Girl Child Marriage, Health, and Well-Being in Sub-Saharan Africa: A Mixed Methods Investigation
The Harvard community has made this article openly available. Please share how this access benefits you. Your story matters
Citable link http://nrs.harvard.edu/urn-3:HUL.InstRepos:37925669
Terms of Use This article was downloaded from Harvard University’s DASH repository, and is made available under the terms and conditions applicable to Other Posted Material, as set forth at http:// nrs.harvard.edu/urn-3:HUL.InstRepos:dash.current.terms-of- use#LAA
GIRL CHILD MARRIAGE, HEALTH, AND WELL-BEING IN SUB-SAHARAN AFRICA: A MIXED METHODS INVESTIGATION
YVETTE OLUSEYI EFEVBERA
A Dissertation Submitted to the Faculty of The Harvard T.H. Chan School of Public Health in Partial Fulfillment of the Requirements for the Degree of Doctor of Science in the Department of Global Health and Population Harvard University Boston, Massachusetts
November 2018
Dissertation Advisor: Prof. Jacqueline Bhabha Yvette Oluseyi Efevbera
Girl Child Marriage, Health, and Well-Being in Sub-Saharan Africa: A Mixed Methods Investigation
Abstract
Nearly 650 million girls and women alive today have experienced girl child marriage, a formal union before age 18. Yet gaps exist in understanding its health consequences for women and children in African contexts. The objective of this multi-study dissertation was to use a mixed methods research approach to improve our understanding of the complex relationships between girl child marriage and the health and well-being of women and their children in sub-
Saharan Africa.
Aim 1 estimates associations between girl child marriage and women’s undernutrition in sub-Saharan Africa. Using household data from Demographic and Health Surveys (DHS), representing 35 African countries (N=249,269 women), this study revealed that girl child marriage was associated with increased odds of early motherhood and being in the poorest versus wealthiest quintile as well as decreased odds of completing secondary school, controlling for contextual factors. Despite these relationships, girl child marriage was associated with reduced odds of being underweight (body mass index less than 18.5).
Aim 2 estimates associations between girl child marriage and children’s development and stunting to examine intergenerational impacts. Using household data from UNICEF Multiple
Indicator Clusters Survey (MICS) for 16 African countries (N=37,558 child-mother pairs), this study showed increased odds being off-track for development (measured using the Early
Childhood Development Index) and stunted (height-for-age z score less than -2) among children born to women who married before age 18, compared to those whose mothers married as adults.
ii
Adjusted models revealed that contextual factors, maternal education, and wealth explained these associations.
Aim 3 qualitatively explores how women married as children in Conakry, Guinea see their marriages as related to their health and their children’s health. Rooted in grounded theory, this study used data from in-depth interviews, brief ethnographic interviews, observation, and participant observation (N=19 women). Open coding was used to identify key themes. In the context of their child marriages, women perceived health disadvantages, categorized under four themes. Women also perceived health advantages in their marriages, categorized under five themes. Further analysis identified three factors influenced women’s perceptions.
In combination with a theoretical chapter that deconstructs the term “girl child marriage,” research findings provide evidence to guide population policies designed to ensure that women and children can achieve their developmental potential.
iii
Table of Contents
Abstract ...... ii List of Figures ...... viii List of Tables ...... ix Preface ...... x Acknowledgments ...... xiii Chapter 1: Introduction ...... 1 Background on Girl Child Marriage ...... 1 Girl Child Marriage and Women’s Health and Well-Being ...... 3 Girl Child Marriage and Children’s Health and Well-Being ...... 5 Investigating Girl Child Marriage and Health Consequences in Sub-Saharan Africa ...... 6 Gaps ...... 6 Purpose ...... 7 Contributions of This Study ...... 8 Research Design and Frameworks ...... 9 Mixed Methods Research Design ...... 9 Other Frameworks Considered ...... 12 Bioecological Framework ...... 13 Biosocial Approach ...... 15 Conceptual Map ...... 16 Chapter 2: Defining and Deconstructing “Girl Child Marriage” and Its Applications to Public Health Research ...... 22 A Historical Journey on Terminology ...... 23 Parsing Out “Marriage” ...... 26 Parsing Out “Child” ...... 31 The Need to Focus on Girls...... 35 Overlaps and Distinctions with Forced Marriage ...... 37 Measuring Girl Child Marriage ...... 38 Conclusions and Implications for Public Health ...... 39 Chapter 3: Girl Child Marriage and Undernutrition: Evidence from 35 Countries in Sub- Saharan Africa ...... 41 Summary ...... 41 Background ...... 41
iv
Methodology ...... 45 Data source ...... 45 Participants ...... 45 Measures ...... 46 Girl child marriage ...... 46 Underweight ...... 46 Secondary outcomes ...... 46 Other covariates ...... 47 Statistical analysis ...... 47 Ethical approval ...... 50 Results ...... 50 Discussion ...... 59 Interpretation ...... 59 Limitations ...... 62 Strengths ...... 63 Conclusions ...... 63 Chapter 4: Girl Child Marriage as a Risk Factor for Early Childhood Development and Stunting ...... 65 Summary ...... 65 Introduction ...... 65 Framework ...... 68 Methodology ...... 72 Study Sample ...... 72 Measures ...... 73 Girl child marriage ...... 73 The Early Childhood Development Index (ECDI) ...... 74 Stunting ...... 74 Other covariates ...... 75 Data Analysis ...... 75 Results ...... 77 Sociodemographic Data ...... 77 Bivariate Models on Associations Between Girl Child Marriage and Child Development and Health ...... 82 Adjusted Models on Associations Between Girl Child Marriage and Child Development and Health ...... 86 ECDI ...... 86 Stunting ...... 86 Hypothesized Mechanisms Explaining the Association Between Girl Child Marriage and Child Development and Health ...... 87
v
Discussion ...... 90 Limitations ...... 93 Strengths ...... 94 Conclusions ...... 94 Chapter 5: ‘It is this which is normal’ A Qualitative Study on Girl Child Marriage and Health in Guinea ...... 96 Summary ...... 96 Introduction ...... 96 Methodology ...... 99 Overview ...... 99 Participant Selection ...... 101 Data Collection ...... 102 Data Analysis ...... 104 Ethical Considerations ...... 105 Results ...... 106 A Brief Account of How Child and Early Marriage was Defined ...... 106 A Profile of the Women Married as Children ...... 108 A Brief Account of Why They Married Early ...... 110 A Brief Account of Whom They Married ...... 113 Impact of Women’s Early Marriages ...... 114 Perceived Health Disadvantages ...... 115 Negative Sexual and Reproductive Health ...... 115 Intimate Partner Violence and Related Long-Term Consequences ...... 117 Poor Mental Health and Psychosocial Well-Being ...... 120 Physical Health Conditions ...... 123 Perceived Health Advantages ...... 125 Having Children ...... 125 “Good health for me and my children” ...... 128 Access to Health Care ...... 129 Delaying First Birth and Birth Spacing ...... 131 Positive Mental Health and Psychosocial Well-Being ...... 133 Factors Influencing Perceptions of Early Marriage ...... 135 Husband’s traits ...... 135 Perceptions and experiences with support...... 136 Individual traits...... 138 Discussion ...... 141 Summary ...... 141 Perceived Health Disadvantages in Child Marriage ...... 141 Perceived Health Advantages in Child Marriage ...... 143 Contextual Factors Matter in Women’s Early Marital Experiences ...... 145
vi
Holding Multiple Truths ...... 146 Limitations ...... 148 Strengths ...... 149 Conclusions and Next Steps ...... 150 Chapter 6: Discussion and Conclusion: Revisiting Findings and Exploring Implications of Research on Girl Child Marriage and Women and Children’s Health and Well-Being ... 152 Summary ...... 152 Findings ...... 153 Contributions to evidence on women’s health and well-being ...... 154 Contributions to evidence on children’s health and well-being...... 155 Relevance of Findings ...... 156 Reconciling Results ...... 157 A Unique Situation for sub-Saharan Africa? ...... 158 Implications for Norm Change ...... 161 Methodological Contributions and Opportunities for Future Research ...... 163 Returning to Human Rights Framework ...... 165 Conclusion ...... 166 Appendix A: Correlations Between Girl Child Marriage (Binary) and Other Covariates ...... 168 Appendix B: Study Flow Chart on How Sample Size was Determined ...... 169 Appendix C: Histogram of Distribution of Age at Marriage Among Ever-married Women Age 20 to 49 Included in Final Sample (N=249,269) ...... 170 Appendix D: Scatter Plot of Mean Age at Marriage and Proportion Underweight by Country of Women Age 20 to 49 Included in Final Sample, with Fitted Line (N=249,269) ...... 171 Appendix E: Associations Between Girl Child Marriage (Categorical Specification) and Underweight for Pooled Analysis (N=249,269) ...... 172 Appendix F: Country-specific Associations by Girl Child Marriage Category and Underweight ...... 173 Appendix G: Regression Results of Unadjusted Associations between Girl Child Marriage and Underweight by Country...... 176 Appendix H: Child Marriage Among Mothers in Sample Compared to Population Rates ...... 177 Appendix I: Summary of Unadjusted Bivariate Analyses to Assess Associations between Girl Child Marriage and Child Development and Health Outcomes ...... 178 Appendix J: List of Supplementary Materials ...... 179 Bibliography ...... 180 Supplementary Materials ...... 207
vii
List of Figures
Figure 1.1: Original conceptualization of a bioecological model of girl child marriage…………14
Figure 1.2: Conceptual map of dissertation………………………………………………………18
Figure 2.1: Number of PubMed articles searchable using “child marriage” published, by year…26
Figure 3.1: Original conceptual model of how girl child marriage can impact female health and nutritional status...………………………………………………………………………………..44
Figure 3.2: Percent of women by age at marriage and country among ever-married women age 20 to 49 included in final sample (N=249,269)………………………………………………………51
Figure 3.3: Country-specific associations between girl child marriage and female underweight...57
Figure 3.4: Map and country-specific associations between girl child marriage and female underweight conditional on full set of covariates………………………………………………...58
Figure 4.1: Original conceptual model of framework for understanding the effects of girl child marriage on children’s early development and health……………………………………………71
Figure 4.2: Logistic regression results of girl child marriage and ECDI binary score, controlling for contextual factors, by country………………………………………………………………...88
Figure 4.3: Logistic regression results of girl child marriage and stunting, controlling for contextual factors, by country………………………………………………………………………………..89
viii
List of Tables
Table 1.1: 20 countries with the highest prevalence of girl child marriage among 20 to 24-year-old women……………………………………………………………………………………………..2
Table 3.1: Descriptive statistics of non-pregnant women aged 20 to 49 included in underweight analyses, by marital status (N=249,269)…………………………………………………………52
Table 3.2: Associations between girl child marriage and childbearing, secondary education, and household poverty (N=249,269)…………………………………………………………………54
Table 3.3: Associations between girl child marriage (binary) and underweight for pooled analysis (N=249,269)……………………………………………………………………………………...55
Table 4.1: Median age of marriage, prevalence of child marriage, and legal minimum age of marriage, by country (N=37,558)………………………………………………………………...78
Table 4.2: Demographic characteristics for children and their mothers in the sample by mother’s age at marriage…………………………………………………………………………………...80
Table 4.3: Percentage of children developmentally on-track and stunted by country (N=37,558)……………………………………………………………………………………….81
Table 4.4: Summary of logistic regression to assess the relationship between girl child marriage and if child is off-track for development based on ECDI binary score…………………………..83
Table 4.5: Summary of logistic regression to assess the relationship between girl child marriage and if their children are stunted based on HAZ…………………………………………………...85
Table 4.6: Summary of unadjusted and adjusted analyses on the relationship between girl child marriage and intermediate variables……………………………………………………………...90
Table 5.1: Participant characteristics……………………………………………………………109
Table 6.1: Demographic characteristics for Guinea and India…………………………………..160
ix
Preface
I first learned about child marriage as a Master of Science in global health student in
2009. I had taken Health and Human Rights and coordinated field-base data collection as part of the first integrated health and human rights program evaluation in Kenya. The idea that Western- led human rights communities created constructs through internationally-binding agreements that trickled down to other cultural settings intrigued me. I completed my Master thesis on child marriage and HIV/AIDS in Malawi, an African country with high rates of both, and felt conflicted when my study results did not reveal higher odds of HIV among women married as children as compared to adult-married counterparts. A tension existed between my empirical public health research and global human rights advocacy when my study results could not overwhelmingly push against a practice understood to violate the rights of girls and women.
Following graduation, I served as Monitoring and Evaluation Coordinator for an HIV prevention non-profit organization in Malawi. While living in the country I had remotely studied as a student, I maintained an interest in human rights. I was shocked, as a volunteer mentor for a college preparation program, that among my 19 students, only two were female. There must be something, I thought, preventing young women from taking advantage of the opportunity to participate in a program that (from a Westerner’s perspective) could be a ticket to success.
These questions of gendered access to opportunities, and the possible injustices that fuel it, continued to emerge when I worked for a private philanthropic organization whose grantmaking aligned with a human rights agenda. Yet as I reflected on who was at the decision-making table and who was not, it struck me that a small group was responsible for setting the norms and standards I assumed to be truth in my previous studies. I began asking more questions:
What is a child? Who decides this, and where? What rights do children really have?
x
What is adolescence, and how is this different from childhood?
What is the role of gender, and what does it mean to be a female versus male, at different
ages, in different settings?
Is marriage a choice, an expression of love, a social contract, a family decision?
What empirical evidence shows that the timing of a woman’s marriage impacts health
and well-being? Is this consistent across all contexts?
When I had the opportunity to return to Harvard for my doctoral degree, I knew that “girl child marriage,” a term that embodied these complex questions, was the topic for me to explore.
Thinking back to my doctoral work over the past five years, something deep within me drove my unrelenting commitment toward exploring the consequences of girl child marriage in
African contexts, as part of my broader research agenda on children, adolescents, youth, and women in adverse settings. My pursuit of this research topic may have been driven by the investments I saw in early childhood development and adult women, populations I have published research on in pursuit of advancing a research and policy agenda targeting a life- course approach to adolescence. My pursuit may have been driven by the dearth of rigorous evidence coming from sub-Saharan Africa despite the continued policy, human rights, and donor-driven efforts against child marriage in these contexts. My pursuit may have been driven by my own realization that I could have been just like the girl child brides I have centered my research on if my parents did not have the opportunities to pursue education and to leave Nigeria, building their lives – and eventually my own – in the U.S.
This dissertation serves as an opportunity for advancing conversations on the integration of research, implementation, and policy to use culturally-competent methods to improve the lives
xi
of young people affected by adversity. It is a public declaration to women like those who participated in my research that their voices are heard.
xii
Acknowledgments
I’ve heard it said,
That people come into our lives for a reason,
Bringing something we must learn,
And we are led,
To those who help us most to grow,
If we let them,
And we help them in return.
Well I don’t know if I believe that’s true,
But I know I’m who I am today
Because I knew you…
-Glinda the good witch, Wicked
Words cannot describe my gratitude to the people, spaces, and other forms of support that have enabled me to complete my dissertation and the Doctor of Science program.
I would like to thank my primary advisor, Professor Jacqueline Bhabha, for taking a chance on me a few years ago when I felt like an academic orphan; her guidance and support has pushed me to reexamine how my work connects with the human rights field and to reconsider how I use language. I would like to thank my other research advisors, Dr. Günther Fink, whose input and contributions strengthened my quantitative skills, and Dr. Paul Farmer, who continues to push me to think beyond traditional qualitative methods and beyond being a researcher. I would also like to thank mentors who ensured my academic success, particularly department chair Dr. Wafaie
Fawzi, who has regularly advised me over the last five years, and collaborator and friend Dr.
xiii
Gretchen Brion-Meisels, whose guidance and support has made me want to be a better researcher, educator, and mentor.
I am grateful to the financial seeds that have been sewn into my doctoral experience. I received training grant support from the Eunice Kennedy Shriver National Institute of Child
Health & Human Development of the National Institutes of Health under Award Number
F31HD090939 and the Initiative to Maximize Student Diversity GM055353-14 through the
National Institutes of Health. I would also like to thank several other sources of financial support including: the Pfizer and Boston Delta Foundation STEM Fellowship, the Harvard Kennedy
School Women and Public Policy Program, Harvard University Center for African Studies, and various funders from the Harvard TH Chan School of Public Health (including the Department of Global Health and Population, Barry R. and Irene Tilenius Bloom Fellowship, and the
Michael von Clemm Traveling Fellowship Fund).
Success on my dissertation research would not have been possible without many people, especially in the field. I would like to acknowledge the data sciences services at the Institute for
Quantitative Social Science, particularly Dr. Simo Goshev and Dr. Steven Worthington, for help with statistical questions, and Dr. Anita Raj, who served as external reviewer for this dissertation. Thank you to members of la presidence (Alpha, Komara, Cissé) who ensured I had a roof over my head and that I was safe over my three months in Guinea; Wiatta Thomas and
Sebastian Cajuste who brought me both research support and joy; Dr. Issiaga Daffe, Soufiana
Kabba, my dearest Hadjia, and other staff of what became my “Guinea office base”; Dr. Aissatou
Diallo and Dr. Hawa Manet, who served as research assistants at different stages of my fieldwork; and a very special group of women who shared their personal experiences with me and opened their homes and their hearts.
xiv
Finally, to my family and friends, I am honored to have you as a part of my journey and am grateful for your unwavering support. To my family (Femi, Dr. Kemi, Temi, and Tayo
Fadayomi): you remind me every day that I am never alone. Patrick Fangen, you have brought laughter, joy, and a new level of rationale thinking into my life. Dr. Kafui Adjaye-Gbewonyo, Dr.
Felicia Browne, and soon-to-be Drs. Jessica Fei and Sonia Alves, you have all been such important parts far beyond the writing communities we created. Jennifer White and Jennifer Gottesfeld, thank you for continuing to be there. I am beyond humbled that the list goes on…
To my mom, whose academic success, commitment to teaching and education, and unfailing love, support, and faith;
To Madame Fanta, who reminds me I have family in Guinea;
To Faith, who inspired me to change global injustices 15 years ago;
And to those that we lost along the way;
You have all made me a better researcher, a better advocate, and a better human being.
This is for you.
We thank God!
-Nigerian saying
xv
Chapter 1: Introduction
Background on Girl Child Marriage
Girl child marriage, or early marriage – defined as a formal union of a girl before the age of 18 – remains a pervasive practice globally. Worldwide, an estimated 650 million women living today married before the age of 18 (UNICEF, 2018b). Developing countries, particularly in the South Asia and sub-Saharan Africa regions, have the largest number of girl child brides
(UNICEF, 2018b). In fact, one in three girls in developing countries is married before age 18
(UNFPA, 2016). (See Table 1.1 for countries with the highest prevalence rates.) A range of international, regional, and national legal instruments, as well as legal documents, such as the
1989 Convention on the Rights of the Child (CRC) prohibit the exploitation or sale of children, including for marriage.
Girl child marriage is believed to persist for several reasons, with some differences across geographic and cultural contexts. Poverty and limited economic opportunities for a girl or her family (de Smedt, 1998; Jain & Kurz, 2007; Myers, 2013; Raj, Gomez, & Silverman, 2014;
Sabbe et al., 2013; Schlecht, Rowley, & Babirye, 2013), low education levels including illiteracy
(Erulkar, 2013a; Erulkar & Muthengi, 2009; Jain & Kurz, 2007; Myers, 2013; Raj, Gomez, et al.,
2014; Sabbe et al., 2013; Schlecht et al., 2013), absence or lack of enforcement of protective laws (Chandra-Mouli et al., 2013; Maswikwa, Richter, Kaufman, & Nandi, 2015), cultural practices directed at protecting the girl child and maintaining family honor (Mourtada, Schlecht,
& DeJong, 2017), and social norms (Haberland, Chong, & Bracken, 2003; Hampton, 2010) increase susceptibility to early marriage. Several negative consequences of girl child marriage have been identified including curtailed schooling and low literacy levels (Nasrullah et al., 2014;
Raj, Gomez, et al., 2014), excessive responsibility with respect to marital duties and childcare
(Nasrullah et al., 2014; Raj, Gomez, et al., 2014), and deleterious impacts on maternal and child health (Myers, 2013; Nasrullah et al., 2014; Raj, Gomez, et al., 2014).
Table 1.1: 20 countries with the highest prevalence of girl child marriage among 20 to 24-year-old women Married Married Country by 15 by 18 (%) (%) Niger 28 76 Central African Republic 29 68 Chad 30 67 Bangladesh 22 59 Burkina Faso 10 52 Mali 17 52 South Sudan 9 52 Guinea 19 51 Mozambique 14 48 Somalia 8 45 Nigeria 18 44 Malawi 9 42 Madagascar 12 41 Eritrea 13 41 Ethiopia 14 40 Uganda 10 40 Nepal 7 40 Sierra Leone 13 39 Democratic Republic of the 10 37 Congo Mauritania 18 37
Note. Percentages are as reported in the March 2018 update of the UNICEF Global Databases for Child Marriage (UNICEF, 2018c).
2
Girl Child Marriage and Women’s Health and Well-Being
A growing body of empirical research suggests that girl child marriage negatively affects married girls’ and women’s health. Relevant factors identified include: (1) maternal and reproductive health and health-seeking behavior; (2) mental health and psychosocial well-being; and (3) HIV status, risk factors, and knowledge. Maternal and reproductive health and health- seeking behavior outcomes strongly associated with girl child marriage include factors relating to reproductive health and family planning such as early age at first birth, decreased modern contraceptive use, short birth intervals between pregnancies, increased unwanted or unintended pregnancies, and increased likelihood of having a stillbirth or miscarriage (Godha, Hotchkiss, &
Gage, 2013; Kamal, 2012; Kamal & Hassan, 2015; Nasrullah, Muazzam, Bhutta, & Raj, 2013;
Raj, Saggurti, Balaiah, & Silverman, 2009). Earlier ages at marriage for women have also been associated with lower use of antenatal care, such as reduced number of antenatal visits, and lower likelihood of delivery by a skilled attendant (Godha et al., 2013). Such relationships have been identified overwhelmingly from data in South Asian countries.
A few studies have measured associations between girl child marriage, mental health, and psychosocial well-being, finding strong associations with the diagnosis of several mental health disorders (Le Strat, Dubertret, & Le Foll, 2011), suicidal ideation and attempts (Gage, 2012), as well as weak relationships between girl child marriage and items in measures of post-traumatic stress disorder, social reactions, abuse attributions, and esteem* (Wondie, Zemene, Reschke, &
Schroder, 2011). Closely linked, girl child marriage has been associated with increased odds of a
*As explained by Wondie et al. (2011), abuse attributions refer to “personal vulnerability, self- blame/guilt, dangerous world, and empowerment” (p. 309). These, as well as post-traumatic stress disorder and social reactions, were measured using the Children’s Impact of Traumatic Events Scale- Revised (CITES-R) while esteem was measured using the Rosenberg Self-Esteem Scale (RSES). 3
woman experiencing physical or sexual intimate partner violence (Raj, Saggurti, Lawrence,
Balaiah, & Silverman, 2010).
There are mixed findings on the association between girl child marriage and women’s objective (non-self-reported) health measures. Some studies have found that girl child marriage is associated with greater odds of HIV infection (Bruce & Rowbottom, 2004; S. Clark, 2004;
Cohen, 2004; Ferry et al., 2001); others found later ages at marriage to be associated with greater odds of HIV infection (Adair, 2008; Bongaarts, 2007; Efevbera, 2012). A country-level analysis did not find significant associations between girl child marriage and national rates of HIV yet did find significant associations to national rates of maternal mortality, fertility rate, and percentage of births with skilled birth attendant across 96 countries (Raj & Boehmer, 2013). A study on cervical cancer risk factors in rural China found that younger ages at marriages were associated with comparatively higher relative risk than those married after 21 years of age in a small sample
(Zhang, Parkin, Yu, Esteve, & Yang, 1989).
Interviews and focus groups with women who married as children, other married and unmarried women in communities, men, children, and local and government leaders have further described negative health outcomes associated with girl child marriage. These include reduced use or understanding of family planning (Mardi, Ebadi, Shahbazi, Esmaelzade Saeieh, &
Behboodi Moghadam, 2018; McClendon et al., 2017) as well as complications with pregnancy and childbirth, maternal mortality, infertility, disruptions in the menstrual cycle, and abortion
(Myers, 2013; Nasrullah et al., 2014; Raj, Gomez, et al., 2014). Additional studies in certain contexts have also captured accounts of frequent pain (Nasrullah et al., 2014), physical weakness
(Nasrullah et al., 2014), and malnutrition and hunger (Myers, 2013). Limited qualitative data asked women married as children about their own health experiences, though one study
4
identified that few women who married early had knowledge of negative health outcomes
(Nasrullah et al., 2014).
Girl Child Marriage and Children’s Health and Well-Being
Limited evidence suggests that girl child marriage negatively affects the health of children born to women who married as children. Measures of malnutrition including stunted growth and being underweight (Raj, Saggurti, Winter, et al., 2010) as well as increased odds of diarrhea in the last two weeks (Mashal et al., 2008) were significantly associated with their mother’s early marriage, although results were not consistent between studies. A country-level analysis found that girl child marriage was significantly associated with national rates of infant mortality across 96 countries (Raj & Boehmer, 2013). A recent study found increased risk of preterm birth among women married before 18 years as compared to married later (Rahman et al., 2018). Child health outcomes that have not been significantly associated with girl child marriage include acute respiratory infection, emaciation and wasting, low birth weight, and child mortality.
Additional evidence toward the possible relationship between girl child marriage and the health of their children comes from literature on adolescent motherhood and children’s health
(Santhya, 2011). Some studies found that earlier ages of motherhood were significantly associated with infant and child mortality even when controlling for sociodemographic variables
(Chen et al., 2007; Raj, McDougal, & Rusch, 2014), while other studies found no significant relationship once potential confounders were included in the empirical models (Lee et al., 2008;
V. Sharma et al., 2008). Lower gestational age and low birth weights were also more common among adolescent mothers (Chen et al., 2007; Kurth et al., 2010). One study hypothesized that the quality of prenatal care may explain the pathway (Chen et al., 2007), while more have cited
5
increased rates of preterm delivery and low birth weight among younger mothers to explain increased mortality (Chen et al., 2007; Kurth et al., 2010; Lee et al., 2008; V. Sharma et al.,
2008), particularly in sub-Saharan Africa (Elshibly & Schmalisch, 2008). A cross-national study found that child growth, measured as length- or height-for-age, was lower among younger mothers, yet most of these associations disappeared once controlling for confounders (Yu,
Mason, Crum, Cappa, & Hotchkiss, 2016). Additionally, stunting, a measure of chronic undernutrition, is associated with reduced child development, particularly along cognitive, learning, and physical domains (Miller, Murray, Thomson, & Arbour, 2015). Stunting is a result of risk factors including pre-term birth and low birth weight; environmental factors such as poor sanitation; child nutrition and infection factors such as diarrhea or nonexclusive breastfeeding; and maternal factors relating to nutrition, infection, and reproductive health (Danaei et al., 2016).
Investigating Girl Child Marriage and Health Consequences in Sub-Saharan Africa
Gaps
Even though the review of literature suggests that girl child marriage leads to adverse health and well-being for women and their children, several aspects of these relationships remain unclear. While most research has been conducted in South Asia, due to the high number of child spouses, sub-Saharan Africa has the highest prevalence of girl child marriage (UNICEF, 2018b;
Walker, 2012). In fact, 18 of the 20 countries with the highest prevalence among 20- to 24-year- olds are on this sub-continent (UNICEF, 2018c). Yet, the empirical evidence on the associations between girl child marriage and health outcomes remain limited in sub-Saharan Africa. Beyond maternal and reproductive health, mental health, and HIV, there are other health outcomes for women and their children that might be impacted by girl child marriage that have not yet been studied. For example, little is known regarding associations between girl child marriage and
6
malnutrition, a leading cause of maternal and child deaths (R. E. Black et al., 2013) and a neglected area of study (Horton & Lo, 2013).
Very limited research looks at the consequences of girl child marriage for children born to women who marry early beyond the context of early motherhood, and no identified studies empirically examine its association with early developmental outcomes of children. Most quantitative studies examine individual-level outcomes drawing from a single country or a few similar countries in a single region, raising questions about the comparability of results across heterogeneous locales. Additionally, qualitative perspectives on women’s experiences within child marriage remain limited, even in some of the contexts most affected. These create research gaps that prevent developing evidence-based policies to improve the well-being of women and their children.
Purpose
The overall objective of this dissertation is to use a mixed methods research approach to improve our understanding of the complex relationships between girl child marriage and the health and well-being of women and their children in sub-Saharan Africa. The author refers to health broadly as defined in the Preamble to the Constitution of the World Health Organization
(1948) as “a state of complete physical, mental, and social well-being and not merely the absence of disease or infirmity” in order to directly connect with overlapping concepts of well-being and development (UN World Health Organization Interim Commission, 1948, p. 100). This dissertation uses all available household data collected across sub-Saharan Africa from the
Demographic and Health Surveys and the Multiple Indicator Clusters Surveys to estimate the associations between girl child marriage and both maternal and child health, making an important contribution to the global evidence base in this field. This dissertation also uses
7
primary data collected and analyzed from the Republic of Guinea, a country with one of the highest rates of girl child marriage, to qualitatively understand the social construct of girl child marriage and perceptions on how it affects health and well-being.
The specific aims of this dissertation are:
Aim 1: To establish the extent to which girl child marriage contributes to socioeconomic
status and underweight, a measure of undernutrition, among women in sub-Saharan Africa
Aim 2: To understand the mechanisms through which girl child marriage affects the early
development and health of children born to women who marry early in sub-Saharan Africa in
order to quantitatively examine its intergenerational effects.
Aim 3: To qualitatively explore perceptions of the social construct of “girl child
marriage” among women in a community, and to explore the perceptions that women have of
the relationship between marriage, health, and well-being.
Contributions of This Study
This dissertation builds on the previous research, making several contributions within public health, global health, and beyond. It jointly analyzes data from several countries and focuses on nutritional status as a non-self-reported measure of health status in women, important toward supporting women’s development as well as understanding key risk factors for early mortality among their future children. This dissertation contributes to the conflicting data on how a mother’s age at marriage relates to their child’s nutritional status. It includes the first study, as far as the author knows, to empirically assess the relationship between girl child marriage and the development (across cognitive, language, socioemotional, and physical domains) of their children early in life. Very little theoretical work has been identified on the relationship between girl child marriage, health, and well-being, which this dissertation aims to contribute to. This
8
dissertation intentionally investigates how women who married as children view their marriages, as related broadly to their health and their children’s health in a setting where high rates of girl child marriage exist and published literature remains absent, leveraging a study using in-depth qualitative research methods.
Through exploring girl child marriage as a potential psychosocial risk factor and elucidating pathways through which women and children’s well-being may be impacted, this dissertation can impact research on other social risks that influence child health and development. This research highlights the importance of examining the intersections between population studies, public health, global health, sexual and reproductive health, child and adolescent development, child protection, gender studies, and human rights. Research findings provide evidence to guide population policies designed to ensure that women and children can achieve their developmental potential. Girl child marriage violates human rights and has, as this study will show, important public health implications.
Research Design and Frameworks
Mixed Methods Research Design
This multi-study dissertation employs a mixed methods research design. A mixed methods research design proposes that multiple paradigms can be used to understand research aims, and in fact, the combination of approaches can lead to a more comprehensive understanding than any one approach itself provides (Creswell & Plano Clark, 2007). A mixed methods research design also enables the strengths of different methods to be capitalized on, while providing alternative ways of researching and understanding to address the weaknesses of each method. Creswell and Plano Clark (2007) provide the following definition:
9
Mixed methods research is a research design with philosophical assumptions as well as
methods of inquiry. As a methodology, it involves philosophical assumptions that guide
the direction of the collection and analysis of data and the mixture of qualitative and
quantitative approaches in many phases in the research process. As a method, it focuses
on collecting, analyzing, and mixing both quantitative and qualitative data in a single
study or series of studies. Its central premise is that the use of quantitative and qualitative
approaches in combination provides a better understanding of research problems than
either approach alone. (p. 5)*
A mixed methods design is imperative in this research on girl child marriage and health in sub-Saharan Africa due to the complex nature of the research aims, while the use of both quantitative and qualitative research methods capitalizes on the strengths of each while balancing their limitations. Quantitative methods can be useful because they help to quantify an issue
(addressing “how much” or “whether”), generating data that can be considered more generalizable and reliable, allowing for standard comparisons. In positivist, scientific communities such as within the field of public health, the results of quantitative methods have traditionally been embraced. Yet quantitative methods often do not allow for a complete understanding of complex issues such as girl child marriage and often cannot provide interpretations for the underlying “why” or “how” in the results. On the contrary, qualitative methods allow for in-depth and more holistic explorations of complex topics and provide human perspectives that can illuminate why a phenomenon such as child marriage and its health
*According to Creswell and Plano-Clark (2007), there are distinctions between a method and a methodology. For this dissertation, the author clarifies that she uses the terms in the following ways: A method refers to the way one collects and analyzes data (e.g., quantitative method). A methodology refers to the selection of a research method (or methods) in a study, based on theoretical justification (e.g., using logistic regression, defining primary and secondary outcomes, etc.). 10
consequences exist. Yet qualitative methods rely on small sample sizes, often limiting generalizability and raising concerns of reliability that may not advance a global understanding.
This complementary relationship of multiple research approaches is described by Teddlie and
Tashakkori (2010) as cornerstone to mixed methods research, where researchers reject the
“either-or at all levels of the research process, which leads to methodological eclecticism (i.e., the researcher as a connoisseur of methods)” and subscribe to “the iterative, cyclical approach to research,” which uses both deductive and inductive techniques (p.16-17). In this dissertation, the quantitative results intend to provide a broad understanding of health and developmental outcomes of girl child marriage in sub-Saharan Africa, addressing existing literature gaps, while the qualitative results intend to support a more nuanced understanding of women’s perspectives through in-depth exploration of one cultural context. In other words, intentionally integrating both quantitative and qualitative methods promote theory generation and verification, a unique advantage of a mixed methods research design.
Specifically, in this dissertation, the author uses an interactive approach to a mixed method design. In other words, this study recognizes the different ways in which integrating both quantitative and qualitative elements are important and additive in the research purpose, theoretical framing, research questions, research methods, and assessment of validity (Maxwell,
2012; Maxwell & Loomis, 2003). In Phase One (Aims 1 and 2, addressed in Chapters 3 and 4), the author planned to examine quantitative data to empirically assess the relationship between age at marriage and health and development outcomes of women who marry early and their children, in comparison to those who marry after age 18. In Phase Two (Aim 3, addressed in
Chapter 5), the author planned to build from Phase One to design a qualitative research study; the author collected and analyzed qualitative data in Phase Two to understand how one local
11
community understands the concept of girl child marriage and how they perceive the relationship between child marriage as a social construct and the health of themselves and their children. In practice, although study design was sequential, the implementation of each research aim had overlap. The quantitative work in Chapter 4 was conducted from 2015-2017, the quantitative work in Chapter 3 was conducted from 2016-2018, and the qualitative work in Ch. 5 was designed in 2016 with data collection completion in 2017 and analysis through 2018. In the conclusion, the author considers both the quantitative and qualitative findings, with the overall objective of understanding the relationship between child marriage and the health and well-being of women and children in sub-Saharan Africa.
Importantly, the author acknowledges that using both quantitative and qualitative elements in this research may not result in a single, cohesive answer to the research aims. In fact, the use of different research methods and paradigms may lead to different conclusions. As mixed methods experts note, “divergent findings are valuable in that they lead to a reexamination of the conceptual frameworks and the assumptions underlying each of the two (QUAL and QUAN) components” (Teddlie & Tashakkori, 2003, p. 17). The author adopts the potential diversity of results as a strength of the mixed methods approach in understanding both regional and local health and developmental outcomes of child marriage in sub-Saharan Africa.
Other Frameworks Considered
The author also acknowledges theoretical frameworks that implicitly informed this research. These frameworks are not necessarily expanded on in the empirical chapters but are instead highlighted here because they informed the author’s development of the dissertation research aims, research questions, and methodological approach.
12
Bioecological Framework. This research has been informed by a bioecological framework. The bioecological framework is adapted from Urie Bronfenbrenner’s layered model to understanding the broader environment in which a child develops (Bronfenbrenner, 1977). It proposes that children do not develop in isolation or in a vacuum, but rather within networks of relationships. In the example of girl child marriage, the development of a girl could be understood to be comprised of the biosystem: a girl child’s unique characteristics (e.g., her age, health, individual behavior); microsystem: a girl child’s direct interactions with her immediate environment (e.g., school, family, peers, community); mesosystem: the linkages between the different microsystem settings of a girl child at a given time in life (e.g., a girl’s parents interacting with community leaders to negotiate a marriage); exosystem: the linkages between the mesosystem and greater social structures, which extend beyond direct inclusion of the girl
(e.g., community education that influenced community leaders’ understanding of the benefits, or harms, of marriage at early ages); and macrosystem: the greater context of culture, beliefs, and structures in which all these relationships take place, creating social norms (Figure 1.1). As this framework reveals, human development is complex and must be understood by considering multiple factors and relationships. This dissertation subsequently focuses on a woman’s marital age as one possible psychosocial risk factor for a woman’s own health, as well as her children’s health and development.
13
Figure 1.1: Original conceptualization of a bioecological model of girl child marriage
14
Biosocial Approach. This research also implicitly draws from a biosocial approach. The biosocial approach to global health calls for examining the interaction between biological and social factors (Hanna & Kleinman, 2013). This approach suggests that in an increasingly global world, health problems cannot be understood without accounting for biological, political, economic, sociocultural, and historical influences. The framework posits that exploring these interactions, using critical social theory, can provide a more comprehensive understanding of health than focusing on certain determinants alone (Hanna & Kleinman, 2013). In Chapters 3 and
4, this dissertation considers both biological components (e.g., women’s physical maturity as measured by age and age at childbearing) and social components (e.g., poverty and education level) of how girl child marriage may affect both women and their children’s health and well- being. This study, moreover, accepts Farmer and colleagues’ understanding that “‘Domestic’ and
‘national’ data often (perhaps always) obscure local inequities” (Farmer, Kim, Kleinman, &
Basilico, 2013, p. 5), using the complementary qualitative research aim to illuminate local perceptions of and experiences with girl child marriage.
A biosocial approach can be effectively realized through the use of critical social theory, including Peter Burger and Thomas Luckman’s theory on the social construction of reality,
Robert Merton’s theory on the unanticipated consequences of purposive social action, Max
Weber’s theory on the three modes of authority, and Michel Foucault’s theory on biopower
(Hanna & Kleinman, 2013). A biosocial approach may also illuminate experiences of suffering, such as through Arthur Kleinman, Veena Das, and Margaret Lock’s theory of social suffering or
Johan Galtung’s theory of structural violence, substantively expanded on by Paul Farmer (Hanna
& Kleinman, 2013). In this dissertation, the author acknowledges that the qualitative study (Aim
3, addressed in Chapter 5), for example, could be viewed through the lens of the Berger and
15
Luckman’s social construction of reality (1966). The social construction of reality proposes that
“knowledge” is created and legitimized in a social context, beginning with habitualization, or repeated actions which become normalized for an individual or set of individuals, and potentially developing into institutionalization, or the practice of habitualization among groups or types of individuals in a given context (Berger & Luckmann, 1966). In other words, a society determines what is considered “knowledge” at a given point in time, and what is valid or invalid, and this understanding takes place within the biological and social constraints of the world. In this study, the author has assumed that girl child marriage is a social phenomenon that has been normalized in Guinea, as illustrated through the frequency of young women who marry before the age of 18.
Moreover, through the example of girl child marriage, the author assumes that a social phenomenon may have biological consequences, and that biological issues such as women’s health and child development must be considered in relation to their social context as well.
Further application of these frameworks could be important for future research.
Conceptual Map
Consequently, the motivations and aims of this dissertation can best be understood through a conceptual map (Figure 1.2). The author’s primary interest is in the relationship between girl child marriage and health and well-being, which she seeks to understand through two conceptualizations of girl child marriage. First, girl child marriage is defined explicitly as a union before the age of 18, following international human rights guidelines. The author hypothesizes a relationship between this conceptualization of girl child marriage and women and children’s health and well-being, specifically focusing on the measures of health and well-being listed in the diagram. Second, girl child marriage is understood through a community’s perspective, which relies on understanding what a child is and what the purpose or motivations
16
for marriage are with a gendered lens. The author assumes that this social construction of knowledge informs women’s perceptions, and that these perceptions will also include local understandings of the relationship between girl child marriage and the health and well-being of women and their children.
As Figure 1.2 demonstrates, this dissertation is divided into several sections. Chapter 1 has introduced girl child marriage and what is known about its impacts on women and children’s health as well as the dissertation research aims and contributions to literature. It has presented a brief overview of the dissertation design and theoretical frameworks underlying this body of work, defining and explaining the selection of a mixed methods research design, providing an overview of two theoretical frameworks that informed the design of the three studies, and concluding with an original conceptual framework illustrating how the remaining chapters connect to one another.
In Chapter 2, a theoretical paper is presented, defining the construct of “girl child marriage” to situate the subsequent empirical chapters. To use such a politicized term requires greater explanation, thus the author interrogates relevant concepts and explains the decision to use this terminology throughout the dissertation. The author also contextualizes the use of the term with a brief historical context and draws primarily from human rights, developmental psychology, and demographic literature.
17
Figure 1.2: Conceptual map of dissertation
18
In Chapter 3, the author presents the first set of empirical findings to address Aim 1, estimating associations between girl child marriage and women’s chronic undernutrition in sub-
Saharan Africa among ever-married women. Using all available household data from 103
Demographic and Health Surveys (DHS) collected between 1991 and 2014, representing 35
African countries, this chapter first shows that child marriage is prevalent, as 55% of women married before age 18, and that it is associated with increased odds of early motherhood, decreased odds of completing secondary school, and increased the odds of being in the poorest
(compared to wealthiest) asset quintile, controlling for primary school completion and enumeration area (or community cluster) fixed-effects. Then, it is shown that girl child marriage was associated with reduced the odds of being underweight (Body Mass Index less than 18.5) both in unadjusted and fully-adjusted models. As a final step, several robustness checks are run including country-specific analyses. The chapter concludes by contextualizing results, encouraging empirically-driven policy decision-making, and calling for further research to understand the determinants of undernutrition and its relationship to socioeconomic status in this context.
Chapter 4 presents the second set of empirical findings to address Aim 2, examining the relationship between girl child marriage and children’s development and health to examine the intergenerational impact of a mother’s age at marriage. Using data from 37,558 mother-child pairs identified through 16 national and sub-national cross-sectional surveys across sub-Saharan
Africa conducted between 2010 and 2014 by the UNICEF Multiple Indicator Clusters Survey program, this chapter shows increased odds being off-track for development (measured using the
Early Childhood Development Index) and stunted growth among 3- and 4-year-old children born to women who married before age 18, compared to those whose mothers married later. Adjusted
19
models reveal that contextual factors (geographic location, primary education, and country fixed- effects) explained most of this relationship, and that early childbearing was not the sole pathway through which girl child marriage affected being on-track for the aggregate measure of child development and health. Final models revealed that disparities in advanced maternal education and wealth explained child development and stunting. The chapter concludes that there are intergenerational consequences of girl child marriage on her child's well-being, and that through association with other contextual, socioeconomic, and biological factors, marrying early does matter for child development and health. These findings resonate with existing literature and point toward important policy considerations for improving early childhood outcomes.
Chapter 5 presents the third set of empirical findings to Address Aim 3, using qualitative research methods, and informed by the previous two empirical chapters, to qualitatively understand the ways in which women married as children in Conakry, Guinea see their marriage as related to their health and their children’s health. Nineteen women whose first marriage occurred before age 18 participated in in-depth interviews, in French or local languages.
Additional “rich data” were collected including brief ethnographic interviews, observation, and participant observation. Rooted in grounded theory and using open coding, this chapter first provides a profile of a diverse group of women married as children. Then, the various perspectives women married as children had on their health are presented thematically. Women described several health disadvantages, falling into inductively-generated themes of negative sexual and reproductive health, intimate partner violence and long-term consequences, poor mental health and psychosocial well-being, and other physical health conditions. Women also described several health advantages, falling into inductively generated themes of having children, “good health for me and my children,” access to health care, delaying first pregnancy
20
and birth spacing, and positive mental health and psychosocial well-being. Further analysis revealed that perceptions of women’s experiences in child marriage corresponded with three factors: husband’s traits, individual traits, and perceptions and experiences with support. While the disadvantages described were consistent with existing literature, the perceived advantages are not frequently highlighted and may provide insights as to why child marriage in Guinea persists.
Chapter 6 revisits the research objective and aims, while providing a summary of key findings from the complete dissertation. Contributions of this dissertation to research, practice, and policy are highlighted. The author also describes some implications and future considerations that the dissertation inspires.
21
Chapter 2: Defining and Deconstructing “Girl Child Marriage” and Its Applications to Public Health Research
Although girl child marriage has been extensively debated in the quantitative and qualitative literature, several gaps remain. Chapters 3 and 4 address gaps in the larger evidence base through presenting quantitative results on associations between girl child marriage, women’s undernutrition, and child development and stunting in children under the age of 5 in sub-Saharan Africa, while Chapter 5 addresses gaps in knowledge of women’s perceptions on their child marriage and health through qualitative work. Before continuing, however, it is important to critically engage with what the author means by “girl child marriage.” In general,
“girl child marriage” refers to the marriage or union of a girl-child before her 18th birthday. It elaborates on the internationally-recognized definition of “child marriage,” or “early marriage,” defined by the United Nations Children’s Fund (UNICEF) as the union of an individual before age 18 (UNICEF, 2014, 2018b). Such terms have been popularized and become normative in the human rights community, used today among governments, non-governmental organizations, advocacy groups, popular media, and researchers alike.
Indeed, references to child marriage appear more frequently than just five years ago, when the author began conducting her dissertation research. A new Google alert regularly shares at least one new article on child marriage crossing disciplines including public health, education, and social sciences broadly. Even in the United States, popular media outlets such as NPR, the
New York Times, and most recently Teen Vogue have increasingly described this social phenomenon, describing its occurrence and consequences for women all over the world
(Belanger, 2017, 2018; Reiss, 2017; Sinclair, 2016).
Yet missing from the flurry of references is an interrogation of what the term “girl child marriage” means.* The author chose not to title the dissertation “early marriage,” though this term still encapsulates a judgement on when marital union is supposed to happen. The author chose not to title the dissertation “women’s age at marriage” because it lacked specificity in the marital age groups that are of interest in this research. The author chose not to title the dissertation “adolescent marriage” or “teenage marriage” because it lacks direct engagement with an expanding human rights and development agenda. As one thinks about its application to women and children’s health, one would be remiss not to acknowledge that it is a value-laden term that has a history and serves a purpose. It is more than just a term to be defined in a single sentence – as most publications do – and instead a set of constructs to further explore. This chapter provides a starting point for this conversation.
A Historical Journey on Terminology
The concept of the term “child marriage” appears to have strong roots in India, perhaps unsurprisingly, as it is a country with high rates that has led rigorous activism for over a century.
Some of the earliest discussions identified in published literature came from India, raising questions about marital unions that were early and with questions surrounding an ability to consent (Bhandarkar, 1893; Roy, 1888). In a letter to an American friend, Roy (1888) describes critical arguments for why child marriage in Hindu culture existed at the time, suggesting thousands of years of history to reconcile. He interchangeably referred to “child marriage” and
“infant marriage,” and though he never defined either term, he reported that 10% of girls and 3% of boys age 8 and below were married at his time of writing (Roy, 1888). Yet at the same time,
Roy explained that the age of consent for marriage in England was similarly low – 12 years for
*Parts of this chapter are inspired by previously unpublished work in Efevbera (2011). 23
girls and 14 years for boys – a considerable departure from England’s majority age of 21 at the time (Roy, 1888). This early reference serves as an important reminder that marriages at young ages were practiced worldwide, in Western and non-Western countries, in economically advantaged and disadvantaged countries, alike. It would not be until the Child Marriage Restraint
Act of 1929 that a legal framework for reconsidering age at marriage laws was formalized in
India, eventually leading to a revised Act that outlawed marriage of girls under 18 years and boys under 21 years in 1978 (M. Black, Haeri, & Moodie, 2001).*
The earliest references to the term “child marriage” in scientific articles in PubMed, a leading database for health-related research, emerged in 1955 and 1957, in the context of Israel and England respectively [personal analysis].* There were no publications again until 1978, and the very limited articles focused on India, neighboring South Asian countries, and the merits of preventing child marriage for population control. The first PubMed mention of child marriage in sub-Saharan Africa was an article in 1984, which hypothesized (though did not test) adolescent sexual exposure, heightened by the common practice of child marriage, may contribute to cervical cancer (Mati, Mbugua, & Ndavi, 1984). Published health-related research on child marriage remained sparse over the next decade. A few studies called attention to child marriage in specific African contexts in the early 1990s (Alabi, 1990; Okwudili, 1993), followed by a slight peak in related research in 1995.*
*Similarly, in England, advocacy in 1929 resulted in raising the minimum age of marriage. However, according to the UK Parliament, the minimum age was raised to 16 years for both girls and boys, where it remains, with parental consent, today (“The law of marriage,” 2018). *“Child marriage” was used as the search term in PubMed. *This peak is consistent with increasing discussions about protecting girls’ rights and promoting their sexual and reproductive health through preventing child marriage as well as the 1990 Convention on the Rights of the Child that laid out special protections for younger populations. 24
By 2000, leading international advocacy organizations based in Europe and the United
States promoted attention toward setting and enforcing a minimum marital age of 18 years, favoring the term “early marriage” in reference to this practice among both girls and boys (M.
Black et al., 2001; Forum on Marriage and the Rights of Women and Girls, 2000; Mathur,
Greene, & Malhotra, 2003). They called for a human rights-based approach toward protecting young people from early unions, which were increasingly recognized as harmful. The
International Center for Research on Women (ICRW) elaborated on its perspective of early marriage. It described that most countries adopted legal frameworks setting childhood until age
18, most countries adopted legal frameworks setting the minimum age of marriage until age 18, and yet the lack of enforcement led to high rates of early marriage, particularly in South Asia and across the African continent (Mathur et al., 2003).
It was not until 2006 that a seminal scholarly article presented an literature overview suggesting that child marriage, defined as before age 18, had health consequences in sub-Saharan
Africa (Nour, 2006). The author used the term “child marriage” in the paper’s title, synonymously referring to “early marriage” or “child brides” throughout the manuscript. Though acknowledging that, by definition, both boys and girls were affected, the disproportionate number of females led Nour (2006) to explicitly focus on girls. Since that time, publications on child marriage and health have grown exponentially; in 2017 alone, there were 28 publications in
PubMed (See Figure 2.1). Raj appears to be the first to publish using the term “girl child marriage” in her paper on child marriage in 2010 and in the title of a subsequent paper in 2012
(Raj, 2010; Raj, McDougal, & Rusch, 2012). Though she provides no rationale for her use of this
25
revised terminology, it makes sense as a framing tool, given her research focus in those studies on the consequences for only women married as children.*
30
25
20
15
Number of articles of Number 10
5
0 1955 1962 1969 1976 1983 1990 1997 2004 2011 2018
Year
Figure 2.1: Number of PubMed articles searchable using “child marriage” published, by year
Parsing Out “Marriage”
To understand what is referred to as “girl child marriage” requires an understanding of the construct of marriage itself. International legal frameworks have sought to define marriage
(Efevbera, 2011). As illustrated by the Convention to Regulate Conflicts of Laws in the Matter
*Personal communication with Anita Raj after preparing this manuscript clarified that her intent was two- fold. First, she sought to ensure child marriage was recognized as a gendered issue that disproportionately affected girls and because boys. Second, the term captured her research focus on only consequences for females, rather than males, in those analyses. 26
of Marriage, drafted by 12 European countries in 1902 (and later dissolved), and the Havana
Convention on Private International Law, drafted by 15 Latin American countries in 1928
(Schwelb, 1963), more than 100 years of international legal agreements have sought to legally define marriage and the rights it guarantees individuals in marriage. The 1926 Supplementary
Convention on the Abolition of Slavery, the Slave Trade, and Institutions and Practices similar to
Slavery also sought to ensure all individuals’ freedoms including that of women and children
(Sarich, Olivier, & Bales, 2016), creating foundations where marriage could only occur with each parties’ consent. These can be viewed as predecessors to understanding modern legal construct around the global definition of “marriage,” and more specifically “child marriage” and
“forced marriage.”
Today, legally-binding international conventions and treaties illustrate a global consensus on the rights and protections humans should be offered, including as related to a legal construct of marriage. The 1948 Universal Declaration on Human Rights provided the first internationally- agreed upon modern legal definition of marriage. Article 16 specifies that all “men and women of full age…have the right to marry and found a family,” that marriage is a union that can be formed with “free and full consent” of participants as well as dissolved, and, by interpretation, that there are rights and protections afforded to married individuals (UN General Assembly,
1948). The 1962 Convention on Consent to Marriage, Minimum Age for Marriage, and
Registration of Marriages reiterates the 1948 guidelines and explicitly calls for registration of marriages and eliminating “child marriages and the betrothal of young girls before the age of puberty,” though further explanation of what these unions are was omitted from this Convention
(UN General Assembly, 1962). A 1965 follow-up to this convention, the Recommendation on
Consent to Marriage, Minimum Age for Marriage, and Registration of Marriages, is the first
27
document to explicitly define “full age” as 15 years (UN General Assembly, 1965), establishing a legal benchmark for who is eligible to marry among signatories.
More recent international agreements further create legal frameworks for defining marriage. The International Covenant on Economic, Social and Cultural Rights (ICESCR) and the International Covenant on Civil and Political Rights (ICCPR), both drafted in 1966, provide additional guidelines for protecting human rights, particularly for women and children, though these were not enforceable until 1976 (UN General Assembly, 1966a, 1966b). Article 23 of the
ICCPR calls for marital union among those who consent and are of “marriageable age.” Other rights protected, such economic and social exploitation of young people (ICESCR, Article 10) and education (ICESCR, Article 13), begin to form a context in which attention must be paid to unions that violate these rights. The 1981 Convention on the Elimination of All Forms of
Discrimination Against Women (CEDAW) articulates that women, especially at younger ages, may be vulnerable in marital practices and prohibits discrimination of women; men and women have differential rights and “the marriage of a child” must be eliminated. More specifically, the
Convention calls for countries to set a minimum age of marriage and require formal registration
(UN General Assembly, 1979).
While marriage may seem clearly defined legally, anthropological and demographic literature reveal that marriage is not a straightforward concept. It is a practice symbolizing a union that can have different meanings in different contexts. Bell (1997) describes marriage as “a construction in a social space whose dimensions are defined by an articulation of rights and responsibilities” (p. 244). “In the structurally simplest case,” Bell writes, “marriage involves the entry of a man into a woman’s domestic unit” (Bell, 1997, p. 239). One could argue that in contemporary societies, this “entry” might be more figurative, rather than literal, as marital
28
arrangements can now take place with someone half-way across the world. Moreover, marriage is no longer exclusively considered as male-to-female relationships in all contexts. Regardless, marriage is an institution, or an established interpersonal relationship, that offers rights to those in this bond, and in traditional societies, will often involve others beyond the individuals themselves. The extent and strength of those rights, and subsequent responsibilities, may differ in different societies because rights only exist in the context of relationships with others (Bell,
1997).
Marriage, particularly in traditional societies, is rooted in socio-cultural and economic contexts involving the joining of two families, rather than just two individuals (Chinwuba, 2016;
Haberland et al., 2003). This union in many societies may have religious roots. One could point toward several religions (e.g., Christianity, Islam, Judaism) for guidance on religious principles that have informed marital practices. As one example, in the context of Hindu religion in India in the 1890s, Roy (1888) describes:
[Marriage] is expressly said to be a divine union. Christ said “What God hath joined
together, let no man put asunder.” We find Solomon calling the wife a “gift from the
Lord,” and in the marriage service appointed by the Church of England some one [sic] is
required to stand as the donor of the bride, as is the case in every Hindoo marriage.
“Marriage,” says an eminent doctor (Hindoo) of law, “is viewed as a gift of the bride by
her father or other guardian to the bridegroom.” The marital union is thus a divine union;
it is an act of God and not of man. It is apparent that marriage is not a civil contract, and
the consent is not the essence of it. The Roman Catholics regard it as a sacrament; so do
the Hindoos. (p.415, italics in original)
29
Indeed, Roy’s writings even today serve as a reminder that colonial legacies and an increasing mixture of cultures are redefining how many, including in African contexts, now conceptualize a marriage. A study in Uganda, for example, illuminates that although concepts like “love” and
“faithfulness” are often used synonymously with marriage in discourse, the reality of marriage, in practice, may be expressed in different ways (Agol, Bukenya, Seeley, Kabunga, & Katahoire,
2014). Marriage is perhaps best viewed as a process – a series of events, decisions, and rites – instead of a dichotomous categorization of a person’s relationship status (Haberland et al., 2003).
Importantly, some relationships considered marriage in local contexts may not result in a legal union (Haberland et al., 2003), creating a challenge for how one measures or documents marital status, particularly in research. Common or civil law provide legislation in support of some marriages, such as based on age and consent of the marrying parties, while customary law and religious teachings may allow divergent unions (M. Black et al., 2001); age at marriage is one way in which these differences manifest. The lack of agreement, even in a single community, of what constitutes a marriage is moreover complicated by a colonial legacy that implemented laws on marriage that directly clashed with customary law (M. Black et al., 2001).
Demographically, perhaps inspired by these challenges, marriage is today conceptualized to include a union through cohabitation. The Demographic and Health Surveys (DHS) program asks women to self-report on their marital status and counts “married women and women living with a partner” as currently married to generate nationally-representative data on marriage estimates (MeasureDHS/ICF International, 2013, p. 81). Similarly, the UNICEF Multiple
Indicator Clusters Surveys (MICS), another major international source of nationally- representative household data, asks women: “In what month and year did you first marry or start living with a man as if married?” or if that information is unknown, “How old were you when
30
you started living with your first husband/partner?” (Multiple Indicator Clusters Survey Program,
2014, p. 115). This conceptualization of marriage adopts an understanding of customary laws and local norms, providing a more comprehensive picture of how individuals themselves view their relationship status.
Definitions of marriage are constructed in different ways, including legally and socioculturally as briefly touched on above. Legal frameworks are perhaps a top-down approach to constructing definitions of marriage and have resulted in broad international policy agreement on the creation of harmonious regional agreements and national legislation. Subsequently, legal definitions contribute to understanding marriage civilly and across geographic boundaries, which has the advantage of being more tangible to conceptualize, practice, and enforce. Yet in societies as complex and multifaceted as those that exist today, where civil law only serves as one influence on how marriage is constructed and understood, marriage cannot be reduced to its legal definition alone. Although understanding marriage from only a legal perspective loses the customary, religious, and broader sociocultural contexts for which individuals in communities may understand and engage in the practice of marriage, it is among the most common ways to define marriage across contexts.
Parsing Out “Child”
The construct of “girl child marriage” also incorporates the concept of a “child,” another ambiguous and difficult term to define. Here, too, international legal frameworks have played an important role in creating shared global norms defining childhood. As early as 1924, the League of Nations, a precursor to the United Nations (UN) established after World War I to maintain world peace, identified children as an important and special population. A first framework was put forth stating that children “have inalienable rights and are not the property of their father” (A.
31
Sharma & Gupta, 1991, p. 12). Sharma and Gupta (1991) further point toward the 1959 UN
Declaration on the Rights of the Child, the 1979 declaration as the International Year of the
Child, and the 1990 World Summit on Children (A. Sharma & Gupta, 1991).
The 1990 Convention on the Rights of Child (CRC), a legally-binding international agreement signed and ratified by all but three countries, was a major turning point. Article 1 defines that “a child means every human being below the age of eighteen years unless under the law applicable to the child, majority is attained earlier” (UN General Assembly, 1990). The CRC creates a framework that establishes children, based on years of life, as a special population for whom specific rights should be granted and protected, which proposes something unique about younger human beings. However, while the CRC proposes the age of 18 years as a benchmark of adulthood; the age of child, or a minor, often legally varies across contexts.
An ethical, or moral, argument supports the establishment of a “child-adult” distinction.
As Schapiro (1999) argues, certain characteristics are ascribed to someone who is a child, which warrants a different treatment or perspective on their actions until they reach the adult threshold.
A child, Schapiro proposes drawing from Immanuel Kant, is “undeveloped” and “dependent”
(Schapiro, 1999, pp. 721-722). In other words, there is a level of immaturity children demonstrate, resulting in a lack of agency and requires additional support until they are able to reason and act independently based on these reasons.
Developmentally, a child is understood to achieve biological, cognitive, psychological, and social milestones over time (Efevbera, McCoy, Wuermli, & Betancourt, 2018). At different ages, children are expected to crawl then walk, talk and express themselves, process increasingly complex forms of information, and develop relationships with others. Children are not expected to behave as maturely as adult counterparts or to clearly assert and describe their own identities,
32
which are shaped overtime as the brain continues to develop and individuals interact with their contexts. Psychology theorists such as Sigmund Freud, Erik Erikson, Jean Piaget, Lev Vygotsky,
Urie Bronfenbrenner, John Watson, B.F. Skinner, Albert Bandura, and John Bolby have written extensively on child development across different domains. From their perspectives, childhood is a fluid construct evolving in an individual’s early lifetime.
In recent years, a further distinction of adolescence, referring to older children who are still not yet adults, has emerged. Defined as individuals aged 10 to 19 (World Health
Organization, 2018), adolescents biologically and socioculturally occupy a gray space between childhood and adulthood. Only recently have adolescents been recognized as distinct, more mature than children, yet less developmentally advanced than adults physically, cognitively, and socially (Bearinger, Sieving, Ferguson, & Sharma, 2007). The need to recognize older children as adolescents may be rooted in changing sociocultural contexts, with different expectations and norms for young people. As Bearinger and colleagues explain:
First, acknowledging wide cultural variation, adolescents are increasingly delaying
marriage—for some, to pursue education or employment options. Urbanisation has an
important role in this societal shift. Second, historically, societies expected childbearing
to follow shortly after marriage; now norms are shifting towards delayed childbearing.
These key changes, which affect all societies by varying degrees, have expanded the gap
between puberty and marriage, and between marriage and childbearing. (Bearinger et al.,
2007, p. 1221)
Importantly, what perhaps all of this points to, defining a “child” has evolved historically and in different spaces and places. These varying ideas of childhood and adolescence are relational concepts and their definitions at a given moment are influenced by culture, history,
33
local ideology, and different levels of law (Macleod, 2003). Macleod (2003) explains that childhood “is not a timeless, transcultural phenomenon” (p. 420). Instead, it should be understood “as the product of a number of cultural processes and modernist ideas, which have come to define a specific life stage as different from others and as in need of special treatment”
(Macleod, 2003, p. 420). Moreover, there may not be a clear or single trajectory from childhood to adolescence to adulthood, and trajectories may vary by context.
In reflecting on the construct of “girl child marriage,” a child cannot be simply viewed as less than adult: small, weak, young, helpless, little, irresponsible. In fact, such stereotypical ideas of a child are, in many cases and contexts, incorrect. Yet a reference to “girl child marriage” or
“child marriage” rather than “adolescent marriage,” “teenage marriage,” or even “early marriage” reminds an audience that a protected population, demonstrated to be less developmentally or socially mature than an adult, is the population being addressed. The author has used this term to explicitly connect this scientific research with the human rights and advocacy discussions, reminding readers that the focus is on social experiences young women in
African contexts have experienced preceding their adult life.
A child, as referenced in “girl child marriage,” should be understood to consist of individuals within the first couple of decades of life, afforded varying levels of legal responsibility and accountability, biological and sociocultural maturity, and agency in a cultural context at a given moment in time. Definitions of a child often overlap with an adolescent, pointing toward a unique period earlier in an individual’s life that is distinct across legal, biological, and developmental domains. Similar to the construct of marriage, international legal frameworks, setting a child at age 18, are most successful in crafting a shared definition of a
34
child that can be operationalized across settings, particularly in the context of consent for marriage.
The Need to Focus on Girls
This dissertation has elected to refer to “girl child marriage,” a departure from most publications by explicitly engaging with the construct of a “girl.” Child marriage, by definition, impacts both boys and girls (Strochlic, 2014), yet existing research and advocacy emphasize girls affected due to the prevalence of the practice and the intensity of its consequences for this population (UNICEF, 2014). Ages at marriage are lower for females, on average, as compared to male counterparts (United Nations Department of Economic and Social Affairs - Population
Division, 2017). Currently, no region is on-track to eliminate child marriage by 2030 to achieve
Sustainable Development Goal 5 Target 3, and nearly 650 million girls and women living today have been affected (UNICEF, 2018b). If rates of girl child marriage remain unchanged, 12 million girls under age 18 will continue to marry each year, in contrast to the prevalence of child marriage among boys, estimated to be one-fifth the level of girls (UNICEF, 2018a). Legal ages of marriage in countries often differ between girls and boys, illuminating a perceived difference at the level of government of when individuals are prepared to marry. In fact, 17 countries where the legal age for boys was higher than for girls were identified, including Mali (16 years for girls,
18 years for boys), Iran (13 years for girls, 15 years for boys), and India, where the majority of women married before age 18 reside (18 years for girls, 21 years for boys) (Women Business and the Law, 2017). Similarly, social expectations on the ideal age of marriage for males and females often differs. An in-depth qualitative study in Guinea revealed that among the 19 participating women married as children, the majority proposed that the ideal age of marriage for women was younger than that of men (Efevbera, 2017b). Thus, while recent literature
35
overwhelmingly focuses on the causes and consequences of marriage at younger ages for females, their use of the term “child marriage” masks an implicit gendered lens.
Indeed, while not discounting that boys also experience child marriage, the potential severity of its consequences makes a focus on girls important. Girl brides were traditionally considered more protected than unmarried counterparts, due to the perceived economic security and reduced risk of sexually-transmitted infections as a result of a perceived reduction in sexual partners (Bruce & Clark, 2003). However, increasing research illuminates that married adolescents may experience limited social support, restricted mobility, and lower levels of education (UNFPA, 2016). From the perspective of the health consequences a child spouse may encounter, the hypothesized pathways for girls and boys differ. Child marriage and early childbearing are closely linked, sometimes even discussed interchangeably. Of the 16 million 15- to 19-year-old girls who bear children each year, 90% are married (Loaiza & Wong, 2012).
Childbearing is not known to directly affect the health of boys through biological pathways yet is of the utmost concern for girls who marry. Additionally, the UN General Assembly recognized explicitly the role that gender inequalities play in the causes and consequences of girl child marriage today (UN General Assembly 71st session, 2017).
The use of the term “girl child marriage” in this dissertation, moreover, makes explicit what other scholars and practitioners who refer only to “child marriage” often fail to: that a focus on the causes and consequences of child marriage for only the girl-child has implicitly applied a gendered lens. The use of this term, in this dissertation, makes explicit the assumption of differences between males and females and identifies the population of focus in this research.
Krieger thoughtfully explains that a girl-boy distinction in health and medical research requires thinking beyond a biological division and engaging in the socially-constructed nature of gender,
36
a concept only introduced in the 1970s (Krieger, 2003). The author of this dissertation further suggests that beyond the experience of marriage, other variables often measured as related to child marriage including education level and wealth are likely shaped by gender norms, and that a girl’s value in her household, community, or society more broadly is influenced by gender norms in the context in which she lives. While further interrogation of this important perspective is beyond the scope of this research, the articulation of “girl child marriage” signals consideration of these socially-constructed norms.
Overlaps and Distinctions with Forced Marriage
Forced marriage refers to a formal union without the free and full consent of both parties
(Glinski, Sexton, & Meyers, 2015). In contrast to definitions of child, or early, marriage, it is not age-bound. Child and early marriage are considered forms of forced marriage by some because a child, by definition, is unable to provide free and full consent. Such perspectives have not yet been universally adopted. Moreover, there are different practices over the use of “child marriage,” “early marriage,” “forced marriage.” Certain agencies and initiatives use a combination of all three terms (Glinski et al., 2015; UN General Assembly 71st session, 2017;
UN General Assembly Human Rights Council 24th session, 2013). To some, “early marriage” serves as a euphemism that hides that children are involved in an act of marriage and “child brides” idealizes a problematic practice (Nour, 2009). To others, “early marriage” captures the
“premature nature of these marital unions” and encompasses an understanding of different cultural concepts of a child that the term “child marriage” misses (Myers, 2013, p. 6). The term
“forced marriage” promotes consideration for the reasons a girl marries early, importantly providing additional context for not only who is affected by why. While acknowledging the variability of legal constructs and socially-constructed understandings, this dissertation primarily
37
uses the term “girl child marriage,” or “child marriage” and “early marriage” among women synonymously, to focus on marriages where the girl-child spouse is below the age of 18. Forced marriage, in its full totality, is beyond the scope of this study.
Measuring Girl Child Marriage
With such possible variations in an understanding of girl child marriage, it is likely clear that measurement is imperfect. UN estimates use data from the DHS, UNICEF MICS, and other national surveys measuring child marriage as the “percentage of women 20-24 years old who were first married or in union before they were 15 years old” and the “percentage of women 20-
24 years old who were first married or in union before they were 18 years old” (UNICEF, 2017, p. 185). This convention has been set because the indicator for adolescent girl marriage
(marriage among girls ages 15 to 19) extends beyond the definition of a child as under age 18 and because measuring marital status among 15- to 17-year-old girls will miss counting women who marry post-survey but still before age 18 (UNFPA, 2012). Prevalence of child marriage is also measured through calculating the median age at first marriage among females, using
UNICEF MICS and UN Statistics Division data (Aspen Planning and Evaluation Program &
Girls Not Brides, 2015). Importantly, as earlier discussions on social constructs of marriage point toward, measuring girl child marriage includes both formal and informal unions. To this end,
UNICEF has importantly and recently articulated that living informally in union, or cohabitation, raises some of the same concerns as marital unions (UNICEF, 2018a).
Girl child marriage relies on self-reported data, which has raised concerns in its validity.
A study in India that compared self-reported age at marriage to two calculated indicators – derived by comparing current age to months in marriage as well as current age to months since menarche and marriage – concluded that self-reported age at marriage was an adequate measure
38
in the absence of additional data (Raj, 2017). Additionally, in many countries, lack of birth registration (confirming the age of the spouse) and lack of marriage registration (confirming a civil marriage and date) complicate a more nuanced perspective of early marital experiences.
Consequently, national marriage registries may severely underestimate the number of marriages considered to include girl spouses.
Finally, although this dissertation adopts the convention of using self-reported data from
DHS and UNICEF MICS to measure girl child marriage as a union before age 18, based on available data, the author further acknowledges aspirations for future measures. Presumably, a binary cutoff of age 18 omits important information about the consequences of marriage at age
15 or age 12; analyses in Chapters 3 and 4 assess health and developmental outcomes using additional categorical specifications of age at marriage. Others have recommended more complex indices that adapt from poverty measurement (Nguyen & Wodon, 2012).* The author of this dissertation encourages, in future efforts, more locally-driven research to better understand and measure girl child marriage from the perspective of those most affected. Chapter 5 provides a glimpse of related qualitative data, emphasizing the importance of understanding women’s own perspectives.
Conclusions and Implications for Public Health
Chapter 2 has presented a theoretical argument deconstructing “girl child marriage.” In doing so, the author has articulated consideration for concepts of marriage, children, and gender, which provide a broader background for empirical Chapters 3 through 5. Public health research
*Nguyen and Wodon (2012) drew from poverty indices to propose a headcount index of child marriage, equal to the number of girls married below the legal age of marriage divided by the total female population (married and unmarried); a child marriage gap, which accounts for the gap between a girl’s marital age and the legal age of marriage; and a squared child marriage age gap, which would place additional weight on girls who married much earlier than the legal age. 39
rarely engages with the broader social, economic, political, cultural, and historical dimensions of key concepts examine and measured. This chapter created space for more critical conversation on the meaning of “girl child marriage,” which will likely continue to evolve over time.
40
Chapter 3: Girl Child Marriage and Undernutrition: Evidence from 35 Countries in Sub- Saharan Africa
Summary
This chapter addresses Aim 1, to establish the extent to which girl child marriage contributes to socioeconomic status and underweight, a measure of undernutrition, among women in sub-Saharan Africa.*
Background
Globally, over 700 million girls and women alive today entered a formal union before age 18 (Lendhardt et al., 2016). In developing countries, one in nine girls marries before age 15, while one in three girls marries before age 18 (Loaiza & Wong, 2012). Girl child marriage, defined as a female entering a formal union before age 18, violates rights guaranteed in international and regional human rights instruments, and has been associated with adverse health behaviors and outcomes. Girl child marriage has been associated with increased fertility and reduced modern family planning, reduced antenatal care, and less safe delivery of their children
(Godha et al., 2013; Nasrullah et al., 2013; Raj et al., 2009). Literature has also documented significant associations between girl child marriage and mental health disorder diagnoses (Le
Strat et al., 2011); suicide ideation and attempt (Gage, 2012); and items in measures of post-
*This chapter was submitted to BMC Medicine in January 2018 and again, following revisions, in August 2018. It is currently under review. The final manuscript was co-authored by the author (first author) with guidance from her research committee. Authors' contributions were as follows: Yvette Efevbera conceptualized the project, reviewed literature, created detailed research strategy, conducted statistical analyses, and was the primary author of the manuscript. Jacqueline Bhabha contributed to the conceptual development of the project and reviewed the manuscript. Paul Farmer contributed to the conceptual development of the project and reviewed the manuscript. Günther Fink contributed to the conceptual development of the project, created the initial database, advised on research strategy and analyses, and edited multiple versions of the manuscript.
traumatic stress disorder, social reactions, abuse attributions, and self-esteem (Wondie et al.,
2011).
Though both ending child marriage and improving nutritional status are key items in achieving 2015 Sustainable Development Goals (SDGs) 2 and 5, much less is known regarding empirical associations between girl child marriage and undernutrition (Goli, Rammohan, &
Singh, 2015; Santhya, 2011). Globally, undernutrition is identified as the main cause for 3.5 million deaths in mothers and children and for 11% of Disability-Adjusted Life Years (DALYs)
(R. E. Black et al., 2013). Among adolescent girls, being underweight, or too thin for age and height, ranges from 1 to 10% across sub-Saharan Africa (Akseer, Al‐Gashm, Mehta, Mokdad, &
Bhutta, 2017). More than 10% of adult women are underweight in sub-Saharan Africa, and despite improvement from 1980 to 1995, where the rate dropped from 18% to 11%, the proportion of women underweight is more than double that of the Americas, Caribbean, and
Europe (R. E. Black et al., 2013). Sub-Saharan Africa also has the highest proportion of countries with girl child brides, as 18 of the 20 countries with the highest percentage of girls married before age 18 worldwide are found in the region (UNICEF, 2018c). If current trends in marriage and population growth continue, sub-Saharan Africa will account for the largest number of girl child brides by 2050 (UNICEF, 2015b).
The relationship between girl child marriage and undernutrition is not conceptually obvious. As illustrated in Figure 3.1, girl child marriage may influence nutritional status through direct and indirect pathways. Marrying at earlier ages is often associated with early and multiple childbearing (Williamson, 2013), leading to biological consequences for women’s nutritional status. Yet early pregnancy has been associated with both increased and decreased maternal weight in different contexts (Hediger, Scholl, & Schall, 1997; Rah et al., 2008), influenced by
42
number of previous childbirths, fat-storing patterns, and dietary intake; the hypothesis for nutritional depletion following adolescent pregnancy has not been consistently observed, revealing that the direction of the relationship between early childbearing and weight change is unclear. Girl child marriage may also affect nutrition through social pathways (Efevbera,
Bhabha, Farmer, & Fink, 2017). Marrying at earlier ages has been associated with lower educational attainment (Lloyd & Mensch, 2008), leading to more limited labor opportunities and income (Filmer & Fox, 2014), weakening women’s socioeconomic status and reducing autonomy. Both lower socioeconomic status and reduced autonomy will affect women’s decisions surrounding diet composition, physical activity, and health-seeking behavior (Goli et al., 2015). A transition to marriage, additionally, can directly influence women’s physical activity levels as they take on new social roles in their contexts (Eng, Kawachi, Fitzmaurice, &
Rimm, 2005). Yet it is unclear in what direction these subsequent changes in women’s behavior would affect their nutritional status. These mechanisms – early pregnancy, lower socioeconomic status, and reduced autonomy – may further lead to adverse health and developmental outcomes for children born to women who marry early (Efevbera, Bhabha, et al., 2017; Williamson, 2013).
Using data from 103 Demographic and Health Surveys (DHS) from 1991 to 2014, representing 35 out of 48 African countries, the aim of this study was to estimate associations between girl child marriage, adult socioeconomic status, and the likelihood of being underweight
(body mass index less than 18.5), as a measure of undernutrition, among women in sub-Saharan
Africa.
43
44
Figure 3.1: Original conceptual model of how girl child marriage can impact female health and nutritional status
Note. Model was developed through extensive review of existing literature on girl child marriage, human development, and nutrition.
Methodology
Data source
This study used data from the Demographic and Health Surveys (DHS) program.
Implemented since 1984 by ICF International under funding of the United States Agency for
International Development (USAID), DHS are cross-sectional household-based surveys administered by national statistical offices and designed to be representative at national, residence, and regional levels (MeasureDHS/ICF International, 2012). DHS typically use a stratified two-stage cluster sampling design, randomly sampling from clusters, or Enumeration
Areas (EAs), followed by households within each EA (MeasureDHS/ICF International, 2012).
All women ages 15 to 49 within selected households are invited to complete the Women’s
Questionnaire. In this study, we focused on DHS that collected height and weight data for women in sub-Saharan Africa.
Participants
Given the focus on medium- to long-term consequences of girl child marriage, we restricted analyses to women ages 20 to 49. According to World Health Organization guidelines, nutritional status in adolescents (age 10 to 19) is measured differently than adults and cannot be directly compared (WHO Expert Committee on Physical Status, 1995). For example, while body mass index is used to measure underweight in adults, low weight-for-age is traditionally used to measure underweight in children; measures of nutritional status in adolescents additionally accounts for age, and it is not until 19 years old that these curves approach convergence with adult models (Corsi, Subramanyam, & Subramanian, 2011; de Onis et al., 2007). We, thus, excluded women less than 20 years of age. Women over 49 years are not interviewed in DHS due to the survey’s historical focus on fertility and reproductive health. We also excluded survey
45
respondents who stated they were pregnant at the time of survey since accurate body mass index assessment would have been difficult and based on DHS convention (WHO Expert Committee on Physical Status, 1995). We further restricted the sample to ever-married women (currently or previously married), given interest in measuring associations between age at marriage and nutritional status. Full information for the exposure and outcomes of interest was also required for inclusion.
Measures
Girl child marriage, the exposure of interest, was defined as a self-reported formal union before the age of 18, based on international human rights guidelines (Organization of African
Unity, 1990). In DHS, women were asked if they were currently married or if they had ever been married. Women were then asked for the month and year they started living with their husband, or how old they were when they first started living with their husband if year could not be provided. We also considered an alternative categorical variable, comparing women married at age 18 years and above (adult marriage) with three child marriage age groups – below age 14, 14 to 15 years, 16 to 17 years – to test for differences by early and very early marital ages.
Underweight status was the primary outcome of interest for this study. The body mass index (BMI), or the ratio of weight (kg) to height (cm), is an objective measure of women’s nutritional status. Weight and height of women were collected in DHS by a trained measurer
(ICF International, 2012). Underweight is defined as having a BMI of less than 18.5. We also created a binary variable for being severely underweight (BMI less than 16) for additional analyses.
Secondary outcomes. We analyzed several potential mediators as secondary outcomes including age at first birth, number of children born, completion of secondary education, and
46
poverty. Secondary education was defined as an educational attainment of secondary school or higher. Poverty was defined as a household being in the lowest asset quintile as compared to the richest asset quintile. We also compared remaining asset quintiles (poorer, middle, and richer) to the richest asset quintile. Asset quintiles were used as proxies for household relative socioeconomic status within a country, and asset scores were computed using principal component analysis of six key household assets following the methodology outlined by Filmer and Pritchett (Rutstein & Johnson, 2004).
Other covariates. To reduce confounding, adjusted analyses controlled for self-reported completion of primary education or higher, age at time of interview (using 5-year age group), and cluster fixed-effects. In some of the empirical models, we also controlled for age gap, calculated by subtracting the woman’s age from her partner’s age to account for age differential, and education level gap, calculated by subtracting the woman’s highest categorical level of schooling from her partner’s, to further reduce confounding concerns. There were no concerns of high correlations between girl child marriage and all covariates, thus we included all variables in fully-adjusted models (Appendix A).
Statistical analysis
To assess the associations between girl child marriage and being underweight, a series of multivariable logistic regression models were estimated. In all models, we included EA fixed- effects to control for local differences in infrastructure, norms, and labor market opportunities that are likely to be correlated with both outcomes and predictors and could thus cause confounding biases. As a first step, we estimated the associations between girl child marriage and secondary outcomes: childbearing, educational attainment, and poverty. These variables were identified as the most likely mediators to undernutrition through an extensive literature
47
review, resulting in the conceptual framework represented in Figure 3.1. As a second step, and for our main analyses, we estimated the associations between girl child marriage and underweight, conditional on an increasing set of variables that may operate both as mediators and confounders in this context. Model 1 (M1) estimated the basic association between girl child marriage and underweight, conditional on EA fixed-effects and primary education completion.
We included primary education in the base model because, based on existing literature, we expect females in African contexts would have had the opportunity to complete primary schooling prior to their marriage, although secondary schooling would likely overlap, for some, with marriage. We thus conceptualize primary education as a predetermined confounding factor influencing the timing of a girl’s marriage (Lloyd & Mensch, 2008; Wodon, Nguyen, & Tsimpo,
2016), and secondary schooling (which we include in a later model) as often disrupted by early marriage instead (Efevbera, 2017e). Model 2 (M2) used the same confounding controls and additionally controlled for early and multiple childbearing as the most commonly described mediators in the literature from girl child marriage to underweight. Model 3 (M3) additionally adjusted for women’s completion of secondary education. Model 4 (M4) additionally included asset quintiles to account for levels of poverty and wealth. Model 5 (M5) additionally included age gap and education level gap to account for partner characteristics that may influence women’s autonomy and decision-making power. We refer to M5 as our fully-adjusted model.
Not all covariates were applicable to all participants (e.g., age at first birth, age gap with partner), thus we also included dummy variables for ever given birth and currently married at time of survey. These coefficients were not reported and were instead used to prevent dropping of censored data. The missing data for age at first birth and age gap with partner variables were imputed using the mean value. We estimated associations for the pooled data and by country.
48
Finally, we ran several sensitivity analyses and additional analyses for country-specific models to further examine findings. We estimated M1 to M5 with severely underweight as an alternative outcome to examine a more extreme measure of undernutrition. We ran analyses restricting the sample to 20- to 24-year-old women to examine results for younger generations in the sample. We ran analyses restricting data to only the most recent wave of data collection from
2011 to 2014 to examine if results were consistent in recent years. We ran analyses controlling for women’s work status in a sub-sample to explore the effect of women’s autonomy. To check for measurement error (i.e., to test if minor error in age reporting, which could lead to girl child marriage misclassification, would change study results), we ran analyses excluding women married at ages 18 and 19 to see if associations were present when accounting for more extreme ages of child and adult marriages. Also measured were associations between girl child marriage and being anemic to see if the same relationship held for another measure of undernutrition.
Results are presented as odds ratios, with 95% confidence intervals and p-values. Given concerns related to potential biases due to missing confounding variables in multivariable logistic models
(Greenland & Morgenstern, 1989), we also estimated alternative Poisson regression models as used in previous literature (Adjaye-Gbewonyo, Kawachi, Subramanian, & Avendano, 2018;
Cummings, 2009; Zou, 2004) (shown in Supplementary Table 3.1). Huber’s cluster-robust standard errors, which assumes that clusters are independent, were used to account for within- group correlation due to the complex survey design used in DHS (Huber, 1967).
Since we include EA fixed-effects (EA-specific intercepts) in all models, our empirical models exclusively explore within-EA variation, comparing women married at different ages within given EAs. From each EA, the DHS samples approximately 20 women who live in close proximity, typically a village in rural areas and a neighborhood in urban areas. These empirical
49
estimates should thus be interpreted as relative outcome differences of women marrying earlier in a community compared to other women from the same community marrying later. Analyses were conducted using StataMP 15 software.
Ethical approval
This study was determined exempt by the Harvard Longwood Medical Area Institutional
Review Board. Permission to use DHS data was obtained from ICF Macro/DHS Program.
Results
The final sample included 249,269 women across 35 African countries (Appendix B).
Fifty-five percent of women married before age 18, with variation in marital age across the sample (Appendix C). The percentage of women marrying before age 18 ranged from 19% in
Namibia to 80% in Niger (Figure 3.2). The median age at marriage for women who married as children was 15 years compared to 20 years for women who married as adults (Table 3.1). Table
3.1 shows differences in the outcome and sociodemographic characteristics comparing women who married as adults and as children. Overall, 18% of women in the sample were underweight while 2% were severely underweight. Sao Tome and Principe had the smallest proportion of women underweight (9%) while Ethiopia had the highest proportion (30%) (Appendix D).
50
100 80 60 Percent 40 20 0 i . . i r l in o d n ic d s p re a n ia a a o a ar w e e ia a e d o a a e n o l a o ep pi o b n ya th ri c ali u bia g d ip ga ne nia g d bi w en as u b h or oi o b a ine n e s la M iq i i er c e o an a o n b ro C R Re Iv i a m u so a a b m N ig an n T a m a B F ur e pu m . , th Ke e g a w ri en Le zil nz a b a B m o m o G Ga Gh G Lib M m N a a Ug Z in a Re g E L N R P S ra w T im k C e te d' da za d r S C D o a o n e Z , Con C M M Si Bur rican o Af ong C Tome a
Central Sao Age at first marriage
Below 14 years 14-15 years 16-17 years 18+ years
Figure 3.2: Percent of women by age at marriage and country among ever-married women age 20 to 49 included in final sample (N=249,269)
51
Table 3.1: Descriptive statistics of non-pregnant women aged 20 to 49 included in underweight analyses, by marital status (N=249,269) Among child Among adult marriage marriage women women (N=111,724) % (N=137,545) % Exposure Age at marriage (median) 20 15 Outcome BMI (median) 21.3 20.9 Underweight 18,736 17 25,035 18 Covariate Age at interview (median) 32 years 31 years Currently married 97,291 87 122,630 89 Mother's highest education level completed None 68,599 61.4 114,880 83.5 Primary 32,582 29.2 20,536 14.9 Secondary or higher 10,543 9.4 2,129 1.6 Total number of births (median) 3 5 Geographic location (urban) 34,111 31 29,217 21 Asset quintile Poorest 38,020 34 55,521 40 Poorer 20,279 18 26,716 19 Middle 20,059 18 26,792 20 Richer 16,310 15 16,693 12 Richest 17,056 15 11,823 9 Partner age (median) 39 40 Age gap with partner (median) 6 8 Partner education None 58,264 52 97,091 71 Primary 35,186 32 30,305 22 Secondary or higher 18,274 16 10,149 7 Education gap (mean) 0.162 0.188
Note. p<0.001 for difference between each variable in adult married and child married populations. Analyses clustered the standard errors at the country and survey cluster level.
52
Table 3.2 shows associations between girl child marriage and the secondary outcomes.
Conditional on primary education and EA fixed-effects, women who married before age 18 were
19 times more likely to give birth before age 18 (OR=18.7, 95% CI [18.1, 19.3], p<0.001) and four times more likely to have four or more children (OR=3.90, 95% CI [3.68, 4.13], p<0.001), compared to those who marred as adults. Girl child marriage was associated with a 69% reduction in the odds of completing secondary education compared to those who married as adults (OR=0.308, 95% CI [0.288, 0.329], p<0.001). Women who married before age 18 also had 12% increased odds of being in the poorest versus richest quintile compared to those who married as adults (OR= 1.12, 95% CI [1.06, 1.18], p<0.001).
Table 3.3 shows the multivariable regression results for being underweight. Conditional on primary education and EA fixed-effects (M1), women who married before age 18 had 8% reduced odds of being underweight as compared to women who married at age 18 or above
(OR=0.921, 95% CI [0.899, 0.944], p<0.01). Once we adjusted for childbearing and biological proxies (M2), the magnitude of the protective effect became marginally smaller (OR=0.933, 95%
CI [0.907, 0.960], p<0.01). When we further adjusted the model for secondary education (M3) and poverty (M4), coefficients changed very little (OR=0.930, 95% CI [0.904, 0.957], p<0.01;
OR=0.928, 95% CI [0.903, 0.955], p<0.01). The same held for models adjusted for the relative age and education of partners (M5) (OR=0.927, 95% CI [0.901, 0.954]). Appendix E shows that similar results are observed when estimating associations among early and very early ages at marriage. The magnitude of the effect slightly increases as age at marriage category increases, and results are not significant for the earliest age at marriage category by the final model.
53
Table 3.2: Associations between girl child marriage and childbearing, secondary education, and household poverty (N=249,269)
Panel 1: Age at first birth Variables Below 16 years old Below 18 years old (16+ years, ref.) (18+ years, ref.)
girl child marriage (18+ years, ref.) 11.0*** (10.5, 11.5) 18.7*** (18.1, 19.3) completion of primary education (no, ref.) 0.627*** (0.600, 0.656) 0.749*** (0.724, 0.775) Panel 2: Number of children Variables 1 child 2 to 3 children 4 or more children (none, ref.) (none, ref.) (none, ref.) girl child marriage (18+ years, ref.) 0.381*** (0.369, 0.394) 2.51*** (2.36, 2.67) 3.90*** (3.68, 4.13) completion of primary education (no, ref.) 1.27*** (1.16, 1.38) 0.944 (0.874, 1.02) 0.495*** (0.459, 0.533)
Panel 3: Secondary education Variables 54 Completion of secondary education or higher (no, ref.)
girl child marriage (18+ years, ref.) 0.308*** (0.288, 0.329)
completion of primary education (no, ref.) 8.62e+09*** (8.05e+09, 9.22e+09)
Panel 4: Poverty Variables Poorest quintile Poorer quintile Middle quintile Richer quintile (richest quintile, ref.) (richest quintile, ref.) (richest quintile, ref.) (richest quintile, ref.) 1.15*** (1.09, 1.21) girl child marriage (18+ years, ref.) 1.12*** (1.06, 1.18) 1.26*** (1.17, 1.35) 1.14*** (1.07, 1.20) completion of primary education (no, ref.) 0.257*** (0.239, 0.275) 0.283*** (0.260, 0.308) 0.361*** (0.337, 0.388) 0.490*** (0.461, 0.522)
Note. Coefficients presented are odds ratios from logistic regression models estimating associations between girl child marriage and possible mediators, with 95% CIs in parentheses, controlling for EA fixed-effects and primary education and clustering standard errors at sample cluster level. Girl child marriage is defined as marriage before age 18. In sample, 55% of women married before age 18. All variables are self-reported by participants. Asset quintiles were based on principal component analysis of household ownership of 6 core items (type of toilet facility, source of drinking water, main floor material, main wall material, and main roof material). Bolded values are significant at the p<0.05 level. *** p<0.01, ** p<0.05
Table 3.3: Associations between girl child marriage (binary) and underweight for pooled analysis (N=249,269)
Variables Model 1 Model 2 Model 3 Model 4 Model 5
Girl child marriage 0.921*** (0.899, 0.944) 0.933*** (0.907, 0.960) 0.930*** (0.904, 0.957) 0.928*** (0.903, 0.955) 0.927*** (0.901, 0.954) (18+ years, ref.) Completion of primary education 0.801*** (0.773, 0.829) 0.792*** (0.765, 0.821) 0.825*** (0.797, 0.855) 0.865*** (0.834, 0.896) 0.801*** (0.770, 0.833) (no, ref.) Current age (20-24 years, ref.)
25-29 years 0.934*** (0.901, 0.968) 0.936*** (0.903, 0.970) 0.942*** (0.908, 0.976) 0.938*** (0.905, 0.972)
30-34 years 0.939*** (0.902, 0.979) 0.942*** (0.904, 0.982) 0.950** (0.912, 0.990) 0.942*** (0.904, 0.981)
35-39 years 1.013 (0.967, 1.061) 1.014 (0.968, 1.062) 1.022 (0.976, 1.071) 1.007 (0.961, 1.055)
40-44 years 1.076*** (1.023, 1.133) 1.077*** (1.023, 1.133) 1.084*** (1.030, 1.141) 1.060** (1.006, 1.116)
45-49 years 1.187*** (1.123, 1.255) 1.186*** (1.121, 1.253) 1.193*** (1.129, 1.262) 1.155*** (1.092, 1.223)
Age at first birth (years) 0.995** (0.991, 0.999) 0.997 (0.993, 1.001) 0.996* (0.992, 1.000) 0.997 (0.993, 1.001)
Number of children ever born 0.963*** (0.957, 0.969) 0.963*** (0.956, 0.969) 0.962*** (0.956, 0.969) 0.964*** (0.957, 0.970)
55 Completion of secondary education 0.701*** (0.651, 0.756) 0.754*** (0.700, 0.813) 0.712*** (0.660, 0.768) (no, ref.) Asset quintile (poorest, ref.)
Poorer 0.933*** (0.901, 0.966) 0.942*** (0.909, 0.975)
Middle 0.866*** (0.837, 0.896) 0.879*** (0.849, 0.910)
Richer 0.776*** (0.744, 0.810) 0.795*** (0.761, 0.829) Richest 0.553*** (0.523, 0.585) 0.577*** (0.545, 0.610) Age gap between partner and woman 0.998** (0.997, 1.000) (years) Education gap between partner and 0.889*** (0.868, 0.911) woman (levels)
Note. Coefficients presented are odds ratios from logistic regression models. 95% CIs in parentheses are based on cluster standard errors. Underweight is defined as body mass index less than 18.5. M1 adjusts for sampling cluster and woman's primary education (as potential confounders). M2 adjusts for all covariates in M1 + woman's age, age at first birth, and number of children born (as biological mechanisms). M3 adjusts for all covariates in M2 + secondary education. M4 adjusts for all covariates in M3 + asset quintile. M5 includes remaining covariates. Bolded values are significant at the p<0.05 level. *** p<0.01, ** p<0.05
Figure 3.3 shows differences in the odds ratio and confidence intervals for girl child marriage by country. Conditional on primary education and EA fixed-effects (M1), eight countries across Eastern, Southern, and Western Africa showed marginally significant and negative associations between girl child marriage and underweight status. Once controlling for all covariates (M5), there were no significant associations between girl child marriage and underweight status in 30 of the 35 countries (Figure 3.4). Only Kenya, Lesotho, Madagascar,
Malawi, and Zambia revealed marginally significant and negative associations (Figure 3.4).
When country-specific models were run using the categorical specification of girl child marriage, even fewer countries revealed significant associations between early marriage and underweight status (Appendix F).
Sensitivity analyses controlling for women’s work status, restricting to more extreme ages of child and adult marriages, and examining anemia as alternative undernutrition outcome confirmed study results (Supplementary Figures 3.1-3.4). Moreover, additional sensitivity analyses suggest that associations were not significant overall at the country level. Girl child marriage was only significant and negatively associated with being severely underweight in one country, Tanzania. When restricting data to the youngest generation of participants (age 20 to 24) and the most recent wave of data collection (2011 to 2014), associations were significant and positive in only one country, respectively.
56
Study ID ES (95% CI) ssize
Benin 0.90 (0.81, 1.01) 14887 Burkina Faso 0.96 (0.86, 1.07) 11261 Burundi 1.17 (0.88, 1.55) 1464 Cameroon 1.20 (0.93, 1.54) 2843 Central African Republic 0.95 (0.70, 1.30) 1018 Chad 1.00 (0.90, 1.12) 10011 Comoros 0.65 (0.46, 0.94) 1412 Congo, Dem. Rep. 0.92 (0.80, 1.06) 5335 Congo, Rep. 0.88 (0.74, 1.05) 4498 Cote d'Ivoire 0.96 (0.65, 1.42) 945 Ethiopia 1.04 (0.97, 1.12) 20928 Gabon 0.70 (0.50, 0.99) 1398 Gambia 1.00 (0.77, 1.29) 1762 Ghana 0.86 (0.72, 1.01) 4237 Guinea 1.08 (0.90, 1.30) 5260 Kenya 0.79 (0.71, 0.88) 10953 Lesotho 0.62 (0.43, 0.90) 1009 Liberia 0.85 (0.69, 1.04) 3690 Madagascar 0.90 (0.81, 1.00) 9531 Malawi 0.79 (0.71, 0.89) 12362 Mali 0.97 (0.87, 1.07) 18879 Mozambique 0.93 (0.81, 1.06) 9391 Namibia 1.08 (0.83, 1.42) 2117 Niger 1.05 (0.86, 1.27) 5970 Nigeria 1.02 (0.94, 1.11) 33758 Rwanda 0.96 (0.80, 1.15) 6752 Sao Tome and Principe 0.84 (0.49, 1.45) 867 Senegal 0.86 (0.72, 1.01) 3666 Sierra Leone 0.99 (0.80, 1.21) 3299 Swaziland 0.39 (0.13, 1.14) 195 Tanzania 0.82 (0.73, 0.92) 10639 Togo 0.83 (0.68, 1.03) 3340 Uganda 0.85 (0.71, 1.02) 3913 Zambia 0.81 (0.73, 0.90) 14183 Zimbabwe 0.82 (0.70, 0.97) 7496 Overall (I-squared = 57.4%, p = 0.000) 0.90 (0.86, 0.94)
-1.55 1 1.55
Figure 3.3: Country-specific associations between girl child marriage and female underweight
57
Study ID ES (95% CI) ssize
Benin 0.96 (0.85, 1.09) 14887 Burkina Faso 0.98 (0.87, 1.11) 11261 Burundi 1.16 (0.81, 1.68) 1464 Cameroon 1.14 (0.86, 1.50) 2843 Central African Republic 0.95 (0.64, 1.40) 1018 Chad 1.00 (0.87, 1.14) 10011 Comoros 0.65 (0.41, 1.02) 1412 Congo, Dem. Rep. 0.92 (0.78, 1.08) 5335 Congo, Rep. 0.95 (0.78, 1.15) 4498 Cote d'Ivoire 1.14 (0.72, 1.82) 945 Ethiopia 0.98 (0.90, 1.07) 20928 Gabon 0.69 (0.47, 1.01) 1398 Gambia 1.09 (0.78, 1.53) 1762 Ghana 0.85 (0.69, 1.05) 4237 Guinea 1.18 (0.94, 1.47) 5260 Kenya 0.79 (0.69, 0.89) 10953 Lesotho 0.59 (0.39, 0.90) 1009 Liberia 0.94 (0.73, 1.21) 3690 Madagascar 0.88 (0.77, 0.99) 9531 Malawi 0.83 (0.72, 0.95) 12362 58 Mali 0.95 (0.85, 1.07) 18879 Mozambique 0.95 (0.82, 1.10) 9391 Namibia 1.07 (0.79, 1.44) 2117 Niger 1.11 (0.90, 1.37) 5970 Nigeria 0.97 (0.88, 1.07) 33758 Rwanda 1.01 (0.82, 1.25) 6752 Sao Tome and Principe 0.70 (0.40, 1.24) 867 Senegal 0.90 (0.73, 1.10) 3666 Sierra Leone 0.90 (0.71, 1.14) 3299 Swaziland 0.34 (0.09, 1.32) 195 Tanzania 0.87 (0.76, 1.00) 10639 Togo 0.89 (0.69, 1.15) 3340 Uganda 0.87 (0.70, 1.07) 3913 Zambia 0.80 (0.71, 0.90) 14183 Zimbabwe 0.86 (0.71, 1.04) 7496 Overall (I-squared = 35.9%, p = 0.020) 0.91 (0.87, 0.95)
-1.82 1 1.82
Figure 3.4: Map and country-specific associations between girl child marriage and female underweight conditional on full set of covariates
Note. Based on 35 country-specific models controlling for primary education, age, age at first birth, number of children ever born, secondary education, asset quintile, age gap, education gap, and EA fixed-effects (M5).
Discussion
Interpretation
The aim of this study was to estimate the empirical associations between girl child marriage, adult socioeconomic status, and the likelihood of being underweight in sub-Saharan
Africa. While results suggest that women who marry before age 18 have substantially increased risk of early and multiple childbearing, lower educational attainment, and living in poverty, analyses revealed that girl child marriage was associated with a slightly reduced risk of being underweight, with variation by country.
While the socioeconomic associations found in this study highlight important negative long-term consequences of girl child marriage, the lack of association with being underweight contradicts recent work from India. The only published study exploring associations between girl child marriage and undernutrition to date found that the percentage of women underweight (thin) was significantly higher among those who married or gave birth before age 18 in two states of
India (Goli et al., 2015). While the authors offer little explanation for their findings, the allusion is made to roles of early childbearing, rural environments, illiteracy, and poverty in their discussion. These broad patterns are also visible in the DHS data used in this paper when women from different communities and districts are compared to each other (Appendix G). However, once we controlled for cluster-level confounding, using EA fixed-effects, these associations reversed, yielding the negative and mostly insignificant associations. The pronounced difference between traditional cross-sectional estimates and models focusing on within-community comparisons only, like the ones presented here, suggests that the general cross-sectional relationship between girl child marriage and being underweight is severely biased by local contextual factors that jointly determine marriage and nutritional outcomes.
59
Findings additionally point toward differential experiences within marriage, particularly in the sub-Saharan African context. An analysis of gender differentials of undernutrition in sub-
Saharan Africa found unique differences from South Asia and cautioned comparisons between the two contexts (Svedberg, 1990). Svedberg (1990) posits that existing social customs including women’s participation in farming and marital practices such as early marriage, polygamy, and bride price paid by a groom’s family to his bride (rather than a dowry paid by the bride to a groom) are more common across sub-Saharan Africa and are distinctly different than the Asian context, possibly leading to more favorable nutritional status among women in these contexts.
These differences, moreover, may lead to differential impact of the potential pathways outlined in Figure 3.1 by context.
One possible explanation for study results is that marriage provides an opportunity for a woman to have access to more food of different nutritional content. Women who marry earlier may more quickly access these nutritional outcomes and do not present as undernourished.
Additionally, women who marry earlier, on average, give birth earlier, which has been associated with weight gain in some contexts (Scholl & Hediger, 1993). Early and repeated pregnancies in adolescent mothers in a long-term prospective U.S. study gained more weight than adult mother counterparts (Hediger et al., 1997). Similarly, weight change in childbearing and marriage may be associated with poorer nutritional status prior to marriage; a study in Pakistan found that women with increased levels of reproductive stress gained more weight, and that those who were malnourished at baseline gained more than marginally-nourished women (Winkvist, Jalil,
Habicht, & Rasmussen, 1994). In the current sample, inclusion of childbearing and biological proxies only minimally affected women’s odds of being underweight, thus no evidence was found for this mechanism.
60
Our results may also point toward the influence of contextual factors on marriage that further connect with a woman’s nutritional status, as identified in Figure 1. For example, social and cultural norms in some African contexts point toward preference for larger body size among women (Ettarh, Van de Vijver, Oti, & Kyobutungi, 2013; Tuoyire, Kumi-Kyereme, Doku, &
Amo-Adjei, 2017), which may extend into bride preference. A news article from Mauritania described the preference for larger girl brides, resulting in force-feeding girls in certain communities in the past (Smith, 2009). Evidence from developed countries further shows that married individuals are more likely to gain weight than unmarried counterparts, possibly explained by marital lifestyle changes (Eng et al., 2005; Sobal, Rauschenbach, & Frongillo,
2003). This study, thus, may capture consequences of changing marital practices and family patterns in African contexts in an increasingly globalized world. Some participants from an ongoing qualitative study on child marriage in Guinea similarly describe how their early marriages, and the security provided, has supported good eating, sleeping, health, and access to health care (Efevbera, 2017a). As such, it would be important to look at associations between girl child marriage and overnutrition outcomes such as obesity, which has been positively associated with marriage in the United States and is on the rise across the African continent. The double burden of malnutrition, another consequence of globalization, may mask negative nutritional and health outcomes in this study; we will explore these associations in a follow-up paper. Beyond sub-Saharan Africa, investigating the associations between girl child marriage and nutritional status would be important in any contexts where girl child marriage and undernutrition are prevalent, including South Asia, where this association has yet to be thoroughly explored.
Additional research will strengthen an understanding on the potential consequences of girl child
61
marriage, or the drivers of undernutrition, and could provide evidence for integrated or alternative programmatic and policy response.
Limitations
This study used cross-sectional data and thus cannot claim causality. Though we control for several variables identified in literature, residual confounding is possible. We were unable to control for other variables including pre-marital factors, such as anthropometry, diet and nutritional intake, and physical activity. Given that our data does not allow us to observe nutritional status prior to girl child marriage, it is possible that differential preferences for better- nourished brides or other contextual and sociocultural factors could be confounding factors in the observed relationships. We were also unable to control for childhood factors, such as early-life socioeconomic status and early-life nutrition, which likely impact women later in life. There may be variability in how data were collected across settings and timepoints. The analysis is also limited by the empirical approach which relies on EA fixed-effects and does not include DHS sampling weights. The main disadvantage of using EA fixed-effects is that only clusters with variation in both girl child marriage and underweight status are observed. The main disadvantage of not accounting for sampling weights is that results cannot be generalized to each country represented if the sample itself was not representative. Importantly, this study does not measure short-term effects of nutritional changes immediately following childbirth and may, consequently, only capture longer-term effects following adolescence. This study does not provide a fully-specified mediation model and analysis but rather reports associations with secondary outcomes of interest; formal mediation analysis in in this setting could be an important area for further research. Additionally, DHS data did not permit us to more thoroughly
62
investigate country-specific sociocultural and historical contexts that may have contributed to marginally significant outcomes, presenting opportunities for future research.
Strengths
Despite these limitations, this study has several strengths. We used a large dataset with nationally-representative data from 35 African countries, greatly increasing generalizability and external validity compared to existing single-country or small sample studies on health consequences of girl child marriage. The outcome of underweight was calculated using measurements externally-measured by trained individuals, rather than self-reported by participants, which could reduce bias. Data collection was standardized across countries and time points under the leadership of the DHS program, and in combination with the caution we exercised in pooling data, we are not overly concerned about data collection variability. The analytical strategy using EA fixed-effects also addressed concerns of residual confounding, as we control for observed and unobserved differences within narrow geographic settings which may affect both girl child marriage and underweight status. By focusing on within-EA variation, we eliminated potential biases due to differences in observable and unobservable factors common to all women in these geographic areas, including local population density, infectious disease environment, poverty levels, availability of food and nutrients, differences in enforcement or presence of laws, cultural and social norms, fragility or effects of conflict, and other environmental factors.
Conclusions
Girl child marriage remains a prevalent practice that threatens women’s socioeconomic status and rights more broadly. It continues to affect millions of women from adolescence, a time where rapid development takes place that has lasting impact. Findings suggest that girl child
63
marriage is likely not a major driver of female underweight, emphasizing the importance of using empirical data to guide program and policy decision-making. Further research is needed to understand the determinants of undernutrition in this context as well as the broader relationship between socioeconomic status and nutritional outcomes.
64
Chapter 4: Girl Child Marriage as a Risk Factor for Early Childhood Development and Stunting
Summary
This chapter addresses Aim 2, to understand the mechanisms through which girl child marriage affects the early development and health of children born to women who marry early in sub-Saharan Africa in order to quantitatively examine its intergenerational effects.*
Introduction
In 2007, more than 200 million children under age 5 in low- and middle-income countries
(LMICs) were estimated to fall short of internationally-accepted minimum outcomes for physical, educational, and cognitive development due to poverty, poor nutrition, and other social risks (Grantham-McGregor et al., 2007). A recent study estimated that nearly one-third of children aged 3 and 4 in LMICs were not on-track for socioemotional or cognitive development, with over 40% of those children in sub-Saharan African countries (McCoy et al., 2016). In the early years of life, competencies are built which may have long-term consequences for child well-being (Efevbera, McCoy, Wuermli, & Betancourt, 2017). Early development is shaped by biological and social factors innate within the child and the child’s greater environment
(Sameroff, 2009). Understanding early childhood outcomes such as poor health, nutrition, and
*This chapter was published in Social Science & Medicine, submitted in August 2016 and published online May 11, 2017. The published version can be cited as: Efevbera, Y., Bhabha, J., Farmer, P. E., & Fink, G. (2017). Girl child marriage as a risk factor for early childhood development and stunting. Social Science & Medicine, 185, 91-101.The final manuscript was co-authored by the author (first author) with guidance from her research committee. Authors' contributions were as follows: Yvette Efevbera conceptualized the project, reviewed literature, created detailed research strategy, created the database, conducted statistical analyses, and was the primary author of the manuscript. Jacqueline Bhabha contributed to the conceptual development of the project and reviewed the manuscript. Paul Farmer contributed to the conceptual development of the project and reviewed the manuscript. Günther Fink contributed to the conceptual development of the project, advised on research strategy and analyses, and edited multiple versions of the manuscript.
developmental delays is important because they can be intervened upon, when necessary, setting children on a positive trajectory for adulthood (Anderson et al., 2003; Engle et al., 2011).
Caregivers, especially mothers, play an important role in their children’s development
(Calkins & Hill, 2007; Efevbera, McCoy, et al., 2017). One maternal risk factor that may affect early childhood development (ECD) and health is a mother’s age at marriage. Girl child marriage, or early marriage, is defined as a formal union of a female before the age of 18
(UNICEF, 2014). It is often considered a violation of the rights of a girl according to international and regional human rights agreements and has been associated with lower educational attainment, economic opportunities, and some measures of health among young women (Erulkar, 2013a; Machel, Pires, & Carlsson, 2013; Nour, 2009; Svanemyr, Chandra-
Mouli, Christiansen, & Mbizvo, 2012).
Limited evidence describes the relationship between girl child marriage and child health and development. Mothers’ early marriage was significantly associated with child stunting, underweight (Raj, Saggurti, Winter, et al., 2010), and increased odds of diarrhea in the last two weeks (Mashal et al., 2008), although results were inconsistent across studies. A country-level analysis found that girl child marriage was significantly associated with infant mortality rates across 96 countries (Raj & Boehmer, 2013).
Empirically, literature on girl child marriage has been closely linked to literature on adolescent motherhood (Santhya, 2011). Some studies found that earlier maternal ages were significantly associated with infant and child mortality even when controlling for sociodemographic variables (Chen et al., 2007; Legrand & Mbacke, 1993), while other studies found no significant relationship once including potential confounders (Lee et al., 2008; V.
Sharma et al., 2008); few studies were in sub-Saharan Africa. Pre-term births resulting in lower
66
gestational age and low birth weight were also more common among adolescent mothers (Chen et al., 2007; Fall et al., 2015; Kurth et al., 2010). Stunting, a measure of chronic undernutrition as a result of risk factors including breastfeeding, maternal education, wealth, and child characteristics, was more likely among infants of adolescent mothers (Fall et al., 2015). Stunting is associated with reduced child development (Teller & Alva, 2008), particularly along cognitive
(Grantham-McGregor et al., 2007), socioemotional (McCoy et al., 2016), learning, and physical domains (Miller et al., 2015).
Very limited research examines the developmental and health consequences of girl child marriage in sub-Saharan Africa, where one in three girls marry before age 18 (UNICEF, 2015b) and some of the largest percentages of young children fall short of developmental and nutritional milestones (Grantham-McGregor et al., 2007; McCoy et al., 2016). Among the 20 countries with the highest prevalence of girl child marriage, 85% are on the African continent (Girls Not Brides,
2017). If current trends persist, sub-Saharan Africa will have the largest number of girl child brides by 2050 (UNICEF, 2015b), illuminating an important region to further investigate.
This paper quantitatively examines the intergenerational effects of girl child marriage, or the developmental and health outcomes of children born to women who marry before age 18.
The overall objective of this study is to understand the mechanisms through which girl child marriage affects the health and well-being of children born to women who marry early in sub-
Saharan Africa, as well as the relative magnitude and impact of these mechanisms. We used data from 16 national and sub-national cross-sectional surveys across sub-Saharan Africa conducted between 2010 and 2014 by the UNICEF Multiple Indicator Clusters Survey (MICS) program.
The Early Childhood Development Index was used to measure child development, while stunting was utilized to measure health. Two hypotheses were tested:
67
Hypothesis 1: Children born to women who married before age 18 have higher odds of being
developmentally off-track and of being stunted.
Hypothesis 2: Mother’s age at childbirth, completion of advanced education, and household
wealth are the primary mechanisms through which early marriage affects child development
and nutritional status.
To our knowledge, this is the first study to elucidate the relationship between girl child marriage and the development of children early in life across cognitive, language, physical, and socioemotional domains.
Framework
A framework for hypothesizing the relationship between girl child marriage and child development and health was developed, informed by existing literature and preliminary qualitative fieldwork (see Figure 4.1). Acknowledging the socioecological context in which children develop, including biological disposition, mother, family, community, and societal influences (Bronfenbrenner, 1977), this study focused on only one risk factor for child development and health: mother’s marital age.
Prior to marriage, several complex and interacting factors create environments in which women marry at earlier ages. A literature review conducted by the first author revealed that risk factors for girl child marriage include poverty (Chandra-Mouli et al., 2013; Human Rights
Watch, 2011; Wolfe, 2013), low education levels and maternal education (Erulkar, 2013a;
Erulkar & Muthengi, 2009; Loaiza & Wong, 2012), lack of laws or enforcement of laws
(Chandra-Mouli et al., 2013; Myers, 2013), cultural and social norms (Hampton, 2010; Nour,
2009; Walker, 2012), and conflict and fragility (Myers, 2013; Schlecht et al., 2013). Some of these same pre-marital risk factors may directly impact a child or a woman’s own health and
68
nutritional status (Gaur, Keshri, & Joe, 2013; Letamo & Navaneetham, 2014), which may directly impact her child’s risk for fetal growth restriction, malnutrition, and poor health at birth with potential long-term health and developmental consequences (Sawant & Venkat, 2013).
Once a woman is married, there are different ways in which she may be vulnerable to biological and social risks, which may subsequently affect her children’s development and health. Using a biosocial analysis, which posits that both biological and social processes interact and influence health and disease (Hanna & Kleinman, 2013), we theorize that through girl child marriage, both biological and social pathways determine her children’s well-being. Girl child marriage often leads to early childbearing, which may have biological consequences leading to poorer health and developmental outcomes for her young children (Williamson, 2013). Beyond biological influence, girl child marriage can also lead to social mechanisms that directly impact her children’s well-being. We conceptualize these direct social mechanisms as maternal caregiving behavior and decision-making, ultimately influencing child development and health
(The Urban Child Institute, 2014).
Maternal behavior may be impacted by early motherhood. Observational data has shown teenage mothers have worse health care behavior for themselves and their children, as compared to adult mothers (Legrand & Mbacke, 1993). Maternal behavior may also be impacted by education and wealth. While earlier primary educational attainment may influence when, why, and how a woman marries, pursuing and completing secondary schooling is often, in part, disrupted by early marriage, particularly in the context of sub-Saharan Africa (Lloyd & Mensch,
2008; Wodon et al., 2016). Early marriage was cited as a reason for up to 28% of secondary school dropouts in some African contexts (Lloyd & Mensch, 2008) and reported as a key reason for not completing secondary school among women in a recent qualitative study (Efevbera,
69
2017a). Lower educational attainment lowers labor market opportunities and average incomes
(Filmer & Fox, 2014), impacting wealth in the marital household. Mother’s advanced education, household income, and early childbearing, in turn, affect a mother’s knowledge and behavior, as well as the resources she has to act on her beliefs. This can lead to poorer nutritional status and medical care as well increased poverty (Dopkins Broecker & Hillard, 2009), initiating a cycle of continued poor health. For example, decisions to breastfeed, ensure that children are immunized, and access health care and early education when such resources are available will affect child development and health and may be impacted by mother’s maturity, knowledge, and ability to invest in related services.
Such maternal decision-making may also be impacted by a woman’s autonomy in her relationship as well as norms in her local sociocultural environment. These may further be complicated by experiences in marriage such as intimate partner violence (IPV), which is more likely to occur among girls who marry early (Raj, Saggurti, Lawrence, et al., 2010) and which may further lead to poor child development and health. A woman’s exposure to IPV may negatively affect her own nutritional, physical, and mental health as well as her children’s response to stress during pregnancy and infancy (Yount, DiGirolamo, & Ramakrishnan, 2011).
IPV has also been negatively associated with child stunting and underweight (Sobkoviak, Yount,
& Halim, 2012) and sub-optimal breastfeeding practices in African contexts (Misch & Yount,
2014), leading to negative health and developmental consequences in the child’s early years.
70
71
Figure 4.1: Original conceptual model of framework for understanding the effects of girl child marriage on children’s early development and health
The causes of girl child marriage are complex, and the potential mechanisms of consequences are rarely discussed in literature. In this study, we empirically adopted a simplified model. First, we tested if girl child marriage was associated with child development and health.
We then examined biological and social pathways that are directly linked to the mother’s age at marriage, and thus mediate the relationship between girl child marriage and child development and health. The three pathways examined were early childbearing (a proxy for biological mechanisms), mother’s completion of secondary education (a measure of advanced educational attainment), and wealth in the marital home (a proxy for resource availability in the marital home).
Methodology
Study Sample
This study used data from the MICS program for all sub-Saharan African countries where exposure and outcome data were available. MICS were first conducted in 1995. To date, 283 surveys have been conducted in 109 countries, collecting cross-sectional, nationally and sub- nationally representative data from women and men aged 15 to 49 (UNICEF, 2015a). MICS data are the most comprehensive compilation of ECD data in LMICs due to the addition of an ECD module in Waves 4 and 5 (Bornstein et al., 2012; UNICEF, 2013a). For this study, all surveys from the fourth and fifth rounds of MICS, conducted between 2010 and 2014 and available as of
February 20, 2016, with exposure and outcome data, were cleaned and aggregated into a single dataset. There were 11 national and 5 sub-national samples in the final dataset including: Central
African Republic, Chad, Democratic Republic of Congo, Ghana, Kenya (Mombasa), Kenya
(Nyanza Province), Madagascar (South), Malawi, Mauritania, Nigeria, Sierra Leone, Somalia
(Northeast), Somalia (Somaliland), Swaziland, Togo, and Zimbabwe. Data included nationally-
72
representative household and women information, as well as data for children under age 5 of mothers/caretakers in the sample who were randomly selected by survey administrators. The sample was restricted to children (aged 3 and 4) born to ever-married women (aged 15 to 49) across sub-Saharan African. Response rates for women were 93% on average, ranging from
88.2% in Somalia (Northeast) to 97.1% in Democratic Republic of Congo (UNICEF, 2016), shown in Supplementary Table 4.1. This study was determined as exempt by the Harvard
Longwood Institutional Review Board (Protocol #: IRB16-0723).
Measures
Girl child marriage, the exposure variable, was defined as a self-reported formal union before age 18. Following on previous studies (Raj & Boehmer, 2013; Raj, Saggurti, Winter, et al., 2010), we conceptualized girl child marriage as a binary variable in our main model. We drew the age limit from the normative framework regarding the legally-permissible age of marriage set out in Article 21(2) of the African Charter on the Rights and Welfare of the Child
(Organization of African Unity, 1990). This Charter has been signed by all members of the
African Union, signifying these states’ intent to not undermine the terms of the agreement. This
Charter has also been ratified by all but 7 members and thus creates a binding legal obligation for these ratifying states. In the MICS, women were asked: “In what month and year did you first marry or start living with a man as if married?” (Multiple Indicator Clusters Survey Program,
2014, p. 115). From these data, age at marriage was calculated by subtracting the year at first marriage from the year of woman’s birth. Women who did not provide a date at marriage were asked: “How old were you when you started living with your first husband/partner?” (Multiple
Indicator Clusters Survey Program, 2014, p. 115). This variable was used when the date of marriage was not available, per MICS survey guidelines. We also conducted robustness checks
73
using a categorical specification of girl child marriage (married at <15 years, 15-17 years, or 18+ years).
The Early Childhood Development Index (ECDI), a measure of child development across the domains of physical, learning, literacy-numeracy, and socioemotional development, was used to assess if a child was off-track for development (UNICEF, 2013a). The child development questions included were developed by UNICEF, in consultation with child development experts (UNICEF, 2013b). Existing measures were first reviewed, and proposed items were tested and validated in multiple countries and in multiple stages (UNICEF, 2013b).
These questions have been validated through confirmatory factor analysis across several LMICs
(Jeong, McCoy, Yousafzai, Salhi, & Fink, 2016; McCoy et al., 2016). Many of the questions in the ECDI are similar to items in validated tools including the Strengths and Difficulties
Questionnaire, Ages and Stages Questionnaire, and the Early Development Instrument (McCoy et al., 2016; UNICEF, 2013b). The ECDI has also been used in recent peer-review studies (Jeong et al., 2016; McCoy et al., 2016; Miller et al., 2015). We also conducted robustness checks using the child development score, a continuous ECD measure calculated as a score of 0 to 10, based on the number of on-track responses for each of the questions. (See Supplementary Figure 4.2 for additional information on the construction of the ECDI in this study.)
Stunting was defined as height-for-age z-score (HAZ) less than -2, based on World
Health Organization Growth Reference Standards (de Onis & Blössner, 2003). HAZ was selected over other measures to look at chronic, rather than short-term fluctuations, in malnutrition and was analyzed as a binary variable in our main model. We also conducted robustness checks using HAZ as a continuous measure.
74
Other covariates. To reduce confounding and to strengthen causal inference, adjusted analyses controlled for several factors identified in the framework of this study. For additional covariates, we followed variables controlled for in related literature (Mashal et al., 2008; Raj,
Saggurti, Winter, et al., 2010). Geographic location, urban or rural, was determined by mother’s self-reported household location. Mother’s age at childbirth was calculated by subtracting the child’s reported age from the mother’s reported age. Binary variables were then created to assess early childbirth if a mother gave birth to the child before age 16, between ages 16 and 17, or between ages 18 and 19. The number of children ever-born to the mother was calculated by mother’s self-reported number of total births. A child was identified as being born in a single or multiple birth by sorting children in the sample by mother ID. Gender, as male or female, and child’s age in months, as calculated by the birth year and month reported by the mother, were also included. Binary variables were generated to measure if a mother completed at least primary school and at least a secondary school. Wealth quintile was also compiled as a relative measure of poverty, using the MICS wealth score generated by comparing enumerator-identified household assets to number of household members (Rutstein & Johnson, 2004). Country fixed- effects were controlled for in all adjusted analyses to account for observable and unobservable differences at the national and sub-national levels.
Data Analysis
To assess the relationship between girl child marriage and child development and health, logistic regression models were run. Two binary outcomes were tested: ECDI, to determine if children were off-track for development, and stunting. Robustness checks were conducted by running linear regression analyses on the child development score and HAZ.
75
Drawing from the framework presented in Figure 4.1, a series of models were run to test pathways explaining associations between girl child marriage and each child development and health outcome. An unadjusted model (Model 1) was first run to compare the exposure of girl child marriage to each outcome. Model 2 (M2) controlled for geographic location and mother’s completion of primary education to assess how much of the relationship was explained by these pre-marital contextual risk factors. Model 3 (M3) added biological factors into M2 including mother’s age at childbirth, singleton/multiple birth, child’s gender, child’s age, and mother’s number of births to assess the relationship when controlling for early childbearing and possible biological pathways. The remaining models assessed causal pathways for the relationship between girl child marriage and child outcomes. Model 4 (M4) built on M3 by adding mother’s completion of secondary school to assess how much of the relationship was explained by advanced educational attainment. Model 5 (M5) built on M4 by adding household wealth quintile to assess how much of the relationship was explained by relative poverty, which is impacted by women’s secondary schooling. M2 to M5 also controlled for country fixed-effects to account for observable and unobservable country-level differences including, but not limited to, differences in fragility or effects of conflict, enforcement of laws, and cultural and social norms. Individual measures of national contextual factors were not included due to the limited availability of strong measures comparable across countries and because this was not the focus of the study.
Results are presented as odds ratios, and 95% confidence intervals are reported.
Hypotheses were rejected if the p-value < 0.05. We also conducted robustness checks and the
Sobel-Goodman test for mediation. We found no concerns of high correlations between our
76
exposure and remaining covariates (See Supplementary Table 4.3), and thus included all variables in the fully-adjusted models.
Results
Sociodemographic Data
This sample included 37,558 mother-child pairs. The median age of marriage for mothers included in this sample was 17 years old. Fifty-one percent of mothers were married before age
18, ranging from 17% in Swaziland, where the median age of marriage was 21 years old, to 75% in Chad, where the median age of marriage was 15 years old in the sample. Trends in marital age by country were consistent with national legislation on the minimum legal age at marriage for girls in these country contexts, as shown in Table 4.1 (World Policy Center, 2016). The prevalence of girl child marriage in the sample is also consistent with current national estimates based on retrospective reporting by women aged 20 to 24 years old. (See Appendix H.)
77
Table 4.1: Median age of marriage, prevalence of child marriage, and legal minimum age of marriage, by country (N=37,558)
Country Year of Sample Median Married before: Minimum age of marriage for girls: survey size age at Age 15 Age 18 Legally With Under marriage (%) (%) (years)a,b parental customary (years) consent or religious (years)b law (years)b Central African Republic 2010 3,170 16 25 64 18 18 18 Chad 2010 4,917 15 34 75 15 15 13 Democratic Republic of Congo 2010 3,269 18 10 45 18 18 18 Ghana 2011 2,525 19 5 33 18 18 18 Kenya (Mombasa) 2009 144 18 15 48 18 16 Unknown Kenya (Nyanza) 2011 1,770 17 15 53 18 16 Unknown Madagascar (South) 2012 977 15 30 75 18 18 18 Malawi 2013-14 6,552 17 10 51 18 15 Unknown Mauritania 2011 2,596 17 22 52 18 18 18 Puberty (Jigwa state); 16 (Akwa- Nigeria 2011 1,077 19 14 40 21 18 Ibom)
Sierra Leone 2010 2,457 16 18 59 18 18 No min. age Somalia (Northeast) 2011 1,543 18 12 44 18 16 18 Somalia (Somaliland) 2011 1,555 19 11 36 18 16 18
Swaziland 2010 579 21 3 17 21 16 Unknown Togo 2010 1,418 19 8 34 18 18 18
Zimbabwe 2014 3,009 18 5 34 18 16 No min. age Total 2010-14 37,558 17 16 51
a Exceptions can be made to lower the age of marriage in all countries except Chad, Democratic Republic of Congo, Ghana, Mauritania, Nigeria, and Swaziland. b Data from World Policy Center, 2016.
78
Table 4.2 shows demographic characteristics by mother’s age at marriage. Fifty-three percent of children in the sample were 36 to 47 months old while 47% of children were 48 to 56 months old. Half of the children surveyed were female. Most mothers in the sample were married at the time of the survey (91%). Mothers had a median age at birth for the child included in the sample of 27 years, ranging from 11 to 46 years. Nearly one out of every three children in the sample lived in an urban setting.
Table 4.3 shows the percentage of children that were on-track for development or who were stunted. Overall, 56% of children were on-track based on the aggregate ECDI. More than
50% of children were on-track for the developmental domains of learning, physical, and socioemotional. However, only 15% of children in the sample were on-track for literacy- numeracy. Variations by country can also be seen in Table 3. Nearly four out of every ten children in this sample were stunted. Seventeen percent of children were severely stunted (HAZ less than -3).
79
Table 4.2: Demographic characteristics for children and their mothers in the sample by mother's age at marriage
p for Among child Among adult difference marriage marriage by mother's Total mothers mothers marital age Covariate (N=37,558) % (N=19,206) % (N=18,352) % status Child characteristics Age p=0.028 3 years old (36-47 months) 19,990 53 10,108 53 9,882 54 4 years old (48-56 months) 17,568 47 9,098 47 8,470 46 Gender (female %) 18,795 50 9,658 50 9,137 50 p=0.440 Multiple birth 1,365 4 677 4 688 4 p=0.464 Mother characteristics Currently married 34,347 91 17,616 92 16,731 91 p=0.517 Mother’s age at birth (median) 27 25 28 p<0.0001 Mother's highest education level completed p=0.001 None/Preschool 15,776 42 9,193 48 6,573 36 Primary 14,018 37 7,806 41 6,212 34 Secondary or higher 7,773 21 2,206 11 5,567 30 Number of children born (median) 4 5 4 p<0.0001 Household characteristics Geographic location (urban) 10,796 29 4,821 25 5,975 33 p=0.003 Wealth quintile p=0.015 Poorest 8,753 23 4,703 24 4,050 22 Poorer 7,958 21 4,352 23 3,606 20 Middle 7,511 20 3,990 21 3,521 19 Richer 7,340 20 3,765 20 3,575 19 Richest 5,996 16 2,396 12 3,600 20
Note. Analyses clustered the standard errors at the country-cluster level.
80
Table 4.3: Percentage of children developmentally on-track and stunted by country (N=37,558)
Sample On-track for On-track by developmental domain Nutritional status Country size development Literacy- Socio- Severely Overall Learning Physical Stunted numeracy emotional stunted
N % N % N % N % N % N % N % Central African Republic 3,170 1,537 48 2,414 71 184 6 3,066 97 1,908 60 1,484 47 703 22 Chad 4,917 1,743 35 2,420 49 313 6 4,766 97 3,385 69 2,009 41 1,092 22 Democratic Republic of Congo 3,269 1,710 52 2,394 73 370 11 2,990 91 2,284 70 1,668 51 884 27 Ghana 2,525 1,858 74 2,256 89 572 23 2,471 98 1,914 76 686 27 228 9 81 Kenya (Mombasa) 144 101 70 137 95 76 53 122 85 83 58 38 26 11 8 Kenya (Nyanza) 1,770 1,171 66 1,678 95 593 34 1,580 89 1,113 63 593 34 257 15 Madagascar (South) 977 694 71 846 87 68 7 942 96 802 82 N/A N/A N/A N/A Mauritania 2,596 1,641 63 2,396 92 550 21 2,498 96 1,562 60 860 33 380 15 Nigeria 1,077 671 62 994 92 203 19 1,035 96 641 60 370 34 166 15 Sierra Leone 2,457 1,182 48 1,890 77 247 10 2,245 91 1,523 62 1,280 52 681 28 Somalia (Northeast) 1,543 554 36 1,111 72 270 18 1,215 79 698 45 N/A N/A N/A N/A Somalia (Somaliland) 1,555 934 60 1,363 88 397 26 1,378 89 974 63 N/A N/A N/A N/A Swaziland 579 365 63 546 94 98 17 572 99 351 61 149 26 39 7 Togo 1,418 778 55 1,134 80 133 9 1,346 95 968 68 520 37 169 12 Malawi 6,552 4,118 63 5,345 82 1,138 17 6,009 92 4,898 75 2,775 42 1,040 16 Zimbabwe 3,009 1,858 62 2,699 90 309 10 2,898 96 1,975 66 739 25 177 6 TOTAL 37,558 20,915 56 29,623 79 5,521 15 35,133 94 25,079 67 13,171 39 5,827 17
Bivariate Models on Associations Between Girl Child Marriage and Child Development and Health
Bivariate analyses showed a significant relationship between girl child marriage and child development and health (see Tables 4.4 and 4.5 for unadjusted and adjusted models by outcome and Appendix I for unadjusted bivariate models). Unadjusted M1 showed a significant relationship between a mother marrying before age 18 and her child being off-track for overall development, as determined by the ECDI (Table 4.4).
The odds of being off-track for development were 25% higher among children born to women who married before age 18 as compared to children born to women who married later
(OR=1.25, 95% CI [1.20, 1.31], p<0.001). Unadjusted M1 for stunting similarly showed a significant association between a mother marrying before age 18 and her child’s nutritional status, as measured by being stunted. The odds of being stunted were 29% higher among children born to women who married before age 18 as compared to children born to women who married later (OR=1.29, 95% CI [1.23, 1.35], p<0.001). Similar relationships were observed for the continuous measures of child development score and HAZ, and for an alternative specification of girl child marriage, where younger marital ages were associated with increased odds of poorer child outcomes (see M1 in Supplementary Tables 4.4, 4.5, and 4.6).
82
Table 4.4: Summary of logistic regression to assess the relationship between girl child marriage and if child is off-track for development based on ECDI binary score
Model 1 Model 2 Model 3 Model 4 Model 5 OR CI OR CI OR CI OR CI OR CI Exposure variable Child marriage (<18 years) 1.25 (1.20, 1.31) 1.05 (1.01, 1.10) 1.04 (0.993, 1.09) 1.01 (0.968, 1.06) 1.01 (0.964, 1.06)
Contextual covariates
Geographic location (urban) 0.752 (0.715, 0.791) 0.755 (0.718, 0.795) 0.790 (0.749, 0.831) 0.934 (0.879, 0.993)
0.820 (0.778, 0.865) 0.830 (0.786, 0.875) 0.893 (0.844, 0.945) 0.928 (0.877, 0.982) Mother completed primary school
Biological covariates
Multiple birth 1.12 (0.998, 1.25) 1.13 (1.008, 1.26) 1.14 (1.02, 1.27)
Gender (female) 0.843 (0.808, 0.880) 0.844 (0.809, 0.880) 0.843 (0.808, 0.879)
83 Age (months) 0.971 (0.968, 0.974) 0.971 (0.968, 0.974) 0.971 (0.968,0.974)
Number of children ever-born to 1.01 (1.004, 1.02) 1.01 (0.999, 1.02) 1.01 (0.997, 1.02) mother
Hypothesized pathways
Mother's age at childbirth
<16 years 1.01 (0.846, 1.22) 0.992 (0.827, 1.19) 0.965 (0.805, 1.16)
16-17 years 1.09 (0.981, 1.21) 1.08 (0.969, 1.20) 1.06 (0.951, 1.18)
18-19 years 1.04 (0.949, 1.13) 1.03 (0.942, 1.12) 1.01 (0.928, 1.11)
Mother completed secondary 0.760 (0.712, 0.812) 0.825 (0.772, 0.883) school
Wealth quintile (reference: Poorest) Poorer 0.895 (0.839, 0.953) Middle 0.825 (0.772, 0.881) Richer 0.777 (0.724, 0.834) Richest 0.614 (0.563, 0.670)
N 37,558 37,557 37,430 37,430 37,430
Table 4.4 (Continued)
Chi-sq 118.78 2252.28 2600.12 2668.4 2795.51
Note. Coefficients presented are odds ratios from logistic regression models with 95% CIs in parentheses. Off-track for ECDI is defined by not being on-track for at least 50% of questions in at least three of four developmental domains. M1 is unadjusted model. M2 adjusts for urban/rural, mother's primary education, and country fixed-effects. M3 adjusts for all covariates in M2 + child characteristics (gender, age, multiple birth), number of children ever born to mother, and mother's age at childbirth. M4 adjusts for all covariates in M3 + mother's secondary education. M5 adjusts for all covariates in M4 + wealth quintile. Bolded values are significant at the p<0.05 level.
84
Table 4.5: Summary of logistic regression to assess the relationship between girl child marriage and if their children are stunted based on HAZ
Model 1 Model 2 Model 3 Model 4 Model 5 OR CI OR CI OR CI OR CI OR CI Exposure variable Child marriage (<18 years) 1.29 (1.23, 1.35) 1.12 (1.07, 1.17) 1.08 (1.03, 1.14) 1.04 (0.992, 1.09) 1.04 (0.986, 1.09) Contextual covariates Geographic location (urban) 0.683 (0.646, 0.721) 0.684 (0.647, 0.723) 0.728 (0.688, 0.771) 0.925 (0.866, 0.987) 0.732 (0.692, 0.775) 0.731 (0.690, 0.774) 0.807 (0.760, 0.857) 0.838 (0.789, 0.891) Mother completed primary school? Biological covariates Multiple birth 1.16 (1.03, 1.30) 1.17 (1.04, 1.32) 1.19 (1.06, 1.34) Gender (female) 0.891 (0.852, 0.933) 0.892 (0.853, 0.933) 0.891 (0.852, 0.933) Age (months) 0.990 (0.987, 0.993) 0.990 (0.987, 0.993) 0.990 (0.987, 0.993) Number of children ever-born to 1.01 (0.999, 1.02) 1.00 (0.992, 1.01) 1.00 (0.991, 1.01) mother 85 Hypothesized pathways Mother's age at childbirth <16 years 1.36 (1.13, 1.64) 1.32 (1.09, 1.59) 1.28 (1.06, 1.54) 16-17 years 1.25 (1.12, 1.40) 1.23 (1.10, 1.37) 1.20 (1.08, 1.34) 18-19 years 1.14 (1.04, 1.25) 1.13 (1.03, 1.23) 1.11 (1.01, 1.22) Mother completed secondary 0.685 (0.640, 0.734) 0.770 (0.718, 0.827) school? Wealth quintile (reference: Poorest) Poorer 0.961 (0.899, 1.03) Middle 0.895 (0.835, 0.959) Richer 0.780 (0.724, 0.841) Richest 0.524 (0.477, 0.576)
N 33,483 33,482 33,359 33,359 33,359 Chi-sq 128.89 1477.16 1570.57 1687.29 1833.53
Note. Coefficients presented are odds ratios from logistic regression models with 95% CIs in parentheses. Stunted is defined as height-for-age z-score less than -2. M1 is unadjusted model. M2 adjusts for urban/rural, mother's primary education, and country fixed-effects. M3 adjusts for all covariates in M2 + child characteristics (gender, age, multiple birth), number of children ever born to mother, and mother's age at childbirth. M4 adjusts for all covariates in M3 + mother's secondary education. M5 adjusts for all covariates in M4 + wealth quintile. Bolded values are significant at the p<0.05 level. Results exclude Madagascar (South), Somalia (Northeast), and Somalia (Somaliland).
Adjusted Models on Associations Between Girl Child Marriage and Child Development and Health
ECDI. Once we controlled for contextual factors, the magnitude of the effect of girl child marriage on child development decreased from OR=1.25 to OR=1.05, suggesting that the increased risk seen in M1 was explained by these three factors. A mother’s completion of primary education and living in an urban setting explained much of the effect of girl child marriage and both were significantly associated with lower odds of a child being developmentally off-track. Figure 4.2 shows interesting variations of M2 by country, though estimates are less precise with smaller country sample sizes. Once we additionally controlled for biological factors, girl child marriage was no longer significantly associated with being off-track developmentally (see M3 in Table 4.4). Both secondary schooling and wealth quintile were also significantly associated with the ECDI. Children born to women who completed secondary schooling or higher had 17.5% reduced odds of being off-track for development as compared to those born to women who did not complete at least secondary schooling (OR= 0.825, 95% CI
[(0.772, 0.883)], p<0.001). Children belonging to the richest quintile had 39% reduced odds of being off-track for development as compared to those in the poorest quintile (OR=0.614, 95% CI
[0.563, 0.670], p<0.001).
Stunting. Similarly, once we controlled for contextual factors, the magnitude of the effect of girl child marriage on child development decreased from OR=1.29 to OR=1.12. Again, mother’s completion of primary education and living in an urban setting were significantly associated with lower odds of a child being developmentally off-track (see M2 in Table 4.5).
Figure 4.3 shows interesting variations of M2 by country. Girl child marriage remained significant when controlling for biological factors, although the magnitude of its effect further decreased to OR=1.08 (95% CI [1.03, 1.14], p<0.01). Early childbearing was significantly
86
associated with stunting, and earlier years of childbirth were associated with increased odds of being stunted (OR=1.36 in M1, p<0.001, OR=1.25 in M2, p<0.001, OR=1.14 in M3, p<0.01). As shown in the final model, both secondary schooling and wealth quintile were significantly associated with child stunting. Children born to women who completed secondary schooling or higher had 23% reduced odds of stunting as compared to women who did not complete at least secondary schooling. Children belonging to the richest quintile had 48% reduced odds of stunting as compared to children in the poorest quintile.
Hypothesized Mechanisms Explaining the Association Between Girl Child Marriage and Child Development and Health
Early childbearing, mother’s advanced education, and wealth quintile were hypothesized as mechanisms accounting for associations between girl child marriage and child development and health. Table 4.6 shows the relationship between girl child marriage and these intermediate variables. Even when controlling for contextual factors, women who gave birth earlier were more likely to be married before age 18. Women who completed secondary school were less likely to be married before age 18. We also see a relationship between wealth quintile and girl child marriage; women who were in the third, fourth, and fifth wealthiest quintiles were less likely to be married before age 18, with the wealthiest women being the least likely to marry early. These associations were significant at the p<0.01 level.
Additionally, we conducted the Sobel-Goodman test to determine if these hypothesized mechanisms were mediators explaining the relationship between girl child marriage and child development and health (see Supplementary Table 4.7). For being off-track for child development, age at childbirth explained 4%, completion of secondary school explained 39%, and wealth explained 13% of the effect of girl child marriage (all significant at the p<0.05 level).
For stunting, age at childbirth explained 9%, completion of secondary school explained 49%, and
87
wealth explained 17% of the effect of girl child marriage (all significant at the p<0.05 level). The results were consistent with the change in effect in adjusted models and reveal that the three hypothesized mechanisms differentially explain the effect of girl child marriage on child development and health.
Study ID ES (95% CI)
Central Africa CAR 0.99 (0.85, 1.14) Chad 0.82 (0.71, 0.93) DRC 1.28 (1.10, 1.46) . Eastern Africa Kenya-Mombasa 1.63 (0.40, 2.85) Kenya-Nyanza 1.10 (0.88, 1.31) Madascar-South 0.93 (0.63, 1.22) Malawi 1.19 (1.07, 1.31) Somalia-NE 1.13 (0.89, 1.37) Somalia-Somaliland 0.91 (0.72, 1.11) Zimbabwe 1.11 (0.94, 1.29) . Southern Africa Swaziland 2.03 (1.12, 2.95) . Western Africa Ghana 1.17 (0.95, 1.39) Mauritania 0.94 (0.79, 1.10) Nigeria 0.85 (0.62, 1.07) Sierra Leone 0.92 (0.77, 1.07) Togo 0.89 (0.69, 1.10) .
-2.95 1 2.95
Figure 4.2: Logistic regression results of girl child marriage and ECDI binary score, controlling for contextual factors, by country
88
Study ID ES (95% CI)
Central Africa CAR 1.02 (0.87, 1.17) Chad 1.10 (0.95, 1.24) DRC 1.14 (0.98, 1.30) . Eastern Africa Kenya-Mombasa 1.47 (0.32, 2.62) Kenya-Nyanza 1.25 (1.00, 1.49) Malawi 1.08 (0.98, 1.19) Zimbabwe 1.15 (0.95, 1.35) . Southern Africa Swaziland 1.44 (0.75, 2.12) . Western Africa Ghana 0.98 (0.79, 1.16) Mauritania 1.29 (1.08, 1.51) Nigeria 1.21 (0.89, 1.54) SL 1.18 (0.99, 1.37) Togo 1.08 (0.83, 1.34) .
-2.62 1 2.62
Figure 4.3: Logistic regression results of girl child marriage and stunting, controlling for contextual factors, by country
89
Table 4.6: Summary of unadjusted and adjusted analyses on the relationship between girl child marriage and intermediate variables
M1 M2 Mother's age at childbirth β CI β CI <16 years 0.444 (0.402, 0.486) 0.363 (0.323, 0.403) 16-17 years 0.422 (0.399, 0.445) 0.364 (0.341, 0.386) 18-19 years 0.222 (0.203, 0.242) 0.209 (0.191, 0.228) Mother completed secondary school -0.287 (-0.299, -0.275) -0.226 (-0.240, -0.211) Wealth quintile (reference: Poorest) Poorer 0.00957 (-0.00553, 0.0247) -0.00325 (-0.0178, 0.0113) Middle -0.00608 (-0.0214, 0.00925) -0.0239 (-0.0389, -0.00892) Richer -0.024 (-0.0398, -0.00893) -0.0382 (-0.0542, -0.0222) Richest -0.138 (-0.154, -0.121) -0.131 (-0.150, -0.112)
Note. Coefficients presented are OLS estimates from linear regression models with 95% CIs in parentheses. M1 is unadjusted. M2 also controls for rural/urban, mother's completion of primary schooling, and country fixed effects. Bolded values are significant at the p<0.01 level.
Discussion
The results of our analyses suggest that children born to women who marry before age 18 had significantly higher odds of being developmentally off-track and higher odds of being stunted than those whose mothers married later. Importantly, contextual factors including mother’s completion of primary education and geographic location, which may create enabling environments for early marriage, explained most of this relationship and may also explain differences observed by country. Lower levels of maternal education and living in a rural setting have been associated with higher likelihood of early marriage, particularly across the African continent (Erulkar, 2013b). Living in an urban setting may result in increased availability of health facilities, antenatal and postnatal care, and early education opportunities (Clifton, 2006), all of which can positively influence child development and health.
We also found evidence that a small part of the relationship between girl child marriage and child stunting was explained by early childbearing. Importantly, these results suggest that
90
early childbearing was not the sole pathway through which girl child marriage affects child development and health, and the results of the Sobel-Goodman test suggest that it explained very little. Other studies have similarly found that early childbearing affects children’s physical growth, which may be the result of a young mother’s physiological immaturity and her body’s competition for nutrients with the fetus (Abdullah, Malek, Faruque, Salam, & Ahmed, 2007; Fall et al., 2015). We found that early childbearing had minimal association with the child being off- track for development based on the four domains assessed. This is somewhat surprising as early childbearing has been linked to child’s poor school performance (Fall et al., 2015), likely influenced by the cognitive development. Though we are cautious to draw strong conclusions from this finding, much evidence in support of our hypothesis came from the South Asian context, which has differences in the practice of early marriage and early childbearing.
Contrarily, in sub-Saharan Africa, early marriage and early childbearing were not always synonymous (Doyle, Mavedzenge, Plummer, & Ross, 2012). The findings of this research may point toward other factors that are more important for mother-child relationships such as mother’s education, which has been found to be disrupted in some African countries through early marriage rather than early motherhood (Lloyd & Mensch, 2008).
Once we controlled for socioeconomic factors, girl child marriage was no longer significantly associated with the child outcomes measured. Instead, the final models revealed that disparities in mother’s completion of secondary education and wealth explained child development and stunting. The Sobel-Goodman test revealed that advanced education explained more than one-third of the association between girl child marriage and the child outcomes assessed. The odds of a child being developmentally off-track were nearly 40% lower among children in the richest wealth quintile as compared to those in the poorest. Similarly, the odds of
91
a child being stunted were nearly 50% lower among the wealthiest, as compared to the poorest quintile. We see further evidence of what might explain the importance of wealth in the current model, as its inclusion reduced the magnitudes of the effect of urban, primary education, and secondary education while we see almost no change in the magnitude of the effect of other covariates. In other words, wealth explains much of the reason why geographic location and education affect child development and stunting in this sample.
These findings resonate with the limited evidence on associations between mother’s age at marriage and child outcomes, and may point towards the detrimental effects of poverty and the importance of maternal behavior in a woman’s child’s health and development. In India, girl child marriage was significantly associated with children’s stunting in both unadjusted and adjusted models, although the magnitude of the effect was higher than this study revealed
(OR=1.85 compared to OR=1.29). Though literature on this relationship is limited, especially in the sub-Saharan African context, researchers have suggested that the relationship between mother’s marriage and child nutrition may be explained through maternal decision-making, including decisions on food for the woman herself as well as her child (Santhya, 2011). A study in Afghanistan did not find a significant relationship between girl child marriage and child stunting; however, in this study, girl child marriage was measured by mother’s age at childbirth
(Mashal et al., 2008), further supporting our findings that child marriage can impact a woman’s child’s health and development through pathways aside from early childbearing. Moreover, in the Afghanistan study, lack of maternal autonomy was significantly associated with child health including stunting, suggesting that women’s ability to make decisions may further impact the decisions she makes for herself and her child (Mashal et al., 2008).
92
Limitations
This study used cross-sectional, observational data, posing a challenge to establishing causality and controlling for changes over time. We used a simplified framework to test mechanisms through which girl child marriage may affect child outcomes, which may not empirically measure all biosocial interactions. Despite controlling for several factors, it is possible that residual confounding exists and that environmental factors not measured in these dataset impact results. Of note, we were unable to include indicators of IPV and child health care in the dataset. This study also does not control for female genital mutilation/cutting, which may be present in settings with child marriage and has been correlated with maternal health (Yount &
Abraham, 2007). However, we are less concerned about residual confounding based on these variables because we control for observed and unobserved differences by geographic location and year using country fixed-effects.
There are also study limitations related to data collection. The exposure and child developmental outcomes were reported by mothers, rather than measured by an enumerator. Data on age of marriage were also collected in two different ways and could introduce measurement error. Even though we cannot directly assess the nature of this measurement error, it seems likely that measurement error would be random, and thus lead to an underestimation of the true causal effects of interest. There is also potential variability in how data were collected in each national and sub-national context, and in accounting for sample designs in the pooling of data across countries. However, UNICEF oversees the design and collection of MICS, providing training, oversight, and consistency across settings that allow the data to be internationally-comparable
(Bornstein et al., 2012). Additionally, our careful review of each individual dataset prior to
93
pooling, as well as the inclusion of country-fixed effects in this model, reduces concerns over data pooling.
Strengths
Despite these limitations, this study has several strengths. We used a large dataset including representative household data across several sub-Saharan African countries, limiting the threat of selection bias. The statistical methods employed allow us to assess not only associations but also the relative magnitude of the mechanisms measured. HAZ was externally- assessed by trained measurers, reducing potential for bias from self-reporting. Additionally, the use of individual outcomes, rather than aggregate cross-country comparisons, is less prone to confounding biases. We also conducted robustness checks through looking at continuous measures of each outcome to assess if there were differences in the results. Though the findings may not be generalizable to all women aged 15 to 49 in sub-Saharan Africa, the large sample size across settings increases confidence on the application of these findings to understanding the relationship between girl child marriage and child developmental and health outcomes.
Conclusions
The findings of this study reveal that there are intergenerational consequences of girl child marriage on child development and health, and that these associations were largely driven by several pathways which could potentially be addressed through targeted policies. Women across the sub-Saharan African countries in the sample who married before age 18 were more likely to give birth early, less likely to have completed secondary schooling, and were more likely to belong to poorer quintiles, all of which were associated with worse child outcomes.
Findings of the importance of maternal education and wealth on child development and health point toward important policy considerations for improving early childhood outcomes within
94
different contexts. If the mechanisms examined could all be controlled for, a theoretical event we acknowledge is unlikely, the negative effects of early ages of marriage on children may be avoided.
As theorized our framework, the social pathways examined may ultimately affect children through influencing maternal behavior, that is, the decisions that a mother makes for herself and her child given the context in which she is situated within, affected by limited resources, knowledge, skills, or agency and ultimately affecting her child’s outcomes. Increasing the age of marriage above 18, or enforcing the legal marital age in most of the African countries included, may have a cascade effect on female opportunities which may, in turn, affect maternal biological and behavioral factors influencing child well-being. National policies and programs could further expand maternal and child health and education to enable access for women and children in rural, poorer, and lower educated environments. Such system changes can also be complimented by ensuring women and other stakeholders in local communities are aware and actively seek out these services. Girl child marriage may, in fact, be a modifiable risk factor for child development and health in the sub-Saharan African contexts explored.
95
Chapter 5: ‘It is this which is normal’ A Qualitative Study on Girl Child Marriage and Health in Guinea
Summary
This chapter addresses Aim 3, to qualitatively investigate perceptions of the social construct of “girl child marriage” among women in a community, and to explore the perceptions that women have of the relationship between marriage, health, and well-being.
Introduction
Eighteen of the 20 countries with the highest prevalence rates of girl child marriage, defined as a union of a girl before age 18, among 20 to 24-year-olds are in the sub-Saharan
Africa region (UNICEF, 2018c). High rates exist despite regional instruments banning the early marriage of females that several countries have signed, ratified, and deposited, including the
African Charter on the Rights and Welfare of a Child (1990) and the Protocol to the African
Charter on Human and Peoples’ Rights on the Rights of Women in Africa (Maputo Protocol,
2005) (African Union, 2003; Organization of African Unity, 1990).* Such startling realities have led the United Nations to describe girl child marriage as a violation of girls’ and women’s rights and to call for the elimination of “all harmful practices, such as child, early, and forced
*According to the African Commission on Human and Peoples’ Rights (2018a, 2018b), for the African charter, 41 countries have signed and ratified, 9 countries have signed but not ratified, and 4 countries have neither signed nor ratified. Countries that have not signed or ratified include: Democratic Republic of Congo, Sao Tome and Principe, South Sudan, and Sudan. Countries that have signed but not ratified include: Central African Republic, Djibouti, Guinea-Bissau, Liberia, Sawhrawi Arab Democratic Republic (partially-recognized state), Somalia, Tunisia, and Zambia. For the Protocol, 36 countries have signed and ratified, 15 countries have signed but not ratified, and 3 countries have neither signed nor ratified. Countries that have not signed or ratified include: Botswana, Egypt, and Tunisia. Countries that have signed but not ratified include: Algeria, Central African Republic, Chad, Eritrea, Ethiopia, Sawhrawi Arab Democratic Republic (partially-recognized state), Sao Tome and Principe, Sierra Leone, Somalia, South Sudan, and Sudan.
marriage” in the 2015-2030 Sustainable Development Goals (UN General Assembly 70th session, 2015, p. 18).
In the Republic of Guinea, a West African country of 12.4 million, 51% of girls are married by age 18; a predominantly Muslim country gaining independence from France in 1958,
Guinea has the eight highest rate of girl child marriage in the world (Central Intelligence
Agency, 2018; UNICEF, 2018c). Continued high rates of girl child marriage exist alongside challenges to socioeconomic development. Guinea is among the poorest countries in the world, with a GDP-PPP of US$2,000, and has the eleventh highest maternal mortality rate (679 deaths per 100,000 live births) (Central Intelligence Agency, 2018). Its rich history is fraught with underinvestment in health systems and health care, a root cause of recent health catastrophes including the 2013-15 Ebola outbreak, enabling further vulnerability of girls and women.*
Importantly, Guinea has signed and ratified both the African Charter on the Rights and Welfare of a Child and the Maputo Protocol (African Commission on Human and Peoples' Rights, 2018a,
2018b), legally binding it to progressively realize the rights guaranteed in these instruments, including preventing girl child marriage before age 18. Community-based organizations and international organizations further advocate against non-consensual marriages in Guinea, including, by legal definition, marriage before age 18 (Canada: Immigration and Refugee Board of Canada, 2015). However, although rates of girl child marriage have slowly declined globally, trends in Guinea have remained relatively unchanged; nationally-representative household data on women ages 20 to 49 years from the Demographic and Health Surveys reveal the median age
*Guinea’s current social and economic context are more comprehensively understood as rooted deeply in historical experience. As Farmer (in press) describes and roots in modern day anecdotes, understanding the resource extraction and trade routes of the 19th century as well as consequences of the slave trade, colonial rule, and conflicts in what was formally known as the Upper Guinea Coast has created an environment where epidemics such as Ebola could unnecessarily occur. 97
at first marriage in Guinea was 16.5 years in 1999, 16.3 years in 2005, and 17.2 years in 2012
(The DHS Program STATcompiler, 2018); marital age was even lower when accounting for 15- to 19-year-olds as well [personal analysis]. Several causes and consequences of girl child marriage have been documented in the literature, yet limited research captures the perspectives of women themselves who experience child marriage. Data from African contexts identify causes including low education levels and poverty, which may be heightened in fragile or conflict settings and rural environments (Efevbera, Bhabha, Farmer, & Fink, 2018a; Lloyd &
Mensch, 2008; UNICEF, 2015b). Girl child marriage has been associated with poor sexual and reproductive health in South Asian contexts (Godha et al., 2013), as well as poor mental health
(Gage, 2012; Wondie et al., 2011), poor child nutrition and development (Efevbera, Bhabha, et al., 2017), increased likelihood of intimate partner violence (Erulkar, 2013b), and reduced secondary education (Wodon et al., 2016) in African contexts. Yet one recent study across sub-
Saharan Africa found that child marriage may be protective for some health outcomes such as underweight (Efevbera, Bhabha, et al., 2018a; Efevbera, Bhabha, Farmer, & Fink, 2018b), complicating the existing narrative of overwhelmingly negative outcomes.
Moreover, despite the frequency of girl child marriage in Guinea, little has been published about its causes and consequences in this setting, especially from the perspective of those who marry early (Callaghan, Gambo, & Fellin, 2015; Myers, 2013; Schlecht et al., 2013).
While existing literature may capture broader stakeholder perspectives, there is only limited empirical evidence on women’s perceptions of how their child marriages specifically affect health consequences (Efevbera, 2017a, 2017d; Nasrullah et al., 2014), with no public health literature identified from Guinea. Understanding how women married as children interpret their own experiences in early marriage may be difficult because their status may be viewed as
98
stigmatized, taboo, and, in places like Guinea, illegal. Yet these are crucial and often missing voices which might provide the missing links to meaningfully inform effective programs and policies across health, education, child protection, and other sectors that seek to incorporate all constituents including women married as children populations.
The overall aim of this paper is to qualitatively explore perceptions of the social construct of “girl child marriage” in an African community, and to explore the perceptions that women have of the relationship between marriage, health, and well-being. Using qualitative research methods, the motivating research question was: In what ways, if at all, do women married as children in Conakry, Guinea see their marriage as related to their health and their children’s health?
Methodology
Overview
Qualitative research as an approach emerged in the 1920s and 1930s to provide structure to understanding human experiences in natural contexts. Denzin and Lincoln (1994) define it as
“multimethod in focus, involving an interpretive, naturalistic approach to its subject matter. That means that qualitative researchers study things in their natural settings, attempting to make sense of, or interpret, phenomena in terms of the meanings people bring to them” (Denzin & Lincoln,
1994, p. 2). Traditionally used in fields of sociology and anthropology, qualitative research provides an opportunity to explore how individuals make meaning of their own experiences. It can provide depth and context where quantitative research cannot. In selecting a qualitative research process, one ideally has thought through (1) the researcher, (2) theoretical paradigm, (3) research strategy, (4) data collection and analysis, and (5) interpretation (Denzin & Lincoln,
1994); explaining the considerations for these alone could be its own anthology. Items 1 through
99
3 are briefly described in this sub-section and Items 4 and 5 are expanded upon in subsequent sub-sections.
This study draws from grounded theory, which involves an inductive process of generating theory from the data through systematic data collection and analysis (Strauss &
Corbin, 1994). Also known as the constant comparative method, grounded theory is distinguished by the iterative practice of collecting and analyzing data, constantly comparing results and using it to inform meaning. Grounded theory also shares the perspective, as other qualitative methods do, that “interpretations must include the perspectives and voices of the people whom we study” (Strauss & Corbin, 1994, p. 274); interpretations, made by the researcher, stay true to participants’ voices and perspectives. In other words, despite the existence of dominant narratives from quantitative and advocacy work on child marriage in other contexts, this research does not apply an existing theory or hypothesis toward understanding how women in Guinea perceive or experience child marriage and its relationship to well-being.
Instead, women who themselves have experienced marriage as children are viewed as authorities to explain how their early marriages are related to health and well-being.
Grounded theory was selected as a principal guiding framework for a few reasons. First, no prior empirical qualitative work on child marriage and health in Guinea was identified, illustrating a knowledge gap. Second, grounded theory is designed to counter other approaches to theory-development which may not be rooted in systematic data collection or analysis. A grounded theory approach with an explicit methodology importantly can complement, or perhaps challenge, existing narratives on child marriage that are rooted in human rights and advocacy work with limited empirical support in different contexts. Finally, selection of this guiding approach makes explicit the author’s bias, as a researcher, toward valuing the experiences of
100
women themselves in understanding why a social phenomenon exists and what its consequences are.
This study also draws from ethnography, at the heart of which is participant observation.
Ethnography, as defined by Atkinson and Hammersley (1994), includes exploring a social phenomenon, working with “unstructured data,” using a limited number of cases, and interpreting human actions (Atkinson & Hammersley, 1994). It is closely linked with the use of participant observation, where the researcher actively participates in activities during field research (Atkinson & Hammersley, 1994). While this study does not provide a comprehensive ethnographic account of experiences with child marriage, the three months spent in Guinea-
Conakry as part of this research advance standard qualitative methods used in public health research, intentionally engaging in more contextual grounding of this work. Drawing from ethnographic methods allowed the author to more explicitly be a part of this work as well. The author does not fully tap into researcher-as-experiencer in this study or use all ethnographic data collected in this chapter. Instead, the author acknowledges upfront that elements of ethnography were considered in this research, as part of a larger study on child marriage in Guinea.
Participant Selection
Qualitative data were collected from in-depth interviews conducted among 19 women whose first marriage occurred before their 18th birthday. The first study participant identified was a community leader in a women’s collaborative that supported community-based economic empowerment and health activities for a project the author simultaneously volunteered with during her time in Guinea-Conakry. Women were purposively sampled from neighborhoods across the capital city of Conakry* and snowball sampling was used to identify additional
*According to the Central Intelligence Agency (2018), Conakry has a population of 1.936 million people. 101
participants. A woman was eligible to participate if she was currently or had formally been married, had her first marriage before her 18th birthday, and was able to schedule an interview time or provide a phone number for follow-up at the time of recruitment. No minimum age was required for participation in this study, and study protocol outlined appropriate consent and assent procedures, based on participants’ ages. Consideration was also given to diversity in current age, age at marriage, duration of time in marriage, and geographic location during recruitment.
Of note, although the study sought to conduct in-depth interviews with women marrying before age 18, following Guinea’s legal definition of child marriage, other married women were invited to participate in brief interviews. These were used to collect general information on the situation of women in Guinea and to screen for experience with child marriage, which was initially left open for participants themselves to define. Through the brief interviews, one additional participant described her marriage as “too early” and requested the opportunity to share more about the challenges of her marriage at age 25. An in-depth interview was, thus, conducted with her, though her data have not been comprehensively integrated into the current manuscript.*
Data Collection
The primary sources of data used in this study were in-depth interviews. These interviews were conducted from August 2016 to January 2017, in French or a local language, led by the author and a local research assistant (RA). RAs were final-year medical students at a local university who had previously participated in public health research, identified through a public
*In future work, comparing her experience to that of women who married before age 18, and collecting additional data from women married after age 18, could provide insight into marital experiences more broadly in the context of Conakry; these perspectives are beyond the scope of this study. 102
health writing workshop the author had volunteered at; the author provided additional training in qualitative research methods and ethical conduct in research to RAs prior to data collection.
Interviews ranged from 45 minutes to 1.5 hours. A semi-structured interview protocol with generic rather than particularist questions was used, allowing the participant to share as much or as little of their experiences as desired (Maxwell, 2012); this structure allowed participants to speak about their own experiences, rather than exclusively about child or early marriage broadly, and probing was used as a technique to elicit additional information. Questions were open-ended, and covered several topics including: personal background, general perspectives and views on relationships and marriage in Guinea, personal perspectives and experiences in marriage, and health history.* With permission from participants, interviews were conducted individually and in-person in a private space at the discretion and comfort of the participant. Most participants requested interviews take place in their home, while a few interviews were administered in an alternative private location based on the participants’ preference. Interviews were audio recorded and files were electronically stored in a secure location, handled by the author. Data were transcribed and translated first to French, when applicable, and then to English, by the author and, when necessary, an RA.
Additional data were collected to increase potential analytical angles including brief ethnographic interviews, observation, and participant observation. Brief ethnographic interviews are used to systematically and rapidly capture community perspectives on a specific topic
*Over one to two interviews conducted with each participant, women were asked to reflect on their lives. They were asked questions about their childhood and families, education, marriages, husbands, children, and their health. They were asked about their current lives, the things they enjoy, and the challenges they feel they face. Aspects of women’s everyday lives, ranging from their homes to their families to their neighbors, were also shared. Through this process, women described several experiences within the context and as consequence of their early marriages. The experiences touched on different social, cultural, and economic aspects of their lives, of which many were related to health and well-being, the focus of this study. 103
(Hubbard, 2008). To this end, rapid, 15-minute interviews were conducted with participants, when possible, which facilitated familiarity with the interviewer and established them as the expert of women’s experiences in Guinea broadly; brief interviews were also conducted with women whose marriages occurred after age 18 and with key stakeholders to provide additional context on marriage, health, and women, though these data are not included here. This study also used observation and participant observation. The author observed and took notes on natural occurrences, interactions, and conversations among study participants. The author also participated in and took note of local activities such as community gatherings, household work, and market interactions, when invited by study participants. Notes excluding participants’ identifying information were hand-written and later typed. Extensive memos were written throughout the process of data collection and analysis, producing additional data on women’s contexts.
Of note, although a wealth of “rich data” were collected as part of a larger study on women’s perceptions and experiences with girl child marriage, in-depth interviews among women who married early, with a focus on health consequences, are the primary sources analyzed and presented in this chapter. Remaining data will be revisited in future work.
Data Analysis
Drawing from grounded theory, data analysis was an iterative process. The author conducted open coding as the categorizing strategy to stay close to participants’ words (Maxwell,
2012). Text was read and reread during and following the translation process to understand key themes emerging. At the first level of coding, key concepts and categories were identified. At the second level of coding (axial coding), concepts and categories were organized and grouped. The author, then, repeated this process to ensure consistency in key themes emerging from the
104
research. Coding categories were also discussed with a collaborator in the process of organizing final codes as a validity check.* These discussions ensured that sufficient evidence was identified in support of each theme developed and that codes were emic, arising from participants’ contributions, rather than etic, coming from outside of the social group studied. Coding categories were also refined following presentations of parts of this work in-progress (Efevbera,
2017c, 2017e). Data were analyzed both by hand and using NVivo 11.
Ethical Considerations
Participants’ safety was seriously addressed in this study, and several considerations were made in study design, data collection, and data management to protect participants. Informed consent was received from all participants prior to research activities. Privacy, anonymity, and confidentiality were upheld throughout data collection and management. A safety plan was also developed to refer participants if they presented signs of distress during the interview or desired support. Prior to conducting interviews, ethical board approval for all study procedures was obtained from both the Harvard Longwood Institutional Review Board (IRB16-0939) and
Guinea’s Comité National d’Ethique pour la Recherche en Santé (CNRES), the National Ethics
Committee for Health Research established by the government (108/CNRES/16). Additionally, the author met with Harvard Longwood IRB for an annual review, given her petitioned status as a student Principal Investigator.
*The collaborator the author consulted with was Gretchen Brion-Meissels, Ed.D., Ed.M., M.A.T., lecturer in the Prevention Science and Practice Program at the Harvard Graduate School of Education. 105
Results
A Brief Account of How Child and Early Marriage was Defined
Before continuing, it is important to highlight that women were not invited to participate in a study on child marriage; instead, this study was framed as researching perspectives on women’s experiences with marriage, health, and well-being in Guinea. Screening for in-depth interviews prioritized enrollment of women whose first marriage occurred before the age of 18, based on the legal definition of marriage in Guinea, yet the terms “early marriage” and “child marriage” were not used in the recruitment script, informed consent forms, or interview guides.
However, throughout the interviews, some women described their marriages as “early” or too young. Their reflections on their marriages, and their perspectives on marriage in their societies, informed an alternative definition of what child and early marriage in Guinea is.
The ideal age of marriage, according to women interviewed, ranged greatly, revealing that “child marriage” or “early marriage” may not best be defined at age 18. The youngest ideal age proposed was 10 years, explained by one woman as the age at which a girl is ready to marry because she “knows the difference between bad and good” and “knows her friends and her enemies.” Other women proposed marriage is best at ages 14 through 17, allowing girls to progress in their studies and gain additional life experience while being able to start childbearing early enough to see their children grow up. Thirteen out of 19 women proposed that women should wait until at least age 18 to marry, providing several reasons beyond the legal framework as their rationale. Seven out of 19 women suggested that marriage should not occur before age
20, with the highest age range recommended as 25 to 30 years. When asked to explain why waiting until up to 30 years was the ideal time to marry, one woman stated:
106
“One finds that a person has good experience, outside even, when she is married [at that
time]. I don’t believe she is going to divorce. One finds that she had a good idea and it is
also her choice… [She has] the experience of a family, how she is going to set up her
home, you see, how she is going to take care of her husband.”
She concluded by suggesting that by age 30, a woman knows how to handle conflicts that may arise between her and husband, allowing the married couple to get along.
Life experience, knowledge, and educational attainment were often raised as reasons for delaying marriage, which resulted in one woman, married at age 25, explaining that her marriage was both undesired and too early. Yet one woman provided both biological and religious arguments for why marriage should occur at younger ages. She explained:
“She must marry at the age of 16 years because it is written in the Qur’an and that is what
is said everywhere. If you marry at this time, you will have honored your family, and you
yourself, you will be honored because you will not have stayed at your parents’ long until
they are annoyed of you.”
She was, moreover, content with marriage at a younger age, when a girl starts her period, based on biological readiness:
“…the woman and the man will not have the same fertility. The woman’s fertility goes
faster than that of the man. The man himself being old, he can have children, but the
woman, as soon as the date of her biological clock arrives, she will be in menopause, you
see? That is why the woman, when you marry early, it is good.”
Although the legal definition, according to the 2011 Child Code, suggests that marriage is illegal before a woman’s 18th birthday, the construct of early or child marriage as understood at the community level, importantly, is much more complex. Many women had knowledge of
107
existing laws, or “the text” as they called it, yet cultural practices and social norms were more salient toward defining what is too early. For some, too early was much younger than the cutoff of 18 years; for others, a marriage at 18 years was still too early. For the purpose of this study, it can be acknowledged that participants perceived their own marriages to be “early,” although how
“early” and what “early” meant for them is varied.* The current limitation of analysis to women who married before age 18 maintains policy relevance.
A Profile of the Women Married as Children
The average age at first marriage was 15 years, among the 19 women whose first marriage occurred before their 18th birthday. Women represented different ages, ethnic groups, education levels, and backgrounds, as viewed in Table 5.1. The average age at the time of interview was 32 years, ranging from 17 years old to approximately 60 years old. Most participants belonged to the Peuhl (Fulani), Sousou, or Malinke (Manika) ethnic groups*; only one ethnic minority (Lélé) was included in this sample. Importantly, not all ethnic groups in
Guinea were represented in the study sample. Many of the women were still married or non- legally separated at the time of interview, and only four explicitly stated having civil marriages.
Most women in this study were married through traditional and religious ceremonies, resulting in unions that do not have the same protections as civil marriage under the law. The women’s current education levels also varied, ranging from none or incomplete primary schooling through university-level training. Although religion was not explicitly asked about in the context of this
*There are more qualitative data on women’s definitions of early marriage for females and males, and how they make sense of relationships more broadly, that were collected as part of the larger study. A future paper will investigate further how women make meaning of the concept of marriage in this cultural context. *According to the Central Intelligence Agency (2018), the Peuhl (Fulani), Sousou, and Malinke (Manika) ethnic groups account for more than 70% of the ethnic make-up of the population.
108
study, nearly all, if not all, of the women in this sample are believed to be Muslim, practicing the predominant religion in Guinea.* This observation was made as participants referred to Islam, studying the Qur’an, or used religious terms (in Arabic) at different times in the interview.
Table 5.1: Participant Characteristics
Characteristic Description Age at interview 17 to 60 years (avg. 32) Ethnic Group Peuhl (Fulani), Sousou, Malinke (Manika), Lélé Age at marriage 13 to 17 years (avg. 15) Current marital status 14 married, 4 divorced, 1 widowed; 4 civil ceremony Education completed or currently 5 none (incomplete primary or koranic) enrolled 4 primary 4 lower secondary 2 upper secondary 3 tertiary 1 unspecified Primary occupation 2 teachers (current and former) 3 market sellers (current and former) 3 students (upper secondary and tertiary) 2 NGO professionals 1 cashier 3 unemployed 1 retired 4 unspecified
Note. This table includes data for women whose first marriage occurred before age 18 (N=19). It does not include data for the final woman interview who married at age 25, who described her marriage as early. At the time of interview, she was a currently-married 32- year-old Peuhl (Fulani) woman who had completed university and technical training and was currently unemployed.
*According to the Central Intelligence Agency (2018), 86% of the population of Guinea are Muslim. 109
A Brief Account of Why They Married Early
Just as there was no singular type of woman who married early, there was no singular reason for their early marriages.* For many women, their marriages were influenced by their families or caretakers. Several women described, as one woman quoted, that “they gave me in marriage.” In these instances, the parents or caretakers identified a husband, often a family acquaintance or relative to the woman, and told the woman she would marry without asking her.
Some women were told in advance of their wedding day; others were informed shortly before the wedding was to occur. For many of these women, there was no question of whether they wanted the marriage. They understood the cultural expectations of being obedient and submitting to their parents, and accepted their marriages accordingly. As one woman explained, and several others similarly described:
“When they called me for the marriage, they told me that now, they gave me in marriage
to the son of my older brother. ‘I gave you in marriage to the home of your older brother
to marry his son.’ But I could not overtake my parents. I was a kid; I could not overtake
them. All that they tell you, you say, ‘Okay.’ All that they say, you must affirm. Even if
you do not want to, you cannot say that you do not want to.”
Another woman explained:
“At the time when I married, I was only almost 15 years old. I knew nothing of life. So,
my parents insisted that I marry, and I married. When I came, I knew nothing of the man,
and he did not know me. I only heard them say that he lived in Conakry. It was my first
time to come to Conakry. I suffered during the first days because I knew no one, and no
*There are more qualitative data on the drivers of child marriage among participants in this sample that was collected as part of the larger study. A future paper will further investigate the factors leading to early marriage in this context, and the brief overview sets the stage for discussing health-related consequences, as per the study research question. 110
one knew me. Since we cannot underestimate the words of parents, that is what made me
accept the marriage.”
Some women even described threats of being cursed by their family, evoking fears of witchery, that further ensured they marry as requested. They perceived they had no choice, and often did not know of any other way; following parents’ requests, including with respect to earlier marriages, was not a matter of personal choice by the woman but a general expectation.
Some women expressed understanding their families’ rationale for ensuring their early marriages. For women who were going out a lot, spending time with friends away from the house, their families were concerned about promiscuous behavior that could lead to an illegitimate child. One woman explained that although she had been hanging out with friends at a community youth center, she had not been seeing men; she had maintained her virginity until early marriage, to the surprise of her parents. One woman explicitly described her marriage as
“forced,” when, at the age of 16, her family quickly arranged for her to marry her then-boyfriend to avoid any chance of out-of-wedlock childbearing. She loved him and was not unhappy with the marriage, but she had not planned to marry at that time. For some women, a family friend or extended family member saw her and asked the parents for her hand, in many cases, without approaching her, the soon-to-be bride, first.
Although many women expressed that they accepted their marriages because they did not believe they had a choice, two women actively contested their early marriages. One woman had been in love with another young man at the time her marriage was arranged, but as per her community’s custom requiring marrying within the family, her parents organized and told her of her wedding. She describes running away for a month, and living with her then-boyfriend, before she eventually accepted the marriage to her cousin, whom she did not love romantically. Another
111
woman similarly contested her marriage and ran away from home. She describes several weeks during which she lived out of the house, hiding from her parents, as she sought to continue her secondary schooling while protesting the arranged marriage to a cousin, whom she had already turned down when he directly asked her to marry him some months earlier. After running out of money and food, and not having the conditions to take the national exams necessary to continue her studies, she returned home, where the traditional wedding had already been celebrated. Yet she remains adamant that she never accepted the marriage:
“I said that maybe there is a reason. If I do not do what [my family] wants, I am going to
stop studying, because if you do not eat, you are not calm; you are violent to all. You
cannot study. I decided to return. They had already ‘attached the Kola.’* But I never
accepted [the marriage], even til now. Because after having given what we call the
dowry, it is necessary to put on the white veil to say that you are married, but I never
accepted it. I said, ‘He is not my husband. I do not love him. You are not going to put me
anywhere.’ But I agreed. I came to Conakry to his home.”
In both cases, the women are now divorced.
Yet not all marriages were organized by the families. Two women described being proposed to by their husbands. Whether meeting their husband while working or through a friend, they developed relationships that led the man to ask them directly for their hand in marriage. In describing their marriages, these women expressed a sense of choice and agency in
*Additional data were collected on the celebration of marriage in Guinea. “Attaching the Kola” is what many women referred to the dowry, or an exchange of gifts and sometimes money between the families of the spouses to celebrate a traditional marriage. One Malinke woman interviewed described that the husband’s family gave Kola nuts to her family to make the marriage official. One Sousou woman interviewed said that her family gave her husband’s family Kola nuts to formalize the marriage. One stakeholder, who worked at a non-governmental organization and participated in a brief ethnographic interview, joked that since he had given his wife’s family 10 Kola nuts to formalize the marriage, he expected her to have at least 10 kids. 112
their decision to marry.* Another woman described her choice to marry the older brother of her friend, whom she had interest in. In contrast to the other two women, she explained that her desire to marry stemmed from wanting to help her mother, who at 35 years old, became widowed, and her mother was left with her and nine other children to care for. The woman was adamant that it was her decision to marry yet explains, “It is this [situation] that pushed me to marry very early to become of help to my mom.”
A Brief Account of Whom They Married
Similar to the way the women themselves varied, there was diversity amongst husbands.
For most of the Peuhl (Fulani) women who participated in this study, their husbands were first cousins or other extended family members, in line with cultural traditions. Some women explained that they did not want to marry a family member, for fear of complications and problems within the family if there were problems in the marriage, but they followed the wishes of their parents or caretakers, even when other family members disagreed. All but one of the remaining husbands were friends or family of close acquaintances, ranging from a friend’s older brother to the son of parents’ colleagues. Only one woman described becoming betrothed to someone outside of her close-knit network of family and friends, marrying a man who noticed her selling items as a street vendor after school.
The education levels and professions of participants’ husbands also varied. Some were described by the women as university educated; some were described as illiterate. The professions of women’s husbands ranged, including a maintenance and logistics manager,
*In one situation, the participant described what could be viewed as a romantic story of how her husband pursued her. Yet, through participant observation, a friend of the participant’s family explained that she viewed the marriage as a scandal, desired by the girl’s mother but not her extended family because she was still in secondary school. Throughout the interview, the participant refered to herself as engaged at times, married at other times; she is included in the study because her traditional wedding ceremony had already been celebrated. 113
shopkeeper, cameraman, and a cook. A couple of husbands were in jobs that required extensive travel or time away from Conakry, including a miner and a trader. A few husbands were teachers, described as a “simple” profession by their wives. The older participants had husbands who were farmers and cultivators. One husband was described as a trained doctor or pharmacist, but his wife, a 17-year-old, was unsure what he was currently doing for work in Dubai, where he was lived without her shortly after celebrating the traditional wedding. There were other husbands who also lived abroad for at least part of their marriages – including in Senegal,
Germany, Mozambique, and Angola. Only in a few circumstances did the girl spouse accompany her husband for part of his time abroad. The remaining husbands were unemployed.
Also important to note, there were some gaps in women’s knowledge of their husbands, particularly with regards to age. Among six women who knew their husband’s age, and one woman who estimated her husband’s age, the average age gap between women and their husbands was 15 years. This ranged from seven years to an estimated 28 years. For other women, they were left approximating their husband’s age. The 17-year-old recent bride was one of several who were uncertain of their husband’s age. When asked, she responded, “Ah, that, I do not know, but nevertheless, he is a youth, eh! Because it is not even two years, if he has, it is three years since he finished his studies. He is a youth. He has a big brother and a sister again, so, he is a youth.”
Impact of Women’s Early Marriages
Women had a variety of perspectives on their experiences with child marriage and its health impacts. Ultimately, women described several health disadvantages of their early marital experiences. These disadvantages fell across four categories: poor sexual and reproductive health, intimate partner violence and long-term consequences, poor mental health and
114
psychosocial well-being, and other physical health conditions. Interestingly, women also described several health advantages of their early marital experiences. These advantages covered five themes: having children, “good health for me and my children,” access to health care, delaying first pregnancy and birth spacing, and positive mental health and psychosocial well- being. Further analysis revealed that perceptions of a woman’s marital experiences corresponded with certain factors. These included husband’s traits, individual traits, and perceptions and experiences with support. Although these outcomes are possible in many marital relations, women’s reflections suggest that the timing of their marriages influenced outcomes and their ability, or lack thereof, to cope. The following sections explain these findings in substantial detail.
Perceived Health Disadvantages
This study found that several women perceived disadvantages of their early marriages.
They described several social, economic, and cultural realities they attribute to their marriages, of which several were related to health and well-being. These unfavorable conditions women discussed in the context of their early marriages are presented across four themes: poor sexual and reproductive health, intimate partner violence and related long-term consequences, poor mental health and psychosocial well-being, and physical health conditions.
Negative Sexual and Reproductive Health
Several women described negative sexual and reproductive health outcomes they experienced within the context of their marriages. For most women, the first sexual experience occurred in marriage. This postponement of sexual contact until marriage was consistent with local sociocultural expectations. Consequently, women experienced first and repeated intercourse at an early age as child spouses, as young as 13 years old among women interviewed,
115
and sometimes before physical or mental maturity. For example, two Peuhl (Fulani) women described details of the test for virginity on the night of the wedding, where family and community members check for blood stains on a loincloth laid on the marital bed to determine if the bride was “devirginized” in marriage. As a 30-year old woman, whose first sexual experience occurred in marriage, shared, “It was at the age of 15 that I started sexual relationships. But to start, it made me tired a lot until I was used to it.”
Early and difficult pregnancies were also described by many women, occurring shortly after engaging in sexual intercourse with their husbands. Among the most severe challenges with giving birth at a young age were miscarriages and stillbirths, resulting in the loss of pregnancy, and obstetric fistula, a condition caused by obstructed labor resulting in a woman’s inability to control urination or defecation. One of the older women, a 60-year old whose family arranged her child marriage, opened her interview by describing the hardships that she faced:
“At 14 years old, they gave me in marriage without my consent, in other words, without
my agreement. At that time, they did not ask for the agreement of the woman, so it was
parents who decided, and they sent you. The month in which I started my period, the
second month, they gave me in marriage. As soon as I entered into the marriage, I had my
first pregnancy. The pregnancy there, the infant, it was a stillbirth. They said that it was
fistula. So, I was treated for this only 10 years after. I was operated on for the disease
they call fistula, and it is at this time that I had four living children, but the first [child],
no.”
She describes suffering throughout her youth and in the early days of her early marriage to her cousin, a common tradition, though at the time of interview, at 60 years, she no longer considers the negative experiences of her marriage.
116
A 27-year-old woman, who married at age 16, continues to suffer from difficulties she experienced from repeated pregnancies. After resisting an early marriage, she submitted to her family’s will and married the man they had selected for her. She explains, “The same year that I came to his house, I had four…three early [miscarriages], independent of my will. For lack of treatment, he [husband] sent me to the village. They gave me some tree bark to drink and there, everything stops there…” In the four years that followed, and in spite of a turbulent marriage, she still hoped to have children. “…But it was impossible. That is because there was this debris in my stomach. I cannot [have children] after that.” She attributes her poor reproductive health to the challenges of her early marriage. Had these women not married early, they believe they would have avoided early pregnancy and some of its unexpected and difficult complications.
Intimate Partner Violence and Related Long-Term Consequences
Physical, sexual, psychological, and economic abuse were experienced by some women in the context of their early marriages. Although experiences of abuse were also perpetrated by in-laws, neighbors, and extended family, some of the most severe and debilitating forms described were inflicted by the partner. Abuse endured within marriage led to poor physical and mental health, and, as was both explained and observed, continued to have lasting consequences years after the wedding. Repeated physical and sexual abuse were most commonly described by women who experienced intimate partner violence. A 40-year-old woman, who had initially been excited about marrying at age 16, described both physical and sexual abuse in her marriage:
“The bad experiences that I had, he plugged the heater to put on me. He hit me. He held
me. He undressed me once at night, for me, it was from the wedding. He told me,
‘Undress as when a man needs his wife.’ He watched me. He told me, ‘Okay.’ Things
117
that I never explained to my mom, for me, it was to go. So, there are things that I cannot
explain.”
She was initially cautious in sharing her experience, as she was taught to keep marital affairs private, like many of the other women who mentioned marital abuse. She did, however, remove her headscarf during a different interaction, revealing half-centimeter thick curly hair with several bald spots, from where her husband abused her. She also showed a black plastic bag of medication that, years later, she continues to take for lasting head pain she still endures from the abuse.
A 27-year-old woman, who had also married at age 16, suffered extensive abuse and multiple physical ailments over five-years as a result of her husband’s violence toward her. She was the only participant to explain in painstaking detail that despite eventually being removed from the marriage by her parents for her safety, she continues to experience poor health:
“Yes, each time that I say it, the day that I am going to die, these are the consequences
that I suffered there. Because my back, one day, he threw me…when I carry something or
I do certain exercises, my back directly stops there and I cannot move. I still stop…Also,
when I feel the head pain that I have, I would say that it is the blows that I received to the
head. Before, I did not have sinusitis…But the blows that I received on the head, I can
say that it is this that transformed and blocked my nose squarely…One day, he poured
hot water on this foot. I am going to show you, if you want, the photo, and this scar
remains today. I am obligated to abstain from seeing certain men because I do not want to
be discriminated against, because if you have scars, they tell you that you have AIDS. So,
I am obligated to abstain, even if, sometimes, I want to be with someone.”
118
She showed pictures saved on her mobile phone of third-degree burns all over her body and, in the middle of the interview, pulled down her blue jeans to reveal cigarette burn- like spots lining both legs, which she explains are the lasting physical scars.
In addition to back pain and difficulty lifting objects, headaches, breathing problems, and lasting physical scars, women described visual impairment and repeated illness resulting in job loss, which they attribute to abuse endured in their early marriages. A 40-year-old woman also described an unusual long-term consequence of stress experienced in her marriage. She explains,
“At my age, I am not, I put on the things because each time I have concerns, I am at the house, and when I think of that, it affects me. At first, they told me that it is blood pressure. Now, they tell me that I am a cardiac, that I have a heart problem.”
Women’s early experiences with intimate partner violence in marriage also included psychological and economic abuse. Women described arguments and indirect behaviors where the husband exerted his control, such as not being given financial means necessary to take care of themselves or their children. One woman, married at age 16 to a cousin she had never previously met, explained how some of her husband’s behaviors left her distressed:
“Always conflicts. We fight each time, arguments. He chases me from the house. Even
when I was nine months pregnant, I only waited for the day of the delivery. They had
chased me from my home at one o’clock in the morning. I went out. I went to my aunt’s.
Despite everything, I went there. It was the next day that I gave birth to my little girl
here. I gave birth. I did not see anyone there. My husband didn’t come. No one! The
things negative, bah, I cannot speak a lot because sometimes, this makes me uneasy.
With my past, I lived many things.”
119
On several occasions, her husband would lock the house doors while she was out, leaving her to sometimes sleep in a truck parked nearby. She says her husband, who is unemployed, insults and taunts her when he earns money from a side job. At 21 years old, she remains married and despondent: “I do not have a choice. I have children.” She explained that she does not talk to others about her experiences with abuse and other challenges in her marriage; she once asked her aunt for milk for her child and was reprimanded because it is her husband’s responsibility to care for her. “It is not necessary to be ungrateful. I am here in his home.” In each of these examples, the women described entering their marriage early, before they had the maturity to handle a relationship with their husbands; they point toward their age at marriage, and their limited life experience prior to union, as ill-equipping them to handle their husband’s behavior.
Poor Mental Health and Psychosocial Well-Being
When asked generally about their current state of well-being, many women described pain, sadness, worry, and despair. Themes of challenges to mental health were raised by all women who experienced abuse in their marriages, and further extended to other women in the context of their early marriages. Some women described internalizing these concerns which altered their own mental health, while others further described worrying about daily challenges such as completing their education or taking care of their children. The stress of emotions such as worrying negatively impacts health, as two women described, and additionally serves as a health challenge itself.
One woman, a university graduate, reflected on negative and often hidden mental health consequences she experienced from her early marriage. Described by her former boss at a small
American non-governmental agency as “vibrant,” “progressive,” and “just lovely,” it quickly became apparent in the interview that she still struggles with the consequences of her
120
unsuccessful marriage at age 17. At 27 years old, she is now financially independent and lives in a small, three-room apartment, which she shares with her sisters and cousin. She confided that although divorced, she now has a boyfriend. Despite what appears to be the hip life of an urban dweller,* she immediately opened up that she has never told anyone, including her sisters, about the challenges of her early marriage. She explains that she wishes she did not have the problems of her early marriage – which eventually led to a formal divorce – so that she would be “free in
[the] head” and have no “sorrow”:
“My mental health [is] a little tormented…because I cannot forget. Sometimes, I think:
he, he already has his life. Now, he works, and he has children. Me, too, if it were me,
voila! I would have my children with him. So, you see? That comes to my mind often.
But I tend to forget some of these times since I tell myself the best is yet to come. One
day, me too, I will have my family, inshallah.”
While her story was one of resilience, other women revealed that negative thoughts and worries have changed their self-image and subsequent relationships. A 45-year-old woman who was initially excited about her marriage at age 16 expressed disdain for early marriage because of her experience. She struggles to reconcile why she was abused and why her family treated her poorly after her separation:
“The pain, the sadness. Before, I did not have this. I was there happy. I thought that I was
in love or all that I did was good. Now, I criticize myself. I reject myself. When I sit
alone, I think that it is those who accepted me as you [interviewer] are in front of me. I
ask you a question: do you feel the attitudes of my brothers and my husband? What is my
*Though it was clear she was not wealthy, her ability to live independently in the neighborhood where she lived signified middle class living. The author noted that her life represented a new generation with Western influence, a reminder of cultural depictions such as in Sex and the City, because of the independent sisters, all unmarried, living in a city apartment together. 121
behavior? Because we know that in these days, I would like you to explain that to me that
so that I, also, I change my manner to be loved. I have this in my head, as it is etched
there. That is why I became suspicious because I told myself that I am wrong, maybe one
does not dare tell me that I am.”
For some, the experiences with worry have led to despair and current struggles.
Descriptions of no longer believing in love and pessimism surrounding traditional customs of marriage emerged. One woman, who married at age 14, tried to articulate the ways in which her early marriage has changed her:
“So, now, if I say that I feel love for a man, no, no, no. I no longer have feelings. I can
follow you. We say to love, to love, when you say, ‘I love you,’ me too, I say, ‘I love
you,’ but I don’t find that—”
She was unable to finish as she began to cry, later explained as a sign of distress in a cultural context where crying is rarely seen in public. After taking some time and deciding to move forward with the interview, she later shared, “I have experienced only hatred, hatred only, because all things happen badly in my heart. This happened, so I hate marriage, even now.”
Sadness and anger further translated into discouragement and lasting despair in the context of some women’s early marriages. In describing her current well-being, one woman explained that while she is physically healthy, she remains hopeless. She stated, “Morally
[mentally], that’s what bothers me. I can say, in other words, I have no future, quoi. I have a home with no future. That is what I can say.” Her sentiments were shared by others, whose mental health declined from stressors experienced in marriage. Another woman explained:
“For me, he had seen the spirit in my head. After there, in his home, I directly died
because there was that. Also, in our home there, just to poison your spirit, after here, you
122
will not make any more of life. After here, you are going to die. After that, you have
problems. So, I had already put that again in my head. After that, you will have problems.
After that, it is finished.”
Shame also reduced psychosocial well-being when a woman’s situation did not turn out to be socially acceptable. For one woman, her early marriage resulted in early childbearing, which increased isolation from her peers:
Participant: …I was even ashamed to be told that I had children. In walking with my
friends, they made fun of me saying that I had children. I was ashamed, I said, even to the
other not to say to people that I had children when I had them. So, I was ashamed of all
that.
Interviewer: Why were you ashamed?
Participant: Yes, because I had had children early. They did not have children. We were
in the same group. We [grew up together]. We did everything together. They gave me in
marriage. It was not my wish, you have seen. Especially in this time of childhood, it will
affect you willingly or by force.
Disappointing experiences, accompanied by strong emotions with lack of support, left lasting impact on women’s lives within and, for some, following their early marriages. Many women described lingering emotions that had a negative impact on their well-being, which plagued them earlier in life due to the timing of their marriage, and which continue to have what they reveal as harmful effects in their adult life.
Physical Health Conditions
A final set of perceived health disadvantages women described experiencing were physical health conditions. While not always explicitly linked to their early marital experiences,
123
some of these conditions may be exacerbated by stress and trauma, as observed in many early marriages for women included in this study. Women suffered from headaches, malaria, gastritis, osteoarthritis, rheumatism, weight gain, and fever. Many women also described having at least one child who is frequently sick.
When directly asked if these experiences were related to their early marriages, many women were unsure. Several women echoed sentiments similar to one woman who stated, “Ah, me, I cannot say anything because man cannot know the future. It is God who knows all. It is He who can do all.” Among the few who were willing to speculate, they described experiencing health problems prior to marriage. A 17-year-old girl and a 24-year-old-woman both suffered from rheumatism, which they described as a chronic disease they grew up with. Conditions such as chronic headaches were similarly discussed by a few women as separate from marital experiences, and women often did not see a link.
Yet to discount the influence of their early marriages on their health may omit pertinent information to understanding women’s current well-being. A 22-year-old woman, married at age
16, described a minor health condition toward the end of the interview. When asked about her current state of health, she explains, “Well, me, I rarely fall sick. Only, I have a little gastritis problem. It is my little problem. That, too, it is when I am too annoyed that it starts.” While she acknowledges that this is a condition she grew up with, she expressed that it worsens under stress and frustration. When previously asked about her marital life and challenges, she describes experiences of stress and annoyance:
“The beginning, for me, really, it was not easy, eh! I was annoyed because I saw it. He
did these things to me that did not please me. I told myself, ‘Ah, when he pursued me, his
behavior, he did not behave like this. Now, he married me. Why does he behave like
124
this?’…I complained a lot at the beginning. Sometimes, I stayed at the house. I cried.
When he went to work, he did not return early. Me, I was alone. I did not have a child.
So, that tired me. We lasted in that. We lasted in that, up to a year, eh! Really, I did not
find myself. I was always in tears. I cried. We did not understand each other.”
The childhood she described as “100% happy” was drastically altered by her marriage at age 16, introducing additional stress that may have aggravated her health condition. These stressors may have been exacerbated by marriage at a young and less experienced time in life.
Perceived Health Advantages
This study also found that several women perceived health advantages related to health and well-being within their early marriages. In fact, some women were adamant that they had no problems within their marriage. As one woman explained, “What I think about marriage, where I am, I never feel bad in marriage because until now, I do not suffer on account of my marriage.”
These favorable conditions, according to their descriptions, are presented across five themes: having children, “good health for me and my children,” access to health care, delaying first pregnancy and birth spacing, and positive mental health and psychosocial well-being.
Having Children
Childbearing and raising children were overwhelmingly described by women as a primary advantage of their early marriages. Childbearing in marriage was described as both a norm and expectation of a successful marriage. Women who were able to conceive immediately following their marriage often described this as a positive outcome. “I was married at the age of
13 but I am going to say thank God because from after my marriage, I conceived. I had children,” a 24-year-old woman described.
125
For some women, having children was the only advantage they experienced in their early marriages. A 21-year-old shared, “Me, my marriage, the only honor that my marriage brought me was my two children. That is it. That is what I can say.” As another woman, who experienced several hardships throughout her marriage explained:
“Well, I will say that the only positive thing that I knew in my home was having children
because they are legitimate. That is what I have as an advantage, my children. So, that is
why I fight today, so my children succeed, so that they take care of me tomorrow and
wipe away all my tears through the sacrifices that I make at this moment.”
The benefits of having children for health and well-being were numerous. Childbearing in marriage brought honor and joy and was the legitimate way to follow customs. For a few women, they explained their marriages had been arranged by their parents due to fears of out-of- wedlock childbearing. In contrast to women who were unable to give birth early in their marriage, childbearing following the wedding was identified as a sign of good health. As one woman articulated, “There are certain people who marry also. They have a problem. They do not have children. This, too, is another sickness, you see? Well, I did not have this problem. I did not have this problem.”
Support and feelings of love accompanied many women’s experiences with childbearing.
A 45-year-old woman, who is a mother of four and now divorced, explained that her children were all she had to live for following her abusive early marriage and aftermath of a bitter divorce. When asked about the benefits of having children, she explained, “Because they love me. They support me. They suffer at my side. They do not want me to cry. There are all these feelings.”
Although childbearing may be viewed as a benefit of any marriage, women described that
126
marriage at an earlier age allowed them to experience its advantages sooner than their peers.
They described the pride in having several children who, in turn, provide support to help when they are overburdened or sick. A 46-year-old woman shared that her marital experiences have been overwhelmingly positive, including that she has birthed several, healthy children:
“At that moment, I saw, uh, I saw the advantage. Because at my age there, sometimes, if I
dress nicely, you cannot see that I have given birth to seven children. And my children
grew up in front of me. I educated them, and they are married also. The girls there, there
is one who did not have children, two who had children. Me, I have seen my grandsons
already at my age. (Laughs.) So, that, it is an advantage.”
She repeatedly shared her pride in her marriage, which she entered at age 15, and described that having legitimate children at a younger age allowed her to live to see and grow up with her children and grandchildren as well. This pride was further displayed by a 24-year-old woman who had four living children at the time of the interview. She described never having the opportunity to be a teenager herself – “At the moment where I started to play around, I was married at the age of 13 years. It is there that I would start to play around a little, but when I was married, I did not play around.” Yet she was required to marry by her parents and submitted to their request. The only worry she had on her wedding day that she could recall was her concern over “marrying early,” which she did not view positively. She had her first pregnancy shortly after her marriage and, over 10 years later, now views her marriage more favorably. As she talked about her children and her daily household responsibilities to maintain the home, she explained, “Soon, my daughters will be big. They start to do my housework. There is a lot of housework that I can no longer do. All this, if you calculate, you will find that to marry early is beautiful.” Some women perceived early marriage to lead to the benefits of successful
127
childbearing more quickly than if their marriage and subsequent pregnancies had occurred at a later time.
“Good health for me and my children”
When asked broadly, many women considered themselves to be in good health. They defined good health as the ability to live within expected cultural and gender roles, including working, completing housework and daily responsibilities, accessing food and shelter, and maintain a marriage.
“Ah, I have many good things. I am not sick, lying down. I am in good health. I do all of
my work. I go everywhere that I want. I eat. I am not frustrated or sick. This is a lot
because it is if you are in good health that you can do all these things yourself, cook to eat
or look for food to eat or think, or think of good things. When you have health, you have
a lot. You have all.”
Access to food and the ability to eat, as well as basic functions of being able to move around, were repeatedly raised as signs of good health.
Many women described positive influences their early marriages had on their health, though for some, being in good health was explained as beyond a woman’s own control. As one woman described, “I am with my husband. God gave me health. I am not sick tired. I am not sick so as to fatigue the parents.” A 30-year-old woman described how her marriage has positively influenced her state of well-being:
“What makes me say that I am in good health at this moment, because a woman, when
you hear that a woman is in good health, in your home, if you have a calm heart, you will
be in good health, but if you do not have a calm heart, you will not be in good health.
Because each time that you are stressed, you will fall sick. Me, my health is related to my
128
marriage at this moment because it goes well in my home, and I feel I am in good
health.”
Her reflection over 14 years of marriage was consistent with experiences of others who assessed their marriages as “calm” or low stress.
Women also described having at least one child in good health, which some women directly linked to their marriages. Good health was determined as the absence of birth defects, chronic illnesses, and worries or daily stressors, the latter of which was made possible by the security of a marital household. As one woman described: “The children are in good health because they are children. They know nothing. They have no worries. They have their health.”
Another woman further revealed how her children’s good health was explained in their lack of daily stressors and in their socially normative behaviors: “My children are in good health at this moment because they study, and they concentrate on their studies. Because they have their father, they do not suffer for anything. All this permits them to be in good health.” The security provided these women and their children in the context of marriage was perceived as an advantage they attained earlier in life.
Access to Health Care
Another theme of perceived health advantages that women described in the context of their early marriages was access to health care, an important catalyst toward supporting and improving the health of women and their children. Through marriage, and the additional relational and financial support it provided, women were encouraged to seek out health services when sick. In other examples, women were provided with additional health services or given access to resources necessary to address health concerns. One woman credited her early marriage with her improved health, an answer to her prayers. The 24-year-old, who married at age 13,
129
described her past health problems:
“Sometimes I have headaches, sometimes. Before that was my illness, before I was
married, when my head hurt, it was like this. (Points to red eyes.) But the fact that I
prayed and prayed to God, they looked for medicine, and God helped me.”
She credited her husband and her in-laws support with her health improvements, as she further explained, “It is when I married that it healed. Before I was married, it pained me. But when I married, people looked for remedies. They sent me to the healers who treated me.”
For some, the benefit of their marital union included access to health insurance. Health insurance enabled access to health care and medications, often described as prohibitive and costly for women suffering ongoing health issues. A 22-year-old woman described how she benefited:
“My husband, where he works, they are insured, so we are taken care of by his service.
So, when I fall sick, I am going to consult. I take medication. It is the same for our
daughter too. When she falls sick, we are going to consult in a clinic that is recommended
by his office. Afterwards, we will look for medication. They take care of us.”
For her, having access to health care has made a difference. Her 2-year-old daughter often suffers from cold-like symptoms, needing regular medical attention. Despite her descriptions of attaining living needs through paying rent and buying food and other necessities as “very, very hard,” she is able to address health concerns using her husband’s health insurance.
These women were not alone in describing how their husbands and their families sought out necessary treatment for health needs. For some, this support was particularly impactful in their earlier teenage years. One woman, who married at age 15, had her first child shortly after her marriage. She thanked God and credited her husband and his family for requiring her to go to
130
the hospital when she was feeling sick; it was through their requests that she found out she was pregnant at age 15:
“Yes, with the parents, my husband. You know when you are a child, your spirit does not
go to a certain level. You understand nothing. It is necessary that one pushes you. It is
necessary to do that. It is necessary to do that. They told me that it was necessary to go to
the hospital. It was necessary to go to the hospital. At that time, I started to feel sorry for
myself, a little lazy, headaches, well, all that there. Well, they said it was necessary to
bring me to the hospital. The parents knew now that I am in this with child, but myself, I
did not know what I had. Well, I arrived there. Now, they said that I am pregnant for
some months.”
She viewed this support as a health advantage of her early marriage at a time when she had little experience with pregnancy and needed guidance. Indeed, some women benefited from increased access to health care earlier because of their early marriage.
Delaying First Birth and Birth Spacing
Themes related to delaying pregnancy were described as a health advantage of a few women’s early marriages. They described their husband’s willingness to postpone pregnancy immediately after marriage, or to permit delays in having subsequent children, as a benefit afforded to them because of the timing of their marriage. In these cases, childbearing was usually delayed due to the woman’s pursuit of schooling or work, given her adolescent age.
A 25-year-old woman, who described an overwhelmingly positive marriage since age 17, explained that she asked to delay childbearing to continue to her studies. In fact, she explained that her husband has given her everything she asked for. She is a third-year medical student at a local university, an accomplishment she was able to achieve without the responsibility of raising
131
children. She did not want children immediately following her marriage, as she knew would be expected, based on norms and traditions:
“As I had not finished schooling, it is me who told him to wait until I finish my studies. If
not, I could not take my child and go to school. There is no one to help me. He accepted.
There is no problem. That’s it.”
At the time of the interview, she now says she is “taking treatments” to enable her to have children.*
Another woman, who first gave birth quickly following her marriage at age 15, explained that she experienced the benefits of birth spacing within her early marriage. Though she had her first child at a young age, she explained time gaps between subsequent pregnancies:
“Yes, for the first time as I studied, there was a gap between the second and the third
child. There were four years. There were four years. The second, between the second and
third because I studied. I was in professional school. Well, uh, after the fifth, fourth
pregnancy and fifth pregnancy, there were 10 years. There was a 10-year interval because
I was looking for work. All that there. Well, I told myself with the four, it was necessary
to develop with the four there, that there first.”
She explains that she and her husband continued birth spacing when she received her first contract as a teacher and later when she was appointed to a full-time public-school position where she was required to teach in whichever community they sent her:
“When I started to teach, as contractual at first before I was committed to a public
function, at this time, these were the [pregnancy] intervals there. With the transfers,
*This raises questions of situations in which delays in childbearing that were described as intentional and beneficial simultaneously appeared to also rationalize the cultural abnormality of not having children after prolonged time in marriage. How women reflect on and interpret their experiences will be considered in the discussion. 132
because you are newly hired, they send you first in the bush, in the villages. With
transfers going and coming, with many children, you do not have support. It is going to
tire you. The children could not even study very well. So, at this time, I stopped to be
stable somewhere first. Well, with the older children there, they were a little bigger. I was
with some [of my children], some also were with my mom. Umm, so, it is the interval
there.”
She was the second (and, at the time of interview, last) wife to her husband, whose family helped arrange the marriage as they searched for an educated wife for their son, whose first wife is illiterate.* With agreement that she should continue and use her education, her husband was supportive of extended gaps between pregnancies. These examples reveal that a few women perceived protective benefits of their early marriages with regards to childbearing, which they attribute to the timing of their early marriage in their educational and professional journeys.
Positive Mental Health and Psychosocial Well-Being
Women described positive sentiments evoked by their early marriages. These sentiments, including happiness, love, and understanding, were further related to relationships emerging from their marriages with their husband, extended families, and in-laws. Some women talked about these positive emotions on the day of their wedding. A 27-year-old who married at age 14, described her recollection of her wedding day thoughts as happiness. For others, these sentiments persisted years after entering marriage:
*Polygyny, a form of polygamy where a man marries multiple wives, is a common practice in Guinea. Though less common today, UNICEF (2005) found that nearly 80% of women in Guinea belonged to polygynous unions. Although interviews in this study probed for belonging to a polygamous household and wife order, this rarely came up as an experience or salient context mentioned by participants. 133
“In marriage, in my home, there is nothing bad except good. For me, it is happiness. But
sometimes, there is not understanding, but I do not consider this. Even when this
[disagreement] arrives, I roll my eyes. Me, it is understanding that I consider. In my
home, it is understanding only. I get along with my husband’s family. When I ask for
help in his family, they help me without problem until I am satisfied.”
Similar sentiments were even described by women who had also perceived several disadvantages of their marriages. One woman, who suffered abuse and later divorced, shared that she enjoyed her marriage at the beginning. She said that she did not experience fear at the time when she married: “No, I wasn’t afraid. Love was there.” Though she now describes lasting impacts of abuse she endured in marriage, she reflected that her marriage was not all negative:
“Yes, the positive things that I lived in my marriage were sincere, kind.”
Moreover, for some, these advantages must be understood as related to the timing of their marriage. Marrying at an early age allowed women to experience these emotional advantages earlier than their peers. As one woman described:
“The fact, to have a person who loves you, it is a benefit also. He will do all things for
you, all that you tell him to do for you, he does it. Throughout where you will tell him to
let you to go, he will let you to go. That, also, is a benefit. In our marriage, between us,
he put democracy. There is no dictator.”
Some women proudly described the way in which they now counsel friends and neighbors who suffer challenges in their lives. They explain that their position as married women, and in some cases mothers, earns them respect in their families and communities and describe how they can use their position to support others, which they ascended to earlier because of their early marriages.
134
Factors Influencing Perceptions of Early Marriage
A final set of findings further contextualize women’s perceptions within their early marriages. Women’s descriptions of their lives and experiences revealed several factors which may help to understand varying descriptions of perceived health disadvantages or advantages.
These broad factors included: husband’s traits, perceptions and experiences with support, and individual traits. Importantly, these factors are likely relevant for all relationships, yet seemed to matter disproportionately to participants who married early because of the earlier timing of their marriage.
Husband’s traits. Characteristics of the husband heavily influenced women’s experiences within their early marriages. Alongside descriptions of perceived health advantages were often descriptions of husbands who listened to them and engaged them on matters pertaining to themselves or the household. One woman described some of the most positive moments in her marriage: “Sometimes, I was very happy. It was the times he consulted me.” Moreover, health advantages were often discussed in relation to the husband’s sharing of resources, getting what she asks for from her husband, or receiving permission from her husband to do what she wants to do, such as continue schooling. One woman described that many men do not give their wives spending money, even when they have the finances to do so; she furthers that her husband provides her with this, and so she has no problems, but advises other women “to wait until the age of 18 years” to marry, in case their husband is not like hers so that they are better able to negotiate with their husband and take care of themselves.
On the contrary, women who described several disadvantages of their early marriages explained their husbands failed to consult them, provide spending money, or spend time with them. “It hurt,” one woman explained when her husband made financial investments for his
135
siblings without informing her or seeking her opinion. She was advised by her mother to submit and “stay behind him,” since her husband was providing for her children, but she described that instance as the first of a series of disappointments which preceded abuse he caused her and allowed others to inflict on her. Inadequate financial means for women’s livelihood was a frequent source of arguments and perceptions of health disadvantages, particularly when a husband earned money but did not provide anything for his wife or children. One woman explained:
“He didn’t give me spending money. I did not have [food] to eat. This is what made me
leave. I stayed nine years with my parents. After nine years, he came to say that he
wanted me to return. Now, he gives me money for spending.”
Women who perceived disadvantages also described husbands who were adulterous, severe, or violent. “It just got worse. He was always violent,” a 27-year-old woman who suffered physical abuse in her marriage explained. “I wanted to try to force my spirit to accept him, but he was more and more violent, arrogant, really difficult.” The personality and behavior of women’s husbands impacted women’s experiences in their early marriages.
Perceptions and experiences with support. Women’s perceptions and experiences with support in their marriages also shaped perspectives on health advantages and disadvantages.
Women who described advantages talked about positive relationships formed around the context of or within their marriage. Women described being welcomed and treated well by their in-laws, a culturally important experience as marriage in this setting, unites not only two individuals but also their families. One woman exclaimed, “Everyone loves me in his family, even my mother- in-law,” signaling that acceptance by in-laws was not a guaranteed reality. Another woman explained:
136
“My in-laws love me. My husband’s big sister loves me. That is what I looked for
because I know that if all these people love me, my husband is going to love me. If your
mother-in-law loves you, your husband will love you. If your husband’s big sister loves
you, your husband will love you, because these are the people who can separate you from
your husband.”
Some women received financial and emotional support from their own families once the marital proceedings began. One woman explained that her marriage was arranged by her parents, and she chose to accept it. She described several advantages she has experienced including support from her family: “At the time of my marriage, I had a lot of money. My brothers contributed for me. My aunts, my uncles, my sisters, my friends all contributed for me.”
Additionally, women described receiving support through being counseled or getting advice from family or neighbors, which helped them overcome situations in their marriage that could have created problems.
Conversely, women who described health disadvantages of their early marriages discussed not having support from others. Women, moreover, shared problems they had with in- laws or neighbors, which not only created hostile living environments for women but often created or exacerbated behaviors from their husbands that led to disadvantages in health and well-being. Problems with co-wives, additional wives for those in polygynous households, sometimes emerged as another challenge to support. One woman remembered the lack of support she received in her first marriage, at age 14:
“When my co-wives wanted to fight me, they did so even in front of my husband. If I
cooked, they did everything to spoil my rice so that my husband fights me. I could do
nothing except leave it up to God for me.”
137
However, not all women who married into polygynous households found lack of support from their co-wife(ves). In fact, one woman who overwhelming described only advantages in the context of her early marriage was the second of two wives in her household.
Individual traits. Finally, women’s personalities, thoughts, and behaviors played an important role in their perception of health advantages and disadvantages within their early marriages. Traits that promoted perspectives of health advantages, or enabled women to persevere following disadvantages, included being courageous, hardworking, and resilient. One woman explained that she ignores less positive aspects of her marriage. She explained:
“I see negative things, but I do not consider this. Where I am there, it is my husband I
look at and the advice that my parents give me. When I sit down, I think to this. It goes
[well]. It goes as if I do not see the negative things.”
Another woman, who was able to overcome the perceived disadvantages of her early marriage, described several ways in which she worked extremely hard to overcome these challenges:
“I lived nearly one-and-a-half kilometers away, and after that, I walked by foot into the
neighborhood on the hill. Each day, I fastened two cans of water, of candy, to send to the
camp here and camp here is so far away. I walked by foot to get back, and that was on my
head, two cans of water tied on the candies. In the morning at 6 am, I returned. I took that
to send to the school to sell the rest that was not finished. I returned to the market. I sold
the rest. I took the spending money there. I bought condiments. I went to cook and in the
evening, I return to the school for the revision classes. It is like this that I lived during
three years with insults, violence, illness, all, all, all.”
Although she describes several disadvantages, she says that the biggest advantage of her early marriage was suffering because it has pushed her forward and shaped her into who she is today.
138
Women also described personal traits of being patient and submissive alongside perspectives of advantages of their marriages. One woman explained how her patience prevented problems that could have arisen from lack of money: “I am here, patient. If my husband earns and he gives me, okay. If he does not give me, I am there, patient, because of God and his prophet Mohammed, peace is given to him.” Additionally, when women saw their situation as better off than someone else’s, it contributed to perceiving advantages. Among women who reflected on overlap between their marriages and what they knew about their mother’s marriages, one woman shared a conversation she had with her mother on her wedding day. She describes,
“My mom, too, had cried, but she had said that I was lucky to marry at the age of 16 because she married at the age of 10.” Her own personal disposition as well as the perception that her situation was better than her mother’s may have influenced descriptions of a perceived advantages within her early marriage.
Women who described health disadvantages in their early marriages often revealed a dimension of naivety in their relationship that they struggled to overcome. Many had no prior experience with men or with managing a household, which led to challenges in their relationships. In the words of one woman: “At the time when I married, I was only almost 15 years. I knew nothing of life. So, my parents insisted that I married, and I married.” Another woman, whose early years of marriage presented many challenges similarly explained:
“I had many difficulties because I was a child. I did not have experience. To find myself
in a home, to take care of her husband before he went to work, to cook, all that there, to
wake up in the morning to cook, clean the house, wash the clothes. I was not used to all
this. And I just found myself in this alone!”
139
Some women had ideas of what marriage would entail, yet their expectations were not met.
Some women had false ideas of what marriage would entail and realized only once in marriage how their lives would change. One woman described her excitement on her wedding day: “I was there singing, ‘I am going to marry! I am going to marry!’” After speaking with her aunt, and telling her that she would soon return to her parents’ home to continue her studies, her perspective started to change:
“She told me, ‘No, after the marriage, you are going to go. We are going to get you
ready. You are going to sleep there. In the morning, you come back.’ This is how they
told me. For me, it was a reality because we didn’t sign the marriage. We did the
religious ceremony. When we finished doing this, during the night, I was with my
friends. I had already cooked. That day there, it was a Saturday, we had to go to the youth
dance hall. Today, we say the youth house. Before [the marriage] was permanent, I had
taken my dresses, all my things, and gave them to my friends for me. After the ceremony,
I was free like I am at my parent’s house. When my parents came to drop me off at my
husband’s parents’ home… They told me, ‘No, here, it is your home now. You can no
longer leave.’ I started to cry. It is like this that I understood. I started to understand
marriage.”
Although she knew her husband prior to marriage, many of the women did not. Yet lack of prior interactions with the husband did not presuppose a challenging marriage, particularly when the husband was understanding and the woman had additional supports to help navigate her marriage. Such individual traits appear to influence the success of early marriage and seemed important for shaping women’s perceptions of their experiences.
140
Discussion
Summary
This study revealed that child marriage is not synonymous with a singular experience; among participants, women who married early ranged in age, ethnic background, education level, and profession, and married a diverse set of men. Findings do not support claims of causality of a prototypical girl child bride in Guinea, or, in other words, that a singular type of girl (e.g., young, poor, out-of-school) is solely affected by child marriage in this context and that there is one type of husband that she marries. Existing literature highlights trends among women who marry as children: in settings such as Guinea, child marriage is associated with girls who are poorer, less educated, and in rural areas (UNFPA, 2012). While this study does not counter such claims, it calls for a more inclusive understanding of who might be affected by child marriage.
Additionally, while findings may point to reduced poverty as one of several drivers that could influence girl child marriage, it most readily supports the roles of family influence and community values as push factors toward early marriage.
Perceived Health Disadvantages in Child Marriage
Women who married as children in this study experienced many perceived disadvantages within their early marriages including poor sexual and reproductive health, intimate partner violence and related long-term consequences, poor mental health and psychosocial well-being, and physical health conditions. These findings were consistent with existing literature, which describes the harmful effects of child marriage on women’s health (Lendhardt et al., 2016;
Machel et al., 2013; UNFPA, 2016; UNICEF, 2011). Some empirical studies show girl child marriage is associated with poor sexual and reproductive health outcomes; women who married as children were more likely to have multiple and rapid childbirths, unintended pregnancies,
141
pregnancy termination, unmet need for contraception, and were less likely to access maternal health services than adult-married counterparts (C. J. Clark et al., 2017; Godha et al., 2013;
Kamal & Hassan, 2015; Raj et al., 2009; Santhya et al., 2010). A recent qualitative study in Iran revealed that women married as children did not understand contraceptive use and felt pressure to become pregnant immediately following marriage (Mardi et al., 2018), which echoed some themes emerging in this research. Intimate partner violence has been empirically associated with child marriage (Erulkar, 2013b; Raj, Saggurti, Lawrence, et al., 2010), and, with early marital ages, can further exacerbate reproductive health outcomes (C. J. Clark et al., 2016). Women married as children across settings are more likely to have poorer mental health outcomes, including suicidal thoughts (Gage, 2012) and diagnosed psychiatric disorders (Le Strat et al.,
2011; Wondie et al., 2011). Women married as children have also described experiencing frequent physical pains and other health conditions more commonly than adult-married counterparts (Nasrullah et al., 2014), consistent with other health concerns emerging from this research.
Similar harmful consequences of girl child marriage have been highlighted in grey literature, including non-profit organization and health services reports. For example, a recent study of sexual and reproductive health among migrants and refugees in Australia and Canada included qualitative evidence that echoed perspectives shared by some participants regarding early marriage following menarche, lack of sexual education, and difficult experiences with first sex (Ussher, Metusela, Hawkey, & Perz, 2017). Although empirical evidence from African contexts remains limited, the harmful effects of child marriage on women’s health have been described in the West African region, where rates of girl child marriage are at their highest
(UNICEF, 2015b; Walker, 2013). Findings from this study are important for illuminating
142
possible harmful consequences of child marriage in the Guinea, where, despite the frequency of marriage, little evidence currently exists.
Perceived Health Advantages in Child Marriage
Importantly, this study illuminates that not all women in Guinea perceived their early marriages to negatively impact health. In fact, in contrast to dominate narratives in public health literature, some women perceived health advantages. Presenting women’s favorable perspectives may provide critical information toward understanding why the practice persists. Though descriptions of health advantages within early marriage are largely missing from public health literature, there are a few studies that raised overlapping themes. Favorable perceptions of child marriage and health were captured by Nasrullah and colleagues (2014), who conducted the only other qualitative study identified that intentionally asked women about the broad health consequences as related to their child marriages. They found that in an urban slum in Pakistan, most women were unaware of negative health outcomes associated with child marriage and expressed a desire to have their children married early as well (Nasrullah et al., 2014). In African contexts, some work on child marriage has found young women see its benefits as a transactional exchange; internally displaced persons and refugees in a Ugandan camp described entering sexual relationships to obtain basic needs and getting formally married if the woman becomes pregnant (Schlecht et al., 2013). Early marital age has additionally been encouraged to protect the girl child and to uphold cultural values; it ensures she gives birth within marriage, preserving local norms, and can be further exacerbated in conflict and fragile state settings (Myers, 2013).
In the context of this in-depth qualitative study in Guinea, having children within marriage was similarly perceived by women as an advantage for a woman’s well-being.
143
Yet other health advantages women described in this study from their early marriages suggest that perceptions of its protective nature are salient in their adolescent and adult experiences. Beyond having children, such perceived advantages – including good personal and child health, access to health care, delaying first pregnancy and birth spacing, and positive mental health and psychosocial well-being – have not been highlighted in the literature, importantly providing insight on how girl spouses themselves make meaning of their experiences. For example, as previously discussed, sexual and reproductive health has been documented as a negative outcome of child marriage in both quantitative and qualitative literature; yet in this study, some women perceived that the opportunity to wait to get pregnant or to take time between pregnancies – decisions largely driven by their husbands – would not have occurred in a marriage, given cultural norms, if they had married at a later age. Similarly, many women perceived additional resources provided within the context of marriage, such as increased access to health care, as benefits they received earlier in life because of the timing of their marriage. Such qualitative findings may provide additional context for research findings that girl child marriage is not only associated with only negative health outcomes (Efevbera, Bhabha, et al., 2018a). As another example, women overwhelmingly saw their early marriage as good for the health of their children, or having no impact, yet recent evidence from sub-Saharan Africa shows associations between mother’s early marriage and mechanisms for poor child nutrition and development (Efevbera, Bhabha, et al., 2017). Importantly, the author does not suggest that these perceived health advantages are benefits of all girl child marriages; instead, they must be understood as perceptions women themselves, who married as children, offer from reflections of their experiences.
144
Contextual Factors Matter in Women’s Early Marital Experiences
This study found that traits unique to a woman and her husband, as well as her perceptions of support outside of the marital household, strongly influenced her perception of experiences in early marriage. Such findings reveal that beyond the timing of marriage itself, there may be contextual factors – some of which are modifiable – that lead to diverse perceptions of and experiences within marriage. One way in which these additional contexts may be understood is within Urie Bronfenbrenner’s bioecological framework, a layered model to understand the broader environment in which a child develops (Bronfenbrenner, 1977).
Bronfebrenner theorizes that individuals do not develop in a vacuum but rather within a web of relationships, as these data from Conakry demonstrate. The characteristics highlighted by women in this study include those within the biosystem, or her own unique characteristics (e.g., her age, her disposition to handling adversity, naivety); microsystem, or her interactions directly within her immediate environment (e.g., parents, other family members, husband, neighbors, peers); mesosystem, or interactions between different microsystem settings (e.g., her husband’s interaction with his in-laws, which, in turn, may influence how she is treated by either of them); and macrosystem, or the broader beliefs, cultures, and structures that create norms within a society. In other words, women were impacted beyond just the experience of marriage itself and these broader layers of context influenced their perception of health advantages or disadvantages within their early marriages.
Few studies identified in public health literature similarly looked beyond consequences of early marriages to understand the factors that might shape women’s experiences and perceptions, missing an opportunity to understand how women construct meaning. From this lens, the outcome of perceiving advantages of one’s early marriage is, thus, rational when understood
145
within the full context of women’s daily lives. Ethnographic literature on early marriage in other
African communities has recognized the importance of “ethnographic contextualization” to understand the proliferation of early marriage practices (Archambault, 2011, p. 632) and women’s descriptions (Callaghan et al., 2015). Archambault (2011), moreover, emphasizes the importance of this in the context of exploring events framed as human rights concerns:
Anthropologists have been on the forefront of criticizing representational frameworks
within human rights discourse (Wilson, 1997). “Legalistic” accounts of human rights
violations are said to strip events of their social meanings and subjectivities and conceal
the ambiguities and contingencies that are at the heart of acts of injustice. This goes
against the very goal of ethnographic investigations of human rights practices, which aim
to restore subjectivity and contextualize rights violations by exploring their local
interpretations and “vernacularizations” (Goodale, 2007; Merry, 2006a,2006b). (p.634)
Similarly, in psychology and education research, individuals are understood to construct their own reality and make sense of their experiences (Ignelzi, 2000). In the context of human development, this meaning construction, or meaning-making as popularized by Robert Kegan, can be understood as an important set of processes and outcomes influenced by a person’s interactions and experiences. Thus, to advance research on phenomenon such as child marriage, which as important global considerations, there is a need to further ground research in participants’ views, experiences, and broader context.
Holding Multiple Truths
Readers may be left wondering how women married as children interviewed in this study could perceive both disadvantages and advantages of their early marriages. Recognizing that meaning-making is fluid, one possibility is that women have engaged in restorying, where they
146
tell and retell tales of their past experiences in different ways, which shape their memories and identities. Restorying can be a powerful tool for overcoming trauma and for survival; it is intentionally used in psychotherapeutic approaches such as narrative therapy today (White &
Epston, 1990 ). It would, thus, be dishonest to neglect acknowledging the possibility that the themes of perceived health advantages are a product of women’s psychological response to their experiences with early marriage.
A second possibility, which the author’s collection of “rich data” in this study supports, is that there are, in fact, benefits experienced by women of their child marriages. Like a woman’s development, these marriages, as has been explained, do not occur in a vacuum but rather in a girl’s bioecological context. Glimpsing into these contexts helps reveal that early marriage, for some women, can fulfil real needs, such as making the family happy, bearing children, accessing health care, and, in Guinea, being a functional member of society. Women’s perspectives of both perceived advantages and disadvantages thus illuminates a gap the World Health Organization has identified: a need to understand both the costs and benefits of child marriage (Svanemyr,
Chandra-Mouli, Raj, Travers, & Sundaram, 2015).
Additionally, it is important to contextualize the major research findings on themes of perceived health disadvantages, perceived health advantages, and factors influencing women’s perceptions through three points. First, the intention of this study is not to defend child marriage, which is illegal in the context of Guinea; the author instead seeks to illuminate perspectives and experiences of women themselves who are affected, as these are voices glaringly missing in current literature. Second, apart from one 17-year-old who had recently married, participants in this study narrated and reflected on their experiences with early marriage after their childhood and, in some cases, young adult years; consequently, most women provided retrospective
147
accounts of their experiences, and for some women, this occurred after their marriage had dissolved. Finally, understanding this study’s results requires that readers hold multiple truths.
What may seem counterintuitive or confusing from a Western perspective, for example, may be what participants from the Guinea context expressed or believed. Thus, their constructions of their reality are taken as truths.
Limitations
This research only tells part of an important and under-discussed story, and it is important to explicitly clarify what it cannot do. This study cannot further disentangle the experiences of early marriage from marriage in general, as it is impossible for women themselves to know their counterfactual life if they had been married at an older age.* The sample of 19 women cannot be used to make nationally-representative claims, as public health research often aims to, yet this was not an intent of the study. Instead, a mid-size study population was identified through purposive sampling, reaching saturation of new themes and ideas. Appropriate qualitative research methods were used, and while the author situates results in the broader context of literature, there is makes no attempt made to generalize beyond the context in which data were collected. While certain researchers and practitioners with an interest in designing scalable policies and programs may view these as study limitations, these instead should be understood as intentional conditions of the study design that can set the stage for future research.
Methodologically, there are some additional limitations of this research to highlight.
Though interviews were conducted in the language of choice by participants, it is possible that
*Additionally, certain contexts of their marriages which may be interesting and relevant – such as polygamy and divorce – were probed for and relied on participants’ willingness to share and understanding of these as relevant experiences to their marital experiences. Consequently, it is possible that not all participants offered this information. These would certainly be interesting areas for future research. 148
language and cultural barriers affected translations; the author is conversantly fluent in French but not a native speaker or Guinean, and translations from local languages to French to English could have had minor impacts on loss of meaning. The author took women’s contributions at face value, and where possible triangulated with observation. Asking women to reflect on their entire lives could be subject to recall bias, and their willingness to share personal and sensitive data with the author, a foreigner, could have been reduced. For example, one woman stopped her interview after 20 minutes and withdrew from the study over concerns of who was collecting information and how it would be used; other women who had been approached to participate in the study within that same community described ethical concerns lingering from foreign researchers during the Ebola outbreak. For most participants, acquiescence and social desirability bias, or concerns over a participant providing the author with a favorable or agreeable answer, was not a concern; in fact, during debriefs following an interview, RAs were surprised with the level of detail participants provided the author on their private life experiences. Attention was given to question framing and ordering, neutrality of facial expressions and voice intonation was used, and cultural differences permitted the primary author to probe from curiosity and uncertainty (rather than, as RAs initially struggled with, countering participants’ questioning why they, coming from the Guinea context, would ask the questions posed). Relationship- building was achieved through brief interviews and participant observation and was essential to building trust.
Strengths
Moreover, this study has several strengths. This research fills several gaps in current literature. It provides rigorous data on girl child marriage in the context of Guinea and is the first study, as far as the author knows, to explore this. This study used multiple ethnographic methods
149
extending beyond single in-depth interviews to more fully contextualize women’s experiences including: brief interviews that enabled multiple contact points with some participants, observation of participants in their homes to corroborate or contradict what they shared in interviews, and participation in aspects of women’s lives when invited to do so as an additional window into their context. “Rich data” were also collected in this study through dozens of memos and pictures, and though not fully integrated into this manuscript, these data informed the preparation of the results.
Methodological strengths included the use of multiple and in-depth qualitative methods is an important contribution to existing quantitative and advocacy work on girl child marriage.
Study results are taken directly from the words of women themselves, illuminating how women affected perceive and experience their situation. Additionally, rooted in a grounded theory framework, this research did not apply theoretical priors assuming what the consequences of child marriage might be for women in this context. Open-coding generating emic themes, which allowed an open curiosity toward women’s perspectives, consequently, allowed for a non- pathologized view of experiences of girl child marriage in Conakry and its relationship to health for these women and their children. As part of a much larger mixed methods dissertation, where prior quantitative study estimated associations between girl child marriage, women’s undernutrition, and child development across African contexts (Efevbera, Bhabha, et al., 2017,
2018b), the author can further interpret and draw inferences on the impact of this social phenomenon.
Conclusions and Next Steps
Through this study, a new picture emerges about how women married as children perceive early marriage as related to their health and their children’s health. First, women
150
perceived health disadvantages of their early marriages, captured through themes of poor sexual and reproductive health, intimate partner violence and long-term consequences, poor mental health and psychosocial well-being, and other physical health conditions. Second, women also perceived health advantages of their early marriages, captured through themes of having children, “good health for me and my children,” access to health care, delaying first pregnancy and birth spacing, and positive mental health and psychosocial well-being. Third, women’s perceptions corresponded with descriptions of husband’s characteristics, individual characteristics, and support. These findings could be considered in the context of many marital experiences in several contexts, yet women’s reflections point toward the importance of the timing of the marriage in their lives.
Addressing women’s health in Guinea will require thinking about women’s complete contexts. One in two women in Guinea marry before their 18th birthday and, as this research shows, this event can profoundly impact their lives. These findings are important to inform evidence-based interventions and policies to address girl child marriage in Guinea through an understanding of how women themselves experience its effects. Women’s perceptions of health disadvantages and advantages may influence not only their own health but intergenerational promotion or disruption of child marriage practices. If women themselves see the experience positively, they may continue to perpetuate the practice intergenerationally, such as through encouraging their children to marry early, impeding efforts to prevent child marriage. Change in preventing this illegal practice and addressing its full set of consequences cannot and will not happen effectively without incorporating the voices and experiences of women affected; listening to women, as this study aims to do, is a necessary first step.
151
Chapter 6: Discussion and Conclusion: Revisiting Findings and Exploring Implications of Research on Girl Child Marriage and Women and Children’s Health and Well-Being
Summary
The romantic language of the international human rights community drew the author, an advocate of social justice issues at heart, to this work. The common discourse in law, advocacy, and research that girl child marriage is wrong, a violation of rights, and harmful to the health and well-being of girl spouses made moral sense. Yet there were gaps in understanding its health consequences, particularly in some of the settings with the highest burden. This dissertation, thus, had the objective to use a mixed methods research approach to improve our understanding of the complex relationships between girl child marriage and the health and well-being of women and their children in sub-Saharan Africa. This objective was accomplished through three aims:
Aim 1: To establish the extent to which girl child marriage contributes to socioeconomic
status and underweight, a measure of undernutrition, among women in sub-Saharan Africa.
Aim 2: To understand the mechanisms through which girl child marriage affects the early
development and health of children born to women who marry early in sub-Saharan Africa in
order to quantitatively examine its intergenerational effects.
Aim 3: To qualitatively explore perceptions of the social construct of “girl child
marriage” among women in a community, and to explore the perceptions that women have of
the relationship between marriage, health, and well-being.
These aims were further connected in the conceptual map in Figure 1.2, which proposed that girl child marriage was related to health and could more comprehensively be understood by assessing how different definitions of girl child marriage were related to different health and well-being outcomes for women and their children. More specifically, the conceptual map
revealed that each chapter of this dissertation illuminated a component to further our understanding of the relationship between girl child marriage and health through:
reviewing literature and identifying gaps (Chapter 1);
presenting a critical discussion of the choice of term “girl child marriage” (Chapter 2);
using all available Demographic and Health Surveys of nationally-representative
household data for sub-Saharan Africa to estimate associations between girl child
marriage and women’s health, in the example of underweight as a measure of
undernutrition (Chapter 3);
using all available UNICEF Multiple Indicators Clusters Survey of nationally-
representative household data for sub-Saharan Africa to estimate associations between
girl child marriage and children’s health, in the example of early childhood development
and stunting (Chapter 4); and
qualitatively exploring women’s own perceptions of the health consequences of their
child marriages in Guinea, a country in the top 10 highest rates of girl child marriage
worldwide (Chapter 5).
In this final chapter, the author summarizes research findings and highlights themes emerging across the empirical chapters of this dissertation. Previous chapters each highlighted study- specific findings, so here, considerations are taken from the dissertation in its totality. This chapter concludes by considering broader implications on child marriage research, practice, and policy.
Findings
This dissertation revealed several findings. In Chapter 1, the extensive review of literature made three points clear. First, girl child marriage is negatively associated with the
153
health of women married as children, as evident through maternal and reproductive health, mental health, intimate partner violence, HIV, and other health-seeking behavior measures.
Second, girl child marriage, particularly as proxied through early childbearing, is negatively associated with children’s health. Third, existing literature had yet to look at all health and developmental outcomes associated with girl child marriage, particularly in sub-Saharan Africa, which informed the outcomes, sample populations, and methods used in subsequent chapters. In
Chapter 2, girl child marriage was described as a politicized terminology, intentionally used to convey this dissertation’s focus on a special population. Chapter 3 described that although there were negative associations between girl child marriage and socioeconomic status, girl child marriage appeared to be protective for undernutrition in some African populations, contrasting with existing narratives and limited empirical work from South Asia. Chapter 4 described that girl child marriage was associated with being stunted and off-track for early childhood development in African contexts, empirically adding to narratives of the intergenerational effects of early marriage; these associations were further explained by mechanisms beyond early childbearing. Chapter 5 revealed that when women themselves, who married before age 18 and described their own marriages as early were asked to reflect on their marriage as related to health, they perceived both health disadvantages and advantages; their perspectives were situated in women’s broader contexts in Conakry, Guinea and contribute to a more complex picture of the consequences of girl child marriage.
Contributions to evidence on women’s health and well-being. In part, this dissertation confirms the dominant narrative that girl child marriage can negatively affect the health and well-being of women. Most compelling, women married as children in Guinea perceived health disadvantages (Chapter 5). Women described experiencing negative sexual and reproductive
154
health, intimate partner violence and related long-term consequences, poor mental health and psychosocial well-being, and other physical health conditions. These findings support existing literature on women’s experiences in girl child marriage, adding evidence from Guinea, a context in which literature is currently missing.
However, contrary to existing literature, this dissertation presented both quantitative and qualitative evidence that the consequences of girl child marriage for women’s health may not be exclusively harmful. In addition to health disadvantages, some women married as children in
Guinea perceived health advantages of their early marriages, including having children, maintaining good health, accessing health care, delaying pregnancy, and experiencing positive mental health and psychosocial well-being (Chapter 5). Additionally, the findings that girl child marriage was associated with lower odds of undernutrition among women using data from 35
African countries suggests that it may not be negatively associated with all health outcomes, a surprising result given the prevailing belief furthered through existing literature (Chapter 3).
These findings note that while girl child marriage is still illegal in many of these settings, its perceived benefits may point toward why the practice still exists.
Contributions to evidence on children’s health and well-being. This dissertation also, in part, confirms that girl child marriage can negatively impact the children born to women who marry early. The pioneering work measuring associations between girl child marriage and early childhood development as well as its associations with stunting, showed that girl child marriage is associated with poorer child outcomes across sub-Saharan Africa (Chapter 4). Through testing the conceptual model, this study, moreover, identified biological and social mechanisms stemming from a woman’s early marriage that help to explain children’s outcomes of being developmentally off-track and stunted. Very little evidence previously existed to inform this
155
theorized relationship, and this study can inform future research on the mechanisms for these associations.
However, this dissertation again contributes evidence in contrast with existing literature, suggesting that the consequences of girl child marriage on children’s well-being may not be exclusively negative. The perceived advantages that women in Guinea described in the context of their early marriages extended to the health of their children (Chapter 5). In fact, some women who married as children described ways in which their marriages supported their children’s well- being yet struggled to identify ways in which their early unions could hurt them.
Relevance of Findings
Girl child marriage is a social phenomenon that occurs during adolescence and has implications throughout a young woman’s life and into the next generation. The dissertation shows that, when investigating it as an adolescent development issue across the life-course, girl child marriage acts as a psychosocial risk factor – and to some women, a protective factor. These findings propose that the way one investigates complex phenomena like girl child marriage may affect what one is able to learn; the multi-facetted, interdisciplinary approach used presents findings that are of interest to disciplines within the fields of public and global health, such as nutrition, social and behavioral sciences, epidemiology, population studies, maternal and child health, adolescent health, and women’s health. These findings reveal that girl child marriage may not be simply bad for women and children’s health and well-being; it is, instead, a more complex issue, thus leading to implications for future research, implementation, and policy, which aim to address girl child marriage.
156
Reconciling Results
Across the five chapters, there are messages that are reinforced. Girl child marriage was frequently associated with lower socioeconomic status, including poverty and lower education, in the African contexts explored. Across empirical chapters, girl child marriage was also frequently associated with earlier age at childbearing, another risk factor for women’s health in existing literature. In the quantitative chapters, a large proportion of women married as children lived in rural environments, suggesting that risk factors identified in existing literature may hold in this context. The qualitative work further illuminated that some women may see early childbearing as an advantage of their early marriages, given customs and expectations in their local communities.
Despite many similarities, there were also differences stemming from the mixed methods research approach. The lack of association between girl child marriage and undernutrition
(Chapter 3) might seem surprising when contrasted with the negative health and developmental outcomes for children born to mothers who married early (Chapter 4), as it prevents a clear picture of girl child marriage as beneficial or harmful. Similarly, women themselves did not identify that girl child marriage might decrease risk of being underweight and had little to say about how child marriage might relate to their weight (Chapter 5); they also did not describe negative developmental consequences their children endured because of their early marriages, and, in fact, saw how a household with two parents present could be advantageous.
Readers might be left wondering: how does one reconcile the dissonance across chapters?
The author would challenge this notion as results that need to be reconciled. Instead, one must understand that girl child marriage may have both disadvantages and advantages related to the
157
health and well-being of women and their children, particularly as understood from meaning constructed by those experiencing it.
Although girl child marriage may serve as a catalyst toward negative outcomes for women and their children, these outcomes are shaped and guided by a woman’s broader situational context. As conveyed in Figure 1.2: Conceptual Map of Proposed Dissertation, poverty, low education, and geographic location, for example, may serve as risk factors that impact how a woman’s timing of marriage affects her and her children’s health and well-being.
The adaptation of Bronfenbrenner’s bioecological framework in Figure 1.1: Original
Conceptualization of a Bioecological Model of Girl Child Marriage revealed the importance of context in another way. Here, although one must consider risk or protective factors at the level of a girl or woman herself, one must also recognize that context extends to families, communities, and societies more broadly. Ultimately, the differences in findings across chapters remind us that health and well-being consequences of girl child marriage can only be understood in a given context or set of contexts.
A Unique Situation for sub-Saharan Africa?
To further illustrate the importance of contextual differences, consider dissonance discussed in previous chapters between this dissertation’s findings on girl child marriage in
African contexts as compared to South Asian contexts, where most literature on child marriage comes from. This difference was explored when findings across 35 African countries revealed a slight protective or null effect of earlier marital age on women’s underweight status, in contrast with a study by Goli et al., 2015, which found distinct differences in weight status by marital age in two Indian states (Chapter 3). One argument explored was that the nutritional baseline of these populations was different; the women in the India sample had higher undernutrition at baseline
158
than the average across the African communities included. Other important influences of nutritional status, including diet, food security, and physical activity, may also have been different between these contexts, explaining conflicting results. This difference could, moreover, be discussed in the context of findings showing that women who married as children saw both health disadvantages and advantages of their early marriages (Chapter 5), in contrast with an ongoing study which found that child marriage is increasingly viewed as both a cultural necessity but social faux-pas (FXB Center for Health and Human Rights, 2018).
African and South Asian cultures, surely, possess several differences – for example, religion, agriculture, diet, natural resources, customs, and traditions. One could certainly try to hypothesize on the sociocultural distinctions within these different settings that may give rise to divergent experiences in and consequences of girl child marriage. In India, for example, class, caste, and patriarchal norms are recognized drivers of girl child marriage (Girls Not Brides,
2018b) and have been documented well beyond the previous century (Roy, 1888). Additionally, dowry, which is paid from the girl’s family to the man’s, may be lower if a girl is married at a younger age and transfers the financial responsibility of raising the girl from her family to her husband (Girls Not Brides, 2018b). Given heterogeneity across sub-Saharan Africa, the study in
Guinea provides one example of why this comparison proves difficult (Chapter 5). The author’s own experiences in both India and Guinea point to both overlaps and cultural distinctions. In
Guinea, drivers of girl child marriage include poverty and lack of education (Girls Not Brides,
2018a). Additionally, marriage is commonly viewed as a joining of families, serving a social purpose (Girls Not Brides, 2018a). Women interviewed (Chapter 5) more commonly shared that the bride’s family received gifts from the man’s family, which they referred to as dowry, to
159
formalize the marriage, a direct contrast to the transfer of resources from the bride-to-groom’s family in India.
However, one must be careful about overgeneralizing broad concepts of sociocultural norms. In fact, if one compares basic demographic data about Guinean and Indian populations, it might be surprising that girl child marriage is viewed more poorly in India, anecdotally, yet proportionally practiced more widely in Guinea. Though the absolute numbers of girl child brides in India is greater, the proportion of girls and women affected in Guinea is larger. Guinea also performs more poorly on economic and social indicators, which further intersect with the practice of girl child marriage (see Table 6.1). It is thus, difficult, to point toward specific country-level differences that explain why the health outcomes of girl child marriage explored in this dissertation may not entirely map onto previous work from other country contexts.
Table 6.1 Demographic characteristics for Guinea and India
Country Characteristics Guinea India Population 12,413,867 1,281,935,911 Age structure: 15-24 19.7% (male 1,236,092/female 17.9% (male 121,879,786/female years 1,212,936) 107,583,437) Median age (female) 19.1 years 28.6 years % marrying before age 18 51 27 % marrying before age 15 21 7 GDP PPP 2,000 7,200 % below poverty line 47 22 Predominant religion Islam (86%) Hindu (80%) Maternal mortality 679 deaths/100,000 live births 174 deaths/100,000 live births Infant mortality 50 deaths/1,000 live births 39 deaths/1,000 live births Life expectancy for 62.6 years 70.1 years women at birth Fertility rate 4.8 children born/woman 2.4 children born/woman Contraceptive prevalence 5.6% 53.5% Health expenditures 5.6% of GDP 4.7% of GDP
Note. Data taken from the Girls Not Brides website and the CIA World Factbook.
160
Implications for Norm Change
A natural “so what” question readers may wonder is: how can child marriage be stopped?
If it is illegal and harmful, as Chapters 1 and 2 lay out, and if it does have negative consequences for women and their children in African contexts, as Chapters 4 and 5, in part, address, what can be done to address it? This dissertation makes no attempt to explore these questions in-depth, as they are beyond the scope of the research objective and aims, yet to study a topic such as girl child marriage and not engage with such big picture ideas would neglect to situate this work more broadly.
Before providing some reflections on what this work can tell us for addressing the practice of girl child marriage, it is necessary to remind readers of a surprising contribution of this research. One of this dissertation’s most novel contributions comes from the qualitative work in Chapter 5, which revealed that in the context of Conakry, Guinea, women who married as children themselves identified not only disadvantages of their marriage on their health and their children’s health but also advantages. To state more plainly, several women themselves perceived positive health experiences in the context of their child marriage. Thus, perhaps it is no surprise that – despite legislation and regional frameworks outlawing marriage before age 18 – the practice persists.
This study exposes that there may be a disconnect between top-down policy approaches to drive changes in social norms and the diversity of community perspectives. Social policy in
Guinea has legally defined girl child marriage as unlawful and harmful. Yet the very fact that it still exists and that many women themselves could point to benefits of their own child marriages reveals that social policies are unlikely to succeed if they do not incorporate local level understanding. What, then, is a social norm and how can it in fact be influenced?
161
A social norm is defined as “a rule of behavior such that individuals prefer to conform to it on conditions that they believe that most people in their relevant network conform to it
(empirical expectations) and most people in their relevant network believe they ought to conform to it (normative expectations)” (Bicchieri, Lindemans, & Jiang, 2014, p. 11). Though everyone experiences preferences and pressures that shape these preferences, they are not universal and are influenced by perceived and experienced autonomy and individual agency. Bicchieri and colleagues (2014), moreover, suggest that a collective practice like child marriage does not necessarily have to be a social norm. It could instead be viewed as a rationale behavior, in response to self-interest such as a girl being too expensive for her family to care for and thus they marry her off at an earlier age; a custom which people continue to practice because generations before did it rather than linked to contemporary significance; a moral rule, driven by the perceptions of the parents or girl herself separately from how they believe society views the practice; or a convention/descriptive norm, whereby empirically, one observes that girls marry early in a community and the practice continues, yet if girls were to marry at later ages others were simply follow suit (Bicchieri et al., 2014). If child marriage is a social norm by this definition, it would require considering that individuals both think it should be followed and that they expect that others think it should be followed, complexifying the type of mechanisms that might be required to lead to change.
From this theoretical lens, this dissertation has generated some evidence in support of each of these pathways. The sheer numbers of girls and women married as children in the
African contexts investigated suggests that girl child marriage is a collective practice, and in the qualitative work, women articulated arguments of self-interest and moral rule, to a lesser degree, and convention and social norms, to a much greater degree. These differences in ways of making
162
sense of one’s preferences, perceived options, and individual beliefs about these options (which may be true or false) could certainly vary across contexts and across time points – explaining differences between Swaziland, a small and predominantly Christian southern African country, and Chad, a more populous central African country with a 52% Muslim majority, which had the highest and lowest median age of marriage (21 and 15, respectively) (Chapter 4), or explaining differences emerging in Africa-specific and India-specific results (Chapter 3).
Increasing evidence has started examination of what works in preventing child marriage, with little consensus on what the answer may be. One rigorous study found that conditional cash transfers increased schooling among adolescent girls in Malawi while only unconditional cash transfers reduced teen pregnancy and child marriage rates. Interestingly, those continuing with school did not experience substantially lower child marriage rates, thus there may have been an economic incentive for girls dropping out to marry, offset by the unconditional cash transfer
(Baird, McIntosh, & Özler, 2011). And a recent study in the different context of Bangladesh, found that providing an incentive of cooking oil, a product with cash value that avoided the risk of theft/improper transfer, reduced child marriage in both in- and out-of-school girls, while an empowerment program did not (Buchmann et al., 2017). Such studies emphasize a need to understand local perceptions of child marriage and the root cause of the practice, in order to consider appropriate incentives that may minimize the practice.
Methodological Contributions and Opportunities for Future Research
The results of this dissertation also make methodological contributions which can inform future research. This research generated evidence on health and well-being consequences of girl child marriage in sub-Saharan African, addressing gaps in the South Asia-dominated literature base. The selection of a mixed methods research framework enabled multiple methods to inform
163
our understanding of a broader phenomenon, increasing confidence in the nuanced nature of results. The two quantitative studies are among the first studies on the health consequences of girl child marriage to exploit nationally-representative data across several African countries, strengthening generalizability of study results. The investigation of potential mechanisms through which girl child marriage affects children’s health and well-being was also the first study of its kind, illuminating a need to think broadly about risk factors for early child development and encouraging child marriage research to move toward causal inference. The qualitative study used grounded theory to avoid a priori conclusions on women’s experiences with girl child marriage and through using multiple qualitative methods illustrates how women’s experiences can increase an understanding of the health effects of girl child marriage. Additionally, this dissertation placed emphasis on defining and deconstructing an understanding of the term “girl child marriage” (through Chapter 2 and qualitative work in Chapter 5), encouraging researchers to do the same. While situated in the fields of public health and global health, this dissertation intentionally took an interdisciplinary approach, drawing from several disciplines that extend beyond these fields.
Future research should continue to examine girl child marriage in sub-Saharan African contexts, where rates remain the highest. There is an opportunity to promote the use of both quantitative and qualitative research methods to strengthen how we understand complex issues such as girl child marriage. Future research must further explore the mechanisms through which girl child marriage impacts health and well-being, such as through longitudinal data and further examining contextual factors influencing women and children’s outcomes. Additionally, future research would benefit from an intentionally interdisciplinary and multisectoral approach, given the complexity of the topic, as revealed through this body of work.
164
Returning to Human Rights Framework
As a final thought, the author returns to the rights-based approach that motivated this research. Overwhelmingly across the literature, and as this dissertation introduced, there is a consistent framing of child marriage as leading to human rights violations (Chapter 2). As was previously explained, in a global context, the most comprehensive treaty on children’s rights –
The Convention on the Rights of the Child (CRC) – sets forward agreed-upon rights that could be violated by child marriage, which nearly every country in the world has signed onto, yet it does not explicitly articulate that child marriage is a violation of human rights. Instead, the global compromise calls against “traditional practices prejudicial to the health of children” (UN
General Assembly, 1990, Article 24.3), which subsequent comments help to further propose as child marriage before age 18, providing governments and vocal advocacy communities the ability to define what those are.
Uniquely, sub-Saharan Africa has a regional instrument designed to protect the rights of children; the African Charter on the Rights and Welfare of the Child (the African Charter) includes language that provides protection against “harmful social and cultural practices affecting the welfare, dignity, normal growth and development of the child” specifically prohibiting child marriage for girls and boys by raising minimum age to 18 years and requiring marriage registration (Organization of African Unity, 1990, Article 21). The result is that in
African countries, there are additional legal frameworks for preventing child marriage. Yet seven countries have yet to ratify the treaty, four countries include reservations (two of which are specific to not following the African Charter’s specifications on child marriage), and – as this dissertation revealed – the practice of girl child marriage remains a common phenomenon among
165
younger generations in sub-Saharan Africa, revealing that there is still much progress to be made.
Conclusion
The author proposes that what can be derived from this exploration is that it is possible for girl child marriage to have different consequences for different populations in different settings; in fact, as has been argued, context for young women’s health and developmental outcomes matters. The results of Chapter 3, for example, may have presented a different picture if South Asian countries had been included. The results of Chapter 5 may not have raised themes of health advantages if it had been conducted in another setting, even within other African contexts. Among public and global health research, further comparisons would be challenging until there is more parity in the research, and more variety of the types of research, from across these distinct contexts.
As this chapter summarized, and as this dissertation has shown, girl child marriage remains a concern now more than ever. Nearly 650 million girls and women remain affected today (UNICEF, 2018b). Progress in some contexts has been achieved, yet it must accelerate to achieve Sustainable Development Goal 5 Target 3, eliminating harmful practices such as child marriage by 2030. Additional attention must focus on sub-Saharan Africa, where the practice remains steady. Girl child marriage is, in fact, associated with consequences for women and their children’s health and well-being across sub-Saharan Africa. These consequences may include harmful health and well-being (as revealed in Chapter 4 on off-track early child development and stunted growth). Importantly, these consequences may also include possible benefits to health and well-being (as described by several women in Conakry, Guinea in Chapter 5). Girl child marriage and its consequences must be explored from multiple methods, as the mixed methods
166
approach used in this dissertation revealed a more nuanced story. Additionally, and perhaps most importantly, the context in which a woman develops and is situated within matters for understanding how an experience such as girl child marriage will affect her and generations to come.
167
Appendix A
Appendix Table A.1: Correlations Between Girl Child Marriage (Binary) and Other Covariates Covariates Coefficient p-value Completion of primary education -0.2496 p<0.001 (no, ref.) Current age (years) -0.0371 p<0.001 Age at first birth (years) -0.5432 p<0.001 Number of children ever born 0.2335 p<0.001 Completion of secondary -0.1786 p<0.001 education (no, ref.) Asset quintile -0.1058 p<0.001 Age gap between partner and 0.1386 p<0.001 woman (years) Education gap between partner 0.0224 p<0.001 and woman (levels)
168
Appendix B
711,964 women aged 15 to 49 Excluded: 140,787 women under age 20
571,177 women aged 20 to 49 Excluded: 61,909 women who were pregnant 795 women with missing pregnancy status 508,473 non-pregnant women aged 20 to 49
Excluded: 58,669 women who had never been married
449,804 ever-married, non- pregnant women aged 20 to 49 Excluded: 751 women who had missing data on BMI 449,053 ever-married, non- pregnant women aged 20 to 49 with BMI data Excluded: 150,836 women who belonged to clusters with no variation 298,217 ever-married, non- pregnant women aged 20 to 49 with BMI data in clusters with variation Excluded: 48,948 women with missing data on covariates in final model 249,269 ever-married, non- pregnant women aged 20 to 49 with no missing data in clusters with variation
Appendix Figure B.1: Study Flow Chart on How Sample Size was Determined
169
Appendix C
Appendix Figure C.1: Histogram Distribution of Age at Marriage Among Ever-Married Women Age 20 to 49 Included in Final Sample (N=249,269)
170
Appendix D
Appendix Figure D.1: Scatter Plot of Mean Age at Marriage and Proportion Underweight by Country of Women Age 20 to 49 Included in Final Sample, with Fitted Line (N=249,269)
171
Appendix E Appendix Table E.1: Associations Between Girl Child Marriage (Categorical Specification) and Underweight for Pooled Analysis (N=249,269)
Variables Model 1 Model 2 Model 3 Model 4 Model 5 Girl child marriage (18+ years, ref.) Below 14 years 0.945*** (0.908, 0.985) 0.974 (0.930, 1.020) 0.974 (0.930, 1.020) 0.972 (0.928, 1.018) 0.970 (0.926, 1.016) 14-15 years 0.921*** (0.892, 0.950) 0.941*** (0.908, 0.975) 0.939*** (0.906, 0.973) 0.937*** (0.904, 0.971) 0.935*** (0.903, 0.969) 16-17 years 0.914*** (0.887, 0.941) 0.923*** (0.894. 0.952) 0.918*** (0.890, 0.948) 0.917*** (0.889, 0.946) 0.916*** (0.887, 0.945) Completion of primary education (0.774, 0.830) (0.766, 0.822) (0.798, 0.857) (0.835, 0.898) (0.771, 0.834) (no, ref.) 0.802*** 0.793*** 0.827*** 0.866*** 0.802*** Current age (20-24 years, ref.) (0.900, 0.966) (0.902, 0.969) (0.907, 0.974) (0.903, 0.971) 25-29 years 0.932*** 0.934*** 0.940*** 0.936*** (0.899, 0.977) (0.909, 0.988) (0.901, 0.979) 30-34 years 0.937*** 0.940*** (0.902, 0.979) 0.948** 0.939*** (0.964, 1.058) (0.965, 1.059) (0.973 1.068) (0.958, 1.052) 35-39 years 1.010 1.011 1.019 1.004 (1.019, 1.129) (1.019, 1.129) (1.027, 1.137) (1.003, 1.112) 40-44 years 1.073*** 1.073*** 1.080*** 1.056** 172 (1.118, 1.250) (1.117, 1.248) (1.124, 1.257) (1.087, 1.218) 45-49 years 1.182*** 1.181*** 1.188*** 1.151*** (0.993, 1.001) (0.994, 1.002) (0.994, 1.002) age at first birth (years) 0.997 0.998 0.998 (0.994, 1.002) 0.998 (0.957, 0.969) (0.957, 0.969) (0.956, 0.969) (0.957, 0.970) number of children ever born 0.963*** 0.963*** 0.962*** 0.964*** Completion of secondary (0.650, 0.754) (0.698, 0.812) (0.659, 0.767) education (no, ref.) 0.700*** 0.753*** 0.711*** Asset quintile (poorest, ref.) (0.901, 0.966) (0.910, 0.975) Poorer 0.933*** 0.942*** (0.849, 0.910) Middle 0.866*** (0.837, 0.896) 0.879*** (0.744, 0.810) (0.761, 0.829) Richer 0.776*** 0.795*** (0.523, 0.585) (0.545, 0.610) Richest 0.553*** 0.577*** Age gap between partner and (0.996, 1.000) woman (years) 0.998** Education gap between partner (0.868, 0.911) and woman (levels) 0.890***
Note. Coefficients presented are odds ratios from logistic regression models with 95% CIs in parentheses, with clustered standard errors at sample cluster level. Underweight is defined as Body Mass Index less than 18.5. M1 adjusts for sampling cluster and woman's primary education (as potential confounders). M2 adjusts for all covariates in M1 + woman's age, age at first birth, and number of children born (as biological mechanisms). M3 adjusts for all covariates in M2 + mother's secondary education. M4 adjusts for all covariates in M3 + asset quintile. M5 includes remaining covariates. Bolded values are significant at the p<0.05 level. *** p<0.01, ** p<0.05
Appendix F
Country-specific Associations by Girl Child Marriage Category and Underweight
Study ID ES (95% CI)
Niger 1.13 (0.89, 1.42) Chad 1.09 (0.93, 1.27) Guinea 1.07 (0.82, 1.38) Mali 0.93 (0.80, 1.09) Cameroon 1.25 (0.90, 1.74) Ethiopia 1.06 (0.95, 1.17) Central African Republic 0.83 (0.46, 1.52) Nigeria 1.04 (0.93, 1.15) Burkina Faso 0.95 (0.74, 1.23) Sierra Leone 1.28 (0.94, 1.74) Senegal 0.75 (0.57, 1.00) Malawi 0.91 (0.73, 1.12) Uganda 0.75 (0.53, 1.06) Mozambique 1.12 (0.91, 1.38) Gambia 0.79 (0.52, 1.23) Zambia 1.04 (0.84, 1.30) Liberia 0.94 (0.61, 1.45) Cote d Ivoire 0.92 (0.45, 1.86) Congo Dem Rep 0.99 (0.74, 1.31) Tanzania 0.82 (0.64, 1.06) Sao Tome and Principe 0.62 (0.13, 3.01) Benin 0.93 (0.75, 1.14) Gabon 1.17 (0.69, 1.97) Madagascar 0.91 (0.73, 1.12) Togo 0.78 (0.50, 1.23) Comoros 0.33 (0.16, 0.67) Ghana 1.05 (0.73, 1.50) Kenya 0.90 (0.74, 1.10) Congo Rep 0.89 (0.62, 1.29) Zimbabwe 0.93 (0.61, 1.42) Lesotho 1.14 (0.35, 3.68) Burundi 1.97 (0.91, 4.27) Namibia 1.32 (0.72, 2.41) Swaziland 0.00 (0.00, 0.00) Rwanda 1.10 (0.56, 2.17) Overall (I-squared = 100.0%, p = 0.000) 0.96 (0.86, 1.06)
-4.27 1 4.27
Appendix Figure F.1: Country-specific associations between girl child marriage (<14 years) and underweight conditional on full set of mediators
Note. Presents odds ratios for marriage <14 years compared to 18+ years. Countries are listed in order of highest to lowest rates of girl child marriage. All models control for primary education, age, age at first birth, number of children ever born, secondary education, asset quintile, age gap, education gap, and EA fixed-effects (M5). Based on 35 independent country-specific models.
173
Study ID ES (95% CI)
Niger 1.02 (0.83, 1.24) Chad 0.98 (0.86, 1.12) Guinea 1.07 (0.87, 1.33) Mali 0.99 (0.89, 1.12) Cameroon 1.30 (0.97, 1.73) Ethiopia 1.05 (0.96, 1.15) Central African Republic 0.98 (0.67, 1.45) Nigeria 1.04 (0.94, 1.15) Burkina Faso 1.02 (0.88, 1.19) Sierra Leone 0.93 (0.72, 1.20) Senegal 0.87 (0.70, 1.08) Malawi 0.80 (0.69, 0.93) Uganda 1.00 (0.79, 1.27) Mozambique 0.89 (0.75, 1.06) Gambia 1.05 (0.76, 1.46) Zambia 0.84 (0.73, 0.96) Liberia 0.84 (0.65, 1.09) Cote d Ivoire 1.00 (0.61, 1.65) Congo Dem Rep 1.00 (0.82, 1.23) Tanzania 0.83 (0.70, 0.98) Sao Tome and Principe 0.61 (0.32, 1.18) Benin 0.90 (0.77, 1.04) Gabon 0.48 (0.30, 0.76) Madagascar 0.96 (0.83, 1.11) Togo 0.81 (0.60, 1.10) Comoros 0.69 (0.44, 1.10) Ghana 0.97 (0.76, 1.24) Kenya 0.85 (0.73, 0.99) Congo Rep 0.89 (0.69, 1.14) Zimbabwe 0.72 (0.55, 0.95) Lesotho 0.29 (0.13, 0.66) Burundi 1.19 (0.70, 2.04) Namibia 1.23 (0.82, 1.83) Swaziland 0.61 (0.12, 3.00) Rwanda 1.00 (0.72, 1.38) Overall (I-squared = 58.9%, p = 0.000) 0.91 (0.84, 0.97)
-3 1 3
Appendix Figure F.2: Country-specific associations between girl child marriage (14 to 15 years) and underweight conditional on full set of mediators
Note. Presents odds ratios for marriage at 14 to 15 years compared to 18+ years. Countries are listed in order of highest to lowest rates of girl child marriage. All models control for primary education, age, age at first birth, number of children ever born, secondary education, asset quintile, age gap, education gap, and EA fixed-effects (M5). Based on 35 independent country- specific models.
174
Study ID ES (95% CI)
Niger 1.12 (0.89, 1.41) Chad 0.99 (0.86, 1.14) Guinea 1.12 (0.90, 1.39) Mali 0.97 (0.86, 1.09) Cameroon 1.18 (0.88, 1.58) Ethiopia 1.06 (0.97, 1.16) Central African Republic 0.98 (0.67, 1.43) Nigeria 1.03 (0.93, 1.15) Burkina Faso 0.96 (0.86, 1.07) Sierra Leone 0.91 (0.68, 1.21) Senegal 0.91 (0.73, 1.13) Malawi 0.80 (0.70, 0.92) Uganda 0.88 (0.71, 1.08) Mozambique 0.87 (0.74, 1.03) Gambia 1.11 (0.80, 1.54) Zambia 0.82 (0.72, 0.92) Liberia 0.85 (0.66, 1.11) Cote d Ivoire 1.00 (0.62, 1.61) Congo Dem Rep 0.86 (0.73, 1.02) Tanzania 0.86 (0.75, 0.99) Sao Tome and Principe 1.01 (0.56, 1.84) Benin 0.91 (0.80, 1.04) Gabon 0.67 (0.42, 1.05) Madagascar 0.94 (0.83, 1.07) Togo 0.86 (0.67, 1.10) Comoros 0.83 (0.54, 1.30) Ghana 0.77 (0.63, 0.94) Kenya 0.82 (0.72, 0.93) Congo Rep 0.89 (0.72, 1.10) Zimbabwe 0.88 (0.73, 1.06) Lesotho 0.78 (0.52, 1.16) Burundi 1.12 (0.81, 1.54) Namibia 1.05 (0.75, 1.46) Swaziland 0.52 (0.11, 2.52) Rwanda 0.96 (0.78, 1.18) Overall (I-squared = 30.6%, p = 0.046) 0.92 (0.87, 0.98)
-2.52 1 2.52
Appendix Figure F.3: Country-specific associations between girl child marriage (16 to 17 years) and underweight conditional on full set of mediators
Note. Presents odds ratios for marriage at 16 to 17 years compared to 18+ years. Countries are listed in order of highest to lowest rates of girl child marriage. All models control for primary education, age, age at first birth, number of children ever born, secondary education, asset quintile, age gap, education gap, and EA fixed-effects (M5). Based on 35 independent country- specific models.
175
Appendix G Appendix Table G.1: Regression Results of Unadjusted Associations between Girl Child Marriage and Underweight by Country
Central African Congo , Dem. Variable Benin Burkina Faso Burundi Cameroon Republic Chad Comoros Rep. Congo, Rep.
Married before age 18 (18+ years, ref.) 1.010*** 1.010 1.010 1.045*** 0.994 1.031*** 0.993 0.988 0.991 (1.004 - (0.999 - (0.979 - (1.035 - (0.959 - (1.015 - (0.979 - (0.973 - (0.977 - 1.016) 1.020) 1.042) 1.055) 1.031) 1.047) 1.007) 1.002) 1.006) Observations 28,002 19,975 2,468 9,145 1,632 13,102 3,763 8,504 7,765
Variable Cote d'Ivoire Ethiopia Gabon Gambia Ghana Guinea Kenya Lesotho Liberia
Married before age 18 (18+ years, ref.) 1.004 1.027*** 0.992 1.016 1.003 1.022*** 1.022*** 0.991 0.993
176 (0.992 - (1.014 - (0.980 - (0.991 - (0.993 - (1.007 - (1.014 - (0.982 - (0.981 - 1.016) 1.039) 1.004) 1.042) 1.012) 1.037) 1.030) 1.000) 1.005) Observations 6,863 22,875 4,928 2,745 12,766 8,541 24,597 6,364 7,192
Sao Tome Variable Madagascar Malawi Mali Mozambique Namibia Niger Nigeria Rwanda and Principe
Married before age 18 (18+ years, ref.) 1.022*** 0.994 1.014*** 1.006 1.026*** 1.042*** 1.068*** 0.995 0.994 (1.007 - (0.987 - (1.006 - (0.999 - (1.007 - (1.026 - (1.062 - (0.985 - (0.973 - 1.036) 1.001) 1.022) 1.013) 1.046) 1.059) 1.073) 1.004) 1.016) Observations 12,658 23,015 24,896 19,053 6,764 12,322 46,800 15,812 1,595
Variable Senegal Sierra Leone Swaziland Tanzania Togo Uganda Zambia Zimbabwe Tanzania
Married before age 18 (18+ years, ref.) 1.008 1.013** 0.994 0.994 1.005 1.003 0.994 0.995 0.994 (0.994 - (1.000 - (0.983 - (0.986 - (0.991 - (0.991 - (0.987 - (0.988 - (0.986 - 1.024) 1.026) 1.004) 1.002) 1.020) 1.015) 1.001) 1.003) 1.002) Observations 8,693 7,303 2,193 19,657 6,115 10,244 24,633 16,073 19,657
Note. Coefficients presented are from unadjusted linear regression models with 95% CIs in parentheses. Underweight is defined as Body Mass Index less than 18.5. *** p<0.01, ** p<0.05
Appendix H
Appendix Table H.1: Child Marriage Among Mothers in Sample Compared to Population Rates
Married before age Married before age 18 Country 18 in sample (%) in population (%)a,b Central African Republic 64 68 Chad 75 68 Democratic Republic of Congo 45 39 Ghana 33 21 Kenya (Mombasa) 48 26 Kenya (Nyanza) 53 26 Madagascar (South) 75 41 Mauritania 52 43 Nigeria 40 43 Sierra Leone 59 44 Somalia (Northeast) 44 45 Somalia (Somaliland) 36 45 Swaziland 17 7 Togo 34 25 Malawi 51 50 Zimbabwe 34 31
a As reported by Girls Not Brides, 2017 b Data calculated among 20 to 24-year-old women
177
Appendix I
Appendix Table I.1: Summary of Unadjusted Bivariate Analyses to Assess Associations between Girl Child Marriage and Child Development and Health Outcomes
Outcome OR/Coefficient p-value Main outcomes ECDI (off-track) 0.797 p<0.001 Stunting 1.29 p<0.001 Secondary outcomes Child development score (out of 10) -0.271 p<0.001 HAZ -0.166 p<0.001 Developmental domain Learning 0.732 p<0.001 Literacy-numeracy 0.692 p<0.001 Physical 0.972 p=0.503 Socio-emotional 1.00 p=0.941
178
Appendix J
List of Supplementary Materials
Supplementary Materials Table 3.1: Relative risk of girl child marriage (binary) and underweight for pooled analysis (N=249,269) Supplementary Figure 3.1: Country-specific associations between girl child marriage (below 18 years) and severely underweight, conditional on full set of covariates Supplementary Figure 3.2: Country-specific associations between girl child marriage (below 18 years) and underweight for women aged 20 to 24, conditional on full set of covariates Supplementary Figure 3.3: Country-specific associations between girl child marriage (below 18 years) and underweight for 2011 to 2014 data, conditional on full set of covariates Supplementary Figure 3.4: Additional country-specific associations Supplementary Table 4.1: Survey response rate for women by country and survey year Supplementary Figure 4.2: Additional details regarding the calculation of the Early Childhood Development Index Supplementary Table 4.3: Bivariate correlations between girl child marriage and other covariates Supplementary Table 4.4: Associations between girl child marriage and child development score Supplementary Table 4.5: Associations between girl child marriage and height-for-age z score Supplementary Table 4.6: Associations between girl child marriage and child development health outcomes Supplementary Table 4.7: Sobel-Goodman Test for mediating influences
179
Bibliography
Abdullah, K., Malek, M. A., Faruque, A. S., Salam, M. A., & Ahmed, T. (2007). Health and
nutritional status of children of adolescent mothers: Experience from a diarrhoeal disease
hospital in Bangladesh. Acta Paediatrica, 96(3), 396-400.
Adair, T. (2008). HIV status and age at first marriage among women in Cameroon. J Biosoc Sci,
40(05), 743-760.
Adjaye-Gbewonyo, K., Kawachi, I., Subramanian, S., & Avendano, M. (2018). Income
inequality and cardiovascular disease risk factors in a highly unequal country: A fixed-
effects analysis from South Africa. International Journal for Equity in Health, 17(1), 31.
African Commission on Human and Peoples' Rights. (2018a). Ratification table: African charter
on the rights and welfare of the child. Retrieved from
http://www.achpr.org/instruments/child/ratification/
African Commission on Human and Peoples' Rights. (2018b). Ratification table: Protocol to the
african charter on human and peoples' rights on the rights of women in africa. Retrieved
from http://www.achpr.org/instruments/women-protocol/ratification/
African Union. (2003). Protocol to the African Charter on Human and Peoples’ Rights on the
Rights of Women in Africa. Retrieved from http://www.achpr.org/instruments/women-
protocol/
Agol, D., Bukenya, D., Seeley, J., Kabunga, E., & Katahoire, A. (2014). Marriage, intimacy and
risk of HIV infection in south west Uganda. Afr J Reprod Health, 18(4), 86-94.
Akseer, N., Al‐Gashm, S., Mehta, S., Mokdad, A., & Bhutta, Z. A. (2017). Global and regional
trends in the nutritional status of young people: A critical and neglected age group.
Annals of the New York Academy of Sciences, 1393(1), 3-20.
180
Alabi, E. M. (1990). Cultural practices in Nigeria. Newsl Inter Afr Comm Tradit Pract Affect
Health Women Child(9), 6-7.
Anderson, L. M., Shinn, C., Fullilove, M. T., Scrimshaw, S. C., Fielding, J. E., Normand, J., . . .
Services, T. F. o. C. P. (2003). The effectiveness of early childhood development
programs: A systematic review. American journal of preventive medicine, 24(3), 32-46.
Archambault, C. S. (2011). Ethnographic empathy and the social context of rights:“Rescuing”
Maasai girls from early marriage. American Anthropologist, 113(4), 632-643.
Aspen Planning and Evaluation Program, & Girls Not Brides. (2015). Measuring progress:
Recommended indicators for Girls Not Brides members working to address child
marriage Washington, D.C.: Aspen Institute and Girls Not Brides.
Atkinson, P., & Hammersley, M. (1994). Ethnography and participant observation. In N. K.
Denzin & Y. S. Lincoln (Eds.), Handbook of qualitative research (pp. 248-261).
Thousand Oaks, California: SAGE Publishers.
Baird, S., McIntosh, C., & Özler, B. (2011). Cash or condition? Evidence from a cash transfer
experiment. The Quarterly Journal of Economics, 126(4), 1709-1753.
Bearinger, L. H., Sieving, R. E., Ferguson, J., & Sharma, V. (2007). Global perspectives on the
sexual and reproductive health of adolescents: Patterns, prevention, and potential. The
Lancet, 369(9568), 1220-1231.
Belanger, A. (2017). Child marriage in the United States, explained. Teen Vogue. Retrieved from
https://www.teenvogue.com/story/child-marriage-in-the-united-states-explained
Belanger, A. (2018). A new study on child marriage is changing the conversation. Teen Vogue.
Retrieved from https://www.teenvogue.com/story/new-study-child-marriage-changing-
conversation
181
Bell, D. (1997). Defining marriage and legitimacy. Current Anthropology, 38(2), 237-253.
doi:doi:10.1086/204606
Berger, P. L., & Luckmann, T. (1966). The social construction of reality: A treatise in the
sociology of knowledge. Garden City, New York: Anchor Books, Doubleday &
Company, Inc.
Bhandarkar, R. G. (1893). History of child-marriage. Zeitschrift der Deutschen
Morgenländischen Gesellschaft, 47(1), 143-156.
Bicchieri, C., Lindemans, J. W., & Jiang, T. (2014). A structured approach to a diagnostic of
collective practices. Frontiers in Psychology, 5(Article 1418).
Black, M., Haeri, V., & Moodie, N. (2001). Early marriage: Child spouses (Vol. 7). Florence,
Italy: UNICEF Innocenti Resaerch Centre.
Black, R. E., Victora, C. G., Walker, S. P., Bhutta, Z. A., Christian, P., De Onis, M., . . .
Martorell, R. (2013). Maternal and child undernutrition and overweight in low-income
and middle-income countries. The Lancet, 382(9890), 427-451.
Bongaarts, J. (2007). Late marriage and the HIV epidemic in sub-Saharan Africa. Popul Stud
(Camb), 61(1), 73-83.
Bornstein, M. H., Britto, P. R., Nonoyama‐Tarumi, Y., Ota, Y., Petrovic, O., & Putnick, D. L.
(2012). Child development in developing countries: Introduction and methods. Child
Dev, 83(1), 16-31.
Bronfenbrenner, U. (1977). Toward an experimental ecology of human development. American
Psychologist, 32(7), 513-531. doi:10.1037/0003-066X.32.7.513
182
Bruce, J., & Clark, S. (2003). Including married adolescents in adolescent reproductive health
and HIV/AIDS policy. Paper presented at the WHO / UNFPA / Population Council
Technical Consultation on Married Adolescents, Geneva, Switzerland.
Bruce, J., & Rowbottom, S. (2004, 4 June). Child marriage, social isolation, and implications
for HIV policy. Paper presented at the Global Health Council 31st Annual Conference,
Washington, D.C.
Buchmann, N., Field, R., Glennerster, R., Nazneen, S., Pimkina, S., & Sen, I. (2017). Power vs
money: Alternative approaches to reducing child marriage in Bangladesh, a randomized
control trial. Boston, MA: Abdul Latif Jameel Poverty Action Lab (J-PAL).
Calkins, S. D., & Hill, A. (2007). Caregiver influences on emerging emotion regulation:
Biological and environmental transactions in early development. New York, New York:
Guilford Press.
Callaghan, J. E., Gambo, Y., & Fellin, L. C. (2015). Hearing the silences: Adult Nigerian
women’s accounts of ‘early marriages’. Feminism & Psychology, 25(4), 506-527.
Canada: Immigration and Refugee Board of Canada. (2015). Guinea: Forced marriages,
including prevalence; legislation affecting forced marriages; state protection; ability of
women to refuse a forced marriage (2012-2015). Retrieved from
http://www.refworld.org/docid/563c5f164.html.
Central Intelligence Agency. (2018). Guinea. The World Factbook. Retrieved from
https://www.cia.gov/library/publications/the-world-factbook/geos/gv.html
Chandra-Mouli, V., Greifinger, R., Nwosu, A., Hainsworth, G., Sundaram, L., Hadi, S., . . .
Braeken, D. (2013). Invest in adolescents and young people: It pays. Reprod Health,
10(1), 51. doi:10.1186/1742-4755-10-51
183
Chen, X.-K., Wen, S. W., Fleming, N., Demissie, K., Rhoads, G. G., & Walker, M. (2007).
Teenage pregnancy and adverse birth outcomes: A large population based retrospective
cohort study. Int J Epidemiol, 36(2), 368-373.
Chinwuba, N. N. (2016). Interaction of customs and colonial heritage their impact on marriage
and children in Nigeria. Anthropos, 111(1), 49-68.
Clark, C. J., Spencer, R. A., Khalaf, I. A., Gilbert, L., El-Bassel, N., Silverman, J. G., & Raj, A.
(2016). The influence of family violence and child marriage on unmet need for family
planning in Jordan. J Fam Plann Reprod Health Care, jfprhc-2014-101122.
Clark, C. J., Spencer, R. A., Khalaf, I. A., Gilbert, L., El-Bassel, N., Silverman, J. G., & Raj, A.
(2017). The influence of family violence and child marriage on unmet need for family
planning in Jordan. J Fam Plann Reprod Health Care, 43(2), 105-112.
doi:10.1136/jfprhc-2014-101122
Clark, S. (2004). Early marriage and HIV risks in sub-Saharan Africa. Stud Fam Plann, 35(3),
149-160.
Clifton, D. (2006). The urban-rural divide in health and development In Population Reference
Bureau (Ed.). Washington, D.C.: Population Reference Bureau.
Cohen, S. A. (2004). Delayed marriage and abstinence-until-marriage: On a collision course?
The Guttmacher Report on Public Policy, 7(2), 1-3.
Corsi, D. J., Subramanyam, M. A., & Subramanian, S. V. (2011). Commentary: Measuring
nutritional status of children. Int J Epidemiol, 40(4), 1030-1036. doi:10.1093/ije/dyr108
Creswell, J. W., & Plano Clark, V. L. (2007). Designing and conducting mixed methods
research. Thousand Oaks, California: SAGE Publications.
Cummings, P. (2009). Methods for estimating adjusted risk ratios. Stata Journal, 9(2), 175-196.
184
Danaei, G., Andrews, K. G., Sudfeld, C. R., Fink, G., McCoy, D. C., Peet, E., . . . Fawzi, W. W.
(2016). Risk factors for childhood stunting in 137 developing countries: A comparative
risk assessment analysis at global, regional, and country levels. PLoS medicine, 13(11),
e1002164. de Onis, M., & Blössner, M. (2003). The World Health Organization global database on child
growth and malnutrition: Methodology and applications. Int J Epidemiol, 32(4), 518-526. de Onis, M., Onyango, A. W., Borghi, E., Siyam, A., Nishida, C., & Siekmann, J. (2007).
Development of a WHO growth reference for school-aged children and adolescents.
Bulletin of the world Health Organization, 85, 660-667. de Smedt, J. (1998). Child Marriages in Rwandan Refugee Camps. Africa, 68(2), 211-228.
Denzin, N. K., & Lincoln, Y. S. (1994). Introduction: Entering the field of qualitative research In
N. K. Denzin & Y. S. Lincoln (Eds.), Handbook of qualitative research (pp. 1-18).
Thousand Oaks, California: SAGE Publications.
The DHS Program STATcompiler. (2018). Retrieved from: https://www.statcompiler.com/en/
Dopkins Broecker, J., & Hillard, P. (2009). Pregnancy in adolesence Global library of women's
medicine.
Doyle, A. M., Mavedzenge, S. N., Plummer, M. L., & Ross, D. A. (2012). The sexual behaviour
of adolescents in sub-Saharan Africa: Patterns and trends from national surveys. Trop
Med Int Health, 17(7), 796-807. doi:10.1111/j.1365-3156.2012.03005.x
Efevbera, Y. (2011). Child marriage in Malawi: A human rights and public health analysis.
(Master of Science thesis), Harvard School of Public Health, Boston, MA.
185
Efevbera, Y. (2012). Child marriage and HIV in Malawi: When human rights can, and should,
meet public health. Paper presented at the AIDS 2012, Washington, D.C. Poster retrieved
from https://library.iasociety.org/Abstract.aspx
Efevbera, Y. (2017a). Experiences of early and forced marriage in Conakry, Guinea: A
grounded theory study. Paper presented at the Society for Adolescent Health and
Medicine Annual Meeting: Cultivating Connections New Orleans, Lousiana.
Efevbera, Y. (2017b). Experiences of early and forced marriage in Conakry, Guinea: An
exploratory study. Journal of Adolescent Health, 60(2), S107.
Efevbera, Y. (2017c). ‘It is this which is normal’ A qualitative study on girl child marriage and
health in Guinea. Paper presented at the Department of Global Health and Population
Doctoral Seminar, Harvard TH Chan School of Public Health, Boston, MA.
Efevbera, Y. (2017d). Recognizing child marriage as a maternal and women's health issue.
Retrieved from Maternal and Child Health Taskforce website:
https://www.mhtf.org/2017/04/10/recognizing-child-marriage-as-a-maternal-and-
womens-health-issue/
Efevbera, Y. (2017e, Oct. 28). ‘They did not give me the chance to finish my studies’ The effects
of child marriage on education in Guinea Paper presented at the Black Doctoral
Network, Black Scholarship Matters: Intellectualism, Race, and the Public Sphere,
Atlanta, Georgia.
Efevbera, Y., Bhabha, J., Farmer, P., & Fink, G. (2017). Girl child marriage as a risk factor for
early childhood development and stunting. Social science & medicine, 185, 91-101.
doi:10.1016/j.socscimed.2017.05.027
186
Efevbera, Y., Bhabha, J., Farmer, P., & Fink, G. (2018a). Child marriage and underweight in
sub-Saharan Africa: A 35 country cross-national study. Journal of Adolescent Health,
62(2), S25-S26. doi:10.1016/j.jadohealth.2017.11.051
Efevbera, Y., Bhabha, J., Farmer, P., & Fink, G. (2018b). Girl child marriage and undernutrition:
Evidence from 35 sub-Saharan African countries. Manuscript submitted for publication.
Efevbera, Y., McCoy, D. C., Wuermli, A. J., & Betancourt, T. S. (2017). Early childhood
development plus violence prevention in low- and middle-income countries: A qualitative
study. Children & Society, 31(2), 98-109. doi:10.1111/chso.12169
Efevbera, Y., McCoy, D. C., Wuermli, A. J., & Betancourt, T. S. (2018). Integrating early child
development and violence prevention programs: A systematic review. New Directions for
Child and Adolescent Development, 2018(159), 27-54.
Elshibly, E. M., & Schmalisch, G. (2008). The effect of maternal anthropometric characteristics
and social factors on gestational age and birth weight in Sudanese newborn infants. BMC
Public Health, 8(1), 1.
Eng, P. M., Kawachi, I., Fitzmaurice, G., & Rimm, E. B. (2005). Effects of marital transitions on
changes in dietary and other health behaviours in US male health professionals. Journal
of Epidemiology & Community Health, 59(1), 56-62.
Engle, P. L., Fernald, L. C., Alderman, H., Behrman, J., O'Gara, C., Yousafzai, A., . . . Ertem, I.
(2011). Strategies for reducing inequalities and improving developmental outcomes for
young children in low-income and middle-income countries. The Lancet, 378(9799),
1339-1353.
Erulkar, A. (2013a). Adolescence lost: The realities of child marriage. J Adolesc Health, 52(5),
513-514. doi:10.1016/j.jadohealth.2013.03.004
187
Erulkar, A. (2013b). Early marriage, marital relations and intimate partner violence in Ethiopia.
Int Perspect Sex Reprod Health, 39(1), 6-13. doi:10.1363/3900613
Erulkar, A., & Muthengi, E. (2009). Evaluation of Berhane Hewan: A program to delay child
marriage in rural Ethiopia. Int Perspect Sex Reprod Health, 35(1), 6-14.
doi:10.1363/ifpp.35.006.09
Ettarh, R., Van de Vijver, S., Oti, S., & Kyobutungi, C. (2013). Overweight, obesity, and
perception of body image mong slum residents in Nairobi, Kenya, 2008–2009.
Preventing chronic disease, 10, 130198. doi:http://dx.doi.org/10.5888/pcd10.130198 .
Fall, C. H. D., Sachdev, H. S., Osmond, C., Restrepo-Mendez, M. C., Victora, C., Martorell, R., .
. . Cohorts Investigators. (2015). Association between maternal age at childbirth and child
and adult outcomes in the off spring: A prospective study in five low-income and middle-
income countries (COHORTS collaboration). Lancet Global Health, 3(7), E366-E377.
doi:10.1016/s2214-109x(15)00038-8
Farmer, P. (in press). Fever, feuds, and diamonds: Ebola and conflict in West Africa.
Farmer, P., Kim, J. Y., Kleinman, A., & Basilico, M. (2013). Introduction: A biosocial approach
to global health. In P. Farmer, J. Y. Kim, A. Kleinman, & M. Basilico (Eds.),
Reimagining global health: An introduction (pp. 1-14). Berkley and Los Angelos, CA:
The Regents of the University of California.
Ferry, B., Carael, M., Buve, A., Auvert, B., Laourou, M., Kanhonou, L., . . . Kaona, F. (2001).
Comparison of key parameters of sexual behaviour in four African urban populations
with different levels of HIV infection. Aids, 15 Suppl 4, S41-50.
Filmer, D., & Fox, L. (2014). Youth employment in sub-Saharan Africa. Africa Development
Series. Washington, D.C.: World Bank.
188
Forum on Marriage and the Rights of Women and Girls. (2000). Early marriage: Whose right to
choose? Retrieved from United Kingdom:
www.khubmarriage18.org/sites/default/files/228.pdf
FXB Center for Health and Human Rights. (2018). Gender and adolescent agency in India.
Retrieved from https://fxb.harvard.edu/gender-and-adolescent-agency-in-india/
Gage, A. J. (2012). Association of child marriage with suicidal thoughts and attempts among
adolescent girls in Ethiopia. J Adolesc Health, 52(5), 654-656.
doi:10.1016/j.jadohealth.2012.12.007
Gaur, K., Keshri, K., & Joe, W. (2013). Does living in slums or non-slums influence women's
nutritional status? Evidence from Indian mega-cities. Social science & medicine, 77, 137-
146.
Girls Not Brides. (2017). 20 countries with the highest rates of child marriage*. Retrieved from
http://www.girlsnotbrides.org/where-does-it-happen/
Girls Not Brides. (2018a). Child marriage around the world: Guinea. Retrieved from
https://www.girlsnotbrides.org/child-marriage/guinea/
Girls Not Brides. (2018b). Child marriage around the world: India. Retrieved from
https://www.girlsnotbrides.org/child-marriage/india
Glinski, A. M., Sexton, M., & Meyers, L. (2015). The child, early, and forced marriage resource
guide task order. Washington, D.C.: Banyan Global.
Godha, D., Hotchkiss, D. R., & Gage, A. J. (2013). Association between child marriage and
reproductive health outcomes and service utilization: a multi-country study from South
Asia. J Adolesc Health, 52(5), 552-558. doi:10.1016/j.jadohealth.2013.01.021
189
Goli, S., Rammohan, A., & Singh, D. (2015). The effect of early marriages and early
childbearing on women’s nutritional status in India. Matern Child Health J, 19(8), 1864-
1880.
Grantham-McGregor, S., Cheung, Y. B., Cueto, S., Glewwe, P., Richter, L., Strupp, B., & the
International Child Development Steering Group. (2007). Developmental potential in the
first 5 years for children in developing countries. Lancet, 369(9555), 60-70.
doi:10.1016/S0140-6736(07)60032-4
Greenland, S., & Morgenstern, H. (1989). Ecological bias, confounding, and effect modification.
Int J Epidemiol, 18(1), 269-274.
Haberland, N., Chong, E., & Bracken, H. (2003). Married adolescents: An overview.
WHO/UNFPA/Population Council Technical Consultation on Married Adolescents, 9-12
Dec 2003.
Hampton, T. (2010). Child marriage threatens girls' health. Jama, 304(5), 509-510.
doi:10.1001/jama.2010.1009
Hanna, B., & Kleinman, A. (2013). Unpacking global health: Theory and critique. In P. Farmer,
A. Kleinman, J. Y. Kim, & M. Basilico (Eds.), Reimagining global health: An
introduction (pp. 15-32). Berkley and Los Angeles, CA: The Regents of the University of
California.
Hediger, M. L., Scholl, T. O., & Schall, J. I. (1997). Implications of the Camden Study of
adolescent pregnancy: Interactions among maternal growth, nutritional status, and body
composition. Annals of the New York Academy of Sciences, 817(1), 281-291.
Horton, R., & Lo, S. (2013). Nutrition: A quintessential sustainable development goal. Lancet
Series Maternal Child Nutr, 382(9890), 371-372.
190
Hubbard, J. (2008). Manual on brief ethnographic interviewing: Understanding an issue,
problem or idea from a local perspective. The Center for Victims of Torture.
Minneapolis, Minnesota.
Huber, P. J. (1967). The behavior of maximum likelihood estimates under nonstandard
conditions. Paper presented at the Proceedings of the fifth Berkeley symposium on
mathematical statistics and probability, Berkeley, California.
Human Rights Watch. (2011). "How come you allow little girls to get married?" Child marriage
in Yemen. Retrieved from New York, New York:
https://www.hrw.org/sites/default/files/reports/yemen1211ForUpload_0.pdf
ICF International. (2012). MEASURE DHS Biomarker field manual. Calverton, Maryland: ICF
International.
Ignelzi, M. (2000). Meaning-making in the learning and teaching process. New Directions for
Teaching and Learning, 2000(82), 5-14. doi:10.1002/tl.8201
Jain, S., & Kurz, K. (2007). New insights on preventing child marriage: A global analysis of
factors and programs. New Delhi, India: International Center for Research on Women.
Jeong, J., McCoy, D. C., Yousafzai, A. K., Salhi, C., & Fink, G. (2016). Paternal stimulation and
early child development in low- and middle-income countries. Pediatrics, 138(4),
e20161357. doi:10.1542/peds.2016-1357
Kamal, S. M. (2012). Decline in child marriage and changes in its effect on reproductive
outcomes in Bangladesh. J Health Popul Nutr, 30(3), 317-330.
Kamal, S. M., & Hassan, C. H. (2015). Child marriage and its association with adverse
reproductive outcomes for women in Bangladesh. Asia Pac J Public Health.
doi:10.1177/1010539513503868
191
Krieger, N. (2003). Genders, sexes, and health: what are the connections—and why does it
matter? Int J Epidemiol, 32(4), 652-657. doi:10.1093/ije/dyg156
Kurth, F., Bélard, S., Mombo-Ngoma, G., Schuster, K., Adegnika, A. A., Bouyou-Akotet, M. K.,
. . . Ramharter, M. (2010). Adolescence as risk factor for adverse pregnancy outcome in
Central Africa–a cross-sectional study. PLoS One, 5(12), e14367.
Le Strat, Y., Dubertret, C., & Le Foll, B. (2011). Child marriage in the United States and its
association with mental health in women. Pediatrics, 128(3), 524-530.
doi:10.1542/peds.2011-0961
Lee, A. C., Mullany, L. C., Tielsch, J. M., Katz, J., Khatry, S. K., LeClerq, S. C., . . . Darmstadt,
G. L. (2008). Risk factors for neonatal mortality due to birth asphyxia in southern Nepal:
A prospective, community-based cohort study. Pediatrics, 121(5), e1381-e1390.
Legrand, T. K., & Mbacke, C. S. M. (1993). Teenage preganancy and child health in the Urban
Sahel. Studies in Family Planning, 24(3), 137-149.
Lendhardt, A., Wise, L., Rosa, G., Warren, H., Mason, F., & Sarumi, R. (2016). Every last girl.
London, England: Save the Children, Retrieved from
http://www.savethechildren.org/atf/cf/%7B9def2ebe-10ae-432c-9bd0-
df91d2eba74a%7D/EVERY%20LAST%20GIRL%20REPORT%20FINAL.PDF
Letamo, G., & Navaneetham, K. (2014). Prevalence and determinants of adult under-nutrition in
Botswana. PLoS One, 9(7), e102675.
Lloyd, C. B., & Mensch, B. S. (2008). Marriage and childbirth as factors in dropping out from
school: An analysis of DHS data from sub-Saharan Africa. Popul Stud (Camb), 62(1), 1-
13. doi:10.1080/00324720701810840
192
Loaiza, S. E., & Wong, S. (2012). Marrying too young: End child marriage. New York, NY:
United Nations Population Fund.
Machel, G., Pires, E., & Carlsson, G. (2013). The world we want: An end to child marriage.
Lancet, 382(9897), 1005-1006. doi:10.1016/s0140-6736(13)61944-3
Macleod, C. (2003). Teenage pregnancy and the construction of adolescence: Scientific literature
in South Africa. Childhood, 10(4), 419-437.
Mardi, A., Ebadi, A., Shahbazi, S., Esmaelzade Saeieh, S., & Behboodi Moghadam, Z. (2018).
Factors influencing the use of contraceptives through the lens of teenage women: A
qualitative study in Iran. BMC Public Health, 18(1), 202. doi:10.1186/s12889-018-5116-
3
Mashal, T., Takano, T., Nakamura, K., Kizuki, M., Hemat, S., Watanabe, M., & Seino, K.
(2008). Factors associated with the health and nutritional status of children under 5 years
of age in Afghanistan: Family behaviour related to women and past experience of war-
related hardships. BMC Public Health, 8, 301-313.
Maswikwa, B., Richter, L., Kaufman, J., & Nandi, A. (2015). Minimum marriage age laws and
the prevalence of child marriage and adolescent birth: Evidence from sub-Saharan Africa.
Int Perspect Sex Reprod Health, 41(2), 58-68.
Mathur, S., Greene, M., & Malhotra, A. (2003). Too young to wed: The lives, rights and health
of young married girls. Washington, D.C.: International Center for Research on Women.
Mati, J. K., Mbugua, S., & Ndavi, M. (1984). Control of cancer of the cervix: Feasibility of
screening for premalignant lesions in an African environment. IARC Sci Publ(63), 451-
463.
193
Maxwell, J. A. (2012). Qualitative research design: An interactive approach: An interactive
approach. Thousand Oaks, California: SAGE Publications.
Maxwell, J. A., & Loomis, D. M. (2003). Mixed methods design: An alternative approach.
Handbook of mixed methods in social and behavioral research, 1, 241-272.
McClendon, K. A., McDougal, L., Ayyaluru, S., Belayneh, Y., Sinha, A., Silverman, J. G., &
Raj, A. (2017). Intersections of girl child marriage and family planning beliefs and use:
qualitative findings from Ethiopia and India. Cult Health Sex, 1-16.
doi:10.1080/13691058.2017.1383513
McCoy, D. C., Peet, E. D., Ezzati, M., Danae, G., Black, M., Sudfeld, C. R., . . . Fink, G. (2016).
Early childhood developmental status in low- and middle-income countries: National,
regional and global prevalence estimates using predictive modeling. PLOS Med, 13(6),
e1002034. doi:https://doi.org/10.1371/journal.pmed.1002034
MeasureDHS/ICF International. (2012). Sampling and household listing manual. Retrieved from
Calverton, Maryland
http://dhsprogram.com/pubs/pdf/DHSM4/DHS6_Sampling_Manual_Sept2012_DHSM4.
MeasureDHS/ICF International. (2013). Description of the Demographic and Health Surveys
individual recode data file. Washington, D.C.: USAID.
Miller, A. C., Murray, M. B., Thomson, D. R., & Arbour, M. C. (2015). How consistent are
associations between stunting and child development? Evidence from a meta-analysis of
associations between stunting and multidimensional child development in fifteen low-
and middle-income countries. Public Health Nutr, 1-9. doi:10.1017/s136898001500227x
194
Misch, E. S., & Yount, K. M. (2014). Intimate partner violence and breastfeeding in Africa.
Matern Child Health J, 18(3), 688-697.
Mourtada, R., Schlecht, J., & DeJong, J. (2017). A qualitative study exploring child marriage
practices among Syrian conflict-affected populations in Lebanon. Conflict and Health,
11(1), 27. doi:10.1186/s13031-017-0131-z
Multiple Indicator Clusters Survey Program. (2014). Instructions for interviewers. New York,
New York: UNICEF.
Myers, J. (2013). Untying the knot: Exploring early marriage in fragile states In H. S. Doornbos
(Ed.), Research Report. Milton Keynes, England: World Vision UK.
Nasrullah, M., Muazzam, S., Bhutta, Z. A., & Raj, A. (2013). Girl child marriage and its effect
on fertility in Pakistan: Findings from Pakistan Demographic and Health Survey, 2006-
2007. Matern Child Health J. doi:10.1007/s10995-013-1269-y
Nasrullah, M., Zakar, R., Zakar, M. Z., Abbas, S., Safdar, R., Shaukat, M., & Krämer, A. (2014).
Knowledge and attitude towards child marriage practice among women married as
children-a qualitative study in urban slums of Lahore, Pakistan. BMC Public Health,
14(1), 1-13. doi:10.1186/1471-2458-14-1148
Nguyen, M. C., & Wodon, Q. (2012). Measuring child marriage. Economics Bulletin, 32(1), 398-
411.
Nour, N. M. (2006). Health consequences of child marriage in Africa. Emerg Infect Dis, 12(11),
1644-1649. doi:10.3201/eid1211.060510
Nour, N. M. (2009). Child marriage: A silent health and human rights issue. Rev Obstet Gynecol,
2(1), 51-56.
Okwudili, O. (1993). Protecting the child. Niger Pop, 15-16.
195
Organization of African Unity. (1990). African Charter on the Rights and Welfare of the Child.
(CAB/LEG/24.9/49). Retrieved from http://www.refworld.org/docid/3ae6b38c18.html.
Rah, J. H., Christian, P., Shamim, A. A., Arju, U. T., Labrique, A. B., & Rashid, M. (2008).
Pregnancy and lactation hinder growth and nutritional status of adolescent girls in rural
Bangladesh. The Journal of nutrition, 138(8), 1505-1511.
Rahman, M. L., Kile, M. L., Rodrigues, E. G., Valeri, L., Raj, A., Mazumdar, M., . . . Christiani,
D. C. (2018). Prenatal arsenic exposure, child marriage, and pregnancy weight gain:
Associations with preterm birth in Bangladesh. Environment International, 112, 23-32.
doi:https://doi.org/10.1016/j.envint.2017.12.004
Raj, A. (2010). When the mother is a child: the impact of child marriage on the health and
human rights of girls. Archives of disease in childhood, archdischild178707.
Raj, A. (2017). Validity of self-reported age at marriage in rural India. Int J Gynaecol Obstet,
137(3), 343-344. doi:10.1002/ijgo.12131
Raj, A., & Boehmer, U. (2013). Girl child marriage and its association with national rates of
HIV, maternal health, and infant mortality across 97 countries. Violence Against Women,
19(4), 536-551. doi:10.1177/1077801213487747
Raj, A., Gomez, C. S., & Silverman, J. G. (2014). Multisectorial Afghan perspectives on girl
child marriage: Foundations for change do exist in Afghanistan. Violence Against
Women, 20(12), 1489-1505. doi:10.1177/1077801211403288
Raj, A., McDougal, L., & Rusch, M. L. (2012). Changes in prevalence of girl child marriage in
South Asia. Jama, 307(19), 2027-2029. doi:10.1001/jama.2012.3497
196
Raj, A., McDougal, L., & Rusch, M. L. (2014). Effects of young maternal age and short
interpregnancy interval on infant mortality in South Asia. Int J Gynaecol Obstet, 124(1),
86.
Raj, A., Saggurti, N., Balaiah, D., & Silverman, J. G. (2009). Prevalence of child marriage and
its effect on fertility and fertility-control outcomes of young women in India: A cross-
sectional, observational study. The Lancet, 373(9678), 1883-1889.
Raj, A., Saggurti, N., Lawrence, D., Balaiah, D., & Silverman, J. G. (2010). Association between
adolescent marriage and marital violence among young adult women in India.
International Journal of Gynecology & Obstetrics, 110(1), 35-39.
Raj, A., Saggurti, N., Winter, M., Labonte, A., Decker, M. R., Balaiah, D., & Silverman, J. G.
(2010). The effect of maternal child marriage on morbidity and mortality of children
under 5 in India: Cross sectional study of a nationally representative sample. BMJ, 340,
b4258.
Reiss, F. (2017). Despite progress, child marraige is still legal in all 50 states. The Opinion
Pages: On the Ground, Nicholas Kristoff. Retrieved from
https://kristof.blogs.nytimes.com/2017/07/26/despite-progress-child-marriage-is-still-
legal-in-all-50-states/
Roy, R. C. (1888). Child marriage in India. The North American Review, 147(383), 415-423.
Rutstein, S. O., & Johnson, K. (2004). The DHS wealth index. Retrieved from Calverton,
Maryland:
http://mics.unicef.org/files?job=W1siZiIsIjIwMTUvMDMvMzEvMTAvMzYvMDAvNz
Q0L0RIU19XZWFsdGhfSW5kZXhfX0RIU19Db21wYXJhdGl2ZV9SZXBvcnRzXy5w
ZGYiXV0&sha=4a283701ee092fbc
197
Sabbe, A., Oulami, H., Zekraoui, W., Hikmat, H., Temmerman, M., & Leye, E. (2013).
Determinants of child and forced marriage in Morocco: Stakeholder perspectives on
health, policies and human rights. BMC Int Health Hum Rights, 13(1), 43.
doi:10.1186/1472-698x-13-43
Sameroff, A. J. (2009). The transactional model. In A. Sameroff (Ed.), The transactional model
of development: How children and contexts shape each other (pp. 3-21). Washington,
D.C.: American Psychological Association.
Santhya, K. (2011). Early marriage and sexual and reproductive health vulnerabilities of young
women: A synthesis of recent evidence from developing countries. Current opinion in
obstetrics and gynecology, 23(5), 334-339.
Santhya, K., Ram, U., Acharya, R., Jejeebhoy, S. J., Ram, F., & Singh, A. (2010). Associations
between early marriage and young women's marital and reproductive health outcomes:
evidence from India. Int Perspect Sex Reprod Health, 132-139.
Sarich, J., Olivier, M., & Bales, K. (2016). Forced marriage, slavery, and plural legal systems:
An African example. Human Rights Quarterly, 38(2), 450-476.
Sawant, L. D., & Venkat, S. (2013). Comparative analysis of normal versus fetal growth
restriction in pregnancy: The significance of maternal body mass index, nutritional status,
anemia, and ultrasonography screening. International journal of reproductive medicine,
2013.
Schapiro, T. (1999). What Is a Child? Ethics, 109(4), 715-738.
Schlecht, J., Rowley, E., & Babirye, J. (2013). Early relationships and marriage in conflict and
post-conflict settings: Vulnerability of youth in Uganda. Reprod Health Matters, 21(41),
234-242. doi:10.1016/s0968-8080(13)41710-x
198
Scholl, T. O., & Hediger, M. L. (1993). A review of the epidemiology of nutrition and adolescent
pregnancy: Maternal growth during pregnancy and its effect on the fetus. Journal of the
American College of Nutrition, 12(2), 101-107.
Schwelb, E. (1963). Marriage and human rights. The American Journal of Comparative Law,
12(3), 337-383.
Sharma, A., & Gupta, S. (1991). Child's right to special care. ICCW News Bull, 39(3-4), 12-15.
Sharma, V., Katz, J., Mullany, L. C., Khatry, S. K., LeClerq, S. C., Shrestha, S. R., . . . Tielsch,
J. M. (2008). Young maternal age and the risk of neonatal mortality in rural Nepal.
Archives of Pediatrics & Adolescent Medicine, 162(9), 828-835.
Sinclair, S. (2016) A photographer gives cameras to child brides. Their images are
amazing./Interviewer: K. Adair. Goats and Soda: Stories of Life in a Changing World,
National Public Radio, Washington, D.C.
Smith, A. D. (2009). Girls being force-fed for marriage as fattening farms revived. The Observer.
Retrieved from https://www.theguardian.com/world/2009/mar/01/mauritania-force-
feeding-marriage
Sobal, J., Rauschenbach, B., & Frongillo, E. A. (2003). Marital status changes and body weight
changes: a US longitudinal analysis. Social science & medicine, 56(7), 1543-1555.
Sobkoviak, R. M., Yount, K. M., & Halim, N. (2012). Domestic violence and child nutrition in
Liberia. Social science & medicine, 74(2), 103-111.
Strauss, A., & Corbin, J. (1994). Grounded theory methodology: An overview. In N. K. Denzin
& Y. S. Lincoln (Eds.), Handbook of qualitative research (pp. 273-285). Thousand Oaks,
California: SAGE Publications.
199
Strochlic, N. (2014). The sad hidden plight of child grooms. The Daily Beast. Retrieved from
http://www.thedailybeast.com/articles/2014/09/18/the-sad-hidden-plight-of-child-
grooms.html
Svanemyr, J., Chandra-Mouli, V., Christiansen, C. S., & Mbizvo, M. (2012). Preventing child
marriages: First international day of the girl child "my life, my right, end child marriage".
Reprod Health, 9, 31. doi:10.1186/1742-4755-9-31
Svanemyr, J., Chandra-Mouli, V., Raj, A., Travers, E., & Sundaram, L. (2015). Research
priorities on ending child marriage and supporting married girls. Reprod Health, 12(1),
80. doi:10.1186/s12978-015-0060-5
Svedberg, P. (1990). Undernutrition in sub-Saharan Africa: Is there a gender bias? Journal of
Development Studies, 26(3), 469.
Teddlie, C., & Tashakkori, A. (2003). Major issues and controveries in the use of mixed methods
in the social and behvioral sciences. In A. Tashakkori & C. Teddlie (Eds.), Handbook of
mixed methods in social & behavioral research (pp. 3-50). Thousand Oaks, California:
SAGE Publications.
Teller, C. H., & Alva, S. (2008). Reducing child malnutrition in sub-Saharan Africa: Surveys
find mixed progress. Population Reference Bureau. Retrieved from
http://www.prb.org/Publications/Articles/2008/stuntingssa.aspx
The law of marriage. (2018). London, England: U.K. Parliament. Retrieved from:
https://www.parliament.uk/about/living-heritage/transformingsociety/private-
lives/relationships/overview/lawofmarriage-/
200
The Urban Child Institute. (2014). How adolescent parenting affects children, families, and
communities. Retrieved from http://www.urbanchildinstitute.org/articles/editorials/how-
adolescent-parenting-affects-children-families-and-communities
Tuoyire, D. A., Kumi-Kyereme, A., Doku, D. T., & Amo-Adjei, J. (2017). Perceived ideal body
size of Ghanaian women: “Not too skinny, but not too fat”. Women & Health, 1-15.
doi:10.1080/03630242.2017.1321607
UN General Assembly 70th session. (2015). 70/1. Transforming our world: the 2030 agenda for
sustainable development Geneva, Switzerland Retrieved from
http://www.un.org/ga/search/view_doc.asp?symbol=A/RES/70/1&Lang=E.
UN General Assembly 71st session. (2017). 71/175. Child, early and forced marriage. Geneva,
Switzerland Retrieved from
http://www.un.org/en/ga/search/view_doc.asp?symbol=A/RES/71/175&referer=http://w
ww.un.org/en/ga/71/resolutions.shtml&Lang=E.
UN General Assembly. (1948). Universal Declaration of Human Rights. Retrieved from
https://www.ohchr.org/EN/UDHR/Documents/UDHR_Translations/eng.pdfs.
UN General Assembly. (1962). Convention on Consent to Marriage, Minimum Age for
Marriage, and Registration of Marriages. Retrieved from
https://www.ohchr.org/en/professionalinterest/pages/minimumageformarriage.aspx.
UN General Assembly. (1965). Recommendation on Consent to Marriage, Minimum Age for
Marriage, and Registration of Marriages. Retrieved from
https://www.ohchr.org/EN/ProfessionalInterest/Pages/RecommendationOnConsentToMa
rriage.aspx.
201
UN General Assembly. (1966a). International Covenant on Civil and Political Rights. Retrieved
from https://www.ohchr.org/en/professionalinterest/pages/ccpr.aspx.
UN General Assembly. (1966b). International Covenant on Economic, Social and Cultural
Rights. Retrieved from https://www.ohchr.org/en/professionalinterest/pages/cescr.aspx.
UN General Assembly. (1979). Convention on the Elimination of All Forms of Discrimination
Against Women. Retrieved from
http://www.un.org/womenwatch/daw/cedaw/text/econvention.htm.
UN General Assembly. (1990). Convention on the Rights of the Child. Retrieved from
https://www.ohchr.org/en/professionalinterest/pages/crc.aspx.
UN General Assembly Human Rights Council 24th session. (2013). 24/… Strengthening efforts
to prevent and eliminate child, early and forced marriage: challenges, achievements, best
practices and implementation gaps. Geneva, Switzerland.
UN World Health Organization Interim Commission. (1948). Official records of the World
Health Organization no. 2: Summary report on proceedings, minutes, and final acts of
the International Health Conference. New York, New York & Geneva, Switzerland.
Retrieved from
http://apps.who.int/iris/bitstream/handle/10665/85573/official_record2_eng.pdf;jsessioni
d=ECC50D71CDC91E0A44765A492C180098?sequence=1.
UNFPA. (2012). Marrying too young: Ending child marriage. Retrieved from New York, New
York: http://www.unfpa.org/end-child-marriage
UNFPA. (2016). State of the world population 2016. Retrieved from New York, New York:
http://www.unfpa.org/sites/default/files/sowp/downloads/The_State_of_World_Populatio
n_2016_-_English.pdf
202
UNICEF. (2005). Early marriage: A harmful traditional practice. New York, New York:
UNICEF.
UNICEF. (2011). State of the world's children 2011: Adolescence: An age of opportunity.
Retrieved from New York, New York: https://www.unicef.org/sowc2011/pdfs/SOWC-
2011-Main-Report_EN_02092011.pdf
UNICEF. (2013a, July 16). ECD monitoring. Early Childhood. Retrieved from
http://www.unicef.org/earlychildhood/index_69846.html
UNICEF. (2013b). The formative years: UNICEF’s work on measuring early childhood
development. Retrieved from https://data.unicef.org/wp-
content/uploads/2015/12/Measuring-ECD-Brochure-HR-10_8_116.pdf
UNICEF. (2014). Ending child marriage: Progress and prospects. New York, New York:
UNICEF.
UNICEF. (2015a). About MICS. Multiple Indicator Cluster Surveys. Retrieved from
http://mics.unicef.org/about
UNICEF. (2015b). A profile of child marriage in Africa. New York, New York: UNICEF.
UNICEF. (2016). Multiple Indicator Clusters Survey. Retrieved from:
http://mics.unicef.org/surveys
UNICEF. (2017). State of the world's children 2017: Children in a digital age. New York, New
York: United Nations Children Fund.
UNICEF. (2018a). Child marriage. Retrieved from https://data.unicef.org/topic/child-
protection/child-marriage/
203
UNICEF. (2018b). Child marriage: Latest trends and future prosepcts. Retrieved from New
York, NY: https://data.unicef.org/wp-content/uploads/2018/07/Child-Marriage-Data-
Brief.pdf
UNICEF. (2018c). Percentage of women aged 20 to 24 years who were first married or in union
before ages 15 and 18. Retrieved from: https://data.unicef.org/wp-
content/uploads/2015/12/Child-marriage-database_Mar-2018.xlsx
United Nations Department of Economic and Social Affairs - Population Division. (2017). World
marriage data 2017. Retrieved from:
http://www.un.org/en/development/desa/population/theme/marriage-
unions/WMD2017.shtml
Ussher, J. M., Metusela, C., Hawkey, A. J., & Perz, J. M. (2017). Sexual and reproductive health
of migrant and refugee women: Research report and recommendations for healthcare
providers and community workers. Retrieved from Western Sydney University:
http://researchdirect.westernsydney.edu.au/islandora/object/uws%3A40355/datastream/P
DF/view
Walker, J. A. (2012). Early marriage in Africa--trends, harmful effects and interventions. Afr J
Reprod Health, 16(2), 231-240.
Walker, J. A. (2013). Mapping early marriage in West Africa: A scan of trends, interventions,
what works, best practices, and the way forward. Retrieved from Lagos, Nigeria:
http://www.girlsnotbrides.org/wp-content/uploads/2014/01/Ford-Foundation-CM-West-
Africa-2013_09.pdf
White, M., & Epston, D. (1990 ). Narrative means to therapeutic ends. New York, New York:
W. W. Norton.
204
WHO Expert Committee on Physical Status. (1995). Physical status: The use and interpretation
of anthropometry. Retrieved from Geneva, Swizerland:
http://apps.who.int/iris/bitstream/10665/37003/1/WHO_TRS_854.pdf
Williamson, N. (2013). Motherhood in childhood: Facing the challenge of adolescent
pregnancy. The State of World Population 2013. Retrieved from New York, New York:
https://www.unfpa.org/sites/default/files/pub-pdf/EN-SWOP2013-final.pdf
Winkvist, A., Jalil, F., Habicht, J.-P., & Rasmussen, K. M. (1994). Maternal energy depletion is
buffered among malnourished women in Punjab, Pakistan. The Journal of nutrition,
124(12), 2376.
Wodon, Q., Nguyen, M. C., & Tsimpo, C. (2016). Child marriage, education, and agency in
Uganda. Feminist Economics, 22(1), 54-79.
Wolfe, L. (2013). Bartered brides. Nation, 296(26/27), 18-20.
Women Business and the Law. (2017). Where girls can get married younger than boys.
Washington, D.C.: The World Bank Group.
Wondie, Y., Zemene, W., Reschke, K., & Schroder, H. (2011). Early marriage, rape, child
prostitution, and related factors determining the psychosocial effects severity of child
sexual abuse in Ethiopia. J Child Sex Abus, 20(3), 305-321.
doi:10.1080/10538712.2011.573458
World Health Organization. (2018). Adolescent health. Retrieved from
http://www.who.int/topics/adolescent_health/en/
World Policy Center. (2016). What is the minimum age of marriage for girls? Retrieved from
http://worldpolicycenter.org/policies/what-is-the-minimum-age-of-marriage-for-girls
205
Yount, K. M., & Abraham, B. K. (2007). Female genital cutting and HIV/AIDS in Kenyan
women. Studies in Family Planning, 38(2), 73-88.
Yount, K. M., DiGirolamo, A. M., & Ramakrishnan, U. (2011). Impacts of domestic violence on
child growth and nutrition: A conceptual review of the pathways of influence. Social
science & medicine, 72(9), 1534-1554.
Yu, S. H., Mason, J., Crum, J., Cappa, C., & Hotchkiss, D. R. (2016). Differential effects of
young maternal age on child growth. Global Health Action, 9(1), 31171.
Zhang, Z. F., Parkin, D. M., Yu, S. Z., Esteve, J., & Yang, X. Z. (1989). Risk factors for cancer
of the cervix in a rural Chinese population. Int J Cancer, 43(5), 762-767.
Zou, G. (2004). A modified poisson regression approach to prospective studies with binary data.
American journal of epidemiology, 159(7), 702-706. doi:10.1093/aje/kwh090
206
Supplementary Materials
Supplementary Materials Table 3.1: Relative risk of girl child marriage (binary) and underweight for pooled analysis (N=249,269)
Variables Model 1 Model 2 Model 3 Model 4 Model 5
Girl child marriage (18+ years, ref.) 0·936*** (0·918, 0·955) 0·946*** (0·924, 0·967) 0·943*** (0·922, 0·965) 0·942*** (0·920, 0·963) 0·940*** (0·919, 0·962) Completion of primary education (no, ref.) 0·826*** (0·802, 0·852) 0·820*** (0·795, 0·845) 0·850*** (0·824, 0·876) 0·884*** (0·857, 0·912) 0·829*** (0·802, 0·857)
Current age (20-24 years, ref.)
25-29 years 0·947*** (0·920, 0·974) 0·949*** (0·922, 0·976) 0·953*** (0·926, 0·981) 0·950*** (0·923, 0·977)
30-34 years 0·951*** (0·920, 0·983) 0·953*** (0·922, 0·985) 0·959** (0·928, 0·991) 0·952*** (0·921, 0·984)
35-39 years 1·01 (0·973, 1·05) 1·01 (0·974, 1·05) 1·02 (0·980, 1·06) 1·01 (0·968, 1·04)
40-44 years 1·06*** (1·02, 1·11) 1·06*** (1·02, 1·11) 1·07*** (1·02, 1·11) 1·05** (1·004, 1·09)
207 45-49 years 1·15*** (1·10, 1·20) 1·15*** (1·10, 1·20) 1·15*** (1·10, 1·20) 1·12*** (1·07, 1·17)
Age at first birth (years) 0·996** (0·993, 0·999) 0·997 (0·994, 1·00) 0·997 (0·994, 1·00) 0·997 (0·994, 1·00)
Number of children ever born 0·970*** (0·965, 0·975) 0·970*** (0·965, 0·975) 0·970*** (0·965, 0·975) 0·971*** (0·966, 0·976) Completion of secondary education (no, ref.) 0·725*** (0·677, 0·776) 0·773*** (0·722, 0·828) 0·737*** (0·688, 0·790)
Asset quintile (poorest, ref.)
Poorer 0·947*** (0·921, 0·975) 0·955*** (0·928, 0·983)
Middle 0·892*** (0·867, 0·917) 0·902*** (0·877, 0·928)
Richer 0·814*** (0·786, 0·843) 0·830*** (0·801, 0·860)
Richest 0·598*** (0·569, 0·628) 0·619*** (0·589, 0·650) Age gap between partner and woman (years) 0·999** (0·997, 1·00) Education gap between partner and woman (levels) 0·906*** (0·888, 0·925)
Note. Coefficients presented are relative risk estimates from Poisson regression models. 95% CIs in parentheses are based on cluster standard errors. Underweight is defined as body mass index less than 18·5. M1 adjusts for sampling cluster and woman's primary education (as potential confounders). M2 adjusts for all covariates in M1 + woman's age, age at first birth, and number of children born (as biological mechanisms). M3 adjusts for all covariates in M2 + secondary education. M4 adjusts for all covariates in M3 + asset quintile. M5 includes remaining covariates. Bolded values are significant at the p<0·05 level. *** p<0·01, ** p<0·05
Supplementary Materials Figure 3.1: Country-specific associations between girl child marriage (below 18 years) and severely underweight, conditional on full set of covariates
Study ID ES (95% CI) ssize
Benin 0.94 (0.61, 1.45) 2131 Burkina Faso 1.05 (0.74, 1.49) 2542 Burundi 3.59 (0.90, 14.29) 193 Cameroon 1.05 (0.44, 2.51) 446 Central African Republic 0.52 (0.11, 2.33) 134 Chad 1.31 (0.90, 1.90) 2651 Comoros 1.29 (0.24, 6.97) 194 Congo, Dem. Rep. 0.63 (0.37, 1.07) 968 Congo, Rep. 1.37 (0.79, 2.38) 751 Cote d'Ivoire 3.63 (0.32, 41.31) 88 Ethiopia 1.21 (0.99, 1.47) 7304 Gabon 0.62 (0.09, 4.12) 145 Gambia 1.36 (0.55, 3.36) 403 Ghana 1.08 (0.50, 2.30) 447 Guinea 1.51 (0.78, 2.93) 731 Kenya 1.06 (0.73, 1.53) 1909 Lesotho 1.07 (0.15, 7.39) 96 Liberia 1.42 (0.56, 3.62) 440 Madagascar 0.83 (0.61, 1.15) 2343 Malawi 1.05 (0.64, 1.72) 1529 Mali 1.09 (0.76, 1.57) 3688 Mozambique 1.22 (0.68, 2.18) 845 Namibia 1.08 (0.50, 2.33) 431 Niger 1.70 (0.89, 3.25) 898 Nigeria 0.80 (0.60, 1.06) 9115 Rwanda 0.91 (0.44, 1.85) 617 Senegal 0.66 (0.38, 1.14) 779 Sierra Leone 0.92 (0.50, 1.68) 572 Tanzania 0.48 (0.28, 0.80) 1338 Togo 0.66 (0.21, 2.02) 378 Uganda 0.81 (0.36, 1.80) 393 Zambia 0.96 (0.65, 1.42) 1977 Zimbabwe 0.80 (0.40, 1.62) 657 Sao Tome and Principe (Excluded) 92 Overall (I-squared = 56.4%, p = 0.000) 1.17 (0.49, 1.86)
-41.3 1 41.3
Note. All models control for primary education, age, age at first birth, number of children ever born, secondary education, asset quintile, age gap, education gap, and EA fixed-effects (M5). Based on 35 independent country-specific models.
208
Supplementary Materials Figure 3.2: Country-specific associations between girl child marriage (below 18 years) and underweight for women aged 20 to 24, conditional on full set of covariates
Study ID ES (95% CI) ssize
Benin 1.19 (0.61, 2.30) 1187 Burkina Faso 0.88 (0.36, 2.13) 1089 Burundi 0.46 (0.00, 48.97) 65 Cameroon 1.18 (0.42, 3.31) 329 Central African Republic 0.16 (0.03, 0.96) 160 Chad 1.32 (0.71, 2.43) 1285 Comoros 0.17 (0.02, 1.53) 89 Congo, Dem. Rep. 2.28 (0.62, 8.32) 495 Congo, Rep. 0.89 (0.29, 2.76) 433 Cote d'Ivoire 2.31 (0.06, 86.70) 59 Ethiopia 1.24 (0.81, 1.89) 2131 Gabon 0.68 (0.16, 2.83) 171 Gambia 15.42 (2.12, 112.23) 190 Ghana 1.28 (0.18, 8.89) 183 Guinea 0.98 (0.40, 2.40) 394 Kenya 0.82 (0.36, 1.83) 927 Liberia 0.89 (0.14, 5.54) 214 Madagascar 1.07 (0.51, 2.23) 933 Malawi 1.25 (0.62, 2.53) 1360 Mali 0.86 (0.56, 1.34) 2194 Mozambique 1.25 (0.60, 2.60) 793 Namibia 0.97 (0.03, 27.16) 110 Niger 2.10 (0.95, 4.63) 677 Nigeria 1.28 (0.86, 1.90) 3290 Rwanda 0.00 (0.00, 0.00) 168 Senegal 1.09 (0.40, 2.96) 369 Sierra Leone 2.98 (0.54, 16.34) 157 Tanzania 0.65 (0.28, 1.52) 797 Togo 0.04 (0.00, 0.34) 259 Uganda 0.67 (0.17, 2.68) 275 Zambia 1.34 (0.65, 2.75) 1232 Zimbabwe 1.64 (0.41, 6.61) 675 Lesotho (Excluded) 81 Sao Tome and Principe (Excluded) 62 Overall (I-squared = 100.0%, p = 0.000) 1.54 (-0.84, 3.93)
-112 1 112
Note. All models control for primary education, age, age at first birth, number of children ever born, secondary education, asset quintile, age gap, education gap, and EA fixed-effects (M5). Based on 35 independent country-specific models.
209
Supplementary Materials Figure 3.3: Country-specific associations between girl child marriage (below 18 years) and underweight for 2011 to 2014 data, conditional on full set of covariates
Study ID ES (95% CI) ssize
Benin 1.46 (1.01, 2.10) 4198 Cameroon 0.94 (0.56, 1.55) 1560 Chad 1.04 (0.80, 1.35) 5304 Comoros 0.43 (0.16, 1.14) 1019 Congo, Dem. Rep. 1.06 (0.70, 1.60) 3481 Congo, Rep. 0.86 (0.48, 1.55) 2030 Cote d'Ivoire 0.92 (0.38, 2.20) 931 Ethiopia 0.89 (0.75, 1.07) 9016 Gabon 1.33 (0.53, 3.30) 830 Gambia 0.89 (0.52, 1.53) 1762 Ghana 2.03 (0.78, 5.28) 764 Guinea 1.47 (0.87, 2.50) 1695 Kenya 1.01 (0.68, 1.50) 3564 Lesotho 0.00 (0.00, 0.00) 166 Liberia 0.65 (0.26, 1.66) 1030 Mali 0.82 (0.50, 1.34) 1949 Mozambique 1.30 (0.89, 1.90) 4307 Namibia 0.28 (0.06, 1.26) 429 Niger 0.79 (0.48, 1.30) 1835 Nigeria 1.07 (0.87, 1.31) 16242 Rwanda 0.00 (0.00, 0.00) 1186 Sierra Leone 1.11 (0.71, 1.74) 2452 Togo 1.05 (0.41, 2.64) 1311 Uganda 0.42 (0.18, 0.99) 656 Zambia 0.82 (0.53, 1.25) 5643 Overall (I-squared = 100.0%, p = 0.000) 0.91 (0.76, 1.05)
-5.28 1 5.28
Note. All models control for primary education, age, age at first birth, number of children ever born, secondary education, asset quintile, age gap, education gap, and EA fixed-effects (M5). Based on 35 independent country-specific models.
210
Supplementary Materials Figure 3.4: Additional country-specific associations
Country-specific associations between girl child marriage (below 18 years) and underweight controlling for work status of woman, conditional on full set of covariates
Note. All models control for primary education, age, age at first birth, number of children ever born, secondary education, asset quintile, age gap, education gap, and EA fixed-effects (M5). We additionally add in whether or not the woman worked in the past year. Based on 35 independent country-specific models.
211
Supplementary Materials Figure 3.4 (Continued)
Country-specific associations between girl child marriage (below 18 years) and underweight, excluding women married at age 18 or 19, conditional on full set of covariates
Note. All models control for primary education, age, age at first birth, number of children ever born, secondary education, asset quintile, age gap, education gap, and EA fixed-effects (M5). Based on 35 independent country-specific models.
212
Supplementary Materials Figure 3.4 (Continued)
Country-specific associations between girl child marriage (below 18 years) and being anemic (mild, moderate, or severe), conditional on full set of covariates
Note. All models control for primary education, age, age at first birth, number of children ever born, secondary education, asset quintile, age gap, education gap, and EA fixed-effects (M5). Based on 35 independent country-specific models.
213
Supplementary Materials Table 4.1: Survey response rate for women by country and survey year
Year of Response rate Country survey for women a Central African Republic 2010 92% Chad 2010 88% Democratic Republic of Congo 2010 97.1% Ghana 2011 97% Kenya (Mombasa) 2009 93.5% Kenya (Nyanza) 2011 89.8% Madagascar (South) 2012 93% Malawi 2013-14 95.3% Mauritania 2011 93% Nigeria 2011 91% Sierra Leone 2010 95% Somalia (Northeast) 2011 94% Somalia (Somaliland) 2011 88.2% Swaziland 2010 95% Togo 2010 91% Zimbabwe 2014 93.7% Total 2010-14 93%
a As reported in MICS program country reports
214
Supplementary Materials Figure 4.2: Additional details regarding the calculation of the Early Childhood Development Index
Ten questions were asked across the four domains, with the option of yes, no, or don’t know. Learning skills assessed if a child could follow directions and work independently and related to preparation for success in school. Literacy-numeracy skills assessed the child’s ability to understand the alphabet/words, numbers, communicate, and cause-and-effect relationships.
Physical skills assessed if a child was able to manipulate small objects and walk unassisted.
Socioemotional skills assessed children’s ability to form positive relationships with others as well as internalizing and externalizing behaviors. All outcomes were coded so that 1 reflected being on-track and 0 reflected being off-track. A child was considered developmentally on-track in each domain if at least 50% of questions were in the desired direction, while the overall ECDI score was calculated following UNICEF guidelines of being on-track in at least three of four of the developmental domains. This was then reverse-coded to assess if a child was off-track for development in all analyses. A continuous ECD measure, which we call the child development score, was also calculated as a score of 0 to 10, based on the number of on-track responses for each of the questions.
215
Supplementary Materials Table 4.3: Bivariate correlations between girl child marriage and other covariates Covariate coefficient p-value Wealth quintile -0.0787 p<0.001 Mother completed secondary education -0.2326 p<0.001 Mother’s age at birth (continuous) -0.1861 p<0.001 Mother’s age at childbirth (binary) <16 years 0.1061 p<0.001 16-17 years 0.18 p<0.001 18-19 years 0.1147 p<0.001 Multiple birth -0.0058 p= 0.2645 Child gender (female) 0.005 p= 0.3338 Child age (months) 0.0173 p<0.001 Number of children 0.1436 p<0.001 Mother completed primary education -0.1221 p<0.001 Geographic location (urban) -0.0824 p<0.001
216
Supplementary Materials Table 4.4: Associations between girl child marriage and child development score
Model 1 Model 2 Model 3 Model 4 Model 5
β CI β CI β CI β CI β CI
Exposure variable Girl child marriage -0.271 (-0.306, -0.236) -0.0651 (-0.100, -0.0299) -0.0470 (-0.0836, -0.0104) -0.00412 (-0.0409, 0.0326) 0.00420 (-0.0323, 0.0407) (<18 years)
Contextual covariates Geographic location 0.395 (0.355, 0.435) 0.386 (0.347, 0.426) 0.312 (0.271, 0.352) 0.0548 (0.00787, 0.102) (urban) Mother completed 0.312 (0.270, 0.355) 0.297 (0.255, 0.339) 0.172 (0.127, 0.216) 0.114 (0.0694, 0.158) primary school
Biological covariates
Multiple birth -0.0702 (-0.159, 0.0186) -0.087 (-0.175, 0.00172) -0.100 (-0.188, -0.0120)
Gender (female) 0.168 (0.135, 0.201) 0.167 (0.134, 0.200) 0.168 (0.135, 0.201)
Age (months) 0.0387 (0.0363, 0.0412) 0.0384 (0.0360, 0.0408) 0.0383 (0.0359, 0.0407)
Number of children -0.0187 (-0.0263, -0.0112) -0.0104 (-0.0180, -0.00283) -0.00772 (-0.0153, -0.000169)
217 ever-born to mother
Hypothesized pathways Mother's age at childbirth
<16 years -0.110 (-0.253, 0.0333) -0.0707 (-0.213, 0.0718) -0.0285 (-0.170, 0.113)
16-17 years -0.114 (-0.198, -0.0306) -0.0936 (-0.177, -0.0104) -0.0645 (-0.147, 0.0182)
18-19 years -0.0698 (-0.138, -0.00128) -0.0567 (-0.125, 0.0116) -0.0337 (-0.102, 0.0342)
Mother completed 0.456 (0.405, 0.506) 0.327 (0.276, 0.379) secondary school Wealth quintile (reference: Poorest)
Poorer 0.150 (0.100, 0.200)
Middle 0.275 (0.223, 0.326)
Richer 0.358 (0.303, 0.414)
Richest 0.731 (0.665, 0.798)
N 37,558 37,557 37,430 37,430 37,430
Chi-square 228.97 207.76 194.61 200.79 192.27
Supplementary Materials Table 4.4 (Continued)
Note. Coefficients presented are OLS estimates from linear regression models with 95% CIs in parentheses. Development is defined by total number of on-track responses out of 10 early childhood development questions. M1 is unadjusted model. M2 adjusts for urban/rural, mother's primary education, and country fixed-effects. M3 adjusts for all covariates in M2 + mother's age at birth. M4 adjusts for all covariates in M3 + mother's secondary education. M5 adjusts for all covariates in M4 + wealth quintile. Bolded values are significant at the p<0.05 level.
218
Supplementary Materials Table 4.5: Associations between girl child marriage and height-for-age z score
Model 1 Model 2 Model 3 Model 4 Model 5
β CI β CI β CI β CI β CI
Exposure variable Girl child marriage -0.166 (-0.198, -0.134) -0.0870 (-0.120, -0.0544) -0.0595 (-0.0939, -0.0252) -0.0326 (-0.0671, 0.00201) -0.0263 (-0.0608, 0.00816) (<18 years)
Contextual covariates Geographic location 0.209 (0.169, 0.248) 0.299 (0.261, 0.336) 0.253 (0.215, 0.291) 0.0706 (0.0259, 0.115) (urban)
Mother completed 0.299 (0.262, 0.336) 0.213 (0.172, 0.253) 0.140 (0.0978, 0.182) 0.112 (0.0695, 0.154) primary school
Biological covariates
Multiple birth -0.164 (-0.246, -0.0816) -0.174 (-0.256, -0.0916) -0.184 (-0.266, -0.102)
Gender (female) 0.0683 (0.0371, 0.0995) 0.0672 (0.0361, 0.0983) 0.0679 (0.0370, 0.0989)
Age (months) 0.000738 (-0.00156, 0.00304) 0.000527 (-0.00177, 0.00282) 0.000452 (-0.00183, 0.00274)
219 Number of children -0.00361 (-0.0108, 0.00362) 0.00174 (-0.00554, 0.00902) 0.00389 (-0.00336, 0.0112) ever-born to mother
Hypothesized pathways Mother's age at childbirth
<16 years -0.242 (-0.376, -0.108) -0.219 (-0.353, -0.0855) -0.191 (-0.324, -0.0575)
16-17 years -0.178 (-0.255, -0.100) -0.165 (-0.242, -0.0873) -0.144 (-0.222, -0.0673)
18-19 years -0.126 (-0.190, -0.0622) -0.118 (-0.182, -0.0545) -0.102 (-0.166, -0.0386) Mother completed 0.261 (0.215, 0.307) 0.172 (0.125, 0.219) secondary school Wealth quintile (reference: Poorest)
Poorer 0.0229 (-0.0240, 0.0697)
Middle 0.100 (0.0520, 0.149)
Richer 0.161 (0.109, 0.213)
Richest 0.483 (0.420, 0.546)
N 33,483 33,482 33,359 33,359 33,359
Chi-sq 104.94 95.31 68.75 71.38 70.78
Supplementary Materials Table 4.5 (Continued)
Note. Coefficients presented are OLS estimates from linear regression models with 95% CIs in parentheses. M1 is unadjusted model. M2 adjusts for urban/rural, mother's primary education, and country fixed-effects. M3 adjusts for all covariates in M2 + mother's age at birth. M4 adjusts for all covariates in M3 + mother's secondary education. M5 adjusts for all covariates in M4 + wealth quintile. Bolded values are significant at the p<0.05 level. Results exclude Madagascar (South), Somalia (Northeast), and Somalia (Somaliland). 220
Supplementary Materials Table 4.6: Associations between girl child marriage and child development and health outcomes
Off track for development (ECDI) Stunting Unadjusted Adjusted Unadjusted Adjusted OR CI OR CI OR CI OR CI Exposure variable Girl child marriage
(reference: 18+ years) <15 years 1.33 (1.25, 1.41) 0.940 (0.878, 1.01) 1.44 (1.35, 1.53) 1.12 (1.05, 1.21) 15-17 years 1.22 (1.17, 1.28) 1.03 (0.984, 1.09) 1.23 (1.17, 1.29) 1.01 (0.955, 1.06) Contextual covariates Geographic location (urban) 0.934 (0.879, 0.993) 0.924 (0.865, 0.987) Mother completed primary 0.925 (0.874, 0.979) 0.841 (0.791, 0.894) school
Biological covariates
Multiple birth 1.14 (1.02, 1.28) 1.19 (1.06, 1.34)
Gender (female) 0.843 (0.808, 0.879) 0.891 (0.852, 0.933) 221 Age (months) 0.971 (0.968, 0.974) 0.990 (0.987, 0.993) Number of children ever- 1.01 (0.998, 1.02) 0.999 (0.988, 1.01) born to mother
Hypothesized pathways
Mother's age at childbirth
<16 years 1.00 (0.833, 1.20) 1.23 (1.01, 1.48)
16-17 years 1.07 (0.960, 1.19) 1.19 (1.06, 1.33)
18-19 years 1.01 (0.929, 1.11) 1.10 (1.01, 1.21) Mother completed secondary 0.825 (0.771, 0.882) 0.771 (0.718, 0.828) school Wealth quintile (reference: Poorest)
Poorer 0.894 (0.839, 0.952) 0.962 (0.899, 1.03)
Middle 0.825 (0.772, 0.881) 0.895 (0.835, 0.959)
Richer 0.777 (0.724, 0.834) 0.780 (0.723, 0.840)
Richest 0.615 (0.564, 0.670) 0.524 (0.477, 0.575)
N 37,558 37,430 33,483 33,359
Chi-sq 126.24 2803.68 152.10 1903.06
Supplementary Materials Table 4.6 (Continued)
Note. Coefficients presented are odds ratios with 95% CIs in parentheses. Unadjusted model includes only girl child marriage and the outcome. Adjusted model includes child characteristics, urban/rural, mother's primary education, mother's age at birth, mother's secondary education, wealth quintile, and country fixed-effects. Bolded values are significant at the p<0.05 level.
222
Supplementary Materials Table 4.7: Sobel-Goodman Test for mediating influences
Off-track for ECDI Proportion of Goodman- Goodman- total effect that is a SE b SE Sobel SE 1 (Aroian) SE 2 SE mediated Age at childbirth -2.676 0.072903 0.072903 0.000362 0.002089 0.000971 0.002089 0.000972 0.002089 0.000971 0.037 Completed secondary school -0.188 0.004068 -0.11414 0.006469 0.021514 0.001305 0.021514 0.001305 0.021514 0.001304 0.385
223 Wealth quintile -0.251 0.014238 -0.02954 0.001849 0.007427 0.000627 0.007427 0.000628 0.007427 0.000626 0.133
Stunting Age at childbirth -2.587 0.077704 -0.00199 0.000375 0.005153 0.000982 0.005153 0.000982 0.005153 0.000982 0.085 Completed secondary - school 0.21262 0.004412 -0.1383 0.006562 0.029406 0.001523 0.029406 0.001523 0.029406 0.001522 0.485 - Wealth quintile 0.26172 0.015061 -0.0398 0.001922 0.010415 0.000783 0.010415 0.000783 0.010415 0.000782 0.172
Note. Bolded values are significant at the p<0.05 level.