Yon Entèvansyon Kominotè pou Anpeche Vyolans kont Fanm ak Tifi ann Ayiti

A Community-Based Intervention to Prevent Violence against Women and Girls in Haiti Institutions for Development Sector Innovation in Citizen Services Division Lessons Learned DISCUSSION PAPER Nº IDB-DP-594 Authors: Patrick Richard Sara Siebert Junior Ovince Alexandra Blackwell Manuel Contreras-Urbina Editors: Robert Pantzer Eric L. Olson Jacquelyn Dolezal With the collaboration of the Wilson Center August 2018 A Community-Based Intervention to Prevent Violence against Women and Girls in Haiti Lessons Learned Authors: Patrick Richard Sara Siebert Junior Ovince Alexandra Blackwell Manuel Contreras-Urbina Editors: Robert Pantzer Eric L. Olson Jacquelyn Dolezal With the collaboration of the Wilson Center August 2018 http://www.iadb.org Copyright © 2018 Inter-American Development Bank. This work is licensed under a Creative Commons IGO 3.0 Attribution-NonCommercial-NoDerivatives (CC-IGO BY-NC-ND 3.0 IGO) license (http://creativecommons.org/ licenses/by-nc-nd/3.0/igo/legalcode) and may be reproduced with attribution to the IDB and for any non- commercial purpose. No derivative work is allowed. Any dispute related to the use of the works of the IDB that cannot be settled amicably shall be submitted to arbitration pursuant to the UNCITRAL rules. The use of the IDB's name for any purpose other than for attribution, and the use of IDB's logo shall be subject to a separate written license agreement between the IDB and the user and is not authorized as part of this CC-IGO license. Note that the link provided above includes additional terms and conditions of the license. The opinions expressed in this publication are those of the authors and do not necessarily reflect the views of the Inter-American Development Bank, its Board of Directors, or the countries they represent.  ABSTRACT* LATIN AMERICAN PROGRAM Violence against women and girls (VAWG) is highly prevalent in Haiti and constitutes a serious public health prob- LATIN AMERICAN PROGRAM lem. Social norms promoting power imbalance between women and men and condoning VAWG are also wide- spread. Changing these norms and curbing the cycle of VAWG in Haiti is an important step toward ensuring health- ier, more productive, and safer communities in Haiti. This report documents the lessons learned from a review of the planning, implementation, and evaluation of community mobilization interventions concerning VAWG in Haiti, namely the SASA! program by Raising Voices and the Power to Girls program by Beyond Borders. The methods used to develop this report consist of a review of literature on VAWG prevention programs, as well as qualitative data collection with key informants in Haiti. Findings from this analysis will contribute to the broader literature on adapting, testing, implementing, and evaluating evidence-based interventions in developing countries. JEL Codes: H76, J16, K14, K42 Keywords: gender, Latin America and the Caribbean, violence, vulnerability * The authors thank the Inter-American Development Bank for its financial support through the Wilson Center. Neither the funding organization nor the Wilson Center has any influence on the findings and conclusions of this report. Gratitude also is extended to Eric L. Olson, Deputy Director of the Wilson Center’s Latin American Program, for his comments, as well as Ulrick Jean Claude, Gladys Mayard, Emanuela Paul, and other key contributors from Haiti and the United States. They graciously volunteered their time to discuss SASA! and Power to Girls. Any errors are solely the responsibility of the authors. TABLE OF CONTENTS Overall Goal.....................................................................................................................................................5 Background......................................................................................................................................................6 Violence against Women and Girls.............................................................................................................6 Violence against Women and Girls in Haiti............................................................................................6 Community-Based Interventions............................................................................................................7 SASA!.....................................................................................................................................................8 SASA! Adaptation in Haiti.......................................................................................................................9 Background........................................................................................................................................9 SASA! in Haiti. ...................................................................................................................................9 Power to Girls. ................................................................................................................................ 10 SASA! and Power to Girls in Haiti. ..................................................................................................11 Methodology................................................................................................................................................. 12 What We Have Learned............................................................................................................................... 13 Adapting Sasa! and Creating Power to Girls........................................................................................... 13 Background. ........................................................................................................................................ 13 Challenges........................................................................................................................................... 13 Promising Practices............................................................................................................................. 14 Implementing the Program. ..................................................................................................................... 14 Background. ........................................................................................................................................ 14 Challenges........................................................................................................................................... 14 Promising Practices............................................................................................................................. 15 Evaluating the Program. .......................................................................................................................... 15 Background. ........................................................................................................................................ 15 Challenges........................................................................................................................................... 18 Promising Practices............................................................................................................................. 20 Recommendations for Adaptation of Community-Based Programs............................................................ 21 Political..................................................................................................................................................... 21 Programmatic........................................................................................................................................... 21 Evaluation. ............................................................................................................................................... 21 References. .................................................................................................................................................. 23 Appendix A. SASA! and Power to Girls: Theory of Change and Logic Mode . ........................................... 25 Appendix B. Key Informant Questionnaires................................................................................................. 28 Abbreviations AIDS acquired immune deficiency syndrome CDCP Center for Disease Control and Prevention GoH Government of Haiti GWI Global Women’s Institute HIV human immunodeficiency virus IFOS Institut de Formation du Sud SDE standard deviational ellipses VAW violence against women VAWG violence against women and girls WHO World Health Organization  OVERALL GOAL T he purpose of this report is to document the This report is divided into four sections. Follow- lessons learned from a review of the planning, ing the background, the second section describes the implementation, and evaluation of the SASA! methodology used to develop this report. The third sec- program by Raising Voices and the Beyond Borders’ tion presents the results of the key discussions held program, Power to Girls, both of which took place in from October 2016 to August 2017 during the planning, Haiti. Findings from this analysis will contribute to the implementing, and evaluation processes. The final sec- broader literature on adapting, testing, implementing, tion presents a series of recommendations based on and evaluating evidence-based interventions in devel- lessons learned for future adaptation and scale up of oping countries. community-based interventions in various settings. 5 BACKGROUND Violence against Women and Girls 2013). This figure, however, underestimates the total economic burden of VAWG, as it does not include the Violence against women and girls (VAWG) is a world- economic burden of VAWG for young girls or the direct wide complex and pervasive issue that has adverse and indirect costs of caregiving, premature deaths, economic, social, and health outcomes for women and disability associated with VAWG. and their families (Duvvury et al., 2013; Ellsberg et al., 2008; Garcia-Moreno et al., 2006). Globally, an Violence against Women and Girls in Haiti estimated 35 percent of women experience physical The Government of Haiti has taken a comprehen- and/or sexual violence at the hands of an intimate sive step toward supporting prevention of VAWG in partner or nonpartner during their lifetime (Contreras detailing a National Action Plan for Gender Equality et al., 2016; Ellsberg et al., 2015; WHO, 2013: 29). for 2014−20. The plan engages various government As a result of this violence, women and girls suffer a ministries and recommends legal reforms, educational number of negative physical and mental health out- system and curriculum changes, and efforts to equalize comes, including death (Duvvury et al., 2013; WHO, rights between women and men more broadly, as well 2013). Families and communities affected by vio- as improving services specific to preventing VAWG lence are deprived of the many benefits associated (GoH, 2014). In addition, the government launched a with healthy family relationships, such as increased National Action Plan in October 2017 on VAWG, valid self-confidence, family unity, and trust, improved for 2017−27. The legal framework on VAWG, however, academic performance among children, increased has many critical gaps, despite years of effort by the participation in community life, community progress women’s movement; these include a law on domes- and development, healthy conflict resolution, and in- tic violence and a sexual violence law that includes a creased community security, among others (Michau, clear definition and specifies the inclusion of sexual undated). Children who witness domestic violence violence in marriage. have been found to experience increased risk of Despite governmental and nongovernmental ef- emotional and behavioral problems, decreased child forts, comprehensive and concrete actions to pre- survival rates, low birth weight, and increased vulner- vent VAWG have been scarce. VAWG remains high- ability to being either victims or perpetrators of vio- ly prevalent in Haiti and constitutes a serious public lence as adults (Garcia Moreno et al., 2003). health problem. According to the 2012 Haiti Mortality, VAWG severely limits a survivor’s potential con- Morbidity, and Service Utilization Survey, 29 percent tribution to society. For this reason, on a macro lev- of ever-married women have experienced violence el, the consequences include the economic costs of perpetrated by their most recent husband or partner lost productivity and an overlooked workforce, strain (GoH, IHE, and ICF, 2013). This rate was highest on response systems and services to treat survivors, among women ages 15−19 (42 percent). Girls also and decreased progress toward the Sustainable De- experience high rates of nonpartner violence. The Vi- velopment Goals. VAWG imposes an economic bur- olence against Children Survey, conducted in 2012 den on some countries of up to approximately 4 per- by the United States Centers for Disease Control and cent of their gross domestic product (Duvvury et al., Prevention on behalf of the Haitian Government, found 6 that more than 25 percent of female children and more eral population, most alarmingly among duty bear- than 20 percent of male children experienced sexual ers and thought leaders. Changing these norms and abuse before the age of 18 (CDCP, 2014). curbing the cycle of VAWG in Haiti is an important The first sexual experience of girls is often forced step toward ensuring healthier, more productive, or coerced, as indicated by a growing number of stud- and safer communities in Haiti. ies (WHO, 2014: 152). For example, nearly one in Community-Based Interventions three girls in the Caribbean (Halcon et al., 2003) and Interventions to prevent and respond to VAWG use one in five girls in Haiti (Bott et al., 2012; 74−75) re- multiple approaches to reduce violence. Several re- ported their first sexual experience was forced or co- cent systematic reviews of the effectiveness of inter- erced. In addition, Haitian girls age 15−24 are twice ventions designed to reduce VAWG in low- and mid- as likely as Haitian boys their age to contract HIV dle-income countries have improved understanding (USAID, 2010), while HIV rates are five times higher in this area (Contreras et al., 2016; Fulu et al., 2015; in girls than in boys aged 15−19 in Trinidad and To- Fulu and Heise, 2015; Arango et al., 2014; Ellsberg bago (UNAIDS, 2014). These figures, however, likely et al., 2015). These reviews have found that most in- underestimate the prevalence of domestic violence terventions to prevent VAWG have occurred in de- due to underreporting on a sensitive issue such as veloped countries such as the United States and the experiences of violence. United Kingdom. However, recent reviews suggest VAWG is a driver and a consequence of HIV (AM- that programs involving community mobilization are FAR, 2005; UNAIDS, UNFPA, and UNIFEM, 2004), among the most promising and show significant evi- with catastrophic implications for the entire country. dence of reducing rates of IPV in low- and middle-in- Haiti’s HIV rate is the highest in the Caribbean, with come countries (Contreras et al., 2016; Ellsberg et nearly 140,000 infected, 56.4 percent of whom are al., 2015). women over 15 years of age (UNAIDS, 2014). One Community mobilization interventions are suc- USAID study indicates that young Haitian women be- cessful in transforming harmful gender norms because tween the ages of 15 and 24 have twice the risk of con- they guide community members at all levels of soci- tracting HIV as their male counterparts (USAID, 2010). ety through gradual and sustainable change. This ap- Compounding the well-documented biological risks proach aims to reduce violence of a population through that link them, VAWG and HIV are also linked through the use of complex interventions that engage multiple gender inequality or power imbalances between wom- stakeholders at all levels (e.g., community men, wom- en and men—also known globally as the root cause of en, youth, religious leaders, police, teachers, political VAWG (AMFAR, 2005). leaders). Community mobilization interventions incor- Social norms promoting power imbalance and porate many strategies, from group training to public condoning VAWG are widespread in Haiti. A 2012 events and advocacy campaigns. They also have been survey of youth in Haiti revealed that nearly half found to be effective at changing attitudes toward pow- of all girls and two out of five boys aged 13−17 er imbalances among boys and girls, thereby prevent- years old believed that a man is justified in beating ing physical, sexual, and emotional violence against a woman for one or more reasons (CDCP, 2014). girls (Lundgren and Amin, 2015). However, due to Similarly, surveys conducted by Beyond Borders in their complex design, very few rigorous assessments and around the southeastern town of Jacmel found of community mobilization programs have been com- that such attitudes were common among the gen- pleted (Ellsberg et al., 2015). A Community-Based Intervention to Prevent Violence Against Women And Girls in Haiti: Lessons Learned 7  Globally, there also is a gap in prevention meth- Based on the stages of change theory scaled up to odologies designed to address the links between a community level, the methodology guides the com- violence against women and violence against chil- munity through a four-phase process of change— dren, and specifically addressing prevention of vio- Start, Awareness, Support, and Action—supporting lence against girls. Traditionally interventions to pre- stakeholders from all levels of the community who vent violence against women and violence against participate in each step of the intervention. children have been implemented separately. Recent I. Start (6−8 months)1: Staff and a community net- reviews, however, have found that these two types work change knowledge and attitudes about VAW of violence (i) have overlapping risk factors, such as and HIV/AIDS as interconnected issues, and foster poverty and community, social, and cultural norms “power within” to address them. that support power imbalances among different gen- II. Awareness (14−18 months): The entire com- erations of men and women; (ii) co-occur in many munity changes knowledge and attitudes, raising families; (iii) lead to intergenerational effects; (iv) awareness about community acceptance of men’s have compounding consequences across the life use of “power over” women. span; (v) are influenced by social norms that support both forms of violence and discourage help seek- III. Support (6−8 months): The community builds ing; and (vi) intersect during adolescence, which is skills to support activists directly involved in these a time of heightened vulnerability to certain types of interconnected issues by joining the movement. violence (Guedes et al., 2016). Authors specifical- IV. Action (6 months+): The community takes action ly noted “the need to harmonize conceptual frame- and changes their behavior, using their “power” to works and instruments used to measure violence prevent VAW and HIV/AIDS. against adolescent girls.” As such, this situation of- SASA! subscribes to the theory of change in fers an opportunity for better coordination and cohe- the social norms approach to preventing VAWG, sion between these types of interventions (Bacchus holding that change in community level knowledge, et al., 2017). To date, however, there has been no attitudes, and skills—and creating social pressure evidence-based methodology on a global scale that toward balanced power and away from violence— engages girls and communities together to prevent successfully changes community behavior and re- violence against girls. duces violence. A rigorous evaluation of the impact SASA! of the SASA! approach on participating communities One of the best-known models to address social in Kampala was conducted by the London School of change and reduction of violence against women is Hygiene & Tropical Medicine. Results from the eval- SASA! The SASA! methodology was developed by uation demonstrate the methodology’s effectiveness Raising Voices, a nongovernment organization based in preventing VAWG and risk behaviors related to in Kampala, Uganda, to prevent VAW and HIV by HIV transmission (Abramsky et al., 2014). In terms of addressing the balance of power in intimate partner IPV experience among participating women, SASA! relationships and broader community dynamics. resulted in a 52 percent reduction in risk of physical 1 Estimates are for implementation time only; additional time may be needed for adaptation of materials. Typically, organizations can imple- ment SASA! in between three and five years. 8 Background violence and a 33 percent reduction in risk of sexual SASA! in Haiti violence.2 These results demonstrate that SASA! was SASA! was adapted to the cultural and language con- effective in changing underlying attitudes and norms, text of Haiti by Beyond Borders between 2010 and and it shows promise in reducing levels of IPV in the 2015, making the range of materials available to Hai- Ugandan context (Abramsky, et al., 2014; Kyegombe tian organizations and communities in Haitian Creole. et al., 2014). Importantly, it shows that prevention at While adapting the methodology, Beyond Borders a community level is possible within programmatic simultaneously implemented it in five communities timeframes; typically, SASA! implementation takes in the southeast of Haiti, integrating feedback on the between three and five years. adaptation throughout the process of implementation. While budgets required to operate SASA! depend The communities were selected based on the exist- on context and organizational style, a recent cost and ing community relationships and recommendations of cost-effectiveness study of SASA! in Kampala showed the local organization that began its implementation. that the average cost per participant in intervention In 2013, in an agreement with Raising Voices as the communities was US$21 over four years or US$5 per creators of the methodology, Beyond Borders also be- participant annually (Michael-Igbokwe et al., 2016). gan providing technical support to other Haitian orga- More importantly, this study found SASA! to be cost-ef- nizations interested in using the methodology. Today, it fective, with an estimated cost for past year IPV avert- has eight long-term local partners in various stages of ed of $460, which was far less than the cost per IPV SASA! implementation and several others making use averted in the Intervention with Microfinance for AIDS & of materials or short courses on VAWG prevention. Gender Equity (IMAGE) study conducted in South Afri- These partners use the adapted SASA! activist kit, ca (Michael-Igbokwe et al., 2016; Jan et al., 2010). and have access to the long-distance and in-person Since the release of the study results, SASA! training and expertise of the people who participated in has gained popularity. It now is being adapted and the Beyond Borders implementation in the Southeast. implemented in over 20 countries by more than An internal process evaluation conducted by the 60 organizations. organization that collected data throughout the im- plementation of the SASA! Haiti adaptation in the five SASA! Adaptation in Haiti implementation communities in the Southeast showed Background positive changes in community knowledge, attitudes, Past interventions to address VAWG in Haiti have skills, and activist behaviors on a number of indicators. focused mainly on response to survivors without in- An end-line evaluation using a random convenience corporating a prevention focus. Those interventions sample of 600 surveys and focus groups across five adopted a case-by-case treatment and did not em- participating communities found significant improve- phasize on the community as a whole. The SASA! ments in community-level indicators from baseline. In adaptation in Haiti in 2010 is among the first interven- a study of SASA! adaptations globally, Raising Voic- tions in the country to focus on prevention of different es, University of California San Diego, and UN Trust types of VAWG at the community level. Significantly, Fund are preparing a research report detailing more it was also the first full adaptation of SASA! information about the Haitian adaptation and success 2 While a large effect was observed, changes were not statistically significant. Evaluators observed an increase in variations of physical IPV levels among control sites at follow-up compared to baseline. This reduced the power to observe statistical significance when analyzed by cluster. A Community-Based Intervention to Prevent Violence Against Women And Girls in Haiti: Lessons Learned 9 of implementation, to be shared in 2018 (Adapting While the potential benefit of SASA! in Haiti was SASA! for Global Communities: A Case Study from clear, a rigorous impact evaluation was still needed Haiti, forthcoming). to assess the impact of SASA! on Haitian intervention In a qualitative program evaluation process con- communities. In addition, program staff feared the pos- ducted in 2016 using a process called Most Significant itive changes experienced by women were not always Change (Davies and Dart, 2005), local researchers in- experienced by girls in the community. Proponents of terviewed community members and activists who had girl-centered programming often remark that methodol- participated in the community change process. They ogies designed for children or women often do not meet were asked what the most significant change (positive the needs of girls (Population Council, 2010). Adult wom- or negative) produced by SASA! was in their communi- en also hold tremendous power in girls’ lives—a unique ty. These interviews were then transcribed and evalu- dimension that needs to be explored in order to effective- ated by a panel of community members, who voted on ly prevent violence against girls. In addition, programs the stories that, to them, represented the most signifi- aimed at helping children often end up benefiting mainly cant change they had experienced in the program. The boys, due to the greater power that boys have to access following quotes were extracted from two of the stories activities, and speak and act freely. Given the connec- selected by that panel. tions between these types of violence, much of the work “There are many changes in my personal life, overlaps. Specific tools, techniques, and spaces, howev- in the way I used to speak to my wife and chil- er, are needed to address the particular needs of girls in dren too, how [I] treated students at school. order to see maximum benefit in their lives. There were certain kinds of jokes I used to tell Power to Girls that I don’t tell anymore. Whenever I yell at the While implementing the Support phase (the third of four children or hit my children or students, my wife phases) of SASA!, Beyond Borders began receiving says, ‘Be careful Pastor! What you are doing is community feedback that more was needed to specifi- violence!’ Now I can’t even remember when I cally address violence against girls, and to engage youth last hit my children . . . Now my relationship with as change agents to interrupt intergenerational cycles of my wife is going well, even when we don’t have violence. In an internal program evaluation conducted at any money. . . . SASA! has allowed us to make that time, Beyond Borders hired local researchers to con- decisions together, and to better manage our duct 700 surveys (350 female and 350 male) and two fo- household. We even made the decision to stop cus group discussions to ask about indicators related to having children together.” (Male religious leader, girls’ lives. Over 70 percent of adults still felt it was best to Meno, Haiti) teach a girl to submit to her husband when she grew up, “My father was very violent. He swore a lot at my and almost as many (69 percent) felt that girls should not mother and really mistreated us women. But to- be allowed to meet together without the presence of boys day, thanks to SASA! training, he has completely or men, particularly as attitudes about adult women’s em- changed . . . and today I feel happy for the beautiful powerment had changed significantly across several in- change in him . . . In the community, people have dicators and communities (Beyond Borders, 2013). The started to understand they shouldn’t do violence, organization consulted with young women leaders and especially to women, and they should value young mentors associated with the Haiti Adolescent Girls Net- girls and balance power well in their families.” (Fe- work to gather ideas for themes specific to preventing vi- male community leader, Kay Jakmèl, Haiti) olence against girls; they drew from evidence-based and 10 Background promising practices and led a participatory co-creation Global Women’s Institute (GWI) at George Washington process with Haitian girls and communities to develop University—with the support of the NoVo Foundation the Power to Girls methodology. and Inter-American Development Bank—are evaluating Power to Girls is a step-by-step methodology that the impact of the implementation of SASA! and Power to combines girl-centered programming with school- and Girls in two communal sections in the southeast of Haiti. community-wide social norm change. Used alone or There are several reasons justifying the continued together with SASA!, Power to Girls builds on SASA!’s implementation and evaluation SASA! and Power to ‘essential’ elements, including gender-power analysis Girls in Haiti, such as: and a phased process of change, involving the entire • The need for rigorous impact evaluation to inform community (including participants from across the so- replicability, scale-up, and sustainability. cio-ecological model) and activism (Raising Voices, 2017). Power to Girls is designed to increase girls’ • The need to integrate both programs to benefit women and girls because of positive and multi- safety and freedom and prevent violence against girls plicative effects or synergy between the two pro- by combining multiple violence-prevention strategies. grams. This combination may also result in reduc- The logic model for the second implementation of tions in implementation and evaluation costs due SASA! Haiti, which includes the Power to Girls meth- to economies of scale. odology, can be found in Annex A. From 2013 to 2016, Beyond Borders piloted Power to Girls in three com- • Effectiveness, appropriateness, and relevance munities where it was continuing to implement SASA!, of this combined approach to address important including one pilot school to help to develop the school- issues of VAW in order to prevent and reduce as- wide component of the community-based intervention. sociated physical, social, emotional, intergenera- While this pilot was not formally evaluated, community tional, and economic adverse consequences. feedback from existing girls’ groups, community lead- • Degree of adoption and acceptability by commu- ers, school leadership, personnel, and students, as nity members and other key stakeholders, such well as community-based activists involved in SASA!, as women’s organizations. was positive. The pilot allowed for the completion of materials being used in the current evaluation. • The implementation cost per participant is low and the SASA! methodology is cost effective. SASA! and Power to Girls in Haiti In 2016, with the support of NoVo Foundation, Beyond • Additional secondary individual and societal benefits include improvement in girls’ schooling Borders engaged a new cohort of communities in south- retention rates and academic achievement, im- east Haiti, using the adapted and piloted SASA! meth- proved labor market outcomes later in life, equity odology in tandem with Power to Girls. This dual model, in gender, and improvement in women’s and girls’ the Rethinking Power program, expects to reduce social self-confidence. acceptance of gender inequality and VAWG; to decrease experiences and perpetration of VAWG; to increase girls’ • Additional secondary community benefits also sense of safety and freedom to make decisions; and to may include enhanced empowerment and activ- decrease HIV/SRH risk behaviors. To measure the ef- isms, access to justice, and other social programs fectiveness of this dual model, Beyond Borders and the by participating community members. A Community-Based Intervention to Prevent Violence Against Women And Girls in Haiti: Lessons Learned 11 METHODOLOGY For the purposes of the review, VAWG includes any terventions for scale up. To learn more about SASA! physical, sexual, economic, or psychological harm and Power to Girls in Haiti, semi-structured interviews perpetrated on a woman or girl by her husband or and focus group discussions were conducted with key intimate partner. It also encompasses the physical, informants, including research staff from GWI, local sexual, economic, and emotional abuse of girls at the research partner staff (Institut de Formation du Sud hands of family members or strangers. (IFOS)), and program staff from Beyond Borders as The methods used to develop this report consist of the implementers of the methodology, as well as local a review of literature on VAWG prevention programs, community leaders and activists. Data gathered from as well as qualitative data collection with key infor- these activities were used to provide a detailed picture mants in Haiti. The literature review informed the back- of the processes of design, adaptation, implementa- ground of this report, and focuses on the effectiveness tion, and evaluation of the SASA! and Power to Girls of VAWG prevention interventions and adapting in- methodologies in Haiti. 12  WHAT WE HAVE LEARNED Adapting SASA! and Creating Power • Funding: Lack of donor support of VAWG work in Haiti and, in particular, long-term funding is a to Girls challenge for many local organizations wishing Background to take on longer-term social change program- The process of planning for the implementation of ming like SASA! This is despite the fact that this SASA! and Power to Girls methodologies was in- program has been found to be cost-effective in formed by the following factors: similar settings. 1. Lessons learned from the implementation of • Need for high staff capacity: There was a lack SASA! in Uganda. For the adaptation of the first of availability of staff with experience with com- implementation of SASA! in Haiti, Beyond Bor- munity mobilization programs aimed at prevent- ders drew from the experience of its implemen- ing VAWG in Haiti; presence of a former Raising tation in Uganda. Voices staff on Beyond Borders staff allowed for continuous process of support and staff training. 2. Lessons learned from the adaptation, test- ing, implementation, and evaluation of the • Lack of guidance on fidelity and adaptation first cycle of SASA! (implementation of all globally: As it was the first full adaptation of four phases of SASA!) and pilot of Power to SASA!, the team in Uganda did not yet have rec- Girls in Haiti, including: ommendations relating to fidelity or the adaptation process. The team in Haiti had to innovate a pro- • positive outcomes from the process evaluation cess and set of guidelines that worked for them. and its potential impact when adapted for Haiti; • Adaptation into Haitian Creole. As a language • community satisfaction with the methodology and with various spellings and with a relatively re- their interest in expanding its activities, demon- cent standardization of writing style, undertaking strating its relevance and potential sustainability; adaptation meant extensive consultation to en- • feedback from the first implementation on the sure readability by the desired population. methodology’s gaps when it came to the lives of Haitian girls, leading to the conception and • Handover of activities after SASA! SASA! works through relationship building between development of the Power to Girls methodol- people in a community to influence each oth- ogy; and er’s social norms. The ‘handover’ to the com- • understanding of requirements for effective munity at the end of SASA! can easily be seen implementation, including staff capacity, re- as an abandonment, and a removal of resourc- sources, and logistics. es. Positive reframing and close community ca- pacity building are needed in order to make the Challenges handover successful to communities of contin- The process faced several challenges, including ued activism around VAWG. the following: 13 Promising Practices Implementing the Program The promising practices that can be identified are: Background • Expanding an innovative intervention. As Implementation of SASA! and Power to Girls in tandem mentioned above, the adaptation of SASA! in through the Rethinking Power program in the commune Haiti was the first of all global SASA! adap- of La Vallée-de-Jacmel, Haiti, began in early 2017. Im- tations, and––particularly with a lack of guid- plementation was delayed by several months due to ance on adaptation and methodology fideli- several factors related to the research study, including ty––managed to achieve a high level of fidelity the length of time it took to obtain adequate resourc- to the original model. Due to the program’s es, identify and hire the right partners, and establish the demonstrated successes in the first imple- planning and baseline data collection for the evaluation. mentation, donors and local institutions alike As both methodologies are community mobilization were interested in adapting and expanding the approaches, the implementation and evaluation areas methodology in Haiti through a second imple- include two communal sections of La Vallée-de-Jacmel mentation and technical support. (Mizak and Tenyé). The methodologies aim to change • Integrating lessons learned on implementa- the knowledge, attitudes, skills, and behaviors of the tion. The wealth of knowledge gained through entire population, although they specifically and directly the implementation of SASA! in Uganda and engage a particular community network comprised of 16 the first phase of the program in Haiti great- girls’ groups run by mentors, 46 (23 female and 23 male) ly informed its smooth implementation during community activists, three schools, a drama troupe, and scale up. a set of religious and local leaders, journalists, and health workers. This network will be trained and mentored • Utilizing a participatory approach. Beyond Borders worked in collaboration with the com- to engage others in their own circles with participatory munity and designed the intervention based activities held regularly and in the course of everyday on their identified needs. Such collaboration community life. The implementation team includes two led to quality adaptation of SASA! as well as Local Activism staff fully present in community activities, the birth of the Power to Girls methodology, a Coordinator, 16 mentors, and two part-time technical which was developed and piloted, based on support staff as well as general support staff. community feedback. This approach demon- Challenges strates good governance and accountability The program team experienced several challenges in within the program structure. implementation, as follows: • Implementing an empowerment model for girls. The Power to Girls model is the first • Navigating the environment. The geographical environment in the implementation communities model focusing on the safety and freedom of is very rural, and it is difficult to reach many of girls to be implemented in Haiti. Integrated into the communities. the SASA! framework, the Power to Girls mod- el uses a comprehensive, community-based • Relying on unpaid activism without strong, approach that engages all stakeholders in the established relationships. Haitian communi- empowerment of girls. ties are in need of livelihoods and financial sup- 14 What We Have Learned  port. However, voluntarism is a significant com- through benefits-based and engaging dialogues ponent of SASA! and Power to Girls. It can be about power and violence. The methodologies a challenge to engage activists and community allow for participation of all sectors with minimal members without being able to promise them fi- community resistance. nancial compensation, in particular when new to • Supporting feelings of ownership and capac- a set of communities and relationships of trust ity building of community network. The adop- are still being built. tion of program activities by the community is • School and teacher availability/ institutional integral for sustainability of SASA! and Power to commitment. It proved very challenging to con- Girls. To avoid the handover of VAWG prevention vene teachers for activities for several reasons, activism being perceived as an abandonment or including exam schedules, heavy rainfall, and the removal of resources at the end of the SASA! the volunteer nature of the program. Teachers and Power to Girls implementation, program staff had participated in previous programs where or- will have to put significant effort into engaging the ganizations had provided financial incentives for community in all aspects of planning and imple- their participation. mentation, facilitate feelings of ownership within the community, and systematically build capaci- Promising Practices ty to prepare them to lead. Implementation staff Several promising practices were identified during the create specific mentoring plans for each phase of implementation of the program: the methodologies toward this end. • Facilitating the high capacity of staff for qual- ity implementation. Because they were heavily Evaluating the Program involved in the adaptation and creation phases, Background the implementation staff had a very high capacity GWI designed and implemented the evaluation of for the effective implementation of the two method- SASA! in partnership with Beyond Borders. Within the ologies and their core activities. first year of the program, GWI designed the evaluation • Establishing good relationships with partners. in collaboration with local partners, conducted quanti- Relationships with program partners were main- tative and qualitative baseline data collection, and be- tained after the first implementation of SASA!, gan preliminary analysis of results. and those partners were included in the second The demand for a rigorous evaluation arose from implementation of SASA!’s scale up and the first community and donor interest. Initial quantitative and implementation of Power to Girls. qualitative findings of the first implementation of the • Utilizing a participatory approach. Program SASA! adaptation in Haiti have been showing promising staff continued working in collaboration with the results and also demonstrate a community interest in ex- community and civil society, which enabled the panding the program within the region. The results of the smooth implementation of the scaled-up program. process evaluation reveal the potential impact of the pro- gram, and conducting an impact evaluation would allow • Engagement of women, men, girls, and boys. stakeholders to attribute the program to a change in out- Both methodologies successfully engage female comes. In addition, the impact evaluation would measure and male community leaders and members A Community-Based Intervention to Prevent Violence Against Women And Girls in Haiti: Lessons Learned 15 the effectiveness of the new Power to Girls component, change in knowledge, attitudes, skills, and behavior were implemented in tandem with the scale up of SASA! pro- also included. Specific indicators measuring changes in gramming. With the support of donors and community knowledge, attitudes, skills, and behaviors among key members alike, Beyond Borders has decided to conduct stakeholders of the Power to Girls component were in- a rigorous impact evaluation in collaboration with GWI. cluded to determine the success of the Rethinking Power GWI developed the evaluation design in collabo- methodology. Each indicator will be assessed at base- ration with Beyond Borders and other local stakehold- line, mid-term, and end-line to measure change. ers. A theory of change for the dual program was de- Based on these indicators and outcomes, quanti- veloped based on the original SASA! framework and tative and qualitative data collection tools were devel- integrated with the Power to Girls methodology. The oped. Data collection tools were designed in English theory of change for the combined program can be as well as in Creole. The quantitative tools include a found in Annex A. household survey for women, a household survey for Utilizing this theory of change and corresponding log- men, an individual girls’ club survey for adolescent girls, ic model, the research team developed a rigorous evalu- and a school-based self-administrated survey with fe- ation design for a community-based program. After con- male and male students. All youth were between 10 and ducting several field visits and discussing the feasibility 19 years old. A summary of the number of people sur- of various evaluation designs with the implementing part- veyed per commune during the baseline is provided in ner, it was determined that randomizing the intervention Table 1. Qualitative tools were developed to conduct fo- setting was not possible for programmatic reasons (e.g., cus groups with male and female, adult and adolescent the site selections, Mizak and Tenyé, had to be done pur- community members including parents, teachers, com- posively). Therefore, the most rigorous evaluation design munity leaders, students, and adolescent girls, and indi- for the specific context was a quasi-experimental design vidual in-depth interviews with female and male stake- that includes the intervention and comparison communi- holders from the public, private, and the third sector. ties. The comparison group was selected from a different The focus group tools include participatory techniques department in Haiti, based on an analysis of sociodemo- such as free-listing, incomplete stories, and Venn Dia- graphic factors and specific selection criteria. Selection grams. A total of 12 and 10 foci group were respectively criteria included being sufficiently far geographically from conducted in La Vallée-de-Jacmel and Marigot. the intervention group that contamination would be un- TABLE 1. NUMBER OF RESPONDENTS SURVEYED likely. This ensured that the intervention community, Jac- LA VALLÉE- mel, and the comparison community, Marigot, had simi- COMMUNE MARIGOT DE-JACMEL lar characteristics at baseline while being geographically Women 819 1,158 distant enough to avoid contamination. Men 317 547 Indicators to measure impact were chosen based on the theory of change and desired outcomes. The indica- Schools 363 358 tors included outcomes at the societal and community Girls’ club 267 _ levels to measure impact at multiple levels of society. At the highest level, the project and evaluation team identi- This mixed-methods approach allows for the triangula- fied a number of key indicators to measure the project’s tion of data and the collection of in-depth information overall impact. To contribute to the overall goal and impact that can be used to inform the program and analysis of level indicators, indicators measuring community-level 16 What We Have Learned the program’s impact. The research protocol and tools the underlying roots of gender inequality. The evalua- were approved by the George Washington University tion focuses on gender inequalities and the systemic Institutional Review Boards and Haiti’s Ministry of Pub- and structural nature of these inequalities. It also en- lic Health and Population. gages women and girls in the participating communi- Conducting research on VAWG is similar to con- ties throughout the evaluation process. The results of ducting research in other sensitive topics; however, the evaluation will not only build the knowledge base there are aspects of this type of research that are unique on the adaptation and scale up of community-based because of the traumatic nature of the subject matter. In prevention programs, it also will inform policies and the past 20 years, a wide variety of international experts funding around preventing VAWG in Haiti and other and researchers have documented a number of best low-resource settings. practices and ethical guidelines for conducting research Baseline data collection occurred from March to on VAWG. In this baseline, we followed the ethical rec- July 2017. It included four components: a household ommendations established by the World Health Orga- representative population-based survey with women, nization to ensure that data was collected in an ethical a household representative population-based survey way and, at the same time, ensure the quality of the with men, a school-based self-administrated survey information. For example, particular attention was given with female and male students, and a qualitative com- to safeguard privacy and confidentiality, availability of ponent that included focus groups and in-depth inter- support services for survivors of violence, and the use views with stakeholders, members of the community, of empathetic and nonjudgmental interviewers who re- and key actors. Fieldworkers received an intensive and ceived intensive training for several days. comprehensive training that included key concepts on In addition, the evaluation applies a participatory gender and VAWG, interactive lessons that involved approach that involves beneficiaries—women and girls significant practice with delivering the research tools, in the intervention communities, in particular—in the and instruction on important methodological and ethi- design and implementation of the research. IFOS, the cal considerations for conducting research on this top- local research partner, was hired to support the qualita- ic. This training helped to build the capacities of local tive and quantitative data collection. Throughout the de- researchers for conducting fieldwork on VAWG. velopment of the evaluation design and data collection Preliminary findings of the baseline (from the tools, GWI, Beyond Borders, and IFOS worked closely women’s household survey) underline that patriarchal together not only to ensure that the design was relevant gender norms and attitudes prevail in both communes. in the local context, but also to strengthen local capaci- Regarding the gender roles in the household, more ty. Local stakeholders, community-based organizations, than 90 percent of women believe that household and local authorities were involved in the planning and chores and childcare are women’s main responsibili- implementation stages through the creation of a techni- ty. Although more than 80 percent appear to support cal advisory group, an independent advising body con- equal authority with their partners in the family, three sisting of experts with knowledge of and experience in out of five women believe that the male partner should VAWG and evaluation of interventions on VAWG. This have the final decision in the household. This dynamic collaborative process will also be utilized in the interpre- is also reflected in the way girls and boys are treat- tation and dissemination of findings ed. For instance, four out five women agree that girls The evaluation also applies a gender approach to should not be allowed to socialize outside the home, achieve the ultimate goal of the program: to transform and half of the women supports that girls should not A Community-Based Intervention to Prevent Violence Against Women And Girls in Haiti: Lessons Learned 17 be able to make the decision on when to get married. ficiently implement the program, while simultane- VAWG is prevalent in both communes, and atti- ously minimizing contamination between control tudes on VAWG reveal a lack of awareness of women’s and intervention sites. Through sociodemographic rights. Four out of five women have suffered controlling and economic data collected at the commune lev- behavior from their partners, and half of the women jus- el, as well as information gathered during several tify physical intimate partner violence (IPV). One-third site visits, the comparison communities were se- of the women agree that IPV is a private matter and 20 lected from a different commune in the same prov- percent believe that IPV is the price to pay to keep their ince as the intervention communities, thereby min- family safe. One-fifth of the women believe that rape imizing the potential for contamination while also is related to the survivors’ behavior. Furthermore, more establishing equivalence at baseline. than half of adolescent girls do not feel safe walking in • Accommodating for limited data and modify- the community. These initial findings corroborate the im- ing the sampling frame. Limited and/or outdat- portance and the need to implement programming that ed official data complicated selection of the con- promotes the transformation of gender norms that can trol site, the development of adequate maps, and lead to a reduction of the different types of VAWG. decisions regarding the sampling frame. Lack Challenges of access to an updated map made it difficult to The research team experienced several challenges determine accurate geographic borders and dis- while designing, planning, and implementing the evalu- tribution of the population, inhibiting the creation ation and baseline data collection. These challenges can of the sampling frame. The research team col- be divided into two phases: design and data collection. lected GPS coordinates in the field to identify the borders among the standard deviational ellipses Design Challenges (SDEs) and create a map. While the research • Attributing results to the program without team was collecting GPS coordinates, it found a bias. Developing a rigorous evaluation design of discrepancy between the official and the real bor- a community-based program that can attribute ders of some communal sections, which led to a outcomes to the program while minimizing bias is revision of the sample selection. difficult in any setting, and was made particularly difficult in this geographic context. Rigorous evalu- • Hiring a local research partner. Finding a lo- ation designs are complex and require a compar- cal research partner with the capacity to conduct ison group. Because the program is a multilevel data collection for a complex study on a sensitive community-based program (including program im- topic was extremely difficult in Haiti. It took sever- plementation through media campaigns, schools, al months of advertising the project and meeting churches, and other service points that draw from with different consultants to identify an appropri- multiple communities), there was a strong poten- ate local research partner. Additional time was tial for contamination of control communities in also required to establish and finalize a contract. a randomized controlled trial design. However, • Establishing a comparison group and delayed especially with a quasi-experimental design, lo- intervention. In a community setting, delivering a gistical and resource limitations complicated the beneficial program to only one group involved in selection of a similar comparison group within a an intervention can raise ethical questions and sufficiently wide area that would allow them to ef- 18 What We Have Learned  heighten tensions at the political and community er sources such as the Haitian Demographic levels, particularly when delayed implementation Health Survey (EMMUS). These higher rates of the program is planned for the comparison may be a result of the rigorous methodological group after end-line data collection takes place. and ethical approaches applied during evalu- The research and implementing partners had to ation, including extensive fieldworker training carefully explain the evaluation design to stake- and piloting, as well as the efforts to reduce holders and community leaders to gain buy-in, risks to confidentiality and privacy. and still received questions from members of the • Ensuring confidentiality and privacy. Due comparison group during data collection. Such a to weather conditions, many people stayed in design requires additional ethical considerations their home and it was difficult for the enumera- when being implemented in locations without ad- tors and respondents to locate a private place equate support services. to conduct the interview. In addition, there were Data Collection Challenges some cases where the enumerators knew the respondents, and some respondents shared • Overcoming logistical and environmental limitations. During data collection, it was more the objectives of the study with their neighbors, difficult than expected for the fieldwork team to compromising confidentiality in the communi- access the households in the mountains. This, ties. The research team worked closely with paired with the high percentage of vacant houses the program and field staff to mitigate any risks in the selected SDEs, led to a decrease in the to participants and to establish strategies for sample size. The research team had to select overcoming these challenges. These issues additional SDEs, namely in the urban areas, to were discussed during training and debriefing achieve the appropriate sample size. sessions, and enumerators were instructed not to conduct an interview unless privacy could be • Managing costs. Implementing a rigorous mixed- guaranteed. When these circumstances arose, methods impact evaluation can be very costly, and enumerators postponed the interview until an the program location is expensive for operational alternative private location or enumerator was planning and conducting fieldwork. GWI and Be- available. Because of these additional proce- yond Borders had to work jointly to raise the finan- dures, however, fieldwork took longer than ex- cial resources for the program and the evaluation pected to complete. alike, which took a significant period of time. The project was implemented with the financial support • Training male enumerators. Training male enu- of NoVo Foundation and the Inter-American De- merators to deliver the survey to other men in a velopment Bank for the baseline and mid-line data sensitive but neutral manner was difficult, as many collection. Funds must be obtained for the end-line of them were not sensitive to the topic of VAWG. data collection. The enumerators participated in interactive train- ing activities and role-playing sessions to sensitize • Supporting disclosure of physical IPV. So- them to the experiences of women and girls, as cial norms and stigma surrounding physical well as to familiarize them with the questionnaires. IPV may have led to underreporting of violence. Special care was taken to discuss these complex Prevalence rates from the baseline, however, concepts and to foster an open environment in are slightly higher than IPV rates from oth- A Community-Based Intervention to Prevent Violence Against Women And Girls in Haiti: Lessons Learned 19  which the male enumerators could expand their IFOS and Beyond Borders teams contributed knowledge of VAWG and practice delivering the to the planning and realization of the translat- interviews in an appropriate and supportive way. ed tools and protocol, as well as the hiring and training of fieldworkers and the conducting of Promising Practices the baseline research. The establishment of the Throughout the design, planning, and implementation technical advisory group and annual meetings of the evaluation, the research and program teams with this group’s members enabled the research identified several best practices that guaranteed the team to obtain technical input on the evaluation successful collection of data. process and programmatic insight, and it also • Evaluating a well-assessed program. The suc- will allow them to contribute to dissemination cessful adaptation of the first phase of SASA! in and uptake strategies throughout the lifecycle of Haiti piqued community and donor interest in ex- the program. panding the program. For this reason, the program • Ensuring researcher-implementer coopera- and research teams decided to undertake an im- tion. Clear and transparent sharing of informa- pact evaluation to measure the effectiveness of tion between GWI, Beyond Borders, and local the second phase of SASA! and Power to Girls research and implementing partners in Haiti es- and to monitor scale up of a community-based in- tablished a collaborative relationship from proj- tervention. Because of the previous evidence on ect inception. The intervention goals and objec- the success of the program and its potential ef- tives, as well as the evaluation methodology, fectiveness when adapted to Haiti, undertaking a were clearly delineated at the beginning. This rigorous evaluation was feasible and appropriate. enabled the local research and evaluation part- • Incorporating a mixed-methods design. The ners in Haiti and other stakeholders to feel that evaluation design employs a mixed-methods ap- that they were an integral part of the process proach that utilizes quantitative and qualitative from the start, as they were able to make sug- methodologies to allow for the triangulation of gestions to ensure program adaptability and fit findings by gathering data using multiple forms of were compatible with the Haitian context while enquiry, giving more depth and certainty to con- also feeling comfortable about the process. clusions made from the data. It also provides op- • Employing ethical research practices. The portunities to collect information on and explore research team, in collaboration with Beyond complex issues that are not easy to quantify. Sup- Borders and IFOS, made significant efforts to porting rigorously collected quantitative data with ensure that the highest ethical standards in the in-depth qualitative findings provides additional field were maintained throughout the design and insights that will increase understanding about implementation of the evaluation and baseline the ways in which impact was achieved and the data collection. Efforts included, among others, factors and conditions that influence them. the selection and training of local enumerators; • Utilizing a participatory approach. GWI en- the establishment of a comparison group; and gaged program partners and other key stakehold- enforced protocols for guaranteeing privacy, ers in all stages of the evaluation design and base- confidentiality, and anonymity while ensuring line data collection. In a consultative manner, the safety and avoiding re-traumatization. 20 What We Have Learned  RECOMMENDATIONS FOR ADAPTATION OF COMMUNITY-BASED PROGRAMS Political • Facilitate a collaborative, participatory pro- cess. Valuable inputs about social norms, cultur- • Collaborate with donors who have a long- al values, practices, and realities of the Haitian term vision and a commitment to prevention communities were gathered during the first im- programs. Maintain an open and communi- plementation of SASA! in Haiti. Both phases of cative process with donors to ensure that they adaptation of SASA! and the creation and pilot have reasonable expectations regarding the of Power to Girls systematically used a collab- timeframe required to design and implement a orative approach by incorporating community community-based prevention program, as well leaders’ and stakeholders’ inputs into the design, as the outcomes that can be expected in the implementation, and evaluation of the two meth- short and long term. odologies. This approach has enhanced and • Provide guidance on keeping political am- maximized potential for effectiveness in the new bitions distinct from activism. While political communities where they are being implemented. office is one avenue for change, it is important to have guidance for staff on keeping political • Maintain fidelity while contextualizing the program. Adaptations of effective interventions ambitions distinct from their roles, to avoid dam- achieve a balance between maintaining the es- aging the reputation of the program. sential characteristics of the original intervention • Ensure the necessary resources. To implement and cultural relevance to a different setting. The a prevention program, and also to evaluate such a most successful adaptions encourage participa- program, it is necessary to have resources in place tion across multiple groups and sectors of society. to support these components, including referral They also allow sufficient time and resources to ef- systems to ensure basic services for survivors. fectively adapt and implement an intervention, re- • Engage with local and national actors. The im- sulting in sustained capacity building and the cre- plementing partner should facilitate a participatory ation of networks that are critical to reducing IPV. process with the national and local governments • Utilize complimentary programs, where pos- and private sector to gain political and financial sible. The SASA! and Power to Girls methodolo- support for the program and to incorporate those gies are designed to be complementary and can actors into the program as valuable stakeholders. work in tandem if an organization has the skills and resources over the long term. This dual pro- Programmatic gram will address the needs of women and girls • Focus the program on the needs of the com- in Haitian communities. munity. Intervention activities should be adapt- • Invest in staff capacity. The experience of the ed, developed, and evolved in direct response to implementing partner and other stakeholders in community priorities, needs, and challenges. For implementing SASA! activities in Haiti was a huge the program to be sustainable, the local commu- asset to the scale up of the program and the de- nities have to embrace it, accept it, and own it. 21  velopment of the Power to Girls model. Providing pact and effectiveness have already been mea- comprehensive training to staff and leading a pro- sured through process evaluations and less rig- cess over which they have ownership can allow for orous evaluations. a more effective expansion of program activities. • Utilize a participatory and gender approach. • Ensure referral services are in place. While a Monitoring and evaluation processes should be program such as SASA! focuses on prevention of approached with the beneficiaries in mind. To VAWG, many community members already may do so, stakeholders—and women and girls, in have experienced violence and may be in need particular—must be engaged in all aspects of of services and other support. Response ser- the evaluation design, implementation, analy- vices, including health and legal support, should sis, and dissemination. Maintaining a participa- be in place before beginning program activities, tory and gender lens will ensure that the eval- and paths of referral from program staff and ac- uation results are useful and applicable to the tivists should be developed to easily connect local population. participants to necessary support. If adequate re- • Assess the pros and cons of evaluation de- sponse programs are not already in place, the im- signs. Not all projects can be evaluated using plementing partner should provide such services randomized controlled trials. The evaluation de- on a temporary basis. sign should be flexible and based on the specific context and environment where the intervention Evaluation is taking place. Programmatic implications need • Conduct an impact evaluation at the appropriate to be taken into account when deciding on the point in the program cycle. Impact evaluations evaluation design. are important for this type of community-based • Foster a close collaboration between re- intervention, as they inform future local policy search and programmatic partners. Establish- and national and international funding, in ad- ing a collaborative and communicative relation- dition to providing a opportunity to learn more ship between the research and programmatic about the feasibility of an approach in different partners from the outset of the project allows for contexts. It also is important, however, to con- a collaborative learning process and a successful duct an impact evaluation at the appropriate adaptation and evaluation. point in the program cycle, when potential im- 22 What We Have Learned  REFERENCES Abramsky, T. et al., 2014. Findings from the SASA! Study: A Clus- Davies, R. and J. Dart. 2005. The “Most Significant Change” ter Randomised Controlled Trial to Assess the Impact of a (MSC) Technique: A Guide to Its Use. Available at www. Community Mobilisation Intervention to Prevent Violence researchgate.net/publication/275409002_The_’Most_Sig- against Women and Reduce HIV Risk in Kampala, Uganda. nificant_Change’_MSC_Technique_A_Guide_to_Its_Use. BMC Medicine, 122(12). Available at https://bmcmedicine. Duvvury, N., A. Callan, P. Carney, and S. Raghavendra. 2013. biomedcentral., com/articles/10.1186/s12916-014-0122-5. Intimate Partner Violence: Economic Costs and Implica- AMFAR. 2005. Gender-Based Violence and HIV among Wom- tions for Growth and Development. Women’s Voice, Agen- en: Assessing the Evidence. cy, & Participation Research Series 2013, No.3. 82532. Arango et al., 2014. Interventions to Prevent or Reduce Violence Washington, DC: World Bank. Available at http://docu- against Women and Girls: A Systematic Review of Reviews. ments.worldbank.org/curated/en/412091468337843649/ Washington, DC: World Bank. Available at http://docu- pdf/825320WP0Intim00Box379862B00PUBLIC0.pdf. ments.worldbank.org/curated/en/700731468149970518/ Ellsberg, M. et al., 2015. Prevention of Violence against Wom- Interventions-to-prevent-or-reduce-violence-against-wom- en and Girls: What Does the Evidence Say? The Lan- en-and-girls-a-systematic-review-of-reviews cet, 385(9977): 1555−66. Available at https://www.ncbi. Bacchus, L. J. et al., 2017. Exploring Opportunities for Coordinat- nlm.nih.gov/pubmed/25467575. ed Responses to Intimate Partner Violence and Child Mal- Ellsberg, M., C. Bradley, A. Egan, and A. Haddad. 2008. Violence treatment in Low and Middle Income Countries: A Scoping against Women in Melanesia and East Timor: Building on Review. Psychology, Health & Medicine, 22(sup1): 135−65. Global and Regional Promising Approaches. Office of De- Beyond Borders. 2013. Program Evaluation Survey Results velopment Effectiveness, AusAID. Canberra: Government from 600 Surveys (300 Male, 300 Female) on Violence of Australia. Available at http://dfat.gov.au/aid/how-we-mea- against Girls in 5 Communities in the Southeast of Haiti. sure-performance/ode/Documents/violence-against-wom- Unpublished data. en-melanesia-east-timor-building-on-global-and-region- Bott, S., A. Guedes, M. Goodwin, J. A. Mendoza. 2012. Vio- al-promising-approaches-2008.PDF. lence against Women in Latin America and the Caribbean: and L. Heise. 2015. What Works to Prevent Violence against A Comparative Analysis of Population-Based Data from Women and Girls Evidence Reviews: State of the Field of 12 Countries. Washington, DC: Pan American Health Or- Research on Violence Against Women and Girls. Paper ganization. Available at https://oig.cepal.org/sites/default/ 1. What Works Programme. London: UK Department for files/violence1.24-web-25-febrero-2014_0.pdf. International Development. Available at https://www.what- CDCP (Centers for Disease Control and Prevention). 2014. Vi- works.co.za/documents/publications/16-global-evidence- olence against Children in Haiti Findings from a National reviews-paper-1-state-of-the-field-of-research-on-violence- Survey 2012. Atlanta, Georgia, USA: Centers for Disease against-women-and-girls/file. Control and Prevention. Available at www.togetherforgirls. Fulu, E. et al., 2015. What Works to Prevent Violence against org/wp-content/uploads/2017/09/2012_Haiti_Findings- Women and Girls Evidence Reviews: Paper 2: Interven- from-a-Violence-Against-Children-Survey.pdf. tions to Prevent Violence against Women and Girls. Pa- Contreras, J. M. et al., 2016. Community-Based Approach- per 2. What Works Programme. London: UK Department es to Intimate Partner Violence: A Review of Evi- for International Development. Available at https://www. dence and Essential Steps to Adaptation. Washington, whatworks.co.za/documents/publications/35-global-ev- DC: The Global Women’s Institute and World Bank. idence-reviews-paper-2-interventions-to-prevent-vio- Available at http://documents.worldbank.org/curated/ lence-against-women-and-girls-sep-2015/file. en/907511467996712161/AUS16688-REVISED-Com- munity-Programs-SinglePages.pdf. 23 Garcia-Moreno, C. et al., 2006. Prevalence of Intimate Partner Michaels-Igbokwe, C. et al., 2016. Cost and Cost-Effective- Violence: Findings from The WHO Multi-Country Study ness Analysis of a Community Mobilisation Intervention on Women’s Health and Domestic Violence. The Lan- to Reduce Intimate Partner Violence in Kampala, Ugan- cet, 368(9543): 1260−9. Available at https://www.ncbi. da. BMC Public Health, 12(196). Available at https:// nlm.nih.gov/pubmed/17027732. bmcpublichealth.biomedcentral.com/articles/10.1186/ Garcia Moreno, C.,et al., 2003. Responding to Violence s12889-016-2883-6. against Women: WHO’s Multicountry Study on Wom- Michau, L. Undated. The SASA! Way to Preventing Violence en’s Health and Domestic Violence. Health and Human against Women. Kampala: Raising Voices. Available at Rights, 6(2): 112−27. http://raisingvoices.org/wp-content/uploads/2013/03/ GoH (Government of Haiti, U.S. President’s Emergency Plan downloads/resources/The_SASA!.pdf. for AIDS Relief (PEPFAR), and Center for Disease Con- Population Council. 2010. Girl-Centered Program Design: trol (CDC). 2012. Violence against Children in Haiti: Find- A Toolkit to Develop, Strengthen & Expand Adolescent ings from a National Survey. Port au Prince: VACS. Girls Programs. New York: Population Council. Available _____. 2014. Plan National d’Action EFH 2014-2020: Re- at www.popcouncil.org/uploads/pdfs/2010PGY_AdolGirl- sume. Port au Prince: Government of Haiti. Available ToolkitResources.pdf. at www.omrh.gouv.ht/Media/Others/1-Publications/Me- Raising Voices 2017. Fidelity to the SASA! Activist Kit. Pro- suresduplansynth%C3%A8se.pdf. gramming for Prevention Series, Brief No.2. Kampala: _____ (Ministry of Public Health and Population (MPHP), Hai- Raising Voices. Available at http://raisingvoices.org/ tian Childhood (IHE), and ICF International). 2013. 2012 wp-content/uploads/2017/05/ProgramBrief2.Fidelity- Haïti Mortality, Morbidity, and Service Utilization Survey: toSASA.RaisingVoices.may2017.pdf. Key Findings. Calverton, Maryland, USA: Government UNAIDS. 2014. Adolescent Girls and Young Women. The of Haiti (MPHP), IHE, and ICF International. Available Gap Report: Geneva: UNAids. Available at www.unaids. http://procurement-notices.undp.org/view_file.cfm?doc_ org/sites/default/files/media_asset/02_Adolescentgirl- id=17735. sandyoungwomen.pdf. Halcón, L. et al., 2003. Adolescent Health in the Caribbe- UNAIDS, UNFPA, and UNIFEM (UNAIDS, United Nations an: A Regional Portrait. American Journal of Public Population Fund, and United Nations Development Fund Health, 93(11):1851−7. for Women). 2004. Women and AIDS: Confronting the Jan, S. et al., 2010. Economic Evaluation of a Combined Mi- Crisis. Geneva and New York: UNAids, UNFPA, and crofinance and Gender Training Intervention for the Pre- UNIFEM. Available at https://gcwa.unaids.org/sites/wom- vention of Intimate Partner Violence in Rural South Africa. enandaids.net/files/UNAIDS-UNFPA-UNIFEM-Wome- Health Policy and Planning, 26(5), 366–72. Available at nAids---Confronting-the-Crisis.pdf. https://www.ncbi.nlm.nih.gov/pubmed/20974751. USAID (United States Agency for International Development). Kyegombe, N. et al., 2014. SASA! is the Medicine that Treats 2010. Haiti HIV/AIDS Health Profile. https://www.usaid. Violence: Qualitative Findings on How a Community Mo- gov/haiti/global-health bilisation Intervention to Prevent Violence against Wom- WHO (World Health Organization). 2013. 16 Ideas for Ad- en Created Change in Kampala, Uganda. Global Health dressing Violence against Women in the Context of Action 7. Available at https://www.ncbi.nlm.nih.gov/pmc/ the HIV Epidemic. Geneva: World Health Organiza- articles/PMC4165071. tion. Available at http://apps.who.int/iris/bitstream/ Lundgren, R., and A. Amin. 2015. Addressing Intimate Part- handle/10665/95156/9789241506533_eng.pdf;jses- ner Violence and Sexual Violence among Adolescents: sionid=C74883D729F58A92A897978FB15FA9B2?se- Emerging Evidence of Effectiveness. Journal of Adoles- quence=1. cent Health, 56(1): S42−S50. Available at https://www. --------. 2014. World Report on Violence and Health. Geneva: ncbi.nlm.nih.gov/pubmed/25528978. WHO. 24 References  APPENDIX A. SASA! AND POWER TO GIRLS: THEORY OF CHANGE AND LOGIC MODEL Theory of Change manner. The responsive parenting (RP) intervention is based on this model. SASA!’s four steps take stake- The original SASA! methodology and Power to Girls holders from all levels of the community through each were developed based on the perspective that IPV and step of the intervention, changing over time their knowl- other forms of VAWG are the result of multiple causal lay- edge, attitudes, and behaviors that perpetuate harmful ers. The framework, originally developed by Urle Bron- gender norms and creating an enabling and empower- frenbrenner (1994) and since adapted by Lori Heise for ing environment that supports safety, nonviolence, and VAWG (1998), presents the causes of IPV at the macro- the dignity of women, men, girls, and boys. social, community, interpersonal, and individual levels. At the community level, imbalance of power is manifested Rethinking Power Model in inequitable norms and harmful practices. These norms To conceptualize the theory of change of the Rethink- are often perpetuated by religious or cultural justification, ing Power project, the evaluation and project teams which deters action by stakeholders who are crucial to reviewed the original SASA! logic model and the com- preventing violence. Lori Michau and colleagues argue plementary Power to Girls to identify which areas were that inequitable norms can be addressed through educa- still applicable for the expanded project and where tion and capacity building, which fosters collective action additions needed to be made to incorporate the new (Michau et al., 2014). This approach allows participants focus on girls. These are detailed in Figure A1. to feel empowered to contribute to an enabling environ- The Rethinking Power model details the logical ment in which a significant part of the population prac- progression that the SASA!/Power to Girls approach tices and advocates for gender equality. Using a similar employs to reduce violence. It begins first by detail- approach, norms that perpetuate gender inequality and ing some of the most important factors that need to drive violence can be addressed in the healthcare sec- be considered when designing a program aimed at re- tor by carrying out educational and behavior change in- ducing violence and providing an overview of how the terventions among health care professionals and other approach seeks to influence these factors at multiple stakeholders (Gennari, 2014). levels. It then describes the initial, intermediate, and Community mobilization interventions are success- long term outcomes of the program approach, as well ful in transforming harmful gender norms because they as the final impact RP seeks to achieve. guide community members at all levels of the ecologi- The model begins by noting some of the most im- cal framework through gradual and sustainable change. portant risk factors for VAWG that should be consid- This approach is founded in the transtheoretical model ered when implementing the program. These include of behavior change. According to this model, there are sociodemographic factors (e.g., sex, age, income, six key stages through which behavior change occurs: education, employment, religion, residence) and so- precontemplation, contemplation, preparation, action, ciocultural factors (e.g., family characteristics, social maintenance, and termination (Prochaska, 1997). support, and alcohol use). For example, education lev- Community mobilization efforts informed by this frame- els correlate with the likelihood of experiencing VAWG. work assist individuals and communities as they move As such, RP may work to increase the safety of girls through these six stages in an organic and empowering 25 FIGURE A1. SASA! ANALYSIS Levels of SASA! Longer activities reaching each Initial Intermediate term Context circle of influence Outcomes Outcomes Outcomes Impact Knowledge Societal Skills • Recognizing VAWG as Individual & problem • Response to women and girls National Policy Makers, experiencing violence Collective Media • Types of VAWG Capacity • Hold boys & men accountable • Consequences of VAWG • Supportive • Promote balanced power environment • VAWG/HIV linkage • Support activists/ couples • Enhanced ability Socio- to prevent and Reduced social demographic respond to acceptance of Factors VAWG gender inequality, Institutional Action IPV and sexual Sex Schools • Intention to act abuse of girls Awareness Age • Personal change: balancing • Root case as imbalance power of power between Income women and men, girls • Public change: sanctions against and boys VAWG Behaviors Decrease in Education • Change can happen • Acceptability of expanded gender experience/ roles • Balancing power perpetration of Employment Community • Communication IPV and sexual Police, Local Leaders, with partner abuse of girls Religion Health Services, • Decreased risk Place of residence/ Girl’s Clubs behaviors levels of mobility • Community Critical Thinking and activism Acceptance and Influence Increase in girl’s Dialogue • Attitudes towards power, gender, freedom and Socio-cultural • Public debate and human rights feelings of safety Factors discussion • Silence broken Relationship • Personal reflection Family character- istics Relatives, Elders, Friends, Neighbors Decrease in Social support Sustained Action HIV/ SRH risk behaviors Alcohol use • Changed policies • Organized Connection groups Participation • Activists/ leaders/ professionals • Changed prac- Individual • Activists at grassroots, in tice in relation- increasingly connected and leadership, in institutions ships, communi- Women, Men, Youth, active Community, Activities ty, institutions 26 Appendix A. SASA! and Power to Girls:Theory of Change and Logic Model on school grounds, as well as to advocate for family who experience violence, holding men and boys ac- and community acceptance of girls’ education. Each countable for their actions, and promoting a balance of of these interventions may increase girls’ educational power. In addition, changes in individual and community attainment and subsequently reduce their risk of ex- behaviors will begin to occur, along with changes in atti- periencing VAWG. Conversely, these factors may pro- tudes on concepts such as power, gender, and human vide contextual knowledge that help improve program rights. Parallel to these achievements, the program will delivery. An example of this would be to determine the continue to strengthen and build its community activist primary employment or income-generating activities network that drives these changes within the community. are used in an area may change how the program These intermediate outcomes lead to long-term approaches dissemination activities (e.g., where and outcomes, including increased capacity that support when to hold activities and which key business leaders longer-term behavior change and sustained action to or other informal sector employees to engage). reduce VAWG. At this level, there will be improved in- The model then presents a visual of the ecological dividual and collective capacity to prevent and respond model approach that informs the RP program, describ- to incidents of violence within the community. Behav- ing the different levels that the methodology seeks to im- ior change will also continue at the community (e.g., pact. As can be seen in the model, the RP methodology increased community activism) and individual levels addresses each of these interlinked drivers of violence (reduced risk behaviors, balanced power, improved by working with community activists to change gender partner communications). In addition, larger societal norms at multiple levels with a variety of groups. The and community-level transformations will take place to initial outcomes expected from the program include in- provide a more supportive environment for women and creased knowledge of community members (male and girls (e.g., improved policies, transformed institutions, female) of the types and consequences of VAWG and and community groups) the linkages between violence and HIV. The program Together, these initial, intermediate, and lon- also expects to increase community awareness of the ger-term outcomes will facilitate a number of expect- root cause of VAWG (an imbalance of power) and build ed program impacts. As detailed in the above model, consensus that it is possible to achieve change within these include (i) reducing the social acceptance of a community. It also hopes to promote critical thinking, gender inequality, IPV, and the sexual abuse of girls; dialogue, and the development of community activ- (ii) decreasing experiences of/perpetration of IPV and ists who can facilitate further community-level change the sexual abuse of girls; (iii) increasing girls’ freedom among men, women, girls, and boys. and feelings of safety; and (iv) decreasing sexual and The initial outcomes give way to intermediate out- reproductive health and human immunodeficiency comes that take longer to develop. This includes building virus (HIV) risk behaviors. This model will guide pro- skills on appropriately responding to women and girls gram and evaluation design for the RP program. A Community-Based Intervention to Prevent Violence Against Women And Girls in Haiti: Lessons Learned 27 APPENDIX B. KEY INFORMANT QUESTIONNAIRE Key Informant Questionnaire for the Adaptation, Implementation, Evaluation, and Acceptability of the SASA! Methodology and Power to Girls Intervention Programs in Haiti Goal and Confidentiality The goal of this interview is to collect important informa- stitute (GWI), and Beyond Borders in Haiti in 2016−17. tion from you as a key informant about the process, the Any information provided here will be anonymous and organization, and the structure used during the adap- any comments or information shared during this inter- tation, implementation, evaluation, and acceptability of view will be personally attributed to you. Your participa- the SASA! Methodology and Power to Girls Programs tion is voluntary and it may take you approximately 45 by George Washington University, Global Women’s In- minutes to complete this questionnaire. The SASA! Methodology and Power to Girls Key Informant Questionnaire A We would like to ask you a few questions about yourself. Please write the following: First and Last Name Organization Name Title Intervention Role Starting Date of Involvement in the Intervention Current Date 28 The SASA! Methodology and Power to Girls B Planning and Adaptation 1. Please discuss the process, the structure, and the organization put in place for planning, select- ing, and adapting the SASA! Methodology. These may range from brainstorming exercises, group discussions, similar interventions implemented in Haiti, literature reviews, or any formal needs assessment activities. 2. Discuss to what extent you have considered such issues (e.g., as the fidelity, the fit, or the cultural acceptance of the intervention during the planning and adaptation phase). 3. What are your insights about some of the lessons learned, challenges faced, and proposed solu- tions during the planning and adaptation phase of the intervention? A Community-Based Intervention to Prevent Violence Against Women And Girls in Haiti: Lessons Learned 29  The SASA! Methodology and Power to Girls C Implementation 1. Please discuss the process, the structure, and the organization put in place during the implemen- tation phase of the intervention. 2. What are your insights about some of the lessons learned, challenges faced, and proposed solu- tions during this phase of the intervention? D Evaluation 1. Please discuss the process, the structure, and the organization put in place during the evaluation phase (quantitative and qualitative) of the intervention. 2. What are your insights about some of the lessons learned, challenges faced, and proposed solu- tions during this phase of the intervention? 30 Appendix B. Key Informant Questionnaire The SASA! Methodology and Power to Girls E Acceptability 1. Please discuss the process, structure, and organization put in place to ensure the acceptability of the intervention by the target communities. 2. What are your insights about some of the lessons learned, challenges faced, and proposed solu- tions during this phase of the intervention? F Final Comments 1. Please add anything else that you would like to add about any of the phases mentioned above. A Community-Based Intervention to Prevent Violence Against Women And Girls in Haiti: Lessons Learned 31 32