For the subsequent statistical analyses, the variable sexual aggression was dichotomized as yes/no. “Yes” was selected when the participant reported to have perpetrated some type of sexual aggression (i.e., sexual contact, sexual coercion, attempted rape, or rape) at least once in his/her life and was assigned the value 1, whereas “no” was selected when the participant reported to have never perpetrated sexual aggression and was assigned the value 0. The same procedure was conducted for the variable “sexual victimization”.

Table 2 shows that perpetrators reported the following prevalence by subtype of aggression: 24.8% sexual contact (i.e., non-penetrative sexual contact, such as kissing, fondling, etc., with an unwilling partner); 14.3% sexual coercion (i.e., sexual intercourse without an individual's consent by means of verbal pressure or use of authority); 12.5% attempted rape; and 10.3% rape. Regarding the number of times sexual aggression was perpetrated, 52% of the participants reported perpetrating aggressive sexual behavior only once, whereas 48% reported engaging in sexual aggression more than once. On average, aggressors reported perpetrating sexual aggression 3.02 times (SD=2.99). In addition, the victim perpetrated by subtype of aggression was indicated, considering the distinction from the SES: stranger, acquaintance, dating partner (casual partner), and ex/current partner (involved in a stable relationship).

The aim of this study was to explore the variables that are associated with and which best predict self-reported sexual aggression in a sample of Peruvian adolescent males. Our results revealed that 33.3% of the sample reported having perpetrated sexual aggression at least once. Of the total sample, according to the subtype of sexual aggression, prevalence ranged from 10.3% for rape to 24.8% for sexual contact with an unwilling partner. The most frequent victim of any type of sexual aggression was a dating partner (casual partner), whereas the least frequent was a stranger or an ex/current partner (for rape). These results are comparable to other prevalence rates of victimization for samples in Spain (Moyano & Sierra, 2015; Santos-Iglesias & Sierra, 2012; Vega-Gea, Ortega-Ruiz, & Sánchez, 2016), the United States (Breiding et al., 2015), and various Latin American countries such as Chile (Lehrer, Lehrer, Lehrer, & Oyarzún, 2007), Mexico (Cortés-Ayala et al., 2014), and Brazil (D’Abreu, Krahé, & Bazon, 2013). Our findings indicate high prevalence rates, taking into account that participants–whose ages ranged from 14 to 17–were asked about aggressions committed since the age of 14.

When comparing sexual aggressors and non-aggressors, we found that sexual aggressors reported having fewer egalitarian gender beliefs and more supportive attitudes towards rape. However, Cohen's d was not very robust. Moreover, sexual double standard was found to be associated with rape supportive attitudes (r=46). This is consistent with previous studies in which the endorsement of traditional gender roles was associated with the justification of forceful actions (Shen et al., 2012; Sierra, Monge et al., 2010; Sierra, Santos-Iglesias, Gutiérrez-Quintanilla, Bermúdez, & Buela-Casal, 2010).

In the logistic regression analysis, data revealed that the best predictors of sexual aggression were self-reported abuse during adolescence and rape supportive attitudes. This was the case for 67% of the sample. These findings support, on the one hand the cycle of sexual coercion (Ward & Beech, 2006), that is, individuals who have been victimized are more likely to become aggressors. On the other hand, males who tend to perpetrate such violence are more likely to accept justifications for rape. As previously shown by Maxwell, Robinson, and Post (2002), attitudes towards violence predicted male dating aggression in a sample of adolescents. Consistently, Sierra et al. (2009) concluded that rape supportive attitudes predicted sexual coercion in a sample of university students from Salvador. Thus, violence might be justified by blaming the victim. Therefore, justification of violence is a relevant variable for aggression (Ruiz-Hernández, García-Jiménez, Llor-Esteban, & Godoy-Fernández, 2015), which has often been associated with hostile sexism (Lila, Oliver, Catalá-Miñana, Galiana, & Gracia, 2014).

Previous research has revealed an association between being a sexual victim and later perpetrating sexual coercion in men (e.g., Moyano & Sierra, 2015; Russell & Oswald, 2002). Prevalence of victimization in Peru is among the highest from Latin America (61%) (Garcia-Moreno et al., 2006). In addition, 40% of Peruvian adolescents and youngsters initiate their sex life by forced intercourse (Cáceres, Marin, & Hudes, 2000). Although in the present study we did not analyze who were the perpetrators of boys’ victimization, data from the National Action Plan for Childhood and Adolescence (2012-2021) from Peru indicate that sexual abuse at these ages tend to be perpetrated by a closely-related individual. In this sense, the use of parents’ violence in Latin America is high, being boys those more highly affected (Gage & Silvestre, 2010). Other data, from Guatemala and Salvador, indicate that 46% and 42% of men respectively, reported being beaten as punishment at younger ages (Speizer, Goodwin, Samandari, Kim, & Clyde, 2008).

In our study, the most reported victim perpetrated was a dating partner (casual partner). It is likely that these participants are frequently involved in violent relationships in which they are victims and aggressors at the same time. Although previous research show that aggression is more frequently presented in committed dating relationships (e.g., Hanley & O’Neill, 1997), research with college students (Klipfel, Claxton, & van Dulmen, 2014) and with adolescents suggests the co-occurrence of sexual risk-taking, such as having casual partners and sexual dating aggression (Gover, Kaukinen, & Fox, 2008).

Contrary to expectations, and in contrast with the study conducted by Sierra et al. (2009), sexual double standard was not found to be relevant to the prediction of sexual aggression. Our findings rather indicate that attitudes that support and justify certain aggressive behaviors are more important than beliefs on the dominance of males over females, at least in adolescence. This is consistent with a review of the dynamic risk factors of dating violence from continental Europe, and from the United States and Canada (Leen et al., 2013). The authors suggest that attitudes towards violence, such as the acceptance of rape myths, tolerance of violence, and justification of the use of violence, were reported as risk factors for dating violence perpetration.

The endorsement of sexual double standard attitudes in itself may not be sufficient to perpetrate aggression. As it is well known, sexual double standard entails the dominance of one group (men) over another (women), which is aligned with the construct of social dominance orientation, defined as the degree to which a person holds anti-egalitarian values and a preference for group-based hierarchies (Sidanius & Pratto, 2004). Similar to sexual double standard, men tend to exhibit more social dominance orientation. In a study of adolescents, Mayeux (2014) found that social dominance orientation did not predict aggression. Another study of Chinese adolescents showed that the strongest predictor of sexual aggression was the boys’ attitudes towards dating violence rather than their traditional gender beliefs (Shen et al., 2012). Taken together, sexual double standard might be more relevant (at least during adolescence) to the prediction of prejudice attitudes, but not to the perpetration of violence.

In a similar vein, a review by Flood and Pease (2009) on the factors that influence attitudes towards violence against women, suggested that general gender attitudes cannot be taken as a simple proxy for attitudes to violence against women. As emphasized by Murnen, Wright, and Kaluzny (2002), measures of general gender-role attitudes have less power to predict male sexual aggression than other measures of patriarchal masculine beliefs in particular.

Another line of interpretations regarding non-significant association of non-egalitarian beliefs with violence is that male's perceptions of the attitudes of the female partner may be important when considering the use of violence as a way for the male to recover dominance. More specifically, some women might endorse similar non-egalitarian attitudes, and thus not provide any resistance to men's attitudes. Based on the confluence model (Malamuth, 2003), one of the most comprehensive explanatory models of male sexual aggression, gratification from dominating women is a predictor of sexual aggression (within the “hostile masculinity” factor). In other words, male sexual aggression may be motivated by dominance and power, as previously suggested (Russell & Oswald, 2001; Zurbriggen, 2000). In a similar vein, male sexual attitudes are often normalized. Consequently, there is also a certain level of constant pressure among boys to behave in sexually aggressive ways, and girls are sometimes likely to accommodate male needs and desires in negotiating their sexual relations (Hird & Jackson, 2001; Tolman, Spencer, Rosen-Reynoso, & Porche, 2003).

This study obtained self-reported data from an adolescent sample of Peruvian boys. To date, little research has approached the study of both gender-attitudes and rape-attitudes in adolescent samples. Therefore, this study provides valuable data for young boys in this age group. Overall, we found that key factors in the prediction of sexual aggression were: (1) having been a victim of sexual aggression and (2) the endorsement of supportive attitudes towards rape. Prevention programs with aggressors and psychosocial interventions should address these factors, in particular, those related to violence-supportive attitudes and to the justification of violence. This is important because until now most interventions have mainly focused on gender inequality, which is often seen as the sole explanatory factor for gender violence (Carbajosa & Boira, 2013). Therefore, there is a special need for multimodal interventions (see Arce & Fariña, 2010), which integrate attitudes towards violence and its justification.

Beyond interventions, these findings should be integrated into legal and political strategies proposed by the Peruvian government. For the last decade, several legal proceedings have been developed for the prevention of sexual abuse in children and adolescents and the promotion of gender equality in Peru. Some of these proposals are materialized in the National Action Plan for Childhood and Adolescence (2012-2021) and the National Plan for Gender Equality (2012-2017). Both programs establish strategies with the priority to reduce sexual violence among children and adolescents, through the promotion of research about the risk factors of sexual abuse and education campaigns based on a gender equality approach.

This study had certain limitations. First at all, the adolescent sample covered a narrow age range (from 14 to 17). A broader picture would have been obtained if younger boys had also been incorporated. Secondly, given that this research has a cross-sectional, correlational design, no causal effects can be established. In this sense, a longitudinal study would have allowed us to better track individuals during their adolescence and even on to their adult transition. This would have undoubtedly enriched the understanding of sexual aggression. Thirdly, it is likely that sexual double standard may not only be related to sexual aggression, but also to psychological or verbal aggression (i.e., disagreements with the partner about role expectations). Fourthly, considering that all information obtained in this study was based on self-reports, social desirability should have been measured and controlled. “Macho” scripts are still robust in Peru, thus males may be exaggerating or minimizing their report of their aggressiveness. Finally, regarding victimization information, more data about the characteristics of the perpetrator of those who report to have committed aggression, should be of interest (being victimized by adults versus peers).

The aforementioned means that other measurements of sexual aggression, which encompass a broader spectrum of violence and perpetrator's characteristics, should be considered for further research. Finally, future studies should consider either male's perception of his partner's endorsement of the same sexual standards or beliefs (as aggression most likely stems from disagreement between both perspectives) or data collection from the partner. This type of dyadic analysis would provide knowledge about whether both partners share the same gender beliefs.

Nonetheless, the results of this study provide valuable information for education and intervention programs with male adolescents, who are a vulnerable target for prevention purposes. Educational programs should pay special attention to and focus on the awareness of rape supportive attitudes.

The authors of this article declare no conflict of interest.