The World Health Organization (WHO) (2017) has described domestic violence and abuse (DVA) as a serious public health problem of global epidemic proportions. For England and Wales, DVA has certainly become a national pandemic, costing approximately £16 billion each year, and statistics consistently show that 1 in 4 women will experience DVA at some point in the lifetime (Guy et al. 2014; Walby, 2009). Whilst physical, sexual, financial and emotional abuses have long been recognized as coming under the umbrella of DVA, coercive and controlling behavior (hereafter called ‘coercive control’), as an insidious form of relationship abuse, is now recognized for its considerable distressing and harmful effects in the UK and beyond (Home Office 2018). This is helpful as it also steers discourse away from the public story of DVA interpreted as physical violence perpetrated by men against women within a heterosexual relationship (Donovan and Hester 2014) to a more nuanced understanding of DVA as complex and multi-dimensional. Over the last decade or so, what has increasing been brought to the fore, is the realization that DVA is not a social problem limited to adulthood, but it is also a problem in the relationships of children and young people. Acknowledging this, in 2013, the UK’s Home Office widened the definition of DVA to include young people aged 16 and 17 to:

Any incident or pattern of incidents of controlling, coercive, threatening behaviour, violence or abuse between those aged 16 or over who are or have been intimate partners or family members regardless of gender or sexuality. (Home Office 2018, para 1)

Moreover, in March 2017 the UK’s Central Government announced that it will legislate to ensure that all schools (primary and secondary) will teach children and young people about healthy relationships in the future (HM Government 2017).

This paper presents the findings of a secondary analysis of the data collected during the delivery of the Change Up programme. Change Up is a secondary school project which promotes healthy relationships and uses social norms theory to explore young people’s attitudes and experiences of DVA. This paper aims to illuminate how social norming theory is beneficial in DVA prevention programs with young people.

Literature review

Whilst methodological and conceptual inconsistencies between studies into young people’s experiences of DVA render it difficult to compare data and findings (Hellevik et al. 2015), there is an emerging body of work in this field. In 2009 Barter et al. (2009) reported concerning levels of physical, psychological/emotional and sexual abuse within the relationships of young people aged 13–17 years after surveying 1353 young people from eight secondary schools across England, Wales and Scotland. Of 88% respondents who had experienced some form of intimate relationship, 22% had experienced moderate physical violence (pushing, slapping or holding down) and 8% had experienced more severe physical violence (punching, strangling, using an object). Three-quarters of the girls and half of the boys had experienced emotional abuse, with the most common form as ‘being made fun of’ and/or the use of surveillance in ‘constantly being checked up on’. One in three girls and 16% of boys reported some form of sexual abuse from a partner with 70% of girls and 13% of boys stating that this had negatively impacted their well-being. Drawing attention to the gender-based framework for understanding the dynamics and impacts of DVA, Barter et al. highlighted that ‘a central issue concerns gender. Girls, compared to boys, reported greater incidence rates for all forms of violence (Barter et al. 2009: 4).

More recently, Broad and Gadd (2014) conducted a survey of 1203 young people, aged 13–14 years old, finding that over half had some direct experience of DVA (whether as victims, perpetrators or as witnesses). They found that 44% of boys and 46% of girls reported that they had experienced at least one of the types of DVA (physical, mental/emotional, sexual abuse or coercive control). The most commonly reported experience of abuse pertained to emotional abuse and controlling behaviors with 38% reporting at least one type of maltreatment falling into one of these categories. Diverging from Barter et al.’s findings, when gender differences were tested for physical abuse, sexual abuse and emotional abuse/controlling behaviors, the only significant difference recorded was for sexual victimization; with girls reporting considerably more than boys.

An international evidence synthesis by Stonard et al. (2014) also identified concerning levels of all types of abuse (physical, sexual and emotional). This study found a high percentage (between 50 and 70%) of young people who reported experiencing abuse through new technologies and there is an emerging body of work reporting forms of exploitation and abuse which use social media and digital technologies in young people’s relationships (Zweig and Dank 2013; Hellevik et al. 2015). Stonard et al.’s study drew together findings from countries in the Global North including the US, Canada, the UK, Europe and New Zealand.

Overall, these findings are unsurprising as the evidence is building. One study across five European countries (the UK, Norway, Italy, Bulgaria and Cyprus) investigated young people’s experiences of face-to-face and digital abuse (Hellevik et al. 2015) again finding significant levels of DVA for young people, but with considerable differences in the prevention and intervention policies and practices between the countries and in the way that the role of gender factors in these responses and each countries’ interpretation of phenomena (Barter et al. 2015). For example, when focusing on sexual abuse in young people’s relationships, in Bulgaria this is often sensationalized in media reporting (linked to pedophilia or stalking) with young people’s sexuality purportedly a hotly debated topic, whereas in Italy data is not systematically collected and consequently there is a tendency for issues, such as sexual abuse in young people’s relationships, to be confused and conflated with others such as pedophilia, familial sexual abuse, child pornography, child trafficking/sexual exploitation, cyber/bullying and gender-based violence more broadly (Barter et al. 2015).

Other work found within the body of international literature reveals that victimhood and perpetration in adolescence are influenced by social, cultural and lifestyle factors (Sabina et al. 2016). For example, when accounting for age differences, the evidence-base in the USA suggests that incidents of DVA in young people’s relationships increase as they get older (Hokoda et al. 2012). Viewing age as an indicative factor is important as research indicates that those young people who are exposed to relationship abuse earlier during adolescence are more likely to experience DVA later in life (Alleyne-Green et al. 2012). There are, therefore, implications with regard to the timing of interventions (Hokoda et al. 2012). Yet in terms of gendered experiences, overall the global literature presents conflicting results, and it is reasonable to conclude that girls and boys are both perpetrators and victims of DVA with more research needed to provide a clearer picture of perpetration, victimhood, risk and protective factors.

Addressing DVA Through Interventions with Young People

Within the DVA sector and across the academy, there is an increasing interest in how social norms theory (SNT) can be harnessed to address gender-based violence, in particular, and other gendered inequalities, more generally (Cislaghi and Heise 2017). This includes identifying a simple way to measure social norms and using SNT to design successful interventions. Between 2016 and 2017 Social Sense, a social marketing agency delivered a pilot project, Change Up, based upon SNT which focused on early prevention work for young people associated with, involved in or at risk of DVA. By using a targeted approach, it was envisaged that the project would, to some extent, address the fragmented support available to young people living in pockets of a city known to have high levels of DVA. Across 2016 and 2017 the Change Up project delivered a high school-based prevention program centering on healthy (non-violent) relationships (described below) to young people aged 13 to 14 years old.

Whilst a national picture of DVA in young people’s relationships has been emerging over the past decade, the location for Change Up was the North West of England which has some of the highest rates of DVA in the UK (CPS 2012). For example, in 2017 across Greater Manchester 22,739 domestic abuse related crimes were recorded, and 67,987 domestic abuse related incidents (incidents not recorded as a crime); combined this made the North West region the third highest in England and Wales (ONS 2017). These statistics pertain to adult experiences as young peoples’ experiences of DVA are not systematically recorded in the UK, and therefore, a pragmatic approach was taken to the location of program delivery with the setting for both high schools in an area of the North West known to have high rates of DVA with Change Up aiming to positively affect future statistics of DVA perpetration. A number of high schools were approached resulting in two schools willing to participate at the time of the project delivery.

In the next section, an explication of a social norms approach is provided to demonstrate its value in prevention work with young people on sensitive topics such as DVA. The project design of Change Up is explicitly described to illustrate how the data was collected. This data was subject to a secondary analysis and the results describe the norms and attitude change between the baseline and repeat surveys towards physical and psychological violence (focusing on coercive and controlling behaviors) and is followed by an account of the experiences of young people. Both qualitative and quantitative data has been triangulated and the discussion synthesizes the results, exploring these in light of current understandings of DVA and young people.

A Social Norms Approach

Since the 1950s the influence of social norms on people’s behavior has been studied across the social sciences. As a consequence, a social norming approach (SNA) is based upon a well-articulated theory of behavior and evidence-based methodology for addressing health and social justice issues (Berkowitz 2012). In work with young people, it has increasingly been utilized in prevention work around alcohol use and smoking (Elsey et al. 2015; Sheikh et al. 2017). More importantly, it is during the last decade that social norms theory has gained momentum as a potentially useful means of addressing gender-based violence and domestic abuse (Cislaghi and Heise 2017).

In a social norming approach, a norm is a belief or custom that is held by the majority of a group or community with three identifiable types: actual; perceived; and misperceived norms. Actual norms are those which are actually believed or shown in behavior, whereas perceived norms refer to what people think or perceive the norm to be. A misperceived norm refers to when the perceived norm is different from the actual norm; that is, when what people think is the norm is not actually the case (Berkowitz 2012). For example, young people wearing hoodies have been portrayed in the media as deviant or criminal (the misperceived norm). The disconnect in relation to the false and the actual norm can be described as pluralistic ignorance (Prentice and Miller 1996). The actual norm is that most hoodie wearing young people are not deviant nor involved in criminality. A SNA draws upon these differences in interventions to demonstrate misperceptions and that actual norms are more commonly held.

Developed by Berkowitz and Perkins (1987), a SNA has been used in various studies and implemented in prevention programs and interventions to change misperceptions and assumptions, addressing problem behaviors. A social norming approach incorporates the following principles:

  • Norms influence behavior yet norms are often misperceived (over or under estimated);

  • Misperceptions lead people to conform to a false norm (attitudes and behaviors are adjusted to conform to what is incorrectly perceived to be true);

  • Correcting misperceptions allows individuals to act in accordance with their actual beliefs, which are most often positive (adapted from Berkowitz 2012).

In the context of prevention work, a SNA reflects a paradigm shift as it focuses upon strengths and positives rather than aspects of problematic behavior. It does so in relation to a reference group; that is, a group of people that have a certain set of rules, and different groups of people have different rules (Bicchieri, 2006). For example, in another social norming project delivered by Social Sense, the RU Different? program, there has been evidence of consistent changes in the perception and behavior of young people around alcohol and tobacco use in pre- and post-test surveys, conducted before and after a digital intervention, highlighting that most young people do not engage in alcohol and tobacco usage (the reference group) (Social Sense 2018).

As a SNA locates people in their social environment, the impact of this is recognized in terms of inhibiting or inspiring healthy norms and behaviors. It also emphasizes the role that individuals play within their environments and communities in terms of prevention. For example, within the context of young people’s intimate relationships, the prevention of DVA can be facilitated by individuals if they recognize friends’ or families’ experiences as abusive and then act to prevent it or seek help to stop it (by telling a trusted adult for example). As such, bystanders play an important role (in what is termed bystander intervention). However, this is closely linked to norms in that if the individual feels that their recognition of abuse would not be shared by others within their social network, and their actions frowned upon, then they are less likely to act. In this way, as Berkowitz (2012: 5) notes ‘the correct perception of the norm’ is the basis of the effectiveness of the social norms approach as a prevention strategy and within a social network or community – the reference group (Bichierri 2006) - where anti-abuse norms are correctly perceived, individuals are more likely to act to prevent violence and abuse of this nature. Essentially, a SNA centers upon aligning behavior and values. This approach to prevention work has been described as ‘cutting edge’ (Berkowitz 2012: 6), but to be effective it requires a particular understanding of the community and the environment. For this reason, scoping work was undertaken in selecting the sites for the delivery of Change Up.

Change Up Project Design

The design of Change Up incorporated a multi-method approach drawing from both qualitative and quantitative methods of data collection. The project design reflected a social norms approach to the design and delivery of prevention programs (Berkowitz and Perkins 1987; Berkowitz 2012) with three core phases: Phase 1 the pre-test (baseline) survey; Phase 2 the intervention (workshops and campaigns); and finally, Phase 3 post-test (repeat survey). The Change Up program was delivered across two high schools (HS1 and HS2) in the North West of England; both neighborhoods were known to have high rates of DVA amongst the adult population.

Phase 1: Baseline Survey

Students accessed the baseline survey during school time (between May and July 2016) and surveys (n = 174 – the population of Year 9 students across both schools) were completed on an anonymous basis in order to minimize any social desirability bias. Parental consent was obtained by Social Sense and, additionally, consent from young people was taken with students informed that the survey was anonymous and that they could withdraw at any time (although any answers given would not be retrieved as these would not be identifiable). The sample was fairly evenly split in terms of gender with 50% (n = 88) self-identifying as female, 44% (n = 77) as male and 6% (n = 11) preferred not to say. Young people were aged 13–14 years old and there was a diversity of ethnic and cultural backgrounds (see Table 1). The sample was more diverse, in terms of the ethnicity, than the national population which reports 86% of the UK resident population as White British in 2011, compared to 68% in the baseline survey (ONS 2017), with 32% representing a number of different ethnic groups.

Table 1 Participant characteristics by gender and ethnicity
Full size table

The survey incorporated 30 (mostly) closed questions. Questions were constructed using SNT with consideration of key issues affecting this age group in relation to healthy and non-healthy relationships (some questions are not reported in this paper as these are not relevant to this topic). For example, some questions were scenario-based; ‘suppose someone hits their partner and says sorry afterwards, do you think this is wrong?’) with multiple choice answers (yes/no/sometimes) or a likert scale used to measure responses. Some attempted to measure gender bias offering scenarios in relation to perpetrating abuse towards males and females (see ‘Results’). Some questions used the pronoun ‘they’, rather than ‘he’ or ‘she’ to be gender neutral and inclusive of all genders and sexual identities, as well as relationship types (opposite-sex/same-sex). In the baseline survey, questions centering on experiences of concepts of DVA (coercive control, psychological abuse and emotional abuse), rather than attitudes, were included some with multiple-indicators (for instance, verbal abuse, controlling behavior, use of threats). In addition to the social norming messages extracted from the survey data, this data served to inform the design of the intervention.

Phase 2: Intervention Design and Delivery

Statistical data from the baseline survey allowed for comparison of the two school samples, and key themes emerged. These informed the design of the interventions which included one that was active (a workshop) and one that was passive (a poster campaign). A key theme of coercive and controlling behavior was identified as there was a greater divergence in norms and attitudes towards these than with physical abuse. Between October and December 2016 delivery of four workshops (two in each school) was undertaken during school hours involving 120 students. Participants were selected in the same way as Phase 1 with a reduced number due to absences from school on the day of the workshop delivery. The workshops combined the viewing and discussion of a short film about young people’s experiences of coercive control, followed by groupwork to design posters. Each poster was used in a peer-to-peer poster campaign. The poster campaigns were delivered in each high school during October 2016 and January 2017. Qualitative data was captured at the end of each workshop through the anonymous completion of pre-printed feedback postcards (with ‘something I’ll do differently after today is…’ (n = 60 completed) or ‘today made me think about…’ (n = 71 completed) (n = 2 were unusable). All feedback card data was anonymous.

Phase 3: Repeat Survey

Between October and December 2016, the repeat survey (post-test) was distributed to the secondary schools with completion by students (n = 171). The survey was open following the delivery of the intervention. See Table 1 for a breakdown of the respondents’ characteristics. Efforts were made to survey the same sample from the baseline survey and workshop participants, but a limitation of the sample is noted as respondents may not have not participated in the workshop and/or poster campaign. The survey was reduced to 20 questions, omitting those in the baseline survey which referred to behavior unrelated to DVA or relationships more generally, whilst including new questions constructed using the key themes contained within the poster campaign. Comparison was made between questions included in both surveys, with analysis of new questions pertaining to norms associated with the content of the intervention (workshop).

Evaluation Methodology

Evaluation Design

The service provider, Social Sense, undertook all data collection as this constituted part of the program delivery. This paper presents a secondary analysis of that data. A secondary analysis has facilitated an extended investigation moving beyond the initial reporting of results by Social Sense to their funding body. Moreover, conducting a secondary analysis of data is now a widely recognized methodology with the intention of extending the analytical depth of the original work; a process which Thorne (1994) terms as analytic expansion (Corti et al. 2005; Bulmer et al. 2009; Rogers and Ahmed 2017). Secondary narrative analysis is also useful when revisiting key themes within the context of contemporary theoretical frameworks (Elliot et al. 2015).

As this is a secondary analysis, it was not possible for controls to be implemented regarding recruitment or sampling. The data has been triangulated to produce a review of the findings but draws principally from the study’s survey data using descriptive statistics (Fisher and Marshall 2008) to provide a summary and picture of young people’s attitudes and norms. Quantitative data from the baseline and repeat surveys is presented to demonstrate if a measurable change had occurred. Survey data has been aggregated from two sources (in HS1 and HS2) but where there is a significant difference in results, this is reported below.

Qualitative data from the feedback cards took the form of concise statements in response to the prompts ‘something I’ll do differently after today is…’ or ‘today made me think about…’ (see below). A thematic approach was used for the analysis of this data and in the reporting of the triangulated data (Braun and Clarke 2006). This involved coding each statement to establish some general themes which were: healthy relationships; coercive and controlling behaviour; the recognition of domestic abuse; help-seeking and speaking out.

Ethical Considerations

Social Sense negotiated ethical approval directly from the participating schools. Participation was voluntary and parental consent was acquired (using an ‘opt out’ strategy) as well as the consent from young people on the day of survey completion or workshops. Participants were guaranteed anonymity, confidentiality and informed that they could withdraw their participation at any time. No official ethical approval was required for the secondary analysis in accordance with the University of Salford’s Ethics Policy.

Safeguarding Protocol

Safeguarding protocols were triggered if a workshop participant experienced distress. Additionally, due to the hidden nature of abusive relationships, a consequence of the delivery of Change Up was that several young people felt empowered to make disclosures. They received targeted support as safeguarding protocols were followed immediately: two disclosures were made following the workshop delivery and three disclosures were made in the survey free text boxes.

Results & Analysis

In this section, relevant data is presented with a summary of results; the first theme is physical abuse. Responses to questions on physical abuse are reported where it is possible to compare these from the baseline and repeat survey data. This highlights changes in young people’s norms and attitudes following the intervention (workshops and poster campaigns). Second, findings which specifically report the norms and attitudes held about emotional abuse are presented following those referring to coercive control. Some refer to questions in the baseline survey whereas questions were moderated in the repeat survey to align with the themes that emerged following the intervention phase. For instance, respondents in the repeat survey were asked if they had seen the poster campaign (31 replied yes, 28 replied no). Students were asked ‘did the workshop and/or posters make you think differently about how you want to be treated’ and 57% (n = 33) replied yes and 43% (n = 25) replied no. However, a more significant number of students reported that they thought differently about how they treated others as a result of attending the workshop or the poster campaign: 69% (n = 40) replied yes and 31% (n = 18) replied no. In the final section, findings pertaining to experiences are presented to provide further context to the sample.

Norms and Attitudes About DVA: Physical Violence

Emotions and Physical Violence

A question was asked that centered on the relationship between love and physical abuse with 79% of young people in the baseline survey indicating that hitting someone you love is wrong, rising to 87% in the repeat survey. Few respondents answered ‘no’ or ‘sometimes’. Therefore, most young people in both the baseline (n = 139) and repeat (n = 143) thought that you should not hit someone that you love (see Q.1 Table 2).

Table 2 Norms and attitudes: physical violence
Full size table

Young people were also asked to consider scenarios where they might experience negative emotions and in response to the question ‘if someone hits their partner because they really embarrass them; is it wrong?’, there was a considerable change from the baseline responses (72% answered ‘yes’) to the repeat survey (90% answered ‘yes). The proportion of the sample answering ‘no’ dropping from 11% to 3% and those responding ‘sometimes’ dropping from 17% to 7% (see Q.3 Table 2).

In order to ascertain a measure of norms around physical violence towards females, young people were asked about hitting a girlfriend if they were found to be irritating (‘getting on your nerves’) with little change across the surveys; there was a small rise of 7% (from 83% to 90%) answering that these was wrong. Similarly, reflecting on the use of physical abuse when angry, more than three-quarters of the sample (78%) felt that this was wrong, and following the intervention, again this rose in the repeat survey (to 89%) (see Table 2).

Behavior Norms and Using Violence: Contrition

The baseline survey indicated that over half (55%) of young people in the sample felt that it was wrong to hit their partner and then apologize with 15% indicating that this was acceptable within the realms of that relationship. Following the intervention, there was a considerable change in the repeat survey with a rise of 19% (from 55% to 74%) of respondents indicating that this was wrong. The number of young people who selected ‘no’ also reflected a sizeable reduction (from 15% to 4%) and with those who selected ‘sometimes’ a fair reduction (from 30% to 22%) (see Q.5 Table 2).

Behavior Norms and Using Violence: Cheating

Table 2 indicates that when asked whether it is wrong for a partner to hit their girlfriend if she had cheated (been unfaithful), almost three-quarters (71%) of young people in the baseline survey were in agreement with a rise to 82% in the repeat survey. In the baseline survey 15% of young people indicated that they felt that it was ‘sometimes’ acceptable to hit their girlfriend in this scenario, but this dropped by approximately half (to 8%) in the repeat survey. Respondents were also asked ‘suppose a boy cheats on his partner – do you think it is wrong for THEM to hit HIM’ with similar findings of 70% of young people agreeing that this is wrong in the baseline survey (indicating only 1% difference in attitudes taking the gender of the victim into account) and again this rose, to 80%, in the repeat survey. Similarly, 17% indicated that this was sometimes acceptable in the baseline survey, with a drop to 8% in the repeat survey. This indicates little difference in terms of gender bias in attitudes to using physical violence after experiencing a partner’s cheating behavior.

Behavior Norms and Using Violence: Alcohol Use

Responding to a scenario whereby a partner is physically abusive whilst under the influence of alcohol, responses were clearly delineated with the majority thinking that this was wrong (81% in the baseline survey, and 87% in the repeat survey). 5% (baseline) and 1% (repeat) reported that this was not wrong, with no change in the repeat survey of respondents who thought that this was sometimes acceptable (14%).

Behavior Norms and Physical Violence: Retaliation

Two questions centering on retaliation were gendered. Responses to these questions showed the most difference in terms of what is considered to be an acceptable behavioral norm in relation to using physical violence. The norm was stated as ‘a girl hits her boyfriend; do you think it is wrong for HIM to hit HER back?’ In the baseline survey, half (50%) of young people felt that it was wrong for a boy to hit his girlfriend in retaliation with the remaining participants spread equally between the attitude that it was sometimes OK (25%) and with the same proportionate (25%) indicating that it was acceptable behavior. There were, however, considerable attitudinal shifts of 16% in the repeat survey (from 50% to 66% indicating ‘yes, it is wrong’ with reductions in participants indicating that this was not wrong or sometimes wrong) (see Fig. 1). However, this is only two-thirds of the sample. Disaggregating the data indicated that in one of the high schools the percentage change was considerable rising from 47% reporting ‘yes, it is wrong’ in the baseline survey to 77% reporting ‘yes, it is wrong’ in the repeat survey.

Fig. 1
figure 1

Behavior norms and physical retaliation (by male)

Full size image

The sample was presented with the same scenario, but in switching the gender identity of the protagonists, so that the norm and question was ‘A boy hits his girlfriend; do you think it is wrong for HER to hit HIM back?’ In the baseline survey there was a modest difference in the proportion who considered this to be wrong (42%), but considerably more respondents (38%) considered that it was acceptable for a girl to hit a boyfriend if he has hit her and 20% felt that it was sometimes acceptable. This indicates a gender difference, in that it is more acceptable for a girl to use physical aggression in retaliation after being hit by her boyfriend. Again, there was a substantial difference in one of the high schools as 36% reported this to be wrong in the baseline survey but this rose to 72% in the repeat survey. Figure 1 positively illuminates the considerable shifts in attitudes between each survey.

Norms and Attitudes: Coercive Control and Psychological Abuse

An additional set of norming statements were included in the repeat survey to ascertain attitudes following the intervention (workshop and/or peer-to-peer campaign). The first statement was ‘emotional abuse is as bad as physical abuse’, with less than half (43%, n = 68) of students in strong agreement, but 51% in agreement (n = 80), 6% disagreed (n = 9) and just 1% strongly disagreed (1) (see Table 3). This suggests that following the intervention there is some convergence in terms of attitudes towards physical abuse and emotional abuse as equally harmful forms of DVA.

Table 3 Norms and attitudes: coercive control and psychological abuse
Full size table

Another statement that students were asked to consider in the repeat survey, concerned the scenario of coercive control in a partner’s choice of friends/clothes/where they go or what they do with just over half (53%, n = 83%) who strongly agreed that you should never control your partner in this way, 40% (n = 63) agreed, 6% (n = 9) disagreed) and 1% (n = 1) strongly disagreed. Both statements 1 and 2 indicate positive norms with 93% of young people in agreement.

To contrast the focus on coercive and controlling behavior, students were asked to comment on positive norms: ‘both partners should always trust and respect each other’ with approximately 98% (n = 155) in agreement. In addition, a statement based on the conception of bystander intervention (Berkowitz 2012) was proposed and the majority of the sample agreed that you should ‘speak out’ if you know someone was being abused with 96% (n = 152) in agreement.

Experiences of DVA: Coercive Control and Psychological Abuses

Young people reported their experiences and those that were not physically violent (but aligned with a definition of coercive control) were more frequently reported in the baseline survey (see Table 4) with 98% in HS1 and 96% in HS2 reporting that they had not experienced physical violence.

Table 4 Experiences: coercive control and psychological abuse
Full size table

Coercive Control

Table 4 demonstrates the reports of behaviors which, if experienced by young people, could be indicative of coercive control. Whilst more overt behaviors (threats, pressure) were not experienced commonly, less aggressive behaviors, which are akin to surveillance (see Q.2 and Q.3), were experienced more often albeit still by only a small proportion of the sample. For instance, students were asked ‘have any of your partners ever told you who you could or couldn’t see and where you could or couldn’t go?’ with 76% (n = 118/156) indicating that they had never experienced this, 6% (n = 10) had experienced this on a single occasion, 14% (n = 22) a few times with just 4% (n = 6) having often experienced this form of control. More young people had experienced a different form of surveillance as 16% had experienced being ‘constantly checked up on’ ‘a few times’, 6% (n = 10) often and 12% (n = 19) had experienced this once within a romantic relationship. Two-thirds (66%) had never experienced this.

Psychological Abuse

In response to the question ‘have any of your partners ever shouted at you, screamed in your face or called you hurtful names?’ 71% (n = 110) indicated that this had never happened to them. Of the remaining sample, 15% (n = 23) of students said that this had occurred once with 10% (n = 16) experiencing this a few times and just 4% (n = 7) reporting that this had occurred often. Focusing more on the content of verbal abuse, as body image is a sensitive issue for young people, respondents were asked about their experiences of receiving derogatory comments about their body and/or appearance. Responses were similar by each category with 79% (n = 123) and 80% (n = 124) never experiencing this with only 1% (n = 2) and 2% (n = 4) often experiencing this within the realms of an intimate relationship. Young people were also asked about whether any partners had made disparaging comments concerning their relationships with family and friends (see Table 4): 65% (n = 101) had never experienced this in relation to their friends, and 83% (n = 130) in relation to their family.

Taking a Social Norming Approach to Interventions and Behavior Change

The feedback cards collected qualitative data following each workshop (with n = 131 usable responses) and illuminated the beginnings of attitudinal shifts with several themes emerging. The data indicated a variety of ways in which participants considered that their behavior would change in terms of how they would treat partners, how they would look for signs of abuse and how they would seek help for their friends or for themselves. The latter point is critical as there are various studies which depict the ways in which people are prevented from recognizing their experiences as abusive and from seeking access from specialist service provision. This is even more difficult for victims who belong to marginalized groups (whether this is because of age, gender, sexuality, disability or culture) (Donovan & Hester, 2014; Rogers 2015). In terms of Healthy Relationships, one young person wrote that the workshop made them reflect upon what is acceptable in terms of another’s behavior as ‘…even if you love someone, don’t let them treat you badly’. Other young people indexed qualities that they would expect in a healthy relationship including equality, respect and trust.

Moving away from the dominant narrative of DVA (the public story) - that it is a problem of physical violence - was a key aim of the intervention. The workshops appeared to be successful on this count and evidently prompted a range of considerations in terms of what counted as healthy and what counted as Coercive and Controlling Behavior. One young person astutely described how abusive behaviors can escalate when commenting ‘[…] that it is not right to control people and controlling starts from small things and can get bigger’. Others reflected that after the workshop experience, something that they would ‘do differently’ is:

‘[…] don’t let anyone control your life’;

‘[…] make sure people don’t control me’;

‘[…] never let a boy control u (sic)’; and

‘[…] remember I’m in charge of my own self’.

The Recognition of Domestic Abuse was another theme as young people commented on being able to identify signs and how the workshops had made them think about ‘[…] how to handle domestic abuse and how to spot it and stop it’. The ways in ways abuse can push relationship boundaries was something that one young person referred to in order to ‘[…] make sure to know when a relationship has been taken too far’. Moreover, the workshop helped young people to acknowledge that DVA takes many forms and is not a singular incident as one young person noted that they were prompted to think about ‘…the patterns of domestic abuse and how it can impact your social life’.

The proposition that DVA can happen to anyone regardless of their background (gender, sexuality, age, ethnicity) was adopted by several young people (‘[…] domestic abuse doesn’t happen to just females but males as well’) and had wide-ranging and serious impacts. These elements of the abusive dynamic were considered in relation to Help-seeking and speaking out as young people wrote considered their own experiences, but also those of their friends:

‘[…] if I see abuse, I’ll help’;

‘[…] to share things with friends and don’t keep it in. And if you don’t think you’re happy in that relationships try to break up’);

‘[…] observe more around your friends and if they get treated weirdly by a gf or a bf (sic), report it’.

As such, the statements suggest the potential for behavioral change resulting from the interventions. In terms of attitudes, safety was referenced by several students as well as the need for more caution in relationships with peers. Responding to the prompt Today made me think about…, in the mature and pragmatic words of one young person ‘…to wait for some time and not sacrifice everything for a short-term relationship’.

Discussion and Concluding Comments

As argued earlier, it is widely accepted that DVA is entrenched in societies across the world and has far-reaching impacts (WHO 2017). Moreover, it is a problem that affects all societal members irrespective of gender, age, socio-economic background or other social characteristic. Given the nature of adolescence as a critical period of human development, it is disheartening, however, that DVA in young people’s relationships has received so little attention until more recently and what has occurred has been fragmented and lacked empirical evaluation (Stanley et al. 2015). However, there is now an emerging body of evidence to signify that it is a problem for a significant proportion of adolescents as well as adult populations (Stonard et al. 2014).

There is a danger, however, that if the problem of DVA in adolescents’ relationships is not adequately addressed through policy, practice and future research then the problem will continue to unfold, at best, and escalate, at worst. A social norming lens helps to explicate this further. Berkowitz (2012) notes how our norms and the ways in which we perceive the behavior of others, which are often incorrect (a misperceived or false norm), influences how we behave. This disconnect, or ‘pluralistic ignorance’ (Prentice and Miller 1996), plays a role in dysfunctional relationship dynamics as it is based on the premise where individuals might have a different attitude or norm to their peers, but then behave in the same way. Highlighting the problem of pluralistic ignorance in terms of young people’s relationships, it is evident that an individual may consider that coercive control is wrong and harmful, but they may tolerate that behavior as their perception is that their peers’ relationships are similarly coercive and controlling. As such, pluralistic ignorance can result in dysfunctional, risky and harmful practices. An implication for research, policy and practice, therefore, is to explore further the value of a SNA to the field of adolescent relationship abuse to substantiate and add to the emerging evidence-base in this regard.

One element of the dominant narrative of DVA, the public story (Donovan and Hester 2014), integrates the gendered notion that DVA is commonly understood to be physical violence perpetrated by men against women. The Change Up survey highlighted how the majority of young people reported that they had not experienced physical abuse from a partner (whatever their gender), contrary to empirical literature in this area (Barter et al. 2009; Fox et al. 2013), albeit the sample size was modest. In addition, the survey did not ask whether respondents had other experiences (as a perpetrator or witness in their homes) and so it is not possible, in the analysis of Change Up data, to make connections between experiencing DVA, directly or indirectly, and the norms expressed by young people.

Yet in terms of attitudes, there were interesting findings in terms of norms and gender bias. One question centered on retaliation after being hit by a partner and it was asked in two ways in order to uncover any bias in norms held about the social acceptability of the behavior of boys and that of girls. In this instance, young people were asked to consider ‘suppose a girl hits her boyfriend, do you think it is wrong for him to hit her back?’ and vice versa. These questions resulted in the lowest score for affirmative responses overall as in the baseline survey as 50% responded ‘yes’ it is wrong for a boy to hit a girl back, and 42% replied it is wrong for a girl to hit a boy back. As with all the other recurring questions, the repeat survey shows attitudinal change with 66% and 59% respectively, but again these were the answers which had the lowest responses in which young people agreed that physical abuse was socially unacceptable in that particular scenario.

Whilst an outlier in relation to most of the other results, this finding is congruent with existing empirical data which has found gendered differences in terms of attitudes towards physical violence (Fox et al. 2013). Fox et al. also found that a considerable proportion of young people regard violence from women to men as more socially acceptable than violence perpetrated by men against women. This reflects a gender norm that persists although Fox et al. (2013) point out that in their sample, those who had experienced DVA (either as a victim, witness or perpetrator) were more likely to consider that hitting a partner was acceptable than those who had no prior DVA experiences. Whilst most of the young people who participated in Change Up had not experienced physical or psychological maltreatment in their own relationships, a finding was that the majority of young people (94%) agreed that emotional and physical abuses are equally harmful. This is heartening as it suggests a departure in young people’s attitudes from the entrenched notion of the public story of DVA (Donovan and Hester 2014).

The Change Up data suggests that the young people see the acceptability of some abuses as contextual. This is highly troublesome as it lends itself to the types of behavior described within the delineation of coercive control; for example, where abusive partners can be manipulative (by frequently saying, for example, ‘I’m sorry. I’ll never do it again’) and effectively exploiting naivety, goodwill and the desire to protect a perpetrator, as well as exploiting the care and love that might exist for a person (Barter 2014). Yet when certain behaviors are explained away by context (it is acceptable to hit someone if they hit you, for example) then this is clearly problematic and can result in the normalization of violence, or a lack of recognition of particular (non-physical) behaviors as abusive. In turn, this can prevent help-seeking and action for change. It can result in a cycle of abuse that can be hard to break.

The concept of coercive control has gained momentum in recent years since the publication of Stark’s (2007) original text, in which he detailed the ways in which the impact of DVA is augmented by gender inequality and how victims are controlled and terrorized in their daily lives. Coercive control is an insidious form of abusive behavior as it can build over time and perpetrators use a variety of means to manipulate, exploit and control. The Change Up project had a number of impacts in terms of raising awareness in young people’s understanding of coercive and controlling behaviors. More importantly, the project enabled some key changes in norms and attitudes about coercive control. To some extent these were evident in the survey data, and the feedback cards demonstrated that the interventions (workshop and poster campaign) provoked considerable reflection about the meaning of a ‘healthy relationship’ in contrast to an unhealthy one; where an unhealthy relationship consisted of behaviors conducive to coercive control (rather than other forms of DVA such as physical or sexual abuse). This is significant as it is coercive controlling behaviors that can be difficult to spot, both in terms of recognizing this in other people’s experiences as well as one’s own. This highlights a further policy and practice implication as being the need to teach about the interconnections between interpersonal violence and gender inequality (resulting from and in an imbalance of power) which lies at the heart of the problem of DVA (Stark 2007).

The findings suggest that there is a policy and practice need for programs, such as Change Up, to be embedded within personal, social and health education (PSHE) programs to enable norms and attitudes to be more firmly rooted to the notion that any interpersonal violence or relationship abuse is unacceptable. Across the UK there are a varied array of programs that reinforce these messages and work with young people to address DVA, but mostly these have not been rigorously evaluated, and support from schools for service delivery has been described as patchy and inconsistent (Stanley et al. 2015). Indeed, in 2008 Coy et al. (2008) described prevention work as being ‘the weakest part of the UK responses to violence against woman’, and Stanley et al.’s (2015) comprehensive mixed knowledge scoping review suggests that this continues to be the case, or more likely, that we do not know what works best. Therefore, it is hoped that there will be an adequate investment in prevention projects, like Change Up, following the Government’s pledge to ensuring relationship education in schools in England and Wales as this would indicate a commitment to a sustainable change in the norms and attitudes of young people in relation to DVA.

In conclusion, this article has presented the data from a prevention program, using social norming theory, designed to address the topic of domestic abuse and healthy relationships with young people. In doing so, we have demonstrated the value of a SNA to changing the attitudes and norms of young people as every question that was included in both the baseline and repeat surveys showed a measurable change. More generally, there is a growing body of evidence that illuminates the centrality of social norms in the development of positive behavior (Elsey et al. 2015; Sheikh et al. 2017), whilst here we provide evidence of its value in relation to healthy relationship norms and behaviors. Moreover, it is widely accepted that the ways in which to successfully address DVA when it presents in relationships, in adolescence or adulthood, is through an approach which targets norms (Stanley et al. 2015; Cislaghi and Heise 2017) within a school environment as ‘shifting social norms in the peer group [is] a key mechanism of change’ (Stanley et al. 2015: v). As such, it is hoped that future policy and practice for young people around healthy relationships is social norm driven, addresses all recognized forms of DVA (including ones using digital technologies) and aims to uncover a deeper level of understanding about the contextual nature of differing norms and attitudes held by young people.

Limitations

The Change Up data does not support the existing prevalence data, which indicates that a considerable proportion of young people have experienced DVA, as the sample was small and the questions focused more on norms, rather than experiences (congruent with a SNA). Moreover, whilst it is not possible to identify all factors (including those external to the project) that may have impacted on the changing norms and attitudes of young people involved in Change Up, the short time between the execution of pre- and post-surveys is taken as a positive in terms of offering a limited time period in which young people were open to external influences.

As this paper reports a secondary analysis, there are limitations in terms of the conclusions that can be made. There are methodological, theoretical and conceptual issues pertaining to the delivery of Change Up as noted above. For example, more demographic information about samples is not available, nor further understandings in terms of how, if at all, concepts were operationalized for the survey respondents. It is also acknowledged that the data was sourced from two schools in one geographical site and thus there are limits in terms of extrapolating findings to the wider population of young people in the UK or beyond.