Preventing violence against women and girls - what works: evidence summary

Footnotes

1 See the full list of effectiveness classifications in Annex C.

2 Accompanying this publication, there is a standalone key findings paper available here, alongside a summary document entitled Effective Investments: A Summary of What Works to Prevent Violence Against Women and Girls for Policy and Practitioners available here.

3 Much of the available high-quality evidence on the effectiveness of primary interventions to prevent VAWG is from high income countries (e.g. USA and Canada). It is important to account for cultural context in applying these interventions in a Scottish context.

4 This WHO (2019) report also highlight the following barriers to successful interventions to prevent and/or reduce VAWG: limitations on women's autonomy; children exposed to violence; social norms that perpetuate male power; inadequate services; inadequate legal and social protections for women; lack of political will and resources; under-resourced women's organizations or movements.

5 For more information see Scottish Government (2020) 'Domestic abuse and other forms of violence against women and girls (VAWG) during COVID-19 lockdown for the period 30/3/20 - 22/05/20'

6 Factors which might facilitate effectiveness or act as a barrier to effectiveness

7 This report does not cover civil protection orders (such as interdicts) or the effect of using criminal justice measures (see Annex E on out of scope interventions). For more information, see Brooks et al. (2014): ‘Violence against women: effective interventions and practices with perpetrators, a literature review’, The Scottish Centre for Crime and Justice Research Report.

8 According to the Scottish Crime and Justice Survey (SCJS), since the age of 16, women were almost twice as likely as men to have experienced partner abuse (20.0% and 10.9%, respectively) (SCJS 2016/18). These results also show that women are more likely than men to have experienced both serious sexual assault and less serious sexual assault (SCJS 2016/18). Likewise, the 2018-19 domestic abuse recorded by the police in Scotland statistics show that where the victim’s gender was known, the clear majority of victims in 2018-19 (83%) were female. Around four out of every five incidents (82%) of domestic abuse in 2018-19 had a female victim and a male accused. In 2018-19, 16% of domestic abuse incidents involved a male victim and a female accused.

9 Much of the available high-quality evidence on the effectiveness of primary interventions to prevent VAWG is from high income countries (e.g. USA and Canada). It is important to account for cultural context in applying these interventions in a Scottish context.

10 see also in this Scottish Government report series, What Works to Reduce Crime (2014)) and What Works to Reduce Reoffending (2015)

11 Comparable to Scotland.

12 For the out of scope list see Annex E.

13 Within the WHO (2019) RESPECT framework, there are interventions that focus on secondary prevention interventions (i.e. those focused on reducing ongoing violence). Due to this report’s focus on primary prevention these areas have been identified as out of scope (see Annex E).

14 The focus on these types of VAWG reflects the specific nature of the violence, and the limited evidence available on prevention-focused interventions that look at preventing HBV or FGM as an explicit outcome. However, as signposted to, other interventions presented within this report seek to change social attitudes and behaviours with the broad aim of preventing various forms of VAWG.

15 For a similar approach to assessing available evidence on violence prevention, see Fulu et al. (2014) report on what works to prevent VAWG.

16 Interventions that have been identified as out of scope for this report are listed in full in Annex E.

17 The WHO ‘s (2002:15) World report on violence and health provides an overview of what constitutes primary, secondary and tertiary prevention. These definitions are summarised in Annex A.

18 For more on prevention within a public health approach, refer to ScotPHN’s (2019) Violence Prevention Framework.

19 Annex D provides detailed information about each of these evidence classifications.

20 Within this review, the interventions presented do not fall into the ‘no effect’ or ‘negative effect/potentially harmful’ categories. However, these have been included here to demonstrate the categories used across this work.

21 By contrast, a no effect classification (of which there are none within this report): has strong or moderate evidence available with no evidence of effect (positive or negative) was found for preventing VAWG.

22 The exception is the section on what works to prevent honour-based violence – this reflects the limited available evidence on interventions that are effective in preventing this form of VAWG.

23 Additional contextual information can be found in Annex F.

24 These characteristics of effective interventions are also present within domestic abuse and sexual violence interventions discussed in sections 2 and 3 of this report

25 This WHO (2019) report also highlight the following barriers to successful interventions to prevent and/or reduce VAWG: limitations on women's autonomy; children exposed to violence; social norms that perpetuate male power; inadequate services; inadequate legal and social protections for women; lack of political will and resources; under-resourced women's organizations or movements.

26 While not directly linked to prevention interventions, it is important to note the broader context of barriers to support women who have experienced a form of GBV.

27 Examples of these protected characteristics include: minority ethnic women and girls; refugees and asylum seekers; disabled women and girls (including those with learning disabilities); LGBTI people; and women at different ages and stages of life (including older women) (COSLA, 2020)

28 See for example Public Health England (2015): Disability and domestic abuse: risk, impacts and response report

29 For more information on this barrier in a Scottish context see for example: Shakti Women’s Aid and Hemat Gryffe Women’s Aid,

30 Partner abuse in the SCJS is defined as ‘any form of physical, non-physical or sexual abuse, which takes place within the context of a close relationship, committed either in the home or elsewhere. This relationship will be between partners (married, co-habiting or otherwise) or ex-partners.’ This definition is consistent with the definition adopted by Police Scotland in recording domestic violence.

31 See also COSLA (2020) Coronavirus (COVID-19) Supplementary National Violence Against Women Guidance

32 For more information see Scottish Government (2020) 'Domestic abuse and other forms of violence against women and girls (VAWG) during COVID-19 lockdown for the period 30/3/20 - 22/05/20' available at: https://www.gov.scot/isbn/9781839608292

33 While this WHO report focuses on both high and low income countries, many of the interventions cited have been identified in relation to high income countries.

34 Referred to as intimate partner violence (IPV) within the WHO report, DA is the tem used within this publication.

35 See WHO description of the ecological model

36 Graham et al. (2019) note that there is limited evidence available on what works with regards to programmes and interventions targeted at men and boys.

37 Awareness raising campaigns are discussed further and are classified as ‘inconclusive’ due to limited evidence on violence-related outcomes.

38 Research indicates that BME victims-survivors may experience a longer wait than white victims-survivors (Femi-Ajao 2020).

39 The Domestic Abuse Scotland Act (2018) extends to cover coercive and controlling behaviours as a form of domestic abuse. Of the primary interventions presented below, those that focus on attitudinal and/or behavioural change relating to VAWG with younger people may have an impact in preventing coercive and controlling behaviours, although whether interventions specifically targeted these behaviours was not always clear from the available literature and the evidence linked to this explicit outcome is limited and could be explored further.

40 While this data refers to both genders, the SCJS reports that women were more likely than men to experience partner abuse since the age of 16 (18.5 per cent) compared to men (9.2 per cent).

41 The SCJS partner abuse figures combine data collected from 2016/17 and 2017/18 survey years. This is referred to as 2016/18, and the data can be found within the SCJS data tables. For more information on the SCJS partner abuse figures, see the 2017/18 SCJS main findings report.

42 While they are not mutually exclusive, stalking and sexual harassment are not included within this report in detail due to limited available evidence (see Annex E for out of scope list).

43 Serious sexual assault: experience of one or more of the following - Forced sexual intercourse; Attempted forced sexual intercourse; Forced other sexual activity (for example, oral sex); Attempted forced other sexual activity. These terms are used within the SCJS report.

44 Less serious sexual assault: experience of one or more of the following - unwanted sexual touching; indecent exposure; and sexual threats. The terms ‘serious sexual assault’ and ‘less serious sexual assault’ are used for ease of reference and do not relate to the seriousness of the impact on an individual.

45 While not directly linked to prevention interventions, it is important to note the broader context of barriers to support women who have experienced a form of GBV.

46 While highlighted by the WHO (2019) framework for VAWG prevention, there is limited evidence about what works within transport settings to make environments safe. See the International Transport Forum’s (2018) report on Women’s Safety and Security: A Public Transport Priority for further information.

47 Children aged 11-14 years old

48 These temporary building-based restraining orders are also known as Respecting Boundaries Agreement (RBA), (Taylor et al., 2012). Details can be found in Stein (2010:10) Shifting Boundaries: Lessons on Relationships for Students in Middle School.

49 These examined outcomes refer to peer physical, sexual, sexual harassment victimisation or perpetration, also looking at these in a date context.

50 Children aged 11-12

51 The researchers measured adolescent relationship abuse and sexual harassment through participant responses to questions on whether they have performed a specific act of physical violence against a partner or peer (e.g. kicking, pushing, hitting).

52 Such limitations around self-reporting are also applicable to other interventions to prevent various forms of violence (including VAWG).

53 Similarly, the Safe Dates intervention did not report results separately by sex of the victim or the perpetrator; therefore it is not clear whether the effect was similar for boys and girls (Ellsberg et al. 2015:1557).

54 A bystander is “somebody who observes an act of violence, discrimination or other unacceptable or offensive behaviour” (Powell, 2011:8). A bystander can therefore be a friend, classmate, team-mate, colleague, relative or a stranger. Bystander approaches aim to encourage ‘active’ or ‘prosocial’ bystanders to intervene in response to violence incidents (Powell, 2011).

55 These behaviours are centred on bystanders actively intervening to prevent or end violent behaviours among peers.

56 For accessible overviews of the MVP programme, see Scottish Violence Reduction Unit website on MVP or Mentors in Violence Prevention (MVP) via Restorative Justice Coventry

57 Within their evaluation Williams and Neville (2017:24) noted that mixed-gender sessions were raised by participants; demonstrating an “appetite to hear the opinions and perspectives of the other gender”.

58 Berkowitz, A.D. (2009). Response Ability: A complete guide to bystander intervention. Chicago: Beck & Company.

59 Within MVP, a bystander is defined as anyone who sees, hears, or has knowledge of an incident, but is not directly involved.

60 Bystander programmes, such as MVP programmes, are one of a number of interventions used within secondary school settings in Scotland, and internationally (see Lombard and Harris, 2017).

61 Peer mentors are individuals who are “older or more senior from the same peer group” as the mentees (Williams and Neville, 2017:4)

62 ‘Prototypical group norms’ refers to those that the researchers would expect to see and/or had observed within this social environment.

63 See Annex B on reviewing and assessing evidence.

64 Mentors in Violence Prevention (MVP) is being implemented in Scottish schools, and some evaluations have been carried out to date. Moreover, in Scotland, the Violence Reduction Unit are also part-funding a PhD examining the effectiveness of MVP in a Scottish context, which will enhance the existing evidence base. Results from this PhD research are expected in late 2020.

65 The programme was piloted within these schools during the 2012/2013 school year, with the evaluation following in early 2013.

66 UNESCO (2016) Global guidance on addressing school-related gender-based violence: https://www.unwomen.org/en/digital-library/publications/2016/12/global-guidance-on-addressing-school-related-gender-based-violence

67 A process that ensures that the evidence is robust and high-quality.

68Sexting is defined as: “sending sexually explicit and (partially) nude pictures of themselves in inter-personal conversations via the internet or smartphone” (Van Oosten & Vandenbosch, 2017). Active sexting was asked about in the survey as: “Have you asked someone to send naked pictures of them to you?”; “Have you sent naked pictures of yourself to another through text, email, or SnapChat?”. Passive sexting is defined as: “Have you been asked to send naked pictures of yourself through text, email, or things like SnapChat?”; “Has anyone sent you a naked picture without you asking?”.

69 See Annex B and Crooks et al.’s (2019) paper on Preventing gender-based violence among adolescents and young adults: lessons from 25 years of program development and evaluation

70 Graham et al. (2019) note that there is limited evidence available on what works with regards to programmes and interventions targeted at men and boys.

71 For more information, see Stoker et al. (2015)

72 Bystander approaches – such as those used within MVP programmes have also been discussed above.

73 Rape myths are defined by Burt (1980:217) as “prejudicial, stereotyped, or false beliefs about rape, rape victims, and rapists”.

74 As Gainsbury et al. (2020:1) note: “campus-based research has found bystander programmes show promise as effective primary prevention of sexual violence. However, evidence regarding domestic violence and abuse bystander prevention specifically, and in community settings generally, is still in development”.

75 This refers to statistically significant change

76 DVA refers to domestic violence and abuse

77 As of December 2019, more than 2600 senior pupils are trained as mentors, with the mentors going on to deliver around 3500 lessons a year (Education Scotland, 2019).

78 see Annex B on implementation fidelity for applying interventions in different contexts.

79 See also Eriksen (2015): The Mentors in Violence Prevention Leadership Training at California State University, Long Beach programme evaluation

80 Backlash effects refer to where prevention efforts may have the opposite outcome to that intended, for example the entrenchment of the attitudes that the programme seeks to change (see Gainsbury et al., 2020).

81 Lombard and Whiting (2018:38) note that so-called gender neutral policies and interventions to prevent VAWG are in danger of averting the focus “structural inequalities in society and thereby avoids addressing the very inequity that the initiative was seeking to address. It is an approach detrimental to women and children but one that equally swerves from engagement with the negative impact gender constructs can have on some men’s well-being and ignores the gendered nature of domestic violence for the minority of men who do experience it”.

82 Family Violence Information Disclosure Scheme (FVIDS) in New Zealand, based on UK initiative. There are currently no evaluations of this scheme available.

83 This refers to the criteria used to identify ‘adults at risk’ based on the Adult Support and Protection (Scotland) Act 2007.

84 Police Scotland has a statutory power under Section 32, Police and Fire Reform (Scotland) Act 2012 to disclose information where it is necessary to prevent and detect crime. It is on this statutory power that the authority of Police Scotland to disclose information under the DSDAS rests. The basis for disclosure of information is recorded within the DSDAS process. The risk assessment and decision making record gives clear justification for every disclosure, as underpinned by existing legislation.

85 A personal safety plan refers to the plan that a victim/survivor of domestic abuse creates in advance, thinking about how they might respond to different situations (including crisis situations) (see Women’s Aid, no date).

86 Duggan’s (2018) empirical research entitled ‘victim hierarchies in the domestic violence disclosure scheme’ highlights barriers and limitations associated with the implementation of this scheme.

87 According to this report, ‘pressing need‟ is one of the criteria the decision-making forum must use to justify the decision to make a disclosure.

88 Page 19 of this report details the approach of practitioners; through their consideration of ‘pressing need’ on a “case-by-case basis and used their professional judgement to assess the ‘pressing need’ for disclosure”.

89 See a forthcoming Scottish Government report on What Works to Prevent Youth Violence

90 See evidence presented below.

91 Safe Dates programme facilitators receive “between 1 and 2 days of training, depending on the implementation plan, and community service providers typically receive 3 hours of training” (Crooks et al. 2019:33).

92 The Safe Dates programme also aims to reduce sexual violence against women and girls.

93 This programme has been found to be effective in both preventing dating abuse perpetration, and reducing victimisation among teens already involved in dating abuse (see Violence Prevention Works, Safe Dates summary, no date).

94 Randomised control trial. See also Annex C on assessment of evidence.

95 Adolescents aged approximately 13-15 years old

96 US eighth and ninth grade students; aged approximately 13-15 years old

97 Often used in health settings and understood as the ‘gold standard’ of evaluations, RCTs can be a powerful tool. RCTs allow the gathering of robust data on the impact of a programme through the comparison of a control group and a group participating in an intervention. A cluster randomized trial uses the same principles, using a particular group (e.g. school setting) as a ‘cluster’.

98 According to Lundgren and Amin (2015: 543), dating violence is used within United States and Canadian scholarship to refer to “physical or sexual violence occurring in the context of a relationship that is neither marriage nor a long-term cohabiting partnership”. Within a Scottish context, the term ‘domestic abuse’ covers these elements.

99 Young people in Canada aged approximately 14-17 years old.

100 Aged 14-15 years old.

101 See also bystander interventions evidence presented earlier in this report.

102 As noted earlier there are overlaps between youth violence prevention (primary) interventions and those aimed at preventing VAWG. More information will be published within a forthcoming Scottish Government report (written by the Scottish Violence Reduction Unit) on What Works to Prevent and Reduce Youth Violence.

103 De Koker et al. (2014:12) also suggest that further research is required to determine whether “a gender neutral approach works better than a focused approach targeting males and females separately”.

104 See DA summary on terminology around domestic abuse, including IPV.

105 See details of these approaches on report sections on Safe Dates, Fourth R and Shifting Boundaries.

106 This intervention is discussed in detail in the section on interventions focused on modifying unsafe physical school environments.

107 Report reviewing evidence on preventing intimate partner and sexual violence against women.

108 See also the Equally Safe delivery plan: year two update report (2019) for additional details on Equally Safe in higher education settings.

109 Examples of common myths around rape and sexual violence can be found on Rape Crisis England and Wales website

110 This project was based in secondary school settings and youth settings (Rape Crisis Scotland online, no date). See also section above on school-based programmes promoting equal relationships; the interventions presented there share similarities with this intervention (i.e. education as a prevention strategy).

111 An evaluation of Equally Safe at School is currently being carried out across a 21-month period between 2019 and 2020 by academics at The University of Glasgow.

112 More information on what these components involve can be found at Rape Crisis Scotland online (no date).

113 See Vladutiu, Catherine J, Martin, Sandra L & Macy, Rebecca J, 2010. College- or University-Based Sexual Assault Prevention Programs: A Review of Program Outcomes, Characteristics, and Recommendations. Trauma, violence & abuse, 12(2), pp.67–86 for more details

114 Also referred to as ‘social marketing’ or ‘social norms marketing’ (Paluk and Ball, 2010).

115 For more on social marketing/social norms marketing campaigns, see Paluck, E.L., & Ball, L. (2010). Social norms marketing aimed at gender based violence: A literature review and critical assessment. New York: International Rescue Committee.

116 However, there are commonalities with some of the approaches and interventions presented in this report to tackle DA and GBV more broadly.

117 FM refers to Forced Marriage

118 See Safe Lives FAQ on MARACs for further information and resources

119 MARACs have been identified as out of scope within this report. However, further information is available in Annex E: out of scope list.

120 MARACs have been identified as out of scope within this report. However, further information is available in Annex E: out of scope list in the main report.

121 Again, the focus on low-income countries within this research may limit how comparable it is with high income countries such as Scotland

122 For more information see Scottish Government (2020) 'Domestic abuse and other forms of violence against women and girls (VAWG) during COVID-19 lockdown for the period 30/3/20 - 22/05/20'

123 ScotPHN (2019:9) note that primary prevention “is distinct from ‘secondary prevention’, intervening to prevent the further escalation of violence where it has not been prevented, and ‘tertiary prevention’ focused on care, rehabilitation and reintegration, post-violence”.

124 These criteria have been informed by Fulu and Kerr-Wilson (2015) What works to prevent violence against women and girls evidence reviews

125 The evidence presented within this report is primarily from quantitative research published in peer-reviewed publications and organisational reports, however qualitative research is identified as important in understanding the effectiveness of an intervention.

126 This review focuses upon studies from high income countries as they are the most directly comparable to Scotland. As such, low- and middle-income countries are not included.

127 This review is limited by the fact that we only drew upon evidence published in the English language

128 Within this review, the interventions presented do not fall into the ‘no effect’ or ‘negative effect/potentially harmful’ categories. However, these have been included here to demonstrate the categories used across this work.

129 By contrast, a no effect classification (of which there are none within this report): has strong or moderate evidence available with no evidence of effect (positive or negative) was found for preventing VAWG.

130 Page 51 of this report discusses limitations around having degendered prevention approaches (e.g. the Green Dot programme). Such programmes should be understood in relation to the gendered approach that the Scottish Government takes to tackling gender-based violence through the Equally Safe strategy.

131 Examples of these scenarios can be found at MVP Strategies online.

132 The interventions listed as out of scope below are predominantly identified as secondary or tertiary prevention (see section 3.2 on prevention level), and have thus been deemed out of scope. For more information regarding primary prevention, see ScotPHN (2019) Violence Prevention Framework.

133 For more information, see Bates, L. & Hester, M. (2020): No longer a civil matter? The design and use of protection orders for domestic violence in England and Wales, Journal of Social Welfare and Family Law, 42:2, 133-153, DOI: 10.1080/09649069.2020.1751943

134 See also Brooks et al. (2014) Violence against women: effective interventions and practices with perpetrators: a literature review for a relevant review of perpetrator-focused interventions.

135 There is some available evidence about use of shelters to reduce harm within Jewkes (2014:17) What works to prevent violence against women and girls? However, they highlight challenges around researching the effectiveness due to the self-reported data rather than robust evaluations, as well as evidence of the potential for re-victimisation by the abusive partner after a period in shelters.

136 For more information, see this Department for Communities and Local Government and University of York report (2010) entitled: The effectiveness of schemes to enable households at risk of domestic violence to stay in their own homes.

137 See also Brown, K.E., Bayley J.E. and Baxter, A. (2015): Evaluation of the Sexual Assault Referral Centre (SARC) based at George Eliot Hopsital, Nuneaton, Centre for Technology Enabled Health Research, Conventry University [accessed 17.06.19]

138 Commercial sexual exploitation is defined by the Equally Safe strategy (2016) as: “activity which includes prostitution, lap dancing, stripping, pornography and trafficking” and other forms of commodification for women’s bodies for sexual purposes (Brooks et al., 2014; Wilson et al., 2015).

139 As with HBV and FGM interventions, primary prevention interventions can be effective or show promise in shaping social norms, attitudes, and behaviours as a broader approach to preventing VAWG from happening.

140 This report looks at school-based interventions; however, these are implemented within secondary school environments predominantly.

141 See also Hetherington (2020) Ending childhood adversity: a public health approach, Public Health Scotland for evidence on prevention and public health approaches to ending childhood adversity.

142 This WHO report (2013a) examines the evidence of effectiveness of interventions aimed at preventing alcohol-related violence. Evidence from this report shows a complex relationship between harmful use of alcohol and domestic abuse for both victims-survivors and perpetrators of domestic abuse (referred to as intimate partner violence within this report).

143 Sources include: (i) Police Recorded Crime, (ii) Scottish Crime and Justice Survey (SCJS), (iii) Emergency Hospital Admissions due to Assault and (iv) Criminal Proceedings.

144 This publication does not include evaluations or information about the effectiveness of these projects.