violent behavior [18–20]. The present findings extend this evidence by showing that the relationship was not limited to the total scores of the two instruments but was also observed across the specific subdimensions of dating violence and gender role attitudes.

The subscale-level findings indicate that the association between dating violence attitudes and gender role attitudes was not uniform across all dimensions. The associations were statistically significant but small to moderate in magnitude. This suggests that total scores may conceal variation in the relationships between particular forms of dating violence attitudes and specific dimensions of gender role attitudes. Examination of both aggregate and domain-specific scores therefore provides a more differentiated description of the relationship between the constructs. Nevertheless, these correlations should be interpreted as descriptive associations rather than evidence of strong relationships, causal mechanisms, temporal pathways, or modifiable intervention targets.

Our findings regarding gender differences partly converge with and partly diverge from previous studies. Several Turkish studies have reported that female students hold more egalitarian gender role attitudes and are less accepting of dating violence than male students [7, 22]. In the present study, male students had higher total Dating Violence Attitude Scale scores, but the effect size was very small. Moreover, the nominal differences in physical and economic violence attitudes did not remain statistically significant after correction for multiple comparisons. These results should therefore not be interpreted as evidence that male students had meaningfully lower acceptance of physical or economic dating violence. The substantial overlap between male and female score distributions indicates that gender alone provided limited differentiation in overall dating violence attitudes.

Recent findings from Türkiye similarly suggest that male students do not invariably report greater acceptance of dating violence. Şahin et al. found that male participants did not support violence across the examined dimensions after covariates were considered, although some small group differences were identified [8]. Research examining dating violence and ambivalent sexism has also suggested that sexist beliefs and traditional gender norms may be more closely associated with dating violence attitudes than gender alone [18]. In the present study, the association between gender role attitudes and dating violence attitudes was more pronounced than the gender difference in the overall Dating Violence Attitude Scale score. This finding suggests that individual beliefs concerning gender roles may provide information beyond a binary comparison between male and female students. However, it does not demonstrate that gender role beliefs are the primary determinants of dating violence attitudes or that modifying these beliefs would necessarily change attitudes or behavior.

The reasons for the differences between the present findings and those of previous studies cannot be established from the available data. Samples may have differed in geographic region, institutional setting, academic discipline, sociodemographic composition, relationship experience, exposure to gender-equality education, and the period in which data were collected. Broader sociocultural changes may also have influenced how contemporary university students understand and report dating violence and gender roles. However, these characteristics were not systematically compared across studies, and several potentially relevant variables were not measured in the present research. Explanations based on cohort-specific socialization or greater internalization of egalitarian norms should therefore be regarded as hypotheses rather than conclusions supported directly by the current data.

The international literature likewise emphasizes the relevance of gender ideology to dating violence. Reyes et al. reported that traditional gender role attitudes prospectively predicted physical dating violence perpetration among adolescent boys [19]. Other studies have shown that adolescents and young adults who endorse rigid or patriarchal norms are more likely to justify, tolerate, or engage in dating violence, whereas individuals with more egalitarian beliefs are less likely to do so [20]. The present study, however, assessed attitudes toward dating violence rather than actual victimization, perpetration, or relationship behaviors. Therefore, the present findings are compatible with this broader literature but cannot independently confirm a predictive or behavioral relationship. The moderate total-score correlation and the small-to-moderate subscale correlations indicate that gender role attitudes should be considered one correlate among multiple individual, relational, family, peer, and sociocultural factors that may influence attitudes toward dating violence.

The gender pattern observed in Gender Roles Attitude Scale scores may reflect the multidimensional nature of gender-related beliefs and the scoring structure of the scale. Türkiye is undergoing sociocultural change in which egalitarian and traditional gender expectations may coexist. Young adults may endorse equality in some domains while maintaining more traditional expectations in others [27]. In the present study, female students had higher scores on the explicitly egalitarian gender role subscale, whereas male students had higher scores on the total scale and on the female, marital, traditional, and male gender role subscales. This pattern may appear inconsistent with the broader literature if the subscale names are interpreted literally. However, after reverse coding according to the original scale instructions, higher scores on all Gender Roles Attitude Scale subscales indicate more egalitarian attitudes. Accordingly, a higher score on the traditional gender role subscale indicates lower endorsement of traditional gender role beliefs, not stronger support for traditional roles.

The finding that male students had a higher total Gender Roles Attitude Scale score should therefore be interpreted cautiously. It does not necessarily mean that male students were uniformly more egalitarian across all aspects of gender ideology, because female students scored higher on the explicitly egalitarian gender role subscale. Rather, the total score reflects the combined contribution of several subscales with different item contents. This result may reflect scale structure, domain-specific response patterns, or sample-specific characteristics. Because measurement invariance across gender was not tested, it also remains possible that male and female students interpreted some items differently. Therefore, this finding should be regarded as a descriptive result specific to the present sample and measurement instrument, rather than as definitive evidence contradicting previous literature.

No robust gender differences were observed across the individual dating violence attitude subscales after correction for multiple comparisons. Emotional and sexual violence scores did not differ significantly before correction, whereas the nominal physical and economic violence differences did not remain significant after Bonferroni adjustment. Previous research has suggested that psychological and sexual forms of dating violence may be normalized, minimized, or not readily recognized as violence by both men and women. Temple et al. reported that adolescents who regarded psychological abuse as acceptable were more likely to experience mental health difficulties and remain in violent relationships, despite frequently failing to recognize such behaviors as abusive [3]. Exner-Cortens et al. also documented adverse long-term outcomes associated with dating violence victimization, including psychological forms of violence [12]. Nevertheless, because the present study did not measure dating violence experiences, the absence of statistically robust gender differences does not demonstrate that male and female students interpret, experience, or behave in relation to these behaviors identically.

The selected instruments may not have captured all such nuanced differences. The Dating Violence Attitude Scale assesses broad domains of general, physical, emotional, economic, and sexual violence but may not fully represent contemporary or less visible forms of gendered control, including digital monitoring, coercive control, jealousy-based restriction, manipulation, social isolation, and context-dependent interpretations of relationship behavior. Similarly, predefined total and subscale scores may conceal item-level differences in how students interpret specific behaviors. Alternative instruments using different item content, response formats, or conceptual definitions might therefore yield larger, smaller, or qualitatively different gender differences. Measures specifically assessing digital dating abuse, coercive control, ambivalent sexism, or gender norm endorsement may be more sensitive to subtle differences not fully represented by the instruments used in this study. The findings should consequently be interpreted as specific to the operational definitions and item content of the selected scales rather than as definitive evidence concerning all dimensions of gender differences in dating violence attitudes.

The subscale-level analyses provided a more differentiated account of the relationship between dating violence attitudes and gender role attitudes. After correction for multiple comparisons, the statistically significant associations ranged from small to moderate in magnitude, and their relative strength differed across the dimensions examined. This pattern suggests that particular forms of dating violence may be associated differently with particular aspects of gender role ideology and that analyses based solely on total scores may obscure domain-specific relationships. However, none of the coefficients represented a very strong relationship. Gender role attitudes should therefore be viewed as a potentially relevant but incomplete correlate of dating violence attitudes rather than as a sufficient explanation of them.

Although the association between age and Gender Roles Attitude Scale scores reached statistical significance after Bonferroni correction, the magnitude of the correlation was negligible and unlikely to be practically meaningful. The correlation between age and Dating Violence Attitude Scale scores was also very small and did not remain statistically significant after correction. These findings suggest that chronological age had little practical relevance to the observed attitude scores in this sample, possibly because the participants were within a narrow age range of 18–24 years.

The large sample substantially increased statistical power and therefore made it possible to identify small differences as statistically significant. Accordingly, p-values were interpreted together with effect size estimates and corrections for multiple comparisons. For the Dating Violence Attitude Scale, only the total score difference remained statistically significant after Bonferroni correction, and its corresponding effect size was very small. By contrast, all Gender Roles Attitude Scale comparisons remained significant after correction, with effect sizes ranging from small to large. These findings illustrate why statistical significance alone should not be regarded as evidence of substantive or practical importance. The gender-related findings should therefore be understood as differences in score distributions rather than evidence that every statistically significant comparison represents a large or consequential distinction between male and female students.

The present study extends the literature in several respects, and its contribution is not limited to its large sample size. First, it provides contemporary evidence from a Turkish university population within a sociocultural context in which egalitarian and traditional gender norms may coexist. Although the study was conducted at a single state university, the large sample provides a detailed description of attitudes toward dating violence and gender role attitudes in a contemporary university setting in Türkiye, where the available evidence remains relatively limited. Second, attitudes toward dating violence and gender role attitudes were examined at both the total-score and subscale levels, allowing patterns to be identified that might otherwise have been obscured by analyses based solely on aggregate scores. Third, effect size estimates and corrections for multiple comparisons enabled statistically detectable findings to be distinguished from differences of greater substantive magnitude. In particular, the finding that the gender difference in overall attitudes toward dating violence was statistically significant but very small adds nuance to research that has primarily focused on whether a gender difference was present. The results further suggest that attitudes toward dating violence among university students may be associated more closely with gender-related beliefs than with gender category alone. Gender role attitudes may be associated with whether controlling, coercive, or violent behaviors are recognized, minimized, tolerated, or normalized within romantic relationships. Taken together, these findings indicate that the relationships among gender, gender role attitudes, and attitudes toward dating violence are multidimensional and context-dependent.

The findings may help define directions for future research but do not provide direct evidence for the effectiveness of specific prevention practices. The association between gender role attitudes and dating violence attitudes identifies a relationship that warrants further investigation. From a preventive perspective, future university-based research may benefit from addressing not only overt physical violence but also less visible forms of emotional abuse, economic control, sexual coercion, jealousy-based restriction, and digitally mediated monitoring. The findings may therefore be relevant for informing the content of future university-based public health and violence prevention studies, particularly by highlighting the importance of gender role attitudes and domain-specific forms of dating violence. However, because the present study assessed attitudes rather than behavior and used a cross-sectional design, it cannot establish that gender role attitudes predict or cause victimization, perpetration, or changes in relationship behavior. Prospective longitudinal studies are needed to establish the temporal ordering of these constructs and to determine whether gender role attitudes predict subsequent changes in dating violence attitudes after accounting for baseline levels of both constructs and relevant personal, relational, and sociocultural covariates. Repeated assessments could also clarify the stability of these attitudes and whether changes in one construct precede changes in the other. Longitudinal studies that additionally assess later victimization and perpetration would help determine whether the attitudinal constructs predict subsequent behavioral outcomes.

Although prospective studies can strengthen evidence regarding temporal prediction, they cannot by themselves establish causality. Experimental and quasi-experimental research is therefore also required. Randomized controlled studies could test whether interventions specifically designed to modify gender-related beliefs produce changes in dating violence attitudes compared with appropriate control conditions. Such studies should include pre-intervention and follow-up assessments, evaluate whether changes are sustained over time, and examine whether attitudinal changes translate into changes in relevant behavioral outcomes. The distinct patterns observed across the subscales further suggest that future experimental research should evaluate physical, emotional, economic, sexual, and digitally mediated forms of violence separately. Until prospective and experimental evidence is available, the current findings should be considered an initial correlational basis for further predictive and causal investigation rather than direct justification for particular campus-based programs.

Several limitations should be considered. First, the cross-sectional design prevents conclusions regarding temporal precedence, prediction, or causality. It cannot be determined whether gender role attitudes precede dating violence attitudes, whether the reverse relationship occurs, or whether both are influenced by shared antecedents. Prospective longitudinal studies with repeated assessments are needed to examine predictive relationships and potential bidirectional effects. Experimental studies are required to determine whether deliberately changing one construct causes subsequent change in the other.

Second, the study relied on online self-report measures, and no separate measure of socially desirable responding or impression management was administered. Dating violence and gender role attitudes are socially sensitive constructs, and participants may have underreported violence-accepting or traditional attitudes or selected answers perceived as more socially acceptable. This may have resulted in an overestimation of non-accepting and egalitarian attitudes and may also have influenced gender comparisons and correlations. The anonymous online format, absence of personally identifying information, and lack of direct interaction with the research team may have reduced evaluation pressure and encouraged honest reporting. However, these safeguards cannot eliminate socially desirable responding, and its extent could not be quantified or statistically controlled. The results should therefore be interpreted as reported attitudes rather than definitive estimates of underlying beliefs or actual behavior.

Third, the sample was drawn from a single state university, which may limit generalizability to students at other institutions, young adults from different regions, or non-student populations. Although the sample was large, its size does not eliminate the possibility of institutional, regional, self-selection, or non-response effects. Thus, the single-center design should be considered a limitation in terms of external validity, even though the large sample size, subscale-level analyses, and effect-size interpretation strengthen the descriptive contribution of the study. Multicenter research involving universities with different demographic, educational, and cultural characteristics is needed.

Fourth, the instruments assessed attitudes rather than actual victimization or perpetration. Participants who reject dating violence at an attitudinal level may not necessarily behave consistently with their stated attitudes in romantic relationships. Future studies should combine attitudinal measures with behavioral assessments, victimization and perpetration histories, qualitative interviews, partner reports where ethically and practically appropriate, and indirect assessment methods.

The internal consistency of the Gender Roles Attitude Scale was lower in the present sample than in the original psychometric evaluation study (Cronbach’s α = 0.75 vs. 0.92). Although an alpha coefficient of 0.75 indicates acceptable internal consistency, the lower value may reflect heterogeneity within the sample, variation across the scale’s distinct dimensions, differences in the interpretation of gender-related statements, or sociocultural changes since the scale was developed. The online administration format may also have contributed. Reduced internal consistency may increase measurement error and attenuate associations involving the total score. Consequently, the magnitude of these associations should be interpreted cautiously.

The selected instruments also impose broader measurement-related limitations. Both scales were originally developed and psychometrically evaluated in Turkish samples, but they have not been examined as extensively across diverse populations and sociocultural settings as some more widely used international measures. Psychometric assessment in the present study was limited to internal consistency; factor structure, measurement invariance across gender, convergent validity, and criterion-related validity were not reassessed. Cronbach’s alpha alone does not establish the overall validity of an instrument. In particular, the absence of measurement-invariance testing means that it remains uncertain whether male and female students interpreted the items and underlying constructs in fully comparable ways. This issue is especially relevant for the Gender Roles Attitude Scale, because the observed pattern of higher male total scores despite higher female scores on the egalitarian gender role subscale may partly reflect the multidimensional structure of the instrument or gender-specific interpretation of particular items. Alternative measures could define, weight, and score dating violence attitudes and gender role attitudes differently and could include forms of control or abuse not represented in the selected instruments.

Despite these limitations, the large contemporary sample and the simultaneous examination of total and subscale scores are important strengths. The inclusion of effect size estimates and corrections for multiple comparisons also enabled statistical significance to be distinguished from substantive magnitude. The study contributes to the literature not only by replicating the association between more egalitarian gender role attitudes and lower acceptance of dating violence but also by demonstrating that gender differences vary across constructs and that statistically significant findings may have limited practical magnitude. The results therefore provide a more nuanced and context-sensitive descriptive account of dating violence attitudes among university students in Türkiye and establish a basis for future prospective and experimental research.

Conclusions

The observed scale scores suggested relatively low reported acceptance of dating violence, although this interpretation was descriptive because no established categorical cut-off point was used. Lower acceptance of dating violence was moderately associated with more egalitarian gender role attitudes. Male students had slightly higher total Dating Violence Attitude Scale scores than female students, but the effect was very small, and none of the individual dating violence attitude subscale differences remained statistically significant after correction for multiple comparisons.

The findings identify gender role attitudes as a correlate of attitudes toward dating violence, but they do not establish whether these attitudes predict actual victimization, perpetration, or behavioral change. The subscale-level findings suggest that the relationship between dating violence attitudes and gender role attitudes differs across specific dimensions and should be interpreted in light of the study sample, operational definitions, and measurement instruments. Future longitudinal and experimental studies are needed to clarify the direction, stability, and behavioral relevance of these associations.

Authors’ contributions

UB and BA: conceptualization, design, and methodology. UB, RT, and AA: data collection. UB and BA: formal analysis, funding acquisition, and supervision. UB, RT, and AA: investigation and project administration. UB, BA, and RT: writing the original draft. BA, RT, and AA: revising the manuscript. All authors contributed to the article and approved the submitted version.

Funding

None.

Data availability

The datasets used and analyzed during the current study are available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

Ethics committee approval was obtained from the Non-Interventional Clinical Research Ethics Committee of Karabük University on 30 September 2025 (decision number: 2025/2468). The study was conducted in accordance with the principles of the Declaration of Helsinki and applicable institutional requirements. All participants provided electronic informed consent before accessing and completing the online questionnaire.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.