Re-conceptualising gender and power relations for sexual and reproductive health: contrasting narratives of tradition, unity, and rights

Sexual and reproductive health interventions in sub-Saharan Africa will be most effective if grounded in emic (insider) perspectives of gender and power in intimate relationships. We conducted eight focus group discussions with 62 young adults in Malawi to explore conceptions of gender and power relations and areas of tension between different perspectives. We framed our enquiry according to the three social structures of the Theory of Gender and Power: the sexual division of labour, the sexual division of power, and social norms and affective attachments around femininity and masculinity. Young adults drew on interrelated and competing narratives to describe the state of gender relations, which we named tradition, unity, and rights. Participants used tradition narratives most frequently to describe patriarchal gender roles, norms and ideals. Some participants challenged this predominant discourse using unity and rights narratives. Unity narratives illustrated how love and couple reciprocity were essential sources of ‘power with’ as opposed to ‘power over’. Rights narratives were more contested than other narratives, with some participants acknowledging that women’s rights were important to the family’s survival and others viewing women’s rights as problematic for gender relations. Gender-responsive interventions should consider the tensions and intersections between multiple narratives on gender and power, including unity as a gender-equitable form of power. ARTICLE HISTORY Received 14 January 2019 Accepted 8 September 2019


Introduction
Public health research documenting the impact of gender relations on sexual and reproductive health has largely framed gender relations as the product of hegemonic masculinities rooted in patriarchy, male dominance and control over 'vulnerable' women, leading to women's poor sexual and reproductive health outcomes including CONTACT Amy A. Conroy amy.conroy@ucsf.edu This article is part of a special issue on Transformative Approaches to Gender Justice in Sexual and Reproductive Health, led and sponsored by the University of California Global Health Institute Center of Expertise on Women's Health, Gender and Empowerment.
increased risk for HIV (Morrell et al. 2013;Dworkin 2005;Jewkes, Flood, and Lang 2015). For example, men's perpetration of intimate partner violence (IPV) against women is often understood as stemming from rigid or traditional gender ideals that perpetuate controlling behaviours over women Dunkle et al. 2006;Jewkes, Flood, and Lang 2015). While such research has advanced our understanding of women's risk for disease and poor health outcomes, the dominance of this discourse has arguably limited a more nuanced and comprehensive understanding of how gender and power dynamics play out in diverse contexts. A paradigm that overemphasises women's vulnerability (referred to as the 'vulnerability paradigm') has been criticised for overlooking the changing nature of gender relations, how variability in social and cultural contexts affects health outcomes, and the role of both sexes in maintaining the gender hierarchy (Higgins, Hoffman, and Dworkin 2010;Dworkin 2005). Social scientists working in sub-Saharan Africa have offered a critical and theoretical lens to understand the complexities of gender relations and to challenge the prevailing discourses on female vulnerability (Wyrod 2008;Hunter 2010). Some researchers have argued that rigid notions of masculinity and femininity can result in missed opportunities to understand practical and nuanced meanings of gender and power relations that change over time (Dworkin et al. 2012;Connell and Messerschmidt 2005).
Women's empowerment is linked to many sexual and reproductive health outcomes in African settings including increased use of contraception and reproductive health services (Blanc 2001). Yet, other research has documented unintended consequences and resistance to efforts to improve women's status through a human rights or female empowerment framework. For example, men in South Africa have perceived that gains in power for women result in equivalent losses for men (Dworkin et al. 2012;Shefer et al. 2008). There is also emerging literature documenting the negative effects of women's empowerment programmes, specifically microfinance, on women's risk for violence or harassment (Slegh et al. 2013;Dunbar et al. 2010;MacPherson et al. 2015), which may be attributed to backlash from male partners. In Uganda, Wyrod (2008) found that attempts to reconfigure gender power relations, such as changing marriage laws or the division of household labour, were viewed as excessive. Also in Uganda, Boyd (2013) found that anxieties about 'rights' stemmed from the perspective that unconstrained freedoms would threaten the normative hierarchy of society, and allow immoral values from the West to infiltrate families. Thus, it would be prudent for gender-responsive health interventions to fully consider understandings of gender relations that challenge the vulnerability paradigm and potential negative responses to women's gains in power.
Moreover, a focus on hegemonic gender roles often treats men's and women's roles as mutually exclusive and pays insufficient attention to the couple relationship. For example, research that seeks to understand archetypes of gender roles and relations may fail to consider shared sources of power within couples, and how beliefs and expressions of gender may allow couples to cope with issues and work towards a set of common goals that are mutually beneficial for the couple. The majority of health research on power in sexual relationships has theorised power as 'power over' (e.g., male control and dominance over women) rather than 'power with' (e.g., shared power within a couple) (Blanc 2001). Yet 'power with' may also be a potent force. Research among South African couples enrolled in an HIV prevention trial found that shared power was a stronger correlate of relationship quality (e.g., trust, intimacy) than women's own power or men's beliefs about equitable gender norms (Conroy et al. 2016). Relationship quality is also associated with positive health behaviours such as condom use and medication adherence (Ruark, Kajubi et al. 2017;Conroy, Ruark, and Darbes 2018). Connell's (1987) theory of gender and power provides the theoretical basis for the vulnerability paradigm and argues that three social structures characterise the relationship between men and women: (1) the sexual division of labour; (2) the sexual division of power; and (3) cathexis, or social norms and affective attachments around femininity and masculinity. We use the term 'sexual' versus 'gendered' to be consistent with the theory's terminology. Wingood and DiClemente (2000) extended this theory to identify important risk factors and exposures that increase women's risk for HIV, and empirical studies have found support for the application of Theory of Gender and Power to explain sexual risk behaviours and sexual and reproductive health outcomes among women (Woolf and Maisto 2008;Blanc 2001;Pulerwitz et al. 2002).
Social norms and affective attachments (cathexis) dictate how men and women should express their sexuality and enforces strict gender roles for men and women. A risk factor resulting from this structure would be men having multiple sexual partners as a result of hegemonic norms of masculinity and the need to assert status through sexual performance, and ideals of femininity that embrace this form of masculinity, thus placing women at risk for HIV (Jewkes and Morrell 2010). The sexual division of labour assigns women to unequal positions (e.g., lower paid positions, unpaid childcare or domestic work) and limits women's economic potential, creating economic dependencies on men. One resulting risk factor would be poverty, which can drive women's decisions to engage in unprotected sex (Krishnan et al. 2008;Farmer, Simmons, and Conners 1996). Finally, the sexual division of power is maintained by social mechanisms such as the abuse of authority and control in relationships and can result in IPV. Women who fear or have experienced abuse are less likely to negotiate condom use if they are concerned about violence (Wingood and DiClemente 1997;Maman et al. 2000).
In pointing out the limitations of using the theory of gender and power understand health, Wingood and DiClemente (2000) suggest that the three social structures are difficult to operationalise and do not take into account local variations across cultures. Expanding on this latter point, we argue that because of its focus on traditional gender roles, the theory of gender and power may fail to capture the variation and complexity of gender relations across cultures, time, and impact on health. The overall objective of this study was to qualitatively describe gender and power relations from the emic (or insider) perspective of young adults in Malawi who had recently married or were contemplating marriage. While recognising the limitations of the theory of gender and power, we used the three social structures of the theory to guide our understanding of gender and power relations in Malawi while highlighting areas of the theory that could be expanded and contexualised.

Materials and methods
The study setting Malawi is a small, resource-poor country located in south-eastern Africa. Marriage in Malawi is quasi-universal and typically occurs around the age of 18 for women and 23 for men (MDHS 2016). The majority of Malawians are engaged in smallholder agriculture, which provide many rural households with a major share of their food supply (Kerr 2005). However, yearly famines are common and have resulted in high levels of malnutrition and food insecurity (Kerr 2005). The country is divided into northern, central, and southern regions. As opposed to the patrilineal/local north, southern Malawians generally follow a matrilineal/local tradition characterised by men living with their wives' families after marriage (Peters 1997). Rates of divorce are also higher in the southern region (Reniers 2003). Yet in spite of this tradition, most households and villages are headed by men (Peters 1997).
According to the United Nations, Malawi ranks 171 out of 189 globally on the gender inequality index (UNDP 2018). As in most African countries, gender relations in Malawi are continually shifting and being re-configured by multiple social forces, notably national laws on domestic violence passed in 2006, recent policies to promote gender equality in the workforce, and the rise of gender-responsive programming in the development and health sectors (Watkins and Swidler 2013). The threat of HIV has also profoundly impacted gender relations and sexualities in Malawi and across the region (Sennott and Angotti 2016;Hunter 2010;Smith 2014).

Study procedures and sampling
This study was embedded within a larger panel study of reproduction and HIV among young adults in Malawi (Tsogolo La Thanzi [TLT]; https://tsogololathanzi.uchicago.edu). In 2011, we conducted focus group discussions (FGDs) with a sample of married and unmarried young men and women in the Balaka district of southern Malawi. To avoid overburdening study respondents and to obtain fresh perspectives, we recruited a new sample of participants who were not part of the TLT cohort. FGDs are an appropriate method for gaining an understanding of social norms and ideals by observing how people interact and debate with each other. Ethical approval was obtained from the Colorado Multiple Institutional Review Board and the National Health Sciences Research Committee in Malawi, and informed consent was obtained from all participants.
After obtaining permission to conduct the study from local village authorities, we enrolled a purposive sample of participants from three villages surrounding Balaka town. Study recruiters approached eligible participants at their homes and explained the details of the study. If participants agreed to participate, they were given appointment cards indicating the time and location of the FGD. We purposively selected both married and unmarried individuals to capture differences in opinions between those who had personal experience of marriage, and those who did not and who might draw more on ideals around marriage. We allowed participants to self-identify as married or unmarried in recognition that marriages can be both customary and legally recognised in Malawi.
We hired and trained two research assistants (RA), one male and one female, to facilitate the FGDs. Both were fluent in English and Chichewa and had a secondary school education and prior qualitative research experience. The male research assistant facilitated men's FGDs, while the female RA facilitated women's FGDs. Prior to the start of the research, the RAs participated in a weeklong training with the lead investigator to become familiar with the study procedures, data collection instruments, and ethical procedures.
We recruited equal numbers of married and unmarried participants, and men and women. Participants were allocated to separate groups by their gender and marital status to facilitate open communication on sensitive topics within relationships. Participants were between 16 and 24 years old, and participants under the age of 18 were married (married individuals are considered maturated in Malawi). We conducted eight FGDs with 7 or 8 participants each, for a total of 62 participants. FGDs were conducted at a teacher development centre in Balaka town, an immunisation clinic in a trading centre village, and under a shady tree in a rural farming village. Privacy precautions were taken to ensure that others could not overhear the conversations. Women and men had average ages of 20 and 21, respectively, and were evenly divided between having a primary school and secondary school education.
We used a semi-structured FGD guide with questions on gender and power dynamics, reasons for testing for HIV, and vignettes on HIV testing for participants to discuss. Findings on HIV testing have been published elsewhere (Conroy 2014b). Questions on gender and power included: Can you describe a perfect relationship? What does it mean to be a 'household head'? What makes men and women feel powerful in their relationships? How much power can women and men have in their relationships?
FGDs lasted between 75 and 120 minutes and were audio-recorded. Each facilitator transcribed their FGDs and translated them verbatim from Chichewa into English. During transcription, participants were assigned a number based on the first time they spoke in the focus group and then assigned a pseudonym to protect confidentiality and anonymity.

Data analysis
The lead investigator (AAC), a US-based researcher with training in qualitative methods, coded the data in Microsoft Word, following the approach outlined by Strauss and Corbin (1998). The first step was to review the transcripts, become familiar with the data and then write memos describing emerging themes. The second step was to conduct open coding in which the transcripts were examined line-by-line and systematically assigned a priori codes based on the FGD guide topics as well as new codes that emerged inductively. As coding progressed, existing codes were removed or modified, and new codes were added. After open coding, axial codes were applied to specify the relationship between codes and to group codes into categories. In the final step, selective codes were used to integrate and refine categories. While we used the Theory of Gender and Power as a starting point to inform our line of questioning around labour, power, and norms, we also allowed new themes to emerge which we used to modify and expand the original theory.
We used several criteria from Lincoln and Guba (1986) to enhance scientific rigor of the analysis. To strengthen credibility of the findings, the research team reviewed each question in the FGD guide and Malawian researchers shared their perspectives as cultural insiders as part of preparing for the study. We also conducted two pilot FGDs that were used to further train the RAs on how to remain neutral while facilitating the FGD, probe effectively, and be reflexive about their roles as researchers. Once data collection began, the research team held debriefing sessions after each FGD to discuss emerging themes and possible interpretations. Throughout data analysis, codes and themes were discussed with an external advisory committee possessing qualitative expertise and knowledge of the Malawi context, to confirm the approach and findings. Finally, the lead investigator spent extended time in the field, which allowed for participant observation, detailed field notes, informal interviews with key informants, and building rapport with the research team. This prolonged engagement facilitated the investigator's reflexivity and checking of internal biases, increasing the credibility of the research.
Participants voiced their perceptions of gender and relationship power using three interrelated narratives, which we named tradition, unity, and rights. We assigned the thematic code of tradition to statements that referenced traditional and patriarchal gender roles, norms, and ideals (including those reflected in religious teachings), such as statements like 'husbands are heads of households' and 'wives must obey their husbands.' When participants discussed couple-level dynamics such as shared power, communication, respect, and collaboration, we assigned the unity code. Finally, we assigned the rights code when participants expressed support for individual freedoms and privileges, especially women's rights. These included discussions of women's right to an education or jobs typically reserved for men (e.g., police officer).

Cathexis: gender and relationship ideals
According to the Theory of Gender and Power, cathexis refers to the social norms and affective attachments that dictate how men and women should express their sexuality and ascribe to traditional gender roles (Wingood and DiClemente 2000). When asked to characterise an ideal spouse, participants drew upon multiple narratives. A set of shared ideals was mentioned by all groups (see Table 1). In reference to tradition, all groups described an ideal wife as one who was virtuous. A unity discourse was also used to characterise an ideal wife as respectful, loving, and trustworthy (i.e., both dependable and sexually faithful). Similarly, tradition and unity narratives were used to describe an ideal husband as a good provider, trustworthy and loving. As shown in Table 1, there were differences by gender and marital status in other conceptions of an ideal wife or husband. Men mentioned desires for wives who were 'God-fearing,' physically attractive, and hardworking. Married women mentioned the desire for a husband who was understanding, loved her relatives, and was obedient (giving the wife what she wants). Unmarried women mentioned the importance of a man's fertility, HIV-negative status, and abstinence from smoking and alcohol use, which were perceived as important precursors to marriage.
When asked to describe a perfect relationship, a majority of participants invoked a unity discourse (see Table 1). Shared ideals across all groups included love, cooperation, mutual respect, and getting tested for HIV together. Married men and women both mentioned the importance of 'understanding each other'. In about half of focus groups, love was noted to be the most important relationship ideal. The meaning of love differed based on gender and marital status. Women associated love with materiality. For married women in particular, love was strongly tied to tradition and the sexual division of labour, that is, whether their husbands lived up to the provider role. In one example, a married woman indicated, 'it shows he has love when he leaves food in the house.' Other participants described love as a means to control a partner, as enacted through the sexual division of power. For example, in accordance with the tradition narrative, women argued that loving a partner meant conforming to what he wanted such as providing sex or dressing a certain way (e.g. not wearing 'miniskirts'). Similarly, men described love as giving women what they asked for. Others talked about the ideal of love as 'working together' in a collaborative act of unity. A few women conceptualised love as sexual faithfulness and dedication to the relationship. Men tended to emphasise unity narratives when describing love and used similar language to women in describing how love translates to working together on household matters and remaining faithful.

Sexual division of labour
According to the Wingood and DiClemente (2000), women are assigned to unequal positions compared to men, which creates economic imbalances that force women to rely on men financially. We asked participants to discuss the meaning of the phrase 'head of the household' and to explain who can take on this role and when.
Consistent with the theory of gender and power and gender roles and ideals in rural Malawi, the tradition discourse was the most widely used narrative to explain the role of head of the household (see Table 2). Women described the household head as the leader (the one who solves problems, makes decisions, and plans for the family's future), the provider for the family (the one who works and takes care of household needs), and the owner of the house who sets the rules. Women generally agreed that Participants provided varied justifications for why the husband is the household head. In general, many believed that the man was always considered the head of the household as dictated by customs and long-standing ideals around respect for men. Some cited Christian teaching from the Bible which they saw as legitimising men's authority. Others defined the head of the household as the one who proposed marriage or the one who was older (typically the man), given cultural rules that dictate respect for elders.
When participants were presented with the scenario of a wife who worked or was the sole 'breadwinner', some indicated that the wife's power was limited. One unmarried woman explained, 'a wife can rule a husband, but not like the way that women are ruled by their husbands' (FGD #5). Men, on the other hand, engaged in lively debate with each other about whether women could be the sole leader of the household: George: She looks after children, buys food in the house, and does everything else that needs to be done? In such a case, the woman is the head of the household.

Moses:
A train does not turn as a car does. When it wants to return, it just moves backwards. But though it moves backwards, we are still able to tell which side is the front, and which side is the back. This is the same in the family. In a relationship where the man does not work but the woman does, more people will say the head of the household is the man. (group laughs) Blessing: What you are saying is not true because in this case, the woman is doing everything. (FGD #4) Rights narratives, on the other hand, described circumstances in which women led the household alone and gained power as a result of education and employment. Both men and women noted how women had the right to be household heads under special circumstances, such as when a husband was 'docile' or submissive, and did not lead the household properly. Some respondents believed that wives could take The decision-maker The leader/ruler of the family The provider, breadwinner, or one who takes care of household needs advantage of husbands and usurp their power, which was frowned upon. One married woman highlighted these examples: Sometimes women declare themselves as head of the household because they feel that their husbands are too quiet (docile) and they cannot take up the role of the household head. Or sometimes it is because their husbands treat them very well that they take advantage and take up the role of the head of the household. (FGD #1) There were mixed views about the benefits of women's increasing rights and effects on the family. For instance, one man believed that increasing women's rights (e.g., education, employment) could lead to increases in couple collaboration (unity): I, for one, disagree because nowadays the world is changing. Women are getting educated and finding good jobs. Let's say I am working and the woman is also working. We will be doing things together. There is no relish [vegetables], maybe the man has no money and the woman will give the money. So, we can't say the household head is the man. We should accept that the household is supposed to be run by two people. That's a household. (FGD #3) Other men (and women) had a negative view about women's increasing power, believing that women's rights took away power from men. One unmarried man described how women could secretly take away the man's power: There are times where you as a man, you know about freedoms. Maybe you went to school and you know that personal freedoms are supposed to [be] like this and that when in marriage. But the woman can have greater freedom such that she can direct the household. And you do not even know that she is directing you. (FGD #3) Similarly, a group of married women discussed that women were taking away jobs from men and some were even directing their husbands to do women's work such as washing nappies [diapers], which they considered to be 'abusive' of men. (FGD #7) In contrast to the tradition narratives that emphasised male authority and legitimacy, unity narratives encompassed ideas of oneness, collaboration, and mutual respect. Although unity was less frequently invoked as compared to other narratives, both men and women used a unity narrative to argue that if both spouses contributed to the household, they should jointly head the household. For example, a group of married women expressed how men and women could lead the household together since both partners were required to make important family decisions.
They need to sit and make decisions together as one. Looking at the fact that these two are working together, [neither] the man nor the woman can be called head of the household without the [other] person … They both need each other's help to be called the head of the household. (FGD #1) Similarly, an unmarried man argued that both spouses could lead the household together, citing the Bible for justification: We are saying that the household head is the man, but I disagree. The household head is both of them because when they say marriage it is both of them. Marriage can't be one person only. Everything concerns both of them, like sex. Two people have sex. In the Bible [it says] women should respect their husbands but also husbands should respect their wives. (FGD #8)

Sexual division of power
According to the theory of gender and power, the sexual division of power reproduces inequities in power between men and women which are maintained by social mechanisms such as the abuse of authority and control in relationships (i.e., 'power over'). Yet the most common narrative used by participants to describe power was a unity discourse (i.e., 'power with'), which included references to love, sex, respect, helping each other, and open communication (see Table 3). Both men and women acknowledged that love made them feel powerful. For men and women, power was also tied to constructive communication in couples. Women (both married and unmarried) said they felt powerful when their partners listened to their advice on important family decisions.
Sex and sexual satisfaction were noted as important sources of power for men and women, but they differed in how power was described. Some men explained how 'being on top' during sexual intercourse made them feel powerful, as a form of male authority or 'power over'. For women, sex as power was described using a unity or 'power with' discourse: Grace: All that matters in a relationship is sex and that is what makes a woman feel powerful. Ellen: When the two satisfy each other's needs. For example, if the woman has strong sexual desires and the man doesn't have them, can we say that anything will work out there? All: No. Ellen: For sure, it cannot work because women are like fishing hooks whereby men get stuck on them and women always want their hook to have something hooked to it. So, if both parties have strong sexual desires and have sex, that's something that makes us feel powerful. (FGD #1) Descriptions of power conflicted somewhat with ideals around the appropriate balance of power in relationships. When asked to discuss how much power men and women should have in their relationships, both men and women used a tradition narrative to conclude that men should have more power. Few participants, if any, concluded that women should have more power than men. Women worried that if a woman had more power than the husband this could 'ruin the marriage' or lead to divorce. Several men said that power in relationships should be more balanced-yet not equal-and used unity and rights narratives to describe the ideal distribution of power. For instance, it was noted that a man should not abuse his power by beating up his wife or treat her like a 'slave', which was believed to violate her rights. Women also ascribed to this belief: A husband may say to his wife, "You should go to the field and I am going to work," leaving the whole field for the woman to work on herself until harvesting without the woman knowing where her husband goes. And, when he comes home, he calls for his wife to have sex, knowing that she has worked in the farm without his help. Some men do things that are not good. (FGD #5) Similarly, some participants espoused the belief that a man shouldn't force his wife to have many children. Others noted that a woman's rights could also be violated if her opinions were not considered in important family matters. Men noted how the couple should practice unity by collaborating on family issues and decision-making. A group of married men explained that when a man used his power wisely, he was respectful of his wife and her different opinions. Some men argued that a lack of unity could limit opportunities for a better life, which was possible only when women's perspectives were considered.

Discussion
Through the use of three narratives, Malawian young adults actively debated ideal gender roles and how much power wives and husbands should have in their relationships. Given the patriarchal gender hierarchy in Malawi, it is not surprising that the majority of young people drew on tradition narratives to make sense of gender and power relations. Yet there was also relatively strong support for a unity narrative highlighting the importance of working together and mutual respect. Overall, participants expressed more acceptance of unity and tradition narratives, and more ambivalence with regards to women's rights to education and leadership roles in the household. In contrast to previous research in the field of sexual and reproductive health, which has described a more absolute gender hierarchy, these qualitative data paint a more nuanced picture of gender relations characterised by 'power with' as well as 'power over' (Blanc 2001). One explanation may be that young people are challenging the status quo around 'power over' given the challenges facing this generation, particularly how to survive the HIV epidemic while safely meeting reproductive and life goals.
The three narratives of tradition, unity, and rights may offer a higher-level framework for categorising power, gender norms, and the sexual division of labour. The theory of gender and power largely assumes that relationships operate under a tradition paradigm characterised by the dichotomy of male dominance and female submissiveness (Connell 1987). While the theory of gender and power is valid and useful across many settings, it may be insufficient to capture cultural change and the multiplicity of social meanings and should be expanded to include all three narratives. Rights narratives emphasised female autonomy and freedom, while unity narratives emphasised collaborative relationship dynamics. Participants' different viewpoints varied according to their perceptions about the impact of tradition, rights, and unity on preserving the family and meeting daily subsistence needs. For instance, women blamed rights for taking away jobs from men or destroying marriages, accepted tradition as the norm unless the husband was failing to provide, and believed that unity could help couples make better decisions.
Unity narratives suggested that power could be reconstructed as a feature of the union in which both women and men gain power together, which diverges from the more typical conceptualisations of power around decision-making dominance and control (Pulerwitz, Gortmaker, and DeJong 2000;Blanc 2001). Previous work has found that unity was a key construct of power among young couples and a predictor of IPV and HIV-related health behaviours (Conroy 2014a;Conroy 2015;Conroy and Wong 2015). Similarly, previous research from Malawi found that shared decision-making was associated with positive relationship dynamics (e.g., intimacy, positive communication), whereas sole decision-making by women or men was associated with negative health behaviours and more IPV (Ruark, Chase et al. 2017).
The theme of love, falling under the unity narrative, was described as an important relationship ideal and form of power, and closely tied to the sexual division of labour. The existing public health literature on relationship power focuses on functional forms of power such as decision-making dominance over household purchases (Blanc 2001;Sriram et al. 2018). In this study, women whose material desires were met by a loving partner could indirectly control how family finances were spent through enticing men with love. This nuance could be missed in quantitative studies measuring men's 'final say' on family decisions, which could lead to incorrect portrayals of power dynamics. However, the findings are consistent with the sociological literature highlighting the inseparable links between money, love, and sex, and the exchange of resources from men and women (Hunter 2010;Mojola 2014). Poor men who cannot provide the things women desire may be unable to achieve life aspirations such as marriage and childbearing, which bring social status (Ruark et al. 2016).
Participants, at times, used these narratives in contradictory ways. While participants were somewhat accepting of the unity narrative to describe power and relationship ideals, most participants agreed that men should have more power than women in practice. Other research on household decision-making in Malawi has found that men and women use a combination of gender-specific cultural scripts (e.g., 'husbands are the head of the household') and non-gender specific scripts that emphasised harmony and open communication (Mbweza, Norr, and McElmurry 2008). This inconsistency reflects the struggle to make sense of changing gender relations including which rights are important and for whom, which has been highlighted by scholars of other SSA settings (e.g. Wyrod 2008;Boyd 2013).
Some sociologists writing on the 'patriarchal bargain' in sub-Saharan Africa have noted that abrupt changes in the existing gender order have at times sparked resistance among women who viewed these changes as a threat to their current accommodations within the male-dominated order (Kandiyoti 2005). In Uganda and South Africa, social scientists have similarly found that some men and women resisted the idea of rights, which they felt would degrade morality and family relationships and bring an infusion of Western culture (Boyd 2013;Wyrod 2008;Dworkin et al. 2012). As Boyd (2013: 698) reminds us, the rejection of liberal human rights underscores the way that 'human rights struggles have become moral debates, predicated not only on what rights mean and do, but how categories of ethical personhood … are experienced differently in different places.' In southern Malawi and similar settings, women's empowerment interventions should acknowledge and consider the potential for women and men to resist change in gender relations; failure to consider this possibility may render interventions ineffective. In some instances, men in our study were more willing to endorse non-traditional conceptions of gender relations than women, bolstering the notion that women are also complicit in maintaining patriarchal gender hierarchies. Critical questions remain regarding how attempts to transform gender relations will affect women's lived realities, which transformations are most meaningful in a given context, and how programmes aimed at transformation should be implemented. While we agree that human rights should be at the forefront of development and public health interventions, programmes in certain socio-cultural contexts might aim to 'meet couples where they are at' or take a more gradual approach guided by community priorities, rather than attempting to transform gender according to models used in liberal, western democracies. Until a rights framework is more socially accepted at the community level, efforts to change gender and power relations might consider focusing on aspects of unity while recognising that traditional gender roles (particularly of the man as main 'breadwinner') serve a functional purpose for families' economic survival. This idea is echoed in women's concerns for how women's rights can destroy marriages. In accordance with intersectionality theory (Hankivsky 2012), we also note that gender and health may intersect in very different ways in other African populations, according to differences in age, culture, economic realities, and other factors. Thus, a critical first step towards the implementation of any gender-responsive intervention would be a rigorous formative component to contextualise acceptance for tradition, unity and rights among the targeted recipients of such programmes.
We note several limitations of this study. In invoking rights narratives, it is possible that participants were reciting language from gender empowerment programmes to which they had been exposed, which typically articulate a human rights framework. Although it is possible that beliefs about rights were socially manufactured, we believe that many participants were expressing genuine beliefs given the many divergent opinions provided and willingness of participants to debate each other. It is also plausible that tradition narratives around social norms such as a man's legitimate and cultural right to power masks what actually happens in the household, as found elsewhere (Tolhurst et al. 2008;Tolhurst and Nyonator 2006). While FGDs could potentially reinforce these ideals rather than emphasising real-life experiences, thus introducing social desirability bias, the debate inherent in FGDs provides a counterweight to this bias. Our findings may also not be transferable to older married individuals or those living in patrilineal settings in SSA. Other studies in South Africa have highlighted a generational divide in the use of rights language (Sennott and Angotti 2016), suggesting that there could be more resistance to change in older cohorts. Likewise, in patrilineal societies, the use of tradition narratives may be even more pronounced if women have less control over land ownership and inheritance.
In conclusion, by listening to the voices of young adults in Malawi, we can gain a more nuanced understanding of the complexity and variation in gender norms, power, and the sexual division of labour among young Malawian couples. Contested narratives around tradition, rights, and unity highlight the fluidity of how men and women relate to each other and their struggles to make sense of gender relations. Our findings on reactions to women's rights and empowerment are consistent with other sociological research in Uganda and South Africa (Dworkin et al. 2012;Wyrod 2008), but bring new attention to the frame of unity as a form of shared power at the couple level. Gender programming could more strongly promote the construct of unity in messaging efforts, and not rely solely on rights as an intervention strategy. Sexual and reproductive health interventions that are gender-responsive and grounded in emic understandings of gender and power relations are more likely to be acceptable, feasible, and effective.

Disclosure statement
There are no conflicts of interest to report for any of the study authors.

Funding
This work was supported by the US National Institutes of Health under grants K01MH107331, T32DA13911, and K01MH106416.