Widening cracks in patriarchy: mothers and daughters navigating gender norms in a Mumbai slum

Abstract Inequitable gender norms can be harmful to girls’ and boys’ health and sexuality. Programmatic approaches that help renegotiate gendered power relationships are sorely needed. This qualitative study reveals how Parivartan, a sport-based intervention in a Mumbai informal settlement, helped families resist inequitable gender norms that limited girls’ mobility in public spaces. Fifteen girl athletes were interviewed in two rounds of face-to-face in-depth interviews. Results identify the strategies girls’ mothers used to support their daughters’ participation in the programme when they feared their husbands’ disapproval. Rather than openly confronting their husbands, mothers worked from within the patriarchal gender order, through its ‘cracks’, for instance initially hiding their daughters’ participation from their husbands. At an appropriate moment, girls’ mothers revealed to their husbands about their daughters playing sports, convincing them of the usefulness of the programme. Girls’ participation profoundly and positively affected relationships between daughters, mothers and fathers. Over time, parents’ trust that girls would not compromise family honour increased, eventually changing the acceptability of girls’ playing sport in public in spite of the patriarchal gender order. Concluding remarks offer key implications for effective interventions, highlighting the historical nature of gender transformation processes.


Introduction
Gender equality is both a valuable end in itself and a means to achieve other global health and wellbeing goals (Magara 2015). Gender norms can be both facilitators (when equitable) and obstacles (when inequitable) to achieving gender justice (Connell 2014;Connell and Pearse 2015). They are unwritten rules of expected action for women and men in a given group or society, that affect their (often unequal) access to resources and freedoms, as well as their voice, agency and power (Cislaghi, Manji and Heise 2018). Inequitable gender norms can be harmful for the health and wellbeing of girls and boys, women and men (John et al. 2017;Lundgren et al. 2018). For instance, expectations that boys and men should be tough and resilient can affect their capacity to seek help and care (Addis and Mahalik 2003), and expectations that girls should not leave the household can affect the aspirations and wellbeing of those girls and women who want to seek employment or further their formal education (Balk 1997;Rao 2012).
In many settings, gender norms distribute power unevenly between men and women (Connell 2014;Schuler et al. 2018). In patriarchal settings, they contribute to the dominance of hegemonic forms of masculinity and to the subordination of femininities and marginalised masculinities (Connell and Messerschmidt 2005). However, the distribution of power in a patriarchal gender system is not dichotomous. Even in extremely patriarchal contexts, women can devise strategies to navigate the context in ways that help them both optimise their life options and generate resistance (Argawal 1997). This resistance, eventually, can contribute to the breakdown and transformation of patriarchal relations (Kandiyoti 1988). Transformation of the patriarchal order can be brought about by collaboration with men too (Jewkes et al. 2015). Not all men hold patriarchal worldviews; some are policed into patriarchy, so that they embody inequitable gender norms with their actions, even when they feel uneasy with those norms (Connell and Messerschmidt 2005). Similarly, not all women hold anti-patriarchal attitudes: some may be protecting a patriarchal order because compliance with it guarantees them social status and respect (Kandiyoti 1988;Catherine and Mannell 2016).
Inequitable gender norms can be particularly harmful to girls and young women, sustaining practices that threaten their health and wellbeing. Child marriageformal or informal marital unions in which at least one of the spouses is younger than 18 (Nour 2009)is one such practice. It disproportionately affects girls and is associated with several harmful health and development outcomes, including, for instance, school drop-out, depression, increased risk of domestic violence and early pregnancy (Godha, Hotchkiss and Gage 2013;Gage 2013;Chandra-Mouli, Camacho and Michaud 2013;Neal et al. 2012;Santhya 2011;Raj et al. 2010;Raj 2010;Nour 2009;Raj and Boehmer 2013;Nour 2006). Interventions that challenge gender norms that sustain child marriage can be particularly effective in reducing the practice (Cislaghi and Heise 2018a;Chattopadhyay and Srikanta Murthy 2018;Prakash et al. 2018). Gender-transformative interventions seek to reshape relations so that they become more gender-equitable, benefiting both men and women by freeing them from the oppression of destructive gender norms (Dworkin, Fleming and Colvin 2015).
One such intervention is Parivartan, a sport-based programme with girls aged 12-16 (Collumbien et al. 2019). Implemented in an informal settlement in Mumbai, South India, Parivartan facilitated family capacity to resist the system of gender norms that helped sustain child marriage in this setting, also helping family members reframe their gender relations and dynamics. Elsewhere, we have reported practical learnings from the programme (Collumbien et al. 2019) and how the young women (20-24 years old), recruited from within the informal settlement to serve as mentors to the Parivartan athletes, negotiated a 'respectable' presence outside the home, thus facilitating opening further normative space in the community for empowerment of girl athletes (Bankar et al. 2018). In this paper, we address how gender-related power relations in the family affected girls' participation in Parivartan and how, in turn, girls' participation affected family relations.

The Parivartan intervention
Implemented by the NGO Apnalaya between December 2014 and March 2016, Parivartan adapted the contact sport known as kabaddi, 1 as part of a strategy to challenge gender norms around mobility and stereotypes about what girls can do and achieve. Sports interventions have increasingly been used to empower girls with an aim towards improving their self-esteem and self-efficacy (Forde and Frisby 2015;Samie et al. 2015). Specifically, the choice of kabaddi was guided by community members themselves (girls, parents, other men and women in the community and NGO representatives). They mentioned practical reasons for choosing it; kabaddi did not require a big field or excessive sport equipment, and it was well known by most children, for instance.
The NGO first recruited 15 mentors, young women from the same community, who were trained to facilitate reflective conversations on gender equality and to coach girls as kabaddi athletes. Mentors attended a five-day residential training that included reflective sessions on gender equality, gender norms and power relations. They also participated in three two-day refresher training events specifically on the curriculum and their role within it. NGO staff held monthly meetings with mentors to plan forthcoming activities, discuss common challenges and identify best practices and solutions.
At the beginning of the intervention, mentors approached 450 families of potential girl athletes (age 12-16) to discuss and negotiate their participation in the programme. Athletes for the intervention were recruited based on age and location (12-16 years and living in the selected blocks of the community). Mentors visited the homes of girls living in the community who were randomly selected from a list derived from a household survey conducted by Apnalaya. Not all families agreed to participate, and 138 girls started the programme. Over the following 15 months, mentors led a carefully sequenced set of life skills and gender training sessions, mixed with weekly games of kabaddi. Girls met twice a week: once on Sunday to play kabaddi and once during the week to participate in the mentoring sessions (that lasted roughly one and a half hours each). On average, mentors held four sessions each month.
While separate adult sessions were planned for both mothers and fathers, men's work commitments and norms dictating that children are a mother's responsibility meant that few men agreed to participate. The mentoring sessions are thoroughly described elsewhere (Collumbien et al. 2019); the curriculum and training guide is available online (ICRW 2014).
The theory of change drew from gender performance theory and social norms theory. The former helped us focus onto the ways that symbolic and visible displays of gender expression at an individual and interactionist level ('doing gender') help to produce and reproduce the gender system (Gonz alez and Rodr ıguez-Planas 2018; Vaitla et al. 2017). Relational theories of gender helped us navigate the gender system as embedded in the complex network of institutions surrounding these girls: the family, the community, the school, etc. Within those institutions, social norms theory suggests that girls are policed for compliance with existing gender norms through social sanctions and rewards (Cislaghi and Heise 2018b). Gender norms, for example, affected where kabaddi could be played. Since norms dictated that girls should not play sports or be seen dressed in sports attire, the weekly games took place at a walled school yard where girls would be protected from public view. While the isolation of the playground protected girls from the public gaze, they nonetheless had to walk 45 minutes to the school. This walk, which girls took dressed in their traditional clothing, meant that neighbours would see them on their way to the playground, accompanied by their female mentors. Community members were, in other words, publicly witnessing a first sign of change: girls leaving their households to travel together (also with some knowing they were going to play sport).
In addition to this weekly walk, two public tournaments further challenged traditional norms against female mobility. The first tournament took place at the end of the 15-month programme, and the second 9 months after that.

Study setting and design
The Parivartan programme was located in four plots at the periphery of Shivaji Nagar, one of Mumbai's largest urban informal settlements. The community consists predominantly of Muslim migrant families mainly from Uttar Pradesh (in northern India) working in the unorganised sector and living in substandard housing with poor water supply and sanitation. Public spaces are dominated by men, and, after menarche, girls' mobility and visibility are restricted, as is their interaction with boys. Burkhas or salwar kameez with a headscarf are the requisite dress code outside the home. Various NGOs implement a range of health and development projects in Shivaji Nagar. Apnalaya, a grass-roots NGO with longstanding presence and credibility in the community, implemented the intervention. The research was implemented by the International Centre for Research on Women (ICRW), as part of STRIVE, an international research consortium. Ethical review and clearance were provided by the London School of Hygiene and Tropical Medicine, ICRW and MAMTA (a local institution that provides ethical clearance to researchers willing to work in the region of the study). We adopted a prospective qualitative research design to study the interaction between the intervention and the changing context within which it was implemented.

Data collection
In this paper, we draw on data from two rounds of face-to-face semi-structured interviews with the 15 athletes, as well as field observations. The second author (SB) conducted observations in these communities during mothers' meetings as well as mentors' debriefing sessions throughout the intervention period. Observations focused on the programme as well as the changes that were observable in terms of gender and power relations in the family. They were recorded with ethnographic fieldnotes. Data from the observations were used to interpret and add to participants' statements provided in the interviews.
Longitudinal semi-structured interviews were conducted with 15 girls during the implementation phase (January 2015 to March 2016). Girl athletes were purposively sampled six months into the programme (rather than before) to ensure representation of three categories of participants based on mentor reports: (1) girls who were consistent in their participation and were putting the new gender knowledge into practice at home; (2) girls who were consistent in their participation and who were not applying this new knowledge; (3) girls whose participation was disrupted by family circumstances not related to the programme (e.g. economic reasons, family illnesses or relocation to another village).
The second round of interviews took place 12 months after the beginning. Five athletes who dropped out of the programme were replaced. Table 1 offers an overview of participants' characteristics.
The interview guides were developed based on formative research and a literature review of relevant gender practices in India. Guides for the first round were developed to understand the aspirations of the girls and their process of joining the programme and their family's attitude towards the programme. Second-round interview guides were personalised to pursue questioning on particular issues of interest arising in the first round. Themes in the interview guides included: athletes' achievements, process of joining, difficulties and challenges encountered throughout the programme; safety and security; outcomes associated with being in a programme; perceptions of changes in themselves and in their relationships with parents and caretakers; and changes in those whose opinion the mentors valued in terms of their own behaviour (their reference group).
Interviews were conducted in participants' first language and at a time convenient for them, by either SB or a research assistant at the ICRW field office; the office was chosen in consultation with participants, as it offered a safe environment that would ensure participants' confidentiality. All interviews were audiotaped, transcribed and translated verbatim into English by a team of transcribers. SB conducted quality checks on both transcriptions and translations and, when English translations were difficult to understand, went back to the original audiotapes and made sure the English version was coherent and clear. Written consent to participate was obtained from all girls' parents; research participants also gave their assent to participate after parental consent was given. The girls' names used in this paper are fictitious.

Data analysis
We combined thematic and narrative analyses (Braun and Clarke 2006). The transcripts were reviewed and themes discussed continuously during data collection by the authors. We followed a step-wise procedure of familiarising ourselves with the data, identifying a thematic framework and developing a coding frame, using Atlas.ti for Ã Categories for girls identified at round one, as defined in the paper, were: (1) Consistent Change -girls who were consistent in their participation and were putting the new gender knowledge into practice at home.
(2) Consistent no change -girls who were consistent in their participation and who were not applying this new knowledge. (3) Disrupted Participation -girls whose participation was inconsistent because of family circumstances not related to the programme.

ÃÃ
This athlete's head of the family was her brother. ÃÃÃ These athletes dropped out before the end of the programme, despite having supportive fathers, and could not be interviewed at round two. Ira and Meher were very close to their mentor, and they left after the mentor who followed them got married and was replaced. Avni and Zoya moved out of the informal settlement. Kashvi decided she was not interested in the programme.
ÃÃÃÃ Support or lack of for these fathers was derived from what athletes said at round 2 since these athletes were not interviewed at round 1. data management. We looked for patterns in how the programme influenced the lives of the athletes and the changes it brought about in various relationships. Specifically, our line of enquiry focused on following athletes longitudinally, looking at their perspectives of change. While we interviewed athletes and mentors over two rounds, we did not collect data directly with mothers. In the analysis, however, mothers' role emerged strongly in athletes' descriptions of the changes in the relation dynamics in their families and of how they worked with their mothers to convince their fathers.

Findings
When the mentors visited the families of prospective athletes for the first time, they sought parental consent for girls to participate in the programme. Some refused, being too concerned by the physical and social risks to their daughters: they feared that their daughters would get injured while playing or assaulted by men. Others maintained that sport was not acceptable for girls, or that religion did not allow girls to play. The overriding concern was that participation in some way might decrease their daughter's marriageability or threaten family honour. Among those who consented to their daughter's participation, none gave consent on the mentor's first visit. It took multiple visits to convince the father, the mother or sometimes both. Some of the fathers became supportive of their daughters' participation relatively soon: they knew Apnalaya and thought that participating in the programme represented an opportunity for their daughters. As the girls reported in their interviews, these fathers were conscious of the norm that girls should not play sport in public, but this ultimately did not dissuade them. Other fathers were against their daughter's participation, fearing the consequences to the family's reputation of contravening existing gender norms. While in most cases fathers' refusal ended the conversation, sometimes mothers were able to change their husbands' minds. When asked who decided if she could attend the sessions, for instance, one athlete, Parinita, noted: 'The final decision to join the program came from my mother; she discussed it with my father and he also agreed. Father is the final decision-maker; however, my mother is very stubborn'.
When a girl's father was not at home at the time of the first visit, mothers would either wait and talk to him later (as most did) or take a decision (either favourable or not) themselves. Some withheld the information regarding participation from the father. The mothers who decided independently to allow their daughters to participate were of particular interest to us. Where gender norms assign the role of final decision-maker to men, women assume considerable risk. Non-compliance with gender norms could compromise the girl's marriageability, with serious economic and reputational consequences for the family.
Mothers' strategies to ensure their daughters' enrolment in the programme Mothers who lived with unsupportive fathers used one of two strategies to ensure their daughters' participation: (1) convincing their husbands to let their daughters participate; and (2) keeping the girls' participation secret.
Mothers who convinced their husbands did so by promising to take responsibility for their daughters' actions. They used four key arguments to convince their husbands: (1) that all girls in the community were playing; (2) that the school yard was enclosed, which ensured that the girls would be protected from physical or social threats; (3) that the sport would help their daughters stay healthy; and (4) that this was a unique opportunity for their daughters to practise a sport they loved, since, once married, they would not be able to do so any more. Both mothers and fathers voiced concern that being seen in public would subject their daughters to gossip, which could bring shame upon the family. But the mothers who argued for participation reassured the fathers, promising they would take personal responsibility for the daughter's behaviour.
Other mothers instead gave consent without consulting their husband, planning to hide their daughter's participation from him. Some of these mothers were not convinced about their daughter's participation from the start and hesitated, considering the risks for both themselves and their daughters. Two things were key in convincing these hesitant mothers: (1) the repeated visits of the mentors, who made themselves available with kindness, conviction and personal examples to answer any question the mothers might have; and (2) knowing that several other girls from the same neighbourhood were participating in the programme. Kavita, for instance, explained how both talking to the mentor and seeing other girls participating convinced her and her mother: 'I didn't want to go initially but once I started seeing other girls going, I got interested. Initially my mum said no, but after the mentor convinced her and she saw other girls participating, she said yes'.
In addition to reputational concerns, data revealed that some parents (both mothers and fathers) were also worried that their daughter might get injured while playing, with immediate consequences for her health and long-term consequences for her marriageability. Finally, parents were aware that their daughters' participation would reduce her availability for household chores, increasing mothers' workload. Given the risks and challenges these mothers anticipated, we were interested to understand what motivated them to support (and in certain cases even encourage) their daughters.

Mothers' motivations to let their daughter participate
A strong motivating force for mothers was their aspiration that their daughter would have a different life than the one they themselves had. These mothers were mostly migrants from rural parts of North and the Eastern India, either employed at the dump yard, 2 as domestic workers or helping their husbands manage the local grocery shop. They knew that their daughters' opportunities would be limited in their natal rural villages due to the restrictive norms that limited women's opportunities. As some mothers discussed during field observations, when they were younger they had had aspirations of going to school, but could not do so. Now, exposed to new prospects offered by the urban environment, they hoped their daughters would seize such opportunities and find good jobs. These mothers hoped that through Parivartan and similar programmes, their daughters would learn new things and would be exposed to new ways of being. Supportive mothers were particularly fascinated by what the girls were learning in the reflection sessions and, week after week, would ask them to share their learning.
These mothers' aspirations for their daughters were powerful motivators. Not only did mothers help their daughters enrol in the programme, they also made sure they would continue participating, which, as we describe in the next section, required specific strategies.
Mothers' strategies to ensure their daughters' sustained participation As the programme started and progressed, mothers who had kept their daughters' participation hidden had to find ways to justify their daughters' absence with their husbands. We identified three strategies they used to do so.
First, some mothers did not say anything to their husbands about their daughters' participation until the end of the programme. In most cases, their husbands were the sole bread winners for the family so they would work outside of the household for several hours each day. The fathers' absence allowed these mothers to keep the girls' participation hidden. When questioned, some lied to their husbands at times, telling them that their daughter was out to visit relatives, or that she had been sent to the market. Others, when questioned, admitted that their daughter was out to participate in an NGO programme, but omitted the fact that the programme included sport. Finally, another group of mothers decided to tell the truth to their husbands. Some had to: the neighbours or relatives would inform him, or he would find the kabaddi gear in the house. To encourage their husband's support for the daughter's participation, these mothers provided evidence of the benefits the girl was receiving and emphasised that many other girls were also participating. As Lakshmi recounted: My mother told my father: 'See how these girls are talking? So, what if she is a girl, can't she do anything? Only the boys can do? … Whatever she wants to become she can do that. Let the family speak, if they like'. (Lakshmi) While the reflective sessions were strategically planned not to interfere with girls' religious and educational commitments, the kabaddi sessions necessarily took place on Sunday, the only day when the playground was available. On Sunday most family members were at home, which increased household chores. To facilitate their daughter's participation, these mothers either helped them with the chores, redistributed the tasks to other female relatives, or completely took the burden on themselves. As Ira observed: 'Yes, I do have household chores, but my mother does that. If I need to go on Friday and Sunday then she doesn't give me any work, she does it on her own'. These strategies were not always successful. While most mothers were able to convince their husbands of the proven benefits of the programme, some athletes (a minority) had to drop out when their fathers found out.
Evidence suggests that mother/daughter relationships strengthened as the programme progressed. Much of what improved, according to the girls, was their capacity to communicate and trust each other. Prisha, for instance, said: 'Earlier I would not talk much to my mother, but now I talk to her about everything'; and Zoya said: 'So, because of all this I felt so good. We can share our secrets with each other'. The sessions became a conversation topic for mothers and girls, who reflected together on the session content. Kashvi, for instance, said: 'My mother and I have a friendly relationship … I discuss the sessions with her, and she likes the topics that we address in these sessions'. These girls also reported an increased mutual trust that followed both the sacrifices of the mothers for their daughters and the fact that daughters respected these sacrifices and behaved in ways that protected their mothers' and family's reputation. Kavita, for example, observed: My mother has started to trust me; now if I am late, she doesn't question me, and trusts that I can handle a situation [with men] if it arises. I am close to mother now … she shares her feelings with me, and I try to cheer her up. (Kavita) Finally, conversations about the programme became an opportunity for mothers to reflect on how the girl was changing thanks to the classes, and likely facilitating mothers' reflection on their own lives. Naira, for instance, said: 'my mother appreciates what I am learning and the changes this has brought about in me … She now shares her worries with me'. Kashvi remarked on the pride that her mother developed because of her daughter's success: My mother feels good that I am aware of a number of things and that I do not hide anything from her; she also appreciates that I have changed, our relationship has changed. Now she says she wants me have a bright future and a dream of making me an IPS officer. (Kashvi) These mothers were changed by witnessing the change in their daughters.

Relationships between fathers and daughters
Girls also reflected on how their relationship with their father had changed through the programme. The daughters of unsupportive fathers said they were uncomfortable with them: they rarely spoke to each other, and some even feared them. Take, for instance, what Prisha said when asked before the programme: 'I am terrified to tell my father about the programme; I am determined to never let him find out because he is very against it and he will beat me'. More generally, the girl athletes reported that it was common in their community for fathers not to interact with their daughters, especially once the daughters had reached puberty. The mother would generally relay any important information to their father and sought his approval for anything that the girl wanted to do.
In the second round of interviews, several girls with fathers who were initially unsupportive said that their relationship with their father had changed considerably. They felt more capable to speak directly to their fathers and their brothers, without need for mediation through their mothers. Even reluctant fathers began to remark on the change in their daughters. The content of the gender sessions as well as her sport successes became a topic of conversation in the household. For instance, Lakshmi described the change in her family relations: 'My father would ask did you play? I would say yes, I played. Then he would say go and wash your hands and legs and then he would give me breakfast. Then he would ask me "how did you play?"' Towards the end of the programme, initially unsupportive fathers encouraged their girls to participate. Some became less restrictive with their daughters' clothing and mobility. Kashvi explained how her father would listen to her dreams and help her plan for the future: [My father] says study more, whatever you don't understand study more, try to do it more. I told him what I want to do continue my education and he said okay and told me to study hard without worrying about the money. He gives us whatever we want. (Kashvi) Lakshmi reflected on how her father supported her, when she said: My father feels proud of me … He now encourages my participation in the program and … encourages me to continue when I loose in the practice matches. Not many fathers in the neighbourhood encourage their daughters as he does. They want their daughters to be in the house and do the household chores; when chores are not done, they beat their daughters. I feel proud of having a father like him. (Lakshmi) That this girl would compare her father with others, and remark on his progressive and compassionate attitude, was a testament to how mothers' strategies to engage unsupportive fathers, paired with a gender-transformative programme, achieved important results. Lakshmi was not an isolated case. Over time, girls were able to establish more direct communication with their fathers, reducing the need for mothers to mediate.
Girl athletes mentioned three reasons for their improved relationship with their fathers. The first was that, over time, they earned their fathers' trust. As they participated in the programme, there were no complaints to the family about them being engaged in any unacceptable behaviour. The second was the transformations that the programme facilitated in the athletes themselves, as witnessed by their fathers. As the athletes improved their negotiation skills, they became more capable to persuade their fathers about the importance of their participation in the programme. The third reason was their participation in the public kabaddi tournament. This was the first time that the girls had played in front of the entire community. A number of fathers attended; those who initially objected to their daughters' participation saw how cherished their daughters were. Even those fathers who did not attend the tournament celebrated the achievements of their daughters at home.
Both mothers and fathers noticed that their girls were changing in how they negotiated their space and rights in the family; sometimes, girls' increased voice in the family led to misunderstandings and tensions. Parents talked about this with the mentors who adapted the programme to help girls negotiate in ways that were harmonious, yet assertive, rather than conflictual. Mentors' weekly interactions with parents also increased family capacity to communicate and helped deal with misunderstandings that emerged during the programme.
The change in the support of the father happened gradually, with some fathers becoming supportive sooner and others later. Some, for instance, asked their daughters to wear their kabaddi gear only at the playground, while others allowed their daughters to leave home only for Parivartan and nothing else. Others were worried about what neighbours would say and publicly would not acknowledge their consent. Still others took on to convince their friends and relatives to let their own daughters to participate: 'First my uncle also didn't treat his daughters good, but seeing my father change, he also has changed and treats his daughters well', said one girl.

Navigating community norms
At the beginning of the programme, some parents worried about what others would say should their daughter be seen outside, dressed in kabaddi gear in a place where she could be harassed by boys or men. As the programme continued, however, these families became better able to resist these pressures.
Supportive fathers were much less worried about what others would say than were unsupportive fathers. Several mothers were concerned about community gossip as well. Even the supportive mothers were worried that neighbours and relatives would deem them responsible of having raised their daughter poorly. Naira, for instance, said that, 'the neighbours used to say [to the mother] that you should not send her, this and that'. In certain cases, neighbours' influence was very powerful; one mother, for instance, kept her daughter from participating in the first sessions after her neighbours questioned her about it. And recall Kavita, who mentioned that her father would not care about what others said; she also added: 'But if someone tells my mother, then she won't let me go'. Not all mothers were, however, sensitive to the neighbours' voices. Naira's mother, for instance, stood up for her daughter since the beginning of the programme: 'My mother and grandfather would tell me that you should not pay attention what others are saying. You should do whatever you believe in'.
The disapproving voices of neighbours and relatives did not weaken as the programme advanced. Rather, as the sessions unfolded, parents became better able to ignore these voices, becoming convinced that their daughter was trustworthy and would not compromise family honour by meeting boys while away from home. Take, for instance, what Pihu said when she described how her mother shushed her neighbours who were trying to raise suspicions about her daughter's decency: My neighbours would say that a girl like me is full of herself, that she will be out of control, this and that. They would tell things like these to my mum … My mum would say that leave it, let her be the way she wants to be. (Pihu) As the trust between parents and the girl athlete increased, the family's willingness to resist community norms increased. These families' gender norms relaxed, improving adolescent girls' space for action and learning. Lakshmi, for instance, explained how after the programme her parents would listen to their neighbours disapproving of them and, yet, would let her leave the household: 'Many people used to tell my father … why do you allow your daughter to play, etc. But my father said that no, if my daughter wants to play … we will support her, and we will always support her'.

Discussion
The Parivartan programme began to create a shift in the gender relations that were embedded in girls' everyday practices, such as playing sport outside of the household. The visibility of these changes facilitated a shift in what girls and parents believed to be acceptable in the community, bringing about gradual changes in girls' social interactions and eventually increasing their freedoms. Girls' presence in a public space, playing sport, could have exposed parents to social disapproval. These parents, instead, witnessed the community accepting their girls. Mothers worked out a wellthought-out long-term strategy to transform family norms and allow their girls to participate in the programme, adding to the larger historical process of transformation taking place around them.
These results were unexpected to the Parivartan staff. During the design phase, the programme staff had anticipated less dramatic changes in a relatively larger group of girls. Instead, the programme achieved greater gains (especially in terms of reducing mobility restrictions for girls). Parivartan was a pilot project to understand the potential of working with mentors and girls to contribute to a gradual process of gender norms change in an informal settlement. Results suggest changes in daily social interactions that are suggestive of transformation in gendered family relations.

Implications for theory and practice
Our findings carry lessons for both theory and practice. Three key insights may help researchers and practitioners working to transform discriminatory gender norms affecting adolescents.
The first insight is related to the role that family communication can play in family members' resilience to gender norms. That family harmony can have positive effects on members' lives and agency is not a new finding, albeit not one that has been studied widely. Allendorf (2012), for instance, used survey data from the State of Madya Pradesh, in Central India, to show that family relationship quality can have great influence on women's agency. On the other hand, how trust among family members affects resistance to restrictive norms in low-and middle-income countries is a relatively unexplored field. In spite of recent theories on how family communication might be influenced by social norms (Le Poire 2006), there is relatively little evidence from South Asia on how improving family communication might work to change gender norms affecting young women's lives. In our study, we found that some girls could not participate in the programme because their parents were afraid that transgressing norms against women's mobility and sexual purity would result in loss of family honour. As mutual trust between parents and girls strengthened, and more generally the quality of their relationships improved, participants described parents' increased capacity to silence relatives and neighbours who tried to pressure them into compliance with dominant norms. Future research should tease out the patterns and mechanisms through which small groups nested within larger groupssuch as families living in a communitycan build reciprocal trust to resist discriminatory gender norms; this potentially facilitates the adoption of more gender-equitable norms in the larger group.
The second key insight relates to the dynamics of transforming patriarchal gender norms. Rather than contesting their husbands' decisions and confronting them openly, mothers who had aspirations for their daughters and wanted them to participate in Parivartan negotiated on their behalf from within the space that the patriarchal gender system allowed; this ultimately expanded the space for both them and their daughters. Some, for instance, first hid their daughters' participation from their husbands and then, at the right time, told their husbands the truth, leveraging both family pride and girls' change to convince their husbands of the usefulness of the programme. In the same way, Parivartan created sport opportunities that, at the beginning, did not unsettle gender norms: the playground was protected from other people's views. This suggests that dynamics of gender norms transformation can begin in the cracks in the patriarchal gender order and that practitioners should identify similar opportunities, helping those who are capable and willing to act within those cracks. Other studies from the same context provided evidence of the nuanced roles that mothers can play in child-related decision-making and protection as, for instance, in Ghosh's ethnography of mothers' roles in protecting their daughters from domestic violence (Ghosh 2015). Importantly, these supportive mothers should not be left unsupported. As mentioned in Parivartan, practitioners should ensure that these risktakers, working to dismantle patriarchy from the inside out, are buttressed and cared for throughout the programme.
The third key insight that emerged speaks to the importance of the two kabaddi tournaments. These public events served as an opportunity to make consensus visible and potentially expand it: not only could local people witness the new behaviour of girls, but they could also witness others accepting and even supporting it. This finding corroborates evidence from studies that demonstrated the importance of public events as a means to challenge inequitable gender norms (Cislaghi and Heise 2018a;Starmann et al. 2018). We suggest that changing gender norms within several small groups first and then gathering the community to witness this change are both important strategies within programmes designed to transform gender norms.

Limitations
Since mothers, by virtue of being at home, were gatekeepers during the recruitment process, our sample excludes couples where the opposite dynamic (with mothers not lending support) pertained. Possibly some reluctant mothers may have withheld the information from the father without ever discussing it. Similarly, the families in our dataset are not necessarily representative of other families in the community. Those who agreed to participate in the programme (and were hence part of our sample) were probably more gender-progressive already than those who categorically refused for their daughters to be enrolled.
The dataset we analysed here does not allow us to understand the extent to which the changes that the programme facilitated were sustained in the long term. What we did find in the data, however, is evidence that norms about what girls could do began to shift. While we are appreciative of the time and effort it takes to bring about longlasting changes in the gender-normative system, we did see some attitudes and family norms begin to shift in a positive direction. No single programme can transform norms on a larger scale; rather, gender transformation is an ongoing historical process that happens over time and is informed by numerous events, including programmatic interventions.
Finally, data were collected by co-author SB, who works for ICRW, the organisation that funded the programme. However, the co-authors in the academic institutions were independent and rigorously cross-validated the analyses, making efforts to identify and uncover biases in their interpretations.

Conclusion
In this paper, we analysed data from a specific subset of participants in an evaluation of the Parivartan sport-based programme. We looked at the ways in which daughters negotiated their participation and at how mothers helped them achieve their aspirations. Daughters' participation was transformative for both mothers and fathers, with certain families increasing mutual trust and, consequently, resilience to discriminatory gender norms. We found that interventions to transform gender norms might increase their effectiveness if they: (1) work to increase family communication and trust; (2) create space and opportunities to challenge patriarchal gender norms in the space that patriarchy provides to them, leveraging cracks within that space; and (3) facilitate emergence of new norms within small groups first, and then make that consensus visible with public events. These three findings signal new directions for research and practice to increase understanding of how men and women can work together to transform inequitable gender norms and achieve greater community wellbeing. Notes 1. On kabaddi and its resurgence in India see Khawaja 2016. 2. The local landfill site for the disposal of waste materials by burial.