Prevalence, Gender and Age Differences of Dating Aggression Among Italian Adolescents

The present study represents an effort to expand and deepen the scant literature on Adolescent Dating Aggression (ADA) within the Italian context; adolescent dating aggression is a public health issue of interest due to its increasing frequency among adolescents. The prevalence of verbal-emotional and physical ADA was examined as well as gender and age differences in a sample of Italian adolescents. Participants included 436 adolescents (47.7% males; 52.3% females) living in northern Italy, aged 16 to 18 years (M = 17.11). Participants completed the Conflict in Adolescent Dating Relationships Inventory measuring abusive behaviors between adolescent dating partners. Non-parametric analyses were computed. Verbal-emotional ADA perpetration and victimization were much more common than physical ADA perpetration and victimization. Females reported higher levels of verbal-emotional and physical ADA perpetration than males. To fully investigate gender differences single behaviors were analyzed and described. Finally, age differences emerged only for perpetrated verbal-emotional abuse with such aggression being highest at age 18. This research suggests that in order to prevent the onset of dating aggression in teens in northern Italy, prevention programs may need to begin earlier than previously provided in junior high school. Another core conclusion is that physical aggression against partners is a problem for both males and females, thus intervention for the empowerment of interpersonal skills are needed.

severe consequences by male partners who perpetrated more severe forms physical aggression (Foshee et al., 2009;Muñoz-Rivas et al., 2007a, 2007bWindle & Mrug, 2009). In particular, O'Leary and colleagues (2008) found that both male and female adolescents reported engaging in physical aggression more than experiencing such form of aggression as victims. Further, when physical aggression occurred, it was largely bidirectional (i.e., both partners were aggressive). In addition, those adolescents who were engaged to be married reported the highest rates of physical aggression. Conversely, a United Kingdom study (Hird, 2000) showed that girls reported higher rates of perpetration and victimization for both psychological and physical ADA, while in New Zealand, similar percentages of psychological, physical and sexual ADA were observed for both males and females (Jackson et al., 2000).
As for age differences, studies presented quite different findings probably due to methodological variability.
A cross-sectional study carried out in Spain with adolescents and young adults aged 16-20 years, showed very high rates of psychological ADA that was stable across the age groups. Conversely, rates of physical ADA decreased significantly across the age groups, while health consequences resulting from aggression became more severe with age (Muñoz-Rivas et al., 2007b). On the contrary, longitudinal studies examining growth trajectories of ADA observed different patterns. O' Leary and Slep (2003) observed a stable pattern for physical ADA from the 10th grade to the 12th grade, while Wolfe and colleagues (2003) found that the mean trajectories for physical and psychological ADA were characterized by a steady decrease from ages 14-16 to ages 16-18. A study by Foshee and colleagues (2009) showed that the development of physical and sexual ADA perpetration decreased from ages 13 to 19, with peak ages ranging from 16 to 17 years of age. As for psychological ADA perpetration, authors observed an increasing pattern from ages 13 to 19.
Concerning Italy, ADA is emerging as a public health issue. A survey carried out by Telefono Azzurro and DoxaKids (2014) with approximately 1500 Italian male and female adolescents attending junior high school and high school, reported that 8% were threatened by a romantic partner; 33% were victims of psychological ADA; 6% were physically beaten at least once by a romantic partner, and 6% were forced to have sex.
To date, research on ADA in Italy specifically focused on relationship's characteristics (e.g., support, power imbalance, conflict). Menesini and Nocentini (2008) found that with age relationships became gradually more supportive, but also more conflicted. In fact, in late adolescence, relationships are more characterized by expectations, emotional commitment and intimacy which may lead partners to overreact thereby increasing the intensity of the conflict (Seiffge-Krenke, 2003). considered that ADA is an emerging and increasing health problem within the Italian context, thus it seems important to expand the knowledge base on this sensitive issue. Therefore, the aims of this study are to: a) describe the prevalence of ADA in a sample of Italian adolescents; b) evaluate gender differences both in relation to each single aggressive behavior and two types of ADA (verbal-emotional and physical) perpetration and victimization. Going through each single behaviors will allow a better understanding of the nature of the phenomenon and to identify which are the most frequent behaviors perpetrated by adolescents, in order to catch the possible meanings and functions that ADA may play; c) evaluate age differences in relation to two types of ADA (verbal-emotional and physical) perpetration and victimization.
No hypothesis will be formulated for the first aim, since its descriptive in nature. As for the second aim, in line with previous findings, we expect to find higher rates of verbal-emotional and physical ADA perpetration among female than male adolescents (Foshee et al., 2009;Muñoz-Rivas et al., 2007a, 2007bWindle & Mrug, 2009) and to find higher rates of verbal-emotional and physical ADA victimization among male than female adolescents (Fernández- González, O'Leary, & Muñoz-Rivas, 2014). As for the third aim, we expect to find stable rates of psychological ADA across the age groups and decreasing rates of physical ADA across the age groups (Muñoz-Rivas et al., 2007b).

Method Participants and Procedure
Participants were 859 adolescents (36.5% males; 63.5% females) living in northern Italy (Lombardy and Piedmont) aged 16 to 18 years (M = 17.01 years, SD = 0.8 years). Within this sample 709 participants (83.3%) were adolescents who have already experienced a romantic relationship (currently or in the past). Starting from this sample, a subsample balanced for gender was randomly extracted, using SPSS (Version 18). Therefore, the final sample included 436 adolescents (47.7% males; 52.3% females) aged 16 to 18 years (M = 17.11 years, SD = 0.8 years). The average age was 17.18 years for males (SD = 0.8 years) and 17.05 years for females (SD = 0.8 years).
There are three types of high school programs students in Italy can select from: one that is more theoretically oriented (a school whose aim is to prepare students to enter university), another that is more professionally oriented (a school whose aim is to prepare students for employment after high school), one that is more technically oriented (a school whose aim is prepare students to enter technical schools). Fifty percent of the participants attended a theoretical program, 32.3% attended a professional program, and 17.7% attended a technical program.
Approval for the study was obtained from the Ethical Commission of the Università Cattolica del Sacro Cuore of Milan. The headmasters and class teachers of the involved schools provided permission for the study, and parents provided written consent for their children's participation. The teenage participants also provided their consent.

Instruments
For this study, a specific assessment self-report instrument with different formats was administered in person: a) A series of questions with various response options that assessed relevant information concerning: • Demographics, e.g., age, sex, nationality, type of school attended.
• Dating relationship variables, e.g., whether they have ever experienced a romantic relationship or not, type of relationship (current vs. past), duration of the relationship, age of the romantic partner, perceived importance of the relationship).
• Sexual behaviors variables, e.g., whether they have ever had a sexual intercourse, age of the first sexual intercourse).
b) A short version of the Conflict in Adolescent Dating Relationships Inventory (CADRI; Wolfe et al., 2001a).
The original version is a 35-item instrument to assess multiple forms of abusive behaviors that may occur between adolescent dating partners. The instrument showed fair psychometric properties for verbal-emotional (α = .81; Wolfe et al., 2001a) and physical abuse perpetration and victimization (α = .76; Wolfe et al., 2001a).
Response choices for each item were measured on a 4-point step scale (1 = never to 4 = often). An interpreter translated the items of the CADRI from English into Italian and the items were preliminary administered to 100 adolescents in a pilot study. Then, a native English speaker performed the back translation. In the present study, we employed only an adaptation of two subscales: verbal-emotional abuse perpetration and victimization (9 items; e.g., "I insulted her/him with put-downs", respectively α = .79; α = .80) and physical abuse perpetration and victimization (4 items; e.g., "I kicked, hit or punched her/him", respectively α = .73; α = .82). Items scores are summed for the score for each subscale, as suggested by the authors (Wolfe et al., 2001a).

Dating Behavior
Among participants, 61.6% reported that they were experiencing a romantic relationship at the time of the study and 38.4% had experienced a romantic relationship in the past. The mean age of the romantic partner was 17.73 (SD = 0.8 years) and no significant differences emerged between male and female participants in terms

Prevalence of Verbal-Emotional and Physical ADA Perpetration and Victimization
Firstly, at a descriptive level, frequency analysis was carried out to report the percentages of adolescents who have perpetrated or experienced at least once verbal-emotional or physical abuse. Percentages of adolescents reporting to have perpetrated at least one act (i.e., participants who scored from 2 to 4) of verbal-emotional abuse was 96.1% while physical abuse was 30.7%. Percentages of adolescents reporting to have experienced at least one act (i.e., participants who scored from 2 to 4) of verbal-emotional abuse was 97.9% while for physical abuse it was 32.1%.
Parametric assumptions were in fact verified for each item and since parametric assumptions were violated, non-parametric analyses were conducted. The Wilcoxon signed-rank test was calculated to describe the prevalence of ADA within our sample. Analyses showed that the median score of verbal-emotional abuse perpetration (Mdn = 2.00) was significantly higher than physical abuse (Mdn = 1.00; Z = -16.959; p < .001).
The same result held for the mean score of verbal-emotional abuse victimization (Mdn = 2.00), which was significantly higher than physical abuse (Mdn = 1.00; Z = -17.440; p < .001).

Differences
To assess gender differences regarding each single ADA behavior, Chi Square tests were calculated on CADRI item describing Verbal Emotional Abuse and Physical Abuse perpetration and victimization. Because multiple comparisons were made, a Bonferroni procedure was used to reduce the risk of inflating the Type I error rate.
The significance of each Chi Square test was evaluated at the .005 level (.05/10).
As shown in Table 1, the most commonly perpetrated behaviors among females were making the partner feel jealous, bringing up something from the past, keeping track of the partner's movements, and accusing the partner of flirting with another person. Making fun of the partner in front of friends was more frequent among males. Dating Aggression Among Italian Adolescents 664 As reported in Table 2, percentages of physical perpetrated behaviors were generally higher among females.
In particular, female participants reported significantly higher rates of engaging in the following behaviors: throwing something at their partner and slapping their partner, or pulling their partner's hair.  Table 3 for ADA verbal-emotional victimization, female participants reported their partners made them feel jealous more frequently than male participants. The male participants reported more their partners made fun of them in front of friends more frequently than female participants. Finally, as shown in Table 4 for ADA physical victimization, male participants reported that their partner threw something at them, slapped them or pulled their hair, and pushed, shoved, or shook them more frequently than female participants. Cuccì,O'Leary,Olivari,& Confalonieri 665  To assess gender differences regarding the two types of ADA (verbal-emotional and physical), Mann-Whitney test was calculated considering the two CADRI subscales for both perpetration and victimization. Gender differences emerged for perpetrated verbal-emotional abuse (U = 18460; p < .001) with females reporting higher levels of aggression (Mdn = 2.11) than males (Mdn = 2.00). The same was for perpetrated physical abuse (U = 19910; p = .001), with females reporting higher levels of aggression (Mdn = 1.50) than males (Mdn = 1.00).

As shown in
No significant differences emerged for the subscales assessing victimization behaviors.

Verbal-Emotional and Physical ADA Perpetration and Victimization: Age Differences
Kruskal-Wallis H tests were carried out considering each of the two CADRI subscales for both perpetration and victimization as dependent variables, and age (divided in three age classes, respectively 16, 17 and 18-year-old adolescents) as the independent variable. Participants for each class were respectively: 118, 151,

167.
Analyses demonstrated significant differences only for perpetrated verbal-emotional abuse, χ 2 (2) = 8.125, p = 0.017. Pairwise post hoc comparisons indicated that 16-year-old participants reported a significantly lower mean rank score (195.36) than 18-year-old males (238.01). No significant differences emerged for perpetrated physical abuse and for verbal-emotional and physical victimization.

Discussion
Since ADA is an emerging and increasing health problem in Italy and the existing literature on ADA in the Italian context is quite scarce, this study aimed at expanding the knowledge base on this sensitive issue. Our work had three different aims.
With the first one we aimed at providing a description of the frequency of ADA in a sample of Italian adolescents at a descriptive level. Our findings confirmed the presence of ADA among Italian dating adolescents, as stated by previous research (e.g., Nocentini et al., 2010;Ortega et al., 2010). In our study, in particular, it emerged that 96% of the participants has perpetrated at least one act of verbal-emotional abuse and around 30% of the sample has engaged in physically aggressive behaviors towards their romantic partner at least Dating Aggression Among Italian Adolescents 666 once. We believe that these data are important because they suggest that adolescents involved in a romantic relationship, even if rarely, may resort to dysfunctional interaction strategies to deal with the partner.
Around 98% of the sample experienced at least one act of verbal-emotional abuse victimization and 32% of the sample reported experiencing physical abuse victimization at least once. These findings are in line with previous evidence emerged in Italy (e.g., Nocentini et al., 2010;Ortega et al., 2010) and Spain (Muñoz-Rivas et al., 2007b). These data are particularly significant when considering that more than 60% of the study participants described their relationships as important, meaning that adolescent participants consider their relationships, even if partially characterized by aggression and conflict, as significant for themselves. This may indicate that these aggressive behaviors are becoming more common within adolescent romantic relationships.
On the contrary, results are quite different from the ones emerged in the survey carried out by Telefono Azzurro and DoxaKids (2014) that showed higher percentages for both verbal-emotional and physical ADA.
This difference could be a function of using different measure use to asses ADA. In the above-mentioned survey ADA behaviors were measured using ad hoc items while in the present study we employed the CADRI inventory (Wolfe et al., 2001a). Another possible explanation could be linked to the age of participants: our sample consisted of adolescents aged 16-18 years while no information was available on the exact age range of participants responding to the previous survey conducted in Italy.
As for the prevalence of the two types of ADA perpetration and victimization, in our study it emerged that verbal-emotional abuse was much more common than physical abuse among adolescent participants. Our participants appear to both perpetrate and experience such abuse, as was observed in other studies (Foshee & Matthew, 2007;Foshee et al., 2009;Hird, 2000;Jackson et al., 2000).
The second aim of our study was to evaluate the gender differences starting from the analyses of each single aggressive behavior, which allowed us to understand the nature of the aggressive behaviors in greater detail.
Moreover, we investigated gender differences across the two types of ADA (verbal-emotional and physical) perpetration and victimization, since in Italian literature little attention has been paid on gender differences across the two types of ADA.
As for single-item analysis, regarding the perpetration of verbal-emotional aggression, girls tend to express their aggression though behaviors of jealousy and control (e.g., "I did something to make him/her feel jealous; I kept track of who he/she was with and where he/she was"), that may be an expression of feelings of personal insecurity towards the self or towards the future of the relationship. Boys, on the other hand, seem to engage more in behaviors involving the social domain (e.g., "I ridiculed or made fun of him/her in front of others"), this may be in order to establish or maintain their social status within the peer context.
Regarding the victimization of verbal-emotional ADA, in a specular way adolescents reported to be victim of the same behaviors they also reported to perpetrate more, suggesting the potential development of regular dysfunctional pattern of interaction within the couple. These behaviors may be perpetrated, indeed, as forms of revenge or imitation between the partners where each partner use the same behaviors to get even with the partner.
Regarding the perpetration of physical aggression, girls tend to engage more than their male peers in moderate aggressive actions, such as slapping and pulling the hair of the partner. However, the severity as well as the consequences of these physical injuries are less severe, as shown in previous studies where male partners Cuccì, O'Leary, Olivari, & Confalonieri 667 perpetrated more severe forms physical aggression (e.g., Cuccì et al., 2019;Foshee et al., 2009;Muñoz-Rivas et al., 2007a, 2007bWindle & Mrug, 2009). Regarding physical aggression victimization, despite the low rates of frequency, for each behavior, male participants reported to be victimized more frequently than female participants. A possible explanation could be that, within the context of a heated situation (e.g., conflict within the couple), a lack of skills in female emotion regulation may resulted in ADA as an expression of emotional discomfort rather than a real intention to physically harm the partner.
These findings, indeed, support the idea that ADA may develop as a dysfunctional strategy of interaction between the two partners.
Concerning gender differences in relation to the two types of ADA (verbal-emotional and physical) perpetration and victimization, girls engaged more frequently than boys in verbal-emotional and physical ADA perpetration.
These findings confirm previous evidence (Foshee et al., 2009;Muñoz-Rivas et al., 2007a, 2007bO'Leary et al., 2008;Windle & Mrug, 2009) showing higher rates of perpetration among female adolescents. These findings could be interpreted considering some relationship's characteristics. The length of the relationship was, indeed, significantly higher for females and that most of them described the relationship as very important.
This may imply that girls feel to be more invested in their relationships. Therefore, it is possible to hypothesize that female adolescents resort to dysfunctional behaviors to maintain and protect their romantic relationships, not being able to engage in more adaptive conflict resolution strategies. Our findings are in line with the developmental perspective that female adolescent aggression toward their partner may result from their relative inexperience in managing romantic relationships. This lack of skill would lead females to adopt more masculine aggressive behaviors to solve their conflicts with the romantic partner, as showed in previous studies (Connolly et al., 2015;O'Leary et al., 2008;Rose & Rudolph, 2006;Wolfe, Scott, & Crooks, 2005). As suggested by some feminist theorists a small but significant percentage of females may also resort to violence as an act of self-defense to contrast male predominance in power (Hettrich & O'Leary, 2007;Stuart et al., 2006;White & Kowalski, 1994). Female adolescents may also resort aggressive behaviors as an effort to restore balance when perceiving an imbalance of power in their relationships. Previous studies have demonstrated that these characteristics are associated with ADA (Connolly et al., 2010;Menesini & Nocentini, 2009). At the same time, it is possible to consider that females engage more frequently in mutual aggressive relationships. This may mean that, within the romantic couple, partners tend to use aggressive behaviors as a form of interaction with each other. Thus, it is possible that partners learn dysfunctional strategies, modeling their behaviors by observing each other (Bandura, 1977).
Finally, it is important to consider that gender differences may also be due to social desirability bias, (i.e., the tendency of people to deny or underreport behaviors that are socially unacceptable). Males may indeed find it difficult to admit they acted physically aggressive towards their female partner, while it might be easier to report ADA victimization. This may also explain why no significant differences were found for ADA victimization, differently from what we expected. As Fernández- González and colleagues (2014) reported, both males and females are less likely to report negative behaviors about themselves than about their partners. With a large sample of adolescents from Madrid (~850), social desirability had a small but significant association with reports of ADA, but covariance corrections for social desirability did not alter the conclusions about such aggression. Because of the small correlation of social desirability with perpetration and victimization for females (-.19, -.16) and males (-.10, -.11), the corrections for social desirability did little to alter any overall conclusions based on uncorrected means. Using uncorrected or corrected means for social desirability, males engage Dating Aggression Among Italian Adolescents 668 in more sexual ADA and females engage in more psychological and physical ADA. If one wants to obtain the maximal prevalence rates, Fernández- González and colleagues (2014) found that maximal dyadic reports (based on either the self or partner) significantly increased the prevalence rates of aggression. However, the overall conclusions about perpetration and victimization did not differ. If one wishes to estimate maximal rates of partner aggression but one only has self-reports, Heyman andSchlee (1997), andWilliams (2006) have used correction factors with married couples and such an approach could be used with dating couples.
Our third aim was to evaluate age differences in relation to the two types of ADA ( (Straus, Hamby, Boney-McCoy, & Sugarman, 1996), includes more severe acts of physical ADA compared with the CADRI, such as beating, choking, burning, and scalding a partner (Cascardi, Blank, & Dodani, 2019). Moreover, though developed with a similar goal, the assessment of psychological and physical abuse towards a romantic partner is different, since CADRI was developed specifically for adolescent samples (Cascardi, Blank, & Dodani, 2019;Wolfe et al., 2001a).
Our findings need to be evaluated with recognition of some limitations. The cross-sectional design, as in other previous studies on ADA (Muñoz-Rivas et al., 2007a, 2007b, limits inferences about the developmental changes. Longitudinal studies could help explain the development across time of this phenomenon. There are very few longitudinal studies that used the same measures of aggression. This need for longitudinal research seems crucial, especially since we have almost no research beginning in the teen years and extending into long-term adult relationships or marriage. Such research is needed to learn which men and women cease their aggression and which continue or escalate their aggressive behaviors. Moreover, in this study, we did not investigate participants' sexual orientation, thus we do not know if they reported ADA occurring in heterosexual or same-sex relationships. Future studies may consider this aspect to Cuccì,O'Leary,Olivari,& Confalonieri 669 obtain more generalizable results. Finally, the CADRI (Wolfe et al., 2001a) is a frequency-based measure of ADA, thus this scale does not investigate motivations for ADA. In the future, qualitative studies, that allow a better understanding of perceptions and attitudes towards a phenomenon (e.g., Olivari, Confalonieri, & Ionio, 2011), are warranted in order to gain more information to better understand the reasons linked to ADA. Future research should be also aimed at collecting dyadic data in order to assess mutuality of ADA within the same couple in detail, considering relationship characteristics such as relationship length, partner age and balance of power.
We believe our work could provide interesting clues for ADA prevention in Italian high schools. It is clear that both male and female adolescents engage in ADA, so programs should target both male and female aggression. Additionally, interventions should begin in high school or earlier since ADA behaviors are already observed among 16-year-old adolescents. Finally, considering that emotional-verbal and psychological abuse are the most common forms of ADA, programs should also target these forms of aggression, as to date, the majority of attention on aggression has been devoted to physical and sexual forms. This change in focus may improve adolescent awareness and knowledge on all forms of ADA and their consequences. Finally, it is important to develop not only prevention interventions, but also programs of skills empowerment aimed at making adolescents better able to manage couple conflict and improving positive and adaptive interaction strategies.

Funding
The authors have no funding to report.