Overview and Analysis of Research Studies on Gender-based Violence Among UK University Students and Staff Working paper 1 from the GW4 project “Investigating GBV Intersectional (Dis)Advantages and Legal Duties - A Scoping Study of UK Universities” The SCuLE Centre Working Paper No. 21-1 Cassandra Jones1 Olivia Smith2 Tina Skinner3 Geetanjali Gangoli4 Rachel Fenton5 1 University of Winchester 2 Loughborough University 3 University of Bath 4 Durham University 5 University of Exeter Correspondence regarding this article should be addressed to Cassandra A. Jones, University of Winchester, Sparkford Road, Winchester, Hampshire, SO22 4NR, UK. Email: Cassandra.Jones@winchester.ac.uk mailto:Cassandra.Jones@mmu.ac.uk The authors would like to thank academics and university support services who provided research studies that were included in this overview and sign posted to additional studies. Contents 1. Introduction ................................................................................................ 3 2. Methodology .............................................................................................. 4 2.1 Search strategies and inclusion criteria ................................................. 5 2.2 Data extraction ...................................................................................... 9 2.3 Analysis ................................................................................................ 9 3. Findings ..................................................................................................................... 9 3.1 Study selection ...................................................................................... 9 3.2 Design and methods ........................................................................... 10 3.3 Sample ................................................................................................ 13 3.4 Quantitative tools ................................................................................ 16 3.5 Results ............................................................................................... 19 4. Limitations of studies .............................................................................. 23 5. Summary and Conclusions ...................................................................... 23 1. Introduction New movements (e.g. #MeToo) brought to the forefront of public consciousness the widespread prevalence of gender-based violence (GBV) and the scale of the impact in women’s everyday lives. GBV is understood to be any act of violence and abuse that disproportionately affects women and is rooted in systematic power differences and inequalities between men and women (Hester and Lilley, 2014). There is an international body of evidence indicating that universities are significant sites for GBV (DeGue, 2014). Understanding the prevalence, characteristics, and impacts of GBV among university students and staff is essential for universities to effectively prevent and combat it. In the UK, a limited number of studies have started to address this gap (e.g. NUS, 2011) but they have not been guided by a contextualised theoretical framework nor have they been reviewed and synthesised to create an overall picture of what is known and not known about GBV. There has been increasing pressure on universities to take action to prevent and combat GBV, including limited legal guidance for investigations (e.g. UUK, 2016; Women and Equalities Committee, 2018). However, no overarching work has been published examining UK universities’ legal obligations on prevention and response in the round and comparative legislative duties in other jurisdictions, nor on the role of legal duties as a facilitator to disrupting GBV and holding universities to account. The role of law within prevention strategies is unknown in the area of GBV. Implementation of prevention and response strategies has been ad hoc and piecemeal (or non-existent) by universities in the light of lack of accountability or enforceable duties An ecological theoretical framework specific to UK universities has not been constructed. Theories have been developed for universities in the U.S. but the history, composition, geography, and culture of UK universities is different (Phipps and Smith, 2012; Stenning et al., 2012). Due to these differences, a theoretical framework relevant to UK universities is needed to guide studies and contextualise findings. As a starting point, the proposers will use and develop Hagemann-White’s et al.’s (2010) framework developed for the European Union. This framework is the most researched, demonstrated and holistic model existing to date. The framework used an ecological model to identify and categorise factors facilitating and scaffolding GBV, including policies, sanctions, redress and implementation of laws, to provide nation states with a framework which have subsequently been adopted to develop and implement policies that would more effectively prevent and combat GBV. The proposed research will develop this model tailored to UK universities, using a more sophisticated understanding of intersectional (dis)advantage (such as ethnicity, gender, sexuality, disability, class, age), men and masculinities, peer-group support for violence, environmental time-space and power relations, and legal duties in prevention and response. In order to fill these gaps, this project aimed to: • Provide an overarching picture of research on GBV in UK universities (Workstream 1) • Provide an overview of UK universities’ legal obligation on prevention and response and the role of legal duties as a facilitator to disrupt GBV (Workstream 2) • Develop a theoretical, ecological model specific to GBV in UK universities that will guide and contextualise future research (Workstream 3) This working paper describes the methodology and results from work undertaken for Workstream 1. The primary aim was to undertake a critical analysis of research on GBV among UK university students and staff to identify the strengths and weaknesses of the current state of knowledge and emerging findings, as well as to make recommendations for future research. This work will feed into the development of an ecological model specific to UK universities (Workstream 3). The focus was not on students’ views on university responses to GBV but on the research processes used by studies to investigate students’ and staff experiences and associated impacts. As such, the key objectives of the research were to gain detailed knowledge on the approaches used in research studies across the UK, with particular emphasis on the methods, research tools, and findings. The following questions guided this work: 1. What types of studies are being conducted across the UK? 2. In particular what methods, designs, and research tools are being used? 3. Are the tools and analyses sensitive to gender and the intersection of gender with other social positions, such as ethnicity, sexual orientation, and disability? 4. What are the studies showing in terms of the prevalence, incidence, characteristics, and impacts of GBV? This working paper presents first the methodology used in the review and then the results of the analysis of studies followed by a discussion and conclusion. More detailed analysis of the studies identified and included in the review will be provided in peer-reviewed, academic articles. 2. Methodology The aim of the overview was to assess the scope of existing UK studies, the range of research designs and tools used, and the results emerging from these studies. It was intended that the results would identify emerging and significant aspects of GBV in UK universities, which would feed into developing an ecological model. We focused on two forms of GBV, sexual violence (SV) and domestic violence and abuse (DVA). To guide the overview, we used the World Health Organisation’s definition of SV and the Home Office’s definition of DVA: Any sexual act, attempt to obtain a sexual act, unwanted sexual comments or advances or acts to traffic, or otherwise directed, against a person’s sexuality using coercion, by any person regardless of their relationship to the victim, in any setting, including but not limited to home and work (WHO, 2011, page 149). Any incident or pattern of incidents of controlling, coercive or threatening behaviour, violence or abuse between those aged 16 or older who are or have been intimate partners regardless of gender or sexuality (Home Office, 2013) We have taken a wide approach and included in the overview all studies (published and unpublished) conducted in the UK on SV and DVA among university students and staff. Studies were stored on a secure server and were given a random ID to facilitate data management and extraction. Searches and data extraction were conducted by the first author and reviewed by the fourth author. 2.1 Search strategies and inclusion criteria We utilised a variety of search methods and approaches because we wanted to access both published and unpublished research. Studies were identified through: • Systematically searching electronic databases for studies on SV and DVA among UK university students and staff • Survey distributed to UK networks of experts and stakeholders (e.g. security services) • Direct contact with university support providers (e.g. students wellbeing), experts, and study authors • Additional searches of university websites We utilised wider inclusion criteria to capture as many studies as possible. Our criteria included • Studies that focused on SV and/or DVA among university students and/or staff • Studies conducted in the UK between January 2005 and January 2019 that were published as formal academic studies; published as grey literature (e.g. government reports, university reports available to the public); or held internally by universities • Studies that reported the methods used. o For studies reporting they used quantitative surveys, quantitative survey queries were provided. • Studies that provided biographical information of research participants Electronic search For the electronic searches, the following search strings were used: 1) (domestic violence or domestic abuse or intimate partner violence or dating violence) AND (university students or university staff) and (location England or Wales or Ireland or Scotland or United Kingdom) and (language English) 2) (sexual violence or sexual assault or sexual harassment) AND (university students or university staff) and (location England or Wales or Ireland or Scotland or United Kingdom) and (language English) Databases searched were: International Bibliography of the Social Sciences (IBSS), PsychInfo, EmBase and ISI Web of Knowledge. Figure 1. Flowchart of electronic searches and screening 1,679 records identified though database searching 1,314 titles and abstracts screened 365 duplicates removed 1272 records excluded 42 full-text studies assessed for eligibility 35 full-text studies excluded, with reason 7 eligible studies included in review Survey + direct contact with study authors and experts A survey was distributed to networks of university researchers and stakeholders, which asked respondents to share reports and publications. Additionally, study authors and experts were directly contacted. This process resulted in 13 studies for inclusion. References of these studies were searched, yielding 1 additional study for inclusion. Figure 2. Flowchart of identifying studies for the overview Search of electronic databases (n = 1314) Survey + direct contact with networks, study authors, and experts + reference searches (n = 14) Records excluded with reasons (n = 1272) Full text articles assessed for eligibility (n = 42) Full text excluded with reason (n = 35) Eligible articles included in review (n = 7) Articles identified for review (n = 21) Excluded (duplicates: n = 2; with reason: n = 4) Unique studies included in final review (n = 15) 2.2 Data extraction An Excel template was used to gather the following information on studies identified for the overview: study reference, focus, methods, and findings. When there were multiple references describing the same study, all references were recorded but the data and findings were counted only once for the analysis. Two categories of information were extracted: 1) Overarching study information • Study ID • Reference • Focus of study: SV, DVA, both • Category of study authors: academic, within universities student organisation; outside of universities student organisation 2) Details of study • Design • Methods: quantitative, qualitative, mixed-method • Sample • Quantitative tools: validated measure, written for study • Results • Limitations 2.3 Analysis Studies were divided into those that focused on SV and those that focused on DVA and then analysed. Only quantitative methods and findings were analysed for this working paper. A narrative synthesis will be conducted on these and will be published separately. An additional narrative synthesis will be conducted on qualitative methods and findings, and published elsewhere. 3. Findings 3.1 Study selection From the electronic search, a total of 1,314 records were generated. Following screening of titles and abstracts, 42 records were assessed for eligibility. Seven records were identified as eligible for the review. These records were compiled with records identified through a survey, direct contact, and searching references of identified records (n=14), for a total of 21 studies. Six articles were excluded, yielding a total of 15 studies in the final review, all of which looked at university students. None of the studies looked at university staff. As shown in Table 1, fourteen studies focused on SV. Of these, seven were conducted by academic researchers, five by student organisations affiliated with a university, and two by student organisations not affiliated with a university. Most of the studies conducted by academic researchers laid out the conceptual framework that guided their studies whereas most of the studies conducted by student organisations did not. All of the SV studies looked at victimisation, yet only nine also asked about the impacts of SV victimisation. Six studies included stalking victimisation. One study focused exclusively on DVA, which was conducted by academic researchers. Two studies that focused on SV included queries about DVA, one conducted by students’ organisations not affiliated with a university and one by a student organisation affiliated with a university. None of the DVA studies presented a conceptual framework. Two looked at students’ victimisation experiences and one looked at student perpetration. Two of the DVA studies included sections on the impact of DVA. 3.2 Design and methods The research design for all studies was cross-sectional. (See Table 1.). For the SV studies, the most common method was a survey with 13 of 14 studies utilising them. One SV study used only interviews. Looking at the studies that used surveys in more detail, 5 studies used surveys with only quantitative queries, 5 studies employed surveys with quantitative and qualitative queries, two studies used surveys with quantitative and qualitative queries and focus groups. The remaining study used a survey and focus groups. For the study which focused on only DVA, the design was cross-sectional and the method was a quantitative survey. For the two SV studies that included DVA queries, the method used was a survey with quantitative and qualitative queries. Table 1 Overview of studies selected for final review Article Topics research Conceptual Framework Recruitment Techniques Design Data Collection Methods Random ID Author category SV Stalking DVA Impact Yes or No (1) Social media; (2) direct face- to-face; (3) survey link sent to email address; (4) Poster/flyers Survey Quantitative Survey Qualitative Interview Focus Groups 1 Outside of universities student organisation X X X X No 1; 3 Cross- sectional X X 4 Outside of universities student organisation X X Yes 1 Cross- sectional X X 6 Within universities student organisation X X No 1; 3; 4 Cross- sectional X 7 Within universities student organisation X X No Cross- sectional X X 8 Within universities student organisation X X No Cross- sectional X X 9 Within universities student organisation X No Cross- sectional X 10 Within universities student organisation X X X X No 1; 3 Cross- sectional X X 13 Academic X Yes 1; 3 Cross- sectional X X X 14 Academic X Yes 1; 3 Cross- sectional X X 15 Academic X No 3 Cross- sectional X 12 Academic X Yes 3; 4 Cross- sectional X 5 Academic X Yes 3 Cross- sectional X 16 Academic X X X Yes 2 Cross- sectional X 2 Academic X X X Yes 1; 3; 4 Cross- sectional X X X 11 Academic X X No 3 Cross- sectional X 3.3 Sample As shown in Table 1, there were four recruitment techniques reported in the studies: 1) social media (e.g. Facebook), 2) direct face-to-face, 3) survey link sent to email address, and 4) physical poster/flyers. Out of the 15 studies, 10 sent survey links to email addresses, 7 used social media, 3 used poster/flyers, and 1 recruited directly face-to-face. Seven of the studies used a combination of recruitment techniques, in which the most common was social media and a survey link. Most of the studies included in the review utilised non-probability techniques, specifically purposive or convenience sampling. Three of the SV studies utilised purposive sampling (e.g. sampled all students in a degree programme; sampled all students attending certain universities) while the rest used convenience. All of the studies looking at DVA employed convenience sampling. See Table 2. Three studies provided information about survey response bias. Sample size varied greatly from more than 100 to over 4,000. While there was variation in the amount of information provided about the samples, all described the gender composition (15 out of 15) and most presented further details (10 out of 15). With the exception of Study 1 and 4 which looked at the experiences of Further Education and Higher Education students, SV studies assessed the experiences of only Higher Education students. See Table 2. The sample for two of the three DVA studies (Study 10 and Study 15) included male and female Higher Education students. The third DVA study (Study 1) included female Further Education and Higher Education students. Table 2 Sample characteristics Source Sampling Technique Size Sample Gender UK status Student status Ethnicity Sexual Orientation Disability Random ID Country FE University Women Men 1 UK X X convenience sampling of target group n = 2001 to 4000 X X X X X 4 UK X X convenience sampling n = 1000 to 2000 X X X X X X X 6 North Ireland X convenience sampling n = 2001 to 4000 X X X X X 7 Scotland X convenience sampling n = 401 to 800 X X X X X X X 8 England X convenience sampling n = 401 to 800 X X X X X X X 9 England X convenience sampling n = 1000 to 2000 X X X 10 England X convenience sampling n = 2001 to 4000 X X X X X 13 England X Purposive sampling n = 100 to 400 X X X X X X 14 UK X convenience sampling n = 401 to 800 X X X X 15 UK X convenience sampling n = 100 to 400 X X 12 England X convenience sampling n = 1000 to 2000 X X X 5 Wales X Purposive sampling n = 4001 to 8000 X X X 2C England X convenience sampling of target group n = 401 to 800 n = 401 to 800 X 11 Scotland X Purposive sampling n = 100 to 400 X X Note. Study 16 was not analysed as it used only qualitative methods and this working paper focuses on quantitative. C Two studies were conducted and described in this record. 3.4 Quantitative tools Out of the 13 SV studies that used quantitative survey tools, eight measured sexual harassment and sexual assault, three measured sexual harassment only, and two measured sexual assault only. See Table 3. The quantitative tools used to measure sexual harassment and sexual assault varied: eight studies measured SV with tools unique to their study and written by the study authors; three used non-validated tools written by another study; one used validated tools; and one used queries from a validated tool and the study authors wrote their own queries. Two of the three DVA studies measured only physical and sexual abuse - one used a non-validated tool written by another study and one used a tool written by the study authors. The third DVA study measured physical and verbal abuse perpetration, with a measure developed by another study. Examining the impact of SV and DVA is essential to understand the different effects of victimisation across social groups, such as gender (e.g. Hester et al, 2017). In this working paper, only seven of fourteen studies examined impact, with queries ranging from mental health to finances to academic, e.g. marks, attendance. Table 3 Quantitative tools used to measure SV, DVA, and Impact Quantitative Tools SV Forms Assessed DVA Forms Assessed Impact Random ID Validated (written by study authors; validated measure) Sexual Harassment Sexual assault* Physical Verbal Sexual SV DVA 1 Written X X X X X X 4 Written X X X 6 Written X X X 7 Written by authors of another study X X 8 Written X X X 9 Written X 10 Written by authors of another study X X X X X X 13 Written X 14 Written by authors of another study & Validated measure X X 15 Measure on aggression developed by another study X X 12 Validated measure X 5 Written X X 2 Written X X 11 Written by authors of another study X X * Rape was included in the sexual assault category. 3.5 Results Prevalence Table 5 below summarises the prevalence of SV and DVA experienced by students while attending university, as well as the impacts. When considering the overall picture of SV victimisation, 10% to 77% of students experienced sexual harassment and 3% to 28% of students experienced sexual assault. The findings showed 2% to 69% of female students and 3% to 39% of male students experienced sexual harassment, and 1% to 34% of female students and 6% to 7% of male students experienced sexual assault. These findings point toward the conclusion that the prevalence of female victimisation is higher than male victimisation. Currently, this conclusion is tentative because of the variations in reporting, with some studies presented the prevalence for individual SV behaviours while others presented the prevalence for all SV behaviours. Of note, one study asked students about revictimization, finding that 46% of those who experienced attempted sexual assault were victimised again and 40% of those who experienced sexual assault were victimised again. DVA studies looked at victimisation and perpetration. Studies on DVA victimisation queried students about different types of victimisation experiences and then asked about the relationship with the perpetrator. Nearly one in five (19%) of all students who experienced physical violence reported the perpetrator was an intimate or romantic partner and 7% to 18% of all students who experienced sexual violence reported the perpetrator was an intimate or romantic partner. Some studies compared the extent of victimisation across social positions, including sexual orientation, ethnicity, and student status (undergraduate vs postgraduate). For sexual orientation, 5% to 15% of gay, queer, and bisexual men experienced SV as compared to 1% to 7% heterosexual men and 8% to 23% gay, queer and bisexual women experienced SV compared to 3% to 15% heterosexual women. BME students were compared to white students, with findings showing 4% to 6% of BME students experienced SV whereas 1% to 4% of white students did. A higher prevalence of postgraduate students (15%) experienced SV than undergraduate students (6%). These findings suggest the prevalence of victimisation is higher for marginalised individuals Impact In terms of impact, mental health impacts were the most common followed by academic and then financial. Only one study provided information about impact experienced by men so it is difficult to make gender comparisons. Characteristics of victimisation experiences Eight studies asked students about the context in which they experienced SV. Specifically, they queried the gender of the perpetrator, location of victimisation experience, and alcohol and drug use. Table 4 presents a summary of these findings. Of immediate note is how findings were reported, i.e. the overall trend was contextual information was reported for both male and female victims’ or for female victims only, leaving a gap in knowledge about the context of male victims’ experiences. The one piece of information reported for male victims was the gender of the perpetrator – approximately half were men. The proportion of male perpetrators increased when male and female victims were amalgamated (76% to 97%) or when female victims were reported on their own (81% to 96%). Studies’ reports on the location of SV experiences tended to ask if the location was on or off university premises and then proceed to ask more detailed questions. These findings suggested more than half of SV experiences occurred off university premises and a significant proportion occurred where students were living. Some studies asked students about their alcohol/drug use and the perpetrators’ use. The different approaches to generating this information limits the extent of what can be said about the influence of alcohol/drugs on sexual violence victimisation and perpetration. For example, one study looked at hazardous levels of drinking alcohol as a risk factor for victimisation while other studies asked female victims if they were under the influence. The former approach used a standardised measure to understand the relationship between alcohol use and SV victimisation whereas the latter approach created the space for students to determine if they were under the influence. Interestingly, the former approach found a higher risk than the latter. Studies asking victims if they thought the perpetrator was under the influence showed consistent findings, with 65% to 81% of male and female victims reporting they thought the perpetrator was and 47% to 60% of female victims reporting the same. Table 4 Summary of context of SV victimisation All Victims Male Victims Female Victims Perpetrator 76 to 97% men 54 to 60% men 81 to 96% men 45 to 74% knew the perpetrator 41 to 84% knew the perpetrator Location 16 to 30% university premises; 26 to 53% university premises; 31 to 78% student residence 41 to 49% non-university accommodation 76% own home or in home of someone known Alcohol/drug consumption 18% of non-hazardous alcohol consumption experienced alcohol related non-consensual sex * 37 to 74% of victims under the influence 82% of hazardous alcohol consumption experienced alcohol related non-consensual sex* 11 to 14% coerced into taking alcohol or drugs 7 to 9% given alcohol or drugs without their consent 17 to 42% perpetrator provided victim with alcohol or dugs 65 to 81% of perps under the influence of alcohol or dugs 47 to 60% perp under influence of alcohol or drugs * The Alcohol Use Disorder Identification Test (AUDIT), a tool developed by the World Health Organisation for students, was used to measure drinking levels. If students’ levels of drinking met the threshold set out by the AUDIT authors, they were considered to be hazardous drinkers. Table 5 Summary of results SV DVA Impact of SV and DVA Sexual Harassment Sexual Assault Physical Verbal Sexual Academic Mental Health Financial Women 2 to 69% 1 to 34% 35 to 38%* 30%** 83.5%** 50%* 10 to 50% 18 to 78% 7 to 8% Men 3 to 39% 6 to 7% 15%** 59%** 3% 5% Total Sample 10 to 77% 3 to 28% 19%* 7 to 18%* 61 to 71% 47 to 85% * These number describe the percent of perpetrators who were identified as intimate/dating/romantic partners. ** Percentages represent the proportion of students who used physical or verbal aggression towards an intimate/dating/romantic partner. 4. Limitations of studies No study could be located that investigated SV and DVA among university staff. None of the studies were conducted with the explicit intention of investigating DVA. There were identified and categorised as DVA studies because they included domestic/dating/romantic partners in their queries about relationship to the perpetrator. None of the studies asked about students’ perpetration of SV or used a holistic understanding of DVA to ask about DVA perpetration, i.e. did not utilise a conceptual understanding of DVA that recognised DVA can occur in more forms than physical, sexual, or verbal. Some studies did not present findings in a way that allowed group comparisons. Studies described impact for entire samples, so it was not possible to see if and how impact differed across student groups. None of the studies reported incidence of SV or DVA. If studies reported incidence of SV and DVA, we would be able to tell how many ‘new’ experiences of victimisation occurred for each time span of education, e.g. during the first year of undergraduate, which could in turn be used to identify risk factors unique to first experiences of victimisation. This information could also be used to identify which forms of SV and DVA co-occur in singular incidents. A dearth of information on response bias made it unclear which recruitment strategies might be the most effective. While all the studies included in this review shed light on the extent of victimisation at one point in time, none looked longitudinally to examine if, when, where, and how revictimization occurs. 5. Summary and Conclusions In total we reviewed fifteen studies, half of which were led by academic researchers and the other half by student led organisations. The predominant focus of these studies was students’ experiences of sexual violence victimisation (14 of 15), with a minority looking at students’ perpetration and experiences of DVA (3 of 15). Studies that looked at university staff’s victimisation or perpetration of SV or DVA could not be located. All of the studies in the review used a cross-sectional design, with 14 studies using online surveys to generate quantitative data. Seven of 14 studies used online surveys to generate qualitative data to complement the quantitative data. Considering the breadth of validated tools available, it is surprising only 2 studies used validated tools to elicit information about sexual violence and none of the DVA studies used validated tools. Most of the studies used tools written by another study or written by the study authors themselves, making it difficult to compare findings across studies and to understand the extent, characteristics, and impacts of SV and DVA among students attending UK universities. Nonetheless, there are important emerging findings which evidence that SV and DVA are pressing social issues among UK university students. Consistent with previous research in the U.S. (e.g. Krebs et al., 2017), approximately 10% to 77% of students experienced sexual harassment and 3% to 28% experienced sexual assault. The majority of these students reported their mental health and academic studies were affected by these experiences, suggesting that university SV prevention work could reduce the need for university resources aiding recovery. Studies included in this review tentatively suggested that the prevalence of female students experiencing SV is higher than for the prevalence of male students experiencing SV. The emerging findings for DVA were less clear, with findings pointing towards the conclusion more female than male students perpetrate DVA. However, the DVA studies did not consider the overall context of the relationship dynamics nor the impact. When these factors were accounted for in previous studies (e.g. Allen, Swan and Ragahavan, 2009), the prevalence rate of perpetration was higher for men than women. Most studies (11 of 14) collected biographical information needed to understand how the intersections of gender with other social positions influenced victimisation. However, only 4 studies used this information to compare groups, e.g. BME vs White, LGB vs heterosexual, and only one study broke down the comparisons further with gender, e.g. male LGB vs male heterosexual. Findings from these four studies were consistent with previous research on GBV (e.g. Coulter et al, 2017) in that they indicated that occupying positions of less social power (e.g. women, BME) had an increased chance of experiencing SV and or DVA. Eight studies elicited information about the context in which SV occurred, focusing on the perpetrators’ gender, location of experience, and alcohol/drug use. In brief, the majority of perpetrators were men, less than half of SV experiences were on university premises, and substantial proportion of victims and perpetrators were under the influence of alcohol/drugs. Recommendations for future research To understand the extent, incidence, characteristics, and impact of GBV in UK universities, as well as how GBV varies across universities, we recommend the following for quantitative studies: - Universities should agree on an understanding of gender-based violence and definitions of SV and DVA, and the forms of each, that can be used as the basis of operational definitions in future research. These definitions should not consider SV and DVA to be mutually exclusive, e.g. SV may occur within DVA. - A quantitative tool specific to UK universities needs to be developed and validated. The tool must include perpetration and impact questions and facilitate extracting information about incidents and prevalence. - Recruitment techniques (e.g. survey link to email address) and data collection strategies (e.g. online survey) need to be tested to find the most effective and consistent. - Sampling strategies need to be more rigorous (e.g. stratified random sampling) and described in more detail in future studies. Additionally, studies must look at all university staff. - Research designs should include longitudinal studies to understand how GBV and its impacts changes over students’ and staffs’ careers at university and beyond. - More sophisticated analyses are needed to understand how experiences and perpetration of SV and DVA differ across social positions, e.g. male, BME, heterosexual students compared to female, BME, LGB students. - Studies should report findings for all students together and then for male students only and female students only. This format should be followed for university staff as well, i.e. all staff, male staff only, female staff only. - More information is need about the context in which SV and DVA occurs, such as year of study or year of employment and type of campus, e.g. city based. References Allen, C.T., Swan, S.C. and Raghavan, C., 2009. Gender symmetry, sexism, and intimate partner violence. Journal of Interpersonal Violence, 24(11), pp.1816-1834. Coulter, R.W., Mair, C., Miller, E., Blosnich, J.R., Matthews, D.D. and McCauley, H.L. 2017. Prevalence of past-year sexual assault victimization among undergraduate students: Exploring differences by and intersections of gender identity, sexual identity, and race/ethnicity. Prevention Science, 18(6), pp.726-736. DeGue, S. 2014. Preventing Sexual Violence on College Campuses: Lessons from Research and Practice. Part One: Evidence-Based Strategies for the Primary Prevention of Sexual Violence Perpetration [online]. Washington DC: White House Task Force to Protect Students from Sexual Assault. Available from: www.notalone.gov/assets/evidence- basedstrategies-for-the-prevention-of-sv-perpetration.pdf [Accessed 4 January 2019]. Hagemann-White, C., Kelly, E. and Römkens, R., 2010. Feasibility study to assess the possibilities, opportunities and needs to standardise national legislation on violence against women, violence against children and sexual orientation violence. Available from: https://op.europa.eu/en/publication-detail/-/publication/cc805fb4-c139-4ac0-99b7- b0dad60179f7/language-en [Accessed 17 April 2017]. Hester, M. and Lilley, S.J. 2014. Preventing Violence Against Women: Article 12 of the Istanbul Convention [online]. Strasbourg: Council of Europe. Available from: www.coe.int/t/dghl/standardsetting/convention-violence/thematic_factsheets/Article 12 English.pdf [Accessed 30 June 2015]. Hester, M., Jones, C., Williamson, E., Fahmy, E. and Feder, G., 2017. Is it coercive controlling violence? A cross-sectional domestic violence and abuse survey of men attending general practice in England. Psychology of violence, 7(3), p.417. Krebs, C., Lindquist, C., Berzofsky, M., Shook-Sa, B and Peterson, K. 2017. Campus Climate Survey Validation Study. Final Technical Report. Washington, DC: Bureau of Justice Statistics. NUS (National Union of Students). 2011. Hidden Marks: A Study of Women Students’ Experiences of Harassment, Stalking, Violence and Sexual Assault. 2nd ed. [online]. London: NUS. Available from: www.nus.org.uk/Global/NUS_hidden_marks_report_2nd_edition_web.pdf [Accessed 16 April 2015]. Phipps, A. and Smith, G., 2012. Violence against women students in the UK: Time to take action. Gender and Education, 24(4), pp.357-373. Stenning, P., Mitra-Kahn, T. and Gunby, C. 2012. Gender-Based Violence, Stalking and Fear of Crime: Country Report United Kingdom [online]. EU-Project 2009-2011 JLS/ 2007/ISEC/415. Bochum, Germany: Ruhr-University Bochum; Keele: Keele University. https://op.europa.eu/en/publication-detail/-/publication/cc805fb4-c139-4ac0-99b7-b0dad60179f7/language-en https://op.europa.eu/en/publication-detail/-/publication/cc805fb4-c139-4ac0-99b7-b0dad60179f7/language-en Project website: www.gendercrime.eu. Available from: http://vmrz0183.vm.ruhr- unibochum.de/gendercrime/project_country_report_united_kingdom.html [Accessed 16 April 2015]. Universities UK Taskforce. (2016) Changing the Culture. Report of the Universities UK Taskforce Examining Violence against Women, Harassment and Hate Crime Affecting University Students [online]. London: Universities UK. Women and Equalities Committee. 2018. Sexual harassment in the workplace. London: House of Commons. Available from: parliament.uk/business/committees/committees-a- z/commons-select/women-and-equalities-committee/inquiries/parliament-2017/sexual- harassment-workplace-17-19/ [Accessed 7 September 2018].