TOWARD an END to CHILD MARRIAGE Lessons from Research and Practice in Development and Humanitarian Sectors

Home , Child marriage, List of child brides

TOWARD AN END TO CHILD MARRIAGE Lessons from Research and Practice in Development and Humanitarian Sectors

SAVE THE CHILDREN AND BEYOND

June 2018

Julie Freccero Audrey Whiting HUMAN RIGHTS CENTER

The Human Rights Center at the University of California, Berkeley, School of Law conducts research on war crimes and other serious violations of international humanitarian law and human rights. Using evidence- based methods and innovative technologies, we support efforts to hold perpetrators accountable and to protect vulnerable populations. We also train students and advocates to document human rights violations and turn this information into effective action.

The Health and Human Rights Program at the Human Rights Center promotes the health and protection of marginalized populations affected by humanitarian crises. Through applied research and technical assistance, we partner with local and international organizations to develop new tools, guidance, and interventions.

Human Rights Center, UC Berkeley School of Law, 396 Simon Hall, Berkeley, CA 94720 Telephone: 510.642.0965 | Email: [email protected] hrc.berkeley.edu, Medium.com/humanrightscenter | @HRCBerkeley

SAVE THE CHILDREN

Save the Children fights for children every single day. We stand side by side with children in the toughest places to be a child. We do whatever it takes to make sure they survive, get protection when they’re in danger, and have the chance to learn. Because every child should be able to make their mark on the world and build a better future for us all.

savethechildren.net

Cover Photo: Rokaya,* 14, stands in the grounds of a temporary learning center for children who fled the conflict in Marawi city, the Philippines, run by Save the Children. At this site, girls and young women face a heightened risk of early or forced marriage as well as other forms of gender-based violence.*The child’s name was changed to protect her identity. Photo by Hanna Adcock / Save the Children. Design and graphics: Nicole Hayward CONTENTS

ABBREVIATIONS / v

EXECUTIVE SUMMARY / 1

INTRODUCTION / 7

IMPETUS AND OBJECTIVES / 9

METHODOLOGY / 11

FINDINGS, PART I: THE LITERATURE / 15 DEVELOPMENT LITERATURE / 15 HUMANITARIAN LITERATURE / 24

FINDINGS, PART II: SAVE THE CHILDREN PROGRAMMING AND ADVOCACY / 25 INTRODUCTION TO SAVE THE CHILDREN’S WORK ON CHILD MARRIAGE / 25 MAPPING / 25 KEY INFORMANT INTERVIEWS / 31

DISCUSSION / 48 LESSONS FROM RESEARCH AND PRACTICE / 48 SAVE THE CHILDREN’S APPROACHES AND THE EVIDENCE / 52 EMERGING ISSUES IN CHILD MARRIAGE PROGRAMMING / 54

RECOMMENDATIONS / 56 DEVELOPMENT / 56 HUMANITARIAN / 60 RESEARCH / 61

APPENDICES / 63

NOTES / 80

ABBREVIATIONS

AMS Awards Management System IDP Internally Displaced Persons ASRH Adolescent Sexual and Reproductive INGO International Non-governmental Health Organization CCT Conditional Cash Transfer IPV Intimate Partner Violence CEDAW Convention on the Elimination of IRC International Rescue Committee All Forms of Discrimination Against NGO Non-governmental Organization Women RCT Randomized Control Trial CEDPA Centre for Development and RH Reproductive Health Population Activities SC Save the Children CEFM Child, Early, and Forced Marriage SGBV Sexual and Gender-based Violence CM Child Marriage SIDA Swedish International Development CRC Convention on the Rights of the Child Cooperation Agency CSE Comprehensive Sexuality Education SRH/R Sexual and Reproductive Health/ CSO Civil Society Organization Rights DRC Democratic Republic of the Congo STI Sexually Transmitted Infection FGM/C Female Genital Mutilation/Cutting UCT Unconditional Cash Transfer GBV Gender-based Violence UN United Nations HIV/AIDS Human Immunodeficiency Virus/ UNICEF United Nations Children’s Fund Acquired Immune Deficiency VSLA Village Savings and Loan Association Syndrome WASH Water, Sanitation, and Hygiene HRC Human Rights Center ICRW International Center for Research on Women

EXECUTIVE SUMMARY

DESPITE RECENT DECLINES in the prevalence of Methodology child marriage globally, an estimated 12 million girls The study methodology has three components: are still married each year.1 Additionally, there has been growing concern around reports of increasing 1. LITERATURE REVIEW: A literature review was child marriage rates among those affected by con- conducted to identify the state of the global evi- flict or natural disaster where underlying drivers dence on child marriage prevention and response such as gender inequality, social norms, and lack interventions. In sum, 355 development and 27 of opportunities for girls are exacerbated, and new humanitarian articles from both grey and aca- drivers related to protection concerns and extreme demic literature were evaluated for inclusion. poverty arise. Consequently, girls who are married Forty-three development and no humanitarian young experience a violation of their human rights articles met the inclusion criteria and were ex- and suffer negative impacts on their education, eco- tracted and evaluated for quality using adapted nomic prospects, social lives, mental and physical STROBE and CONSORT scales.2 Data was re- health, and the health and nutrition of their chil- viewed and findings relating to programs and dren. Effective interventions are urgently needed to initiatives for combatting child marriage were address this critical issue. identified and analyzed for common themes. The purpose of the following study is to gather findings from efforts to prevent and respond to 2. MAPPING EXERCISE: A mapping exercise of Save child marriage in both development and human- the Children’s past and current programs to address itarian contexts and determine what Save the child marriage was undertaken using internal and Children and other organizations can do to improve published documents from three sources: submis- their response to this critical issue. The study adds sions from Save the Children member and country to existing research by synthesizing the most up-to- offices, Save the Children’s Awards Management date literature, bringing the voices of practitioners System Database, and the program and project and experts into the discussion, and examining the Evaluation Repository database. Any evaluations topic of child marriage interventions in a humani- meeting inclusion criteria for the literature review tarian context for the first time. In addition, through were included in the literature review extraction and a mapping and analysis of Save the Children’s pro- analysis process. Program information from the 181 gramming, it assesses the organization’s approaches resulting documents that met the inclusion criteria against the available evidence. was extracted to a standardized form for analysis.

1 Methodology

1 A literature 2 A mapping 3 Key informant review to exercise of Save interviews with 27 Literature identify the Mapping the Children’s Key practitioners active Review state of Exercise past and current Informant in child marriage the global programs to Interviews prevention and evidence on address child response in 21 child marriage marriage using countries or regions prevention internal and across Africa, the and response published Middle East, South interventions documents from Asia, Europe, and three sources North and South America

3. KEY INFORMANT INTERVIEWS: Key informant and legal and advocacy. Empowerment approaches interviews were conducted with 27 practitioners were the most frequently evaluated and consistently active in child marriage prevention and response effective, but covered a broad range of activities that in 21 countries or regions across Africa, the Middle were rarely evaluated independently to determine East, South Asia, Europe, and North and South which components were most effective. Incentive/ America. Interviews were conducted over Skype us- asset transfer approaches had mixed results, but ing a semi-structured interview guide. Twenty-one were more likely to be effective when focused on interviewees were from Save the Children, while six younger adolescents and conditioned on or used practitioners were from other non-governmental to promote educational outcomes (such as school organizations (NGOs). Transcripts were coded and attendance). Community sensitization and engage- analyzed thematically to identify patterns in par- ment approaches, while promising, were poorly ticipant responses. represented and evaluated in the literature, but programs focusing on deeper community engagement Findings were generally more successful than those that used The literature review identified 43 experimental and more light-touch sensitization approaches. Finally, quasi-experimental evaluations of 30 distinct inter- legal and advocacy approaches were found to be ventions in development settings that had as their the least represented in the literature, likely due to primarily objective or a significant outcome im- substantial challenges in evaluating such programs pacts on child marriage outcomes including related using experimental and quasi-experimental meth- behavior, knowledge, and attitudes. Studies focused ods. We identified two studies evaluating programs primarily on African and South Asian contexts. designed to address child marriage in humanitarian Interventions were typically designed as single-com- contexts, but neither fit the criteria for inclusion in ponent or multi-component and targeted girls alone the study. or multiple levels of the social-ecological model, The mapping of Save the Children’s program- including families, communities, policy environ- ming revealed that Save the Children has been ac- ment, and society. Approaches to addressing child tive in addressing child marriage since at least 2001 marriage included empowerment, incentive/asset when the Ishraq program for out-of-school girls was transfer, community sensitization and engagement, instituted in Egypt. Since then, the organization has

2 | Toward an End to Child Marriage conducted over 52 projects, programs, and advocacy Recommendations initiatives in more than 41 countries. The majority of The following recommendations are targeted at Save programs are designed as multi-component, multi- the Children and intended to help move the organi- level interventions, and all four approaches identified zation forward in addressing child marriage in de- in the literature are recognized in Save the Children’s velopment and humanitarian contexts. They are also programming. However, empowerment and com- intended to help other stakeholders, including re- munity sensitization and engagement approaches are searchers, policymakers, and practitioners, to more the most frequently utilized, with legal and advocacy effectively prevent and respond to child marriage. efforts increasingly common. Incentive/asset transfer approaches were used to address child marriage Development only twice by the organization.3 Program evaluations Cross-thematic Recommendations ranged from highly rigorous randomized controlled 1. A dopt a gender equality lens when designing, im- trials (RCTs) to quasi-experimental and qualitative plementing, monitoring, and evaluating child evaluations, and less-rigorous baseline-endline eval- marriage programs and initiatives. uations. Four past and current programs in human- 2. E xpand the geographic coverage of child mar- itarian contexts were identified. riage programming and advocacy to include Key informants discussed interventions or strate- countries or regions with very high rates of gies to address child marriage, outcomes or impacts child marriage not currently covered by Save (anecdotal or based on formal evaluation), imple- the Children’s efforts. mentation challenges, risk factors and intervention 3. T arget programs and advocacy to address the adaptation strategies for humanitarian settings, in- following under-reached populations: 1) girls formation or guidance needs of practitioners, and in urban contexts, 2) very marginalized girls, 3) key programming gaps. A number of themes arose faith-based and customary/traditional leaders, from the interviews. On strategies and promising 4) men and boys, and 5) married girls. practices, practitioners discussed the importance 4. Consider addressing context-specific risk fac- of multi-level, integrated interventions, mobilizing tors at multiple social-ecological levels; how- communities to change social norms, conducting ever, these activities should complement efforts formative research to inform intervention design, to support, empower, and build the capacity of and contextually sensitive framing of child marriage. girls themselves. They also identified implementation challenges and 5. Contextualize empowerment approaches to gaps, including the need for a holistic, coordinated address local drivers of child marriage as well strategy across sectors, the diverse and context-spe- as the specific needs and interests of the girls cific drivers of child marriage, the resource and time themselves. intensive nature of social norms change, limited 6. U se incentive/asset transfer approaches to build legal frameworks and enforcement, and the diffi- girls’ human capital by incentivizing educa- culty of translating national level legal and policy tional outcomes or combining them with com- successes into meaningful change at the local level. plementary approaches for long term impact. Researchers identified certain groups—such as eth- 7. Ensure community sensitization and engage- nic minorities, faith-based leaders, men and boys, ment approaches are long-term, participatory, and married girls as needing greater inclusion in fu- and engage a broad range of actors within a ture efforts. community who have been identified through formative research.

Toward an End to Child Marriage | 3 8. I dentify and engage key decision-makers and/or 8. I ncrease the rigor of evaluations of child mar- people of influence within communities to serve riage advocacy initiatives using creative, con- as champions for child marriage prevention and text-appropriate methodologies. response. 9. D esign programs with the following consider- Child Poverty ations in mind: 1) scalability and cost, 2) length, 1. M easure the impact of child poverty program- 3) formative research, 4) community input, and ming on behavioral child marriage outcomes. 5) rigorous evaluation. 2. C onsider the use of incentive/asset transfer pro- 10. D evelop a coordinated strategy at the organi- grams which incentivize educational outcomes zational level to address child marriage by es- to reduce child marriage. tablishing joint objectives, outcomes, and in- 3. Ensure that incentive/asset transfer programs dicators; unifying terminology; and clarifying are well-evaluated and that impact is measured where the issue sits within Save the Children by age. and how coordination within the organization 4. Consider the use of age-appropriate employ- will take place. ment and livelihood opportunities, as well as 11. D evelop a coordinated strategy at country lev- training, to delay marriage among older adoles- els with all stakeholders within and outside of cent girls. Save the Children active in addressing child marriage. Child Protection 1. E valuate the impact of child protection inter- Advocacy ventions on child marriage-related outcomes 1. P lace youth voices at the center of all advocacy ef- using rigorous experimental and quasi-exper- forts, with a particular emphasis on girls’ voices. imental methodologies or alternative rigorous 2. Develop high-level partnerships and identify evaluation approaches. champions to expand political will and buy-in. 2. En courage other sectors to include child mar- 3. C oordinate with well-established women’s rights riage outcomes when evaluating interventions organizations to strengthen advocacy efforts and to address drivers of child marriage in the local maximize impact. context. 4. C ombine advocacy efforts with technical assis- 3. C oordinate with other thematic areas to address tance to support governments in developing and child marriage through approaches such as edu- implementing legislative and policy changes. cation, health, or livelihoods. 5. En sure advocacy efforts complement program- 4. P rovide assessment and measurement guidance ming directly benefitting girls where possible, to other sectors to support the identification of that programs integrate advocacy components, contributing factors to child marriage. and that programming and advocacy are coordinated and work in synergy with each other. Child Rights Governance 6. E xpand legal advocacy initiatives to include ed- 1. Support opportunities for child-led advocacy ucational goals for adolescent girls. and accountability on child marriage, such as 7. Assess the impact of advocacy achievements via child clubs and other forums for children on women and girls, rather than strictly focus- to receive information about their rights, gain ing on whether legal or policy goals have been advocacy skills, and lead activities in their com- achieved, to strengthen evaluation efforts. munities and countries.

4 | Toward an End to Child Marriage 2. B uild gender equality and child marriage into 4. M easure the impact of nutrition programming child-friendly budget analysis and advocacy on child marriage behaviors. work, to increase resourcing for child marriage 5. E valuate the impact of SRH programming, in- policies and programs. cluding CSE, SRHR information, and youth- 3. Integrate gender equality and child marriage responsive SRH services, on behaviors, knowl- into efforts to protect and expand civil society edge, and attitudes around child marriage. space from local to international levels, includ- 6. Offer youth-responsive SRHR education pro- ing through working with girls’ and women’s grams and services, as well as nutrition pro- groups. grams, to already-married girls.

Education Humanitarian 1. M easure the impact of educational programming 1. C onduct assessments to identify both pre-existing on behaviors, knowledge, and attitudes on child drivers of child marriage, as well as new risk fac- marriage. tors related to crisis coping, such as economic 2. R emove barriers to school participation and in- and protection concerns, when developing child crease access to education for adolescent girls. marriage interventions. 3. C onsider additional incentives to lower barriers 2. D etermine which forms of child marriage inter- to school attendance and increase the oppor- vention are appropriate for various phases of an tunity cost for older adolescent girls to leave emergency. school. 3. Identify key entry points for mainstreaming 4. Integrate comprehensive sexuality education child marriage into existing humanitarian re- and/or child marriage, gender equality, and sex- sponse efforts. ual and reproductive health and rights (SRHR) 4. P ilot stand-alone child marriage programs and into standard education curriculum. integrate sector-specific approaches to address 5. R emove barriers to education for already-mar- child marriage in humanitarian settings. Build ried girls and pregnant and adolescent mothers. in rigorous evaluations to develop evidence in acute and protracted displacement settings. Health and Nutrition 5. B uild child marriage prevention and response 1. D etermine whether food-insecurity is a driver of capacity in fragile contexts whenever possible to child marriage and consider how nutrition pro- ensure that the underlying drivers of child mar- grams may improve child marriage outcomes. riage are being addressed prior to acute crises. 2. Determine whether lack of access to youth- 6. En able girls to return to school as quickly as responsive sexual and reproductive health possible following acute crisis. (SRH) services is a driver of child pregnancy, and thus child marriage, in a given context and Research consider how such programs may improve child 1. E xpand research on child marriage drivers and marriage outcomes. interventions to include settings that are poorly 3. Integrate child marriage into comprehensive represented in the literature, including child sexuality education (CSE), SRHR, and gender- marriage in urban contexts, child marriage in based violence (GBV) education and awareness- regions such as Latin America and Southeast raising activities for in- and out-of-school youth. Asia, and child marriage in the context of gang conflict in Central America.

Toward an End to Child Marriage | 5 2. C onduct more rigorous impact evaluations of achievements on women and girls and 2) em- child marriage programming and advocacy ploying creative, context appropriate method- initiatives. ologies to improve rigor of non-experimental 3. Disaggregate evaluation data by age to better methodologies when necessary. understand what works to prevent child mar- 10. E valuate the impact of other approaches such as riage among younger and older adolescents and health, nutrition, and education on child mar- among different social, economic, and identity riage as an outcome. groups. 11. Build a formative research phase into piloting 4. L engthen evaluation time to capture individuals’ and evaluation efforts to tailor interventions to significant life events and social norm change. the context-specific drivers and needs and pref- 5. E valuate each component of multi-component erences of the community. and multi-level programming individually as 12. Prioritize research addressing support needs, well as together to determine which components outreach strategies, and effective interventions are most successful at addressing child marriage. for already-married girls to expand child mar- 6. C onsider cost, sustainability, and scalability when riage response efforts. evaluating program impacts. 13. B uild an evidence base in humanitarian settings 7. E valuate distinct components of empowerment that addresses the following topics: 1) why child programming separately to determine the im- marriage increases in some contexts, 2) decision- pact of each component on child marriage making factors of parents, 3) pre-existing and outcomes. crisis-specific drivers of child marriage, 4) sup- 8. I mprove evaluation of community sensitization port needs of girls and their families, and 5) what and engagement to contribute significantly to interventions work to address child marriage. organizational and global evidence on changing 14. Engage youth in participatory research during social norms. formative and pilot design phases to ensure that 9. Improve evaluation of advocacy initiatives to the needs and interests of young people are pri- contribute to organizational and global evi- oritized and to improve buy-in. dence by 1) assessing the impact of advocacy

6 | Toward an End to Child Marriage INTRODUCTION

CHILD MARRIAGE, defined as formal or informal as armed conflicts and natural disasters, and that union before the age of 18,4 is a pervasive global phe- where it does, adolescents under age 15 are often the nomenon. According to the United Nations Children’s most significantly impacted.10 In Jordan, for exam- Fund (UNICEF), more than 650 million women alive ple, UNICEF reported that rates of child marriage today were married as children, and more than a among displaced Syrian communities increased third of those unions took place before those women from 12% in 2011 to nearly 32% by the end of 2014.11 were 15. Currently, 12 million girls are married each In Bangladesh, Human Rights Watch found that year, 5 meaning globally one in six adolescent girls families were hurrying to marry off their daughters (aged 15–19) is married or in union.6 While some before river erosion took away their homes.12 progress has been made with regards to decreas- Drivers of child marriage vary significantly across ing the prevalence of child marriage, it is often not regions and even within countries, but in nearly ev- enough to offset population growth in regions such ery context, child marriage is rooted in gender in- as West and Central Africa where the practice is the equality and discrimination. While statistics on most common.7 In Niger, for example, where women rates among boys are poor, current data shows that have an average of 7.3 children each,8 prevalence of girls are at nearly five times the risk compared with child marriage is 76%. According to UNICEF, given boys.13 Other drivers of child marriage include so- the impact of population growth, at the current rate cial and cultural norms and traditions with deeply of decline, countries are not on track to reach the entrenched beliefs around virginity and purity as Sustainable Development Goal of ending child mar- they relate to sex and pregnancy outside of marriage by 2030, and an additional 150 million girls will riage. Economic factors also play a significant role, be married by that date.9 with girls from poor families nearly three times In addition, there has been increasing concern more likely to marry before age 18 than those from around child marriage in humanitarian contexts, wealthy ones.14 Parents who are unable to adequately where practitioners say a complex combination of meet the financial needs of their families may see factors leads parents to turn to early marriage as marriage of a daughter as a way to lighten the finan- a negative coping mechanism. While research in cial burden, obtain a bride price, or keep dowries this context is minimal thus far (several large-scale more affordable. Finally, where education is inac- prevalence studies are due for publication this year), cessible or of low quality and where job prospects one multi-country study found that early marriage for girls are poor, girls may become brides simply may increase markedly during some crises such because they have few alternatives.15

7 In humanitarian contexts such as armed conflict emphasizing that child marriage, among other or natural disaster, child marriage is driven by many harmful practices, impairs “the recognition, enjoy- of the same factors; however, these factors are often ment and exercise of the human rights and funda- severely aggravated by significant loss of resources, mental freedoms of women and children.”21 Second, crumbling infrastructure, breakdown of the rule of child marriage has a serious impact on girls’ edu- law, and uprooted social structures. Families report cation and future economic prospects. Girls who marrying off their daughters because there is not marry before 18 years of age are less likely to stay enough food to go around and few educational opin school and consequently, have 9% lower earn- portunities or job prospects in displacement camps. ings as adults.22 Young brides also face increased In addition, protection concerns become primary, social isolation and are at significantly higher risk and parents arrange marriages hoping to protect of physical and sexual violence at the hands of their their daughters’ safety and honor in volatile situa- partner than girls married after 18.23 Finally, child tions where living quarters are often close and leav- marriage is correlated with a number of serious ing home unsafe.16 health consequences. Complications from preg- Unfortunately, girls who marry young face con- nancy and childbirth are the number one reason sequences that impact nearly every aspect of their for the death of girls ages 15–19. In addition, babies lives. First and foremost, child marriage is a violation born to adolescent mothers are more likely to have of children’s rights. While the Universal Declaration low birth weight, and thus suffer from malnutri- of Human Rights does not explicitly address child tion and underdevelopment. This often perpetuates marriage, Article 16(2) states that “Marriage shall the cycle of poverty and early marriage for another be entered into only with the free and full consent generation.24 of the intending spouses.”17 The Convention on the Effective interventions to prevent and respond Elimination of All Forms of Discrimination Against to child marriage and its consequences are sorely Women (CEDAW) adds that any marriage of a needed. This report outlines the findings of a three- child, which the Convention on the Rights of the part study, conducted in collaboration with Save the Child (CRC) defines as a person under the age of Children, to determine what we know and what we 18,18 shall have no legal effect.19, 20 still need to learn about preventing and responding Recently, committees for both CEDAW and CRC to child marriage in both development and human- reaffirmed 18 as the minimum age of marriage, itarian contexts.

8 | Toward an End to Child Marriage IMPETUS AND OBJECTIVES

Impetus the experiences of their practitioners have not yet Save the Children is a global leader in addressing been synthesized and analyzed to inform a clear the health and protection needs of children in both strategy to guide their work on this issue. development and humanitarian contexts. Through Further, the global evidence addressing child its member and country offices, Save the Children marriage interventions is not available in an acces- implements programs and advocacy initiatives to sible format for practitioners and has several limita- prevent and respond to child marriage. The Human tions (Box 1). First, previous literature reviews ad- Rights Center (HRC), at the University of California, dressing child marriage interventions include only a Berkeley, School of Law, through its Health and small handful of studies due to strict inclusion crite- Human Rights Program, conducts research to pro- ria, often excluding grey literature or quasi-experi- mote the health and protection of marginalized mental studies that may offer useful information for populations affected by humanitarian crisis. HRC programming. Second, previous reviews also do not undertook this research, in partnership with Save include interventions in humanitarian contexts in the Children, in order to provide the organization addition to development contexts. Finally, research with a synthesis of available evidence both within and evaluations addressing child marriage inter- and beyond the organization to inform Save the ventions lack the perspectives of practitioners, their Children’s strategy development and to strengthen strategies, and challenges encountered in the field, programming to address child marriage. In addi- in order to ensure that we are building on lessons tion, this research responds to key gaps in the evi- learned from practice. dence base more broadly and aims to provide organizations that address child marriage with accessible Objectives information to inform intervention design. This study has three main objectives: Although Save the Children implements a broad 1. To provide a summary of the state of the evidence range of child marriage prevention and response in- regarding child marriage prevention and response terventions around the world, the organization has interventions, including promising practices and not yet undertaken a comprehensive mapping of evidence gaps; their previous and current work on the issue of child 2. To assess the nature, scope, and coverage of Save marriage to ensure strategic coverage and coordina- the Children’s current and recent child marriage tion across member and country offices. In addition, work within the past 10 years; and lessons learned from their program evaluations and

9 3. To identify strategies, implementation challenges, BOX 1 and the research priorities of practitioners within Reviews and Reviews-of-Reviews on Save the Children and beyond. Interventions to Address Child Marriage 1. Child Outcomes of Cash Transfer Programming Research findings are intended to inform Save by Anjini Mishra and Francesca Battistin (Save the Children, 2018) the Children’s strategic focus, future programming, 2. The Global State of Evidence on Interventions to research, and advocacy, further strengthening the Prevent Child Marriage by Sophie Chae and Thoai organization’s ability to prevent and respond to Ngo (Girl Center, 2017) child marriage in development and humanitarian 3. “Preventing Gender-Based Violence Victimization settings. in Adolescent Girls in Lower-Income Countries: Systematic Review of Reviews,” Social Science & Medicine, by Kathryn Yount, Kathleen Krause, and Stephanie Miedema (2017) 4. “Interventions to Prevent Child Marriage Among Young People in Low- and Middle-Income Countries: A Systematic Review of the Published and Gray Literature by Amanda Kalamar,” Journal of Adolescent Health, by Susan Lee-Rife, and Michelle Hindin (2016) 5. “Interventions to Prevent or Reduce Violence Against Women and Girls: A Systematic Review of Reviews” by Diana J. Arango, Matthew Morton, Floriza Gennari, Sveinung Kiplesund, and Mary Ellsberg (The World Bank Group, 2014) 6. “An Evidence Review of Gender-Integrated Interventions in Reproductive and Maternal-Child Health,” Journal of Health Communications, by Joan Kraft, Karin Wilkins, Guiliana Morales, Monique Widyono, and Susan Middlestadt (2014) 7. “The Efficacy of Interventions to Reduce Adolescent Childbearing in Low- and Middle-Income Countries: A Systematic Review,” Studies in Family Planning, by Kate McQueston, Rachel Silverman, and Amanda Glassman (2013) 8. “What Works to Prevent Child Marriage: A Review of the Evidence,” Studies in Family Planning, by Susan Lee-Rife, Anju Malhotra, Ann Warner, and Allison McGonagle Glinski (2012) 9. Solutions to End Child Marriage: What the Evidence Shows by Anju Malhotra, Ann Warner, Allison McGonagle, and Susan Lee-Rife (ICRW, 2011)

10 | Toward an End to Child Marriage METHODOLOGY

THE STUDY METHODOLOGY includes three com- professional networks were consulted to identify ad- ponents: a review of the academic and grey liter- ditional research or grey literature that had not been ature addressing child marriage prevention and found through the previous processes. The resulting response interventions; a review of project and pro- 355 development and 27 humanitarian articles were gram documents available in Save the Children’s stored using RefWorks web-based reference manage- global database; and key informant interviews with ment software. practitioners from Save the Children and other or- All titles and abstracts resulting from the searches ganizations involved in child marriage prevention were screened according to the inclusion criteria and response. provided in Table 1 below. Criteria for literature addressing interventions in humanitarian contexts were Literature Review expanded to include all years and humanitarian in- Researchers identified peer-reviewed and grey liter- dicators (such as armed conflict or natural disaster) ature that evaluate interventions to prevent and re- in addition to low- and middle-income country in- spond to child marriage in development (i.e. low- and dicators. The remaining 43 evaluations from develop- middle-income country contexts) and humanitarian ment contexts were subject to full-text review, quality (i.e. armed conflict, natural disaster, or displace- assessment, and data extraction. No evaluations for ment) contexts. Relevant databases were combed us- interventions in humanitarian contexts met inclusion ing search terms appropriate to each database’s native criteria. Extraction was conducted using predefined controlled vocabulary. Searches were built by com- criteria outlined in a comprehensive data extraction bining terms that indicate a young person in a mar- template (Appendix 2). In addition, researchers ap- riage-like arrangement, with terms that indicate pre- plied ranking criteria, adapted from the STROBE and vention, intervention, reduction, or response in both CONSORT standards for observational studies and humanitarian and development settings (Appendix randomized trials,25 to assess the quality and report- 1). In addition, websites for non-governmental orga- ing of intervention and evaluation design (Appendix nizations (NGOs), international non-governmental 3). Disagreements were recorded and resolved by fur- organizations (INGOs), and governmental organiza- ther discussion among the research team. Findings tions known to be active in child marriage program- from the included studies relating to programs and ming were hand-searched for relevant publications. initiatives for combatting child marriage in humani- Searches were expanded by screening bibliographies tarian and development settings were identified and of published and grey literature matching eligibility analyzed for patterns. Findings are outlined in the criteria to identify new sources. Finally, experts and following report.

11 TABLE 1 INCLUSION/EXCLUSION CRITERIA FOR LITERATURE

DEVELOPMENT SETTINGS Inclusion criteria Exclusion criteria

Publication type Peer-reviewed journals or grey literature from organizations known to be active in CEFM

Publication date 2007–2017

Study type Intervention studies, clearly reported Descriptive studies methodology

Population Girls or boys <18 years of age

Intervention Targeting CEFM, ≥1 month in length Addressing SGBV broadly

Outcome Measure At least one CEFM-specific Other forms of SGBV broadly

Location Low- or middle-income countries High-income countries

Language English Other languages

The literature review is subject to several limita- over the age of 18, which may have impacted find- tions. First, the search strategy may have failed to ings with regards to age at marriage or proportion capture literature on interventions that address the of girls married, where that was not taken into ac- root causes of child marriage under the greater um- count. Furthermore, most studies relied on self-re- brellas of sexual and gender-based violence (SGBV) ported data on age at marriage, which may have un- or child protection if child marriage language was derestimated true rates, especially in countries with not clearly used. In addition, search results pertain- strong child marriage laws. Finally, evaluation stud- ing to already-married girls were limited to those ies were often too short to capture long-term pro- that were clearly marked as preventing or respond- gram impacts and did not follow participants over a ing to child marriage, though it is very likely that long enough period to capture age of marriage for a many promising interventions of this type don’t use majority of participants. child marriage response language and/or include In general, published literature is frequently bi- married adolescent girls in programs alongside ased towards interventions that had positive out- married adult women. Finally, the search excluded comes (known as publishing bias), while interven- studies not reported in English. tions with ineffective or harmful outcomes may be At the program level, participation in initiatives less likely to be published. Likewise, interventions represented in the literature was generally volun- funded by international NGOs or government orga- tary, meaning there was likely a selection bias to- nizations are more likely to be published and may be ward participants who were more inclined to delay biased in favor of the program for funding purposes. marriage at baseline, thus showing an exaggerated Additionally, given that child marriage research has program effect. Ages of participants also varied been conducted in only a few global regions, practi- widely between studies. While researchers initially tioners should exercise caution in generalizing find- tried to limit the literature to those studies that ings to other regions, making it important to invest evaluated programs with participants under age 18, in strong data collection and analysis within all fu- they found this criterion to be too restrictive. Thus, ture programs. several programs include young adult participants

12 | Toward an End to Child Marriage Mapping Once collected, program name, implementation Researchers reviewed and synthesized Save the dates, objectives, target population, intervention com- Children’s internal and published documentation, ponents, theory of change, and evaluation methods focusing primarily on project and program evalu- and outcomes were input into a standardized map- ations and research on child marriage interventions ping form for analysis. Any evaluations which met across the movement since 2012. Documents were the inclusion criteria for the literature review were collected from a number of sources including: put through the same data extraction, quality review, and analysis process used for published literature. 1. Submissions by Save the Children member and There are a few notable limitations to the mapping country offices; exercise. First, use of Save the Children’s Evaluation 2. Save the Children’s Awards Management System Repository is voluntary, which means member of- (AMS) Database; and fices have not necessarily submitted all evaluations of 3. Save the Children’s Evaluation Repository (a data- child marriage-related programming to the reposi- base where members and country offices may vol- tory. Second, not all grants or “awards” identified in untarily submit program or project evaluations). the AMS Database were associated with strong program documentation. Thus, not all programs related A request for program documents was circulated to child marriage were well-understood. Third, par- among Save the Children member offices, asking for ticipation in the mapping exercise was not manda- all relevant proposals, donor reports, and evalua- tory. Though all member offices were encouraged to tions with a child marriage component within the submit documents and fill in documentation gaps, past five years. The AMS search was conducted by ex- some member offices were more actively engaged in porting all grants or “awards” to Excel and applying this process than others. Given these limitations, this filters to identify active, closed, and approved awards mapping exercise should not be considered exhaus- relating to child marriage. Researchers then down- tive or in any way representative of all of Save the loaded logframes, final project reports, or evaluation Children’s programming related to child marriage, reports associated with each award. Together, these especially in regards to non-current interventions. resulted in 175 program documents. The Evaluation Repository was searched for documents related to Key Informant Interviews child marriage, yielding a total of 89 documents. All Researchers conducted key informant interviews mid-term reviews, baseline studies, needs assessments, with 27 practitioners currently engaged in child or formative research documents were excluded, leav- marriage prevention and response efforts. The -ma ing 31 evaluation documents. Finally, transcripts from jority of these interviews (21) were conducted with key informant interviews were reviewed for project Save the Children staff. Six practitioners at other in- or program information relevant to mapping. (For ternational and local NGOs implementing child mar- key informant methodology, see below.) Documents riage programming were also interviewed (Box 2). from all sources which met inclusion criteria were Researchers selected an initial set of key infor- those related to interventions (including programs, mants within Save the Children through purposive tools and curricula, and campaigns) that sought to sampling based on their work on child marriage address child marriage as an explicit objective or prevention and response programming, research, those that measured child marriage indicators as an or advocacy. Sampling aimed to include Save the outcome in any year. In total, 206 Save the Children- Children staff focusing on gender, child protec- generated documents were collected and reviewed tion, advocacy, and program development across and 181 included in the final analysis. member, regional, and country offices. From these

Toward an End to Child Marriage | 13 BOX 2

Key Informant Offices and Locations

SC Member Offices SC Country/Regional Offices Outside NGOs • Canada • Asia • Nepal • Arab Region • International • Bangladesh • Niger • Bangladesh • Netherlands • Ethiopia • Sierra Leone • Ethiopia • United Kingdom • Latin America • Somalia • Lebanon • United States and Caribbean • South Sudan • Somalia • Malawi • West and Central Africa • South Sudan

interviews, snowball sampling was used to identify interviews were not as directly involved as others in additional key informants involved in direct imple- child marriage program design or implementation. mentation of child marriage programming within Although these informants could provide general Save the Children and other organizations, priori- information about their programming and advo- tizing those working in humanitarian contexts to cacy, they were not able to speak about promising address the significant gap in the literature on inter- strategies, challenges, and research needs based on ventions in these contexts. direct experience. Third, interviews were conducted Researchers conducted all interviews by Skype primarily with staff focusing on gender and child using a semi-structured interview guide which was protection, while only a few focused on advocacy. developed based on research objectives and informed Save the Children staff working in the areas of ed- by key findings from the literature review. Interviews ucation, poverty / livelihoods, and nutrition were explored the following topic areas: nature of child not interviewed, and therefore strategies to address marriage, context-specific risk factors, nature of inter- child marriage through these sectors were not fully ventions or strategy, outcomes or impact (anecdotal explored. Finally, the vast majority of interviews or based on formal evaluation), implementation chal- conducted were with practitioners from Save the lenges, adaptation strategies for humanitarian set- Children, and their perspectives reflect their experi- tings, information or guidance needs of practitioners, ences within this organization. Promising practices, and key programming and research gaps. Researchers challenges, and research priorities of practitioners took detailed notes during interviews and followed in other organizations were not fully explored and up with thematic coding and analysis of those notes may follow different trends. Given limited time and to identify patterns in participant responses. resource constraints, and the lack of literature ad- This qualitative research has several limita- dressing child marriage interventions in these con- tions. First, interviews were conducted in English texts, individuals working in humanitarian settings only. Second, some key informants in our initial were prioritized.

14 | Toward an End to Child Marriage FINDINGS, PART I THE LITERATURE

Development Literature Fourteen countries are represented in the evalua- Study Characteristics tion literature, including Bangladesh, Burkina Faso, Researchers identified, extracted, and analyzed a to- Egypt, Ethiopia, India, Kenya, Malawi, Mexico, tal of 43 experimental and quasi-experimental eval- Pakistan, Senegal, Tanzania, Uganda, Zambia, and uations to determine what is known about prom- Zimbabwe (Figure 1). Program settings included 23 ising practices, programs, and evidence gaps to rural, 2 urban, and 18 mixed rural, semi-urban, or address child marriage in development contexts urban settings. Program location settings were gen- (see Appendices 4 and 5). Of these documents, 31 erally chosen (where reported) based on low liter- were grey literature, including United Nations (UN) acy rates, high poverty rates, and high prevalence of reports, NGO reports, working papers, and dis- child marriage. sertations, and 12 were academic (peer-reviewed). Of the programs included in the evaluation Evaluations included 15 randomized controlled trials studies, a majority of interventions (29) targeted (RCTs), 23 quasi-experimental studies, and 5 natural adolescent girls, with ten programs targeting both experiments which evaluated 30 distinct projects or adolescent girls and boys, and four targeting fami- programs. Six studies evaluated programs imple- lies of adolescents or the community more broadly. mented by Save the Children. Across studies, the term adolescent is used gener- Of those studies that met inclusion criteria, re- ously, and where defined, may include children and searchers determined that 24 were of high quality, 18 youth ranging in age from 10 through 24 years. Most of medium quality, and 1 of low quality according to programs targeted some broad span of ages within modified STROBE and CONSORT scales.26 Medium this range (i.e. 12–19 years); however, five programs and low-quality evaluations generally scored lower specifically targeted younger adolescents (ages 10– because the intervention was poorly described, 14 years), including Save the Children’s “CHOICES” inclusion/exclusion criteria for participants was curriculum and four school stipend programs for poorly outlined, the sample size was not justified girls beginning in grade 6 (roughly age 11). In addi- on a statistical basis, no adjustment for confound- tion, six studies aimed to respond to the needs of al- ing was used, and limitations were not reported. We ready married adolescent girls either exclusively or retained all studies that met inclusion criteria for in addition to preventing child marriage. For clarity, learning, regardless of quality. However, only me- these are separated from the remaining analysis, and dium and high-quality studies were included in the findings are examined in Box 3. analysis of successful approaches.

15 Number of Studies

7–9

2–4

FIGURE 1: Locations of child marriage program evaluations represented in this study.

In general, studies sought to measure program- early pregnancy or contraceptive use. The remain- matic impact on child marriage via behavior change, ing 18 interventions seek to reduce child marriage change in knowledge and attitudes, or both. Of the as one of a range of objectives, including improving evaluations measuring behavior change, indicators educational outcomes; improving economic oppor- generally included age at first marriage or propor- tunity; decreasing SGBV; improving SRH outcomes; tion of girls married between baseline and endline. changing gender norms, attitudes, and practices; Of the evaluations that sought to measure change strengthening social networks; and improving life in knowledge and attitudes, a wide range of indica- skills. tors were evaluated for both girls and their parents Programs to address child marriage may be single- or guardians, including knowledge of the legal age or multi-component by design. A single-component of marriage and legal consequences for violations, intervention is one in which the program activities perceptions regarding marriage before 18 years of are focused around addressing one primary driver age, belief in girls’ right to choose both marriage age of child marriage. For example, the Zomba cash and partner, knowledge of advantages or benefits of transfer program sought to address child marriage marrying after 18 years of age, and knowledge of a in Malawi by giving a cash transfer to families with range of consequences of marrying before 18 years never-married daughters between the ages of 13–22 of age. years.27 Multi-component interventions, on the other hand, are those that seek to address multi- Intervention Design ple drivers of child marriage using several different The evaluation studies outline a number of differ- components as part of the same program. These may ent intervention designs and approaches used to all be targeted directly towards girls, or they may ad- prevent child marriage in development contexts. dress multiple levels of the social-ecological model Interestingly, only five of the interventions aimed from a girl’s family, to her community, her policy to address child marriage alone, while five inter- environment, and her society. The Empowerment ventions combined child marriage objectives with and Livelihood for Adolescents (ELA) program by the objectives around fertility decision-making such as NGO BRAC is one such program. It combines five

16 | Toward an End to Child Marriage different components into one program, including 1) and 16 assess 13 distinct multi-component interven- adolescent development centers, 2) life skills train- tions, 7 of which evaluated all components as a sin- ing, 3) livelihoods training, 4) community participa- gle intervention and 9 of which evaluated compo- tion, and 5) microfinance (Case Study 1).28 Overall, nents separately and against a control to determine of the 37 evaluation studies included in this report, the impacts of individual components. 21 assess 14 distinct single-component interventions

BOX 3

Responding to Child Marriage

The prevention of child marriage is a also evaluated response components ly being piloted by IRC in Lebanon to critical focus of child marriage interven- (Appendix 5). Response programming address the specific needs of married tions, yet the stark reality is that despite in these studies addressed a broad array adolescent girls in displacement set- efforts to the contrary, 12 million girls of topics, including use of modern tings. The program uses a 26-module are still married each year.29 While contraceptives, use of SRH services, life skills curriculum tailored to the local some of the challenges faced by young pregnancy-related care seeking, egali- context through participatory assess- brides are similar to those of their tarian marital relations, and measures of ments to cover topics such as self-es- unmarried peers, many, including those autonomy and social support networks. teem/self-worth, healthy relationship having to do with SRH, family planning, Evaluations measured outcomes such skills, financial management, family IPV, parenting, social isolation, educa- as increased contraceptive use, delay planning, and SGBV. With girls’ approval, tion, and livelihoods, are unique to this and spacing of pregnancies, seeking husbands and mothers-in-laws are en- population. For the nearly one in six out RH services and pregnancy-relat- gaged through one-on-one meetings adolescent girls (aged 15–19) married ed care, absence of marital violence to ensure girls are allowed to attend or in an informal union, it is critical that and physical punishment of children, sessions regularly. interventions and support are tailored improved partner communication and Generally speaking, Save the Chil- to meet their unique needs.30 Unfor- knowledge of SRH, and improved eco- dren is well-poised to contribute to tunately, many organizations are not nomic outcomes. addressing the needs of married girls focusing on supporting married girls An example of an effective response through new and existing program either because they do not feel it falls project for married girls is TESFA (To- models. As one key informant told us, under their mandate or because they ward Economic and Sexual Reproduc- fear supporting married girls could be Response falls firmly under our man- tive Health Outcomes for Adolescent viewed as condoning child marriage. date. We’re talking about anyone under Girls), which worked with married girls Of those that do provide programming, the age of 18. . . . We also do a ton in the Amhara region of Ethiopia to many include married and unmarried of work with adults. . . . This is really achieve four primary aims: to 1) em- girls in the same cohorts, failing to relevant work through the themes we power girls with information, skills, and address their distinct needs. have: livelihoods, life skills, post- a support network, 2) provide essential secondary education, through health There are so many organizations work- health information and services, 3) we’re doing work around SRH rights, ing on prevention of early marriage, educate and rally parents and commu- maternal health. . . . It’s really not a but not a lot are working at response to nity members, and 4) provide financial stretch for us, [it’s] just connecting the early marriage. . . . You cannot promote and livelihoods training. Over the three dots. That is something we haven’t awareness and prevention of early years it ran, the program saw a signifi- done as well as we could. marriage in a culture that does not cant improvement among married girls —PRACTITIONER, SAVE THE CHILDREN, want to stop early marriage. . . . I think with regards to improved husband and CANADA it would be much more effective, when wife communication, decreased SGBV, you’ve passed the point when you can improved mental health, increased As programming to address child do prevention, to do response. So our investment in productive economic marriage continues to develop, inter- job is to empower these girls, empower assets, and improved knowledge and ventions that respond to the needs these individuals in that specific cul- use of SRH services, including family of already married girls, in addition ture. —PRACTITIONER, NGO, LEBANON planning.31 to those that prevent child marriage, should receive increased attention and This review identified 10 articles eval- In humanitarian contexts, key in- funding. uating programming for married girls, formants identified a program known including 4 prevention intervention simply as the “early marriage tailored studies from the literature review that package curriculum,” which is current-

Toward an End to Child Marriage | 17 An analysis of child marriage program evaluations Intervention Approaches over time shows that there has been a significant in- Given the many drivers of child marriage, organi- crease in rigorous evaluations of child marriage pro- zations and governments take a broad range of ap- gramming in the past 10 years. Programming prior proaches to addressing the issue. However, across to 2013 was typically single component by design; interventions, there are four primary approaches however, since 2013, there has also been a recog- evaluated in the literature: 1) Empowerment, 2) nizable increase in the number of multi-compo- Incentive/asset transfer, 3) Community sensitization nent programs, likely in response to the increasing and engagement, and 4) Legal and advocacy. acknowledgement of the complex range of drivers that may promote child marriage in a given context. Empowerment Around the same time, evaluations of multi-com- Social economist and women’s empowerment spe- ponent programs began to separate components cialist Naila Kabeer defines empowerment as, “The for individual evaluation against a control to better expansion in people’s ability to make strategic life determine which program components were most choices in a context where this ability was previ- effective. ously denied to them.”32 Thus, empowerment ap- Overall, program periods ranged in duration proaches are those that seek to improve a child’s from 1–16 years, with nearly half of all evaluated agency, choice, and freedom of action with regards programs falling between 1–2 years in length, and to marriage and other aspects in his or her life.33 more than three quarters between 1–5. Given that Given the broad definition, empowerment pro- longer programs were all run by government en- grams encompass a wide range of interventions. tities, the high number of 1–5 year interventions Common themes across programs are gender and likely reflects the funding cycles to which NGOs child-rights awareness, life skills education, health and INGOs are subject, rather than ideal program education (particularly SRH), financial literacy pro- length. Likewise, intervention time—the amount of grams, and livelihoods skill-building (Box 4). These time a single participant might be engaged in the hold a common theory of change that by improv- program—ranged in length from 3 days to 5 years, ing a girl’s knowledge, skills, self-confidence, and with nearly three quarters of all interventions com- self-awareness, she will have the tools necessary to pleted in 24 months or less. There are no apparent reduce her vulnerability to early marriage, improve trends in length of programs or interventions over her position within her home and community, and time. expand her alternatives outside of the home. Interventions were generally introduced at the Of the programs evaluated, 22 include empow- level of the child, household/village, community, erment interventions either as a sole component region, or nation. Size or scope of the interven- or in addition to other programmatic components. tions were measured in different ways depending Empowerment interventions are generally held in on the target, while program scope was typically designated youth-friendly spaces which have been measured using number of villages/communities called safe spaces, girls groups, help groups, devel- reached, number of households reached, num- opment clubs, or adolescent development centers ber of girls reached, and number of girls and boys in the literature. These spaces typically function as reached. Programs’ scopes ranged from 100–460 meeting spaces for learning, recreation, and social villages, 600–150,000 households, 860–40,000 activities, are often stocked with books, art sup- youth or children, and 183–2 million girls more plies, and recreation equipment (like balls or table specifically. games), and are generally managed by peer edu- cators or older adolescent or youth leaders who

18 | Toward an End to Child Marriage Intervention Approaches

1 2 3 4 Empowerment Incentive/asset Community Legal and transfer sensitization and advocacy engagement Empowerment programs Incentive/asset Sensitization Legal and advocacy encompass a wide transfer approaches approaches are those approaches may range of interventions. are those approaches that seek to disseminate include ensuring laws Common themes that use the transfer information and raise are in place which set across programs are of cash or goods, awareness around child 18 as the minimum gender and child-rights whether conditional or marriage to counteract age for marriage, awareness, life skills unconditional, to a girl inaccurate and harmful assisting governments education (Box 3) health or her family, to achieve beliefs. Engagement in developing policy education (particularly a decrease in child approaches, on the frameworks that SRH), financial literacy marriage either through other hand, seek support girls’ economic, programs, and incentivizing behaviors to collaborate with educational, and livelihoods skill-building. known to decrease community members social opportunities, child marriage (such as to create long-term strengthening birth and school attendance) or change, including social marriage registration to incentivize the delay norms change, around systems, and improving of marriage. child marriage and monitoring and related topics in their accountability. own communities.

provide information, tutoring, counseling, support, thought to work by removing economic barriers to or referrals as needed. In these spaces, curricula, education and increasing the opportunity cost of readings, lectures, discussions, games, songs, and early marriage. In the process, organizations hope to role play may be used to educate girls or mixed-gen- address gender inequality by changing attitudes of der groups on a large range of topics (Box 4). families and communities around the value of girls. Incentives/asset transfer approaches evaluated by Incentive/Asset Transfer this study found that while unconditional transfers Incentive/asset transfer approaches are utilized in 17 were occasionally employed, conditional transfer of of the child marriage interventions included in this cash and/or goods was by far the more common ap- report. Incentive/asset transfer approaches are those proach. Assets transfers were most frequently con- approaches that use the transfer of cash or goods, ditioned on either 1) school attendance, or 2) delay whether conditional or unconditional, to a girl or of marriage. Cash transfers for school attendance in- her family, to achieve a decrease in child marriage cluded payment of tuition and monthly or quarterly either through incentivizing behaviors known to stipends, often combined with transfer of goods in decrease child marriage (such as school attendance) the form of books, school supplies, or uniforms, and or to incentivize the delay of marriage. Where eco- conditioned on school attendance rates of 75–85%, nomic motivators such as poverty, the relatively with one program adding the additional require- high cost of schooling, or higher dowry costs for ment of at least a 45% score on annual exams.34 older adolescent brides are seen as drivers of child Incentives to delay marriage were more commonly marriage, incentive/asset transfer approaches are in-kind incentives, though cash was occasionally

Toward an End to Child Marriage | 19 BOX 4

Themes in Empowerment and Life Skills Programming

Education Leadership and Communication Rights and Protection • Programs for out-of-school girls • Communication and negotiation • Child and women’s rights • After school literacy/numeracy • Critical thinking • Early marriage rights and effects • After school tutors or mentors • Decision-making skills • SGBV Financial Education/Literacy • Teambuilding • Land rights • Budgeting and saving • Volunteering Sports and Recreation • Accounting • Good citizenship training • Physical fitness • English/French language skills Livelihoods Education • Mental wellbeing Gender Awareness • Computer skills • Teamwork and cooperation • Gender and sexuality • Market assessment • Self-confidence • Power dynamics • Enterprise skills • Gender roles and division of labor • Income generating opportunities • Mentorship and jobs assistance Health • Microfinance • WASH and nutrition • FGM/C Psychosocial • HIV/AIDS • Mental and social wellbeing • ASRH and contraception • Group discussions on feelings • Safe motherhood • Group discussions on gender discrimination

employed. These goods-based incentives took the her community has the potential to place her at inform of chickens, goats, or oil and were given to girls creased risk of violence and abuse. or their parents once yearly for every year they de- Interventions that fall within the category of layed marriage, upon completion of the program or community sensitization and engagement are very on a girl’s 18th birthday if she was unmarried. diverse and may engage actors from every level of society from parents or guardians to neighbors, Community Sensitization and Engagement friends, and relatives, to community and faith-based Community sensitization and engagement approaches leaders. Forums for engaging with the community were represented in nine of the interventions in- range across studies from one-on-one meetings cluded in this report. Sensitization approaches are with parents and community leaders, to community those that seek to disseminate information and raise meetings or conversations, print or non-print mass awareness around child marriage to counteract inac- media communications, and public ceremonies or curate and harmful beliefs. Engagement approaches, events such as theatre or video shows. Two interven- on the other hand, seek to collaborate with commu- tions required parents or guardians to publicly sign nity members to create long-term change, including a document committing to allow their daughters social norms change, around child marriage and re- to attend the program and to delay their marriages lated topics in their own communities. The theory during the program period,35 while a third had women of change for this approach type is based on the un- of the community make a public statement that they derstanding that discriminatory gender norms, tra- would abandon the practice of child marriage alto- ditional practices, and expectations are significant gether.36 Master trainers, also called community fa- drivers of child marriage in many contexts, and that cilitators or influencers, may be trained to lead activi- empowering a girl without educating and engaging ties within the community, or special subcommittees

20 | Toward an End to Child Marriage CASE STUDY 1

Empowerment and Livelihood for Adolescents (ELA) Program by BRAC

The Program: The ELA program was developed in 1993 in Bangladesh to address female disempowerment by improving vocational skills and building knowledge to enable girls to make informed choices about sex, reproduction, and marriage. The program is delivered in after-school adolescent development clubs to reach school dropouts in addition to those still enrolled. Since 1993, BRAC has established more than 9,000 clubs reaching one million adolescent girls in Bangladesh and has begun implementation in other countries, including Tanzania and Uganda. Outcomes: An evaluation of ELA in Bangladesh showed a decrease in child marriage among the intervention group, as well as increased engagement in economic activities, though researchers warned that there were differences between treatment and control groups that may have impacted the outcome.37 In Uganda, however, two independent research studies found the program had a significant impact on child marriage,38 with treatment groups showing an overall reduction of 62% in one study.39 In addition, the program was found to reduce teen pregnancy and increase girls’ likelihood of engaging in income generating activities. Challenges: Researchers faced a number of challenges when rolling out the program in Tanzania, including resource limitations and preference among participant girls for educational activities rather than income-generating activities. In the end, the evaluation in Tanzania found the program had no significant impact.40 Conclusions: The ELA program highlights the potential for promising programs to be replicated globally, yet serves as a warning that such rollout won’t always be successful due to a number of factors. It is especially critical that programs are adequately resourced and adapted to the local context if interventions are to be successful.

with representation from various segments of the approaches generally focus on the importance of community may be utilized to share child marriage having structural frameworks that enable large-scale messaging and impact norms and practices. In these social change. Two of the three interventions which programs, child marriage messaging is frequently utilized legal and advocacy approaches to address combined with other forms of education and sensi- child marriage were the Ishraq program in Egypt, tization around other harmful traditional practices which collaborated with local and national policy- such as female genital mutilation/cutting (FGM/C), makers to build support for policies that benefitted gender equality, girls’ education, and adolescent girls, and Deepshikha, which established gender re- sexual and reproductive health (ASRH) needs. source centers to support regional and national adolescent and gender programming. 42 Legal and Advocacy Legal and advocacy approaches are those that seek Characteristics of Successful Approaches to address child marriage by developing a strong le- When examining what works to address child mar- gal and policy framework at national, regional, and riage in development contexts, we examine themes international levels. Legal and advocacy approaches in both intervention design and approaches and may include ensuring laws are in place which set their correlates with success. With regards to inter- 18 as the minimum age for marriage, assisting gov- vention design, the number of program components ernments in developing policy frameworks that did not clearly correlate with success in the litera- support girls’ economic, educational, and social op- ture. In general, single-component approaches ap- portunities, strengthening birth and marriage reg- peared to be more successful more frequently, while istration systems, and improving monitoring and multi-component interventions more often had accountability.41 Theories of change behind these mixed results. However, this is likely due to publication

Toward an End to Child Marriage | 21 bias—evaluations with positive outcomes are more likely necessary to delay marriage for older adoles- likely to be reported—and due to the fact that multi- cent girls.44 component programs, especially those with com- When examining the success of incentive/asset ponents evaluated individually, are more likely to transfer programs on preventing child marriage, have at least one ineffective component. Only two we generally find mixed results. However, by look- studies evaluated a multi-component intervention ing at individual study branches in evaluations with intended to address more than one level of the social- mixed outcomes, we find that depending on design, ecological model. They did not find that multi-level, these programs can be very successful at achieving multi-component interventions were necessarily the intended outcomes. Incentive/asset transfer pro- more successful than a simple, single-component grams were successful in Mexico, Africa, and South intervention.43 Asia. They frequently worked for targeting girls in Turning to approaches, we find that empower- grades 6–8 and in high school. When we evaluated ment programming is the most consistently effec- impact by age, cash or in-kind transfers conditioned tive programming approach across evaluations— on educational outcomes were particularly effective both as stand-alone programs and when evaluated for younger adolescents (ages 12–14). While impacts separately as part of multi-component programs. of education-conditioned transfers dropped off for Empowerment programs were successful at achiev- older adolescents (ages 15 and older), incentives ing both positive behavioral and knowledge and at- for delaying marriage seemed more helpful among titude-related outcomes in Africa and South Asia, this cohort. Effective programs tended to last on for girls between ages 10–24 years, and in both average 24 months to 5 years, were conditioned on rural and urban contexts. Themes within success- school attendance of 75–85%, and included money ful empowerment programs included designated given to families and/or directly to girls themselves. meeting spaces, life skills, and livelihoods educa- Interventions with mixed results showed that cash tion, educational support, and gender and child or in-kind transfers conditioned on attendance are rights awareness components. They also tended to helpful for girls not likely to be in school but have be 1–3 years in length. Very few studies report unsuc- essentially no effect for girls likely to be in school cessful outcomes; however, among those that do, already (although cash may encourage them to stay themes include a short intervention time (3 days– in school, the noted effect is small). When compar- 6 months), a lack of programming resources such ing cash with in-kind transfers, we find more infor- as designated space, supplies, or well-trained facil- mation about cash. However, transfer of uniforms itators, and an emphasis on topics in which girls and school supplies worked particularly well with were not interested (e.g., livelihoods training over younger adolescents, while incentives of cooking oil education). Among empowerment programs that and chickens worked to delay marriage among some did not show a statistically significant effect, most older adolescents. Only two interventions employed reported early positive trends, but began program- unconditional cash transfers (UCTs).45 One did not ming with girls who were very young and/or did find UCTs to be effective at delaying marriage, while not evaluate over a long enough time period to the other found them to be effective for girls already capture most participants’ transitions into mar- in school, but only for as long as the intervention riage. Interestingly, one evaluation that measured persisted. However, unconditional transfers of outcomes separately for younger (aged 12–14) and school uniforms and school supplies were found older (aged 15–17) adolescents found that empow- to be consistently helpful in incentivizing school- erment programming worked particularly well in ing and thus delay of marriage. Cash and in-kind the earlier years, but that further intervention was transfers worked to delay marriage; however, gains

22 | Toward an End to Child Marriage CASE STUDY 2

Second-Generation Research on Child Marriage Interventions by Population Council

Background: In 2012, Population Council, together with Findings: In summary, researchers found the following: the Ethiopia Ministry of Youth and Sports, reported that • Community dialogue/sensitization: Correlated with a girls who took part in Berhane Hewan, a multi-component significant reduction in child marriage among 12–14 program designed with community consultation to meet year olds in Ethiopia and 15–17 year olds in Burkina the needs of married girls and prevent child marriage, Faso. reduced rates of marriage for girls ages 10–14 to one- tenth of those of the comparison group.46 With such • Education promotion: Correlated with a significant promising results, many expected the program to be reduction in child marriage among 12–14 year olds in scaled up nationwide. However, scale-up efforts met Ethiopia and Tanzania. resistance due to concerns around cost and feasibility of • Conditional asset transfer: Correlated with a significant expending a multi-component intervention. To address reduction in child marriage among 15–17 year olds in this challenge, Population Council designed a second- Ethiopia and Tanzania. generation research study based on a modified version of Berhane Hewan which simplified the program into three • Comprehensive model: Correlated with a significant basic components, taking cost into consideration.47 reduction in child marriage at all ages in Tanzania and among 15–17 year olds in Ethiopia.48 The Study: Research was conducted in Burkina Faso, Ethiopia, and Tanzania and involved the following Conclusions: The study found that simple, cost- components: effective interventions can be highly effective in delaying child marriage. Interventions like these are 1. Community dialogue/sensitization ($9–$20/person) particularly promising for scale-up over large regions 2. Education promotion with school supplies ($13–$20/ and demonstrate potential as simple, cost-effective girl) interventions in emergency contexts. 3. Conditional asset transfer of two chickens or a goat ($32–$107/girl) Interventions were tested over 27–28 months, both individually and as a comprehensive model compared against a control. Cost effectiveness was also analyzed.

faded immediately when girls turned 18 or the pro- to be poorly resourced, shorter in length, and tended gram stopped. Researchers expressed concern that to focus on more superficial forms of engagement this was because such transfers do not promote the like sensitization and short information sharing development of human capital through education or sessions. Most programs with a community-facing livelihoods training. component showed some positive impact on knowl- Among community sensitization and engage- edge/attitudes around child marriage; however, of ment programs, outcomes are mixed. In general, those programs that measured impact on behavior, programs are poorly explained in the literature and just two demonstrated positive outcomes directly notably heterogeneous with regards to design, ob- attributable to community engagement.49 jectives, target populations, forums, and program or Legal and advocacy approaches are poorly de- intervention length. Successful programs are gener- scribed and evaluated in the literature, perhaps due ally 1–6 years in length, use mass media campaigns, to the significant challenges in evaluating such pro- or focus on in-depth forms of community engage- grams using experimental and quasi-experimental ment such as community dialogue or participatory methods. Of the three studies that met our inclu- action. Those with poor outcomes were more likely sion criteria and utilized advocacy approaches,50 the

Toward an End to Child Marriage | 23 most common method for evaluating success was fit our inclusion criteria. The first, a multi-country simply to determine whether advocacy goals had evaluation of a participatory video project called been achieved. The Deepshikha model was the only Through Our Eyes, used qualitative methods to as- program to measure an advocacy outcome (arguably sess changes in awareness around several SGBV also an empowerment outcome). It demonstrated a outcomes, including child marriage. The results 5% increase in girls’ civic engagement when evaluated were promising, but the evaluation was not rigor- against the comparison group, as well as reporting ous enough to meet our criteria.53 The second study some impact on the design and implementation of is a multi-country RCT of an empowerment in- similar government programs. The latter, however, tervention called Creating Opportunities through was not evaluated using experimental methods.51 Mentorship, Parental Involvement and Safe Spaces The Ishraq program, which aims to “increase local (or COMPASS), which is currently being conducted and national policymakers’ support for girl-friendly by International Rescue Committee (IRC) in east- measures and policies,” among its many program ern Democratic Republic of the Congo and several activities, states only that it was successful at gain- western Ethiopian refugee camps.54 The study in- ing buy-in and active engagement from governors tends to measure early marriage as a secondary out- and relevant ministries which it felt were crucial to come to changes in sexual violence; however, at the national-level scale-up of the program.52 time this report was published, findings related to child marriage had not yet been released. Humanitarian Literature Only two studies evaluate interventions for child marriage in humanitarian contexts. Of these, neither

24 | Toward an End to Child Marriage FINDINGS, PART II SAVE THE CHILDREN PROGRAMMING AND ADVOCACY

Introduction to Save the Children’s Work child marriage in both development and humani- on Child Marriage tarian settings produced the following findings. Save the Children, founded in 1919, is comprised of 29 member organizations and Save the Children Geographic Coverage International and works in over 60 of the world’s The map in Figure 2 below illustrates the geographic least developed countries. The organization fights coverage of Save the Children programs and advo- for children’s rights and sees child marriage as a cacy initiatives across the globe as identified by the fundamental violation of those rights. Child mar- mapping exercise. Interventions have primarily been riage cheats girls of their education. It leaves them at implemented in the Middle East and North Africa, risk of early pregnancy and vulnerable to domestic West and Central Africa, East and Southern Africa, violence and social isolation. It often traps them— and South Asia, but parts of Latin America and the and their children—in poverty. Save the Children Caribbean, Eastern Europe and Central Asia, and programs and advocates—directly and through Southeast Asia are also included. Many of the coun- partners—in humanitarian and development set- tries in which Save the Children works have some of tings to both prevent and respond to child marriage, the highest prevalence rates of child marriage in the through a gender equality lens and drawing on their world, including Niger (76%), Bangladesh (59%), expertise in programming on Child Poverty, Child Burkina Faso (52%), South Sudan (52%), Mali (52%), Protection, Health and Nutrition, Education and and Somalia (45%). However, there are a number Child Rights Governance. of countries with equally high prevalence where Save the Children has the opportunity to contribute Mapping significantly to the reduction of child marriage. In Save the Children has been active in addressing Africa, these include the Central African Republic child marriage since at least 2001 when the Ishraq (69%), Chad (67%), and Guinea (51%) in West and program, which aimed to change harmful gender Central Africa; Madagascar (41%) and Eritrea (41%) norms including child marriage, was introduced in in East and Southern Africa; and Sudan (34%) in the Egypt. Since then, the organization has conducted north. In Latin America and the Caribbean, Brazil more than 52 projects, programs, and advocacy ini- has the highest rates (36%), while Nicaragua (35%), tiatives in more than 41 countries to meet the needs Guatemala (30%) and Guyana (30%) fall close be- of both married and unmarried girls in developing hind. Other places like Lao People’s Democratic countries (Appendix 7). A mapping exercise of Save Republic (35%) in Southeast Asia also present op- the Children’s past and current efforts to address portunities for expansion of efforts.55

25 Programing and Advocacy

Programming

Advocacy

FIGURE 2: Save the Children’s program and advocacy initiative locations.

Target Populations the rates of child marriage or changing knowledge Programs and initiatives addressing child marriage at and attitudes related to child marriage. Programs for Save the Children have generally targeted girls alone married adolescents have a broader range of objec- or boys between the ages of 10–19, and often include tives including reduction in domestic violence; im- parents, the community, and other stakeholders. proved reproductive, maternal, newborn, and child At least nine programs identified in the mapping health outcomes; increased positive parenting skills claimed to take a social-ecological approach, which and practices; and empowerment. targets more than one level, including adolescent girls, their families, their communities, policy envi- Thematic Areas ronments, and/or their societies.56 Four programs in Save the Children focuses on five Global Thematic five countries, along with the “Choices, Voices, and Areas: child rights governance, health and nutrition, Promises” curriculum which has been implemented education, child protection, and child poverty (Box in at least 10 countries, sought to specifically address 5). Researchers attempted to analyze child marriage the needs of very young adolescents (ages 10–14),57 programs by thematic areas; however, program doc- while 13 programs implemented in nine countries umentation rarely identified which thematic areas sought to meet the needs of married adolescent girls were involved and in what capacity. When relevant or girls and boys either exclusively or together with thematic areas were indicated (less than 37%), child unmarried adolescents.58 Programs geared towards protection and child rights governance were the very young adolescents include multi-component most commonly cited, particularly in humanitarian empowerment type programs with social, economic, contexts where child protection was the primary educational, ASRH, and community sensitization thematic area involved. and engagement components, all aimed at lowering

26 | Toward an End to Child Marriage We also analyzed the cross-thematic area of gen- BOX 5 der equality. No programs were identified as gender equality programs per se and none of the docu- Save the Children’s Five Global Themes ments explicitly stated that programs were designed 1. Child Rights Governance: Save the through a gender-equality lens (though it is possible Children works to identify and addresses that some programs were in practice). However, a the systemic and underlying factors that can help or hinder the realisation of children’s few programs, including Marriage: No Child’s Play rights. By addressing a lack of open, inclusive (multiple countries), CHANGES (Somalia), Life Steps and accountable governance for child rights and a lack of resources to fund obligations (Nigeria), Keep It Real (Ethiopia and Uganda), and to children, Child Rights Governance Bien Grandir! (DRC) include at least one gender programming and advocacy helps to remove these obstacles and create conditions for equality-related objective or program component. long-term sustainable change.

Summary of Design and Approaches 2. Child Protection: Save the Children works to prevent violence from happening and Intervention Design respond efficiently by: strengthening families and caregivers, implementing laws and Of the programs and initiatives included in this map- policies, working with governments and other ping, 12 aimed solely to address child marriage. These partners to develop strong child protection systems, changing attitudes and social norms, included flagship programs such asMarriage: No listening to and involving children, building the Child’s Play (Case Study 3) and The Right to Be a Girl. social work workforce and providing services for children. Nine additional programs sought to address child 3. Health and Nutrition: Save the Children marriage as one of a few objectives, including reduc- focuses on ensuring that all young children re- ing child labor, addressing trafficking, and improving ceive the health and nutrition care and services outcomes related to ASRH or SGBV. However, the they require to survive. To achieve this goal, Save the Children collaborates with local gov- majority of programs bundled child marriage objec- ernments, multilateral institutions, grassroots tives along with several other aims including educa- organizations and others to strengthen health systems, and works with individuals, families tional, economic, health, social well-being, gender and communities to encourage the uptake of norms, and child protection objectives. In this vein, healthy behaviors. the majority of Save the Children’s programs are 4. Education: Save the Children believes that multi-component and multi-level in nature and have all children have the right to learn in a safe and happy environment. They work with families been since the first program to address child mar- and communities to ensure that all children riage was introduced in Egypt in 2001. receive the support and care they need in the years leading up to school. Once in primary Intervention Approaches education, they work to ensure that all children learn to read and write, and that no child’s Save the Children has employed each of the ap- learning stops because they are caught up in a crisis situation. proaches to addressing child marriage identified in 5. Child Poverty: The Child Poverty Global the literature to varying degrees, including empow- Theme addresses families and children affected erment, community sensitization and engagement, by and suffering from deprivations relating to legal and advocacy, and incentive/asset transfer extreme, absolute and relative poverty. Poverty exists in societies at all income levels, including approaches. middle and high-income countries with signif- The vast majority of Save the Children’s programs icant forms of inequalities. Save the Children’s efforts are focused on the most deprived girls addressing child marriage involve empowerment as and boys of all ages and may include children the primary approach. Empowerment approaches who lack adequate family care, who have are most frequently employed in the context of recently migrated or have been displaced.

Toward an End to Child Marriage | 27 youth-led clubs or adolescent centers. Programs attendance.62 An RCT comparing this among four may be in school or after school and most frequently intervention packages is currently underway.63 expose adolescents to topics such as life skills, live- It is difficult to pinpoint the exact number of ded- lihood training, and educational programs; child’s icated advocacy campaigns that Save the Children rights, gender equality, and norm change; and SGBV has conducted on child marriage, as advocacy ef- and ASRH education in particular. For married ad- forts may be undertaken directly by advocacy teams olescents, these topics are generally adapted to have or as a component of programming, with a contin- a greater focus on livelihoods training; literacy and uum between programming and advocacy work. numeracy education; maternal, newborn, and child Currently, the organization’s primary advocacy ini- health; and positive parenting skills and practices. tiative is Every Last Child, a global umbrella cam- As many as half of child marriage programs iden- paign which seeks to contribute to the achievement tified in the Save the Children movement involved of the Sustainable Development Goals by pushing a community sensitization and engagement compo- for an increased focus on the most excluded children nent. This has been consistently true since 2001. While by governments and the international development some programs engage families or communities only community. Child marriage is a focal theme under superficially through awareness raising, others, such the campaign, which is operating in over 15 devel- as Marriage: No Child’s Play (Case Study 3), seek to oping countries, plus a number of Save the Children promote long-lasting social norms change by engag- member or donor offices and advocacy offices- tar ing a variety of stakeholders, “promoting community geting the African Union and United Nations. The dialogue, facilitating social mobilization, and sup- SIDA-funded Local to Global initiative also addressed porting collective action.”59 As part of those efforts, advocacy around child marriage, among other issues programs like Marriage: No Child’s Play, CHANGES relevant to Millennium Development Goals 4 and 5. in Somalia, and six others have specific focuses on The program has since transitioned to an advocacy engaging men and boys.60 Tools for community en- approach which supports children and civil society gagement vary widely between programs and many organizations in standing up for their rights, ap- include flipbooks, storybooks, activity cards, com- proaching local decision-makers, and holding their munity games, and radio discussions as employed governments accountable. Advocacy is also one com- by Bien Grandir! in the Democratic Republic of the ponent of many large multi-component programs in- Congo (DRC), or peer-led discussion groups and tes- cluding Marriage: No Child’s Play (Case Study 3) and timonial videos as used by the “Choices, Voices, and The Right to be a Girl. Efforts often center around -ad Promises” curriculum for very young adolescents, vocating for laws to establish 18 as the minimum age their families, and their communities to promote of marriage without exception, strengthening the ex- deeper engagement and lasting social change. isting legal framework, improving policy and budget Incentive/asset transfer approaches were the least commitments for child marriage prevention, increas- commonly employed of all approaches within Save ing accountability and monitoring, and influencing the Children’s child marriage programming. Only the attitudes and commitment of high-level political two programs were found to have used this approach. and influential stakeholders. The first program, Kishoree Kontha in Bangladesh, incentivized delay of child marriage by awarding Evaluations girls cooking oil for every three months marriage Evaluations are not available for all programs included was delayed.61 The second program,Adolescent Girls in the mapping. Of those that are available, various Initiative—Kenya, is currently using household level methodologies were utilized including random- cash transfers to incentivize school enrollment and ized controlled trials (RCTs), quasi-experimental

28 | Toward an End to Child Marriage CASE STUDY 3

Marriage: No Child’s Play

Program: Marriage: No Child’s Play is Save the a comprehensive program. Standardizing the Children’s most comprehensive and largest evaluation process has also been challenging. funded child marriage program. This five-year For example, due to cultural sensitivities, not program, which began in 2016, is currently all countries started with outcome one. In implemented in India, Pakistan, Malawi, Mali, and some contexts, such as Mali and Niger, it has Niger by the More than Brides Alliance which been critical to start with creating an enabling includes Oxfam, Simavi, Population Council, and environment before engaging girls. In Mali, they Save the Children. Program activities, based on a began with less culturally sensitive economic detailed theory of change, target multiple levels and employment activities and will follow up of the social-ecological framework and aim to with SRHR activities. Some components of achieve five key outcomes: the theory of change suffer from funding gaps 1. Young people are better informed about and the varying nature, dose, and frequency of SRHR and empowered to voice their needs activities across contexts makes comparison and rights. difficult. Staff have found that countries with more 2. Increased access to formal education, experience implementing child marriage economic opportunities and child programming have run into fewer challeng- protection systems for girls at risk of es as they are able to draw on staff capacity, and affected by child marriage, and their partnerships, and a more supportive environ- families. ment for this work. They are also able to draw 3. Increased utilization of SRHR services on context-specific curriculum and activities that are responsive to the needs of young that have also proven effective. Linking child people, particularly girls at risk of, or marriage programming to existing social affected by, child marriage. protection schemes, whenever possible, has been an important strategy in countries such 4. Increased engagement and collective social as India where such schemes exist. To this end, action against child marriage and in support program staff inform girls and their families of ASRHR. about available social protection structures and 5. Development of a supportive rights-based ensure they have access those services. Finally, legal and policy environment against child conducting formative research is essential marriage. to informing ongoing social and behavioral Strategies to achieve outcomes include change communication methodologies. empowering adolescent girls with life skills and Research and Learning Agenda: Marriage: youth-responsive SRHR education, increasing No Child’s Play offers an ideal opportunity to access to education and economic opportu- increase our understanding of a multi-compo- nities as alternatives to marriage, engaging a nent, multi-level intervention across contexts. variety of stakeholders to change social norms The program includes a detailed research and build a supportive community environ- agenda that aims to build the evidence base ment, and working with local partners and the regarding what strategies work, in which government to conduct policy dialogue and contexts, and why. Population Council is advocacy activities to improve national legisla- conducting the research and evaluation in all tive and policy environments. five countries to include impact assessments of Implementation: The program, now reaching programs, analysis of existing evidence in each its midpoint, initially set out to implement the context, formative research to fill gaps in under- complete theory of change in each country, standing of context-specific drivers, baseline, with different members of the alliance respon- midline, and endline surveys, and country- sible for different components of the program specific research questions where feasible. The at a given location. In practice, however, mid-term evaluation report is expected to be members have implemented program activities available at the end of 2018. The program will in different locations within the countries, run through 2020. making it difficult to evaluate components as

Toward an End to Child Marriage | 29 evaluations, simple baseline/endline evaluations with advocacy component to enhance the capacity no control, qualitative evaluations, and mixed-method of local officers and communities to protect evaluations. RCT methodologies, which are consid- girls from marriage and enhance the capacity ered by many as the gold standard for impact evalua- of girls to protect themselves from abuse, tions, are only utilized in evaluations of two programs: exploitation, and harm through marriage. The Kishoree Kontha in Bangladesh (completed in 2010) program had a duration of one year and was and the Adolescent Girls Initiative—Kenya which is not formally evaluated, but it was reported to currently ongoing. Quasi-experimental methods uti- be effective in creating dialogue at all levels lizing baseline and endline methodologies with a con- on a topic previously considered taboo in the trol are more common and are represented in evalua- region. tions of eight programs, both completed and ongoing. • Building a Resilient Youth among Syrians and Three evaluations which met the inclusion criteria for Jordanians, Jordan (2015–2017) focused on the literature were included in our review and evalu- protecting children from the risks of child ated as part of the literature review. Qualitative meth- marriage and child labor by providing them ods were used in four evaluations, and simple baseline/ with the tools of awareness, education, and endline evaluations with no clearly defined control institutional support to protect them and were used in 12 others. There are no clear trends re- help them find a pathway to continuing their flected in the type of evaluation methodologies utilized education. The program included psychosocial over the years. Simple baseline/endline surveys were support, life skills training, informal education just as likely to be used since 2015 as before. curriculum, and community awareness raising.

Programs in Humanitarian Settings • CHANGES (also known as SNaP), Somalia Save the Children has been conducting program- (2016–2020) is an ongoing program funded by ming to address child marriage in humanitarian Save the Children UK which has just com- contexts since at least 2013 when Save the Children pleted the design phase and is currently in the Canada began Going Against the Grain, a one-year first year of implementation. It aims to achieve program to address child marriage in Somaliland. broader social norms change with a strong Since that time, the organization has conducted component of addressing child marriage. The programs to address child marriage with refugees, program includes economic empowerment, internally displaced persons (IDPs), and conflict- engages men and boys, engages community affected adolescents in Egypt, Ethiopia, Iraq, Jordan, leaders, and conducts specific activities around Kenya, Lebanon, Nepal, Rwanda, Syria, and in other FGM/C and child marriage. regions of Somalia. For most programs in these con- • The Right to be a Girl, Jordan (2018–2019) is texts, child marriage is a secondary objective or fo- an ongoing program supported by Save the cus under child protection. However, four programs Children Spain which utilizes empowerment based in Jordan and Somalia feature child marriage approaches with married and unmarried ado- as a primarily objective (Appendix 7). lescent girls, involves communities in trans- • Going Against the Grain, Somalia (2013– forming behaviors, and conducts advocacy 2014) aimed to address child marriage using around child marriage in Za’atari camp to a social-ecological model with a strong enhance girls’ resilience and wellbeing.

30 | Toward an End to Child Marriage Key Informant Interviews implemented by the More than Brides Alliance com- Semi-structured key informant interviews were con- prised of Oxfam, Simavi, Population Council, and ducted with practitioners both within and outside of Save the Children in five countries (Case Study 3). Save the Children. Researchers identified common In addition, the five-year CHANGES program im- themes within the following areas: 1) strategies and plemented by Care, IRC, and Save the Children in promising practices, 2) implementation challenges 15 districts in Somalia addresses discriminatory and gaps, 3) considerations for humanitarian set- gender norms at the structural, cultural, and indi- tings, and 4) research priorities. vidual levels. In addition to advocacy and technical support to the government to strengthen child mar- Strategies and Promising Practices riage laws and policies, they implement community Multi-level Integrated Interventions dialogue sessions with men, women, girls and boys to discuss gender inequality and enhance civic en- When asked about promising or effective practices, gagement, gender training and rights education for most key informants discussed the importance of very young adolescents, and village saving and loans multi-level, integrated interventions that address the associations (VSLAs) and social and economic em- complex drivers of child marriage at the individual, powerment activities for women and girls. family, community, and structural (law and policy) Although evaluations are limited or not yet avail- levels. Through an ecological approach, these pro- able for these programs, informants highlighted the grams involve multiple activities, across various sec- importance of combining empowerment activities tors, aimed at achieving a set of common outcomes. for adolescent girls with efforts to create an enabling Given the complexity of these interventions, they family and community environment in which girls may require implementation through a partnership can freely communicate and advocate for their of multiple organizations. rights. Similarly, they also spoke about the need to There’s also a need for more integrated program- combine community mobilization strategies with ming. . . . child marriage isn’t an issue you can think national level advocacy efforts to strengthen legal about from one lens only. It is not only driven by frameworks and improve access to educational and poverty or a lack of access to education. It is about employment opportunities for women and girls. For multiple risk factors that come together and that example, the Life Steps program in Northern Nigeria are all driven by gender inequality and gender- included a school enrollment campaign to encourage based discrimination. It is, therefore, important to parents to send their daughters to school as well as focus on integrated programming that places a state and national level advocacy to increase access to strong focus on gender equality and transforming education and strengthen laws and policies address- the discriminatory social norms that drive child ing child marriage. Similarly, through the Marriage: marriage. It is also important for all stakeholders No Child’s Play program in Mali and Niger, pro- to work together to see how we can complement gramming and advocacy are jointly implemented to each other, as opposed to trying to reinvent the mutually reinforce program objectives. wheel or do everything at once. Despite widespread acknowledgement among —Senior Advisor, Save the Children, Canada practitioners that multi-level, integrated approaches are necessary, they are not always feasible. Some Save the Children’s largest and most comprehen- informants involved in their implementation ex- sive example of a multi-level, integrated program pressed that these interventions are very challeng- is Marriage: No Child’s Play, a five-year program ing to coordinate and implement well given their

Toward an End to Child Marriage | 31 complexity, are expensive, and can be difficult to Working with the communities, consulting com- evaluate. Others mentioned that, given resource munity members, consulting community groups, constraints, it is important to simplify interventions religious leaders, all different actors in that commu- and understand which components are most effec- nity is really what’s been working for us. When the tive. A few informants noted that it is critical to keep community doesn’t adhere to your ideas, there’s no girls at the center of programming and to prioritize way you can get results. When the community is on the majority of resources for building their knowl- board with the vision, you can make progress. edge, confidence, and capacity. —Practitioner, Save the Children, Niger

One final thing, you’ll hear about integrated Informants described key characteristics of suc- “multi-faceted” approaches, but I think we have to cessful norm change interventions. First, community sort this out. We need to make the case. What do mobilization strategies should be long-term, going we mean? And what is an integrated approach? well beyond typical one- to two-year child marriage What is the evidence base it works better? We need prevention projects in order to shift entrenched so- to be more explicit about what works and how. cial norms. Second, they should engage a broad range —Practitioner, Save the Children, Malawi of actors within a community—including teachers, For me, I think the programs need to really focus police officers, healthcare providers, local govern- on getting resources to the girls and increasing ment officials, religious leaders, men and boys, and the status and the resource base for the girls. A lot adolescent girls—in order to reinforce key messages of times when we talk about child marriage, they and build a movement for social change. Informants say, ‘Oh we have to do it with boys and men and noted that identifying opinion leaders within a com- community leaders.’ I think that’s important, but it munity, particularly religious or other community should not be the central focus of the program. You leaders, to act as champions for ending child marriage can still reach out to men, religious leaders, commu- and promoting women and girls’ rights is critical. nity leaders and pass your message, but the bulk of Some informants discussed the need to identify your programming resources should be directed at and sensitize the “owners” of child marriage, who girls, otherwise, nothing will ever change. are not always parents, but may be grandparents, —Practitioner, NGO, Ethiopia mothers-in-law, aunts, or uncles in a given context. Other practitioners engaged and trained “providers” Community Mobilization to Change Social Norms of child marriage. For example, in Somaliland, Save the Children had success in developing a code of con- Until you can make a practice socially unaccept- duct and training religious leaders who perform mar- able, no matter what the law is, it will continue. riages to recognize and oppose child marriages. In —Practitioner, Save the Children, Bangladesh Bangladesh, practitioners are working with marriage Community mobilization to change social norms registrars to identify false birth certificates and ensure was cited as the most important strategy to prevent that they do not perform marriages of girls who are child marriage by many key informants. Practitioners below age 18. Finally, community mobilization strat- highlighted the importance of investing in commu- egies should be generated and driven by community nity sensitization and engagement activities in order members themselves. Some practitioners explained to address discriminatory gender norms, attitudes, that they had the most success when they developed and practices, and change the culture of acceptance context-specific messages and strategies, in partner- of child marriage. ship with communities, which promoted local ownership and leadership in their implementation.

32 | Toward an End to Child Marriage What we did in Somalia is we went around and to program implementation. Given the diversity of asked various people in the community, ‘Who do the drivers of child marriage across settings, con- you listen to most?’ And we compiled the most ducting an in-depth assessment to understand the frequently mentioned people and those were the complex individual, family, community, and policy- people we considered our opinion leaders. Then level drivers is essential to informing the design or it’s important to change those people and get adaptation of effective child marriage interventions. them on board because those are the people most Assessment findings can help practitioners to more people said they listened to. effectively target interventions, develop communi- —Practitioner, Save the Children cation strategies, and identify the influential opinion leaders to bring on board. Several informants within and outside of Save the Information about social norms and practices Children reported that promoting dialogue and dis- within a community at baseline also provides practi- cussion on issues related to child marriage and girls’ tioners with the foundation for developing an effec- rights is essential to changing social norms. Save the tive theory of change. Some informants explained Children’s work to prevent child marriage in Nepal that this evidence is rarely available and that theories offers a strong example of community dialogue and of change often rely heavily on assumptions. Several mobilization to change social norms. Program staff informants noted that going forward, it is essential work with the police to increase awareness of the that adequate time for a formative research phase, law banning child marriage and develop prevention ranging anywhere from a few months to a year, is strategies. They also work closely with district child built into project periods. welfare authorities and religious leaders to address child marriage and promote dialogue on the issue We need to look into causal relationships and among teachers, parents, and community members. theories of change and make sure that we have an Finally, through child clubs, girls and boys engage evidence base before we charge in with our inter- in social activities, training, and discussions re- ventions. Sometimes we’re not even talking to or garding leadership skills and children’s rights and targeting the right people. implement prevention activities in their own com- —Practitioner, Save the Children, munities. According to a staff member overseeing West and Central Africa work in the region, the prevalence of child marriage Formative research also provides program staff has declined in some districts where they are imple- with the opportunity to engage with community menting community-based prevention activities. members for a period of time prior to program What I’ve found is that it’s best to work with com- implementation. This engagement can also serve munities to identify solutions from within. Ensure to build the trust of various stakeholders, such as that community members are deciding the solutions religious and community leaders, parents, and ad- as opposed to approaching them with ‘what works.’ olescent girls, which informants said is essential to —Practitioner, Save the Children, South Sudan program attendance and uptake. One practitioner stressed that at least two months of activities should Formative Research to Inform be conducted by field staff in a location to build trust Intervention Design before implementing any intervention. A participa- Key informants from Save the Children and other tory community assessment process helps to ensure organizations emphasized the value of conducting that interventions are aligned with the needs and quality formative research on the context-specific preferences of the community and designed in part- drivers of child marriage within a community prior nership with community members.

Toward an End to Child Marriage | 33 Framing the Issue by political and community leaders, and ease barri- Framing the issue of child marriage and related inter- ers to girls’ participation in program activities. ventions strategically, according to social and political sensitivities in each context, is essential to gaining Advocacy Strategies and Lessons Learned community support and overcoming barriers to im- Save the Children has implemented a broad range of plementation. For example, informants explained the advocacy initiatives and campaigns to strengthen or different framing required in various locations for their introduce new laws, policies, and coordinated efforts work on child marriage in West and Central Africa. In to reduce child marriage in Asia, Africa, and Latin the Hausa Islamic communities in Northern Nigeria, a America. Several key themes emerged from discus- conservative cultural context, it can be challenging to sions of successful advocacy strategies. First, infor- discuss girls’ rights and child marriage. For this reason, mants spoke of the need to develop strategic partner- Save the Children program staff have framed most of ships with high-level government officials and political their child marriage prevention work around promot- leaders in order to build political support and create ing girls’ education. By contrast, in Senegal, staff are structural change. For example, Save the Children able to have open discussions about the stigma associ- works closely with the Vice President’s office and state ated with early pregnancy as a driver of child marriage governors in Nigeria, the Prime Minister’s office in and brainstorm ways to improve access to contracep- Senegal, and the First Lady’s Office in Burkina Faso, tives for adolescent girls. Mali, and Sierra Leone. Although evaluation of advo-

In Mali, once girls’ groups were formed, we decided cacy efforts is extremely limited, informants reported to start with life skills and economic empowerment that partnerships with, and the endorsement of, polit- because it’s less sensitive, and then we’ll follow ical leaders have helped to push policy and legislative with the sexual and reproductive health and rights changes forward, develop coordination structures to component. So, we’re trying to find entry points facilitate multi-sectoral work on the issue, and increase and angles that are more acceptable to the context public awareness and support. and to the people involved. Second, successful advocacy strategies should en- —Practitioner, Save the Children, Netherlands gage youth and ensure their voices are at the center of campaigns. Informants explained that adolescent In Niger, Save the Children has engaged the girls have a platform to speak out and to advocate for Ministry of Interior in addressing child marriage themselves and their peers. In Mexico and Guatemala, by framing the issue as a matter of national secu- for example, Save the Children has worked to engage rity. After the minister expressed concerns about adolescent girls as champions in advocacy coalitions the increasing number of forced marriages in areas to ensure that their voices and priorities are reflected affected by violence by Boko Haram, program staff in advocacy initiatives. In Bangladesh, in the recent, viewed this as an opportunity to make the case for multi-agency advocacy campaign against the new pro- strengthening national efforts to prevent child mar- vision allowing for child marriage in special circum- riage nationally. In addition to gaining high level politi- stances, Save the Children brought a focus on children, cal support, this Ministry also oversees the Department emphasizing the need to listen to children’s views on of Religious Affairs, which provided Save the Children the issue and enable their participation in legal and staff with the opportunity to address child marriage policy decisions affecting their lives (Case Study 4) among religious leaders at key forums. Child clubs are a common model for engaging Informants noted that the appropriate framing of youth in children’s rights discussions and advocacy im- child marriage interventions can improve the feasi- plemented by Save the Children in several countries. bility of implementation, generate greater ownership Through these clubs, girls and boys engage in social

34 | Toward an End to Child Marriage CASE STUDY 4

New Child Marriage Laws in Bangladesh and Lessons Learned from Advocacy

Background—The “Special Circumstances” Pro- circumstances provision cannot be applied in vision: Bangladesh has one of the highest rates of cases of rape. Save the Children will focus next on child marriage in the world with over half of girls raising community awareness of the regulations (59%) married by their 18th birthday.64 Child mar- through briefing papers and information sessions riage is prevalent in both rural and urban settings, to ensure that community leaders, parents, and driven by cultural and religious beliefs, poverty and children understand what is stated in the new law SGBV, and dowries that often increase as girls get and regulations. older.65 While child marriage rates have gradually Through more than two years of sustained declined in recent years, progress has been slow. advocacy, Save the Children was also able to In February of 2017, the Bangladesh Parlia- develop strong partnerships that will lay the ment passed a bill that allows for a new provision foundation for future collaborative advocacy work in the Child Marriage Restraint Act, Bangladesh’s and increased public awareness of the issue and 1929 law against child marriage. The provision consequences of child marriage. allows girls under 18 to marry in “special cir- We had a true spirit of solidarity and activ- cumstances” when parents or guardians obtain ism. So, for me, that was a huge gain. I al- permission from a court. Prime Minister Sheikh ways view rights work as a journey. It’s the Hasina was supportive of the provision, arguing friendships you develop, the networks, that it would protect girls who become pregnant the solidarity that will help you in future before age 18 from stigmatization and social movements also . . . and the discussions isolation, particularly in more conservative, rural we were able to generate on children’s areas. Women and children’s rights organizations rights, girls’ rights not to be discriminated argue that the provision is a major step backward, against, valuing girls’ dreams and realizing as the undefined “special circumstances” allow their potential, [and] the status of women for a broad range of exceptions within a law that and girls. I think those conversations are previously set the minimum age of marriage for valuable to society. You have to have girls at 18. those conversations. If it was not for this Advocacy: Save the Children, in partnership with advocacy, we wouldn’t have discussed the international and local women and children’s rights issue this much. organizations, engaged in more than two years of —Practitioner, extensive, sustained campaigning and advocacy Save the Children, Bangladesh against the new provision. They implemented advocacy activities at various levels, organized Lessons Learned: protests in local districts, developed position papers • Combine broad public awareness and for national government committees, held com- mobilization strategies with technical munity dialogue sessions, and engaged the media. assistance to the government. Throughout the campaign, Save the Children promoted a “child first” lens, placing children’s voices • Collaborate with women’s rights organizations at the center of advocacy efforts to ensure their and networks to strengthen and expand views and needs were represented. advocacy efforts. Outcomes: Although their advocacy was not • Engage the media to generate discussion and successful in preventing the new law’s passage, increase public awareness. Save the Children continued to engage with the • Following an advocacy initiative, regardless government on this issue. They served as one of of the outcome of that initiative, continue to seven members to draft the regulations for the support and provide technical assistance to new law which focused on limiting the use of this the government to mitigate further protection provision and ensuring additional safeguards are risks and strengthen legal and policy in place. For example, advocates were concerned frameworks addressing child marriage. that the provision could be used to force girls to • Engage youth and ensure their voices are at marry their perpetrators in cases of sexual violence. the center of campaigns. The regulations, which were being finalized at the time of writing this report, state that the special

Toward an End to Child Marriage | 35 activities, receive education and training in repro- the constitutional court to harmonize domestic leg- ductive health and rights, leadership skills, and child islation with international standards, and effectively protection issues including child marriage. With these engaging the media. skills, club members engage in advocacy and awareness raising efforts in their own communities. According to Strategies Through Other Sectors an informant in Nepal, child clubs have become very Ending child marriage is not a child protection active in both advocacy and prevention efforts, acting objective, it’s a foundational intervention. If we as vigilant groups in their communities and even inter- don’t raise the age at first pregnancy or ensure that vening and talking to parents when they learn of girls women have better birth spacing and that they at risk of marriage. Despite numerous positive experi- have the education to care for their children and ences shared by Save the Children practitioners, one start reducing their fertility rates, then we’ll never informant from an outside NGO mentioned that child address the issue of malnutrition which is a huge clubs can compete with homework time and domestic problem in this region. work at home, and may be challenging to attend for —Practitioner, Save the Children, girls who travel long distances home from school. For West and Central Africa these reasons, they may be more appropriate and accessible in urban areas. When asked about child marriage interventions, Third, informants highlighted collaboration with many key informants mentioned that although women’s rights organizations as a critical strategy for they are not implementing targeted child marriage strengthening advocacy efforts and maximizing impact. projects or programs, they are addressing the issue through programming in the education and health What I have noticed is that working with wom- sectors or through SGBV programming more gen- en’s rights organizations on issues related to girl erally. In particular, efforts to increase access to children is really important, because in this advo- education and incentivize girls to stay in school cacy we really collaborated very well with women’s as long as possible were discussed as critical child rights organizations and networks, and that made marriage prevention interventions. For example, the advocacy stronger, because women’s rights Save the Children’s Adolescent Girls Initiative— have a longer history than child rights in the coun- Kenya (AGIK), implemented in collaboration with try and they are stronger as a movement. Population Council, focused on reducing barriers —Practitioner, Save the Children, Bangladesh to education for girls by providing SRH education Practitioners in Latin America and South Asia and a household-level cash transfer. In addition, the expressed that Save the Children’s current advocacy Keeping Girls in School program implemented in 10 partnerships with civil society organizations address- districts in Malawi provided cash to households as ing women’s rights are limited and highlighted this an incentive to keep girls in secondary school. The as a priority area for further development. Other program also aims to improve the learning environ- strategies and lessons learned regarding effective ment for girls by training female teachers to serve advocacy campaigns included combining policy as positive role models and to more effectively en- advocacy with technical support to the government gage girls who are at-risk of dropping out. Although to implement changes, simplifying legal terms to often referred to as education projects rather than develop clear messaging to mobilize the public, us- child marriage interventions, efforts to ensure the ing both prevalence data and qualitative stories to safety of girls both while traveling to and at school; convey urgency to policymakers, filing a petition in strengthen legal frameworks to guarantee access to

36 | Toward an End to Child Marriage education until the age of 16; integrate child mar- to incidents. Some informants noted that measur- riage, gender, and girls’ rights into standard educa- ing child marriage as an outcome when evaluating tional curriculum; and invest in improving the qual- programming in other sectors would enable them to ity of education were all cited as promising child better identify promising practices for prevention. marriage prevention strategies. Implementation Challenges and Gaps In recent times, we’ve been looking at the girl child, An Integrated, Coordinated Strategy and for us it’s all about keeping girls in school for those 12 years, developing their aspirations, making Several informants spoke of the need for a holis- sure that they understand that there’s a world out tic, coordinated strategy to address child marriage there and so exposing them to role models and across sectors, both internally at Save the Children helping them step into leadership roles. We arrange and with other organizations at the country level. what we call “takeovers” to build aspiration and Within Save the Children, coordination at the leadership, so a large part of our programming now global level is limited as there is not currently a cen- turns toward the girl to advance gender equality. tral coordinating body for work on this issue. Child —Practitioner, Plan International, Bangladesh marriage is addressed by both child protection and gender; however, programming within each sector/ Save the Children’s programming through the area is often disconnected from that within other health sector may also serve to prevent and respond sectors. One informant explained that there is often to child marriage, particularly in contexts where a lack of a gender perspective in child protection early pregnancy is a primary driver. For example, programming. Another practitioner mentioned that through the Marriage: No Child’s Play program in sexual violence is addressed by child protection, Niger, program staff train local nurses and midwives while SGBV is addressed by gender, which further to provide youth-responsive SRH services and sup- inhibits coordination. There is also a disconnect be- port health centers with necessary equipment and tween advocacy and programming efforts. Other supplies. In the Somaliland region of Somalia, they informants emphasized the need to more effectively provide payments to enable girls with pregnancy mainstream child marriage prevention and response complications to go to the hospital. Through their across sectors. SRH program in the Northern Region of Ethiopia, child marriage and other SGBV issues have been Child marriage just isn’t something we’re main- integrated into their awareness-raising activities for streaming across our thematic areas. Child marriage in- and out-of-school youth, parents, and commu- and reproductive health and rights are critical com- nity members. ponents under every heading, and we miss oppor- Finally, informants from Save the Children in tunities for this to be part of the work we’re doing. some locations with a high prevalence of child —Practitioner, Save the Children, Canada marriage, such as Bangladesh and Ethiopia, men- At the country level, there are often disparate efforts tioned that although they do not currently have to address child marriage through health and nutrition, active child marriage projects, child marriage is ad- education, child poverty, child protection, and child dressed through their broader child protection and rights governance thematic areas. Informants spoke of SGBV prevention and response mechanisms such the importance of developing an integrated child mar- as community-based child protection committees riage strategy focused in a particular geographic area comprised of community members who implement and bringing together various components of this work. activities to raise awareness of SGBV and respond

Toward an End to Child Marriage | 37 A challenge is for Save the Children to have this concerns and more by poverty, which may lead to kind of holistic approach. We tend to think of it transactional sex and pregnancy, and result in many [child marriage] in thematic areas… but we’re not of these girls dropping out of school and marrying very strong as an organization in having a more for economic reasons. Informants working on child integrated and holistic approach, which is really marriage issues in Malawi, Sierra Leone, and Latin what’s most needed in addressing this issue. America explained that early pregnancy is a primary —Practitioner, Save the Children, Asia driver of child marriage due to the stigma associated with pregnancy outside of marriage, making In addition to a lack of a common strategy with access to SRH services and information a primary joint objectives and outcomes, the absence of com- form of child marriage prevention. In areas of El mon terminology or language regarding child Salvador, Honduras, and Guatemala affected by gang marriage across the Save the Children movement violence and organized crime, girls are forced by further inhibits coordination at both global and na- gang members to get married to them and may be tional levels. threatened or killed if they refuse. Addressing the Developing a holistic approach at the country diverse drivers of child marriage in each of these level, in collaboration with other organizations, was contexts requires unique and tailored prevention also cited as a promising practice. Recognizing that interventions. addressing the complex drivers of child marriage at I think because the drivers are very different, it’s multiple levels is not feasible for any one organiza- very difficult to say ‘this works!’ Some things work tion alone, informants spoke of the need to partner in some circumstances… Really what we need to with other organizations to implement different have is the whole list of all the different things you components of a multi-level intervention in a given could potentially do and under what kind of cir- location in order to maximize impact. Coordination cumstances do these ones work best. is often challenging given that organizations have —Practitioner, Save the Children different agendas and priorities. Establishing a community of practice to address child marriage at the country level can help to prioritize and move pre- Norm Change and Resource-Intensive, Longterm vention and response work forward. Interventions Many informants explained that changing social Diverse and Context-Specific Drivers norms at the root of child marriage requires long- One of the most commonly cited challenges among term interventions. However, this need does not practitioners was the diversity of drivers of child align with the typical model of short-term, donor- marriage and the need for tailored, context-specific funded projects. interventions. Child marriage is driven by a com- Social norms that shape gender roles and pro- bination of cultural, social, economic, and struc- mote acceptance of child marriage are deeply en- tural factors that vary widely by location, making it trenched and require extensive time and resources challenging for practitioners to draw on a relevant to change. For example, one informant spoke about evidence base and to develop replicable, scalable the dowry system in the Terai region of Nepal in interventions. For example, in Pakistan, India, and which the dowry becomes higher as the girl ages, Bangladesh, drivers are often related to concerns incentivizing her parents to marry her off earlier. In about girls’ virginity and the need to preserve fam- this context, despite laws, policies, and a national ily honor. In some areas of West Africa, informants strategy to end the practice, early marriage is widely noted that child marriage is driven less by morality accepted as the norm, not only among community

38 | Toward an End to Child Marriage members, but also among many government repre- the Constitutional Court to set the marriageable age sentatives and NGO program staff. An informant in at 18 or above. Informants explained that the appli- Bangladesh emphasized that social acceptance that cation of customary law remains very strong. is so ingrained in society makes it extremely chal- In other contexts, despite a strong legal frame- lenging for families to choose alternatives to mar- work and widespread public awareness of the riage for their daughters. law, enforcement and compliance are limited. In Resource constraints were consistently cited as a Bangladesh, for example, although the law prohibits key challenge in the implementation of child mar- the marriage of girls below age 18, child marriage is riage interventions. The majority of child marriage widespread and parents may present false birth cer- programs and projects are implemented for one to tificates to circumvent the law. Similarly, one prac- two years due to short-term and inconsistent fund- titioner in Ethiopia explained that despite a clear ing streams. Although it was widely acknowledged penal code that prohibits child marriage, parents by practitioners that shifting deeply entrenched continue to arrange marriage ceremonies for their social norms required holistic, long-term interven- underage daughters even though they are aware of tions, sustained investment in the same communi- the legal consequences. Others emphasized that im- ties over time has rarely been feasible. proving the law is a necessary but insufficient step in ending child marriage. Another challenge is obviously that these kinds of interventions have to take place over a significant At the advocacy level, a legislative framework— amount of time, and our capacity to think long- a law banning child marriage—is important, but term about sustaining these programs for a long it’s very far from being enough to reduce child time is also a challenge. A lot of our funding is marriage. We have to focus on the community dependent on donors. Very seldom do donors say level. . . . We really need to address social norms they’ll support a program or project for 10 years. and give girls a voice to say it’s not right to be —Practitioner, Save the Children, Asia married before age 18. —Practitioner, Save the Children, Asia

Limited Legal Frameworks and Some informants spoke of the need to improve Enforcement of Laws the enforcement of existing child marriage laws. Limited legal frameworks addressing child mar- One informant highlighted that those involved in riage, as well as poor enforcement of existing laws, the marriage, including leaders of the faith-based were highlighted as key challenges in child mar- community, village chiefs, or parents of children riage prevention and response. In some contexts, should be subject to prosecution to send a clear and laws addressing child marriage are vague or conflict consistent message to others that child marriage with customary law. In South Sudan, for example, will not be tolerated. He explained that state courts, informants explained that there is currently no di- rather than customary justice mechanisms, should rect legislation addressing child marriage and the handle all child protection cases. On the other hand, laws related to child marriage are not harmonized. another informant cautioned that holding parents The Child Act protects the rights of children; how- accountable under the law is too heavy-handed and ever, customary laws are also recognized by the can be dangerous for girls. By encouraging girls to Constitution and can undermine protections set report the actions of their mother or father, they forth in the Child Act. To address this inconsistency, may be at risk of being thrown out of their house Save the Children, in partnership with national or losing the breadwinner of their families if their NGO Advocates without Borders, filed a petition in father is arrested. Strict enforcement of the law may

Toward an End to Child Marriage | 39 also “drive the practice underground,” making it ministers are fragile and not really empowered. . . . more difficult to reach at-risk or married girls with So, you have ministers who have very limited ca- support services and information. pacity, scale, and resources to be able to operate, and they don’t have the political gravitas to put Advocacy Challenges this as a priority theme in the political agenda of In the area of advocacy, informants explained that the country. it is often challenging to translate national level law —Practitioner, Save the Children, and policy successes into concrete changes at the lo- West and Central Africa cal level. Finally, informants noted that methodologies and We have a lot of information on the legal frame- tools for evaluating advocacy efforts are needed. work, and we’re moving forward on that. The prob- Advocacy efforts and campaigns are rarely rigor- lem is how to bring these regulations and this infor- ously evaluated, and without understanding what mation of the law to lower levels of policy making, makes an impact, practitioners struggle to deter- especially when we have countries where the mine how best to invest time and resources in ad- policy making is very decentralized in terms of vocacy activities. education policies or health policies or decisions on where to provide sexual and reproductive health Implementation Gaps: Marginalized Communities, services for whom, when, etc. Faith-Based Leaders, Men and Boys, and —Practitioner, Save the Children, Married Girls Latin America and Caribbean When asked about key gaps in programming, informants highlighted marginalized communities, lead- In addition, it is important for advocacy strategies ers of the faith-based community, men and boys, to find ways to target powerful stakeholders who and married girls as important groups not suffi- are blocking change. For example, in Bangladesh, ciently reached by Save the Children’s interventions. despite extensive advocacy against it, the new legal Practitioners commonly cited the urgent need to provision allowing for the marriage of children un- engage leaders of the faith-based community in child der age 18 in certain circumstances was eventually marriage prevention and response. Religious leaders passed by Parliament, largely due to strong sup- are instrumental in shaping public perceptions and port of the Prime Minister (Case Study 4). In Latin norms in many communities, particularly regarding America, one informant noted that strong religious marriage ceremonies and other cultural traditions. and patriarchal norms and powerful conservative Yet despite recognizing this group as a priority, some groups are a challenge to advancing SRHR. informants explained that they do not know how to Further, limited capacity and resources of the engage them effectively in child marriage preven- relevant government ministries and the judiciary tion efforts. For example, one informant working in to develop and implement child marriage laws and West Africa explained that getting faith-based lead- policies was cited as an advocacy challenge in sev- ers on board can be time and resource-intensive, eral contexts. particularly in religions that are more decentral- One of the lessons learned from the high-level ized. In the Sahelian cultural tradition, for example, meeting we had . . . in Dakar is really that child Islam is decentralized and program staff need to “go marriage very often is a topic which is under the re- imam by imam to start a movement,” one leader at a sponsibility of the Ministry of Family, Ministry of Social time, which can be difficult with limited resources. Affairs. Too often in West and Central Africa those Practitioners in Malawi and South Sudan explained that it can be very difficult to bring leaders of the

40 | Toward an End to Child Marriage faith-based community fully on board because they provide shade to women waiting outside of health are often invested in upholding the status quo and facilities. In Bangladesh, Save the Children program view carrying out these marriages as a source of staff found that by facilitating community-based employment and income. Several practitioners said discussion groups with men and boys about gen- that they would like to learn evidence-based strate- der and SGBV, their attitudes related to dowries and gies for working with the faith-based community in child marriage changed. order to expand their work with this population. I also think that the biggest gap exists when we Unfortunately, even though everybody has iden- say that we want to empower girls. We com- tified the traditional and religious leaders as some pletely forget the other part which is working with key actors of change, nobody has been seriously males. . . . All the things that she’s learned, she can’t engaging with them in our day to day work. I strongly really apply because there’s no response from the believe in the region that there is a huge opportunity people that she’s surrounded by. to start engaging them to try to make a difference —Practitioner, NGO, Lebanon because they’re the ones at the end of the day that Practitioners in Asia, Africa, and Latin America influence the behavior of the people in our country spoke of the challenges in reaching ethnic and reli- through TV shows, radio shows, through sermons, gious minorities and other marginalized adolescent etc. We believe if we can get them with us, we can girls. For example, among Dalit communities in make a difference. If we can’t, maybe we will have a Nepal and India, child marriage rates are high, yet bottleneck we won’t manage to unlock. practitioners have a hard time reaching these remote —Practitioner, Save the Children, communities. In Latin America, practitioners strug- West and Central Africa gle to reach indigenous adolescent girls in very rural Men and boys are rarely meaningfully engaged areas where child marriage is widely accepted by the in child marriage prevention and response inter- community and is not a priority for local govern- ventions, which practitioners viewed as an import- ments. In Ethiopia, one practitioner mentioned that ant gap to address moving forward. Informants in the Western part of the country, communities are explained that interventions to empower girls and very traditional and adolescent girls are at high risk change social norms often neglect to involve men of child marriage. However, successful implementa- and boys entirely. Without understanding the mo- tion of child marriage prevention interventions in tivations and experiences of men and boys involved these areas would first require extensive preliminary in early marriage, and bringing men and boys into work to educate the community on basic concepts of norm change efforts, progress is stalled. While many gender, power, and women’s rights, which requires practitioners felt that they were not working enough more time and resources than are usually available. with men and boys on this issue, a few offered ex- In several contexts, practitioners expressed the con- amples of successful strategies. For example, in cern that although Save the Children aims to tar- Niger, through “husband schools,” groups of men in get the most marginalized adolescents as an orga- the community receive training and then sensitize nizational priority, they are often not the target of their peers on topics of gender equality and wom- interventions. en’s rights and serve as role models in the commu- Practitioners also highlighted the gap in pro- nity. According to one informant, with enhanced gramming for married girls, as the majority of child understanding of SRH, men are working to facili- marriage efforts focus on prevention rather than re- tate, rather than inhibit, their wives from accessing sponse. Married girls are often difficult to reach with health services. For example, they built structures to adolescent support services because they are less

Toward an End to Child Marriage | 41 visible in schools and their communities. However, as we could. There is a huge amount of potential because they are often socially isolated, providing there, and maybe we’re actually doing ourselves them with life skills programming and opportu- a disservice by having this issue specifically under nities to meet with their peers is critical. Reaching child protection when it actually cuts across the- them with SRH information and services should matic areas. also be a priority, as they are at high risk of com- —Practitioner, Save the Children, International plications related to early pregnancy. Married girls Considerations for Humanitarian Settings may also face heightened barriers to participation in programming. In India, for example, one infor- Given the limited literature addressing child mar- mant explained that married girls were often pro- riage in humanitarian settings, interviews were hibited by their husbands from attending discussion conducted with practitioners who have worked, groups. In addition, it can be challenging to gain or are currently working, in humanitarian settings support for child marriage response as a priority in Bangladesh, Ethiopia, Iraq, Lebanon, Uganda, both at community and organizational levels. Once Nepal, Niger, Nigeria, Somalia, and South Sudan. girls are married, they are often viewed by their Observations regarding the nature and drivers of communities as women and their participation in child marriage, intervention approaches, and imple- child marriage programming or other adolescent mentation challenges in humanitarian contexts are support programming can be viewed as contradic- summarized below. tory. Focusing on programming for married girls Nature and Drivers of Child Marriage may also be seen as condoning or accepting child marriage. Informants mentioned that although the Practitioners working in humanitarian settings con- focus of Save the Children’s child marriage work has sistently reported that child marriage is driven by been on prevention, they are well-positioned to also acute economic and protection needs in crisis. Many address response through their work with adoles- informants discussed child marriage as a form of cents across thematic areas. coping with extreme poverty and financial instabil- ity. Even where child marriage is not a normative Unfortunately, girls wanted to have children in cultural practice, parents marry off their daughters order just to have their marriage registered. So this, because they simply cannot provide for them. for me, is like what are we doing? . . . If you are responding to child marriage, you aren’t reinforc- When you ask them, ‘If you hadn’t been living in ing child marriage, you are recognizing the reality this situation would you still consider marrying off and reaching out to vulnerable communities and your daughter?’ They say, ‘No, I think I would much providing the services that are needed. prefer that she would go to school, get an educa- —Practitioner, NGO, Lebanon tion, finish her education, and then get married. But under these circumstances, we don’t have money to I think it [response] falls really firmly under our support our children. We don’t have money to send mandate . . . we’re doing programs globally them to school, and things are out of our control.’ around livelihoods, life-skills, and different types —Practitioner, NGO, Lebanon of secondary education or alternative vocational learning. Through our health piece, we are do- Protection concerns are also a primary driver ing work around sexual and reproductive health of child marriage in displacement contexts where and rights . . . Overall, I would say, it’s really not a traditional family and community support systems stretch for us, it’s just connecting the dots—that is have broken own. In South Sudan, for example, one something that we haven’t necessary done as well informant observed that when families flee conflict,

42 | Toward an End to Child Marriage children are exposed to new protection risks be- Interventions cause they are separated from their families, often Most Save the Children practitioners interviewed lack proper accommodation, and are increasingly reported that they were not implementing or aware vulnerable to sexual exploitation and abuse. One of child marriage-focused projects or program- practitioner in Lebanon explained that refugee girls ming in humanitarian settings. However, many are at increased risk of sexual violence, and child mentioned that child marriage should be addressed marriage often takes place to avoid the shame and through their child protection strategy and inter- stigma associated with sexual violence or pregnancy. ventions. Child protection is consistently a major Informants who had experience working with IDPs component of Save the Children’s response efforts, in Iraq, Rohingya refugees in Bangladesh, and IDPs as program staff work to identify children at risk in in Nepal during the 2015 earthquake response ex- humanitarian settings, assess their needs, and refer plained that many families married their daughters them to protection and support services. In several because they felt they would be safer and protected humanitarian contexts, such as Somalia, Iraq, and from sexual violence. Nigeria, Save the Children has established child pro- Others working with Syrian, Somali, and Rohingya tection committees comprised of community mem- refugee communities shared the observation that cri- bers who identify and respond to children at risk of sis exacerbates the pre-existing cultural drivers of SGBV and other protection issues. However, one child marriage. For example, in Bangladesh, an in- informant in Somalia pointed out that they are not formant explained that child marriage was already documenting the number of marriages prevented as practiced in Rakhine State in Myanmar, and that an outcome, and therefore do not know if this inter- crisis and displacement to Bangladesh simply inten- vention is working well in cases of child marriage. sifies this risk. In some cases, child marriage has been incor-

Child marriage was already very prevalent in the porated into community sensitization and aware- Rohingya community. So, what we’re noticing in this ness-raising activities regarding children’s and wom- humanitarian crisis is that you already have accep- en’s rights issues. For example, through a program tance of child marriage, and then in dealing with a to build resilience in conflict-affected areas of Iraq, crisis like this, it just makes the situation worse. Save the Children implemented community educa- —Practitioner, Save the Children, Bangladesh tion activities, many of which were led by children themselves. Child marriage was not initially part of this program; however, upon learning that child Informants generally spoke about child marriage marriage was a primary protection concern for girls, increasing in crisis. Acknowledging the lack of of- children and program staff worked to sensitize camp ficial data, practitioners working in Diffa, Niger, leaders and parents about the importance of pre- Northeast Nigeria, and Bangladesh shared anec- vention. In Somaliland, program staff sensitized the dotal evidence that child marriage was on the rise Ministry of Justice and developed a code of conduct in these settings. Inconsistent with these narratives, for traditional marriage providers to prevent child however, one informant in Somalia explained that marriage and educate key actors about the conse- child marriage had actually decreased throughout quences. In Bangladesh, program staff addressed the country, including in the drought-affected areas. the misconception among Rohingya parents that He noted that one possible theory for this decline in girls could not wear a hijab to school, which was a humanitarian context is that due to the prolonged preventing parents from sending their daughters to droughts, the ability of households to afford the school. According to one informant, working with costs of marriage has also declined.

Toward an End to Child Marriage | 43 parents to ensure that girls continue their education sectors through existing humanitarian response ef- is an essential child marriage prevention strategy in forts. Some informants explained that meeting the displacement settings. minimum standards in humanitarian response Child-friendly spaces, safe areas designated for across sectors and ensuring the holistic needs of children to engage in structured play, educational communities are met, rather than establishing par- activities, and psychosocial support were also cited allel child marriage programming, is the most im- as interventions that can address child marriage in portant way to prevent child marriage. This includes humanitarian settings. Save the Children imple- basic needs and identifying and supporting girls at ments this model in several humanitarian contexts. risk of child marriage through adolescent SRH pro- In these spaces, program staff address child marriage gramming, SGBV response efforts, and through the through adolescent SRH and SGBV education and education sector. training. In Bangladesh, one informant had found Some informants highlighted that nutrition/ it important to create adolescent-friendly spaces food security and livelihoods sectors offer key en- focused exclusively on offering protection, rights try points in humanitarian response, particularly education, and capacity building of adolescent girls. in the acute phase of an emergency, for addressing Regarding life skills curriculum addressing child child marriage. For example, informants working in marriage, International Rescue Committee (IRC) Bangladesh after the 2017 influx of Rohingya refu- has developed curricula for adolescent girls, includ- gees observed an increase in child marriage driven ing “My Safety, My Well-being” and “Girl Shine,” of- by a lack of food. They reported that parents married fering life skills education and coping strategies for off their children because they were unable to pro- humanitarian settings, both of which address the vide food for them. In this context, child marriage issue of early marriage. After recognizing the very prevention efforts took the form of increasing food distinct needs of married girls, IRC also developed a distribution. One informant from Save the Children tailored life skills curriculum for engaged, married, explained that the livelihoods sector offers critical and previously married girls. The life skills package opportunities to obtain information about house- is currently being piloted and evaluated in Lebanon. hold composition and identify children at risk of Finally, integrating child marriage and other marriage. For example, through the household eco- protection issues into cash transfer programming, nomic assessments and surveys conducted by food also referred to as child sensitive social protection, security and livelihoods teams in emergencies, child was also highlighted as a potential child marriage marriage has been identified as a coping mechanism prevention strategy in humanitarian settings. In to manage deteriorating livelihoods. Additional Somalia, for example, Save the Children is currently questions aimed at identifying child marriage as a piloting unconditional cash transfers as part of a coping strategy, which may be associated with the child sensitive social protection strategy that in- advent of a crisis or shock, could be integrated into cludes education and discussion sessions for parents emergency assessments. This information should on issues of child protection and rights. They are be used to guide targeted child marriage preven- evaluating the impact of the program on a range of tion programming through cash-based and other child protection outcomes. interventions. Discussions with key informants also reflected We have such enormous opportunities through an emerging debate as to whether child marriage the household economy analysis and through the prevention and response warrants specific, targeted coping strategies index and through much more programming or should be mainstreamed into other

44 | Toward an End to Child Marriage advanced and sophisticated tools that we have in In general, I think we’re well set up to do it in our the livelihood sector, but we’ve never fully utilized structure, mandate, and approach. As far as the them for monitoring protection issues and evaluat- kind of humanitarian response approaches, we do ing the impacts of our programming and developing a lot of work around child-friendly spaces, educa- integrated strategies. . . . When we register a house- tion, protection, and all of these areas that could be hold, we should be asking, ‘Are there any children such amazing entry points. We have the structure, who don’t normally live there?’ to identify separated we need the capacity and vision or approach. children. We can also ask if any of the family have —Practitioner, Save the Children, Canada recently left and why, potentially simple questions, Save the Children practitioners generally agreed to get a sense of which girls have gone. that the organization should do more to address —Practitioner, Save the Children, UK child marriage in humanitarian settings and is well- One informant emphasized that approaches to positioned to do so through their current response child marriage prevention and response in human- efforts in these contexts. itarian settings depends on the stage of the emergency. For example, targeted child marriage preven- Challenges tion programming aimed at changing social norms Key informants expressed that they are not currently may not be feasible or appropriate for the acute stage addressing child marriage in humanitarian con- of an emergency, but is important to introduce after texts because they do not know what programming the situation stabilizes. should look like and are not aware of evidence-based approaches. One informant explained that Save the I’m not saying social norms change doesn’t belong Children does not yet have standards or protocols in emergency contexts. But in acute emergency addressing child marriage in emergencies, which [the first 3–6 months], I wouldn’t advocate for a inhibits mainstreaming of child marriage into hu- parallel program where prevention/mitigation of manitarian response efforts. A practitioner in South child marriage is the ultimate goal, or that there’s Sudan highlighted the challenges in enforcing child a whole program focused on this end. I would marriage-related laws in settings where police and instead focus on making sure basic needs are met: judicial systems have been weakened by conflict. that girls are enrolled in education, that there are Further, data on the drivers, nature, and prevalence opportunities for girls to meet safely. . . . And those of child marriage is extremely limited in conflict- are those early footprints of what could potentially affected settings. Marriages may be less likely to be be longer social norms change efforts . . . that roll registered and therefore it can be more difficult to out six months into an emergency. track trends. In Latin America, research on child —Researcher, NGO, USA marriage drivers and interventions is particularly Aligned with this perspective, another informant limited. One practitioner discussed the need for ev- explained that displacement actually offers an im- idence to inform interventions in contexts affected portant opportunity for prevention programming as by gang activity and organized crime in El Salvador, gender norms and expectations can change rapidly Guatemala, and Honduras. in humanitarian contexts where traditional commu- The short-term nature of humanitarian funding nity structures may have broken down and people cycles also poses a significant challenge to addressing are in a new environment. child marriage, which often requires long-term investment in activities intended to target deeply-rooted drivers such as social norms.

Toward an End to Child Marriage | 45 To create such big change like reducing child More rigorous evaluations of child marriage in- marriage, it’s a long-term type of work. And it is terventions, built into program designs from the also an issue that needs to be addressed across the start, were commonly cited research priorities. In social-ecological model and over a long period particular, informants highlighted the need for im- of time. It is a challenge for sure in humanitarian proved methodologies to evaluate the long-term contexts because it’s very rare that we have funding impact of advocacy campaigns and initiatives on that goes beyond a year because of the nature of the reducing child marriage to inform future resource context in which we’re working. And that’s why it’s allocation. Measuring rates of child marriage as really hard to do the groundwork we need to do to an outcome in evaluations of efforts through other really have a meaningful and long-term impact on sectors such as health, education, and livelihoods child marriage in humanitarian settings and fragile programming would also be useful in building the contexts. evidence base for prevention. —Senior Advisor, Save the Children, Canada Other research priorities in development contexts included: evaluations of interlinked interven- Finally, practitioners in humanitarian settings ex- tions addressing child marriage and teen pregnancy, plained that insecurity creates challenges in implemen- understanding how child marriage cases are han- tation and access to program sites. In South Sudan, for dled by formal and customary justice systems, the example, practitioners explained that despite weeks of impact of social media on the nature of child mar- planning, they could no longer carry out advocacy ac- riage, and improving marriage registration data and tivities and community dialogues in some communi- registration systems. ties because of conflict-related violence. Practitioners emphasized that child marriage interventions in cri- Humanitarian sis settings need to have flexible time frames and be When asked about research priorities, practitioners adaptable in their design and delivery. in humanitarian settings consistently reported on the need for piloting and evaluating new interven- Research Priorities tions to prevent and respond to child marriage. Development Intervention research is needed to build the evi- Noted earlier as a promising practice, informants high- dence base and establish what works in different lighted formative research on context-specific drivers stages of an emergency. of child marriage as a key research priority to improve Informants also discussed the need for research child marriage interventions. This foundational re- on the decision-making factors of parents to marry search, which should be built into project periods, off their daughters in humanitarian settings. They recognizes the diversity of drivers across contexts also requested information on the support needs of and allows for tailored interventions. It also helps to families and adolescent girls that may put girls at risk ensure that interventions are aligned with the needs of child marriage. One informant explained that re- of community members and establishes trust among search should focus on understanding the benefits of stakeholders. One informant stated that a standard child marriage from the perspective of families. Doing community assessment tool to help practitioners to so would help humanitarian actors to identify mea- collect data on the drivers of child marriage, under- sures to offset these incentives through humanitarian stand perceptions and motivations of communities, response efforts such as food distribution, cash-based and identify influential community members may be interventions, or protection programming. useful in designing more responsive interventions.

46 | Toward an End to Child Marriage We would like to study those choices that parents Practitioners in some humanitarian contexts re- are making now and whether those choices to marry quested research to better understand trends, as well off their daughters are influenced by economic as factors contributing to the rise or decline of child livelihoods, the need to make ends meet, or is it the marriage in emergencies. For example, in Somalia, same that influenced them before the conflict to one practitioner expressed that understanding why marry off their daughters? We want to draw out that child marriage has declined in some areas is critical distinction. to understanding how to design effective interven- —Practitioner, Save the Children, South Sudan tions and avoid negative, unintended consequences

There’s been a lot of research done on the negative of other interventions that may incentivize mar- economic impacts of child marriage and girls not riage. In addition, research on the nature and driv- finishing school . . . but what are the positive im- ers of child marriage in conflict-affected settings in pacts? If we see that as crisis coping, can we quantify Latin America is absent from the literature and ur- the positive impacts to the family of marrying off gently needed. their daughter?. . . . They might not want to do it and Finally, evidence-based, youth-driven program- might see it as the only option, but let’s get into the ming was cited as a priority need in crisis settings. detail there of why. Research to understand how to more effectively en- —Practitioner, Save the Children, gage adolescent girls and boys in programming in West and Central Africa humanitarian settings through peer-led activities, SRH services, and context-specific life skills programming is necessary.

Toward an End to Child Marriage | 47 DISCUSSION

Lessons from Research and Practice fell neatly into 1) empowerment programs, which This three-part study, including a literature review, focused on after-school and school enhancement key informant interviews, and a mapping exercise activities as part of a broader empowerment objec- of Save the Children’s programming and interven- tive, or 2) incentive programs, which sought to pro- tions, identified promising practices, programs, and vide incentives to encourage girls to stay in school. evidence gaps to address child marriage in both It should also be noted that other approaches, in ad- development and humanitarian settings. A total of dition to the four identified in this study, likely exist 43 evaluations were identified in the literature, in- and simply have not been evaluated for their impact cluding experimental and quasi-experimental eval- on child marriage or published. For example, com- uations. Findings from the literature combined with prehensive sexuality education (CSE) in schools findings from 27 key informant interviews help to may also help reduce rates of child marriage, as may outline what we know, and what we still need to youth-responsive sexual and reproductive health learn, to make headway on this critical issue. This services in contexts where early pregnancy is a key study adds to existing research by synthesizing the driver. Likewise, many programmatic approaches most up-to-date literature, bringing the voices of that effectively address drivers of child marriage, practitioners and experts into the discussion, and such as educational interventions for girls in low- examining the topic of child marriage interventions and middle-income country contexts, have not been in a humanitarian context for the first time. In ad- evaluated for their impact on child marriage. These dition, through a mapping and analysis of Save the may offer promising approaches outside of those Children’s programming, it assesses the organiza- identified thus far. tion’s approaches against the available evidence. What we know now is that empowerment ap- The review identified four approaches to child proaches are the most frequently evaluated, the marriage programming currently evaluated in the most likely to show positive impacts in those eval- literature: 1) Empowerment, 2) Incentive/asset trans- uations, and the most common approach within fer, 3) Community sensitization and engagement, the Save the Children movement. There are likely and 4) Legal and advocacy. These are similar to a number of reasons for this. These approaches are programmatic approaches identified by two other generally easy to implement on a small scale, rela- reviews;66 however, they differ in that both included tively inexpensive, and work not only to prevent an approach involving education. We felt that purely child marriage, but to equip adolescents, especially educational approaches were not represented in the girls, with skills to improve many aspects of their literature that fit our inclusion criteria. Rather, they lives from health to livelihood opportunities. There

48 are a few significant challenges to empowerment attempts as indicative of the usefulness of the ap- programming, however. First, such programs tend proach more broadly. This review did not find that to address a large range of topics and skills. This to be the case. Incentive/asset transfer programs makes it difficult to compare various program mod- were found to play an important role in improving els and to understand which portions of the pro- outcomes related to child marriage when such program are having the most significant effect. Second, grams are complementary to other approaches that these programs can be very difficult to scale up and address gender inequality (such as empowerment sustain due to expense, issues with quality man- or community sensitization and engagement) and agement, and the time and resources required to are focused on building adolescents’ human capital implement them successfully. In-school programs rather than simply incentivizing parents to delay such as CSE offer systematic, scalable approaches marriage until a girl’s 18th birthday. Again, practi- to achieving empowerment and may contribute to tioners are encouraged to carefully assess their local preventing child marriage, assuming government context before implementing cash programming, as support for CSE; however, these approaches may at least two practitioners warned that cash payments fail to reach vulnerable out-of-school adolescent to parents may actually increase child marriage by girls. Third, a few studies which separated impacts providing parents with the funds necessary for of empowerment programming for younger (10–14 young daughters’ dowries, ceremonies, and other year-old) and older (15–19 year-old) adolescents wedding expenses. found that empowerment approaches are less effec- The most significant difference between findings tive in reaching older adolescents, perhaps because from the literature and those from practitioners and the economic and social costs of delaying marriage other experts were with regards to community sen- in later years are much higher. Finally, there is a dan- sitization and engagement approaches. While prac- ger that empowerment approaches place the onus of titioners championed such approaches as the way preventing marriage on the girl herself, which may forward for child marriage programming, the evalu- put her at increased risk of violence or abuse if her ation literature on this approach was extremely lim- family or community are not supportive of her deci- ited. Most of the programs or program components sion. As noted by key informants, the content of em- in this category were poorly evaluated, described in powerment programs should be tailored to the lo- very little detail, and treated by many study authors cal context based on initial formative research, and as secondary to other approaches. There are a few complementary approaches that engage family and reasons why this may be the case. community should be implemented as appropriate • Changing social norms takes a significant to increase the chance of positive outcomes and de- amount of time. Given the funding cycles crease risks for adolescent program participants. on which most NGOs and civil society Other program types present their own unique organizations rely, a program and evaluation benefits and challenges. Incentive/asset transfer cycle may be too short to create or measure programs are particularly appealing because they any significant social norms change. are relatively easy, if not relatively inexpensive, to implement at a large scale (assuming the transfer is • Community programs vary with regards to in cash or easily obtained goods, rather than live- depth of engagement. While some programs stock or goods which are more difficult to procure held regular community meetings with and distribute). This approach has received a some- leaders and established subcommittees what negative response in at least one recent litera- focused on the issue, others consisted only of ture review,67 as practitioners point to a few failed a simple one-time meeting with parents or

Toward an End to Child Marriage | 49 short film screening. Short term, superficial and Gender-based Violence offers guidance, tools, engagement such as the latter may not be and practical examples to assist practitioners in di- sufficient to produce meaningful change. agnosing social norms.71 Legal and advocacy approaches were represented • It can be challenging to evaluate long-term the least in the evaluation literature and few key community sensitization and engagement informants shared learnings on successful advo- efforts through traditional experimental and cacy initiatives. This is likely due to the difficulty quasi-experimental methods. At least one of evaluating the effectiveness of such efforts -us study evaluating a mass media campaign ing experimental methods—several practitioners on child marriage in Ethiopia was excluded noted that tools are sorely needed. That does not, from this review because it lacked a valid however, negate the important role that such efforts comparison group—although results from the play in creating long-term, sustainable change with study were highly promising.68 regards to child marriage. For example, advocacy Community-level interventions to change social on child marriage laws can not only provide criti- norms aimed at preventing other forms of SGBV fell cal legal protection for children, but also establish outside of our review criteria but may offer import- a foundational commitment upon which national ant evidence for changing social norms around child strategies and action plans can be built and help to marriage. For example, SASA, a long-term commu- create a supportive policy environment with suffi- nity mobilization intervention aimed at reducing in- cient resource allocation. Research examining child timate partner violence (IPV), was rigorously eval- marriage laws in Sub-Saharan Africa found that uated in Uganda through a cluster randomized trial countries with consistent laws establishing 18 as and found to be effective in lowering social accep- the minimum age of marriage had 40% lower child tance of IPV, reducing physical and sexual violence marriage and 25% lower teenage childbearing preva- against women, and improving supportive commu- lence than in countries without consistent marriage nity responses.69 Practitioners shared that effective laws.72 Yet legal approaches have their limitations— norm-change interventions, from their experiences, joint research by Save the Children and the World must be long-term (more than two years), should Bank found that 68% of child marriages still occur in involve a broad range of influential actors in a given countries where the practice is illegal.73 Practitioners community, and should be developed and driven by interviewed in this study highlighted other advo- community members themselves. Many felt that for- cacy approaches such as increasing monitoring and mative research was a powerful tool for understand- policy implementation, strategic partnerships with ing the social determinants of child marriage as well high-level government officials and other influen- as for establishing the level of community trust and tial people, ensuring that the voices of youth are at buy-in necessary for achieving social norms out- the center of campaigns, and greater collaboration comes. However, researchers express concern that with local organizations, like women’s rights groups. a traditional knowledge, attitudes, and practices These are important strategies for achieving results (KAP) survey may not be sufficient to capture social at scale, tackling structural barriers, and building expectations and collective practices driving SGBV. policy and financial environments that are condu- Cristina Bicchieri et al. offer alternative tools practi- cive for sustainable change. tioners might consider to capture these social drivers of child marriage.70 In addition, the London School Evidence Gaps of Hygiene and Tropical Medicine’s (LSHTM) re- While research in the last 10 years has added con- cently established Learning Group on Social Norms siderably to what we know about addressing child

50 | Toward an End to Child Marriage marriage, there are still significant gaps in the ev- while new drivers such as protection concerns, de- idence. Existing research largely captures the ef- creased access to educational and livelihoods op- forts of governments, large NGOs, and INGOs, portunities for women, and increased isolation are and fails to reflect the innovations, experiences, also seen. Less is known about prevalence in these or perspectives of grassroots or civil society orga- settings. According to UNICEF, of the 10 countries nizations. Research locations are heavily centered with the highest child marriage rates, 9 are consid- around Africa and South Asia, creating challenges ered fragile or extremely fragile states.74 One review for generalizability of findings to under-represented found prevalence rates ranging from 3–51% in con- regions, such as Latin America or Southeast Asia. flict settings,75 while another found clear evidence Samples frequently include girls who are greater of both increases and declines in marriage for those than 18 years in age and suggest that findings are rel- under 20, but noted that where increases were seen, evant to all adolescent girls, in spite of evidence that they were generally concentrated among younger interventions may have varying effectiveness based adolescents.76 New research on prevalence is expected on age. Many evaluations of successful programs to be published in 2018. do not provide enough information about program Unfortunately, a significant gap still exists with content for effective replication. Evaluations of em- regards to interventions to address child marriage powerment programs, in particular, fail to specify in humanitarian contexts. Practitioners informed which empowerment approaches are being used us that child marriage is driven by extreme poverty and rarely break down program components for in- and heightened protection risks in humanitarian dividual evaluation to determine which portions of settings, and that in some settings, it exacerbates programs are having the largest effect on marriage pre-existing cultural drivers. Although most prac- outcomes. Likewise, multi-component programs are titioners were not implementing specific child mar- only beginning to be evaluated on a component-by- riage programming, they reported that they were component basis, rather than as a single program. addressing the issue to some extent through child All approaches suffer from evaluation timelines protection case management, community-based which are generally too short to measure long-term child protection committees, community sensiti- impact or capture key life transitions such as age zation efforts addressing children’s rights, child- at marriage for younger adolescent participants. friendly spaces, and general life skills curriculum. Finally, while simple baseline/endline evaluations However, these efforts were not being evaluated remain standard for many NGOs and civil society for their impact on child marriage. Practitioners organizations, this type of evaluation design is sim- underscored the urgent need for evidence of effec- ply not rigorous enough to provide generalizable tive interventions in humanitarian settings to guide evidence to improve child marriage programming. their work. While there is little evidence on how this should be undertaken given the complex bar- Humanitarian riers faced by researchers and practitioners in these In examining the state of the evidence on child mar- settings, recent research examining methodologies riage in humanitarian settings, this review found to evaluate GBV interventions in humanitarian set- that this critical research is still largely in its infancy. tings may offer guidance for child marriage research A small but significant body of literature demon- as well. The authors found that where traditional strated that pre-existing drivers such as traditional impact evaluation methodologies such as RCTs are practices, poverty, weak legislation and enforce- not appropriate or feasible, other mixed methods ment, pregnancy, and concerns over protection of approaches such as realist evaluations, longitudi- honor are aggravated in humanitarian contexts, nal designs with a qualitative component, or other

Toward an End to Child Marriage | 51 observational designs offer rigorous and informa- child marriage both individually and combined in tive alternatives.77 three African contexts concluded that multi-level interventions were not necessarily more successful Save the Children’s Approaches than simple, single-component interventions due and the Evidence to poor cost-effectiveness and difficulties of effec- When we compare Save the Children’s child mar- tively implementing these complex programs.78 riage programming with the evidence on effective This was echoed by some interviewees who added child marriage interventions from the literature re- that coordination and implementation of such pro- view and key informant interviews, we find a num- grams is incredibly challenging, and that given lim- ber of ways that Save the Children’s efforts are well- ited resources, it is important to understand which aligned with the evidence and offer suggestions for components are working and prioritize those that how future programming can be brought into closer directly benefit girls themselves. Save the Children alignment. can contribute to a broader understanding of the During the project inception phase, evidence role of multi-component, multi-level interventions from key informant interviews pointed to the im- in addressing child marriage by improving evalua- portance of formative research to better understand tion of these programs. the context-specific drivers of child marriage and In regard to approaches used to address child community input to ensure programs are aligned marriage, the majority of Save the Children’s pro- with community needs and preferences, build trust, grams focus on empowerment. This approach is and gain buy-in from parents, guardians, and local well-backed by the literature, which shows that em- leaders. Yet, Save the Children staff often recalled powerment approaches can be very effective in im- only a brief community assessment prior to improving knowledge and attitudes around child mar- plementing programming in new regions. Save the riage and delaying age of marriage. Given the broad Children can strengthen its programming and out- range of topics which often fall under empower- comes by considering how to more directly engage ment programming, both within Save the Children the community in the early project-inception phase and within child marriage programming generally, and build local ownership over its programs. Save the Children is well-poised to contribute to In recent years, Save the Children’s child marriage learning in this area by evaluating the individual im- programming has increasingly moved away from pact of various topics on child marriage outcomes to simple, single-component interventions towards better understand which components of empower- multi-component interventions intended to address ment programming are the most impactful for re- many levels of the social-ecological model, from the ducing child marriage. Additionally, early research girl herself to her relationships, her community, her suggests that empowerment programming may be policy environment, and her society. Evidence from more beneficial for reducing child marriage among many key informants pointed to multi-level inter- younger adolescents than for older adolescents. ventions as the future of child marriage program- Save the Children can contribute to learning around ming, arguing that programs are more likely to be age-specific empowerment interventions and addi- successful if they are working to address the multi- tional approaches that might be required to reduce ple and interlaced drivers of child marriage at every child marriage rates among older adolescents. level. Unfortunately, this type of program design is Incentive/asset transfer approaches to address- poorly represented and rarely evaluated in the lit- ing child marriage are rarely utilized by Save the erature. One study which evaluated the impact of Children. However, the literature indicates that in- adolescent- and community-level interventions on centive programs focused on promoting education

52 | Toward an End to Child Marriage can be a powerful tool for reducing rates of child with governments and other stakeholders to pro- marriage, particularly among out-of-school girls mote change around child marriage legislation and or older adolescent girls, where they can be used policy environments and has reported a number to increase the opportunity cost of early marriage. of promising successes worldwide. However, Save Save the Children could significantly expand ef- the Children’s evaluations of these programs reflect forts around this type of cash or asset transfer pro- a weakness found in the literature broadly—they gramming, particularly in areas where empower- focus on whether the program achieved its goals, ment-type programming falls short of achieving rather than the impact on child marriage after these the desired impact and economic factors are deter- goals are achieved. This is likely because advocacy mined to be major drivers of child marriage. Care efforts do not lend themselves well to traditional should be taken to ensure that quality education is experimental and quasi-experimental evaluation readily available in program areas, and any efforts methodologies and suffer from limitations posed should be closely monitored and evaluated to ensure by timescales and data collection. UNICEF offers safety and both immediate and long-term effective- recommendations for how monitoring and evalua- ness. Additionally, Save the Child should consider tion, including impact evaluation, can be more re- adding child marriage indicators to evaluations of sponsive to advocacy programs. Among its recom- existing cash programming around health, nutri- mendations, UNICEF suggests increasing the rigor tion, or emergency response. of appropriate non-experimental methodologies, Community engagement around child marriage focusing on measuring outcomes such as improved is the second-most widely utilized approach to ad- services and systems and positive social conditions, dressing child marriage within Save the Children’s and working to demonstrate an organization’s con- programming. This is also aligned with the litera- tribution to outcomes, rather than proving direct at- ture, which finds that meaningful community en- tribution.79 In addition to evaluation challenges, we gagement can reduce rates of child marriage by identified advocacy challenges related to interven- increasing the perceived value of girls, changing tion length and structural coordination. Advocacy gender norms, and challenging traditional practices. initiatives attempt to create a long-term, sustainable Both the literature and key informant interviews, impact in environments which are often resistant or however, indicate the importance of long-term, even hostile to change. Thus, advocacy efforts ben- deep level engagement and mobilization rather than efit from long-term horizons over which to conduct shorter, lighter-touch sensitization approaches. Save campaigns. Yet Save the Children’s largest global the Children should ensure that community pro- campaign runs on three-year strategy periods. We gramming focuses on activities which encourage also found that global advocacy and programming significant engagement, involve a broad range of structures within Save the Children are often si- actors within a community, and are long enough to loed, limiting opportunities for joint planning and promote social norm change. In addition, by care- implementation. This not only creates the potential fully describing and rigorously evaluating these for confusion over roles and responsibilities, but programs, Save the Children can contribute sig- also creates the risk of missing synergies between nificantly to the literature around what works with programming and advocacy and that efforts by one community engagement approaches to reduce child structure might duplicate or even undermine efforts marriage and why. of another. In addition, both structures could serve Legal and advocacy efforts are the final approach as valuable resources for each other, with advocacy to addressing child marriage evaluated in the liter- staff contributing a high level of expertise to country ature. Save the Children has been active in working programs, and program staff providing first-hand

Toward an End to Child Marriage | 53 perspectives to inform high-level advocacy efforts. not all forced marriage is child marriage; adults can In general, program and advocacy structures would also experience forced marriage. Forced marriage benefit from an integrated theory of change, and can involve physical, psychological, or financial co- mechanisms to facilitate joint strategy, planning, ercion and can occur in a variety of circumstances and implementation from national to global levels. such as human trafficking, mail order marriages, or Evidence on child marriage prevention and re- arranged and customary marriages, which require sponse in humanitarian settings is still in its infancy, unique and targeted interventions. This report uses and many practitioners point to the gap that needs the term “child marriage” both for clarity of defini- to be filled here in terms of research and practice. tion and to reflect the term most commonly used by Save the Children is well-positioned to advance the practitioners in our interviews, as well as in the this work, given the scope of their humanitarian literature reviewed. response efforts in the area of child protection and In addition, there remains a lack of clarity regard- across sectors, and their long-term work in coun- ing what constitutes a child marriage intervention, tries with both acute and protracted emergencies. As which creates challenges when assessing the ev- noted by one practitioner within Save the Children, idence. For example, some of the practitioners in- the Protecting Children in Conflict thematic group terviewed explained that they are not implementing may be a mechanism for strategy development child marriage programs or interventions; however, and coordination on this issue within the Save the they are working to prevent and respond to child Children movement. marriage in other ways, such as through efforts to increase girls’ access to education. For the purposes Emerging Issues in Child Marriage of this study, we define child marriage interventions, Programming programs, and initiatives as those addressing child Over the course of this research, several issues marriage as an explicit objective or those measur- emerged that warrant further discussion and debate ing child marriage as an outcome. This enabled us both within Save the Children and among stake- to capture some of the programming that may have holders in this field at large. impacted, but was not entirely focused on, child A lack of consistency and clarity in the termi- marriage. On the other hand, efforts such as those nology used to address the issue of child marriage to ensure girls remain engaged in education or pro- creates challenges in strategy development, coor- vide sexual and reproductive health services, which dination, and research. For example, while some may serve to prevent child marriage or respond to stakeholders within Save the Children and other the needs of married girls, were not included if child organizations prefer to use the broader term “child, marriage was not explicitly stated within program early, and forced marriage” (CEFM), others use sim- objectives or outcomes. ply “child marriage” when addressing this issue. As Finally, regarding terminology, there is also con- noted earlier, child marriage is defined as a formal fusion among stakeholders, both within Save the or informal union before the age of 18. The term Children and more broadly, regarding response to early marriage is often used interchangeably with child marriage. When asked about preventing and child marriage, referring to marriage before age 18. responding to child marriage, informants inter- Forced marriage, however, is defined as a marriage viewed often spoke about prevention only, or noted that occurs without the free and full consent of one that they have focused primarily on prevention and or both parties.80 Child marriage can be considered there is not enough focus on response. Those who forced marriage because individuals below 18 do spoke about response explained that response re- not have the ability to provide consent. However, fers to support services for married girls. Under this

54 | Toward an End to Child Marriage definition, child marriage response is an incredibly community leaders, or policymakers are important, broad umbrella which could include any education some expressed concerns about diverting already initiatives, SRH services, psychosocial support, live- limited resources and underscored the need to in- lihoods training, or any other support services in vest in building the capacity of girls. which married girls take part, either exclusively or Despite growing evidence and awareness among in addition to other populations. This creates chal- practitioners that child marriage can increase in hu- lenges in conducting research and assessing the ev- manitarian crises, there remains debate regarding idence base, because it can be challenging to isolate how to prevent and respond to child marriage in these efforts and synthesize what works for this spe- these settings. Some stakeholders assert that rather cific population. In addition, without a clear concept than implementing parallel programming, the most of child marriage response programming, preven- effective way to prevent and respond to child mar- tion may continue to take precedence while tailored riage is to ensure that existing minimum standards support for married girls is overlooked. in humanitarian response are met, such as the Sphere Another emerging trend that has generated de- standards in food security and nutrition, health bate among practitioners is the movement toward and protection, or the UN Interagency-Standing multi-component interventions which target risk Committee’s (IASC) gender-based violence guide- factors at multiple levels of the ecological frame- lines. Other practitioners said that it is important work. Many practitioners emphasized the need for to identify key entry points within existing human- multi-level interventions, recognizing that in order itarian response efforts, such as household vulnera- to prevent child marriage, interventions must aim bility assessments used in food security, livelihoods, not only to educate and empower girls but also to nutrition, and cash programming, in which addi- create an enabling family, community, and policy tional measures to identify and support girls at risk environment that is protective and supportive of of child marriage are integrated. Some practitioners their rights. However, as noted previously, these expressed that child marriage-specific programming more complex interventions are often challenging is appropriate and necessary in humanitarian set- to coordinate, implement, and evaluate, or are sim- tings, as displacement by conflict or natural disaster ply not feasible given resource constraints. Further, offers an ideal opportunity to shift social and cultural while many key informants in this study promoted norms. Different phases of an emergency likely re- multi-level interventions, some emphasized the quire different intervention approaches. Across prac- need to simplify programming and identify and pri- titioners, the urgent need for evidence and guidance oritize the most effective components of interven- regarding child marriage prevention and response tions in the context of scarce resources. Although strategies in both acute emergencies and protracted most practitioners recognized that child marriage displacement settings was clear. prevention efforts targeting parents, men and boys,

Toward an End to Child Marriage | 55 RECOMMENDATIONS

THE FOLLOWING RECOMMENDATIONS, based on Nicaragua, Guatemala, and Guyana in Latin findings from the literature review, mapping exer- America; Sudan in North Africa; Central cise, and key informant interviews, are targeted to- African Republic, Chad, and Guinea in West wards Save the Children and intended to help move and Central Africa; and Madagascar and the organization forward in addressing child mar- Eritrea in East and Southern Africa. riage in development and humanitarian contexts. 3. Target programs and advocacy to address the However, these recommendations are also relevant following under-reached populations: outside of Save the Children and are intended to a. Girls in Urban Contexts—Given the lack make a meaningful contribution to the efforts of of evidence on effective interventions in other stakeholders—including researchers, poli- urban contexts, conduct formative research cymakers, and practitioners—who are working on to understand the unique drivers of child this issue. marriage in these settings, gain girls’ input Development in program design, and implement programming directly tailored to their needs. For ease of use across the Save the Children move- b. Very Marginalized Girls—Focus ment, we offer recommendations under each of the programs on often-neglected girls living in organization’s five global thematic areas (Box 5). remote, difficult to reach areas or fragile Cross-thematic Recommendations contexts, as well as those belonging to racial, religious, or ethnic minorities where 1. A dopt a gender equality lens when designing, prevalence of child marriage may be partic- implementing, monitoring, and evaluating child ularly high. marriage programs and initiatives to address c. Faith-Based and Customary/Traditional the deep roots of gender inequality which drive Leaders—Given their role as gatekeepers, child marriage. influencers, and decision-makers in many 2. Consider expanding the geographic coverage communities, engage key faith-based, of child marriage programming and advocacy customary and traditional leaders, as identi- to include countries or regions with very high fied in formative research, in community rates of child marriage not currently covered outreach, sensitization, and advocacy efforts by Save the Children’s efforts. These include to reshape public perceptions and norms Lao People’s Republic in Southeast Asia; Brazil, around child marriage.

56 d. Men and Boys—Engage men and boys in 7. En sure that community sensitization and efforts to address child marriage to better engagement approaches are long-term, par- understand motivations and experiences ticipatory, and engage a broad range of actors of men involved in child marriage, help within a community who have been identified transform gender and social norms, and through formative research. Build in rigorous educate them on women’s participation in evaluation of behavior, knowledge, and atti- programming to ensure buy-in and mitigate tude change outcomes over time to ensure the backlash, such as the potential for increased efficacy of community engagement approaches domestic and intimate partner violence. is reflected in the literature. e. Married Girls—Invest in programs for al- 8. I dentify and engage key decision-makers and/ ready-married girls to address the dearth or people of influence within communities to of programming targeting this population. serve as champions for child marriage preven- Identify opportunities to reach married tion and response. girls with information and support services, develop outreach strategies, and reduce 9. Design programs with the following consider- barriers to married girls’ participation. ations in mind: a. Scalability and Cost—When deciding 4. C onsider addressing context-specific risk fac- between single- and multi-component tors at multiple social-ecological levels such as interventions, consider the feasibility, engaging families, religious and other com- sustainability, and cost of program scale-up munity leaders, and local and national govern- and whether a simple, single-component ment leaders; however, these activities should intervention may prove as effective in complement efforts to support, empower, and achieving the desired results as a complex, build the capacity of girls themselves who multi-component design given organiza- should remain at the center of programming. tional limitations. 5. Contextualize empowerment approaches to b. Length—When possible, design programs address local drivers of child marriage as well and corresponding evaluations to be con- as the specific needs and interests of the girls ducted over longer time periods to promote themselves. Consider combining empower- and capture changes in social norms and ment with other approaches, such as incentive/ long-term outcomes for girls. asset transfer programs, to fully address the c. Formative Research—Conduct formative needs of older adolescents. Conduct targeted research to understand context-specific outreach strategies and tailored programming risk factors, including social expectations, as needed to engage married girls. knowledge, attitudes, and perceptions related to gender roles and practices of child 6. U se incentive/asset transfer approaches to build marriage; identify barriers to abandoning girls’ human capital by incentivizing educa- the practice of child marriage, influential tional outcomes or combining them with com- community members and decision-makers; plementary approaches such as empowerment and develop tailored interventions. and community sensitization and engagement d. Community Input—Encourage community for long-term impact. Ensure such approaches participation, with a focus on girls, during are responsive to different age groups. the program-inception phase to ensure

Toward an End to Child Marriage | 57 programs are responsive to local needs and Advocacy interests and have local buy-in. 1. P lace youth voices at the center of all advocacy e. Rigorous Evaluation—Build in a rigorous efforts, with a particular emphasis on girls’ monitoring and evaluation plan from the voices, to ensure youth’s needs and interests are outset of program implementation, based prioritized. on an experimental or quasi-experimental 2. methodology, wherever appropriate. Develop high-level partnerships and iden- Consider alternative ways to achieve rigor tify champions, such as national ministers or when evaluating advocacy initiatives or offices of the first lady, to expand political will social norms change such as those identi- and buy-in on the issue of child marriage. fied by UNICEF and the LSHTM’s Learning 3. Coordinate with well-established women’s rights Group on Social Norms and Gender-based organizations to strengthen advocacy efforts 81 Violence. Consider longer program and maximize impact. evaluation periods to capture a majority of program participants’ age at marriage. 4. C ombine advocacy efforts with technical assistance to support governments in developing and 10. D evelop a coordinated strategy at the organiza- implementing legislative and policy changes. tional level to address child marriage. 5. a. Joint Objectives, Outcomes, and Indicators: Ensure advocacy efforts complement program- Develop an organization-wide set of objec- ming directly benefitting girls where possible, tives, outcomes, and indicators for address- that programs integrate advocacy components, ing and evaluating child marriage. and that programming and advocacy are coor- . b Terminology: Unify terminology to address dinated and work in synergy with each other. child marriage prevention and response 6. E xpand legal advocacy initiatives around child across the movement. marriage to include educational goals for c. Coordination: Clarify where the issue sits adolescent girls, such as guaranteed access to within Save the Children and how coordi- quality education through secondary school. nation will take place across development 7. and humanitarian sectors, across thematic A ssess the impact of advocacy achievements on areas, and at the organizational and coun- women and girls, rather than strictly focusing try levels. Further coordinate and integrate on whether legal or policy goals have been advocacy and programming structures to achieved, to strengthen evaluation efforts. unify and reinforce efforts and synergize 8. I ncrease the rigor of evaluations of child expertise. marriage advocacy initiatives using creative, 11. Develop a coordinated strategy at country context-appropriate methodologies, such as 82 levels with all stakeholders, within and outside those identified by UNICEF, to significantly of Save the Children, active in addressing child contribute to the global literature on the im- marriage to improve national outcomes and pacts of advocacy approaches. Consider how minimize duplication of efforts. indicators for child marriage outcomes, such as reduced prevalence or changes in community knowledge and attitudes, can be included in these evaluations and monitored over time.

58 | Toward an End to Child Marriage Child Poverty 4. P rovide assessment and measurement guidance 1. M easure the impact of child poverty program- to other sectors to support the identification ming on behavioral child marriage outcomes of contributing factors to child marriage, for such as age at first marriage or proportion of example, by developing tools which build mea- girls married. surements of social expectations and collective practices into traditional knowledge, attitudes, 2. Consider the use of incentive/asset transfer and practice (KAP) studies to identify the so- programs which incentivize educational out- cial determinants of child marriage. comes to reduce child marriage among families for whom poverty is a primary driver of child Child Rights Governance marriage. 1. S upport opportunities for child-led advocacy 3. En sure that incentive/asset transfer programs and accountability on child marriage, such as are well-evaluated and that impact is measured via child clubs and other forums for children by age to determine whether younger or older to receive information about their rights, gain adolescents are impacted more significantly. advocacy skills, and lead activities in their communities and countries. 4. C onsider the use of age-appropriate employment and livelihood opportunities, as well as 2. Build gender equality and child marriage into training, to delay marriage among older adoles- child-friendly budget analysis and advocacy cent girls in contexts where limited educational work to increase resourcing for child marriage and income generating opportunities and policies and programs. poverty are primary drivers of child marriage. 3. Integrate gender equality and child marriage into efforts to protect and expand civil society Child Protection space from local to international levels, includ- 1. E valuate the impact of child protection interven- ing through working with girls’ and women’s tions on child marriage-related outcomes using groups. rigorous experimental and quasi-experimental methodologies or alternative rigorous evalua- Education tion approaches whenever relevant. 1. M easure the impact of educational program- 2. Encourage other sectors, such as education, ming on knowledge and attitudes around child health, or livelihoods to include child marriage marriage as well as child marriage behaviors, outcomes when evaluating interventions to such as age at first marriage or proportion of address drivers of child protection issues in the girls married. local context to better understand the role other 2. R emove barriers to school participation, thematic areas can play in addressing the issue. such as direct and indirect costs of schooling, 3. Coordinate with other thematic areas to ad- gender-discriminatory norms, or poor health dress child marriage through programming and nutrition, and increase access to education and measurement/evaluation approaches such for adolescent girls of all ages, including older as education, health, or livelihoods which are adolescents. critical for child marriage prevention and re- 3. Consider additional incentives such as cash sponse, yet do not fall under the child protection transfer programming or free school uni- thematic area. forms and supplies to lower barriers to school

Toward an End to Child Marriage | 59 attendance and increase the opportunity cost 5. Evaluate the impact of SRHR programming, for older adolescent girls to leave school. including CSE, SRHR information, and youth- responsive SRH services, on knowledge and 4. I ntegrate comprehensive sexuality education attitudes around child marriage as well as child and/or child marriage, gender equality, and marriage behaviors such as age at first marriage sexual and reproductive health and rights into or proportion of girls married, particularly in standard education curriculum. regions where early pregnancy is a significant 5. Remove barriers to education for already- driver of child marriage. married girls and pregnant and adolescent 6. Off er youth-responsive SRHR education pro- mothers, including addressing legal and policy grams and services, as well as food security and barriers, counteracting discriminatory social nutrition programs, to already-married girls norms, and providing child care, as well as to delay pregnancy and reduce risk of STIs, offering after-school literacy and numeracy encourage healthy spacing of pregnancies, pro- training and bridge programs where necessary. mote safe pregnancy and delivery, and respond Health and Nutrition to the health and nutrition needs of married girls’ infants and children. Develop targeted 1. D etermine whether food-insecurity is a driver outreach strategies to increase service utiliza- of child marriage in a given context and con- tion by married girls. sider how food security, livelihoods, and nutrition programs may improve child marriage outcomes—for example, where parents choose Humanitarian to marry daughters early because they cannot 1. C onduct assessments to identify both pre- afford to provide food for all family members, existing drivers of child marriage, such as tra- or where girls are unable to attend or succeed ditional norms and practices, as well as new risk in school because of poor nutrition. factors related to crisis coping, such as economic and protection concerns, when developing 2. Determine whether lack of access to youth- child marriage interventions. Identify key responsive SRH services is a driver of child decision-making factors and support needs of pregnancy, and thus child marriage, in a given adolescent girls and their families to inform context, and consider how such programs may intervention design. improve child marriage outcomes by increasing access to SRHR education and contraceptives. 2. D etermine which forms of child marriage in- Develop targeted outreach strategies to intervention are appropriate for various phases of crease service utilization by unmarried girls. an emergency. For example, nutrition programs to reduce child marriage among food-insecure 3. Integrate child marriage into CSE, SRHR, and households may be most appropriate in acute GBV education and awareness-raising activities crisis, while empowerment programs and com- for in- and out-of-school youth. munity sensitization to change social norms 4. M easure the impact of food security, liveli- may not be realistic or appropriate until the hoods and nutrition programming on child long-term recovery stages. marriage behaviors such as age at first marriage 3. Identify key entry points for mainstreaming or proportion of girls married. child marriage into existing humanitarian response efforts including household economic

60 | Toward an End to Child Marriage surveys, SGBV prevention and response, and participants such as age at first marriage and food distribution and nutrition activities. long-term program impacts, particularly with social norms change. 4. P ilot stand-alone child marriage programs and integrate sector-specific approaches to address 5. Evaluate each component of multi-component child marriage in humanitarian settings. Build and multi-level programming individually as in rigorous evaluations to develop evidence in well as together to determine which compo- acute and protracted displacement settings. nents are most successful at addressing child marriage. 5. Build child marriage prevention and response capacity in fragile contexts whenever possible 6. C onsider cost, sustainability, and scalability to ensure that the underlying drivers of child when evaluating program impacts. marriage are being addressed prior to acute 7. E valuate distinct components of empowerment crises. programming such as SRHR education, finan- 6. En able girls to return to school as quickly as cial literacy, or sports programming separately possible following acute crisis by incentivizing to determine the impact of each component on schooling and reducing barriers such as lan- child marriage outcomes. guage, transportation, and safety concerns— 8. I mprove evaluation of community sensitization particularly in displacement settings—to and engagement to contribute significantly to increase access to education for all girls. organizational and global evidence on changing social norms around child marriage. Use Research creative methodologies to achieve this when 1. E xpand research on child marriage drivers and experimental and quasi-experimental methods interventions to include settings that are poorly are not appropriate. represented in the literature, including child 9. Improve evaluation of advocacy initiatives to marriage in urban contexts, child marriage in contribute to organizational and global evi- regions such as Latin America and Southeast dence by doing the following: Asia, and child marriage in the context of gang a. Assess the impact of advocacy achieve- conflict in Central America. ments on women and girls by measuring 2. Conduct more rigorous impact evaluations of changes in knowledge and attitudes or child marriage programming and advocacy ini- behaviors rather than strictly focusing on tiatives. Use experimental and quasi-experimen- whether law and policy goals have been tal evaluation methods or alternative rigorous achieved. approaches whenever possible. . b Employ creative, context-appropriate methodologies, such as those identified by UNICEF,83 3. Disaggregate evaluation data by age to better to improve the rigor of non-experimental understand what works to prevent child mar- methodologies when experimental and quasi- riage among younger and older adolescents and experimental methods are not appropriate. among different social, economic, and identity groups. Likewise, report findings for adolescents 10. E valuate the impact of other approaches such under age 18 as distinct from youth over age 18. as health, nutrition, and education on child marriage as an outcome. 4. L engthen evaluation time to capture individuals’ significant life events for program

Toward an End to Child Marriage | 61 11. Build a formative research phase into piloting c. Drivers—Expand local and global knowl- and evaluation efforts to tailor interventions edge around the pre-existing and crisis- to the context-specific drivers and needs and specific drivers that promote child preferences of the community. marriage. d. Support Needs—Understand the support 12. P rioritize research addressing support needs, needs of adolescent girls and their families outreach strategies, and effective interventions to prevent and respond to child marriage. for already-married girls to expand child mar- . e Interventions—Evaluate interventions to riage response efforts. determine what works to address child 13. B uild an evidence base in humanitarian set- marriage in humanitarian settings. tings that addresses the following topics: 14. Engage youth in participatory research during a. Increases and Decreases—Determine why formative and pilot design phases to ensure child marriage increases in some humani- that the needs and interests of young people are tarian contexts and decreases in others. prioritized and to improve buy-in. . b Decision-Making Factors—Determine what decision-making factors by parents and adolescents lead to child marriage.

62 | Toward an End to Child Marriage APPENDICES

Appendix 1: Literature Review Key Search Terms

KEY SEARCH TERMS

Child youth, young people, young person, adolescent*, teen*, juvenile*, minor*, pre-teen*, preteen, girl*, boy*

Marriage marriage*, marital, matrimony, nuptial*, wedlock, union, wedding*, bride*, spouse*, “conjugal relationship”, “consensual union”, cohabit*

Child Marriage “child early and forced marriage”, “early marriage”, “forced marriage”, “child bride”

Intervention program*, solution*, service*, approach*, method*

Prevention prevent*, eliminat*, reduc*, decreas*, end*, stop*, delay*, address*, respon*

Setting Humanitarian: humanitarian disaster*, conflict*, natural disaster*, disaster zone*, refugee camp*, IDP camp*, emergenc*, crisis, crises Development: development, developing countr*, low-/lower-/middle-income countr*, global south, regions or countries (by name)

* Indicates a wildcard character

Appendix 2: Literature Review Data Extraction Form

DATA EXTRACTION FORM

Background Information 3. Theory of change 1. Abstractor name 4. Frequency and duration (exposure period) 2. Article title 5. Number and size of intervention groups

3. Author or agency name(s) Outcomes 4. Publication title 1. Operational definition of CEFM 5. Year of publication 2. Primary and secondary outcomes by sector 6. Type of publication 3. Key conclusions by sector 7. Country(ies) where the study took place 4. Negative impacts reported 8. Study setting/situation (school/community, rural/urban, etc.) Evaluation 9. Humanitarian crisis type (armed conflict or natural disaster) 1. Study design 10. Crisis stage (acute, chronic, recovery, etc.) 2. Method of data collection 11. Participants/population type (refugees, IDPs, etc.) 3. Adapted STROBE score OR Adapted CONSORT score (1–8) 12. Participant characteristics (age, sex, etc.) Analysis Intervention 1. Strengths/weaknesses of evaluation 1. Type/description 2. Any other comments 2. Objectives

63 Appendix 3: Literature Quality Review Criteria and Scoring84

QUALITY REVIEW CRITERIA

STROBE Criteria (Observational Studies) CONSORT Criteria (Randomized Trials)

Intervention: Eligibility: 1. Is the intervention clearly described? 1. Did study state # not meeting inclusion criteria? Selection of participants: 2. Did study state # declined to participate? 2. Is the target population defined? Once Randomized (allocation): 3. Is there a comparison group (e.g. baseline, control?) 3. Did study state # receiving intervention? 4. Are the inclusion and exclusion criteria defined? 4. Did study state # not receiving intervention? Statistical methods: Follow-Up: 5. Is the sample size / method justified with statistical basis? 5. Did study state # lost to follow-up? 6. Is there a statistical test (p-value or confidence interval)? 6. Did study provide reasons for loss to follow-up? 7. Is there adjustment for confounding variables? Analysis: Limitations: 7. Did study state reasons participants were excluded from 8. Are study limitations explained (e.g. biases)? analysis? 8. Are limitations of the study explained (e.g. biases)?

LEVEL OF QUALITY (PER STROBE/CONSORT)

High 7–8 criteria met = high quality evidence Medium 4–6 criteria met = moderate quality evidence Low 1–3 criteria met = low quality evidence

64 | Toward an End to Child Marriage Appendix 4: Summary of the Literature on Child Marriage Interventions in Development Contexts85–121

Program, Target Country, Implementing Population, Evaluation Study Details Years Organization Setting Objectives Approach Design, Quality Outcomes

Alam et al., Pakistan, Punjab Female Girls in grades To improve school Incentive Quasi- • Age at marriage 2011 2003–09 School Stipend 6–8, urban/ participation of Experimental, (increased) Program (FSSP), rural middle school-aged High • Middle school completion Government of girls and improve (increased) Pakistan gender equality

Amin et al., Bangladesh, BALIKA Project, Married and To delay marriage Empowerment RCT, Medium • Probability of child 2016 2012–15 Population unmarried by increasing marriage (decreased) Council, girls aged girls’ skills, • Drop-out rates PSTC, CIDIN, 12–18, rural self-confidence, (decreased) mPower Social and strategies • Knowledge of SRH topics Enterprises to advocate for (improved) themselves

Ara and Das, Bangladesh, Adolescent Girls and boys To increase aware- Empowerment, Quasi- • Child marriage knowledge 2010 2008–10 Development aged 11–19, ness of HIV/AIDS Community Experimental, (not significant) Program (ADP), rural border related issues, Medium • Knowledge of HIV/AIDS BRAC districts marriage laws, (increased) gender equality, • Schooling outcomes (no SRH, and dowry impact) • Awareness of legal, social issues (increased) • Positive attitudes towards women (improved)

Austrian et al., Zambia, Adolescent Girls Girls aged To reduce social Empowerment RCT, High • Timing of marriage, edu- 2016 2013–16 Empowerment 10–19, urban/ isolation and build cational attainment, preg- Program (AGEP), rural assets to facilitate nancy/birth, experience Population intermediate and of violence, or HIV/HSV-2 Council long-term pos- prevalence (no impact) itive change for • Transactional sex girls’ transitions to (decreased) adulthood • Use of condom at first sex (increased) • Empowerment outcomes (improved)

Baird et al., Malawi, Zomba Cash Unmarried To improve school Incentive RCT, Medium • Marriage rates (mixed) 2010 2007–09 Transfer girls aged enrollment, educa- • Pregnancy rates (declined) Program, 13–22, urban/ tional attainment, • Sexual debut (delayed) Multiple rural marriage, and child- • School enrollment bearing outcomes (improved) • Number of sexual partners (reduced)

Toward an End to Child Marriage | 65 Appendix 4: Continued

Program, Target Country, Implementing Population, Evaluation Study Details Years Organization Setting Objectives Approach Design, Quality Outcomes

Baird et al., Malawi, Zomba Cash Unmarried To improve school Incentive RCT, High • Marriage rates, UCT arm 2011 2007–09 Transfer girls aged enrollment, educa- (decreased) Program, 13–22, tional attainment, • Pregnancy rates, UCT arm Multiple urban/rural marriage, and child- (decreased) bearing outcomes • Dropout rates, both arms (declined) • Educational achievement, CCT arm (improved)

Baird et al., Malawi, Zomba Cash Unmarried To improve school Incentive RCT, Medium • Marriage rates (mixed) 2013 2007–09 Transfer girls aged enrollment, educa- • Pregnancy rates Program, 13–22, urban/ tional attainment, (decreased) Multiple rural marriage, and child- • Access to financial er - bearing outcomes sources (increased) • Schooling outcomes (increased) • Health outcomes (improved) • Household agency (improved)

Baird et al., Malawi, Zomba Cash Unmarried To improve school Incentive RCT, Medium • Marriage rates (no impact) 2016 2007–09 Transfer girls aged enrollment, educa- • Schooling outcomes Program, 13–22, urban/ tional attainment, (mixed) Multiple rural marriage, and child- • Fertility outcomes (no bearing outcomes impact) • Wage work (little impact)

Bandiera et Uganda, Empowerment Girls aged To improve Empowerment RCT, Medium • Age at marriage (delayed) al., 2012 2008–10 and Livelihood 14–20, urban/ cognitive and • Age at first birth (delayed) for Adolescents rural non-cognitive • HIV, pregnancy related (ELA), BRAC skills knowledge (improved) • Risky behaviors (reduced) • Likelihood of self-employment (increased)

Bandiera et Uganda, Empowerment Girls aged To improve the Empowerment RCT, Medium • Income generating activi- al., 2017 2008–10 and Livelihood 14–20, urban/ economic and ties (increased) for Adolescents rural social empower- • Adolescent pregnancy (ELA), BRAC ment of adolescent rates (decreased) girls • Early marriage, cohabita- tion rates (decreased) • Non-consensual sex (decreased)

66 | Toward an End to Child Marriage Appendix 4: Continued Appendix 4: Continued

Program, Target Country, Implementing Population, Evaluation Study Details Years Organization Setting Objectives Approach Design, Quality Outcomes

Brady et al., Egypt, The Ishraq Out-of- To change gender Empowerment, Quasi- • Child marriage attitudes 2007 2001–07 Program, school girls norms, improve Advocacy Experimental, (improved) Population aged 13–15, literacy, life skills, High • Schooling outcomes Council, Caritas, rural health knowledge (improved) CEDPA, Save the and attitudes, and • Health knowledge Children civic participation, (increased) and gain policy- • Gender equitable attitudes makers’ support (improved)

Buchmann et Bangladesh, Kishoree Kontha Unmarried To reduce CM and Empowerment, RCT, High Conditional incentive only: al., 2017 2007–15 and Conditional girls aged teenage child- Incentive • Likelihood of marriage Incentive 10–19, urban/ bearing through (reduced) Program, rural an empowerment • Adolescent childbearing Save the program and/or (reduced) Children a conditional incen- Combined: tive intervention • Likelihood of being in school (increased)

Buehren et al., Tanzania, Empowerment Adolescent To assist adoles- Empowerment, RCT, Medium • Child marriage (no impact) 2015 2009–11 and Livelihood girls, urban/ cents in achieving Community • Participants’ empower- for Adolescents rural greater economic ment (no impact) (ELA) Plus, BRAC and social empow- • Reproductive outcomes erment through life (no impact) skill and livelihood training and safe spaces

Chow and Ethiopia, N/A, Girls aged To reduce child Community, Quasi- • Child marriage (reduced) Vivalt, 2016 2012–14 Government of less than 18, marriage through Incentive Experimental, • Women’s empowerment Ethiopia their parents economic High (mixed) and commu- incentives and • Educational outcomes nities, rural information about (mixed) the potential harms • Polarization in beliefs of early marriage about CM (increased)

Duflo et al., Kenya, Education Girls and boys To decrease STI Empowerment, RCT, High Education subsidy only: 2015 2003–05 Subsidy pro- in grade 6, prevalence and Incentive • Early marriage (reduced) gram and the urban/rural teenage pregnancy • Adolescent childbearing HIV Education using an education (reduced) program, subsidy and/or ab- • School drop-out (reduced) Government of stinence-only HIV Abstinence-only HIV educa- Kenya and ICS education program tion program: Africa • Counteracts positive effects of subsidy

Toward an End to Child Marriage | 67 Appendix 4: Continued

Program, Target Country, Implementing Population, Evaluation Study Details Years Organization Setting Objectives Approach Design, Quality Outcomes

Erulkar and Ethiopia, Berhane Hewan, Married and To improve educa- Empowerment, Quasi- • Preferred marriage age Muthengi, 2009 Population unmarried tional attainment, Community, Experimental, >18 years (increased) 2009 Council and girls aged RH knowledge, Incentive High • Age at marriage (mixed) Ethiopia Ministry 10–19 and contraceptive use • RH knowledge, contra- of Youth and their parents, and age at first ceptive use (increased) Sport rural marriage • School enrollment (increased)

Erulkar et al., Burkina N/A, Unmarried To delay CM Empowerment, Quasi- Child marriage rates: 2017 Faso, Population girls aged through simplified Community, Experimental, • Community dialogue Ethiopia, Council 12–17, rural interventions, in- Incentive High (decreased for girls 12–14 and cluding community in ET, 15–17 in BF) Tanzania dialogues, school • Education promotion promotion, and (decreased for girls aged asset transfers 12–14 in ET) • Asset transfer (decreased for girls 15–17 in ET and TZ) • Comprehensive model (decreased for girls 12–14 in TZ, girls 15–17 in ET and TZ)

Gulemetova- Mexico, Oportunidades, Adolescents To reduce poverty Incentive Natural • Age at marriage (delayed) Swan, 2009 2002–2004 Government of and families for families meeting Experiment, • Onset of premarital sex Mexico in poverty, the official criteria High (delayed) urban of poverty • Age at first and second birth (delayed)

Hahn et al., Bangladesh, Female Unmarried To increase female Incentive Natural • Age at first marriage 2016 1994–2011 Secondary girls in grades enrollment and re- Experiment, (delayed) School Stipend 6–10, rural tention in second- High • Age at first birth (delayed) Program (FSSSP), ary school, enhance • Years of schooling Government of female employment (increased) Bangladesh opportunities, • Probability of complet- and delay age of ing secondary school marriage (increased) • Likelihood of women working (no impact)

68 | Toward an End to Child Marriage Appendix 4: Continued Appendix 4: Continued

Program, Target Country, Implementing Population, Evaluation Study Details Years Organization Setting Objectives Approach Design, Quality Outcomes

Hallfors et al., Zimbabwe, N/A Never To impact HIV Incentive RCT, Medium • Likelihood of marriage 2015 2007–10 married or infection, sexual (reduced) pregnant debut, marriage, • HIV infection (no impact) orphaned educational • Sexual debut (delayed) girls in grade achievement, • School achievement 6, rural health-related (improved) quality of life, and • Health-related quality of socioeconomic life (improved) status • Socioeconomic status (improved)

Handa et al., Kenya, Cash Transfer Girls aged To reduce poverty Incentive RCT, Medium • Likelihood of early mar- 2015 2007–11 for Orphans 12–24, urban/ for families caring riage (no impact) and Vulnerable rural for orphans and • Probability of being preg- Children vulnerable children nant (reduced) (CT-OVC), Government of Kenya

Heath and Bangladesh, N/A Women To analyze the Other Natural • Age at marriage (increased) Mobarak, 1980s–2009 employed in welfare effects of Experiment, • Child-bearing (declined) 2015 the garment access to factory High • School enrollment industry, jobs for women (increased) urban/rural • Employment (increased)

Hong and Bangladesh, Female Unmarried To increase Incentive Natural • Age at marriage Sarr, 2012 1990–2012 Secondary girls grades enrollment rates in Experiment, (increased) School Stipend 6–10, rural secondary school, High • Labor force participation Program (FSSSP), enhance enroll- (increased) Government of ment opportunities, • Educational attainment Bangladesh and delay age at (increased) marriage

Independent Pakistan, Punjab Female Girls in grades To improve educa- Incentive Natural • Age at marriage (delayed) Evaluation 2003–08 School Stipend 6–10, urban/ tional attainment Experiment, • Adolescent pregnancy Group, 2011 Program (FSSP), rural among girls and High (suggestive decrease) Provincial decrease gender • Enrollment rates (increased) Government of gaps at the middle • Completion of high school Pakistan school and high (increased) school levels • Labor market participation (decreased)

Jensen, 2012 India, N/A Girls aged To examine the Empowerment RCT, Medium • Likelihood of getting mar- 2003–06 15–21, rural effect of an ried (decreased) exogenous change • Likelihood of childbearing in labor market (reduced) opportunities for • Employment gains women • School enrollment (increased) • BMI (increased)

Toward an End to Child Marriage | 69 Appendix 4: Continued

Program, Target Country, Implementing Population, Evaluation Study Details Years Organization Setting Objectives Approach Design, Quality Outcomes

Kanesathasan India, The Married and To increase access Empowerment, Quasi- • Age at marriage (increased) et al., 2008 2005–07 Development unmarried to contraceptive Community Experimental, • Attitudes around marriage Initiative girls and boys and SRH services, High (improved) Supporting aged 14–24, delay marriage • Contraceptive use Healthy rural and childbearing, (increased) Adolescents teach livelihoods • Livelihoods skills and (DISHA) skills, and develop perceived self-value Program, leadership capacity (increased) ICRW, Local of adolescents • SRH knowledge, attitudes NGOs (improved) • Uptake of SRH services (no impact)

Landesa: India, Security for Girls Girls, rural To break the cycle Empowerment, Quasi- • Age of marriage (delayed) Rural 2012–13 Through Land, of rural poverty Community Experimental, • Dowry practices Development Landesa: Rural using a curriculum Low (improved) Institute, 2013 Development on land-based • School drop-outs Institute livelihoods and land (decreased) rights • Economic and land assets (increased) • Legal knowledge and life- skills (increased) • Likelihood of having own income (increased) • Participation in land-based livelihoods (improved)

Nanda et al., India, Apni Beti Apna Unmarried To delay child mar- Incentive Quasi- • Marriage before age 18 2016 1994–98 Dhan (ABAD), girls born riage and improve Experimental, (no impact) Government of from 1994– attitudes related to High • Completion of 8th grade India 98 to families the value of girls (increased) below the • Aspirations for post-sec- poverty line, ondary school (increased) urban/ rural

Pandey et al., India, PRACHAR Girls and boys To provide ado- Empowerment Quasi- • Timing of marriage (no 2016 2010–11 Project (Phase 3), aged 13–21, lescents with SRH Experimental, impact) Pathfinder rural knowledge and High • Age at first pregnancy (no International communication impact) and negotiation • SRH knowledge skills (improved) • Contraceptive use (increased)

70 | Toward an End to Child Marriage Appendix 4: Continued Appendix 4: Continued

Program, Target Country, Implementing Population, Evaluation Study Details Years Organization Setting Objectives Approach Design, Quality Outcomes

Pathfinder India, PRACHAR Girls and boys To provide ado- Empowerment Quasi- • Child marriage attitudes International, 2002–11 Project (Phase aged 13–21, lescents with SRH Experimental, (improved) 2011 1 and 2), rural knowledge and Medium • Knowledge of SRH Pathfinder communication (Improved) International and negotiation • Contraceptive use skills (Increased) • Years of schooling (Increased)

Rahman et al., Bangladesh, Skills Training Out-of- To enhance em- Empowerment Quasi- • Early marriage (no impact) 2017 2012–17 for Advancing school girls ployment of school Experimental, • Employment (increased) Resources and boys drop-outs through High • Earning levels and financial (STAR), aged 14–18, practical training assets (increased) BRAC urban/ rural and empowerment • Health knowledge (increased)

Sambodhi India, Deepshikha Girls aged To engage ad- Empowerment, Quasi- • Child marriage (decreased) Research 2008–15 Adolescent Girls 12–18, olescent girls in Advocacy Experimental, • Political, economic em- and Groups (AGG), urban slums/ protecting their Medium powerment (improved) Commu- UNICEF rural rights and provide • Social impact (significant) nications, life-skills education • School enrollment 2014 (increased) • Health knowledge (increased)

Save the Bangladesh, Adolescent Girls aged To improve knowl- Empowerment, Quasi- • Confidence to advocate Children, 2014–15 Health and 10–19, edge of SRHR, delay Community, Experimental, with parents for delayed 2015 Empowerment urban slums marriage, reduce Advocacy Medium marriage, refuse marriage to Reduce Child harmful practices, before 18, help prevent Marriage (RCM) improve adoles- marriage of other girls Project, cents’ economic before 18 (increased) Save the situation, and • Awareness on CM and Children strengthen policies SRHR (increased)

Scales et al., Bangladesh, Kishoree Girls aged To build Empowerment RCT, Medium • Marriage status (excluded 2013 2006–10 Kontha, Save the 10–19, developmental from evaluation) Children rural assets of rural • Developmental assets girls through peer (increased) education in social skills, literacy, and school learning

Toward an End to Child Marriage | 71 Appendix 4: Continued

Program, Target Country, Implementing Population, Evaluation Study Details Years Organization Setting Objectives Approach Design, Quality Outcomes

Selim et al., Egypt, Ishraq Program, Out-of- To improve school- Empowerment, Quasi- • CM knowledge and atti- 2013 2001–13 Population school girls ing, health, and Advocacy Experimental, tudes (improved) Council, aged 12–15, economic develop- Medium • Desire to delay marriage Caritas, rural ment by improving until 18 (increased) CEDPA, self-confidence, • Knowledge and attitudes Save the self-awareness, and related to SRH, child bear- Children community gender ing, FGM/C, and gender norms roles (improved) • Functional literacy and life-skills (improved) • Mobility (improved)

Shahnaz and Bangladesh, Empowerment Girls aged To create empow- Empowerment Quasi- • Age at marriage Karim, 2008 2005–07 and Livelihood 10–24, urban ering opportunities Experimental, (increased) for Adolescents and rural for girls and equip High • STI and HIV knowledge (ELA), them with psy- (no impact) BRAC cho-social and in- • School enrollment rates terpersonal skills to (no impact) make decisions that • Economic empowerment promote well-being (improved) • Mobility (improved) • Extracurricular reading (increased)

Yoder, 2008 Senegal, Tostan Women aged To encourage Community Quasi- • Age at first marriage 2000–06 Programme, 15–49, rural social change by Experimental, (unclear) Tostan developing capacity Medium • Female genital cutting in human rights, (small decrease) problem resolu- • Health status of mothers tion, basic hygiene, and children (improved) financial manage- • Use of services (improved) ment, and health

72 | Toward an End to Child Marriage Appendix 4: Continued Appendix 5: Summary of the Literature on Interventions for Married Girls122–127

Program, Target Evaluation Country, Implementing Population, Design, Study Details Years Organization Setting Objectives Approach Quality Outcomes

Ashburn et al., Uganda, Responsible, Married To address gender Empowerment Quasi- • All forms of IPV (decreased) 2016 2013–15 Engaged, fathers aged norms that promote Experimental, • Confidence in non-violent and Loving 16–25 with use of violence with High resolution (increased) (REAL) Fathers children children and intimate • Positive parenting (mixed) Initiative, Save aged 1–3, partners through • Communication (improved) the Children, rural promotion of pos- Institute for itive parenting and Reproductive partnership skills Health building

Edmeades Ethiopia, TESFA, CARE Married girls To improve eco- Empowerment Quasi- • SRH knowledge, use of services and Hayes, 2010–13 and ICRW aged 10–19, nomic empower- Experimental, (increased) 2014 rural ment and SRH out- Medium • Contraceptive use (increased) comes for married • Decision-making, adolescent girls communication (improved) • Income-generating activities (increased)

Erulkar and Ethiopia, Berhane Hewan, Married and To improve educa- Empowerment, Quasi- • Preferred marriage age >18 years Muthengi, 2009 Population unmarried tional attainment, RH Community, Experimental, (increased) 2009 Council and girls aged knowledge, contra- Incentive High • Age at marriage (mixed) Ethiopia Ministry 10–19 and ceptive use and age • Reproductive health knowl- of Youth and their par- at first marriage edge and contraceptive use Sport ents, rural (increased) • School enrollment (increased)

Erulkar and Ethiopia, Meseret Hiwott, Married To increase social Empowerment Quasi- • Husbands’ support with Tamrat, 2014 2008–13 Population girls aged networks and pro- Experimental, domestic work, accompani- Council 12–24 and vide knowledge and High ment to clinic, family planning, husbands, skills to improve RH voluntary counseling, and testing rural and prevent HIV (improved) • Sexual violence (mixed)

Kanesathasan India, The Married and To increase access Empowerment, Quasi- • Age at marriage (increased) et al., 2008 2005–07 Development unmarried to contraceptive Community Experimental, • Attitudes around marriage Initiative girls and and SRH services, High (improved) Supporting boys aged delay marriage and • Contraceptive use (increased) Healthy 14–24, rural childbearing, teach • Livelihoods skills and perceived Adolescents livelihoods skills, and self-value (increased) (DISHA) develop leader- • SRH knowledge, attitudes Program, ship capacity of (improved) ICRW, Local adolescents • Uptake of SRH services (no NGOs impact)

Pandet al., India, PRACHAR Unmarried To provide ado- Empowerment Quasi- • Timing of marriage (no impact) 2016 2010–11 Project and married lescents with SRH Experimental, • Age at first pregnancy (no (Phase 3), girls and knowledge and High impact) Pathfinder boys aged communication and • SRH knowledge (improved) International 13–21, rural negotiation skills • Contraceptive use (increased)

Toward an End to Child Marriage | 73 Appendix 6: Summary of Save The Children’s Programs and Initiatives to Address Child Marriage

Program or Target Intervention Country, Years Populations Description Evaluation

Adolescent Bangladesh, Adolescent girls Livelihood interventions, awareness-raising • 65% increase in adolescents receiv- Development 2014–ongoing and secondary about gender norms and CM, SRH, and ing ASRH services from govern- school students menstrual hygiene management to pro- ment health facilities mote equal opportunity to adolescent girls.

Adolescent Vietnam, Adolescents Multi-pronged approach to improve • None available Development 2016–ongoing and community adolescents’ SRH knowledge, attitude and members practices, the quality and accessibility of health information and services, and the social and policy environment

Adolescent Girls Kenya, Marginalized Includes educational, health, wealth • Project ongoing—early evidence of Initiative - Kenya 2015–ongoing adolescent girls creation, and prevention of violence changes in attitudes toward CM in (AGI-K) aged 11–14 interventions, including CM one community

Adolescent Bangladesh, Adolescent girls Interactive sessions on SRH, CM, and early • Increased girls’ ability to name Health and 2015 aged 10–19 pregnancy; strengthened local services; harmful effects of CM Empowerment vocational training; and working with male • Increased girls’ self-confidence and to Reduce Child decision-makers strength to refuse marriage before Marriage (RCM) age 18 Project • Increased the likelihood girls will refuse to marry before 18

ASPIRE Malawi, Adolescents aged Activity to improve educational achieve- • Project ongoing 2014–18 10–19 and their ment for girls, addressing literacy, SRH communities education, and social norms and behavior change

Bien Grandir! DRC, Very young In-school and out-of-school club activities • None available (Growing Up 2015–20 adolescents as well as family and community engage- Great!) aged 10–14 and ment to improve gender-equitable norms, parents, newly family planning, and SRH practices and married couples, outcomes first-time parents, and teachers

Choices Voices Multiple, including Very young Curriculum for very young adolescents, • Girls felt empowered to talk to Promises Bangladesh, Bolivia, Egypt, adolescent girls their families, and their communities their parents about continuing their El Salvador, Ethiopia, aged 10–14, to increase knowledge, skills and studies and avoiding early marriage Kyrgyztan, Malawi, Nepal, adolescent opportunities to improve gender norms • Significant positive impact on Uganda, and Zambia brothers, creating a shift towards more parents, and their gender-equitable attitudes communities

74 | Toward an End to Child Marriage Appendix 6: Summary of Save The Children’s Programs and Initiatives to Address Child Marriage Appendix 6: Continued

Program or Target Intervention Country, Years Populations Description Evaluation

Choices Voices Bolivia, Bangladesh, Egypt, Unclear Unclear • None available Promises Ethiopia, El Salvador, Malawi, (Adapted Zambia Curriculum)

Combatting Malawi Children, Comprehensive social-ecological approach • None available Child Marriage communities, to strengthening children, communi- and Human community and ties, community and district structures to Trafficking in district leaders prevent and respond to child marriage and Mwanza and trafficking Neno

Combatting Ethiopia, Unclear Program to prevent child marriage and • None available Child Marriage 2012–13 contribute to the attainment of girls’ right to in North Gondar education and protection Zone of Amhara Region, Ethiopia

Every Last Child Bolivia, Burkina Faso, Cote Various, including Global advocacy initiative to contribute • Varies by country d’Ivoire, Ethiopia, Honduras, adolescent girls to SDG targets and SC’s breakthrough Malawi, Mali, Mozambique, ambitions, focusing on exclusion. Select Nepal and Bhutan, Niger, countries focus on CM. Nigeria, Peru, Senegal, Uganda, Senegal, Sierra Leone, Somalia, South Sudan, Tanzania, Zimbabwe, as well as advocacy offices in New York (UN); Geneva (UN); Addis Ababa (AU), and donor countries 2016–ongoing

FACT with Pragati Nepal, Hard-to-reach Series of games to diffuse information • Project ongoing 2016–ongoing married and about fertility awareness and family unmarried planning (including messaging around CM) adolescent girls and boys aged 15–25

Ishraq Egypt, Out-of-school, A literacy, life-skills, and sports approach • Positive effects on participants’ 2001–13 rural, adoles- to foster girls’ self-awareness and build knowledge and attitudes related cent girls aged self-confidence to RH, marriage and child bearing, 12–15 and their FGM/C, and gender roles communities • Participants more likely to want to delay marriage and limit childbearing

Toward an End to Child Marriage | 75 Appendix 6: Continued

Program or Target Intervention Country, Years Populations Description Evaluation

Keep it Real Ethiopia and Uganda, Very young Comprehensive sexuality education • External evaluation by 2013–16 adolescents aged program for in-school and out-of-school ICRW—evaluation reports 10–14 and youth youth with improved service access, provi- available aged 15–24 sion, and coordination

Kishoree Kontha Bangladesh, Adolescent Community mobilization, SRHR and literacy • Substantial improvement in girls’ 2006–10 girls aged for girls, financial literacy, and cooking oil developmental assets 10–19 and their incentive to improve girls’ health, educa- communities tion, economic opportunity, and social well-being

Kishori Abhijan Bangladesh, Adolescent girls Awareness raising, peer-to-peer participa- • Evidence that girls targeted under 2001–10 aged 10–19, tory education and life skills training, and this project waited significantly boys, and their community mobilization to lower school longer to marry communities dropout rates and delay marriage

Life Steps Nigeria, Adolescents, Comprehensive model to address child • Evaluation ongoing 2014–17 families, marriage at every level from adolescent communities, empowerment and vocational skills training government and to family and community knowledge en- religious leaders hancement, to training CSOs and strengthening the legislative framework

Local to Global Afghanistan, Bangladesh, Children, civil Program and advocacy approach to foster • Qualitative evaluation Ethiopia, India, Mexico, society organi- dynamic advocacy efforts from the com- • Stakeholders were able to Sierra Leone, Yemen, zations, media, munity up to national and relevant global strengthen their capacity and hold Zambia, as well as governments, platforms to ensure governments are on duty-bearers to account for deliv- advocacy offices in and other tradi- track to prioritize and significantly invest in ery on strengthened policy, legal, Brussels (EU), New York tional leaders children, especially the most marginalized; and budget commitments (UN), Addis Ababa (AU), (has CM focus) and Geneva (UN), 2013–16

Marriage: No India, Malawi, Mali, Niger, Adolescent girls Model social-ecological program to en- • Evaluation ongoing Child’s Play Pakistan, and stakeholders able girls to decide if and when to marry 2016–20 and pursue their SRHR in a supportive environment

Norad Mozambique, Adolescents Multi-pronged approach of equipping ado- • None available Cooperation 2015–18 lescents with necessary skills and informa- Agreement tion, increasing their access to services and opportunities, and building their agency, while shifting social norms related to CEFM, SRHR, and gender equality

76 | Toward an End to Child Marriage Appendix 6: Continued

Program or Target Intervention Country, Years Populations Description Evaluation

Pan-African CSE Cote d’Ivoire, Kenya, Liberia, Community SRHR education to change restrictive • Child marriage criminalized in and Information Malawi, Nigeria, Senegal, members, perceptions around gender, sexuality, and Zimbabwe Project South Africa, Swaziland, including child marriage Tanzania, Uganda, Zambia, parents, and Zimbabwe teachers, community and religious leaders

Reducing Bangladesh, Children and Program aimed to reduce the vulnerability • Communities with program ac- Exploitation 2011–16 stakeholders of children from abuse, exploitation and tivities reported little to no cases and Abuse of other forms of violence with a small CM of CM as compared with control Children through component groups where nearly all girls were Strengthening married by 18 National Child Protection System in Bangladesh (REACH)

Reducing Child Sierra Leone, Burkina Faso, Unclear CM prevention integrated with SRH, • None available Early and Forced 2018–21 education and child protection Marriage components

Reducing Mozambique Unclear Integrated child protection and health inter- • None available Teenage ventions to support completion of quality Pregnancy by basic education by addressing barriers to Keeping Girls in access including CM School

Resources Cote d’Ivoire, Girls aged 12–24, Awareness raising and sensitization of • Project ongoing Towards 2015–20 parents, and parents, caregivers and community mem- Elimination community bers and empowering girls with accurate of Child members information and education covering CSE, Vulnerability cultural pressures, early pregnancy, gender (REVE) norms, GBV, and consequences of CM

Strengthening Malawi, Families, Program to strengthen both the formal • Increase in knowledge of Community 2014–16 communities, and informal child protection systems to adults who have the capacity to Child Protection and institutions be able to prevent, detect, and respond detect and respond to all forms of Systems to fight to all forms of exploitation, corporal violence against children, Violence Against punishment and neglect of children, including CM Children (VAC) including CM

Toward an End to Child Marriage | 77 Appendix 6: Continued

Program or Target Intervention Country, Years Populations Description Evaluation

That’s No Way to Mali, Niger, Bangladesh, Adolescents, SRH and CM training in existing youth clubs, • Increased knowledge about the Marry Pakistan, India, communi- training of teachers and other professionals detrimental effects of CM 2015–16 ties, teachers, on child rights and protection, and a radio • Increased awareness of SRH ser- and other campaign vices, though not service utilization stakeholders

The Right to be Dominican Republic, Sierra Married and Social-ecological approach to prevent • Project ongoing a Girl Leone, unmarried girls and reduce the acceptance and practice of 2017–21 early marriage, while ensuring support for adolescent girls who are already married

TZA Supporting Tanzania, Adolescent girls Program to empower and enable civil • Project ongoing Civil Society to 2018–19 aged 10–19, society in Shinyanga region to counter Counter Harmful men and harmful traditional practices that violate Traditional boys, civil rights of the most vulnerable children, Practices in society groups, notably adolescent girls, and share Sinyanga Region community learning with other regions of Tanzania members, elders, and policymakers

78 | Toward an End to Child Marriage Appendix 6: Continued Appendix 7: Summary of Save the Children’s Programs and Initiatives to Address Child Marriage in Humanitarian Contexts

Program or Intervention Country, Years Target Populations Description Evaluation

Building a Jordan, Vulnerable youth and Informal education, life skills, and psycho- • None available Resilient Youth 2015–17 their communities social support to improve vulnerable youth among Syrians outcomes and protect them from the risks of and Jordanians child labor and CM

CHANGES (or Somalia, Women and girls, Harmonized package of approaches to en- • Project ongoing SNaP—Social 2016–ongoing families, communities, hance civic engagement to support gender Norm and and leaders equality, improve knowledge and attitudes Participation) on FGM/C and CM, and increase social and economic empowerment of women and girls

Going Against Somalia, Girls, boys, Program to strengthen institutions and civil • No formal evaluation the Grain: 2013–14 family members, society in crisis affected states by working to • Found enhanced capacity of local Protecting Girls and communities build capacity of government, community officials and community to protect girls Against Early and organizations and local partners to work to from CM and enhanced capacity of Forced Marriage prevent CM girls to protect themselves from abuse, in Somalia exploitation, and harm through CM

The Right Jordan, Married and Girls empowerment, advocacy, and involve- • Project ongoing to be a Girl 2018–19 unmarried girls ment of communities in transforming behaviors to end child marriage and empower married girls/young women to enhance their resilience and wellbeing

Toward an End to Child Marriage | 79 NOTES

1 “Child Marriage,” UNICEF, March 5, 2018, https://www.unicef 10 Sarah Neal, Nicole Stone, and Roger Ingham, “The impact .org/protection/57929_58008.html (accessed May 15, 2018). of armed conflict on adolescent transitions: a systematic re- 2 Previously adapted by Karl Blanchet et al. in An Evidence view of quantitative research on age of sexual debut, first Review of Research on Health Interventions in Humanitarian marriage and first birth in young women under the age of Crises (London: LSHTM, 2015). 20 years,” BMC Public Health 16, no. 1 (2016): 1–11. 3 Programs include Kishoree Kontha and Adolescent Girls 11 A Study on Early Marriage in Jordan 2014 (Amman, Jordan: Initiative—Kenya (AGIK). See Nina Buchmann et al., Power UNICEF, 2014), 8. vs Money: Alternative Approaches to Reducing Child Mar- 12 Marry before Your House is Swept Away: Child Marriage in riage in Bangladesh, a Randomized Control Trial (2017), Bangladesh (New York: Human Rights Watch, 2015), 9. 1–42; Eunice Muthengi et al., Adolescent Girls Initiative— 13 “Child marriage is a violation of human rights,” UNICEF Data. Kenya: Qualitative Report (Nairobi: Population Council, 14 Edilberto Loaiza and Sylvia Wong, Marrying Too Young: 2016), 1–40. End Child Marriage (New York: UNFPA, 2012), 34. 4 The term “child marriage” is used throughout the report to 15 “Child Marriage,” UNICEF. denote a formal or informal union where at least one of the parties is under 18 years of age. The authors acknowledge 16 Child Marriage in Humanitarian Settings (London: Girls that individuals or organizations may prefer other terminol- Not Brides, 2017), 1–7. ogy including “child, early, and forced marriage” (CEFM) or 17 UN General Assembly, The Universal Declaration of Human simply “early marriage.” However, as early marriage is fre- Rights, 1948, 217 (III) A, Paris. quently inclusive of adolescents and young adults above 18, 18 UN General Assembly, Convention on the Rights of the and forced marriage may happen to anyone, regardless of Child, 1989, United Nations, Treaty Series, vol. 1577. age, we have chosen to use “child marriage” for clarity and 19 UN General Assembly, Convention on the Elimination of accuracy. For further discussion on terminology, see Susan- all Forms of Discrimination Against Women, December 18, na Greijer and Jaap Doek, Terminology Guidelines for the 1979, United Nations, Treaty Series, vol. 1249, 13. Protection of Children from Sexual Exploitation and Sexual Abuse (Luxembourg: Interagency Working Group on Sexu- 20 It is worth noting here that at least one key informant ex- al Exploitation of Children, 2016), 62–64. pressed concern that the Convention on the Rights of the Child does not provide sufficient protection against child 5 “Child Marriage,” UNICEF. marriage, as the document does not contain a specific pro- 6 “Child marriage is a violation of human rights, but is all too vision banning the practice of child marriage. common,” UNICEF Data, March, 2018, http://data.unicef. 21 United Nations Convention on the Elimination of All org/topic/child-protection/child-marriage/# (accessed May Forms of Discrimination against Women, Joint general rec- 15, 2018). ommendation/general comment No. 31 of the Committee on 7 “Child marriage is a violation of human rights,” UNICEF Data. the Elimination of Discrimination against Women and No 18. 8 “Fertility rate total (births per woman),” The World Bank of the Committee on the Rights of the Child on harmful prac- Group, https://data.worldbank.org/indicator/SP.DYN.TFRT tices, CEDAW/C/GC/31-CRC/C/GC/18, 2014. .IN/ (accessed May 15, 2018). 22 Quintin Wodon et al., Economic Impacts of Child Marriage: 9 “Child Marriage,” UNICEF. Global Synthesis Brief (Washington, DC: The World Bank and International Center for Research on Women, 2017), 8.

80 23 United Nations Children’s Fund, Hidden in Plain Sight: A ington, DC: World Bank, 2017), 1–28; and Oriana Bandiera Statistical Analysis of Violence Against Children (New Work: et al., Empowering Adolescent Girls: Evidence from a Ran- UNICEF, 2014), 130–137. domized Control Trial in Uganda (Washington, DC: World 24 World Health Organization, WHO Guidelines on Preventing Bank, 2012). Early Pregnancy and Poor Reproductive Health Outcomes 39 Bandiera et al., Women’s Empowerment in Action, 1–28. among Adolescents in Developing Countries (Geneva: WHO, 40 Niklas Buehren et al., Evaluation of Layering Microfinance 2011), IX. on an Adolescent Development Program for Girls in Tanzania 25 Previously adapted by Karl Blanchet et al. in An Evidence (2015). Review of Research on Health Interventions in Humanitarian 41 Girls Not Brides, ‘A Theory of Change on Child Marriage’ Crises (London: LSHTM, 2015), 18. (London: Girls Not Brides, 2014). 26 Blanchet et al., An Evidence Review of Research on Health 42 See Martha Brady et al., Providing New Opportunities to Interventions in Humanitarian Crises, 18. Adolescent Girls in Socially Conservative Settings: The Ish- 27 Sarah Baird, Craig McIntosh, and Berk Ozler, When the raq program in rural Upper Egypt (New York: Population Money Runs Out (Washington, DC: World Bank, 2016), Council, 2007), 1–38; Mona Selim et al., The Ishraq Pro- 1–93. gram for Out-of-School Girls: From Pilot to Scale-up (Cairo: 28 Niklas Buehren et al., Evaluation of Layering Microfinance Population Council, 2013), 1-13; Evaluation of Empowering on an Adolescent Development Program for Girls in Tanzania Young Girls and Women in Maharashtra, India (New Del- (BRAC, 2015). hi: Sambodhi, 2014), 1–89; and Adolescent Health and Em- powerment to Reduce Child Marriage (RCM) Project: End 29 “Child Marriage,” UNICEF. Line Survey Report (Dhaka, Bangladesh: Save the Children, 30 “Child marriage is a violation of human rights,” UNICEF Data. 2015), 1–28. 31 Jeffrey Edmeades, Robin Hayes, and Gillian Gaynair, Im- 43 Erulkar et al., The Impact and Cost of Child Marriage Pre- proving the Lives of Married Adolescent Girls in Amhara, vention, 23. Ethiopia: A Summary of the Evidence (Washington, DC: In- 44 Erulkar et al., “Evaluation of Berhane Hewan,” 12. ternational Center for Research on Women, 2014), 1–12. 45 See Sudhanshu Handa et al., “Impact of the Kenya Cash 32 Naila Kabeer, “Reflections on the Measurement of Wom- Transfer for Orphans and Vulnerable Children on early en’s Empowerment,” in Discussing Women’s Empowerment: pregnancy and marriage of adolescent girls,” Social Science Theory and Practice, ed. Anne Sisask (Stockholm: Swedish & Medicine 141 (2015): 36–45; and Sarah Baird et al., When International Development Cooperation Agency, 2001). the Money Runs Out, 31–33. 33 Deepa Naryan, ed., Measuring Empowerment: Cross-Disciplinary 46 Erulkar et al., “Evaluation of Berhane Hewan,” 6–14. Perspectives (Washington, D.C.: The World Bank, 2005), 72. 47 Erulkar et al., The Impact and Cost of Child Marriage Preven- 34 Seo Yeon Hong and Leopold Remi Sarr, Long-term Impacts tion, 1–29. of the Free Tuition and Female Stipend Programs on Educa- tion Attainment, Age of Marriage, and Women’s Labor Market 48 Erulkar et al., The Impact and Cost of Child Marriage Preven- Participation in Bangladesh (Washington, DC: World Bank, tion, 14–20. 2012), 6. 49 See Erulkar et al., The Impact and Cost of Child Marriage 35 See Annabel S. Erulkar and Eunice Muthengi, “Evaluation Prevention; and Eva Vivalt and Vinci Chow, “Challenges in of Berhane Hewan: A Program to Delay Child Marriage in Changing Social Norms: Evidence from Interventions Tar- Rural Ethiopia,” International Perspectives on Sexual and geting Child Marriage in Ethiopia,” (2016). Reproductive Health 35, no. 1 (2009): 8; Annabel Erulkar, 50 See Brady et al., Providing New Opportunities to Adolescent Girmay Medhin, and Eva Weissman, The Impact and Cost of Girls in Socially Conservative Settings; Selim et al., The Ishraq Child Marriage Prevention in Three African Settings (Addis Program for Out-of-School Girls; and Adolescent Health and Ababa: Population Council, 2017), 10. Empowerment to Reduce Child Marriage (RCM) Project: End 36 Paul Stanley Yoder, Long-Term Evaluation of The Tostan Line Survey Report. Programme in Senegal: Kolda, Thiès and Fatick Regions 51 Evaluation of Empowering Young Girls and Women in Maha- (New York: United Nations Children’s Fund, 2008), 1–74. rashtra, India, 67. 37 Jinnat Ara and Narayan Das, Impact Assessment of Adoles- 52 Selim et al., The Ishraq Program for Out-of-School Girls, 2. cent Development Programme in the Selective Border Regions 53 Tilly Gurman et al., “‘By seeing with our own eyes, it can of Bangladesh (Dhaka, Bangladesh: BRAC, 2010), 1–33. remain in our mind’: qualitative evaluation findings suggest 38 Oriana Bandiera et al., Women’s Empowerment in Action: the ability of participatory video to reduce gender-based Evidence from a Randomized Control Trial in Africa (Wash-

Toward an End to Child Marriage | 81 violence in conflict-affected settings,” Health Education Re- et al., Solutions to End Child Marriage: What the Evidence search 29, no. 4 (2014): 690–701. Shows (Washington, DC: ICRW, 2011), 1–32. 54 Kathryn L. Falb et al., “Creating opportunities through 67 Chae and Ngo, The Global State of Evidence on Interventions mentorship, parental involvement, and safe spaces (COM- To Prevent Child Marriage. PASS) program: multi-country study protocol to protect 68 Anastasia J. Gage, “Child marriage prevention in Amhara girls from violence in humanitarian settings,” BMC Public Region, Ethiopia: Association of communication exposure Health 16, no. 1 (2016): 1–10. and social influence with parents/guardians’ knowledge and 55 “Access the Data,” UNICEF Data, March, 2018, http://data. attitudes,” Social Science & Medicine 97 (2013): 124–133. unicef.org/topic/child-protection/child-marriage/# (accessed 69 Tanya Abramsky et al., “Findings from the SASA! Study: a May 18, 2018). cluster randomized controlled trial to assess the impact of a 56 Programs include Bien Grandir! (DRC); “Choices, Voices, community mobilization intervention to prevent violence Promises” curriculum (Multiple); CO Senegal (Senegal); against women and reduce HIV risk in Kampala, Ugan- Combatting Child Marriage and Human Trafficking in d a ,” BMC Medicine 12, no. 1 (2014): 122. Mwanza and Neno (Malawi), Keep It Real (Ethiopia and 70 Cristina Bicchieri, Jan W. Lindemans, and Ting Jiang, “A Uganda); Marriage: No Child’s Play (India, Malawi, Mali, Structured Approach to a Diagnostic of Collective Practic- Niger, Pakistan), REVE: Resources Towards Elimination of es,” Frontiers in Psychology 5 (2014): 1418. Child Vulnerability (Cote d’Ivoire); The Right to Be a Girl 71 Ben Cislaghi and Lori Heise, “Measuring Gender-related (Dominican Republic and Sierra Leone); Going Against the Social Norms: Report of a Meeting, Baltimore Maryland, Grain (Somalia). See Appendix 6 for further information on June 14–15, 2016,” Learning Group on Social Norms and each program. Gender-based Violence of the London School of Hygiene and 57 Programs include Adolescent Girls Initiative (Kenya), Bien Tropical Medicine, http://researchonline.lshtmac.uk/46469 Grandir! (DRC), Ishraq (Egypt), and Keep it Real (Ethiopia 72/ (accessed May 17, 2018). and Uganda). See Appendix 6 for further information on 72 Belinda Maswikwa et al., “Minimum Marriage Age Laws each program. and the Prevalence Of Child Marriage and Adolescent Birth: 58 Prevalence rates of child marriage by country obtained Evidence from Sub-Saharan Africa,” International Perspec- from “Child marriage is a violation of human rights, but is tives on Sexual and Reproductive Health; New York 41, no. 2 all too common,” UNICEF Data. (2015): 64. 59 More Than Brides Alliance,Marriage: No Child’s Play Pro- 73 Quentin Wodon et al., Ending Child Marriage: Child Marriage gramme Document (MTBA, 2014), 4. Laws and Their Limitations (2017), 6. 60 See also LAHIA (Niger), Nchanda ni nchanda pa umi wam- 74 “Child Marriage,” UNICEF. bone (Malawi), REACH (Bangaldesh), REAL Fathers (Ugan- 75 Alys McAlpine, Mazeda Hossain, and Cathy Zimmerman, da), TZA Supporting Civil Society to Counter Harmful Tradi- “Sex trafficking and sexual exploitation in settings affect- tional Practices in Sinyanga Region of Tanzania (Tanzania), ed by armed conflicts in Africa, Asia and the Middle East: and the “Choices, Voices, and Promises” curriculum. See Ap- systematic review,” BMC International Health and Human pendix 6 for further information on each program. Rights 16 (2016), 5. 61 See Buchmann et al., Power vs Money: Alternative Approach- 76 Sarah Neal, Nicole Stone, and Roger Ingham, “The Impact es to Reducing Child Marriage in Bangladesh, a Randomized of Armed Conflict on Adolescent Transitions: A Systematic Control Trial. Review of Quantitative Research on Age of Sexual Debut, 62 See Eunice Muthengi et al., Adolescent Girls Initiative—Kenya: First Marriage and First Birth in Young Women Under the Qualitative Report. Age of 20 Years,” BMC Public Health 16, no. 1 (2016): 5. 63 See Karen Austrian et al., “The Adolescent Girls Initia- 77 Mazeda Hossain and Alys McAlpine, Gender Based Violence tive-Kenya (AGI-K): study protocol,” BMC Public Health 16, Research Methodologies in Humanitarian Settings: An Evidence no. 1 (2016): 1–14. Review and Recommendations (Elhra: Cardiff, 2017), 25. 64 Brian Keeley and Céline Little, ed., Children in a Digital 78 Erulkar et al., The Impact and Cost of Child Marriage Pre- World (New York, NY: UNICEF, 2017) vention, 23. 65 “Bangladesh,” Girls Not Brides, https://www.unicef.org/ 79 Julia Coffman, Monitoring and Evaluating Advocacy: Com- protection/57929_58008.html (accessed May 31, 2018). panion to the Advocacy Toolkit (UNICEF, 2011). 66 See Sophia Chae and Thoai Ngo, The Global State of Evi- 80 UN General Assembly. Preventing Child, Early, and Forced dence on Interventions To Prevent Child Marriage (New Marriage, Report of the Office of the United Nations High York: Population Council, 2017), 1–14; and Anju Malhotra Commissioner for Human Rights, April, 2014, 4.

82 | Toward an End to Child Marriage 81 Coffman, Monitoring and Evaluating Advocacy; Cislaghi and sions about marriage and fertility: The case of Oportuni- Heise, “Measuring Gender-related Social Norms.” dades” (Ph.D. diss., University of Pennsylvania, 2009), 82 Coffman,Monitoring and Evaluating Advocacy, 6–7. 1–59. 83 Coffman,Monitoring and Evaluating Advocacy, 6–7. 103 Youjin Hahn et al., Education, Marriage and Fertility: Long- Term Evidence from a Female Stipend Program in Bangla- 84. Blanchet et al., An Evidence Review of Research on Health desh (The University of Chicago, 2015), 1–35. Interventions in Humanitarian Crises, 18. 104 Denise Hallfors et al., “The Impact of School Subsidies on 85 Andaleeb Alam, Javier Baez, and Ximena Del Carpio, Does HIV-Related Outcomes Among Adolescent Female Or- Cash for School Influence Young Women’s Behavior in the p h an s ,” Journal of Adolescent Health 56, no. 1 (2015): 79-84. Longer Term? Evidence from Pakistan (discussion paper, Insti- tute for the Study of Labor [IZA], Bonn, Germany, 2011), 1–55. 105 Handa et al., “Impact of the Kenya Cash Transfer.” 86 Sajeda Amin et al., Delaying Child Marriage Through Com- 106 Rachel Heath and A. Mushfiq Mobarak, “Manufacturing munity-Based Skills-Development Programs for Girls: Re- Growth and the Lives of Bangladeshi Women,” Journal of sults from a Randomized Controlled Study in Rural Bangla- Development Economics 115 (2015): 1–15. desh (Dhaka: Population Council, 2016), 1–50. 107 Hong and Sarr, Long-term Impacts of the Free Tuition and 87 Ara and Das, Impact Assessment of Adolescent Development Female Stipend. Programme. 108 Independent Evaluation Group, Do Conditional Cash Trans- 88 Karen Austrian et al., Adolescent Girls Empowerment Pro- fers Lead to Medium-Term Impacts?: Evidence from a Female gramme: Research and Evaluation Mid-term Technical Re- School Stipend Program in Pakistan (Washington, DC: port (Lusaka, Zambia: Population Council, 2016), 1–15. World Bank, 2011), 1–61. 89 Sarah Baird et al., “The Short-Term Impacts of a Schooling 109 Robert T. Jensen, “Do labor market opportunities affect Conditional Cash Transfer Program on the Sexual Behav- young women’s work and family decisions?” The Quarterly ior of Young Women,” Health Economics 19, no. 1 (2010): Journal of Economics 127, no. 2 (2012): 753–792. 55–68. 110 Anjala Kanesathasan et al., Catalyzing Change: Improving 90 Sarah Baird, Craig McIntosh, and Berk Özler, “Cash or Con- Youth Sexual and Reproductive Health Through DISHA, an dition? Evidence from a Cash Transfer Experiment,” The Integrated Program in India (Washington, DC: Internation- Quarterly Journal of Economics 126, no. 4 (2011): 1709–1753. al Center for Research on Women, 2008), 1–40. 91 Sarah Baird et al., “Girl power: Cash Transfers and Ado- 111 Landesa: Rural Development Institute, Security for Girls through lescent Welfare. Evidence From a Cluster-Randomized Ex- Land: Pilot Evaluation (2012–2013) (Landesa, 2013), 1–35. periment in Malawi,” National Bureau of Economic Research 112 Priya Nanda et al., Making Change with Cash? Impact of a Working Papers (2013): 139–164. Conditional Cash Transfer Program on Age of Marriage in 92 Baird, McIntosh, and Ozler, When the Money Runs Out. India (Washington, DC: International Center for Research on Women, 2016), 1–8. 93 Oriana Bandiera et al., Empowering Adolescent Girls : Evi- dence from a Randomized Control Trial in Uganda (World 113 Neelanjana Pandey et al., Effects of the Prachar Project’s Re- Bank, Washington, DC, 2012), 1–33. productive Health Training Programme for Adolescents: Find- ings from a Longitudinal Study (New Delhi: Population 94 Oriana Bandiera et al., Women’s Empowerment in Action. Council, 2016), 1–72. 95 Brady et al., Providing New Opportunities to Adolescent Girls 114 PRAGYA Multisectoral, Gendered Approach to Improve in Socially Conservative Settings. Family Planning and Sexual and Reproductive Health for 96 Buchmann et al., Power vs Money. Young People: A Research Study (Watertown, MA: Path- 97 Buehren et al., Evaluation of Layering Microfinance. finder International, 2011), 1–35. 98 Eva Vivalt and Vinci Chow, “Challenges in Changing So- 115 Rehnuma Rahman et al., The Effects of Skill Training on cial Norms.” Livelihoods: Evidence from BRAC’s Intervention on School 99 Esther Duflo, Pascaline Dupas, and Michael Kremer, “Ed- Dropout Adolescents (Dhaka: BRAC, 2017), 1–44. ucation, HIV, and Early Fertility: Experimental Evidence 116 Sambodhi, Evaluation of Empowering Young Girls and Women from Kenya,” American Economic Review 105, no. 9 (2015): in Maharashtra. 2757–2797. 117 Save the Children, Adolescent Health and Empowerment to 100 Erulkar and Muthengi, “Evaluation of Berhane Hewan.” Reduce Child Marriage. 101 Erulkar et al., The Impact and Cost of Child Marriage Prevention. 118 Peter C. Scales et al., “Building Developmental Assets to 102 Michaela Gulemetova-Swan, “Evaluating the impact of Empower Adolescent Girls in Rural Bangladesh: Evalu- conditional cash transfer programs on adolescent deci-

Toward an End to Child Marriage | 83 ation of Project Kishoree Kontha,” Journal of Research on Northern Uganda,” Prevention Science 18, no. 7 (2016): Adolescence 23, no. 1 (2013): 171–184. 854–864. 119 Selim et al., The Ishraq Program for Out-of-School Girls. 123 Edmeades et al., Improving the Lives of Married Adolescent Girls. 120 Rizwana Shahnaz and Raihana Karim, Providing Microf- 124 Erulkar and Muthengi, “Evaluation of Berhane Hewan.” inance and Social Space to Empower Adolescent Girls: An 125 Annabel Erulkar and Tigest Tamrat, “Evaluation of a Re- Evaluation of BRAC’s ELA Centres (Dhaka: BRAC, 2008), productive Health Program to Support Married Adoles- 1–38. cent Girls in Rural Ethiopia,” African Journal of Reproduc- 121 Yoder, Long-Term Evaluation of The Tostan Programme. tive Health 18, no. 2 (2014): 68–76. 122 Kim Ashburn et al., “Evaluation of the Responsible, En- 126 Kanesathasan et al., Catalyzing Change. gaged, and Loving (REAL) Fathers Initiative on Physi- 127 Pandey et al., Effects of the Prachar Project’s Reproductive cal Child Punishment and Intimate Partner Violence in Health Training Programme for Adolescents, 1–72.

84 | Toward an End to Child Marriage

HUMAN RIGHTS CENTER UC BERKELEY SCHOOL OF LAW 396 SIMON HALL BERKELEY, CA 94720 510.642.0965

[email protected] HRC.BERKELEY.EDU MEDIUM.COM/HUMANRIGHTSCENTER

@HRCBERKELEY 86 | Toward an End to Child Marriage