Pathways of change: qualitative evaluations of intimate partner violence prevention programmes in Ghana, Rwanda, South Africa and Tajikistan

Abstract

A critical component of evaluations of the effectiveness of intimate partner violence prevention programmes involves understanding pathways of change among individuals who participate in such programmes, and the intervention or contextual elements that support or hinder these. This paper draws on qualitative evaluations of four intimate partner violence prevention programmes in Ghana, Rwanda, South Africa and Tajikistan conducted as part of the What Works to Prevent Violence against Women and Girls Programme. Using a comparative case study approach, a secondary analysis was applied to thematically analysed data to explore how and why men and women change in response to different types of programmes across diverse contexts. Similar pathways of change were identified including the value of learning and applying relationship skills to support equitable, non-violent relationships; the importance of participatory approaches to challenge harmful gender norms and allow for group rapport; and the integration of economic empowerment activities to reduce drivers of intimate partner violence and conflict, and promote participants’ self-confidence and status. These findings provide insights regarding intervention design and implementation factors pertinent to bring about changes in intimate partner violence.

Keywords:

• Intimate partner violence
• gender
• intervention
• principles for success

Introduction

The proliferation of intimate partner violence prevention programme evaluations has generated significant insights into what works to prevent intimate partner violence. A recent review summarised the core elements of more successful interventions as including: a strong contextually rooted theory of change; the use of participatory, group learning methods; working with men, women and their families collectively; focusing on multiple drivers of violence; and rigorous implementation (Jewkes et al. 2019). However, we have much less understanding of how interventions work, the pathways to change among individuals who participate in such programmes and the intervention or contextual elements that support or hinder this. Analysing pathways of change is particularly important given the complexities of shifting harmful gender norms and acceptability of the use of violence. Moreover, effective intimate partner violence prevention interventions are typically quite complex and often work at multiple levels, or with different interacting components (Davis et al. 2019). Participants’ interaction with programmes, including the negotiations and struggles that lead to change (Sullivan 2004), are not well understood.

The UK Department for International Development (DFID)-funded What Works to Prevent Violence against Women and Girls (VAWG) Programme has generated evidence around the effectiveness of intimate partner violence prevention programmes from 13 countries in Asia and Africa. This paper explores pathways of change through qualitative evaluations of four of the studies included in the portfolio. We included studies with quantitative evidence of reducing intimate partner violence and/or domestic violence for some participants, with ample qualitative data available, targeting heterosexual adults as the primary beneficiaries. The included studies cover a diversity of approaches: community mobilisation (Rwanda, Ghana), livelihood strengthening and gender-transformative training (South Africa, Tajikistan, Rwanda) and working with couples (Rwanda).

Using a comparative case study approach (Goodrick 2014), we explore how and why men and women changed in response to diverse programmes across different contexts, and the intervention and contextual enablers and/or barriers underlying this. A comparative case study approach involves the analysis and synthesis of similarities, differences and pathways across two or more cases that share a common focus or goal (Goodrick 2014), which in doing so offers a more nuanced consideration of processes and context (Bartlett and Vavrus 2017). We further draw on Heise’s (1998) socio-ecological model of gender-based violence to identify salient risk and protective factors underlying violence. This model focuses on factors that emerge across individual, relationship, community and society levels.

Methods

A description of the included interventions, theoretical frameworks for the interventions and accompanying qualitative evaluations is presented in Table 1. The primary methodological approaches used differ as they were evaluating different programmes and populations, and this study was conducted as a secondary comparative analysis. In all contexts, experienced qualitative interviewers external to the programme implementation team were hired and trained in qualitative research on violence and conducted the interviews in local languages. The majority of interviews across the contexts were gender-matched between interviewer and interviewee. Prior to work commencing on this paper, data from all of the studies had been coded, analysed and published in country-specific reports and papers, where further details of the interventions and methods can be accessed (Addo-Lartey et al. 2019; Gibbs et al. 2017; Mastonshoeva et al. 2019; Stern et al. 2018). This includes details and results of the accompanying quantitative evaluation research conducted in all settings.

Table 1. Interventions and qualitative research.

The Rural Response System in Ghana

Context: The intervention was delivered in two districts of central Ghana, one inland and mostly rural where participants lived in small villages and towns and were mainly farmers, and one coastal with small towns and many households living from fishing.

Intervention: A six-person Community Action Team (COMBAT) per village was nominated by the community, trained for 2 weeks with a refresher training after one year, and deployed per community. They worked through community activities that raised awareness about the types and consequences of violence against women and girls (VAWG), challenged inequitable gender attitudes, provided mediation and counselling for couples, engaged leaders and state actors and strengthened linkages to response services. The intervention was implemented over 18 months.

Theoretical/conceptual framework: Through sensitisation, awareness creation, case management and referrals, leading to pathways to improve attitudes to violence against women (VAW) at the individual and community levels; and a reduction in VAW.

Implementation team: Gender Studies and Human Rights Documentation Centre.

Research methods: In-depth interviews (IDIs) were conducted with 20 adult women and 20 adult men who had reported experience (women) or perpetration (men) of intimate partner violence in the baseline survey; IDIs repeated at midline (12 months) and endline (24 months) with 54 men and 57 women in total. Scope: VAWG knowledge and perceptions, experiences with the intervention, gender attitudes, stigma and violence for survivors; changes in violence against women and gender attitudes, the role of and seeking help from COMBATs, sustainability of the intervention. Focus group discussions (FGDs) were held with 20 men and 20 women from the communities, following the same scope of inquiry at baseline, midline and endline. Groups were stratified by gender and age (18–34 and 35–49 years). There were 2 FGDs with COMBATs at baseline, midline and endline focusing on their perceptions of activities regarding awareness creation and support provided to survivors of violence.

Indashyikirwa in Rwanda

Context: Indashyikirwa (Agents of Change) was implemented in seven districts across fourteen sectors in the Eastern, Northern and Western provinces of Rwanda, in predominantly rural, widely dispersed communities.

Intervention: The intervention was delivered through a participatory curriculum with couples recruited from CARE’s micro-finance village savings and loans associations (VSLAs) that spanned 21 sessions of 3 hours and was delivered over 5 months. Community activism was then led by 500 men and women who attended the couples curriculum and volunteered for a further 10 half days of training to equip them to implement activities, in a programme that drew on SASA! established by Raising Voices. They facilitated 3–4 activities per month over 19 months. Safe space facilitators were trained for 10 days. The safe spaces were open 3 times a week for 22 months. They offered counselling, income-generating activities (IGAs), activism activities and supported community members to access formal response services. Ten days of training and ongoing quarterly meetings were provided for opinion leaders, who provided support for survivors and raised awareness over 30 months.

Theoretical/conceptual framework: The Indashyikirwa programme drew on the socio-ecological model to address violence (Heise 1998), and thus purposefully included programme elements to address risk and protective factors for violence at the individual, relationship and community levels.

Implementation team: CARE International, Rwanda Men's Resource Centre (RWAMREC) and Rwanda Women's Network (RWN).

Research methods: IDIs were held with 15 male and female partners of couples at baseline, midline (7 months) and endline (19 months) across 3 districts. Some were married and some cohabiting. Scope: Couples’ experiences of conflict and intimate partner violence, communication skills and joint decision-making; whether and how Indashyikirwa impacted their relationships. Twelve individuals who trained as community activists were interviewed at 12 months and 30 months across 3 districts (2 men and 2 women per district). Scope: Impressions of the activism training and support received, motivation to continue as activists, what they had been doing recently as activists, and whether they faced any challenges. Opinion leaders were interviewed from across 3 districts: 9 at baseline, 6 at midline (9 months) and 9 at endline (12 months). Scope: Impressions of the training and whether their involvement with the programme influenced their intimate partner violence prevention and response actions. Three women’s safe space facilitators were interviewed at baseline (5 months), midline (18 months) and endline (30 months) across 3 districts. Scope: Motivations to be facilitators, impressions of training and support received, perceived impact of the safe spaces. Women’s safe space attendees were interviewed across 3 districts at 9 months (n = 6),18 months (n = 8) and 30 (n = 6) months into the study. Scope: Why they visit the safe spaces and the difference the spaces make in their lives. Fifteen RWAMREC and RWN Facilitators were interviewed across various districts at 7 months (n = 14) and 19 months (n = 12). Scope: Successes and challenges engaging opinion leaders and supporting the safe spaces (RWN staff), and from training couples, community activists and supporting community activism with trained couples (RWAMREC staff).

Zindagii Shoista in Tajikistan

Context: Rural villages with high levels of poverty and limited economic opportunities. Many men migrate to Russia in search of work, and women mostly live with their husband's family.

Intervention: Zindagii Shoista (Living with Dignity) was implemented with up to four members of multi-generational households and comprised 21 sessions (each 2.5 hours), over 15 months: 10 sessions focused on building gender equity, respect and intrahousehold communication and reducing violence; and 11 sessions worked with each family to develop a concept for an IGA, offer training in business skills and materials to support starting an IGA (∼US$370/household). There was supervision and business support for each IGA.

Theoretical/conceptual framework: The intervention drew heavily on the Stepping Stones and Creating Futures broad theoretical framework described below.

Implementation Team: International Alert, Cesvi and three Tajik NGO partners – ATO, Farodis and Zanoni Sharq.

Research methods: 21 IDIs (two – in one case three – men and three women in each village) and eight FGDs (one per gender per village) were conducted with 9–13 participants. These were repeated at midline at 8 months (n = 25), endline at 15 months (n = 25) and at 15 months post-intervention (n = 8). Scope: Prevalence of VAWG, socio-economic situation, health and mental well-being; participants developed ‘everyday change indicators’ and tracked these, allowing them to define and monitor issues they saw as pertinent. Two FGDs were conducted with facilitators at each of the target districts involving four facilitators per district at the same time points.

Stepping Stones and Creating Futures in Durban, South Africa

Context: Urban informal settlements with exceedingly high levels of poverty, limited work opportunities and high levels of violence and harmful alcohol use. Young people tended not to live with a partner, but on their own or with friends.

Intervention: Stepping Stones and Creating Futures (SSCF) intervention was delivered in single-sex groups of approximately 15–20 participants, with sessions lasting 3 hours, over approximately 3 months. Men and women were recruited individually, not as couples. The 21 three-hour sessions included: 10 sessions focused on building gender equity, respect and inter-personal communication and reducing violence; and 11 sessions from Creating Futures, an intervention focused on getting and keeping work, saving and understanding the opportunities and constraints for livelihoods.

Theoretical/conceptual framework: The intervention drew on theories of gender and power and how materiality intersects with these. Processes of change were based on Freire’s ([1968] 2014) theory of dialogical communication and critical thinking.

Implementation Team: Project Empower.

Research Methods: IDIs were conducted with 19 men at baseline, midline at 12 months (n = 5) and endline at 20–24 (n = 7) months, and 15 women were interviewed at baseline, midline at 12 months (n = 12) and endline at 18 months (n = 13). They were living in two informal settlements participating in the programme. Scope: Young people’s livelihoods, experiences of violence, intimate relationships, gender attitudes and whether and how the intervention leads to changes in these issues. Interviews were supplemented by participant observation. Women’s photovoice was conducted at baseline and 18 months with 8 and 6 women. Scope: Relationships, reproductive decision-making, intimate partner violence, experiences and perceptions of power, livelihoods and future aspirations.

Ethics

The projects received ethical approval from the South African Medical Research Council (Ghana: EC031-9/2015; Rwanda: EC033-10/2015; South Africa: EC006-2/2015; Tajikistan: EC012-5-2016). Additional national ethical approvals were obtained for studies in Ghana, Rwanda and South Africa. In Tajikistan, given the absence of a functioning national review board at the time of data collection, approval letters to certify compliance with research ethics standards were obtained from the Tajik Academy of Dialogue and from local authorities. Across all studies, written informed consent was obtained from participants, and findings were anonymised. All studies included mechanisms for referring participants either to professional counsellors or organisations to support survivors of violence. Pseudonyms are used for quotations to protect the confidentiality of participants.

Joint data analysis

Qualitative data from all studies were audio-recorded, transcribed and translated verbatim into English. The data were then analysed thematically by the respective investigative teams, using different thematic codes and areas of interest, although all with the primary research question of assessing the impact of the local programme on intimate partner violence. Respective teams, including the authors of this paper, coded the data using a priori codes developed from the research topic guides. Further codes that emerged during the process of thematic analysis were added to provide a rich, detailed and holistic account of the findings.

For this paper, authors from Rwanda, South Africa and Tajikistan studies used thematically analysed midline and endline data (and for Tajikistan post-endline data) to answer the following secondary questions: (1) How did participants make sense of the impact of the intervention, and how did this understanding evolve throughout the programme? (2) Which components of the intervention appeared to support changes in intimate partner violence and why, and which did not, and with whom in particular? and (3) What contextual factors facilitated or impeded the programmatic changes and how? The first analytic question implies the use of an inductive approach to access participants’ perspectives on pathways and processes of change. The second and third analytic questions imply the use of a deductive approach with respect to investigators’ perspectives on pathways of change. Original quotations from the data set are included in investigators’ responses to these analytical questions. For this paper, the first author analysed the Ghana midline and endline data to respond to these questions as the Ghana investigation team was unavailable for this task.

The first author subsequently identified common patterns across the four studies. A comparative case study approach guided the analysis and the presentation of the findings, focused on similarities and differences around pathways and processes of change to prevent and mitigate intimate partner violence.

Findings

Across the different studies, a range of similar pathways of change, potentially leading to reductions in intimate partner violence, were identified. We discuss these using the socio-ecological model’s structure.

Individual level pathways to change

Greater self-confidence

Participants in all settings reported improved self-confidence resulting from participation in the intervention, a significant individual-level change, which could be an important protective factor against intimate partner violence. As Azin, a young woman in Tajikistan, reflected at endline: ‘I became more brave and lively and have a vision for my future’. Greater self-confidence was related to the pathway of intervention elements promoting empowerment. In South Africa, some women described how the intervention gave them more power over their lives, especially to address the challenges they face: ‘SSCF [Stepping Stones Creating Futures] changed me because I am now able to be independent, I am able to handle any challenges and accept that they exist’ (Philani, midline, South Africa). Women also appreciated that the SSCF sessions reflected on goal-setting; for many it was the first time they had space to reflect on their lives, dream and set goals: ‘I learned from the intervention that I must know what I want in life, so I can succeed’ (Oyama, midline, South Africa). One of the underlying concepts of Indashyikirwa (adapted from SASA!) in Rwanda concerns the four types of power: positive power (power to, power within, power with) and negative power (power over). Most trained couples, women’s safe space facilitators and attendees emphasised the value of the concept of ‘power within’ to recognise and use one’s power positively: ‘Now I openly speak out and I use the power that I have in me and I feel there is something that I can do to make my family developed. That is a very big thing’ (Justine, midline, Rwanda).

Increased self-confidence for participants was also related to the perception that they had gained greater recognition in their communities. The majority of activists and safe space facilitators in Rwanda reported gaining community status for their efforts, took pride in being recruited for this work and regularly wore Indashyikirwa vests and badges to be identifiable. Similarly, several community members in Ghana identified greater status and respect towards the Community Action Teams (COMBATs). In Tajikistan, many women and men also reported greater respect and acceptance in their communities and noted how this built their self-confidence. As one young woman emphasised:

Prior to engagement in the intervention, I always felt bad and some form of guilt. I thought I was the reason for the violence in our family. Before the intervention, no one was even considering me as a person worth talking to. The neighbours were labelling our family as a problematic family. Now I feel very happy, our family restored the respect of the community. (Darya, post-endline, Tajikistan)

Shifts in attitudes towards violence and gender

Participants reported how the interventions supported them to change their attitudes around the acceptability of violence and gender norms and roles. Participatory group-based approaches encouraging sharing and critical reflection of experiences were highly valued and enabled participants to collectively engage with new ideas and shift harmful attitudes underlying intimate partner violence and gender inequality.

In Rwanda and Ghana, participants reported how the acceptability of violence was reduced as they came to realise the negative impacts of violence. Emphasising the value of peaceful relationships and communities also supported individuals to condemn violence:

Violence is not a good thing when you live in a peaceful community, you grow healthy, when you have peace in your marital home, it helps you, it is not always about money. Violence destroys your peace, especially we women when you experience violence, it can affect your children, and your children will go astray, so we always pray to God to let peace prevail. This is why we do not tolerate violence. (Focus Group with Women, endline, Ghana)

A significant attitude shift in Tajikistan was the increasing acceptability within many families for young women to work outside the home, which was encouraged through the programme, and after families witnessed the financial benefits first-hand:

After the project sessions, the way my husband and mother-in-law relate to me has changed. I found a job in a bathhouse as a cleaner. My husband and my mother-in-law are not against my work before they used to be opposed to this idea. My mother-in-law now supports me; whenever I go to work, she takes care of my children, helps with daily chores. My husband supports me too. (Farhana, endline, Tajikistan)

Similarly, several men in Rwanda reported being increasingly supportive of women’s economic activities after considering the household benefits and identifying this as women’s rights.

In South Africa, many men described how they increasingly endorsed gender equality, particularly greater respect for women, which led to reduced use of intimate partner violence and controlling behaviours against their intimate partners. Some women in South Africa described resisting their partners’ controlling and violent behaviours and shifting their expectations about gender relations to no longer tolerate such behaviours: ‘I will not get back together with him [ex-partner], he is abusive and quick to beat me. He wanted me to stay at home and cook and clean and do all the house chores. I can’t live like that, not anymore’ (Buhle, PhotoVoice endline, South Africa). While most women in South Africa remained in violent relationships, they shared changing relationship expectations and reduced normalisation of intimate partner violence. One woman noted how she increasingly considered her relationship satisfaction as a priority before considering the needs of her partner, which equipped her to leave a violent relationship:

One of the things that helped me get out of that relationship was some of the things I learned from the intervention. We learned that you must live life for yourself and that your happiness matters more than the other persons’ and you must not stay with someone who doesn’t make you happy. (Thandi, midline, South Africa)

However, limited changes in attitudes around some gender norms were also identified. In Tajikistan, beliefs around gendered domestic roles were difficult to change for some participants, with childcare continuing to be seen as (young) women’s responsibility. The continued salience of the norm of men as heads of households and primary breadwinners was evident across settings. As one participant in Rwanda noted: ‘the last decision is taken by a husband because he is a man and a man is the head of his family … the head of the family is the one to take the final decision’ (Valentine, midline, Rwanda). In South Africa, the continuation of norms around men’s provider roles following SSCF could hinder women from engaging in income-generating activities.

Economic empowerment

In the three interventions that included livelihoods strengthening (South Africa, Tajikistan and Rwanda), the impact of stronger livelihoods, combined with gender-transformative sessions, reduced stress and built confidence and self-esteem. Most young women in Tajikistan gained their first experience of involvement in economic activities through the project, enabling them to earn and contribute to their family budget. The combination of access to resources, an ability to provide and self-confidence raised their self-esteem, status and respect from other family members. Strengthened livelihoods meant less economic stress and related family conflict in households, as one older man emphasised:

The major benefit is that I could start making money for my family’s needs. My brother and sister supplied food for my family. With the support of the project, I could take over that responsibility. They are also happy that I could do it. There is still a long way to go for the complete economic stability of our family, but I am happy I can at least cover some basic needs. (Doro, post-endline, Tajikistan)

In South Africa, men and women appreciated the livelihood skills which enabled them to improve their access to income-generating and education opportunities. For men, the focus on livelihoods included developing plans to generate small incomes and make progress towards longer-term goals. This aspect reportedly improved their confidence, reduced their anger and frustration at their situation and in turn supported the movement towards reduced intimate partner violence. Several women appreciated how SSCF enhanced their ability to look for and secure work and increase their income: ‘The writing of the CV and letter, I learned how it’s done and how to dress for an interview plus which points are important and what to expect from an interview’ (Woman, midline, South Africa). In Rwanda, the stipend offered to each partner for attending the couple’s curriculum session (2000 RWF, approximately US$2.5) was identified as a motivating factor for participation, especially among men. In many cases, couples used the micro-finance village savings and loans association (VSLA) savings and curriculum stipend towards a household investment, such as purchasing a fridge, bicycle, roof or livestock. Many couples in Rwanda reported improved financial security as the result of the programme, which reduced conflict, and couples made more joint decisions and communicated more openly about resources:

When there is a conflict between a man and his wife, they cannot reach financial security, they instead lose and consequently, their household remains in poverty. But when they share ideas and give up quarrelling, there is peace at home, they work together and then there is no more conflict. (Jean, endline, Rwanda)

Economic empowerment activities were critical to supporting change as all the programmes worked with communities from vulnerable socio-economic backgrounds. Indeed, a few participants in Ghana, the only programme without an economic element, highlighted the financial determinants and triggers of intimate partner violence and suggested the intervention would have been strengthened by providing economic empowerment activities.

Relationship-level pathways

Improved intimate relationship skills and quality

Across all settings, participants reported that the programme strengthened their relationship skills, particularly constructive communication and conflict resolution skills, which in turn supported their ability to mitigate and prevent violence. Several women in South Africa reported less frequent arguing and being more strategic about when to communicate with their partner, including to avoid arguments if they or their partner were drunk: ‘I learned that a person can be assertive and have a calm conversation with someone, without it turning into a fight’ (Nambusi, midline, South Africa). Several men in South Africa reported the impact of learning conflict resolution and communication skills for improving their relationships with others, including their partners, and helping reduce their use of violence:

My girlfriend and I discuss things. I don’t just make decisions without communicating with her, it’s a 50/50 agreement, we listen to each other, we are not in a perfect relationship, we argue here and there, but there are ways we deal with those things. You must understand your partner and not just hit her whenever you get into an argument. (Thando, midline, South Africa)

Similarly, in Rwanda, women and men described improved relationship skills including being able to identify and manage triggers of violence, asking for forgiveness, constructive communication and conflict resolution skills, leading to less violence and conflict in relationships. The take-home activities from the couples’ curriculum, in which couples were asked to practise a skill related to each session and reflect on this at the beginning of the next session, helped participants apply new skills at home. Couples completing the curriculum together said it helped to hold each other accountable to changes in their relationships: ‘We have the same opinion about the curriculum. We understand sessions the same, we even revise or remind the other in case one of us goes wrong. This has helped us not to repeat the wrong’ (Marie, endline, Rwanda).

In Ghana, some female community members mentioned that, through awareness and knowledge gained from the programme, they felt more equipped to control their anger and impatience and had a better understanding of how some behaviours (e.g. anger) could lead to intimate partner violence:

Some of us were impatient about certain things, we used to disrespect our husbands, but we have been taught that exchanging words with your husband is not the best, those who accepted the teachings have benefitted. (Focus Group with Women, endline, Ghana)

In Tajikistan, as well as reduced violence between partners, there was a reported reduction in violence from in-laws, particularly reduced emotional abuse by mothers-in-law against daughters-in-law, a common form of violence. This was connected to improved relationships as a result of greater communication within families.

Improved relationship quality was also identified as an important pathway to reduce intimate partner violence, particularly as a result of the interventions in Rwanda and Tajikistan for working at the household level. In Tajikistan, the intervention created spaces for the entire family to talk to each other openly, which improved men’s understandings of women’s needs and everyday household burdens and helped develop mutual respect. In Rwanda, several participants noted how completing the curriculum together allowed for rare, positive time together, and fostered closeness and intimacy.

Because we were in a secure and discrete place, everyone was free to say what he/she thinks. So after that lesson, because everyone had said freely what was on his/her mind, everyone went back to his/her home knowing what his/her partner likes and dislikes. (Henriette, midline, Rwanda)

Despite positive change around respecting partners and building healthier relationships, some women in Ghana and Rwanda misinterpreted the focus on respect to mean it was their responsibility to prevent intimate partner violence by respecting and acquiescing to men so as not to provoke violence. This was related to the strong norm of men as heads of households, as well as participants learning ways of communication and resolution skills to manage conflict:

I was disrespecting him, concerning the sexual intercourse I was not treating him as my husband. Even when he told me: ‘come here and I tell you’ sometimes I would decide just to leave him but now it has changed. I humble myself and I respect him. (Garuka, midline, Rwanda)

I learned that as a woman, you have to respect your husband because he is the head of the family, so you have to know how to talk to him even if you’re angry with him. I have learnt how to deal with my temper because our inability to manage our anger well can bring violence. (Focus Group with Women, endline, Ghana)

Gender-equitable roles and decision-making

Participants in each programme reported more gender-equitable relationships between couples and in families, an important pathway to violence reduction. In Tajikistan, many participants reported more gender-equitable household decision-making:

The project sessions on violence had a good impact on my husband. His attitude and behaviour towards me and my daughters changed. Now he pays attention to family problems. Now he asks for my opinion about family issues. (Gawhar, endline, Tajikistan)

In Rwanda and Ghana, a focus on families working together was central to the changes in practices, with men reconsidering their role in the home, and engaging in new behaviours, including care and domestic work, and couples making more equitable decisions. The concept of positive and negative types of power strongly resonated with participants in Rwanda, and was said to underlie transformative changes, including around inequitable gender roles:

The lesson that I see that has been the most useful and that we discuss so much at home is the lesson about power. That lesson helped him because the problem of power was about to destroy our marriage. He thought he was holding all the power. (Immaculee, midline, Rwanda)

The programme in Ghana also emphasised the benefits of peaceful relationships and couples working together, which similarly motivated changes in gender roles. For instance, one participant related how a COMBAT member emphasised the relationship benefits of men and women sharing domestic roles, which supported her husband to participate in domestic work after being teased by his friends:

My husband stopped helping me with house chores at some point because his friends were teasing him any time they see him helping me to wash. When the COMBATs came to speak to us, they made it clear there is nothing wrong when a man helps his wife with house chores. He said when we do it, we help each other and that will result in a peaceful marriage. (Adwoa, endline, Ghana)

Yet, across all programmes many participants shared the challenges of transforming power dynamics and gender roles in their relationships, with men continuing to have the final say in decision-making, and women being predominantly responsible for domestic and care work. In Rwanda, although women reported greater decision-making power in relationships as a result of the programme, several men and women shared how men increasingly helped their wives with domestic and care work, but on men’s terms. Although domestic care continued to be identified as women’s responsibility in Tajikistan, young women stated that men increasingly helped them with this, indicative of how individual attitudes and behaviours and roles in relationships are not always aligned.

Community-level pathways

Group-based participatory approaches

Across all the interventions, the participatory nature of the intervention facilitated processes of change, as participants could share and critically reflect on key issues and built a strong rapport with others. This approach, which enabled reflection and self-expression, contrasted with the didactic and authoritarian education system many participants in South Africa had experienced: ‘I was able to raise my opinion because we all allowed each other to speak … I was comfortable and felt free if I wanted to speak, I would speak, plus there were no wrong answers, every answer was right’ (Luphumlo, midline, South Africa). Similarly, in Rwanda, most participants were highly appreciative of the participatory approach of the curriculum, supporting a safe space in which to share and reflect on difficult/taboo issues without feeling judged.

Groups also provided an important space to build relationships with others. Despite living in dense informal settlements in South Africa, many women and men reported feelings of isolation and valued the interactions provided by the groups alongside having something more meaningful to do with their time: ‘During the intervention, we had a break from everyday life. It’s better to do something, not just sit at home all day, not talking and interacting with anyone’ (Nolwazi, endline, South Africa). The intervention in South Africa offered a rare opportunity for men to share painful experiences with others and in turn develop emotionally supportive relationships:

This brother, the way he was struggling and the situation he was in, he showed me that if you have nothing in life it doesn’t mean that it’s over. You have to try and struggle by all means until you are successful. That’s the story that stuck with me, and other stuff like you must share your problems with other people, you have to speak in order to get help in the end. (Langa, midline, South Africa)

In Rwanda, couples commented on the strong rapport they built with other couples by the end of the curriculum and valued how they could draw on each other for support. This group rapport was especially important to bolster the confidence and capacity of the participants who transitioned to become community activists. A similar process was seen in Tajikistan where the participatory curriculum enabled greater communication between age hierarchies to emerge and challenge inequitable gender dynamics. Facilitators in Tajikistan noted at baseline that younger women tended to criticise dominant gender norms and mother-in-law/daughter-in-law relationships, but only when speaking in a smaller group without older women present. By midline and especially endline, facilitators observed this had shifted, and in the larger group discussions younger women expressed their views and questioned dominant gender norms. Older women at times encouraged younger participants to speak and were more attentive and respectful of their comments.

Yet difficult group dynamics, particularly in South Africa and Tajikistan, sometimes impacted intervention effectiveness. While many women in SSCF reported they found the friendships and groups supportive, others reported they did not feel safe or supported in the groups. Group tensions included participants arguing or gossiping about what had occurred in the session, or being laughed at when they spoke in sessions: ‘It was when we were sharing the true stories about our lives, and when I tried to speak, Ndoni began laughing and I felt uncomfortable and I told them that I will not speak anymore’ (Woman, midline, South Africa). Although the intervention in Tajikistan separated groups according to age and gender hierarchies to promote sharing and dialogue, other community dynamics were present that could hinder such openness:

In our village the success of the project is challenging. See, I am local, and even if we live in the same village, we are from different avlods [extended families/clans]. I would never share my family problems in front of people from this other avlod. (Feda, post-endline, Tajikistan).

In Rwanda, ensuring equal engagement by men and women was initially challenging, and men often dominated the curriculum discussions, but through careful facilitation and greater group rapport this issue reduced over time.

Enabling environments

A key component that supported change in the community was how programmes worked at a community level, engaging leadership and working to connect families to important forms of support. In Tajikistan, improved community relations as a result of the programme were widely reported through the post-intervention data collection. Older women, in particular, mentioned an increase in mutual support and help for resolving issues, such as accessing drinking water, irrigation and other infrastructure. For younger women, an indication of better community relations was reduced social control through gossip, more supportive attitudes towards women’s employment outside the household and increased mobility.

In Ghana and Rwanda, several participants reported the value of the programmes in sharing information on property rights and inheritance laws for women to support more equitable access for women and children to property:

Before COMBAT, what was happening in this community was that when a man dies, his extended family members will claim all the man’s properties leaving the widow and the children with nothing. But since COMBAT started these things have stopped. (Kisi, endline, Ghana)

The strong relationships that developed between local leaders, activists and safe space facilitators through the Indashyikirwa programme (including through quarterly coordination meetings) enabled activists in Rwanda to access existing community forums for activism. This allowed for wider reach and diffusion of the messages and was said to enhance the credibility of the activists. In Ghana, activists were similarly supported by local leadership and as a result had access to a variety of community occasions (e.g. funerals, weddings) to allow diffusion of messages.

Support for the survivors of intimate partner violence in Ghana and Rwanda was critical for programme success. Activism emphasised that survivors should not be blamed for experiencing violence, but rather should experience support from their communities:

When there is violence, the community intervenes which is different from before. They no longer say ‘it is her fault, she is the one who provoked that violence that she is suffering.’ They have understood that even if she may be wrong, it should not make her a victim of violence. (Valentine, endline, Rwanda)

These efforts were important in changing conceptualisations of violence and fostering individual and community responsibility to intervene. The women’s safe spaces provided in Rwanda were highly appreciated and actively used. In Ghana, COMBATs were identified by several men and women as uniquely able to resolve domestic disputes and provide supportive responses to intimate partner violence survivors: ‘COMBATs are the only group in this community to teach us marital issues. Since they are trained to handle marital issues, I know whatever the case may be, they can help solve it’ (Kofi, endline, Ghana).

In Rwanda and Ghana, participants stressed the value of referral to social, health or justice services, community leaders or women’s safe spaces in Rwanda, to ensure longer-term and varied solutions as well as immediate responses. In addition, it was important for activists (and safe space facilitators in Rwanda) to accompany individuals to seek services, to reduce potential intimidation and provide needed support: ‘A woman can go to court with some of the COMBAT members. Because there will be a need for witnesses that is why it will be the best to go with the COMBAT so they could get the truth out’ (Focus Group with Women, endline, Ghana). However, a few men in Ghana reported being uncomfortable seeking help from COMBATs, for fear of being reported or prosecuted. Several participants lamented how men were not as actively engaged in the programme because they were too busy or uninterested, or perceived the programme as favouring women:

The men complain to us that COMBAT takes sides with the women. I am suggesting the next time, issues from men should also be addressed in the intervention. On the logo it is written ‘violence against women and girls or children’, and that means they are not involved. They feel they are not favoured. (Focus Group with Women, endline, Ghana)

Discussion

Across varying contexts and programmes, similar pathways of change to reduce intimate partner violence were identified through this comparative analysis. At the individual level, participants, particularly women, identified how through participation in interventions they gained greater self-esteem and confidence, reduced their tolerance of abuse and inequalities in relationships and better prioritised their own relationship needs. At the relationship level, participants across the studies reported shifts in harmful gender norms and roles, including men and women more equally sharing household and provider roles, and more equitable decision-making. Yet the limits of the programmes’ abilities to effect fundamental transformation around gender norms within programmatic timeframes were also evident.

Other papers from these data sets have discussed the difficulty of shifting men’s headship roles and patriarchal power in Rwanda (Stern, Heise and McLean 2018) and South Africa (Gibbs et al. 2020). An earlier evaluation of Stepping Stones with adolescents in South Africa found that women showed greater assertiveness and some agency in HIV risk reduction, yet the majority did not challenge their male partners or the cultural norms of conservative femininities (Jewkes, Wood and Duvvury 2010). In the analyses presented here, some female participants reported respecting and acquiescing to men’s headship role as a strategy to prevent violence, and more widely there was a fundamental inability to adequately challenge patriarchal power. This affirms the importance of regularly and meaningfully targeting inequitable gender norms and power relations, and for individual-level interventions to be complemented by structural initiatives to challenge broader patriarchal norms (Willan et al. 2019).

A critical pathway to reduced intimate partner violence was the integration of economic empowerment activities into programmes, which had multiple benefits for participants and generated structural-level change. Economic components helped reduce economic stress, a critical driver of intimate partner violence (Buller et al. 2018). In addition, women’s engagement with livelihoods enabled them to be seen as contributing financially to the household, earning them greater respect and authority and increasing their sense of self-worth. In Tajikistan and Rwanda, the involvement of women’s family members supported women to become involved in the economic empowerment components, and reduced risk of this fuelling conflict or intimate partner violence. For men, improved livelihoods also supported their self-esteem and confidence, especially given the prioritisation of men’s provider roles in all settings.

Another pathway to reducing intimate partner violence was the emphasis on healthier, more equitable relationships, by understanding the causes and consequences of violence and applying relationship skills to mitigate it. This supports other data suggesting that poor communication and handling of partner conflict are contributing factors to intimate partner violence (Conroy et al. 2016). Relationship skills learned within programmes included identifying triggers of violence and more assertive and constructive approaches to communication and conflict resolution. Applying these skills allowed participants to experience and reflect on the benefits of greater cooperation, communication and non-violence among intimate partners, or within families. It was particularly effective to strengthen relationship skills within households (families in Tajikistan and couples in Rwanda), enabling participants to have a shared understanding of and hold each other accountable to these skills. Indeed, the literature suggests that when both partners of couples receive information about conflict resolution tools, they are more likely to feel invested in applying these to negotiate conflicts (Todahl et al. 2012). The development of skills further supported participants’ self-confidence, together with the programme elements that promoted empowerment, and greater community status of participants. Emphasising the benefits of equitable, non-violent relationships also supported motivation and active engagement of participants, especially men. This aligns with other literature suggesting that men need the space to consider how gender inequalities can be harmful to both women and men and the benefits of gender equality (Clowes 2013).

The participatory nature of the programmes was important for the development of bridging and bonding social capital (Campbell and MacPhail 2002), which were additional pathways to change. The groups offered rare spaces for participants to reflect on harmful behaviours and develop alternative norms, supporting the literature that such approaches can provide platforms for change through reflexivity and consciousness raising (Campbell and MacPhail 2002). Group-based approaches additionally allowed for powerful rapport and the emergence of a supportive environment for change in ideas and practices (Stith et al. 2004). However, in our analysis challenges were also identified to build group rapport and trust, which speaks to the importance of facilitators being supported and trained to secure this, and to carefully monitor the dynamics of group-based interventions.

The analysis also highlighted the importance of bridging social capital between activists and local leaders, especially to ensure an enabling environment for activism approaches. In Tajikistan, an unexpected consequence of the programme was greater social cohesion, which created an enabling environment for programmatic changes. Providing support to female survivors of intimate partner violence in Rwanda and Ghana was an identified pathway of change for more empowering responses and shifting community conceptions of violence. This speaks to the importance of working at the intimate partner violence prevention and response continuum, especially when applying activism approaches to foster community-level changes. Men were identified as more difficult to engage in community-level activities in Rwanda and Ghana, including to shift rigidly held attitudes and beliefs around gender and violence, and to be encouraged to seek support to reduce their perpetration of violence, which speaks to the importance of programmes carefully considering how community-level efforts can support and engage men.

Limitations

Like all research, this analysis has its limitations. There may be social desirability bias with participants wanting to report favourably on interventions they clearly valued. Each study attempted to mitigate this by using researchers external to the programme and emphasising the confidentiality of answers. The fact that this secondary analysis found similar domains of change, and openness from participants around programme limitations or challenges, enhances the validity of the findings. Moreover, observational data, and data from staff, where available, helped triangulate the perspectives of participants.

An additional limitation of this study is that the first author, who was external to the Ghana research team – conducted the secondary analysis of the thematically analysed data from Ghana. However, the findings were interpreted and validated by the Ghanaian research team, including two co-authors of this paper.

We also recognise the potential investigator biases and limitations of secondary analysis. However, investigators were asked to share original data as part of their secondary analysis and report on barriers and enablers to change, as well as strengths and limitations of changes, which helped mitigate such biases.

Conclusion

This qualitative comparative case study provides important insights into programmatic pathways of change. Tracing how participants interacted with intervention ideas and opportunities, as they took place in different contexts through varying approaches, can generate understanding of the design and implementation elements pertinent to bringing about change in intimate partner violence.

Acknowledgements

The views expressed and information contained are not necessarily those of, or endorsed, by DFID, which accepts no responsibility for such views and/or information, or any reliance placed on them.

Disclosure statement

We wish to confirm that there are no known conflicts of interest associated with this publication.

Additional information

Funding

This paper is an output from the DFID UK What Works to Prevent Violence Against Women and Girls Programme funded by UK Aid from the UK Department for International Development (DFID) with funds managed by the South African Medical Research Council.