Study Site and Study Population
The study was conducted in Attiak sub-county, Amuru district in the Northern region of Uganda. This district was heavily affected by the 20-year war waged by the Lord’s Resistance Army (LRA) against Uganda’s national armed forces. At the height of the conflict, nearly two million people were relocated to internally displaced person (IDP) camps. A high level of violence was documented in these camps during the conflict (Gelsdorf et al. 2012) and has likely contributed to the existing levels of violence in the region (Saile et al. 2014). After the ceasefire, brokered during 2006 to 2008, families began to leave the IDP camps and move back to their traditional villages, reclaiming their land, and rebuilding their homes. The war deeply affected social relationships, leaving younger generations without the support of parents or elders to guide and advise them in important roles in family and community. Traditionally, land was handed down from father to son, but with an increasing scarcity of land, families are pressured to seek income generation opportunities where few are available. With the advent of basic education reforms and mandatory primary and now secondary schooling, parents and caregivers have the added burden of paying school fees. Now, in a period of post-conflict development, the government has been rebuilding schools and strengthening services in the region.
Participants were 500 young fathers aged 16 to 25 who have toddler-aged children (1–3 years) and are married or cohabitating with their wife or partner. They were recruited from villages across all eight parishes of Attiak sub-county, an area where the implementing partner, Save the Children International (SCI), previously worked and was well known as a reputable organization. Within the sub-county, Local Council 1 (LC1s), the lowest level administrative elected leader, participated in generating a list of all eligible young fathers from selected villages based on eligibility criteria. The study was conducted in two cohorts. A total of 340 men were recruited into cohort 1 in July 2013 and 160 men in November 2013. Power calculations suggested that a sample size of 500 was sufficient to detect a 5 % reduction in the use of physical violence against their wife/partner between baseline and end line.
Measures included in the survey instrument were based on the content of the curriculum and constructs in the literature that influence IPV or physical punishment. Perpetration of IPV and physical punishment measures were adapted from the Conflict Tactics Scale (Strauss et al. 1996) and Parent-Child Conflict Tactics Scale (Straus et al. 1998), respectively. Selection of items was based on those behaviors that are reported most frequently in the Demographic and Health Surveys (DHS) domestic violence module in Uganda and include physical, psychological, and verbal forms of violence. Sexual violence was omitted from the survey because it was not covered in any depth in the curriculum.
IPV: Men were asked, “In the past 3 months have you…?” for each of eight possible actions including “shouted or yelled at your wife?,” “slapped your wife,” and “pushed or shoved your wife,” with possible response options never, sometimes, often, or no response (alpha = 0.76).
Physical punishment: Men were asked, “In the past month, have you done the following to discipline the child…?” for each of seven possible actions including “shook him/her”; “shouted, yelled, or screamed at him/her”; and “spanked, hit, or slapped him/her on the bottom with bare hand,” with response options yes, no, or do not remember (alpha = 0.73).
Independent variables cover constructs known in the literature to predict IPV and physical punishment. Justification for using IPV measures were adapted from the DHS domestic violence module used in Uganda.
Justification for IPV: Men were asked, “In your opinion, is a husband justified in hitting or beating his wife if she…?” for each of three items including “goes out without telling him,” “neglects the children,” or “argues with him” (alpha = 0.66).
Justification for physical punishment: Men were read five statements including “stubborn children need to be hit to teach them right and wrong”; “if a child is old enough to defy a parent, then he/she is old enough to be hit”; and “if you love children, you will hit them when they misbehave” (alpha = 0.79). Response options were agree, partially agree, or disagree.
Two parenting practice measures were developed by the study team and based on constructs in parenting literature such as warmth and involvement that are correlated with child behavior and discipline strategies (Stormshak et al. 2000). These include positive parenting and spending time with the child.
Positive parenting: Men were asked, “In the past month, if the child did something you liked, did you…?” for each of six potential responses including “say something nice about it or praise the child” and “give the child physical affection, e.g., pat on the back” (alpha = 0.77).
Spending time with the child: Men were asked, “In the past 3 days, did you or anyone in the household over 15 years of age engage in any of the following activities with the child…?” for each of six items including “read books or look at books,” “tell stories,” and “sing songs” (alpha = 0.81).
Couple communication measures were developed by the study team. Men were asked, “In the past month, did you…?” for each of five items including “tell your wife you appreciated her” and “take time to listen to your wife” (alpha = 0.77).
All measures are additive indices with binary outcomes using cutoffs at the mean. A single item was developed by the study team to measure confidence in using nonviolent discipline, and several items were adapted from the Gender Equitable Men Scale (Pulerwitz and Barker 2008) to measure attitudes about women’s and men’s roles in child care and decision making. Program exposure was measured as attendance of at least one individual and one group mentoring session.
The REAL Fathers Initiative uses a mentoring program and a community poster campaign. These intervention components are grounded in social cognitive theory, which posits that gender differentiation is a social phenomenon with both psychological and social-structural influences (Bussey and Bandura 1999). Individuals are able to differentiate gender by the early ages of 2 to 4 years. Gender knowledge is one aspect of gender differentiation, but social cognitive theory proposes that individuals self-regulate, adapting their behaviors to adhere to gender roles and expectations across a range of social experiences and contexts. Central to social cognitive theory is learning by observation through modeling of behaviors that allow individuals to expand their knowledge and skills. The REAL Fathers Initiative uses modeling of alternative strategies for nonviolent discipline and conflict resolution to improve fathers’ parenting and communication skills and confidence in adapting nonviolent strategies. Promoting these strategies leads to reductions in the use of either form of violence at home, thus reducing child exposure to IPV and maltreatment over time. Social learning theory takes a life course perspective such that gender conceptions and roles are not static but change across the life span. Intervening with young fathers who are learning new roles as parents and partners is an ideal time to promote nonviolence in parenting and partner relationships as there is still ambiguity in the normative expectations about these roles and behaviors. The theory of change is based on the assumptions that increased knowledge and skills in positive parenting and exposure to alterative nonviolent discipline strategies leads to fathers practicing more positive parenting and improved parent-child interaction. Targeting couple communication skills, joint problem solving, and nonviolent responses to couple conflict reduces perpetration of IPV in the long term. Self-reflection on gender roles, by husbands and wives, and at the community level through exposure to posters, leads to improvement in acceptance of an expanded role for the father over time.
A total of 64 mentors were recruited, each guiding up to four mentees. Each mentor/mentee pair met twice a month for 6 months, once in an individual session and once in a group session of three or four mentors and their mentees. Two of the individual sessions and one group session included the young father’s wife or partner. Each session followed a standard format and took approximately 40 to 90 min. During each session, mentors gave the men assignments to practice new skills. Men’s experiences in completing their assignments and practicing the new skills were discussed during the following session. The curriculum used a “yellow card” strategy adapted from soccer’s yellow warning card system. Designed to avoid the escalation of disagreements into physical violence, the yellow cards could be used by either partner to communicate to their partner that an important issue needed to be discussed or to pause a discussion that has the potential to turn violent. In addition to mentoring sessions, a series of six posters were displayed on sign boards at locations in the community frequented by the young fathers. Posters changed monthly and corresponded with the themes and messages presented during the mentoring sessions. Posters included a photograph of a local father performing one of the desired behaviors, such as reading to his child, and a statement indicating that others approve of that action. After the final mentoring session, an open community meeting, or “community celebration,” took place in each study community and were attended by LC1s, program participants, and their wives and families. These celebrations supported norm change at the community level by providing fathers a public forum to commit to continue practicing new skills and for the LC1s and family members to commit their support for the men’s adoption of positive change.
Mentors were trained in delivering a structured curriculum covering three key themes of fatherhood, parenting, and couple communication. Before the intervention began, mentors received a 5-day training covering the content of the curriculum and values clarification exercises on violence and gender. A 2-day mentor training was held midway through implementation of the curriculum to assess the mentoring and to prepare mentors for the remaining sessions. Mentors were volunteers from the community selected by the young fathers. The selected mentors were interviewed by SCI staff to provide information about the project, ensure their willingness and motivation to participate as a mentor, and to map the location of the mentor's residence relative to the mentee. Mapping the residences of the mentor/ mentee pairs helped facilitate the mentoring visits and in some cases young fathers were asked to select a different mentor who was geographically closer. A meeting was organized by SCI with LC1s to introduce them to the study and solicit their support in identifying eligible young men in their communities. Bicycles were provided to the mentors in recognition of the value of their work and to facilitate meeting regularly with their mentees. Incentives were not provided to mentors or participants.
Effectiveness of the REAL Fathers Initiative was evaluated by comparing men who participated in at least one individual and one group mentoring session (exposed) versus men who did not (unexposed), in two independent samples at end line and at long-term follow-up by using cross-sectional data. Originally designed as an RCT, men were assigned to control and intervention groups by using a lottery method. Unique identification codes were not used due to concerns about confidentiality and the sensitive nature of the research topic. Therefore, the study team was not able to track group assignment. For purposes of this evaluation, men who attended at least one individual and one group mentoring session were included in the exposed group, even if originally assigned to the control group.
Men were surveyed at baseline prior to implementation of the intervention and again after a 10-month intervention period, including a 6-month period of mentoring, posters, and community celebrations and a 4-month post-implementation period. All 500 men were surveyed at long-term follow-up in July 2015, 12 and 8 months for cohorts 1 and 2, respectively, after the end of the intervention period.
The study received ethical approval from institutional review boards at Georgetown University and the Uganda National Council for Science and Technology.