Sexual violence is no longer an unspoken issue in the world of youth sport. It is not only a potentially criminal issue, but prevalence estimates also clearly illustrate the problem as a health issue. Mountjoy et al. (2016), p. 3 define sexual violence as “any conduct of a sexual nature, whether non-contact, contact or penetrative, where consent is coerced/manipulated or is not or cannot be given.” In this study, we likewise use the term as an umbrella concept encompassing verbal and non-verbal contact forms of sexual harassment and abuse. Depending on the study designs and measures used, prevalence rates for sexual harassment range from 19% to 92% and for sexual abuse from 2% to 49% (Mountjoy et al., 2016; Bjørnseth and Szabo, 2018). In a recent European study, surveying over 10,000 young adults in six European countries, 32% of the respondents in Flanders (Belgium) reported having experienced at least one form of non-contact sexual violence and 20% reported having experienced at least one form of contact sexual violence within the context of organized sports before the age of 18 (Hartill et al., 2021; Vertommen et al., 2021). In most cases, it did not concern single events but repeated incidents that took place over longer periods of time (Vertommen et al., 2019).

The often severe physical, psychological, social, and societal consequences of such incidents render sexual violence in sport a serious public health issue (Koontz et al., 2021). Yet, despite its high prevalence and impact, signs or incidents of sexual violence in sport are often overlooked or not acted upon. Not intervening or offering help allows the violence to continue, facilitating the perpetrator and thus aggravating and prolonging the victim’s suffering (Cense and Brackenridge, 2001; Spaaij and Schaillée, 2019). The role of bystanders (i.e., individuals that are not directly involved as a victim or perpetrator that have the potential to intervene and amend the situation; Banyard, 2008) can be important in the context of sexual violence.

Banyard et al. (2020) distinguish two types of bystander interventions: reactive and proactive. Reactive bystander actions concern both the negative or positive actions that a person undertakes in response to a high-risk situation (McMahon and Banyard, 2012; Rothman et al., 2019), where negative reactive bystander action is defined as either a lack of action or unhelpful actions (e.g., making fun of the victim, applauding the perpetrator’s conduct) immediately before, during or after an incident of sexual violence. Conversely, positive reactive bystander behavior encompasses all potentially helpful actions directly before, during or after an incident, i.e., at the time of an immediate risk or an ongoing, witnessed event of sexual violence.

Proactive bystander behavior is, in contrast, not restricted to one specific incident but comprises all supporting behaviors aimed at ameliorating or preventing sexual violence in general, (which may, for instance, include taking a course on sexual violence or remaining aware and perceptive of its risk or occurrence; Banyard et al., 2020). In the literature, people who take responsive and helpful actions in emergency situations are also labeled as “actionists” (Rothman et al., 2019), upstanders’ (Ferrans et al., 2012), “defenders” (Pozzoli et al., 2012), or “prosocial bystanders” (Banyard, 2011). In the present study, the terms positive bystander(s) and positive bystander behavior include both reactive and proactive onlookers and interventions.

Positive bystanders play an important role in maintaining a safe environment, and over the last decade the development, implementation, and evaluation of bystander interventions has received more attention in research, policy-making and practice than ever before. By educating people about sexual violence and bystander options, peer norms, attitudes, and beliefs can be changed. Adequately intervening in situations of sexual violence is not only beneficial to the victim and might stop the perpetrator, it also models the desired behavior, potentially encouraging fellow-bystanders to do the same on future occasions. In addition, positive bystander behavior may also reduce the risk of victimization and perpetration in bystanders themselves (Banyard, 2015).

A variety of bystander intervention programs in the context of violence is available (e.g., “Bringing in the bystander,” “Coaching boys into men,” “Green dot”), but only a few focuses on the context in sport. Most were developed in the United States and mainly focus on the prevention of sexual and dating violence on college campuses (e.g., Miller et al., 2012; Palm Reed et al., 2015; Fenton and Mott, 2018). Some of these bystander programs have been systematically reviewed and were found to be effective with small to moderate effects on different components. Some bystander interventions were effective in increasing or promoting actual bystander behavior (Katz and Moore, 2013; Kettrey et al., 2019; Kettrey and Marx, 2019), while others were effective in increasing knowledge, pro-social attitudes and beliefs about sexual violence (Jouriles et al., 2018). The authors concluded that the main reason for those small effects was the use of small samples (Jouriles et al., 2018; Kettrey et al., 2019).

Banyard et al. (2004, 2007), for example, showed that one- or three-session programs had improved attitudes, increased bystander efficacy and knowledge, and, most importantly, engendered prosocial bystander behavior in 389 male and female undergraduates. The recent meta-analysis of Jouriles et al. (2018) confirmed these findings and found a greater effect on bystander attitudes and beliefs for programs with a longer duration, but the effects diminished over time.

Studies regarding perceived norms, which includes the perception of other’s willingness to intervene show that these norms play a significant role in bystander behavior (e.g., Brown and Messman-Moore, 2010; McMahon, 2015; Santacrose et al., 2020). People may be concerned about a violent situation but feel not confident to act, as the group norm implies it is not a problem (Fenton and Mott, 2018).

In their study, Levine et al. (2019) explored key factors involved in violence reduction and bystander interventions. Firstly, they emphasize that intervening in violent or hazardous situations is already the social norm among bystanders, even though interventions may not always be successful. Secondly, the authors stress that people do intervene even when others are present, thus showing that the popular belief that people do not act in the presence of other witnesses is a fallacy, making this a vital message to convey in bystander campaigns. Another major issue they describe is that people often think it is not safe to intervene. However, the researchers found the opposite: the chance of being victimized while intervening is quite low. The final finding they report on concerns the social identity relationship between bystander and victim or perpetrator: bystanders are more likely to intervene in situations involving members of their own social or ethnic group are being victimized (e.g., friends, family, and acquaintances).

To make prevention programs even more effective, Moynihan and Banyard (2008) recommend instructing course participants to not focus on possible perpetrators or victims only but rather on all stakeholders in the organization or environment given the notion that all members of a community have a role in preventing violence or harm and that everyone can become a bystander. Banyard (2015) states that prevention programs should be directed at promoting proactive and reactive behaviors among all members of a given community.

When aiming to promote positive coach-bystander behavior, we first need to analyze its constituent parts. Before taking action, any bystander will, consciously or unconsciously, go through several, interlaced stages. First of all, they will have to note and interpret a situation as a problem. Next, they should feel the responsibility to do something about it, for if they do not, they will not take action. When they do experience a sense of responsibility, they will subsequently need to figure out what to do and, if a decision is reached, eventually choose to take action (Banyard, 2011). Each of these constituent stages can be influenced by different environmental conditions and psychological constructs (i.e., behavioral determinants; Bartholomew et al., 2006). A first step in developing successful coach-bystander intervention programs is translating these steps in the context of sport and subsequently gaining insight into the determinants of each sub-behavior.

Knowing that signs of sexual violence in sport are often overlooked or not acted upon (Vertommen et al., 2019), understanding this behavior is crucial. The literature does not provide these insights in this context, therefor this paper aims to identify the determinants that contribute to coach-bystander behavior regarding sexual violence following the Intervention Mapping Protocol. This framework outlines steps, tasks and processes to help develop practical health promotion and education programs through stimulating desirable behaviors (Bartholomew et al., 2006). These insights will inform the development of a dedicated coach-bystander training program to prevent sexual violence in youth sports.

To this end, we adopted the reasoned action approach (RAA; Fishbein and Ajzen, 2010) as a theoretical framework. The RAA has been used multiple times to predict a variety of behaviors, but not in the context of coach-bystander behavior regarding sexual violence. Identifying the essentials predictors of coaches’ intention, constitutes the basis for developing a coach-bystander intervention, and the RAA offers these predictive insights as the theoretical framework links people’s intention to underlying attitudes, norms and beliefs. The practical contribution of this study lays in the next step, where the found determinants will be translated into applications that are part of the actual intervention.

The RAA model builds on the theory of planned behavior (Ajzen, 1991) and the theory of reasoned action (Fishbein and Ajzen, 1975) which explains why people decide to perform (or not to perform) a certain behavior (Fishbein and Ajzen, 2010). The RAA states that the intention to carry out a behavior, such as intervening in case of a violent event, is determined by a person’s attitude toward the behavior, their idea of how relevant others perceive the behavior (perceived norms), and their sense of behavioral control. For people with high intention to intervene, actual and successful execution of the behavior also depends on their abilities and skills, as well as environmental factors. The RAA explicitly does not address automatic behaviors (e.g., habits). Since positive bystander behavior is only infrequently called for, it firmly belongs to the domain of reasoned action.

When behavior is explained by a theory, the variables are often called determinants. A determinant is a psychological variable on global scale, which can be further specified in subdeterminants. Subdeterminants are situated on a lower level of the psychological generality to predict the overarching determinant (Peters and Crutzen, 2018).

Attitude, the first determinant of the RAA, is defined as a person’s subjective evaluation of the consequences of a behavior, conceptualized as favorable or unfavorable. Fishbein and Ajzen (2010) distinguish instrumental and experiential attitude, where in the first the individual sees the outcome as instrumental to their goal (e.g., “If I intervene, this is likely to end the sexual violence.”), whereas experiential attitude is about the anticipated affective consequences of the behavior (e.g., “If I intervene, it is likely that I will feel confident”).

Perceived norms refer to a person’s perception of the approval or disapproval (injunctive norms) and behavior (descriptive norms) of relevant socials referents. Injunctive norms are all about what you think others expect of you in a particular social context (e.g., “People important to me would approve of me protecting children from sexual violence”). An example of descriptive norms would then read “Most people like me will protect children from sexual violence” (Fishbein and Ajzen, 2010).

Perceived behavioral control is a person’s perception of their capacity and autonomy over their behavior, where capacity entails the level of confidence a person has in being capable to perform a behavior (e.g., “I am confident that, if I want to, I can protect children from sexual violence”). Autonomy refers to people’s belief that they have control over the behavior and that performing the behavior is up to them (e.g., “It is me who decides whether I protect children from sexual violence”; Fishbein and Ajzen, 2010).

Because educational programs cannot directly change behavior but can change behavioral determinants, understanding which determinants primarily influence a given behavior in a given population is an important first step in developing behavior change interventions (Crutzen et al., 2017). Besides establishing which determinants are the most promising targets for an intervention, it is important to identify critical subdeterminants, i.e., determining underlying beliefs. Thus, if perceived norms play a prominent role, it is vital to learn which social referents matter the most (i.e., peers, family, or other persons or groups). If behavioral control is the problem, we need to identify the perceived barriers or the (sub)skills people feel they lack.

The study presented below focuses on youth sport coaches as key players in the prevention of sexual violence toward child athletes in local sport settings. Almost all athletes begin their sport careers in local clubs at an early age, making safe sport environments crucial for their wellbeing and development. Since coaches play such a prominent role in youth sport, they are necessarily critical in their protection from harassment and abuse, most particularly sexual violence. We therefore conducted a survey among youth sport coaches to systematically identify the most relevant determinants of positive bystander behavior for future inclusion in a coach-bystander training program. The research question guiding this study is: what are the most relevant determinants of positive bystander behavior in the prevention of sexual violence in sport?

The majority of the coach-participants reported not to have noted or witnessed any case of sexual violence in their sports club in the previous 12 months (n = 1,184, 90.3%), while 78 (5.5%) had witnessed one such event, of whom 52 (n = 52/78; 66.7%) had intervened. Another 49 coaches (3.4%) reported to have witnessed several situations involving sexual violence, with six (n = 6/49; 12.2%) coaches not taking any action and nine (n = 9/49; 18.4%) having only rarely done so; 11 (n = 11/49; 22.4%) had acted on some occasions, 10 (n = 10/49; 20.4%) had intervened in most instances, while 13 (n = 13/49; 26.5%) coaches had intervened in each event. Most common reasons for not intervening were a failure to (correctly) identify the behavior as sexual violence (“it looked harmless”), the lack of knowledge and skills to name and/or address the issue, and an inaction due to the perpetrator’s hierarchical position.

Next, we will describe the (sub-)determinants our analyses identified as the most relevant for the promotion of positive coach-bystander behavior. Figures 2–4 provide examples of the analysis process for several of these target behaviors.

We adopted the CIBER approach to evaluate determinants of positive coach-bystander behaviors to develop a bystander intervention and therefore help to prevent sexual violence in youth sport. The majority (90.3%) of the 1,422 coaches responding to our online survey reported not to have witnessed any incident involving sexual violence in the last 12 months. This could mean that there were no actual incidents or that coaches may not have been successful in detecting (signs of) sexual violence (Mountjoy et al., 2016; Vertommen et al., 2016a), which in turn could be explained by the fact that coaches do not have the knowledge to correctly identify and assess a situation of sexual violence. Additionally, due to the COVID-19 lockdown that coincided with a large part of the study period, most organized sport activities were cancelled and therefor the opportunities to witness any incidents may have been relatively scarce. Nevertheless, the fact that 10% did report witnessing one or more incidents in the given timeframe highlights the importance of raising awareness and educating coaches about sexual violence in sport.

By combining CIBER plots with the potential for change we were able to identify the most relevant factors contributing to the four coach-bystander behaviors that could be targeted in sexual-violence prevention programs for coaches. Reassuringly, we found high intention for all four target behaviors, reflecting that the coaches in our sample would be inclined to intervene when witnessing sexual violence during their watch, with the majority of those having observed an actual incident following up on this intent by taking action. Still, some did not intervene.

In line with previous studies (Banyard et al., 2007; Jouriles et al., 2018) attitude was found to be an important and changeable determinant in bystander intervention. In our study, the experiential attitude in terms of belief expectation and instrumental attitude in terms of belief evaluation are put forward as promising determinants to focus on in the development of coach-bystander interventions. The same applies for perceived norms, which has been pointed out in the literature as an important determinant (e.g., Brown and Messman-Moore, 2010; McMahon, 2015; Santacrose et al., 2020), and was also confirmed in our results.

Of perceived behavioral control (i.e., perception of their capacity and autonomy on their behavior), subskill presence, played only a role in two target behaviors regarding taking action: setting firm boundaries and reporting to the safeguarding officer. Although various studies on topics ranging from child maltreatment to sexual violence show that higher levels of confidence are related to a greater willingness to offer help (e.g., Mudde et al., 2007; Moynihan and Banyard, 2008), our study does not confirm this. Perceived behavioral control proved the least relevant attribute to target and therefor this determinant is not considered to as a key element to target when developing a coach-bystander intervention.

Like Witte et al. (2017), who observed that many bystanders who had intervened in a violent situation reported their actions had made them feel positive, proud and empowered (with far fewer reporting negative sentiments), we found in our study that all but one of the items gauging the participants’ expectations about their behaviors, and most specifically their experiential attitude on belief expectations (i.e., affective consequences), were selected and thus core targets for a prevention program. We wish to note that the subdeterminant “As a coach I feel comfortable being vigilant for signs of sexual abuse,” not approached the threshold score (0.05) but may still be relevant as an intervention target.

Of perceived referent behavior, the second most relevant factor we identified, only the (sub-)determinants of the target behavior on being vigilant were not relevant. All other items (e.g., “Most people like me will set boundaries in case of sexual violence” and “Most people like me would intervene in a situation of sexual violence”) appeared important for fostering positive bystander behavior. Levine et al. (2019) earlier highlighted the importance of utilizing shared group membership through which intervening becomes the ‘social’ norm among peers when threatening or violent events are being witnessed. A sexual-violence prevention program for coaches of local sport clubs should then focus on this sense of group membership and prevailing (pro)active peer behavior.

Looking at instrumental attitude, we found belief evaluation (i.e., how one rates the value of a particular behavior; Fishbein and Ajzen, 2010) to be relevant for all four target behaviors—and the only determinant that was significant for vigilance (target behavior 1). This latter would suggest that a prevention program should also be aimed at reinforcing coaches’ willingness to be watchful for (signs of) sexual violence.

Perceived referent approval (i.e., appreciation by relevant referents) was the last most important determinant found in our study, and was relevant for all subdeterminants of all three target behaviors, except for vigilance. Negative social group norms can hamper one’s intention to intervene as a bystander, which refers to the approval of others. In the context of sport this mainly refers to fellow-coaches and club managers. Based on our findings, not all coaches in this sample can count on the approval of their fellow trainers and managers when intervening in case of an incident of sexual violence.

Summing up, our study shows that the coaches in Flanders that responded to our survey already show high levels of intention to engage in the desired positive bystander behaviors we evaluated but that there is room for improvement. Our results will aid the development of a well-founded sexual violence prevention program for youth sport coaches. Methods and applications to effectuate positive attitudes (experiential attitude in terms of belief expectation and instrumental attitude in terms of belief evaluation) and positive perceived norms (perceived referent behavior and perceived referent approval) will be key in the development of effective intervention programs. Finally, although determining features of target behaviors can differ per population, region, and over time, the patterns we identified in the youth sport context may still prove useful for prevention professionals in other settings who lack the means to conduct determinant studies themselves (Fishbein and Ajzen, 2010).

The CIBER approach in combination with the potential change index was applied to gain a better understanding of the predictors of positive coach-bystander behavior with respect to sexual violence and to subsequently select the determinants in order to develop a coach-bystander intervention. The results showed that for the prevention of sexual violence in sport, coach-bystander interventions should target specific (sub-)determinants of their attitude toward and perceived norms of the desired (pro)active behaviors. The aims of such interventions should focus in the first place to increase the positive affective consequences coaches’ experience when engaging in the target behaviors. When intervening in situation of sexual violence coaches must feel comfortable and determined in order to act as a prosocial bystander. Secondly, the sense of group membership among coaches and manager in the sport club needs to be reinforced. Intervening in case of an incident of sexual violence needs to be the social norm in the sport club. And lastly, the focus should lay on the evaluation of the value of the specific target behaviors. When intervening becomes the social norm, fellow-coaches and managers will give their approval when a coach intervenes in case of an incident of sexual violence. Although determining factors may vary for different types of abuse and sport settings, the outcomes presented can also inform other violence-in-sport prevention campaigns.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

The studies involving human participants were reviewed and approved by Thomas More University Ethics Committee (G-2020 12 2035). The patients/participants provided their written informed consent to participate in this study.

HV was involved in the conceptualization, funding acquisition, investigation, and methodology and wrote the original and final draft of the manuscript. TV was involved in the conceptualization, funding acquisition, investigation, methodology, and project administration and contributed to the revision of the original draft of the manuscript. G-JP was involved in the conceptualization, investigation, and methodology and contributed to the revision of the original draft of the manuscript. All authors contributed to the article and approved the submitted version.

This project received internal funding from Thomas More Research (M-OHC-190086).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.