Shifting and transforming gender-inequitable beliefs, behaviours and norms in intimate partnerships: the Indashyikirwa couples programme in Rwanda

Abstract

This paper explores the relationship between changes in individual beliefs and behaviours, couple relationship dynamics and gender norms – and how interventions can influence these. It draws on longitudinal qualitative research with heterosexual couples who participated in the Indashyikirwa programme in Rwanda. The couples followed a curriculum designed to improve relationship skills and reduce the gender-inequitable beliefs, behaviours and norms that underpin intimate partner violence. Qualitative findings show that the programme resulted in moderate, but significant, positive ‘shifts’ in individual beliefs and behaviours, couple relationship dynamics and levels of inequality - increasing men’s engagement in domestic duties, women’s participation in household decision making, and women’s access to economic resources. They also suggest which parts of the couples’ curriculum were most effective in catalysing these changes. However, the data also show that these ‘shifts’ occurred without fully transforming deeply-entrenched beliefs and norms around gender roles and male authority over economic resources. The paper suggests that the persistence of these beliefs and norms constrained the extent of changes among couples – and could potentially constrain their longevity and act as an obstacle to longer-term, larger-scale changes in gender inequalities and violence.

Keywords:

• Gender norms
• gender inequalities
• intimate partner violence
• gender roles
• household decision-making
• male authority
• Rwanda

Introduction

There is increasing evidence to show that gender-inequitable¹ norms are powerful determinants of health and wellbeing (Heise et al. 2019). Intimate partner violence (IPV) is sustained by a number of inter-related beliefs and norms, particularly those that justify male authority over female behaviour and men’s violence against their partners (Heise and Kotsadam 2015). IPV can lead to negative health consequences including serious injuries, death and increased risks of poor maternal, reproductive and mental health outcomes (Ellsberg et al. 2008). Several programmes have therefore been developed to challenge gender-inequitable beliefs, behaviours and norms as a strategy to reduce IPV. While some of these programmes have succeeded in changing individual beliefs and behaviours and reducing levels of IPV (Ellsberg et al. 2015; Fulu and Kerr-Wilson 2015), research on the relationships between changes in individual beliefs and behaviours, couple dynamics and gender norms – and how external interventions can influence these - is still emerging.

This paper draws on qualitative research with heterosexual couples who participated in the Indashyikirwa programme in Rwanda and followed a curriculum designed to improve relationship skills and shift the gender-inequitable beliefs, behaviours and norms that underpin IPV. It analyses whether and how beliefs and behaviours around the gender division of labour and dynamics of household decision-making shifted during the course of the intervention and which parts of the curriculum contributed to these changes.

The paper reveals moderate shifts in individual beliefs, behaviours and couple relationship dynamics which reduced gender inequalities and are likely to have contributed to the reduction in IPV among couples found by a parallel quantitative study (Dunkle et al. 2019). However, the data also show that these shifts occurred without changes to deeply-entrenched beliefs and norms around gender roles and male authority over household resources. The paper considers the consequences of this for the longer-term impact of the Indashyikirwa intervention and other programmes that seek to bring about long-term, large-scale ‘transformations’ in gender inequalities and violence.

The relationships between gender-inequitable beliefs, behaviours and norms

Social norms are the unspoken rules that govern behaviours within a group, defining which behaviours are considered acceptable or desirable and which will be condoned, tolerated or condemned. Social norms theories argue that a person’s behaviour depends on both their expectations about other people’s behaviour and their expectations of the reaction of others (e.g. approval, condemnation) to their own behaviour (Bicchieri 2010; Mackie et al. 2015). This means that even when an individual’s personal beliefs diverge from the predominant norms, they will often - although not always - behave in accordance with those norms to avoid negative social sanctions (ibid.).

Social norms sustain inequitable power relationships including those based on gender. Indeed, ‘gender norms’ are seen by scholars as ‘the means by which gender-inequitable ideologies, relationships and social institutions are maintained’ (Marcus and Harper 2014). Gender norms are deeply ingrained in most societies and have proven particularly resistant to change because they are usually justified on the basis of biological sex differences (Correll, Thebaud, and Bernard 2007) and trigger deeply entrenched cognitive schema that associate different roles and statuses with different genders (Heise et al. 2019). Children are socialised into particular gender beliefs and practices from an early age, these associations become automatic and gender inequalities are normalised. Gender norms are then reproduced and reinforced through everyday interactions and by institutions such as marriage, schools, the market, media and government bodies (Marcus and Harper 2014).

Nonetheless, gender scholars stress that gender is not something that simply ‘is’; rather gender is continuously ‘done’ or ‘performed’ through social interaction (Butler 2004; West and Zimmerman 1987). Men can ‘do gender’ by using violence against a partner to demonstrate their control in contexts where male authority over women is a hegemonic norm; women can ‘do gender’ by minimising or justifying male violence (Merry 2010). Similarly, men can ‘do gender’ by maintaining control over household resources and decision-making in contexts where men are seen as heads of households; women can simultaneously ‘do gender’ by submitting to male authority in contexts where women are expected to be submissive. In this way, hegemonic gender beliefs - or gender norms - are simultaneously (re)produced through individual behaviour (Ridgeway and Correll 2004) and act as a ‘cultural frame’² for that behaviour (Ridgeway 2009).

Deutsch (2007) argues that if gender is repeatedly ‘done’ through social interaction, it can also be ‘undone’. For example, where individual men and women fail to live up to normative ideals of womanly or manly behaviour - whether out of necessity or purposeful acts of resistance or innovation – this can reduce gender inequalities in their relationships and eventually contribute to wider changes in social norms and inequalities (West and Fenstermaker 2002). Historically, large-scale economic dislocations, wars and social movements have sometimes made it extremely difficult for individuals to act in line with expected gender norms (Legerski and Cornwall 2010). Equally, interventions such as policy and legal reforms have supported women to enter the workforce and challenged men’s roles as primary breadwinners, leading to shifts in both behaviours and expectations of behaviours (norms) (Chant and Gutmann 2002).

Can external interventions change gender-inequitable beliefs, behaviours and norms?

In recent years, there has been increased focus on implementing community-based programmes designed to ‘shift’ or ‘transform’ the beliefs, behaviours and norms that contribute to gender inequalities and violence (Alexander-Scott, Bell, and Holden 2016; Haylock et al. 2016). Several of these programmes have worked with small groups of women and/or men to improve their knowledge, risk awareness and communication skills around HIV, gender, violence and relationships. Evaluations have often found positive changes in the knowledge and behaviours of programme participants and their close contacts; yet little impact on influencing wider beliefs, behaviours and norms around gender roles behaviours across the community (e.g. Stepping Stones in India - see Bradley et al. [2011]).

Community mobilisation programmes such as SASA! - first implemented in Uganda - have more explicitly attempted to apply behavioural change theories to connect individual change to wider social norm change (Michau 2012). SASA! applies the Transtheoretical Model (Prochaska and Velicer 1997) to create a network of trained community activists who engage, inspire and support community members and institutions over an extended period of time, seeking to reach a critical mass of ‘adopters’ of new beliefs and behaviours. SASA! approaches also allow people to observe and practice new behaviours and to listen to public testimonials which endorse new behaviours and help introduce new positive norms (Mackie el al. 2015).

SASA! is also informed by social network theory which posits that when a specific sub-group adopts new behaviours, these eventually diffuse to a wider social network and when a ‘critical mass’ of adopters is reached, ‘transformation’ of the predominant norm takes place (Institute for Reproductive Health and FHI360 2016). In this respect, it is analytically useful to distinguish the terms ‘shift’ and ‘transformation’ which are often used interchangeably and undefined. In this article, we use the term ‘shift’ to refer to significant but finite changes in individual beliefs, behaviours and relationship dynamics – recognising that these may be temporary or reversed. We use the word ‘transformation’ to refer to large-scale changes in beliefs, behaviours but also – critically - norms among the majority of people in a specific context, which are more likely to be sustained than shifts (although are not irreversible).

An evaluation of SASA! in Uganda found community-wide (population-level) changes in beliefs and behaviours as reported by women and men – including reduced tolerance of physical IPV; increased acceptability of women’s right to refuse sex; and reductions in women’s experience of physical IPV (Abramsky et al. 2014). It also resulted in increased joint decision-making among couples, greater male participation in household tasks and more open communication (Kyegombe et al. 2014a). Analysis of qualitative data collected from couples in SASA! communities sheds light on the dynamics of change in their relationships, showing that both women and men were significantly motivated by ideas of working together to improve their relationships and the situation of their households (Kyegombe et al. 2014b). This study suggested that these may have been the main motivations for men to consult women in decision-making and participate in housework; rather than this resulting from fundamental changes in their beliefs about gender.

This raises important questions about the drivers of and pathways to behaviour changes and how changes in individual beliefs, behaviours and gender norms affect each other. For example, it is possible for a man to change his behaviour towards his wife – such as consulting her about decisions – without him or indeed his wife - changing their underlying belief that he is ultimately in charge of decision-making. Equally, it is possible for a man to participate in household tasks, without he or his wife changing their belief that domestic labour is ultimately the woman’s role. Thus, there may be important changes in behaviours among couples – even the majority of couples in a community – without changes in individual beliefs or wider norms about gender roles or male authority. We therefore need to better understand the relationships between changes in beliefs, behaviours and norms, why some are more susceptible to change than others, and what types of contextual changes and interventions can influence these processes.

The Indashyikirwa programme and how it aimed to change beliefs, behaviours and norms

Indashyikirwa (‘agents for change’ in Kinyarwanda) was a four-year programme (2014-2018) developed to reduce gender inequalities and IPV in Rwanda. It was implemented across 14 sectors in the Eastern, Western and Northern provinces and used CARE’s Village Savings and Loans Associations (VSLA) (which targets low income households) as an entry point for recruiting 840 heterosexual couples to attend a 21-session, five-month curriculum designed to build skills for healthy, equitable relationships and to shift the beliefs, behaviours and norms that underpin male dominance and violence in relationships.

The core curriculum covered concepts of power and gender; human rights; managing triggers of IPV - including alcohol abuse, jealousy and economic stress; healthy relationships; activism and providing empowering responses to those experiencing IPV (Stern et al. 2018). The theory of change posited that - through group reflection and support; developing greater awareness of types of power, the benefits of gender equality and non-violence; as well as developing key relationship skills such as communication, negotiation and managing conflict - couples would change both their beliefs and behaviours and manage relationship tensions in non-violent ways (ibid).

The impact evaluation of Indashykirwa included a cluster randomised control trial (cRCT) with a cohort of control and intervention couples (Dunkle et al. 2019). The results from the couples’ cohort were extremely positive with women participants in the Indashyikirwa curriculum reporting lower levels of physical, sexual and emotional IPV compared to the control group. This was matched by data from male participants who reported significantly lower perpetration of physical and/or sexual IPV compared to men in the control group (ibid). In addition, both women and men in the intervention groups demonstrated lower acceptance of violence and reported improved relationship quality, better communication, improved conflict management, improved household economic development and better mental health than in the control groups (ibid).

Methods: longitudinal qualitative research with couples

Alongside the cRCT, longitudinal qualitative research was conducted in three intervention sectors (Rurembo Sector, Western Province; Gishari Sector, Eastern Province; and Gacaca Sector, Northern Province) chosen to represent rural, urban and peri- urban locations. Before the programme was implemented, 15 couples – five in each of the three sectors – were recruited from the list of all the couples who had volunteered to participate in the couples curriculum (who all had at least one partner in a Village Savings and Loan Association group, were between 18-49 years of age, and had been married or living together for at least 6 months). All couples enrolled in the programme had also given their agreement to participate in the research study. From the list of all those enrolled, we took a convenience sample, approaching couples sequentially until enough couples agreed to participate. The only purposive criteria used was marriage status as this was hypothesised as potentially influencing risk and protective factors for IPV (Stern and Mirembe 2017). The research team purposely selected an equal mix of formally and informally married couples as they went down the list. Basic demographic information on the interviewees is given in Table 1.

Table 1. Demographic information of partners of couples interviewed.

Pseudonym^a	Sex	Age (endline)	Province	Marital status	Relationship duration (years)
Béatrice	Female	29	Western	Formal	10
Thierry	Male	38	Western	Informal	10
Ritha	Female	29	Western	Formal	8
Richard	Male	32	Western	Formal	8
Solange	Female	37	Western	Formal	17
Edouard	Male	38	Western	Formal	17
Betty	Female	33	Western	Informal	14
John	Male	34	Western	Informal	14
Mireille	Female	35	Western	Formal	15
Justin	Male	34	Western	Formal	15
Nadège	Female	29	Northern	Formal	8
Célestin	Male	37	Northern	Formal	8
Alice	Female	22	Northern	Informal	3
Eric	Male	24	Northern	Informal	3
Mary	Female	28	Northern	Informal	5
Paul	Male	30	Northern	Informal	5
Patience	Female	29	Northern	Formal	9
Janvier	Male	30	Northern	Formal	9
Valentine	Female	43	Northern	Formal	22
Charles	Male	45	Northern	Formal	22
Grace	Female	31	Eastern	Informal	10
Gatete	Male	42	Eastern	Informal	10
Kalisa	Female	28	Eastern	Informal	5
Jean-Michel	Male	29	Eastern	Informal	5
Sidonie	Female	26	Eastern	Formal	4
Joseph	Male	37	Eastern	Formal	4
Pacifique	Female	31	Eastern	Formal	10
Yves	Male	32	Eastern	Formal	10
Winnie	Female	25	Eastern	Informal	3
Robert	Male	38	Eastern	Informal	3

^aThe names of individual research participants are not used to protect confidentiality and anonymity.

Male and female partners were interviewed in kinyarwanda by experienced Rwandan researchers with a male researcher interviewing men and the same female researcher interviewing women at baseline, midline and endline, which was helpful for building rapport. Participants were interviewed at places they preferred where they could have privacy - often at a clinic or local government centre. In terms of ethical and safety considerations, individual women and men were interviewed separately and were informed that their responses would not be shared with their partner. They were all participating in the intervention and, in the case of a request or report of IPV, were asked if they wished to be referred to the local women’s safe space or meet with the research team’s dedicated counsellor. Ethical approval for the study was obtained from the Rwandan National Ethics Committee, the London School of Hygiene and Tropical Medicine, and the South African Medical Research Council.³

30 baseline interviews were conducted in November 2015 with couples enrolled in the programme (who had not yet started the curriculum). The interviews explored gender roles by asking partners about their expectations and realities of each other’s roles and behaviours. Partners were each also asked about their quality of communication, household decision-making, relationship difficulties and what happens when there is a disagreement. If IPV was mentioned, interviewees were asked about the perceived causes and implications of their last episode of violence. Otherwise, they were asked for their views about IPV.

28 midline interviews were conducted with the same couples immediately after completion of the Indashyikirwa curriculum in May 2016 (one couple moved to Uganda and dropped out) and 28 endline interviews one year later in May 2017.⁴ The same questions were asked around gender roles, communication, household decision making and managing disagreements. Midline interviews also probed for participants’ experiences of the couples’ curriculum and whether their expectations and/or behaviours had changed as a result of their participation. At endline, couples were also asked whether changes had been sustained, whether there were further changes or if behaviours had reverted.

Using the audio files, the data were transcribed and translated verbatim into English by a professional translator. We first coded the data using a priori codes developed from the interview guides. We then added further codes that emerged during the process of thematic analysis to provide a rich, detailed and holistic account of the data (Braun and Clarke 2006). We used matrices to organise the data across codes which enabled us to compare across all respondents as well as between and within couples.⁵ We also looked at whether participants’ responses varied by marital status or age but found no significant differences. Specific attention was paid to any inconsistencies in how interviewees described their relationship and key events over time. For example, it was common for individuals to be more forthcoming at midline and endline about problems in their relationships compared to baseline. Thus, careful analysis and triangulation was undertaken across individual interviews and between partners. Emerging findings and their interpretation were regularly discussed with Indashyikirwa programme staff through participatory workshops.

Key findings

The gender division of labour

At baseline, there was strong concordance across the group and between male and female partners that the primary role of a husband was to provide for his family – contributing a sufficient amount of his income to meet household needs for food, clothing and shelter – and to pay the costs of the mutuelle de santé (health insurance) and schooling.

As a leader, you understand… a man should provide what is needed at home… to get the wellbeing… their daily food. (John, Western Province, baseline)

During these initial interviews, as well as in later interviews, several male and female partners admitted that the men were often not meeting these expectations prior to the programme. A common complaint was that men were not contributing sufficient income to the household, instead using it for themselves:

Sometimes the family might starve… because he may earn his pay and buy alcoholic drinks for other people and he finishes all the money… then he comes at home without money to provide food for the family. (Pacifique, Eastern Province, baseline)

However, following participation in the curriculum, at midline and/or endline, one or both partners among 12 of the 14 couples testified that the man was now bringing more of the money he earned home and providing more for the family:

He brings the money to support the family whereas before I would not even know if he has got his salary or not! He would not even leave 100 Rwandan Francs for me to buy soap; I didn’t even know what his salary was. (Nadège, Northern Province, midline)

Previous research in Rwanda has confirmed the deeply entrenched social expectation that it is a man’s role to provide economically for the household and be the main breadwinner (Slegh and Kimonyo 2010; Stern, Heise and McLean 2017). The Indashyikirwa curriculum was not designed to directly challenge men’s role as primary breadwinner; rather to increase acceptance of both partners working to earn income and deciding together on the use of that income. With regards to the former, after the curriculum there were two cases in which women said that their husbands now permitted them to work outside the home. Five male partners also spoke about the value of their wives working to contribute to the household through work outside the home or growing crops for them to sell:

When it is about [her] working for money I don’t have any problem because it is beneficial to our family. On contrary I appreciate it so much. (Edouard, Western Province, midline)

In terms of expectations of women’s roles, at baseline, both men and women concurred that a wife’s primary role was to take care of the children and domestic duties:

Women are expected to be actively involved in childcare over other responsibilities… (Edouard, Western Province, baseline)

Preparing and cooking food is wives’ responsibility (Valentine, Northern Province, baseline)

Following the curriculum, the midline interviews reveal changes in both beliefs and behaviours with respect to domestic duties and childcare with both women and men stressing the importance of men engaging in these activities. Among all couples interviewed, one or both partners testified that the man had started to engage to some degree in domestic activities and the endline data suggests that these changes were sustained a year later:

I used to think there were tasks designed for men and others designed for women, but when Indashyikirwa came I realised that a man can do any task… I thought that a man could not make the bed but now we both make the bed… I thought that a man cannot wash children but now I wash them without any problem. (Célestin, Northern Province, endline)

However, there were important nuances. First, there were certain tasks that men seemed to prefer - largely those within the household such as bathing or cooking for the children or outside the household, fetching water and firewood. There seemed to be more resistance around undertaking tasks such as cleaning or sweeping the compound, or going to the market, which are more visible to others apparently more entrenched as ‘women’s work’:

[What he] has not yet done in the household is sweeping. He can trim the compound - that is for men - but sweeping is typically for women only… Bathing children, washing utensils, yes, but sweeping for men is not possible. (Mary, Northern Province, endline)

He gives me money and I go shopping at the market. He can’t go in the market to shop. They say that women are better to do the shopping. (Kalisa, Eastern Province, midline)

Second, several men and women spoke specifically about when men did domestic tasks - mostly when their wives were away, busy with other tasks or sick:

[T]oday … he even cooks food. But he cannot cook food when I am there doing nothing; he only cooks food when he knows that I am doing other work. (Grace, Eastern Province, endline)

[S]uppose that she has gone to the market, if a child is sleeping I bathe him and I cook some food and when she comes she cooks other food. (Jean-Michel, Eastern Province, endline)

Third, the way interviewees—especially men—spoke about their engagement in domestic and childcare activities suggests that they made a conscious choice to ‘help’ their wives:

He can wash the clothes if there is a reason for that… if I am sick or if I travelled somewhere. That is when he can do it… (Ritha, Western Province, midline)

In other words, men still believed that these tasks are fundamentally the woman’s responsibility, but that a man could choose to help on his own terms (when he has time, when he wants to relieve her burden or when she is out). At endline, many women also seemed to agree that domestic duties were still ultimately the primary responsibility of women:

Cooking, cleaning, washing clothes, fetching water, looking for the firewood, shopping, bringing up the children, caring for the sick people or old people… all those are my responsibilities. (Sidonie, Eastern Province, endline)

Household decision-making

Interviews confirmed that, before participation in the curriculum, men were seen as the primary and final decision-makers; that they rarely consulted their wives about decisions over income and assets; and that women should not be seen to question their husbands’ authority.

Even if I would ask her for her opinion - she could tell me for example ‘I think we should do like this’ - but then I would say, ‘I am the husband here, things have to be like this’ and things were done my way. (John, Western Province, endline)

Some responses – including following the curriculum - explicitly referred to predominant beliefs and norms about male authority over economic decision-making in Rwanda and the consequences of men being seen to let their wives participate in decision-making (e.g. being called ‘bewitched’):

[A]s a husband… you feel that if you have sold something, your wife should not ask you about the money – you are the one who decides the share that she has to take. (Robert, Eastern Province, midline)

They say that a man who seeks his wife’s advice is a man who is not respected, who is not significant among other men. (Edouard, Western Province, midline)

The interviews also revealed the lack of trust present in many relationships prior to participation in the curriculum. At midline and endline, several interviewees revealed that one or both partners had previously been concealing their income and spending from their partner:

Before, he could even tell me that he didn’t withdraw any money from the bank whereas he had taken the money… you wouldn’t know when he was paid, that was his own secret… (Grace, Eastern Province, endline)

I was showing her a part of the money and… hiding the rest… and when she also had money, she was telling me less than what she had. (Richard 2, Western Province, midline)

After participating in the curriculum, both men and women interviewees stressed the importance of talking about the use of income and the purchase and sale of household assets, attesting to the fact that beliefs around decision-making had somewhat shifted:

I thought that my wife shouldn’t have any right about the household resources… I thought they were all mine alone. But I understood the lessons and… realised that my wife should also have right on what belonged to me so we can share them. (Gatete, Eastern Province, midline)

In over half of couples interviewed, men and women spoke specifically of changes in decision-making behaviours such as increased transparency and discussion around decisions over income and assets. The following couple concurred about this change:

[I]nstead of consulting me… he made all decisions alone… Today we sit together and make decisions, we talk and we mutually share feelings. The training really was very useful for us… Honestly, we have changed so much. (Pacifique, Eastern Province, endline)

While I used to make the decision alone because I thought that the cow belonged to me, today I cannot come from work and just sell the cow. If my wife is not there, I wait for her so that we can have time to talk about it… (Yves, Eastern Province, endline)

Some couples gave concrete examples of joint decisions and economic projects. This husband and wife attested to the importance of her input into two recent decisions.

[S]ince the training…I told him let us buy a cow with that money. So we bought an ox for 85,000 francs… and we raised it and we sold it recently… at 120,000 francs. (Sidonie, Eastern Province, midline)

I realised that we had planted enough beans and I respected her suggestion… I gave up imposing my decisions as a husband and I agreed to respect her… then we planted groundnuts… when I look at them today, I see that they are flourishing… (Joseph, Eastern Province, endline)

There were some examples of women acquiring a degree of decision-making autonomy, although usually curtailed to ‘smaller’ decisions such as selling a bundle of bananas to buy some food or clothing, rather than major decisions such as buying and selling land or livestock.

Although the midline and endline data suggest that couples increasingly engaged in discussion over the use of income and assets, there were limitations to this. Interviews strongly indicate that, in most cases, men consulted their wives and allowed them to express their views before taking the decision. Ultimately, however, men retained control of the decision-making process and had the final say. Men went along with their wife’s suggestion, but only if they thought it was a good idea:

We decide together and agree on what to do. Although… being he is the head of the family he can make the final decision… (midline) The things we decide together are the things we both agree on… When he decides and I see it is true, I agree with it. But in that case, he already has made the decision… (Grace, Eastern Province endline)

Grace makes explicit reference to the belief that men are heads of households in Rwanda. This was referred to frequently including at midline and endline, suggesting that it was not ultimately transformed:

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. (Célestin, Northern Province, midline)

However, a related belief did seem to be challenged as a result of the intervention – about male ownership of household economic resources. In midline and endline interviews, both men and women spoke about their prior belief that all income earned as and household assets (aside from a woman’s clothes or what she brought to the household) belonged to the man:

I knew that a husband can leave home and go to work for his own money that he can use for his own interest… he would buy whatever he wanted… It was like that - everything including a machete and a hoe belonged to the husband… I thought that she didn’t have any right on the properties in our household, be it domestic animals or crops harvest. (Gatete, Eastern Province, endline)

Midline and endline interviews also demonstrated a shift in knowledge about women’s rights to access or use household property and examples of specific actions taken by men or jointly to improve this. For example, two couples established joint bank accounts and three couples bought land together in both their names:

I hadn’t given her the right on my bank account but nowadays I have given her full right… I added her as a signatory; when I have money on the account, I can also tell her to go and withdraw it. (Jean-Michel, Eastern Province, endline)

[B]efore when we bought a land, he would write his name as buyer of the land but now when we buy a land, he writes that it is for both of us. I see that these lessons have changed him. (Solange, Western Province, midline)

Which parts of the curriculum are likely to have contributed to change?

Further analysis of how women and men talked about their experiences provided insights into which parts of the curriculum were most influential in ‘shifting’ their beliefs and behaviours around the gendered division of labour and household decision-making.

First, the session on gender socialisation and norms⁶ helped many participants, especially men. This is likely to be because - rather than blaming men individually - it stimulated them to reflect critically on where their ideas about male control of assets and decision-making authority originated:

[B]efore… the training, I used to think that I am the one who has the right to make decisions. my father behaved the same way in their time… ‘My goat, my cow, my banana plantation’ instead of ‘Our cows, our banana plantation, our plots of lands’ … everything belonged to the husband. So, after getting married I still had that mindset because I developed it as I grew up. (Joseph, Eastern Province, endline)

Second, the participatory exercise on daily tasks performed by men and women⁷ made a strong impression. The take-home work for that week was for men to engage in domestic tasks. As a result, many men realised that their wives suffered with a large burden of domestic work (see also Stern and Nyiratunga 2017).

He said: ‘You had really suffered! Now I will help you with housework’… Now when I am cooking and the baby cries he tells me, ‘Let me cook for you and you breastfeed the baby’. (Sidonie, Eastern Province, midline)

What I learned… is that I will not let my wife do all the work alone anymore because I realised that my wife was very exhausted…. (Jean-Michel, Eastern Province, midline)

Third, another motivating factor mentioned – related to the exercise on daily tasks and the session on ‘healthy relationships’ - was how sharing the burden of work could improve the quality of a couple’s relationship, including their sexual relationship:

We saw that because a woman works so hard and she gets tired, sexual intercourse doesn’t go well… That used to happen… After that lesson he told me: ‘So women do really exhausting work! All the tasks we have seen, you do all of it!… I realised that I don’t even do any work… From now on I will come early and help you’. (Nadège, Northern Province, midline)

Fourth, in terms of shifts in decision-making, the session that distinguished negative ways of using power (power over) from more positive ways (power to, power with, power within) resonated with many of those trained and was explicitly mentioned in several interviews:

We used to think that the man holds all the power, but… We saw that the power over is the power that oppresses others and commits violence. Therefore, I no longer use that type of power with my wife. Instead I use the power with… we first talk and we do things we agree on. (Thierry, Western Province, midline)

Finally, the emphasis on working together for ‘household development’ clearly resonated with many couples. The curriculum stressed the benefits of a ‘partnership model’ - involving open communication, joint decision-making and sharing the burden of work - to improve the economic situation of the household.

When two people have a good relationship, their household develops. But one person only cannot achieve anything. (Gatete, Eastern Province, endline)

We were taught about making savings… how we can manage our resources … we found out that a family where there is good relationship… is a family which sits down to discuss family problems and plans ways of solving poverty issues… that a family with conflicts is a family where development cannot be achieved. (Richard, Western Province, midline)

Discussion

Changing beliefs, behaviours and norms: ‘shifts’ vs ‘transformations’

These research findings document some promising changes in individual behaviours and couple dynamics related to the gendered division of labour – specifically: men participating more in domestic tasks; women participating more in decision-making around the generation and use of household income and assets; and in the access of women to household economic resources. There is also data to show that both women and men had somewhat shifted their beliefs to recognise the rights of women to own and access resources.

These are important shifts which reduced inequalities between male and female partners who participated in the Indashyikirwa programme and are likely to have contributed to the reductions in IPV found by the parallel study. At the same time, this research reveals limitations in the extent of these shifts. Men who participated in the couples’ curriculum, to varying extents, improved dialogue with their wives, consulted them about economic decisions, listened to and valued their advice and even implemented their proposals. Yet, they still retained their role as the ultimate arbiter of how income and assets were generated and used. Equally, while men did engage in domestic activities, this was usually limited to specific tasks and situations and they did so on their own terms.

Thus, the findings show that men and women will – individually or together – stray to some degree from expected behaviours, ‘bending’ or resisting dominant norms (see also Stern, Heise and McLean 2017). However, couples continued to refer – even at endline - to assigned male/female roles – especially men as head of the household, male providership and male authority over economic resources. This suggests that the intervention was insufficient to fully transform these deep-rooted inequitable beliefs and norms and shift power relations.

This also shows that the relationship between beliefs, behaviours and norms is not necessarily linear. Women and men’s behaviours and - to a lesser extent - beliefs did change despite the persistence of norms around gender roles and male authority. Indeed, in this case reductions in violence occurred without those norms being transformed. However, the findings of this study suggest that the very persistence of these norms also constrained the extent of changes in beliefs and behaviours among couples. It also potentially leaves these new behaviours and beliefs vulnerable to reverting - under normative pressure.

The role of programme interventions in catalysing change

The findings – along with those of the quantitative evaluation (Dunkle et al. 2019) - show that shifts in individual beliefs, behaviours and relationship dynamics can be catalysed by engaging couples in a well-designed, well-facilitated and sufficiently intensive curriculum. The findings also give insights into the specific approaches likely to have been most effective.

First, they confirm previous findings (e.g. Kyegombe et al. 2014b) that using the lens of power - helping men and women to understand how it can be used positively and negatively and to examine the consequences - was a successful approach.

Second, the findings support the proposition that it is easier to introduce new behaviours with a view to fostering a new positive norm rather than to challenge an entrenched one (Kincaid 2004). Rather than directly challenging male authority, the curriculum used the notion of ‘household development’ – also advocated by the Rwandan Government - to make the case for men and women to work as ‘partners’ for the benefit of their households. The sessions outlined the tangible benefits of working together, taught new skills and gave couples take-away work to practice new behaviours, allowing them to experience and reflect on the benefits of better communication and cooperation.

Third, the research highlights the value of this positive approach - outlining the benefits of new behaviours. For men, in particular, this tapped into a number of incentives to try out new behaviours – the positive feelings they may experience in relieving their wives’ domestic burden, the prospects of better sexual relations and improved household income. This aligns with other literature which suggests that men need the space to consider that gender inequalities can be harmful to men and that – accordingly - gender equity can be beneficial to men (Clowes 2013).

However, our findings also encourage reflection on this ‘benefits-based’ approach. They indeed suggest that changes in men’s behaviours are likely to have been primarily motivated by these perceived positive benefits rather than a fundamental shift in their beliefs or corresponding norms around gender roles and male authority. Yet decades of feminist research suggest that to achieve long-term transformational change in gender inequalities, discrimination and violence requires dismantling these patriarchal norms.

The question for the Indashyikirwa programme therefore is what happens next? Are the shifts in behaviour and - to a lesser extent - beliefs that occurred steps on the path to wider transformational change? Will couples who participated in the curriculum experiment further with new roles and forms of cooperation which will eventually lead them to challenge beliefs and norms around male/female roles and male authority? Or will behaviours revert as entrenched beliefs and norms have been insufficiently challenged and not transformed?

Inspired by SASA!, a subset of individuals who completed the Indashyikirwa curriculum received further training and mentoring to become ‘community activists’, with the goal to diffuse messages about the positive uses of power and the benefits of non-violent relationships more widely via modelling, testimony and advocacy. Indashyikirwa also worked with opinion leaders and established ‘safe spaces’ for women to create an ‘enabling environment’ for change. A separate article will examine the impacts of this community activism component and help us to explore these questions further.

Limitations

It is important to acknowledge the limitations of the study. First, the data were collected to assess the impacts of the Indashyikirwa curriculum on multiple outcomes and processes of change, not only on beliefs, behaviours and norms related to the gendered division of labour and decision-making. Second, as for all studies of social behaviour – qualitative and quantitative - we cannot eliminate the possibility of social desirability bias where participants decide to report favourably on an intervention they valued. We attempted to mitigate this by engaging field researchers identified as ‘external’ to the programme and emphasising confidentiality. We also undertook an intensive process of cross-referencing data between different interviews with the same individuals and between couples. Third, we did not use full dyadic analysis for this paper, as we found the individual narratives more pertinent to answer the research questions for this article. Nonetheless, this could be useful, especially as part of mixed methods analysis triangulating with the quantitative survey data for the same couples.

Conclusion

The ‘shifts’ documented in this paper – largely in behaviours, but also to some extent in beliefs - are important and have contributed to reductions in gender inequalities and IPV among couples targeted by the Indashyikirwa programme. However, they fall short of fully transforming entrenched beliefs and norms around gender roles and male authority over resources. It is possible that these shifts in behaviours and couple dynamics – as well as the introduction of new ideas about working together for better relationships and household development – will be steps on the road to transformational change. However, this paper suggests that the persistence of these patriarchal beliefs and norms may continue to constrain the extent of change among couples and may potentially act as a significant obstacle to longer-term, larger-scale ‘transformations’ in gender inequalities and violence.

We therefore need further research to examine the patterns of change that programmes like Indashyikirwa catalyse and which conditions help new ideas and behaviours to consolidate and spread. Rwanda has the advantage of existing laws, policies and systems to promote gender equality and address violence against women. This bodes well should the government and donors agree to expand implementation of Indashyikirwa. Such wider structural supports for normative change and women’s movements are likely to be key for consolidating shifts at an individual and community-level.

Acknowledgements

This paper is an output from the What Works to Prevent Violence Against Women and Girls Programme funded by UK Aid from the UK Department for International Development (DFID). The funding was managed by the South African Medical Research Council. The views expressed and information contained within it are not necessarily those of, or endorsed, by DFID, which accepts no responsibility for such views and/or information, or for any reliance placed on them.

Disclosure statement

All authors have declared that no conflicts of interest exist.

Notes

1 We use the term ‘gender inequity’ to refer to unfair, avoidable injustice and ‘gender inequality’ to refer to an imbalance in the distribution of resources on the basis of gender. ‘Gender-inequitable’ beliefs, behaviours and norms justify, promote or result in the unfair treatment of men and women based on their needs.

2 Ridgeway defines a ‘cultural frame’ as a ’shared cultural system for categorising and defining ‘who’ you and the other in the situation ‘are’, and based on that categorisation, how you are expected to behave’ (ibid p315).

3 RNECREF: 340/RNEC/2015); NISR REF: 0738/2015/10/NISR). Secondary approval was also obtained from the South Africa Medical Research Council (REF: EC033-10/2015) and the London School of Hygiene and Tropical Medicine (REF: 10339).

4 The retention of 14 or 15 couples for the duration of the research reflects the 98% completion rate of the curriculum. This can be explained by a combination of factors. First, Rwandan citizens are expected to participate regularly in community meetings and do so. Second, the qualitative and monitoring data attest to how much participants valued the programme. Third, following curriculum completion, RWAMREC staff followed up with couples twice a year and were regularly in the communities.

5 All the transcripts were coded by the third author and then analysed for this article by the first author.

6 The ‘G is for Gender’ sessions include a ‘gender lifelines’ exercise where participants discuss the difference between gender and sex and gender socialisation. They then discuss why men and women do or don’t perform certain activities and whether they could.

7 During ‘The 24–hour day’ exercise, participants work in small groups to imagine a typical day in the lives of a husband and a wife in their community. They list the typical tasks performed by each and whether these tasks are paid or unpaid. This is used to prompt a discussion about the quantity of tasks done by men and women, which are considered as ‘work’, who has leisure time and the impacts of this workload.