How Mothers in Poverty Explain Their Use of Corporal Punishment: A Qualitative Study in Kampala, Uganda

Corporal punishment in the early years is associated with antisocial behaviour and violence, but little is known about its social and cultural context in low-income countries. This paper analyses how 12 deprived women in Kampala, Uganda, perceived corporal punishment, drawing on repeated semi-structured interviews. All thought it was sometimes necessary, for three main reasons. First, it was an important strategy to ensure good behaviour and maintain their and their child’s, respectability, crucial to self-respect given severe poverty. Second, it was a means of establishing household routines and managing scarce resources. Third, it was a way to protect children from health risks. However, all mothers thought corporal punishment could be excessive, and most said it can be counter-productive, making children ‘stubborn’. There appeared to be considerable variation in their degree of harsh parenting and emotional support. These findings could inform culturally appropriate interventions to reduce violence against children.


Introduction
Developmental psychologists have emphasised that parenting practices in the child's early years are a key mechanism in the development of persistent antisocial behaviour and violence (Murray and Farrington, 2010;Hawkins, 1998;Farrington, 1998;Smith and Stern, 1997). Meta-analyses highlight the importance of harsh and coercive behaviour, especially corporal punishment, and the lack of parental positive reinforcement and responsiveness as contributors to antisocial behaviour (Gershoff, 2002;Rothbaum and Weisz, 1994).
The evidence for these mechanisms comes overwhelmingly from High Income Countries (HICs) and far less is known about parenting and child outcomes in Low and Middle Income Countries (LMICs) (Lansford and Deater-Deckard, 2012). However, increasing research evidence from LMICs suggests that the associations are the same as those in HICs (WHO, 2007;Fulu et al, 2013;Murray et al, 2013;Devlin et al, Submitted). Thus, one important strategy of reducing violence throughout society is to reduce parental violence against children.
In this paper, we examine the parenting practices of mothers living in poverty in urban Uganda. As in much of SSA, in Uganda, there is a marked contrast between policy and practice on corporal punishment. The government has expressed a commitment to prohibit corporal punishment in all settings and it has been banned in schools since 1997 (Devries et al, 2013). The Children Act and other laws protect children from violence and abuse in the home, but under common law there is right to administer 'reasonable chastisement'. In practice, there is near universal social acceptance of corporal punishment and children's exposure to violence is common in both homes and schools (Devries et al, 2013). The prevalence of corporal punishment varies considerably between surveys in different parts of the country, with findings from 33 to 90 per cent for exposure to corporal punishment in the home and from 35 to 81 per cent in schools. Heavy work, such as digging, cleaning pit latrines or collecting water, and the denial of food are similarly widespread punishments (Global Initiative to End All Corporal Punishment of Children, 2017). A particularly disturbing survey of 500 18-24 year old women in Uganda found that 94 per cent had been subjected to physical violence in childhood (Stavropoulos, 2006). Beating with an object was the most prevalent form of physical violence (86 per cent), while 55 per cent reported being punched, 53 per cent being denied food, 48 per cent forced to do hard work, 27 per cent being kicked, 20 per cent being choked/burned/ stabbed, and 18 per cent being locked or tied up. At school, girls were told they were beaten for being late, for misbehaving, or for being disrespectful. At home, the reasons given for beating or hitting girls included for breaking/losing something, being disrespectful to their elders, or not doing housework (Stavropoulos, 2006).
In many HICs, programmes have been delivered to try and modify parenting practices associated with child aggression, with some degree of success (Piquero et al, 2009). The expert consensus is that such programmes are more effective with younger children and should ideally be implemented in infancy, before child problems become established (Tremblay, 2008), when parents can more easily be helped to influence their child's behaviour (Kazdin, 1993), and when key developments in young children's social behaviour and emotion regulation can be exploited (Tomasello, 2007;Kochanska et al, 2001).
Recently, there has been increasing policy concern to address youth interpersonal violence in LMICs (WHO, 2013). Evidence is emerging that parenting programmes, delivered by trained lay workers, can be effective in improving child outcomes in LMICs, at least for child cognitive development and attachment (Engle et al, 2011), and there are calls for this approach to be applied to tackle aggressive child behaviour to prevent the development of youth violence (WHO, 2013;Knerr et al, 2013). However, in order to assess the appropriateness of transferring early years parenting programmes from HICs to LMICs, or to develop new ones for LMICs, it is first important to understand the nature of the problem, the factors perpetuating it, and which of these are malleable (Wight et al, 2016). In the case of harsh parenting, this requires understanding mothers' aspirations for their children and their beliefs about child behaviour and discipline, and identifying how they talk about child rearing, i.e. their 'emic' concepts. Crucially this should clarify what their motivations might be to maintain or change problematic behaviours. Considering that the mothers in our study were living in poverty, we also pay attention to their accounts of resource constraints, and how these may affect their ability to change their parenting behaviour.
The mothers were part of the 'Managing Difficult Behaviour' project, a study of how mothers in deprived neighbourhoods in Kampala, Uganda, manage their young children's behaviour, specifically, behaviour that they found challenging or 'difficult'. The project is part of a wider exploratory study, 'Young Children's Behaviour and Local Constructs of Parenting' in Kampala, Uganda and Cape Town, South Africa, collecting both quantitative and qualitative data to inform the development of parenting interventions for early prevention of male violence. In this paper, we report qualitative findings from repeated interviews with 12 mothers in Kampala to answer the following research questions: -What do mothers perceive as the greatest challenges in bringing up young children? -What strategies do they have for managing the challenging behaviour of their children? -In what circumstances do mothers practice corporal punishment? -How do they perceive the advantages and disadvantages of corporal punishment?

Research Setting
This study was conducted in Kampala, the capital of Uganda. In 2014, the population of Uganda was 34.6 million persons (Uganda Bureau of Statistics, 2016). Various cultural and ethnic groups inhabit the country. The largest ethnic group in Uganda is the Baganda, living primarily in central Uganda, especially around Kampala, followed by the Banyankole. Amongst the numerous languages spoken, English is the official language, although the use of Swahili is being promoted and integrated within the East African Community (EAC). Approximately two-thirds identify as Christian and one-third as Muslim, with Roman Catholics the largest Christian denomination followed by Anglicans (Uganda Bureau of Statistics, 2016).
The international poverty line is set at $1.90 a day per person (using 2011 prices) (World Bank, 2016) and The World Development Indicators show that in Uganda in 2012 the Poverty headcount ratio at $1.90 was 34.64 per cent (World Bank, 2016). As such, it is vital that those developing interventions in this context take the constraints of poverty into consideration. In LMICs generally, and Uganda specifically, it is paramount that programmes are designed and implemented with poverty in mind. In this paper, we explore the possible impacts of poverty on parenting practices.

The Sample
We draw on findings from qualitative repeat interviews conducted with 12 mothers aged 19-45 years in Kampala, Uganda. Interviews were conducted between 2014 and 2016. Ethical approval for the study was obtained from the Uganda Virus Research Institute (UVRI) and the Uganda National Council for Science and Technology (UNCST).
In 2008, a large cohort of women, highly vulnerable to HIV, was established for both research and the provision of health services (The Good Health for Women Project). By 2011, approximately 1027 women had been identified from 'hot spots' for sex work in Kampala, and recruited to the cohort. They were followed up and offered treatment for HIV and other sexually transmitted infections (STIs), as well as other health conditions.
For the purpose of this qualitative study, we recruited 12 mothers of boys or girls within the age range 18-42 months, consisting of six older mothers ([28) and six younger mothers (\22). A random sample of women from the cohort meeting these criteria was generated by the data manager and approached by a member of the study team with a request to participate. Due to inconsistences in reporting of age, and/or unsuccessful attempts to reach them through phone contact, insufficient women under the age of 22 years were identified in the initial sample. To address this issue, women aged 22 years and under were identified from records of those attending clinic services; demographic and clinical data collected as part of the clinic triage system enabled the research team to identify a random sample of women who met the study criteria. In total, out of the 26 women approached, 12 women agreed to participate. Reasons for non-participation included potential participants being away from Kampala at the time of the study, or being ineligible due to not living with the child in the age range targeted by the study.
In line with the original sampling criteria, half of these mothers were aged 22 years and under with the other half aged 28 years and over. Within each of the two age groups, three of the mothers had at least one boy within the age range with the other three mothers having at least one girl within the age range. A more detailed summary of participant characteristics is found in Table 1 below.

Data Collection
After obtaining informed consent, semi-structured interviews were conducted with participants by the second author, an experienced female qualitative interviewer based within the Good Health for Women Project. Interviews took place within the clinic providing health services for the wider cohort of women, a setting familiar to participants. An interview schedule was used as a basis for discussion about key topics, and included questions about: the mothers' background; household composition and material circumstances; mothers' relationships with their children and wider family; their parenting and disciplinary practices; and their views on parenting interventions. The data analysed for this paper were confined to responses to questions explicitly focused on young children aged 18-36 months. Nine of the 12 women had more than one child, but during the interview they were asked explicitly to focus on their child aged between 18 and 36 months when answering questions. Repeat interviews were conducted. All mothers were interviewed twice, and two mothers interviewed three times. A key aim of this approach was to enable trust and rapport to develop between the mother and the interviewer. Benefits of conducting repeat interviews relate to the quality of the relationship established with participants and the opportunities afforded to develop greater understanding about mother's parenting practices. In addition, repeat interviews allowed the researcher to seek clarification and additional information about issues raised in earlier interviews.
Interviews were conducted in Luganda, the most widely spoken language in the study area. With the permission of participants, interviews were audio recorded and later translated and transcribed into English. Participants' were assigned pseudonyms to protect their identity.

Data Analysis
A thematic framework approach to qualitative data analysis was applied . Initially, all twelve individual interviews were read multiple times by WN, NB and ES, and written memos around recurring patterns and themes recorded. To enable systematic comparisons to be made across the data, all transcripts were imported into the qualitative data analysis software, NVivo. All authors contributed to the preparation of a coding framework, which was piloted independently by WN and NB on two transcripts. The initial coding was reviewed with the aim of identifying and incorporating any emergent issues or themes. After consensus was achieved around the coding framework it was then systematically applied to each transcript. Credibility and trustworthiness of the data were ensured by discussing and comparing codes and themes within the research team. For the purposes of the analysis reported here, data were examined by DW and NB to identify common reasoning around disciplinary practices, specifically corporal punishment, and to identify illustrative quotes (White et al, 2003). Throughout the process of analysis, specific attention was paid to divergent and contradictory cases, and efforts were made to contextualise findings within the mother's accounts of their everyday lived experiences, as well as the wider sociocultural context in which they are embedded.

Findings Poverty
Poverty, specifically lack of money, was seen as a key challenge in mother's roles as parents, most experiencing major difficulties paying for rent, food, healthcare, education and material goods for themselves and their children.
My children are used to the miserable state. Sometimes we mingle posho [maizemeal] with salt and have it without sauce. We also take tea with no sugar. My children are used to it. (Marjani,30) Nabirye (45) commented that her 'biggest problem is looking at the children go hungry all day', while Taifa (33) said hers 'is when this young one falls sick and I don't have any money [for health care]'.
Lack of money for children's school fees was a primary concern for Mangeni (22): 'Paying their school fees is the hardest thing. Sometimes the older one is chased from school and I don't have a single penny'. Lack of money for school fees was perceived as particularly problematic  (Nabirye,45). Finally, Ziraili (28) described the consequences of prioritising spending on her child over herself: I get very little financial support. I can't take very good care of myself: I can't buy expensive clothes if my child has no clothes. I can't buy any good things to eat before the child eats. I don't get much from my business.
Nine of the mothers described themselves as single, one being widowed, and the other three received little, if any, financial support from their husbands: 'He didn't support us at all… The man mistreated me but did not support me in any way' (Marjani,30). Indeed, regardless of their marital status, all participants described themselves as the main income provider. The need to provide materially for themselves and their children meant many of the mothers reported engaging in transactional sex, specifically commercial sex work, 1 as a way of moving out of extreme poverty and gaining autonomy. An important impact of this occupation on their roles as mothers was the difficulty in spending time with their children, given their unsocial working hours. Several commented on this: That is why I said we don't have enough time. We dedicate very little time to our children. Think of the time I spend with them in the morning before school. I know it is very little. I can't be sure whether the child was bathed or not or had a mosquito net spread over her bed at night. All I do is seeing them wake up to go to school. I haven't lied to you. I told you the truth about time. (Ife,32) A further concern was that children would become aware of their work, potentially undermining their authority as a parent, impacting on their relationship with their child(ren).

Child Care
Most mothers expressed frustration with the lack of reliable support for childcare from relatives, neighbours and, to a lesser extent, maids. They reported that their children's fathers provided inconsistent or no financial support or child care: 'My children don't get any care from a father' (Masani,29). Their preference was to receive support from their own mothers, but for many this was not possible due to the mother living at a great distance, having other commitments, or having died. Resorting to support from other relatives was not always satisfactory. For instance, Dembe (19) commented that she did not want to burden her mother with her grandchild but that she would 'worry so much through the day' about whether her cousin would feed her child on time or take care of her baby. Those who employed maids for childcare could also face problems. Taifa (33) described how her child: 'told me that the maid used to beat them a lot and prohibited them from ever telling anyone. She used to take men home disguised as uncles. From that day, I gave up on maids'. Consequently several participants left their children with a particular 'lady' (child minder) in the neighbourhood, who was known for caring for multiple children in the local area. However, some mothers said they had little or no help with childcare.

Rationale for Discipline
The reasons mothers gave for disciplining their children can be grouped into three broad categories: to establish good, or 'respectable', behaviour; to establish necessary domestic practices; and to protect children from health risks. Key aspects of good behaviour included not fighting, stealing or swearing, and showing appropriate respect to adults. When asked which child behaviours are most problematic to parents, five mothers referred to children being rude to their parents (primarily mothers) and/or not willingly following their instructions. Mothers were particularly concerned that their children's behaviour in public should demonstrate good upbringing, this being critical to the mother's respectability and that of the family in general.
INT: Are you happy with … your child's behaviour?
I want her to remain that way. She doesn't disrespect me, she is very obedient. If you tell her to go greet someone, she will go, kneel down and say, 'how are you?' since that's the only greeting she knows. She cannot greet someone while standing, she always kneels down…. She knows where the place, where to do her long calls [defecating]. If she can't take off her clothes by herself, she asks me to help her and then goes to the right place. … She can't beg for tea in the neighbourhood. She always tells me when she wants tea. She cannot eat left over pieces of food without asking for permission. She isn't badly off yet….
You are supposed to teach the child not to respond rudely while an adult is talking. If you don't you will be embarrassed in public when you tell him something and he announces that he won't do it. To other people, you will be the one to blame because it is your duty to discipline the child. (Marjani,30) Mothers placed particular emphasis on the importance of children 'greeting' adults appropriately. This was understood as being critical not only in demonstrating respectable behaviour, but also in maintaining good social relations within the wider community. When asked about why greeting was important mothers explained: I may greet you in the morning and find out how your night was. I may not have anything to do for you if you had a bad night but at least I will know. I will also be glad if you had a good night. It is a sign of togetherness. You cannot greet someone you hate. You may not have a solution to someone's problem but you will know their status. That is why we greet. If I went passed you without greeting, I won't know if you are going through a tough time. (Mangeni,22) It's important to learn how to greet because it teaches to give respect to people and he helps to bring connection to one another. (Eshe,21) Several mothers commented on how annoyed they were if other adults informed them that their child had misbehaved somewhere in the neighbourhood. Their concern probably related, at least in part, to a perception expressed by one mother: A child's behaviour is adapted from the mother. It is said that a child's bad behaviour is a display of her mother's behaviour. I don't think a child's behaviour can be any different from her mother's. (Afia,22) A particular anxiety mentioned by half the mothers was that their children might be indiscreet about adult conversations overheard at home. It was clear from a few comments that this related to inadvertently disclosing their mothers' occupation: A child shouldn't talk [participate in conversation] when adults are talking. She will later tell her friends all she had. Remember these are private conversations. Our talk is vulgar. We tell each other what the streets were like the previous night. How can a child listen to such things? (Marjani,30) There are some conversations that children are not supposed to listen to at all…. You might talk about the streets where you worked the previous night. The children shouldn't listen to such. Those are adult Many examples given of incidents that prompted corporal punishment related not to wilful misbehaviour but disruptions to 'correct' household management, such as not doing domestic chores properly or following hygienic practices. Mothers were particularly concerned about wasted resources, as in children being careless and breaking household utensils, or losing their shoes or dinner money.
If a child steps on a plate and you don't blame him while still a minor, he will grow up thinking that it is okay. If you forbid him to do it from childhood, he will always know that it is bad….. For example if there are plates lying on the floor, he shouldn't jump them but rather remove them and place them in their right position before passing. (Nabirye,45) A far less common cause for punishment was to protect children from potential health risks, two mothers mentioning sexual health risks (HIV and defilement) and another playing with electrical appliances. However, the latter also related to concern about resources: There are some actions that are extreme and call for a beating. I would definitely beat my child if she played with electric appliances or dissembled the radio or DVD player. In case you are away and the child does such things, can be fatal. I punish her there and then. If the electricity is off and she bothers with the appliances without switching off first, it can return any time during the process and shocks her. So, I beat her with a lot of brutality so she won't do it again.

INT: What do you mean by brutal beating?
Slapping her once or twice isn't enough. Instead I use a cane to strike her seriously. I point out to her that the electronic is expensive to replace if spoilt.
INT: What if she spoils someone else's item?
I still punish her. Remember you have to replace that item. That person might give you two or three days to have it replaced but if it belongs to you, it could even take a year without being replaced.

Disciplinary Strategies
Different options for disciplining children, beyond corporal punishment, were explored during interviews. The mothers varied considerably in the number of options they mentioned, from one to over five. All but one referred to withholding food, and all but two to 'counselling'. The latter was universally recommended, and most said it should be the first response to bad behaviour before corporal punishment. However, it was not always clear what 'counselling' entailed. Some mothers talked of the need for calm advice -'When you tell a child something gently, it will stick to her mind' (Taifa, 33) -but others might have meant scolding and threatening punishment.
I have a friend that I usually argue with for being too harsh on her child. She says she scolds her because she wants her to learn how to do chores. She barks at her so hard that she drops anything she has in her hand at the time. I always tell her to teach her gently to avoid scaring her. (Ife,32) There were polarised views on withholding food. Two mothers recommended it, in moderation (e.g. not for the whole day): INT: How can a child be punished?
There are so many ways to do that. It is possible that you buy your child a sweet every day, if he does something wrong, you can go three days with buying the sweet. That's enough punishment even though you haven't used a cane. You could cook meat and give everyone else apart from him. That's a punishment too. My grandmother used to deny us food sometimes when we messed up. You would never do the same mistake again because you wouldn't want to go hungry again. (Nabirye,45) However, all the other interviewees condemned withholding food as too severe, corporal punishment being preferable.
I cannot go hungry all day. It makes me feel so weak and lazy. How can I deny a child food then? It isn't good…. How can you cook food and serve the rest of the family except for one child, simply because he made a mistake? (Namono,22) One mother said that withholding food is dangerous for a child's health and another that it may push the child to steal out of hunger. Beyond 'counselling' and withholding food, other means of punishment mentioned included: showing disappointment in the child's behaviour, for instance through 'a particular look'; withholding special treats, such as sweets or new clothes; restricting them from pleasurable activities, for instance by locking them in the house, sending them to bed early or stopping them playing with friends; and referring them to others to be disciplined, such as their father, grandmother or religious leaders.

Corporal Punishment
All mothers regarded corporal punishment as inevitable in child rearing, being a primary means to respond to bad behaviour, stop it being repeated, and to 'get rid of naughtiness' (Eshe, 21). Without corporal punishment children become 'spoilt', and were described as 'haughty' (Taifa, 33), 'unruly' (Masani, 29) or having 'a superiority complex' (Mangeni,22). Several mothers observed that without beatings threats of punishment are soon learnt to be empty, making the management of difficult behaviour more challenging. Several mothers also commented that firm discipline should start from early childhood.
Despite the universal acceptance of corporal punishment, all mothers thought that it can be excessive and considered that a balance is needed between leniency and harshness. Everyone thought that children should receive at least one verbal warning before being hit, but what was considered appropriate corporal punishment varied. Marjani (30) thought one should initially only strike a child twice, it should not be done in anger and beatings should not be done daily. At the other end of the range, Mangeni (22) thought one should only give one warning before corporal punishment, and should then use three to five strokes. In between, Ife (32) described how two strokes can be sufficient, and caning is better than beating violently. She described how she first instructs her child, then threatens, and then beats: 'I usually give him five strokes but if he gets me more upset, I beat him much more'. Some mothers specified how physical punishment should be administered, for instance only on the buttocks, not before they have eaten and not in public. Amanka (21), who acknowledged that she 'can be very harsh sometimes', said that 'If the situation is not clear (about which child is in the wrong) I beat both of them' since it avoids one child becoming a favourite.
All mothers provided examples of what constitutes excessive corporal punishment, but their criteria varied considerably. Some referred to the intensity of the physical punishment, condemning 'ruthless' or 'terrible' beatings, too many strokes, striking all over the body, rather than just the buttocks, and leaving visible marks on the head or face. Several said that parents should only use a cane, not slapping or hitting with fists: Some parents do not use canes to punish children. They slap them without considering the amount of energy they have in their hands. At times you can even notice the traces of the slaps on the child's face. Such a parent beats badly. Often times they slap up to the ears and the child's face swells up. (Ife, 32) Some referred to the child's circumstances, saying that beating is only appropriate if the child understands what they did wrong, or that it is wrong to beat intensely if a child steals from hunger. Several mothers said one should not beat children simply for 'making mistakes', and one regretted doing so with her 24-month-old daughter: Another time, she defecated in her potty and left it inside the house. I beat her on that incidence too…. I actually apologised to her afterwards. She explained to me that she put it inside the house because there were people outside. She feared to get embarrassed. I apologised to her and she apologised too. I had beaten her out of anger. (Marjani,30) Four mothers observed that one should not beat one's child when angry, since it leads to excessive beating, some drawing on their own experiences as children: Adults have their own problems. The child may annoy an adult at a time when she has something bothering her. She will shift all the anger to the child. The reason my mother beat me a lot is because my father married another woman yet I really took my appearance after my father…. She used to chase him away from home instructing him to take his child with him….. So, adults can shift their problems to the child. (Masani,29) Other examples given of extraneous parental factors leading to unjustifiably severe beatings were being drunk, resenting step children, or accusing a child of misbehaviour to pursue a vendetta with the other parent.
Seven mothers commented that excessive corporal punishment can be counter-productive, leading children to become 'stubborn' or 'rigid'. Indeed, it was believed that children can become so hardened to harsh beatings that they are no longer a deterrent.
I have a sister who comes here also. The first time I came here, I came with her son. He is used to being beaten all the time. You may refuse him to do something and he does it deliberately because he knows he has to be beaten. He has been raised in a manner whereby he is beaten whenever he makes a mistake. (Masani,29) When a child gets accustomed to being beaten, he becomes rigid. You should beat him once in a while, for example once in a month, so that he forgets the pain. If you beat him too much, he will become stubborn and won't be scared of the canes so he will deliberately do wrong. (Ife, 32)

Degree of Emotional Support
In discussing approaches to discipline, some mothers seemed to endorse principles of what Western analysts term 'positive parenting', that is 'involved, nurturing, and accepting behaviour' (Lachman et al, 2014). For instance, three mothers said that showing disappointment in their child's behaviour is, in itself, a means to encourage good behaviour, implicitly recognising that children greatly value their parents' approval. Taifa (33) recommended that one should not respond to a rude child while angry but: 'Later, when you feel better, call her and counsel her. Tell her that you will only give your attention to the well-behaved children'. Taifa observed that 'children listen more if you discipline them lovingly', but countered this with the need for balance: 'You should correct a child lovingly but you shouldn't pamper her too much either. There must be a limit to love'.
Masani (29) commented on her nephew: He has been raised in a manner whereby he is beaten whenever he makes a mistake… That is violating a child's rights. When he grows older, he will feel unloved. He will wish to go to other relatives' homes.
In this way mothers recognised and attached importance to the provision of emotional support to their children. Indeed, the importance of close physical proximity and nurturing practices were also discussed by participants. 'Nurturing' practices reported by the mothers included educating and 'counselling' their children, and some participants also reported the necessity of creating a friend like bond between a mother and a child in order to eliminate fear from the child. Nevertheless, mothers need to provide materially for their children meant that they were limited in the amount of time they could spend on these more 'nurturing' activities.

Discussion
Corporal punishment was seen as essential to child rearing by all mothers, though they all thought a balance is needed between leniency and harshness. In similar settings such as Kenya, physical discipline is used frequently (Oburu and Palmérus, 2003). Although all mothers in our study had a concept of excessive physical punishment, what was considered appropriate was generally far more harsh than might be considered acceptable in Western Europe or America (Lansford and Deater-Deckard, 2012;Lansford et al, 2010), such as two or three strokes with a cane after one warning, especially since in many cases 2 the children being discussed were between 24 and 36 months. These findings suggest that it is far too simplistic to dichotomise mothers between those who do or do not endorse corporal punishment, and to attribute to the former all the negative features and poor child outcomes of 'harsh' or 'authoritarian' parenting (Baumrind, 1967(Baumrind, , 1971(Baumrind, , 1996. Rather, within this sample, and presumably amongst poor Ugandan mothers in general, there appears to be a wide spectrum of harshness in parenting practices, from highly empathetic mothers whose practices are in line with 'positive parenting' to severe parents with little evidence of empathy for their children, and all these mothers see corporal punishment as a legitimate and at times necessary option for maintaining discipline. Thus, the findings are in line with Baumrind's later work (Baumrind, 1996) in which she moved away from an absolute condemnation of corporal punishment of children and highlighted the importance of, among other factors, the cultural context and the development stage (or level) in which physical discipline occurs.
Although some quantitative studies have shown associations between corporal punishment and poor child outcomes with little differentiation between degrees of harsh parenting (Meinck et al, 2017), other studies have shown that such associations are moderated by parents' warmth and emotional support for children (Lansford et al, 2005(Lansford et al, , 2014, reflecting that parent-child relationships are rather more complex, as demonstrated in these findings. In a longitudinal study of approximately 2000 American children, McLoyd and Smith (2002) found that spanking was associated with subsequently increased behavioural problems in the context of low levels, but not high levels, of emotional support, and this pattern held across all three ethnic groups. They hypothesise that: 'The child may be less likely to view spanking as harsh, un-just, and indicative of parental rejection when relations with the parent are generally warm and supportive ' (McLoyd and Smith 2002, p. 51). Furthermore, a series of multi-country studies led by Lansford and colleagues (Lansford et al, 2005(Lansford et al, , 2010(Lansford et al, , 2014 examining parental disciplinary strategies and the effects of physical discipline, have highlighted debates around the orthodox mainstream Western position that corporal punishment is totally inappropriate for the socialisation of young children (Straus et al, 2014). For example, in their 2005 study of mothers and children in six countries -Kenya, India, China, Philippines, Thailand and Italy -Lansford and colleagues found that although high levels of physical punishment were related to adjustment problems in children, specifically child aggression and anxiety, these associations were weaker in those countries in which the practice of physical punishment was more normative, specifically Kenya (Lansford et al, 2005). These studies suggest that cultural normativeness, along with warm and supportive parenting, may mediate the deleterious effects of corporal punishment on child outcomes. Nevertheless, Lansford and colleagues (2014) found that among children whose mothers used corporal punishment and exhibited high levels of warmth anxiety increased over time, leading them to conclude that ''corporal punishment may be especially harmful in the context of high warmth'' (pp. 14). These contrasting findings highlight the need to pay continued attention to mothers' positive engagement with their children, in particular their responsiveness and emotional support, and material factors that inhibit such engagement, as well as their punitive behaviour. The extreme poverty in which most of these mothers brought up their children was frequently referred to in their discussions of managing their children's behaviour, either explicitly or implicitly. Material hardship was emphasised as one of their greatest challenges as parents, and they frequently alluded to concerns about poverty. Of the three broad rationales for disciplining their children, the two main ones were exacerbated by deprivation. For these children, respectful behaviour involved not betraying the family's poverty, for instance by begging for food, or by discussing their mother's occupation, and learning to abide by household rules was largely about the careful use of precious material resources.
The heavy emphasis placed on material poverty by study participants is in many ways reminiscent of Ngwaru's (2014) account of rural Zimbabwean and Kenyan parents' perceptions of their role in early childhood education. Reflecting on the findings of three studies around childhood education and literacy, Ngwaru notes that parents were ''preoccupied with poverty, socioeconomic insecurity, morbidity (usually arising from the scourge of HIV/AIDS), […] and lack of education'' (pp. 67-68). Parents in Ngwaru's study tended to invoke arguments reflective of Maslow's (1943) hierarchy of needs: that only after their families' needs for food, water, warmth and shelter had been satisfied would it be possible to engage in wider school and educational activities. Following Ngwaru, it could be argued that although our participants recognised the importance of their children's socio-emotional needs, describing the need for counselling and educating children, they too emphasised the imperative to meet material needs such as food, water, warmth and shelter, and the need for disciplinary strategies that supported this endeavour.
Researchers exploring child development in SSA have identified that the primary care of children is often shared beyond the biological parents (Serpell and Marfo, 2014a, b;Pence and Nsamenang, 2008). To some extent our findings support this, with mothers' noting that grandmothers, aunties, older siblings, and to a lesser extent, fathers, all play some role in child care. However, similar to findings from Zalwango et al's (2010) study exploring Ugandan sex workers accounts of parenting, for participants in this study, caregiving, both materially and in terms of disciplinary practices, was understood primarily as a responsibility of motherhood and they had little expectation of support from either the fathers of their children, or the wider community. It is also important to contextualise discussions of disciplinary practices within women's accounts of their anxieties around health risks for them, and their children. Mothers' accounts suggest that they are well aware of health risks, specifically risk of HIV. Their concerns around who would support their children if they could no longer care for them, through illness or death, shaped their understandings of the need for well-behaved, obedient and respectable children, who would be more likely to be cared for by others.
This research is part of larger study intended to inform the development of interventions. All of the mothers drew on their own experiences as children, to a greater or lesser degree, to both point to what constituted good disciplinary practices and excessive corporal punishment. Prompting adults to reflect on their own experience as children is a useful technique to encourage empathy with their own children, already used in programmes intended to reduce harsh parenting, such as Parenting for Respectability (Kazemi, 2016) or to modify norms of masculinity, such as Program H (Promundo, 2017).
The importance of motivation is recognised in many models of behaviour change (Michie et al, 2011) and designing interventions to be 'culturally compelling' (Panter-Brick et al, 2006) is one way to harness existing motivation. These findings suggest that appealing to mothers' desire for their children's good behaviour and respect, a key hallmark of family respectability, could help ensure that parenting programmes are perceived as relevant to parents, specifically mothers. Interventions are already being developed that appeal to the desire for family respectability while demonstrating disciplinary strategies that help parents move away from 'harsh' corporal punishment to a focus on responsiveness and positive reinforcement (Kazemi, 2016;Lachman et al, 2016;Siu et al, 2017). 'Counselling', which most mothers reported engaging in, is consistent with such approaches as it implies encouraging children in a supportive way to reflect on the implications of unwanted behaviours.
Since corporal punishment is clearly perceived as socially and culturally appropriate, if not necessary, as a parenting strategy to achieve respectable behaviour, it will require more than individual level interventions to change practices. The Ugandan government recognises this in so far as their expressed intention is to extend the recent ban on corporal punishment in schools to the home. Legislative interventions need to be complemented by work at the cultural and community levels (Frieden, 2010) in addition to parenting interventions, and, very importantly, the material factors that act as barriers to positive discipline strategies need to be addressed.
A key strength of this study is the rich qualitative data generated, which provide insights into the diverse ways that mothers talked about disciplinary practices used in managing difficult child behaviour. Although more research is needed, this study provides insights into the practices of a particular group of women, which may also be applicable in similar contexts (Lewis and Ritchie, 2003), and offers a basis for future large-scale studies. However, several limitations of the study should be acknowledged. These findings come from a small sample, although they are likely to be typical of Kampalan mothers living in conditions of poverty with similar social and occupational positions. It is always challenging to collect rich data on issues that are not salient to one's interviewees. Some participants were not very forthcoming about the use of corporal punishment presumably because it was not an issue about which they had deliberated much prior to the interviews. The use of repeat interviews was intended to facilitate the development of rapport between the interviewer and participants, and to encourage them to speak openly about their parenting practices. Nevertheless, it is likely that participant's accounts were shaped by a desire to present themselves in socially desirable manner, aligning their responses with socially and culturally acceptable narratives of how to manage difficult child behaviour.
Indeed, a particularly important methodological limitation is the way social desirability may have shaped women's accounts. Women were recruited within a context that emphasises the hazardous nature of widespread sexual practices related to their social and occupational positions, and in which they may feel beholden for receiving ongoing health services. In such a context, the social position of interviewees and the researcher is likely to be different, which  (Hawkins et al, 2009;Price and Hawkins, 2002;Elmusharaf et al, 2016Elmusharaf et al, , 2017, in which members of the group targeted in the research take on an active role, as 'peer researchers', in interviewing their peers about their experiences of the issue being explored. The potential benefit of this approach is that peer researchers have already built a relationship of trust with those they interview, which facilitates open discussion. This approach has been used successfully in exploring sexual and reproductive practices in Zambia (Price and Hawkins, 2002), age-disparate transactional sexual relationships in Mozambique (Hawkins et al, 2009), cultural dimensions of maternal health in South Sudan (Elmusharaf et al, 2016(Elmusharaf et al, , 2017 and in a range of ongoing projects in LMICs around the development of family planning services (Options, 2017).

Conclusions
This article has explored how mothers in deprived neighbourhoods in Kampala, Uganda, manage their young children's difficult behaviour, with a particular focus on corporal punishment. In the anthropological tradition of ''making the familiar strange'' (Myers, 2011), we have sought to highlight both the rationality and the subtlety of mothers' views of corporal punishment, views that are widely condemned by the mainstream of international development orthodoxy as dysfunctional. Understanding mothers' perspectives on corporal punishment of young children is an essential precondition for ethical and effective interventions to reduce potentially harmful disciplinary practices that are likely to contribute, in the long term, to male violence. The findings suggest three main motivations to use corporal punishment: to ensure good behaviour and maintain respectability; as a means of establishing household routines and managing scarce material resources; and as a way to educate children about, and protect them from, potential health risks. A key finding is that although corporal punishment was universally accepted and seen as essential to child rearing, there was considerable diversity in how it was administered and in the emotional connection mothers reported for their children. This implies that those designing parenting interventions to reduce violence against children should not equate the practice of corporal punishment with an absence of empathy for, or emotional connection to, their children, and interventions should build on mothers' existing aspirations to be supportive of their children. In order to do this, it is also necessary to recognise the limits which extreme poverty places on mothers' ability to spend time with their children in nurturing, counselling and facilitating their education.
supported by the European Union. The wider study was funded through an anonymous donor grant from which WN and GS received additional funding. Requests for data should be sent to the MRC/UVRI Unit (info@mrcuganda.org). However, the datasets generated during the current study are not publicly available due to lack of participant consent for public access. Notes 1. As described in the methods section of this paper, all study participants were drawn from a cohort of women at high risk for HIV infection. This cohort was established in an area of Kampala characterised by high levels of sex work. Although there was an awareness among the local study team that women recruited to the study were all involved in some form of 'transactional sex', exchanging sex for material support, is important to note that not all study participants defined themselves as 'sex workers'. While beyond the scope of the analysis reported here, the women's accounts of what can be described as 'transactional sex' speak to complex issues around agency, gender relations, cultural expectations and structural inequality explored in ethnographic research around transactional sex in the context of SSA (Verheijen, 2011;Wamoyi et al, 2010). 2. It is important to note that although mothers were asked to focus on their child aged between 18 and 36 months when answering questions about parenting and disciplinary practices, they nevertheless drew on their wider experience and knowledge of local parenting practices during interviews.