Internal Migration and Premarital Reproductive Behavior Among Adolescents in Nigeria
By
Blessing Uchenna Mberu
B.Sc. Hons. Sociology, Imo State University, Nigeria, 1987
M.Sc. Sociology, University of Ibadan, Ibadan, Nigeria, 1990
M. A. Sociology, Brown University, 2004
A dissertation submitted in partial fulfillment of the requirements for the degree of
Doctor of Philosophy in the Department of Sociology
Brown University, Providence, RI 02912
May 2008
© Copyright 2008 by Blessing U. Mberu
This dissertation by Blessing Uchenna Mberu is accepted in its present form by the
Department of Sociology, Brown University as satisfying the dissertation requirements for the degree of Doctor of Philosophy.
Date……………………………… ………....……………………………………...... Michael J. White, Advisor and Committee Chair
Recommended to the Graduate Council
Date………………………………… ……………………………………….. Daniel J. Smith, Committee member
Date……………………………….. ……………………………………….. Nancy Luke, Committee member
Approved by the Graduate Council
Date……………………………….. ……………………………………………….. Sheila Bonde, Dean of the Graduate School
iii
CURRICULUM VITAE BLESSING UCHENNA MBERU [email protected];
Address Population Studies and Training Center and Department of Sociology Brown University Providence RI 02912 Phone: 401-331-2749
Education Ph.D in Sociology, Brown University, Providence, RI, USA, May 2008 (Expected). MA in Sociology, Brown University, Providence, RI, USA, May 2004. M.Sc. in Sociology, University of Ibadan, Ibadan, Nigeria, Decemeber 1990. B.Sc (Hons.) in Sociology, Imo State University, Okigwe, Nigeria, August 1987.
Research Interest Migration, Urbanization, Adolescent Reproductive Behavior, and Poverty in sub-Saharan Africa.
Books • Mberu, B.U. 1999. Infrastructural Sociology. Enugu, Nigeria : John Jacobs Classic Publishers Ltd. • Mberu, B.U. 1999. Social Structures and Institutions: Nuclear Themes in Sociology. Abakaliki, Nigeria: Willyrose and Appleseed Publishing Company.
Recent Peer Review Journal Articles • Mberu, B.U. 2005. Who Moves and Who Stays? Rural Out-Migration in Nigeria. Journal of Population Research, 22 (2): 141-161. • Mberu B.U. 2006. Internal Migration and Household Living Conditions in Ethiopia. Demographic Research, 14(21): 509-540. • Mberu, B.U. 2007. Household Structure and Living Conditions in Nigeria. Journal of Marriage and the Family, 69 (May 2007): 513–527.
Work in Progress • Mberu, B.U., Migration Circulation and Rural Development in Sub-Saharan Africa: The Case for a Multi-Factor Study Framework (under review). • White, M.J., Mberu, B. U. and Collinson, M.A., African Urbanization: Recent Trends and Implications. Book chapter prepared for Cities, Poverty and Environment: Addressing the Future Now. (UNFPA/IIED Book Project).
iv Blessing Mberu CV
• Mberu, B. U. and R. Pongou, Ethnic identity and human capital accumulation: Educational attainment among Nigerian youths (under review). • Mberu, B. U., A.M. Terraza, and R. Pongou, Nigeria Migration Profile: Shaped by political evolution and divergent regional economies. Washington DC: Migration Policy Institute (under review). • Mberu, B. U., R. Pongou, and A. M. Terraza, Crossing Boundaries: Internal, Regional and International Migration in Cameroon (under review).
Book Chapters • Mberu, B.U. 1995. Effective Planning and Execution of Population Health Programmes: Issues Relating to Community Participation in Nigeria. In A. Anyanwu (ed.), Enabling Environment for Economic Growth. Benin, Nigeria: Barloz Publishers Inc. • Mberu, B.U. 1996. The Problem of Accountability in Contemporary Nigeria. In C. Nwahunanya (ed.), Higher Perspectives on CitizenshipEducation. Owerri, Nigeria: Springfield Publishers. • Mberu, B.U. 1998. The Development of Science and Technology in the Third World: A Sociological Perspective. In A. Ebeogu and U. Nwokocha (eds.), Issues in Humanities and Social Sciences. Okigwe, Nigeria: Fasmen Communications. • Mberu, B.U. 2001. Religion and Citizenship Education in Nigeria. In O. Okereke (ed.), Readings in Citizenship Education. Okigwe, Nigeria: Whythem Publishers. • Mberu, B.U. 2001. The Causes and Social Effects of Drug and Alcohol Abuse. In D.Njoku (ed.), Readings in Citizenship Education 11. Enugu, Nigeria: John Jacobs Classic Publishers Ltd.
Working Papers • Mberu, B.U. 2005. Internal Migration and Household Living Conditions in Ethiopia. African Population and Health Research Center, Nairobi, Kenya, Working Paper No 31. • Mberu, B.U., F. Nduka, and B.E.B. Nwoke.1998. Community Participation in Onchocerciasis Control Programme in the Imo River Basin of Nigeria. L’Institut de Formation et de Recherche Demographiques, Universite de Yaounde. Les Cahiers de l’IFORD No 21.
v Blessing Mberu CV Theses and Dissertation
• Mberu, B.U. 2008. Internal Migration and Adolescent Premarital Reproductive Behavior in Nigeria (PhD Dissertation, Department of Sociology, Brown University). • Mberu, B.U. 2004. Rural Out-Migration and Migrants’ Destination in Nigeria (M.A. Thesis, Department of Sociology, Brown University). • Mberu, B.U. 1990. Fertility Levels and Family Planning in Rural Nigeria: A Study of Akokwa Community, Imo State Nigeria (M. Sc. Thesis Department of Sociology, University of Ibadan, Ibadan, Nigeria (funded by CODESRIA, Dakar, Senegal). • Mberu, B.U. 1987. The Effects of Increases in School Fees on Primary School Enrolment: A Case Study of Akokwa Community, Imo State, Nigeria (B.Sc. Thesis Department of Sociology, Imo State University, Okigwe, Nigeria).
Presentations in Professional Meetings
• Mberu, B.U. 1995. Effective Planning and Execution of Population Health Programmes: Issues Relating to Community Participation in Nigeria. (Paper presented at the Conference on Enabling Environment for National Growth, Abia State University, 1994). • Mberu, B.U. 2006. Predictors of Premarital Sexual Initiation among Adolescents in Nigeria. (Poster presented at the 2006 Annual Meeting of the Population Association of America, Los Angeles, CA). • Zulu, E. M., A. Konseiga, E. Darteh, and B. U. Mberu. 2006. Migration and Urbanization of Poverty in sub-Saharan Africa: The Case of Nairobi City, Kenya. (Paper presented at the 2006 Annual Meeting of the Population Association of America, Los Angeles, CA). • Mberu, B.U and R. Pongou. 2007. Ethnic Identity and Human Capital Accumulation: Educational Attainment among Nigerian Youths. (Paper presented at the 5th Wits-Brown-Colorado-APHRC Colloquium on Emerging Population Issues, Brown University March 24-26, 2007) • Mberu, B. U. and R. Pongou. 2007. Ethnic Identity and Human Capital Accumulation: Educational Attainment among Nigerian Youths (Poster presented at the annual meeting of the Population Association of America, New York, March, 2007)
vi Blessing Mberu CV Collaborative Interdisciplinary Research Projects
• Nwoke, B.E.B., B. U. Mberu, O. Oha, I.N.S. Dozie and C.N. Ukaga. 1999. Lymphatic Filariasis and Onchocerciasis in the rainforest of southeastern Nigeria: Knowledge, Attitudes and Practice of Endemic Communities. WHO/ TDR/ ID: 931087. Geneva: WHO. • Nwoke, B.E.B., B. U. Mberu, O. Oha, I.N.S. Dozie and C.N. Ukaga. 1999. Lymphatic Filariasis and Onchocerciasis in the rainforest of southeastern Nigeria: the Social Effects of Genital Complications among Women. WHO/ TDR/ ID: 931087. Geneva: WHO. • Nwoke, B.E.B., B. U. Mberu, O. Oha, I.N.S. Dozie and C.N. Ukaga. 1999. Lymphatic Filariasis and Onchocerciasis in the rainforest of southeastern Nigeria:Clinical and Parasitological Studies. WHO/ TDR/ ID: 931087. Geneva: WHO.
Work Experience • Research Intern, African Population and Health Research Center, Nairobi, Kenya, September 2004 –June, 2005. • Senior Lecturer in Sociology, Abia State University, 1999- 2002. • Lecturer 1 in Sociology, Abia State University, 1996-1999. • Lecturer 11 in Sociology, Abia State University, 1993-1996. • Assistant Lecturer in Sociology, Abia State University, 1991- 1993. • Graduate Assistant, Imo State University (Now Abia State University, Uturu) August, 1988 - December, 1989.
Grants and Awards 1) Compton Fellowship, 2007/2008. 2) Ford Foundation Fellowship, 2007. 3) Compton Fellowship, 2006. 4) Poster Award Population Association of America, 2006. 5) Mellon Foundation Africa Grant, 2003-2004. 6) Mellon Foundation Africa Grant, 2002-2003. 7) WHO/L’Institut De Formation Et De Recherche Demographiques, Universitie du Yaounde Onchocerciasis Research Grant, 1994. 8) CODESRIA (Council for the Development of Economic and Social Research in Africa) Small Grant for Thesis Writing, 1990. 9) Deans prize for Sociology, Imo State University, Okigwe, Nigeria, 1987. 10) School of Social Sciences prize for Sociology, Imo State University, Nigeria, 1987.
Review Service to Peer-review Journals 1) Journal of Marriage and Family. 2) Migration Information Source. 3) Population Review.
Computer Skills: Microsoft Office Programs, STATA, SPSS.
vii Abstract
The high rate of migration of young people to urban and economic centers in the context of persistent poverty and devastating HIV/AIDS burden raise intricate social policy questions on early sexual behavior in Nigeria. This dissertation explores the linkages between patterns of adolescent internal migration and premarital sexual initiation, condom use at premarital sexual debut, risk perception for sexually transmitted infections
(STIs) and being sexually active before marriage. The study analyzes data from the 1999 and 2003 Nigeria Demographic and Health Surveys, utilizing bivariate statistics, Kaplan-
Meier survival curves, discrete-time hazard and binary logistic regression models.
The results link rural-urban migrants with the strongest independent association to
early premarital sexual initiation, lending support to the notion that exposure to a more
sexually lenient urban environment different from normatively traditional rural origins
may increase the propensity of rural-urban migrant adolescents to premarital sexual
involvement. Other significant covariates are adolescent age, gender, religious affiliation, ethnic origin, independent living arrangement, and employment in the formal sector.
The analyses of predictors of condom use, perception of risk of STIs, and being
sexually active, collectively quantify the high-risk context that premarital sexual
engagement entails in Nigeria (15.8% of condom use at sexual initiation, 64% being
sexually active, and 45% denying any risk of STIs). The findings highlight the
complexity of social contexts that exist within groups in one country, the inadequacy of
policy perspectives that neglect contextual nuances and the need to look beyond sexual
behavior to address limited livelihood opportunities that fuel massive and often distress
adolescent rural out-migration to urban destinations.
viii Acknowledgements
I would like to express my deepest gratitude to my advisor and chair of my mentoring and dissertation committee, Michael J. White, for his support and encouragement throughout my stay in the graduate program. His technical guidance and pointed critical questions were invaluable in the shaping and completion of this dissertation.
Daniel J. Smith deserves my special thanks for bringing me to Brown and for his gracious mentorship throughout my program. He brought his amazing knowledge of
Nigeria to bear on his consistent guidance of my graduate education generally and my dissertation work in particular. He is always pushing me for more contextually nuanced interpretations of results. I only hope that I can live up to his high levels of academic standards and human kindness.
I would like to thank Nancy Luke for the warmth and commitment she brought to my dissertation committee. I am grateful for her painstaking guidance from my preliminary examinations, through dissertation proposal and completion.
I am deeply grateful to Patrick Heller for stepping up to my help as a member of my mentoring committee. I learnt so much from his theory class and his encouragements and guidance were invaluable through the process of my preliminary examination and dissertation. Great thanks go to my outside readers, Dennis Hogan and Gregory Elliot.
I would like to specially thank Frances and Calvin Goldscheider. Fran and Calvin believed in me and told me so. They encouraged me all the way from my first year at
Brown. My paper on Household Structure and Living Conditions in Nigeria published in the Journal of Marriage and the Family, 69 (May 2007): 513–527, began in Fran’s class and a tribute to her inspiration and encouragement.
ix I want to thank Mary Fennel, for her great assistance; Jonathan Kelley, Mariah
Evans, Susan Short, Paget Henry, Jose Itzigsohn, David Lindstrom and all the professors of the department of sociology for their great contributions throughout my years at
Brown. I would like to thank William Simmons, Mark Pitt and David Kertzer for their invaluable assistance and encouragements.
I am grateful to the management and colleagues at African Population and Health
Research Center (APHRC) Nairobi Kenya under Alex Ezeh and Eliya Zulu for offering me a home between 2004 and 2005 as a Brown travel scholar. Special thanks to Nyovani
Madise, Zewdu Woubelam, J-C Fotso, Joseph Gichuru, Chi-Chi Undie, Lillian Okoth,
George Kidiga and all. Thanks to Nkee Ezeh and family for home care.
I want to thank Abia State University, Nigeria and Elder (Prof.) Ogwo Ekeoma
Ogwo for study leave and financial assistance that enabled my trip to Brown. I am grateful for prayers and friendship of the brethren of August 14 Fellowship, Abia State
University and Full Gospel Business Men’s Fellowship International, Okigwe chapters.
I am deeply indebted to friends and colleagues at Brown: Olumide Taiwo, Roland
Pongou, Daniel Schensul, Becky Altman, Afra Chowdhury, Salome Wawire, Thandie
Hlabana, Holly Reed, Adrianna Lopez-Ramirez, Kassahun Admassu, Julia Drew, etc.
I want to specially thank the wife of my youth Chi-Chi, daughter Favour, and sons
Emmanuel and Daniel for their patience and support. Special thanks to Daddy and
Mummy Ajaegbu, Ifeanyi Emeje, Chima and Ngozi James, Emeka Ohagi for prayers and blessings and support on every turn. Special thanks to Wilson Duru, Cletus Agwunobi,
Hassan Saheed, and most of all my great American host family-Billie and Gil Mason of the International House of Rhode Island. God bless you all abundantly.
x TABLE OF CONTENTS
Abstract
Acknowledgements
Table of Contents
List of Tables
List of Illustrations
List of Appendices
Chapter 1: Introduction
1.1. Research Questions
1.2. Background to the Study
1.3. Nigeria, Migration and Adolescent Reproductive Behavior
1.4. Primary objectives
1.5. Significance of the study
1.6.Limitations of the study
1.7. Organization of the dissertation
Chapter 2: Theory, literature and hypotheses
2.0. Introduction
2.1. Individual level theories on migration and reproductive behavior
2.2. Family and social environmental models
2.3. The life course perspective
2.4. Premarital Sexual Behavior: Empirical Research Review in Sub-Saharan Africa
2.5. Conceptual framework
2.6. Statement of preliminary hypothesis
xi Chapter 3: Data and methods
3.1. Data
3.2. Definition of Variables
3.3. Methods and statistical models
Chapter 4: Data Analysis: Premarital Sexual Initiation
4.1. Characteristics of adolescents study population
4.2. Premarital sexual initiation: Descriptive analysis
4.3 Premarital sexual initiation: Discrete-time event history analysis
Chapter 5: Data Analysis: Condom use at premarital sexual initiation
5.0. Introduction
5.1. Condom use at premarital sexual initiation: Descriptive analysis
5.2. Condom use at premarital sexual initiation: Logistic models
5.3. Summary
Chapter 6: Data Analysis: Risk Perception for Sexually Transmitted Infections
6.0. Introduction
6.1. Risk Perception for Sexually Transmitted Infections: Descriptive analysis
6.2. Risk Perception for Sexually Transmitted Infections: Logistic models
6.3. Summary
Chapter 7: Data Analysis: Active Premarital Sexual Behavior
7.0. Introduction
7.1. Active Premarital Sexual Behavior: Descriptive analysis
7.2. Active Premarital Sexual Behavior: Logistic models
7.3. Summary
xii Chapter 8: Summary and Conclusion
8.1. Summary of study
8.2. Key findings and implications for policy, program and theory
8.3. Direction of future research
References
Appendices
xiii List of Tables
Tables 1: Operational definition of independent variables, direction of effects, suggested
mechanisms and citations……… ……...…………………………………………… 46-47
Table 2: Univariate characteristics of the study population and bivariate association with
premarital sexual initiation……………………………………………………………....54
Table 3: Discrete-time hazard models predicting the odds of premarital sexual initiation
among Nigerian adolescents……………………………………………………………..59
Table 4: Odds Ratios for the Effects of Ethnicity and Household SES Interaction with
Gender on the Hazard of Premarital Sexual Initiation among Nigerian adolescents…....66
Table 5: Characteristics of adolescents who used condom at sexual initiation………….71
Table 6: Estimates predicting condom use at premarital sexual initiation among Nigerian adolescents……………………………………………………………………………….74
Table 7: Perception of risk of STIs among the premarital sexually initiated……………84
Table 8: Odds ratios predicting risk perception of STIs among premarital sexually
initiated Nigerian adolescents………………………………..………………………....87
Table 9: Characteristics of sexually active never married adolescents three months prior
to survey………………………………………………………………………………….97
Table 10: Logistic regression estimates predicting active premarital sexual behavior
among never married Nigerian adolescents………………….………………………..100
xiv List of Illustrations
Figure1: Conceptual Model of Migration and Adolescent Premarital Reproductive
Behavior…………………………………………………………………………….32
Figure 2: Kaplan-Meier survival estimates of premarital sexual debut by adolescent migration status...... ……………………………………………………………………142
Figure 3: Kaplan-Meier survival estimates of premarital sexual debut by gender……..143
Figure 4: Kaplan-Meier survival estimates of premarital sexual debut by age cohorts...143
Figure 5: Kaplan-Meier survival estimates of premarital sexual debut by religion…....144
Figure 6: Kaplan-Meier survival estimates of premarital sexual debut by childhood place of residence……………………………………………………………………………..144
Figure 7: Kaplan-Meier survival estimates of premarital sexual debut by Current place of residence…………...... 145
Figure 8: Kaplan-Meier survival estimates of premarital sexual debut by Ethnic groups…………………………………………………………………………………...145
Figure 9: Kaplan-Meier survival estimates of premarital sexual debut by education.....146
Figure 10 Kaplan-Meier survival estimates of premarital sexual debut by media exposure index…………………………………………………………………………………….146
Figure 11: Kaplan-Meier survival estimates of premarital sexual debut by Living arrangement………………………………………………………………………...... 147
Figure 12: Kaplan-Meier survival estimates of premarital sexual debut by HH SES.....147
Figure 9: Kaplan-Meier survival estimates of premarital sexual debut by Status and nature of employment……………………………………………………………….....148
xv CHAPTER 1
INTRODUCTION
1.1 Research Questions
Drawing from theoretical and empirical research on the relationship between migration and reproductive behavior, this dissertation seeks answers to the following key questions:
Why is internal migration and adolescent reproductive behavior a relevant area of research in modern Nigeria? What do we know about adolescent premarital reproductive activities in the country and its relationship to migration and what gaps in knowledge exist? To what extent are the prevalence and timing of adolescent premarital sexual activities related to adolescent migration status? What gender differences exist in the relationships between adolescent migration and the prevalence and timing of premarital sexual debut, and how different are the outcomes among adolescents from Nigeria’s different ethnic nationalities? What other individual, household and community level variables mediate the relationship between internal migration and adolescent initiation of premarital intercourse? Among adolescents who initiated premarital sex, what are the predictors of condom use at premarital sexual initiation, perception of risk of sexually transmitted infections, and being sexually active (defined as premarital intercourse within three months of the survey). These questions, which this dissertation seeks to address, builds on the overall objective of understanding the particular linkages between the high rates of migration of young people to urban and economic centers and the challenge of adolescent premarital sexual behavior and perceptions, in the context of increasing poverty and health threats, particularly the HIV/AIDS epidemic, in Nigeria.
1 2
1.2 Background to the study
The increases in internal migration and the rate of urban growth associated with economic and political transitions in countries of Africa, Asia, Latin America and the
Pacific (Gurmu et al. 2000) have placed migration at the center of development policies and programs in these countries. The rapid urbanization process in sub-Saharan Africa
(associated with migration and the unprecedented growth in the size of the population of young people mostly through natural increase), in the context of declining economic performance of most African countries, have created a new face of poverty characterized by a significant proportion of the population living below the poverty line in over- crowded slums and sprawling shanty towns around major cities (NRCIM, 2005; Isiugo-
Abanihe and Oyediran, 2004; APHRC, 2002; Brockerhoff and Brenan, 1998; UN-
HABITAT, 2003). This pattern of urban living is linked to high levels of unemployment and underemployment, crime, general debilitating environmental conditions, risky reproductive behavior and poor health outcomes (NCRIM, 2005; APHRC, 2002;
Mugisha and Zulu, 2004; Zulu et al. 2002).
Further, a corpus of research has particularly focused on adolescent reproductive behavior and the changing context of adolescent premarital sexual initiation in developing countries. Evidences from these studies indicate increased prevalence of premarital sexual relationships over time (Adedimeji, 2005; Isiugo-Abanihe and
Oyediran, 2004; Blanc and Way, 1998). This outcome is explained as a logical consequence of rising age at marriage, increased schooling rates, the falling age at puberty, the penetration of Western mass media and entertainment, increasingly common ideas about individualism, and the erosion of traditional social controls (NCRIM, 2005;
3
Bongaarts and Cohen, 1998). The increasing research focus on adolescent initiation of sexual activities is significantly motivated by findings that the timing of first sexual intercourse is highly associated with exposure to sexually transmitted infections (STIs) and HIV/AIDS, human papilloma virus and precancerous changes of the cervix, use of contraception, pregnancy and pregnancy complications (Gage, 1998; Blum, 2002,
Abanihe and Oyediran, 2004). In developing countries, it is estimated that as many as 60 per cent of all adolescent pregnancies and births are unintended, and early pregnancy and unsafe abortion can compromise young women’s health (Onifade, 1999). Early parenthood can also interrupt schooling which can lead to fewer job opportunities and lower income (McCauley et al. 1995). Every year, millions of young people contract sexually transmitted diseases (Population Reference Bureau and Center for Population
Options, 1994) and it has been established that the presence of an STI leads to possible infertility and increased HIV transmission (Elias and Heise, 1993). Further, HIV/AIDS has become a disease of young people, with half of all new global HIV infections shown to occur in men and women less than 25 years of age. In many developing countries, recent evidence indicates that up to 60 percent of all new HIV infections are among 15-
24 year olds, with females outnumbering males by a ratio of 2 to 1 (Family Care
International 1995; UNFPA, 2003).
While studies on migration and reproductive behavior have existed as bourgeoning distinct spheres, there is a dearth of nationwide research in sub-Saharan
Africa that directly connects adolescent migration patterns to adolescent reproductive behaviors. However, research evidence from other places, particularly the United States, has informed us of linkages between geographical mobility, problematic sexual
4 behaviors, and health outcomes in different sociocultural settings (Evans, 1987;
Landsdale and Oropesa, 2001). Central to this linkage are the social disruptions which characterize migration and the relationships between migration and behavioral change for migrants and non-migrants. In particular, studies on the impact of social control factors on adolescent sexual involvements, such as religious systems (Beck et al. 1991; Davidson and Leslie, 1977), socioeconomic class position, and family bonds (Miller and Moore,
1990; Clayton and Bokemeier, 1980; Davidson and Leslie, 1977) have identified migration as an index of weakened social control, which can fracture bonds of integration at family and community levels (Stack, 1994). The weakening of bonds to conventional institutions is in turn associated with enhanced probability of deviant behavior (Sampson and Laub, 1993; Vold and Bernard, 1986; Liska, 1981). Further, in this era of the
HIV/AIDS epidemic in sub-Saharan Africa, population migration has become a central theme in the discussion of the epidemic, and a few but growing body of studies, mostly in
Eastern and Southern Africa, have recognized geographic mobility as one of the main facilitating conditions for HIV transmission (Lurie, 2004; Brockerhoff and Biddlecom,
1999; Fontanet and Piot, 1994; Hunt, 1989). In particular, massive migration of young, unmarried adults from presumably conservative rural environments to more sexually permissive African cities in recent years has been suggested as partly responsible for the much higher AIDS seroprevalence levels observed in urban than rural areas (UN,
1994:8). The dynamics of transmission becomes more complex as frequent movements between cities, towns and the home villages remain the norm for many urban migrants in sub-Saharan Africa (Andersson, 2001; Bartle, 1981; Geschiere and Gugler 1998; Gugler,
1971, 1991; Hanna and Hanna, 1981; Smith, 1999). This pattern of circulation
5 complicates the direction of the influence of migration in both urban and rural areas.
Consequently, migrants in both urban and rural areas are associated with higher propensity of sexual practices that elevate their risk of acquiring HIV/AIDS (Caldwell et al. 1997, Chirwa, 1997; Lurie et al. 1997). However, migrants with previous exposure to urban environments are linked with increased likelihood of high-risk sexual behavior in rural areas through socialization to less restrictive sexual norms or acquisition of enabling characteristics such as wealth in cities (Brockerhoff and Biddlecom, 1999). Despite the increasing body of research on these themes, particular focus on adolescent migration and its implications for reproductive behavior and health-related outcomes has received scant attention.
Building on this backdrop, this dissertation examines the linkages between the high rate of internal migration of young people in Nigeria and their premarital sexual activities and perceptions. The significance of understanding these linkages in Nigeria is premised on the demographic profile of the country, with an unprecedented growth of young people, their high rate of migration to urban and economic centers in search of livelihood opportunities in the context of persistent poverty, the fast pace of socio- cultural and economic changes unleashed by globalization, and the recent policy and program attention being focused on adolescents in the wake of the devastating effects of the HIV/AIDS epidemic (NRCIM, 2005; Isiugo-Abanihe and Oyediran, 2004; UN-
HABITAT, 2003; APHRC, 2002; Brockerhoff and Brenan, 1998). The study systematically reviews available theoretical and empirical works that relate migration to reproductive behavior and develops a comprehensive multi-factor study framework that
6 examines individual, household, and community/national level factors, using data from the nationally representative 1999 and 2003 Nigeria Demographic and Health Surveys.
1.3 Nigeria, Migration and Adolescent Reproductive Behavior
Nigeria, with a population growth rate of 3.2 per cent and a 2006 provisional census population of 140 million, is the most populous country in Africa and the tenth most populous in the world (Population and Development Review, 2007; National Population
Commission and ORC/MACRO, 2003; United Nations, 2004). The country’s population is young, with a median age of 17.4 years; over 45 percent are aged 0-14 and about 63 percent are currently less than 25 years of age (Isiugo-Abanihe, 2003).
Data on migration in Nigeria is scanty. The 1991 national census indicated that about 38 per cent of the population lived in urban areas, and this increased to 45 per cent in 2000, and projected to reach 55.5 per cent by 2015 (UN, 2003). This projection comes close to a 1996 World Bank report that suggests that the pace of urbanization in Nigeria is such that 60 per cent of the population will be urban by 2010 (World Bank, 1996).
Ekpunobi (2003) suggests that the population of Abuja the national capital triples every year, in the fastest process of urbanization in West Africa. Similarly, Lagos, with an estimated population of 13.2 million in 2003, is the 10th largest city in the world, and if
the estimated annual growth rate of 4 per cent is sustained, the metropolis will become
the world’s third largest city, averaging 24 million persons, by 2020 (World Bank, 1999).
Evidence from the region identified young men and women aged 15-29 as the most likely
to migrate from restrictive local environments to economic centers in search livelihood
opportunities (Adedimeji, 2005; NISER, 1997; Adepoju, 1986; Oucho, 1998; Smith,
7
2004a). A 1993 national representative migration survey identifies about 58 per cent of the 85,777 sample population as lifetime migrants (defined as those who have lived or are living outside their places of origin for at least six months), with 28% of all migrants aged 15-24 (NISER, 1997).
Competing traditional and modern values in Nigeria, as in several other African countries, have located premarital sexuality within the domain of moral ambivalence.
Ethnographic and demographic studies in the country have shown that while parental, family, and religious messages often assert that premarital sex is immoral, powerful peer pressure and forces of modernization and globalization equate premarital sex with modern, educated, urban lifestyles- identities for which many young people yearn (Smith,
2000, 2003). While these views will continue to generate wider discussions about social change and sexuality in Nigeria, the reality is that premarital sexual relationships are increasingly common in contemporary Nigeria. Adolescents in the country are increasingly linked to risky sexual behaviors such as casual sex and keeping of multiple partners, some of whom may include commercial sex workers (Isiugo-Abanihe, 2003;
Arowujolu, et al. 2002; Smith, 2000). Such behavior portends grave implications for reproductive health and current prognosis of the situation in the country is hardly optimistic. The U.S. National Intelligence Council (2002) projected the next wave of the
AIDS epidemic in Nigeria to grow to 15 million cases by 2010. The current HIV/AIDS prevalence rate is 5 per cent with about 3 million adults living with HIV/AIDS. Other estimates confirm rising rates of HIV prevalence in the country, from 1.8 per cent in 1992 to 3.8 per cent in 1994, 4.5 per cent in 1996, 5.4 per cent in 1999 and 5.8 per cent in
2001, with most of the new cases found among adolescents (Federal Ministry of Health,
8
2001). Other problems linked to adolescent sexual activities in the country include unwanted pregnancies and clandestine abortions by untrained or poorly trained providers
(Aja-Nwachuku, 2004; Aziken et al. 2003; Bankole et al. 2006; NPC and ORC/MACRO,
2003; Otoide et al. 2001).
The Federal Government of Nigeria launched a new national policy on population for sustainable development in January 2004, which seeks, among other objectives, to integrate adolescents into development efforts and effectively address their reproductive health needs. This follows similar population policies in 1988 and 2000. However, the dearth of research and scientific evidence informing most reproductive health policies and programs by the government continues to hinder the achievement of stated objectives. Slap et al. (2003) pointed out that the 2000 national policy on reproductive health that focused on preventing risky behaviors during adolescence was particularly hampered by outdated and incomplete information on the sexual knowledge, attitudes and behaviors of Nigerian adolescents. Therefore, achieving the objectives of the most recent
Nigerian 2004 population policy, particularly as it concerns adolescents, requires an updated understanding of factors that influence their reproductive behaviors in the country.
The bulk of studies on the relationship between migration and reproductive behavior in sub-Saharan Africa focus on countries of Eastern and Southern Africa, the so called “AIDS Belt” (Kaufman et al. 2002; French and Dishion, 2003, Brockerhoff and
Biddlecom, 1999). Relative to its population size and composition, very little is known about Nigeria. For the few studies undertaken in the country, most are largely constrained by limited coverage. Some are based on data collected from adolescents in schools,
9 leaving off non-school youths, who are estimated to consitute over 60 percent of all youths in the country (Slap et al. 2003; Amazigo, et al. 1997; Arowujolu, et al. 2002).
Other studies are limited to urban Nigeria leaving off the rural areas where an estimated
65% of the population lives (Smith, 2004b; Makinwa-Adebusoye, 1992; Feyisetan and
Pebley, 1989). Other studies are constrained by focus on ever-married women or unmarried female adolescents, entirely leaving off the male gender (Isiugo-Abainihe and
Oyediran, 2004; Ajuwon et al. 2002). Such efforts fail to incorporate the increased recognition of the need to involve men in sexual and reproductive health initiatives, as well as understand their needs, perceptions and motivations in reproductive health matters. The importance of involving men follows the social reality of daily living in the
African context, where considerable authority and power is vested on men generally as decision makers in the home and society (Isiugo-Abanihe, 2003). Reports from countries in the region indicate that the neglect of men/boys in this context is a costly omission. For instance, despite age asymmetry among sexual partners in the region (Luke, 2003), the
Kenyan Ministry of Health reports that school boys are responsible for most school girls’ pregnancies in that country (Bledsoe and Cohen, 1993). Further, recent studies from a spectrum of developing countries including South Africa, Nigeria, and Cameroon suggest the vulnerability of adolescent men to some of the problems faced by young women such as sexual coercion, unwanted sexual touch, penetrative sex and being “rented” as prostitutes by older men and women (Ganju et al. 2004; Jejeebhoy, S.J. and S. Bott.
2003; Ajuwon, A., 2003).
Further, very little national level research on migration has been undertaken in
Nigeria and none have simultaneously addressed premarital sexual behavior of men and
10 women. In particular, research on the direct role of migration on premarital reproductive behavior among adolescents has received scant attention and no national analysis to the best of my knowledge has been undertaken. Consequently most of what we know about the relationships between adolescent migration and reproductive behavior, as well as the social and behavioral mechanisms underlying these linkages is based on small-scale studies on small geographical areas and distinct local groups (Aja-Nwachuku, 2004;
Oladele and William, 1994; Smith, 2000, 2003, 2004ab).
This dissertation addresses most of the identified gaps, using data from the 1999
and 2003 nationally representative Demographic and Health Surveys, with samples of
school and non-school men and women, aged 15-24, from all the regions of the country
including rural and urban areas. The study specifically examines the relationships between adolescent migration status and the prevalence and timing of first premarital sexual intercourse, accounting for the roles of individual, household and community factors. The study further addresses other important dimensions of adolescent premarital reproductive perceptions and behaviors. In particular the study examines the independent
association between migration status and adolescent use or non-use of condoms at first
premarital sex, perceptions of risks of sexually transmitted infection (STIs), and being
sexually active before marriage. In examining these sub-themes among those who have
initiated premarital sex, this study seeks to shed new light on the levels of risk associated
with premarital sexuality and the extent to which models of understanding sexual debut
predict other potentially problematic adolescent sexual attitudes and behaviors.
11
1.4 Primary Objectives
The specific objectives of this present study are:
To identify the prevalence and timing of premarital sexual debut and how it varies
among adolescent migrants and non-migrants;
To assess the extent to which adolescent premarital sexual debut varies in urban
and rural areas, among men and women, and Nigeria’s heterogeneous ethnic
groups;
To identify other individual, family and community variables that significantly
predict premarital sexual initiation;
To shed light on the levels of risk associated with premarital sexuality and
examine whether the predictors of premarital sexual debut predict other risky
premarital sexual perceptions and behaviors: adolescent use or non-use of condom
at premarital sexual debut, the perception of risk of sexually transmitted infections
including HIV/AIDS, and adolescent transition from premarital sexual initiation
to being sexually active.
1.5 Significance of Study
The focus on migration and premarital sexual initiation is predicated on the fact that the advent of AIDS and AIDS-related studies have confirmed that HIV/AIDS infection specifically involves migration and some attempts to limit the spread of the disease in the last two decades already involve the notion of regulating migration and the living conditions of migrants (Lurie, 2004; Evans, 1987). Understanding the role of migration in adolescent sexual behavior in Nigeria creates a linkage not often emphasized but
12 necessary for widening the scope of control of the epidemic to include the streams of population movement in the country and the forces that generate them at places of origin.
Further, the study generally finds relevance in providing an evidence base to enrich behavior change programs. The particular focus on timing of premarital sexual debut and other potentially risky reproductive behaviors is relevant in identifying important characteristics of youths who abstain and those who initiate premarital sex.
Heterosexual contact represents the primary mode of HIV/AIDS transmission in sub-
Saharan Africa (Caldwell and Caldwell, 1993; Langstone, 1989; Isiugo-Abanihe, 2003;
Orubuloye et al. 1991), and the consensus among researchers is that promoting positive behavior, particularly delaying sexual debut, remains a key to addressing significant hindrances to adolescent successful transition to adulthood, including stemming the spread of sexually transmitted diseases and HIV (Caldwell, 2000; UNAIDS, 2000;
Cleland and Ferry, 1995; van der Straten, et al. 1995; Konde-Lule, 1991). This dissertation holds the potential of providing new and nationally representative evidence that will inform such behavior change policies and programs in Nigeria.
While modern sexual values continue to gain influence in the face of sweeping social changes, traditional norms in most Nigerian cultures still place emphasis on the importance of premarital sexual abstinence (Feyisetan and Pebley, 1989). Marriage is a normatively valued rite of passage among all Nigerian groups and premarital sex remains negatively sanctioned. This project finds significance in its promise of relevance to behavior change congruent with norms and values in Nigeria around premarital sexual chastity (Feyisetan and Pebley, 1989). These research themes and programs they will
13 engender are most likely to elicit a wider consensus and support among the country’s diverse sub-groups.
Finally this dissertation makes a major contribution to literature on the trajectory from premarital sexual initiation to other risky reproductive attitudes and behaviors. It hopes to generate empirical evidence to support or refute the perspective that premarital sexual initiation is part of a pattern of other risky behaviors, and offers a unique opportunity to examine the extent to which other adolescent problem behaviors can be understood using models of etiology of sexual debut (French and Dishion, 2003).
1.6. Organization of the dissertation
The present document is organized into eight chapters. Chapter 1 presents the problem statement, a background on Nigeria and the key subjects of the study, objectives of the study and expected contribution of the study to theoretical and empirical knowledge.
Chapter 2 reviews relevant theoretical and empirical literature that justifies and explains the expected relationships between migration and premarital sexual behaviors.
Individual, household and community/national level theories, together with empirical studies reviewed, inform a multifactor conceptual framework that guide the study and key hypotheses that are examined. Chapter 3 presents an overview of data sets used for the analyses, definition of variables, and methods adopted for statistical analyses. The limitations of the study conclude this chapter. The following four chapters present the results of the analyses of the role of migration on premarital sexual initiation and other risky premarital reproductive behaviors. Chapter 4 presents the characteristics of the study population, results of the descriptive and discrete-time hazard analysis of the
14 association between migration and premarital sexual debut. Chapters 5, 6 and 7 examine risky reproductive attitudes and behaviors among adolescents who have initiated premarital intercourse. Chapter 5 introduces condom use at sexual initiation and presents the results of the descriptive and logistic models predicting adolescent use of condom at sexual debut. Chapter 6 contextualizes risk perception and presents descriptive and multivariate regression results predicting perception of risk of sexually transmitted infections. Chapter 7 examines being sexually active and presents the bivariate and multivariate predictors of transition to being sexually active. Finally, Chapter 8 presents a summary of the study, discuss its key findings and their policy implications and the direction of future research.
CHAPTER 2
THEORY, LITERATURE AND HYPOTHESES
2.0 Introduction
Theoretical and empirical research has associated migrants with risky heterosexual
behavior as compared to non-migrants. This difference in behavior is attributed to three
broad factors summarized by Brockerhoff and Biddlecom (1999) as: “1) predisposing
individual characteristics; 2) changes in individual attributes due to migration, notably
separation from a spouse or partner; and 3) exposure to a new social environment,
featuring different sexual norms, opportunities and constraints that result in behavioral
modification.” These three perspectives essentially derive from the concepts used by
demographers in fertility studies to account for fertility differences between migrants and
non-migrants: selectivity of migration, life disruptions associated with moves, and
migrants’ adaptation to life norms in places of destination (Goldstein and Goldstein,
1983; Lee and Faber, 1984; Hervitz, 1985). In brief, the selectivity hypothesis posits that migrants are not randomly selected, reflecting the entire population, but are typically selective of personal characteristics such as higher education, young age, and desire for upward social mobility. They also possess distinct marital and occupation characteristics.
The disruption hypothesis usually associates the period immediately following migration with both physiological stress due to moving and the separation from spouses or significant others. The adaptation principle proposes that migrants adapt to the new economic, social and cultural environment at the places of destination, resulting in behavioral changes. These concepts are closely related to theories and empirical research
15 16 findings that cover individual, family, and community level explanations reviewed in detail under the following sub-sections.
2.1 Individual Level Theories
The Health Belief Model (HBM) is a psychological model that attempts to explain and predict health behaviors by focusing on the attitudes and beliefs of individuals. The model was developed in the 1950s as part of an effort by social psychologists in the
United States Public Health Service to explain the lack of public participation in health screening and prevention programs. Over the years since its formulation, the HBM has been adapted to explore a variety of health and contraceptive behaviors, including sexual risk behaviors and the transmission of HIV/AIDS in diverse populations (Janz and
Becker, 1984). One of the areas in which the model has been adapted is in the study of the relationship between migration and reproductive behavior. The core postulate of the model is the assumption that: “an individual’s characteristics, prior experiences and current surroundings shape his or her perceptions concerning the risks and severity of behavioral outcomes- such as contracting the AIDS virus through sexual conduct – and thereby influence behavior”(Brockerhoff and Biddlecom, 1999)1. Consistent with this
model is the perspective that the act of voluntary movement - often over long distances
between radically different sociocultural environments and with uncertain consequences
and social support networks at destinations - defines migrants, to a greater or lesser
degree as innovators or risk-takers (Peterson, 1958; Massey et al. 1994). Further,
econometric studies of migrant behavior routinely attribute differential outcomes to
migrants and non-migrants’ unobservable risk tendencies or characteristics (Moreno,
1 (For details of key themes of model, see Rosenstock et al. 1994; Bandura, 1989).
17
1994; Rosenzweig and Wolpin, 1988). Consequently, migrants may be predisposed towards heightened risk-taking behavior including riskier sexual behavior than others in their new social settings, owing to personal traits established before migration
(Brockerhoff and Biddlecom, 1999).
The role of individual characteristics of migrants and their consequent sexual behavior is consistent with the migration selectivity hypothesis identified above.
Empirical evidence supports the view that migration to economic centers in East and
Southern Africa during colonial and the apartheid government involved young men aged between 15 and 30. Though migration composition has significantly changed since the end of colonial rule to include substantial proportion of women (Hugo, 1993; Peil, 1985), the overall conclusion in the region is that rural-urban migration is selective of the young, the educated, the innovative and the energetic, leaving behind the very young, the apathetic, the retired and the tired, the illiterate and the infirmed (Makinwa, 1981,
Adepoju, 1983, Oucho, 1998). Research findings have linked imbalances in sex and age ratios at places of destination to largely a product of net migration, and suggested its particularly marked effects on migrants’ sexual practices (Quinn, 1994). Hunt (1989) describes specific instances where the forced concentration of male migrant workers in rural extraction industries inevitably fed demand for migrant female sex workers, with catastrophic results for the spread of veneral diseases in migrant-populated areas. An identical process based on the continued imbalanced age/sex ratio of migrant laborers has been cited for the spread of AIDS in contemporary African cities (Decosas and
Pedneault, 1992; Larson, 1989; Lurie, 2004).
18
One major limitation of individual characteristics models is the scant attention to broader environmental and economic forces that may influence individual behavior. The models do not adequately incorporate social and cultural norms and peer influences on people’s decisions regarding their health-related behaviors particularly sexual behavior.
Despite the recognition of contextual factors in HBM for instance, their linkage to the broader social structure and the important interactions that exist between them are very limited. In fact, structural variables in the model are limited to those that affect individual perception and therefore indirectly influence health-related behavior. In contrast, sociological explanations go beyond micro-level analysis to include meso and macro level explanations, thereby avoiding the potential bias of “individualistic fallacy”, which occurs when population patterns of outcomes of interest are erroneously presumed to be explained only by individual-level characteristics (Krieger et al. 1997). Further, evaluations of interventions based on the HBM have shown consistent disappointing effects on risky behavior, giving strength to criticism that the individual is an inadequate unit of analysis (Auerbach et al. 1994).
2.2 Family and Social Environmental Theories
One relevant theoretical model used in the study of the relationship between migration and premarital sexual behavior at the family level is the social control theory.
This model posits that ties to conventional institutions lower the probability of deviant behavior, and that a stake in conformity to behavioral norms can be enhanced through bonds to school, family and career (Liska, 1981; Vold and Bernard, 1986). Researchers within this framework show that the process of migration weakens such bonds and
19 increases the probability of deviance (Liska, 1981, Sampson and Laub, 1993; Vold and
Bernard, 1986; Stack, 1994). Accordingly, migration can weaken bonds to familiar institutions such as schools and significant others who would normally act as mechanisms of control (Sampson and Laub, 1993). Long distance migration can sever ties to the extended family and the influence of older kin who are apt to provide negative definitions and negative reinforcements of premarital sexual activity (Clayton and Bokemeir, 1980).
Further, the supply or amount of parental emotional resources for control can be weakened by moving. Researchers suggest that there is considerably more psychological depression, feelings of loss, frustration and anxiety among migrants than non-migrants, especially among mothers (Fichter et al. 1988; McCollum, 1990). In a longitudinal study in New England, United States, McCollum (1990) shows that after two years of moving; only 17 percent of his sample of mothers had re-established deep friendship in their new community. New migrants are generally located at the fringe of social circles and have to continually seek for acceptance into the inner circle by proving themselves in a variety of ways. The preoccupation of parents with the challenges of the resettlement process weakens their supervision of their children, thereby increasing the influence of peers who tend to be more permissive (Reiss and Miller, 1979). Given the many stressors placed on women who move, migration also is predicted to lead to low maternal supervision of children (Sampson and Laub, 1993). To the extent that women are the primary socializers of children, and with researchers suggesting that the influence of mothers outweighs that of fathers as it relates to adolescent sexual behavior (Blum, 2002), the burdens of resettlement can have an important impact on overall social control (Stack, 1994).
20
The demand or need to control may increase with migration. According to
Sampson and Laub (1993) moving tends to lower children’s attachment to parents and increases parental rejection of children. In turn estrangement between parents and children is linked to relatively high probability for premarital sexual activity (Davidson and Leslie, 1977). Given the difficulties in establishing friendships in their new environment, children of migrants are prone to loneliness and might use sexuality as a vehicle to hasten friendship formation, or as a means to combat loneliness (Stack, 1994).
Evidence from the United States confirms that indicators of social isolation contribute to premarital sex (Davidson and Leslie, 1977).
Closely related to the social control model is the social support theory, which lends further credence to the perspective that longstanding social ties are broken when migrants leave their home communities (Portes, 1998). As a result, migrants lose a major source of social capital –the network of social relations that entail reciprocal commitment, social support and social control (Zhou, 1997). Coleman (1990) has insisted that social capital is dependent on stable social relationships and migration reduces both the size of the social network and the level of support potentially available through it
(Vega and Amaro, 1994). Overall these arguments suggest that migrant adolescents will have lower levels of social support in their places of destinations relative to non-migrants in their places of origin (Landsdale and Oropesa, 2001).
Further, the social support theory addresses what happens to social support and control as migrants assimilate into host societies in urban or rural destinations, offering a competing explanation on the consequences of migration for social disruption and therefore sexual behavior. Researchers have frequently indicated strong family bonds
21 among immigrant groups, and migration as a family-related process that is often undertaken either by family groups or by individuals joining family members in the destination area (Landsdale and Oropesa, 2001). In developed and developing countries, there is ample evidence that family connections form the backbone of the networks that link origin and destination areas and facilitate migration (Rambault, 1997; Fix and
Zimmerman. 1997; Watts, 1984; Nigeria Institute for Social and Economic Research,
1997). The stress of relocation and establishing a new life in the destination area is associated with motivation for family cohesion. It has been noted that the shared adversity of migration and subsequent adaptation experiences can motivate the family to close ranks cohesively and productively, honing in group solidarity and adaptive culture of efforts and efficiency (Rambault, 1997).
Another explanation for the likelihood of strong family ties among migrants is related to the cultural orientation from origin areas carried over to places of destination, which often emphasize the primacy of the family commitment (Oropesa and Gorman,
2000; Rambault and Weeks, 1996; Sherraden and Barrera, 1997; Zambrana et al. 1997).
Related to the above is the recognition that migration disrupts the family system that was in place at the origin, making migrants to feel especially the lack of sustaining social ties and the hope that as new family ties are established over time and in subsequent generations via marriage and childbearing, there may be a return to higher levels of social support. Under these circumstances the loss of social capital associated with migration and its implications for sexual behavior may be mitigated if migration involves families rather than individuals or if migration involves some living arrangements that will promote group solidarity. This expectation underscores a key limitation of these
22 perspectives, particularly when applied to cultural frontiers outside the United States, where they have been widely applied and validated.
Though the core themes of the social control/support perspectives are consistent with the migration disruption and adaptation hypotheses employed in explaining the role of migration in reducing migrant fertility, their validity in a culturally different African context such as Nigeria is not well understood. Evidence from Africa generally and
Nigeria in particular, suggests that migration takes place within family and community networks, affording migrants some of the supports they need to adjust to life in a new environment. And for adolescents, moves are often with family members or close kin as business apprentices or house helps (Chukwuezi, 2001; Smith, 1999). It follows therefore that migration may not create high levels of social disruption and anonymity postulated for developed societies. The logical expectation is that living arrangements of adolescent migrants at destinations and the structure of migrant households may moderate the relationship between migration and adolescent premarital sexual behavior. If an adolescent migrant moved with a family group or if an adolescent joins a relative and lives within an enclave of his migrant group, the potential social disruption generally linked to migration may be less likely when compared to situations where an adolescent migrates alone or lives independent of his family or outside the enclave of migrants from place of origin. This study will further address this outcome by examining how the living arrangements of adolescent migrants and non migrants differ relative to their sexual behavior outcomes.
23
2.3 The Life Course Perspective
Despite their invaluable contributions to our understanding of the influences related to the immediate social environment of the migrant, particularly at the family and kin levels, the social control and social support theories did not address broader society-level influences that are proven predictors of behavior. This gap informs the review of the life-course theoretical perspective, whose core postulations incorporate broader societal level explanations. The life course perspective as a paradigm refers to the framework of a set of interrelated presuppositions, concepts, and methods that are used to study age-graded, socially embedded roles (Elder et al. 2003). The central premise of the paradigm is that no period of life can be understood in isolation from people’s prior experiences, and their aspirations for the future. The perspective emphasizes the inter-locking and sequencing of events in the life of different age cohorts, linking modernization of societies as the underlying process that drives long-term trends and differentiate successive cohorts. In relation to adolescents, the life course framework emphasizes the study of their biological, cognitive, and emotional development “at the nexus of social pathways, developmental trajectories, and social change” (Elder, et al. 2003). Consequently the study of adolescent behavior takes into consideration not only individual adolescent characteristics but also those of his particular context - including his peers, his family, his school, and his church, his community, his national and global environment. While the previous reviewed theories relate adolescent sexual behavior to rational action based on adolescents weighing of the pros and cons of their actions, influenced by their immediate social environments, the life-course perspective promotes a holistic understanding of adolescents’ lives over time and across changing social contexts. In sum, the Health
24
Belief Model posits that individual characteristics select migrants and risky sexual behaviors. The Social Control and Social Support Models emphasize that disconnections to social control and support mechanisms at the places of origin predispose migrant adolescents to deviant sexual behavior. But the life-course perspective extends these explanations to include the social structural hindrances that not only deny access to knowledge through lack of education, but also the social inhibitions like poverty and powerlessness that leaves adolescents no option than to engage in sexual activities. It also links to unemployment in rural areas that occasion adolescent migration for employment in urban settings, where they may lack parental supervision, but have access to personal money to finance personal preferences including sexual exploration. The perspective points to the globalized world of culture and communication that transmits sexual stimuli worldwide. It also brings into focus the economic stipulations of the World Bank that helped introduce Structural Adjustment Programs (SAP) in many developing countries that is linked to economic downturns in many urban centers, with implications for unemployment and urbanization of poverty. These structural forces exacerbate deviant behavior among youths, which may include transactional sex and drug abuse. In sum, the life course perspective help us view adolescent reproductive behavior less as a discrete set of experiences, but more as an integral part of a biography that reflects the early experiences of youth and also that shapes the later life (Shanahan, 2000). It not only incorporates the view that adolescents exercise agency in the construction of their biographies, but also that national and global forces beyond the rational calculations of individuals are important influences in shaping behaviors.
25
The major challenge of the life-course perspective is the availability of data to test its major postulations. Since migration is a process, migrant behavior should also be considered as a process and the life course perspective is adequate for such analysis.
However, despite this recognition, most studies compare migrants and non-migrants at certain points in time, while hardly any explore migrant behavior over time. Yet, this limitation is hard to overcome, even in this study, as studies that track migration and behavior as processes must attempt to follow individuals over their life course. This requires study designs that have a longitudinal data gathering framework, which is yet to be available in many developing societies. Consequently, most studies, including this project, utilize and adapt available cross-sectional data sets in the study of migration and behavior despite known limitations.
2.4 Premarital Sexual Behavior: Review of Empirical Research in SSA
Beyond migration, a growing body of studies in sub-Saharan Africa has identified significant independent predictors of adolescent sexual behavior. Researchers have shown that as chronological age increases, the likelihood of premarital sexual initiation increases as well (Gage-Brandon and Meekers, 1993; Meekers, 1994; Slap et al. 2003).
Available explanations point to the fact that older adolescents have more exposure to the opposite sex with increased pressure associated with biological and social maturity
(Kiragu and Zabin, 1993).
Gender, particularly being male, has been identified as a significant predictor of premarital sexual activities in the region. Mensch et al. (2001) argue that boys have
26 relatively few hazards and many potential benefits, including gratification and social prestige from engaging in early/premarital sexual experience. In fact, Brockerhoff and
Biddlecom (1999) argue that underlying the relationship between migration and high risk sexual behavior are substantial gender differences in sexual behavior; namely having multiple sex partners is far more predominant among men than among women. Men have strong cultural prerogative and biological drive over women in initiating and negotiating sexual relationships, but economic rationale usually apply to women. However, other
findings contradict this proposition with emphasis that girls are more exposed to early
sexual activities than boys. Researchers in Kenya point to the vulnerability of girls to
escalating cases of rape and defilement, which is linked to the high incidence of HIV
among women in the country. The mechanism that produces such outcomes is identified
as the social contexts in which girls mature into puberty earlier than boys, and often have
older sexual partners than boys (Konde-Lule et al. 1997). Luke (2003) concludes that
adolescent girls’ sexual relationships with older partners that involve economic
transactions are common, and that receipt of financial benefits is a major motivation for
such relationships.
Family economic situation is linked to the sexual behavior of adolescents. The
stagnant - if not dwindling - economic fortunes of countries in sub-Saharan Africa, including Nigeria, through the 1980s are estimated to have resulted in widespread poverty. Desperate survival strategies and coping mechanisms are linked not only to weakened moral values that moderate sexual behavior, but also aggravated several adolescent antisocial activities, which include indiscriminate and clandestine sexual involvements (Isiugo-Abanihe and Oyediran, 2004). The most vulnerable female
27 adolescents have to contend with the allure of financial gratifications and sexual overtures by relatively richer peers and adults. Consequently, poverty or the need to survive became a driving force and motivation behind the sexual activity of adolescent girls (Isiugo-Abanihe, 1993). Much of the evidence in sub-Saharan Africa suggests that many secondary and university students engage in premarital sexual relations to meet basic needs, such as school fees and supplies, as well as for economic security (Luke,
2003). The primary gap in these assessments, however, is the focus on young women, while very little attention is directed at understanding the circumstances of young men.
Notwithstanding, it is plausible to expect that the meaning of and propensity to engage in premarital sexual activities by adolescents is likely to be influenced by the socioeconomic status of their households, and mediated by gender.
Education has been identified as relevant in creating an enabling environment for adolescent sexual debut. The increased exposure hypothesis suggests that being enrolled in school and prolonged years of schooling increases the risk of premarital sexual engagement. In the same light, rising education is linked to limiting family control of young people’s behavior and exposure to peer influence on sexual behavior. Feyisetan and Pebley (1989) identified education as a measure of exposure to modern society and identified women who completed higher levels of schooling to be more independent of their families in determining their own behavior, with some young women moving away from their parents’ home to pursue secondary and post secondary education. In the same token, Kaufman et al. (2002) suggest that schools may be a locus of sexual harassment and an environment providing opportunities for sexual contact. On the other hand, human capital and role incompatibility hypotheses argue to the contrary, that women who are
28 enrolled in school are at a lower risk of initiating first sex following, not only the opportunity costs associated with potential unplanned pregnancy, but also the role conflict associated with early motherhood and studentship. Schools provide supervised time, thereby offering students, especially girls, an alternative to higher-risk activities available outside the school yard gates. Education may also bring knowledge of disease transmission and risks and lead to safer behaviors. These assessments are limited not only by the focus on school youths without comparative non-school youths that are quantitatively more important in Nigeria, but also by an over emphasis on young women.
This dissertation will seek to bring more clarity on the relationship between educational attainment and adolescent premarital reproductive behaviors by simultaneously examining school and non-school youths, as well as young men and women.
Related to education is the availability and nature of adolescent employment, which may offer independence and greater sexual access in places where the supervision of elders can no longer reach. Urbanization is particularly implicated, as it provides employment possibilities, which typically draw young men and women away from parental authority as wage workers, apprentices and self employed hawkers and hustlers.
At the nodes of the transport system and traffic bottlenecks are the numerous teenage boys and girls who load lorries, entice passengers, collect fares and sell anything: phone cards, bottled water, mousetraps, peanuts, clothes, etc. (Bledsoe and Cohen, 1993;
NISER, 1997). For young girls, this may substantially intensify their exposure to sexual demands, and for young men it offers exposure to the world of sex workers, who offer sexual services for a living, and resources to finance sexual exploration. The flip side of adolescent employment opportunities is unemployment. In Western Nigeria, an increase
29 in sexual expression among youths is attributed to one effect of unemployment of youths- youths with “nothing to do” (Bledsoe and Cohen, 1993). This study offers a unique opportunity to use current national data to compare the relationships between employed and unemployed adolescents and reproductive behavior.
Islamic religion is associated with a consistent emphasis on premarital virginity and
Muslims are less likely to engage in premarital sexual activities when compared to
Christians (Feyisetan and Pebley, 1989; Meekers, 1994). However, Slap et al. (2003) found no significant difference between Muslims and other religious groups in their study of secondary schools in Plateau State of Nigeria. Rather they found the extent of religiosity to be inversely related to premarital sexual initiation. The role of religiosity finds support in earlier works among school age adolescents in Kenya (Kiragu and Zabin,
1993) and adolescent women in Northeast Brazil (Gupta, 2000). Following the limited coverage of these studies, and exclusion of male adolescents, these results may be regarded with caution. However, this gap will be addressed in this current study and a more comprehensive result is expected. Notwithstanding the available data will only enable an examination of religious affiliation but lacks measures of religiosity.
Cultural values regarding sexuality and gender roles exert powerful influences on the decision making process. Through the enforcement of positive or negative sanctions, and by altering the individual’s assessment of risk, social groups can affect an individual’s assessment of the relative costs and benefits of engaging in a particular sexual activity.
The cultural norms governing premarital sexual activity, and the availability of social and economic support networks to unmarried mothers, as well as the consequences of unsanctioned childbearing, often strongly influence adolescents’ likelihood of engaging
30 in premarital intercourse (Gage, 1998). Among the Fulani of Cameroon for instance, premarital sexual relations of men is taken for granted and there is no expectation that young men remain celibate before marriage, nor is any value placed upon the fidelity of married men particularly when away from home (Bledsoe and Cohen, 1993). Similarly,
Hollos and Leis (1989) observed that among the Ijo of the Niger Delta of Nigeria, parents have no objection to their sons’ sexual exploits and usually welcome a child fathered by him as a member of the family. They pay the circumcision fees and buy the appropriate gifts for girls’ circumcision, thereby affiliating the child to their patrilineage. Examining how adolescents from all cultural regions of Nigeria relate to premarital sexual activities offers a unique opportunity to empirically understand variations in cultural attitudes and behaviors in the country.
Urban and rural environments are associated with different influences on adolescent reproductive behavior. There is however no consistent and clear evidence as to direction of this relationship (Kiragu and Zabin, 1993). Evidence from North East Brazil supports this indiscernible relationship between current residence and premarital sexual initiation, but pointed out that place of childhood residence have significant effects (Gupta, 2000).
One important finding in empirical literature is the evidence that adolescent premarital sex is part of a pattern of similar risky behaviors. Consequently, French and
Dishion (2003) suggest that precocious sexual initiation can be understood using models of the etiology of other problem behavior and identified deviant peer involvement as a salient dimension of this trajectory. Mugisha et al. (2003) found that the slum environment in Nairobi, Kenya breeds a group of other adolescent problem behaviors: alcohol, substance and drug abuse. Examining the extent to which this proposition is true
31 of Nigeria forms the second part of this study. Specifically the study will examine the trajectory from sexual initiation to other potentially problematic reproductive behaviors, and will use the same models for the study of sexual debut to seek understanding of these risky/protective reproductive behaviors in the country.
Despite the limitations of most of the studies reviewed (limited coverage of most studies in Nigeria and the implications of different cultural and institutional settings of those done outside Nigeria), they offer relevant insights into the multi-factor predictors of adolescent sexual behavior. The extent to which empirical data will support the independent effects of these variables in Nigeria is part of the primary goal of this study.
Though available Nigerian data would not allow for an examination of all the predictors reviewed, this study builds on the multi-factor approach revealed by this review to seek advancement of our knowledge of adolescent reproductive behaviors beyond the limitations of previous studies.
2.5 Conceptual Framework
The theoretical and empirical literature reviewed above informs a multi-factor conceptual framework to guide this study. The framework, presented in Figure 1, recognizes the principle of human agency, the roles of family context and their interlocking with community, national and global factors, in shaping the behavior of adolescent sexual behavior. The framework incorporates the perspective that global/national forces such as globalization, westernization, and national policies and programs influence the characteristics of places of origin and adolescent family and individual characteristics, which in turn influence migration to urban areas or other economic centers by an
32 adolescent or a family. The framework shows that migration exposes the adolescent to a
Figure 1. Conceptual Model of Migration and Adolescent Premarital Reproductive
Behavior.
Global/National Forces: Global economic policies/programs Global/national mass media Behavior Change Programs/ health awareness campaigns, etc
Community Characteristics Individual Characteristics (Pre-migration): (Post-migration): Premarital Rural/Urban Age, Gender, Education, Reproductive Economic Opportunities Religion, Status/Nature of Behaviors Migration History Employment, Ethnicity
Household Characteristics (Pre-migration) Household Characteristics Household Structure (Post-migration): Household SES Household Structure Household SES
Individual Characteristics (pre-migration): Migrant’s New Social Environment: Age, Gender, Education, Age/Sex Structure Religion, Ethnicity Sexual permissiveness Social Control/Support Networks, Urban/Rural, Exposure to Global Migration to a New media/culture and Modernization Destination
new social environment characterized by different age and sex structures, sexual norms that may border on permissiveness, autonomy or maturity, demand and supply of new mechanisms of social support and control, different status and nature of employment, and
33 various forms and level of exposure to the media, modernization and global forces. These new environment influences post-migration family characteristics including household structure and socioeconomic status, which in turn affect individual adolescent characteristics, and subsequently premarital sexual behavior.
The framework recognized other important pathways to adolescent sexual behavior. For instance global and national policies and programs and community level characteristics can directly influence adolescent post-migration characteristics, which will in turn affect adolescent post-migration family and individual characteristics, which again influence premarital sexual behavior. Accordingly, a community with a long history of out-migration to a particular destination will influence the level of human capital available for social support and control for a new migrant to such a destination. This is expected to mediate the anonymity and social disruptions associated with migration and will in turn influence premarital sexual activities.
2.6 Statement of Preliminary Hypotheses
Although my framework is comprehensive, I am not able to analyze all the relationships it embodies due to limitations of data. The study will therefore concentrate on the analysis of the following preliminary hypotheses:
That generally, migrant adolescents are more likely to initiate early premarital sex
than non-migrants. This expectation follows the general disruptions and
challenges associated with the migration experience.
34
That the relationship between migration and adolescent reproductive behavior is
mediated by gender, with female adolescents more likely to initiate sexual
activities earlier than males.
That adolescents with higher education are more likely to engage in premarital
sexual activities relative to others. This assumption is predicated on their being
more likely to have exposure to the opposite sex, be more physically and socially
mature, and delay marriage long after puberty.
That household poverty is an independent risk factor for adolescent premarital
sexual engagement, with adolescents from poor households more likely to initiate
premarital sex than those from rich households.
That adolescent current living arrangement (household structure) is a significant
predictor of adolescent premarital reproductive behavior, with transition to
independent living more likely to promote premarital sexual relationship than
living with parents or relatives.
That employment in the formal sector is a significant independent risk factor in
adolescent premarital reproductive activities.
Following their perceived conservative cultural norms and values, that
adolescents from Hausa/Fulani/Kanuri ethnic groups and Muslims are expected to
less likely engage in premarital sexual relations compared to other ethnic and
religious groups in Nigeria.
That migration status and other predictors of adolescent sexual debut are
significant predictors of adolescent use of condoms at premarital first intercourse,
perception of risk of STIs and transition to active premarital sexual behavior.
CHAPTER 3
DATA AND METHODS
3.1 Data
This study uses data from the1999 and 2003 Nigeria Demographic and Health Surveys
(NDHS), whose samples were designed to provide estimates of population and health indicators for urban and rural areas and all the regions of the country.
The 1999 NDHS sample was designed as a probability unbiased nationally representative sample of eligible respondents within all regular households in the entire country. The sampling frame used for selecting the Primary Sampling Units (PSUs) was the Enumeration Areas (EAs) into which the country was delineated for the 1991
National Population Census. The frame contains 212,079 EAs that are mutually exclusive and collectively exhaustive of the territorial land area of Nigeria. The EAs were classified into one of the two strata - (i) rural and (ii) urban, where urban EA (U) is defined as an
EA within a locality having population of 20,000 and above, while rural EA (R) is an EA within a locality with population less than 20,000 persons. After arranging the EAs in each stratum of a state in their geographic order, the PSUs allocated to the stratum in the state were selected by systematic sampling procedure. Selection was therefore independent both within stratum and between states. The second stage sampling of households and eligible women followed. All eligible women (EWu) within the selected sample units (households) are included in the sample. That is, Pr (EWIj\SUij) = 1.
Whereas for the eligible men's (EMij) interview, a sub-sample (n~), one in every three selected sample units (households) selected for women's questionnaire is systematically selected. In all, the 1999 NDHS interviewed 8,199 females aged 10-49 years and 2,680
35 36 males aged 15-59. Out of this number a total 3,793 adolescent men and women aged 15-
24 were drawn into the study.
The 2003 NDHS sampling design follows the same procedure as the 1999 survey but focusing on all women age 15-49 and all men age 15-59. A probability sample of households was selected and all women age 15-49 identified in the households were eligible to be interviewed. In addition, in a sub-sample of one-third of the households selected for the survey, all men age 15-59 were eligible to be interviewed. The sample frame for the survey was as in the 1999 survey the list of approximately 212,080 enumeration areas (EAs) developed for the 1991 Nigeria Population Census. The 2003
NDHS sample was also selected using a stratified two-stage cluster design and a total of
365 clusters were selected, 165 in urban and 200 in rural areas. The 2003 survey
interviewed 7,620 women aged 15-49 years and 2,346 men aged 15-59. From the male
and female data files, 4,006 adolescent men and women aged 15-24 were selected into
this study. Together the two surveys yielded a study population of 7,799 men and women
aged 15 – 24 who have complete information on key variables regarding migration and
premarital sexual behaviors. Among the overall study population, 4,956 adolescents
(63.5%) have never married, while the complementary 36.5% are either currently or
previously married.
The surveys used household questionnaires to obtain information on housing
characteristics, living facilities and household composition. Individual questionnaires for
men and women provided information on current and previous places of residence,
duration of stay in current residence, respondent’s age, education, religion, ethnicity,
nature of employment, reproductive behavior and intentions, knowledge and use of
37 contraception, breastfeeding, marriage, knowledge of and perception about AIDS, etc.
From the information gathered, variables relevant for this study were selected and defined: migration status, premarital sexual activities and control variables.
3.2 Methods
3.2.1 Definition of Variables
The primary goal of this study is to determine the prevalence and timing of initiation of premarital intercourse by adolescents aged 15-24 at the time the survey. The comparative group is those of the same age bracket and characteristics who have not initiated sexual activities or who did so at marriage. Male and female questionnaires used for the 1999 and 2003 surveys asked respondents about their age at first sex. The response categories were: 0 for not had intercourse; real age at first intercourse for those who had initiated sex, and 96 for intercourse at first union. From this category of responses those who initiated premarital sex relative to those who did not were determined. To account for exposure time and censored cases, I further restructured the data set following the event history framework, into person-year format, in which each adolescent contributes one record each year he/she is exposed to the risk of premarital sexual initiation. The retrospective nature of the survey enables the observation period for premarital sexual initiation from ages 10 - 24. The periods before premarital sexual initiation are interpreted as survival time until premarital sex is initiated. Adolescents are right censored if they never initiated premarital sex throughout the observation period or if they got married before sexual initiation. Adolescents who initiated premarital sex and got married in the same year, are censored a year prior both events. The outcome
38 variable is treated as a series of dummies at each age and coded 1 if the event occurs at a given age and 0 if otherwise. The series of data management operations yielded 59,089 person years that were analyzed for premarital sexual initiation.
In the second stage of the analysis, other potentially risky reproductive behaviors are examined among adolescents who have initiated premarital intercourse. The key outcome variables examined at this stage are condom use at first premarital intercourse, perception of risk of STI/AIDS infection, and being sexually active before marriage.
Adolescent premarital risky sexual behavior is further examined in terms of use and non- use of condoms at first premarital intercourse. The relevant question on condom use at first sex was only asked in the 2003 survey, and elicited a binary response category: Yes
= 1, No = 0. These outcomes were analyzed to identify the level of risk associated with adolescent premarital sexual debut in terms of condom protection. Perception of risk of sexually transmitted infection is defined as a binary outcome and intended to identify the characteristics of adolescents that have initiated premarital sex and yet deny their susceptibility to sexually-related infections and differentiate them from those who admitted susceptibility to sexually transmitted infections. The surveys asked respondents:
Do you think your chances of getting AIDS are small, moderate, great, or no risk at all?
The responses enabled a binary classification of those who perceived no risk at all relative to those who perceived some level of risk. Those who indicated that they had
AIDS were dropped in this analysis. Finally, transition to active premarital sexual engagement (defined as adolescents who have initiated premarital sex and had premarital sex in the three months or less prior to the survey) is intended to examine whether those who initiated premarital sex continue to have premarital sex afterwards. I defined this
39 outcome as binary, seeking to distinguish the characteristics of adolescents who remain sexually active from those who practice some level of secondary abstinence. The relevant questions asked men and women to indicate the last time they had sexual intercourse prior to the surveys. The responses, which are in days and months enables the distinction of sexually active adolescents from others.
The key independent variable is migration status. The two rounds of Nigeria
Demographic and Health Surveys have no direct questions on migration status of respondents. However there are questions on childhood place of residence before age 12, current place of residence, duration of stay in current residence and type of previous and current places of residence (rural or urban; city, town or countryside). Using responses to these questions (questions are available on request), the migrant status, origin and destinations of adolescents were determined. This operation yields six migrant categories:
Rural and Urban Non-migrants, Rural-Rural Migrants, Rural-Urban Migrants, Urban-
Urban Migrants and Urban-Rural Migrants. Non-migrants are adolescents whose childhood place of residence remains their current place of residence. Migrants are those whose current places of residence are outside their places of childhood residence prior to the survey.
To explore the multi-factor predictors of premarital sexual behavior, empirical and theoretical literature reviewed inform the selection of other predictor variables. In particular, the study examined the relationships between adolescent sexual behaviors and individual characteristics: current age, years of exposure to the risk of premarital sex, gender, educational attainment, and religious affiliation. The surveys provided information on age in single years and five-year age cohorts. For the analysis of
40 prevalence and timing of premarital sexual initiation the age variable is used to determine the number of years an adolescent is exposed to the risk of premarital sex. This variable takes account of not only the age of adolescents at the time of survey but also the censoring in the data. For the second part of the analyses among adolescents who have initiated premarital sex, age at survey is recoded into 5 age groups of two year gaps.
Gender is coded 0 if male and 1 if female. The surveys provided information on highest educational attainment of respondents, which for this analysis is recoded into four educational attainment categories: No education, primary education, secondary and higher education. The surveys provided information on religious affiliation of respondents in six response categories: Catholic, Protestant, Other Christian, Islam,
Traditionalist and Other. For this study religion was recoded into four categories:
Catholic, Protestant/Other Christians, Muslim and Traditional/Other. In examining the role of the nature/status of employment, adolescents were classified on the basis of the nature of independence from parental supervision which each employment confers: formal employment, self/agricultural employment, and the unemployed. Adolescent in formal employment are most likely to be economically independent and spatially separated from parents and relatives when compared to others. Unemployed youths on the other hand are most likely to depend on parents and relatives for sustenance and therefore more likely to reside with them.
The association between household variables and sexual behaviors is examined using living arrangements (household structure) and household socioeconomic status.
The surveys provided information on the relationship between adolescents and the head of households. Adolescents who are heads of households are defined as having transited
41 into independent living outside the authority and supervision of their parents. They are compared to adolescents who are living under their parents as children, or with a relative or a spouse. In determining household economic status, the 2003 Nigeria Demographic and Health Survey constructed wealth index and quintiles through principal components analysis of household possessions. However, the 1999 survey have no such index.
Following this approach (see also Montgomery, et al. 2000; Gwatkins et al. 2000; Filmer and Pritchett, 1998, 1999; Mberu, 2006, 2007), a household wealth index is constructed for the pooled 1999 and 2003 data sets. The factor scores generated for all adolescents are recoded into wealth quintiles.
The role of community-level variables in predicting adolescent sexual behavior is examined through place of current residence (rural or urban), place of childhood residence before age 12, and ethnic origin. While place of current residence is provided in the survey as urban or rural, place of childhood residence was provided in four response categories: “Capital, large city”, “City”, “Town”, and “Countryside”. The countryside was defined as rural following other studies in the region, and the complementary categories were defined as urban. To deal with the problem of multicollinearity, current place of residence was not included in the multivariate models. The surveys provided information on 112 ethnic groups in the country. To assess variations in adolescent sexual behavior among these diverse cultural groups, they were classified into six major cultural categories, approximating the nation’s geopolitical zones and following the Nigerian
Institute for Social and Economic Research’s categorization (1997). The three main groups (Hausa-Fulani/Kanuri, Yoruba and Igbo), represent the core North, South West and South East cultural regions respectively. The Minority tribes of Central Nigeria and
42 those of the Niger Delta were respectively grouped into two cultural zones. Finally the myriad of groups that could not fit into these five clear categories were grouped together as others. Initial exploratory analysis of data indicates high correlation between region of origin and ethnic origin. To avoid the problem of multicollinearity, the region-of-origin variable is dropped. This decision is taken because ethnicity is a variable for cultural identity and its implication for behavior is more important for this study than region of origin. Other control variables include the index of media exposure and year of survey.
The surveys asked whether respondents listen to radio every day, read newspaper and magazines once a week and watch television every week. Using the simple summation of frequency method, 1 for yes and 0 for no, I constructed three levels of exposure to the mass media. A critical examination of these questions indicates a design problem. One who did not listen to the radio everyday may listen every two days or once a week. So answering no to that question does not markedly distinguish those who listen from those who do not. The same applies to reading newspaper, magazines or watching television once a week. It is therefore doubtful whether this variable adequately delineates levels of adolescents’ media exposure.
It is important to point out that the variables operationalized are primarily guided by the first stage of the study, analyzing the relationship between migration status and premarital sexual debut. The same operational definitions applied to the second stage of the analysis with other variables particular for this stage of the analysis also identified and defined. In particular, in the analysis of condom use at first premarital intercourse, the defined age predictor is age at first intercourse. In examining condom use at premarital sexual initiation and perception of risk of sexually transmitted infections, the
43 relevant questions were only asked in the 2003 Nigeria DHS survey and were so analyzed.
3.2.2 Statistical methods and models
This study employs univariate (frequency tables), bivariate (cross-tabulations, chi-square tests of associations), Kaplan-Meier survival curves, discrete-time event history and binary logistic multivariate techniques for the analyses of data.
For premarital sexual debut, frequencies and cross-tabulation techniques were used to identify its distribution among migrants of all streams and non-migrants, as well as the distribution by gender, ethnicity and other individual, family and community variables. The chi-square test of association was employed to test the statistical
significance or otherwise of these bivariate distributions across independent variables.
Following the transformation of the data into an event history framework, I generate
Kaplan-Meier estimates of the survival curves for premarital sexual initiation. The
Kaplan-Meier estimator is a non-parametric estimate of the survivor function S(t), the
probability of survival past time t. The estimate at any time t is given by
where nj is the number of adolescents at risk of premarital sexual initiation at time tj and
dj is the number of failures or events occurring at time tj. The Kaplan-Meier estimator
generates the ages at which a given percentage of adolescents initiate premarital sex. The
44 procedure accounts for right censoring in the data, and therefore yields unbiased probability estimates of the timing of premarital sexual initiation.
For examining the role of migration on first premarital sex, net of other relevant factors, I use discrete-time hazard regression models. The discrete-time hazard for interval t is the probability of an event during interval t, given that no event has occurred in a previous interval, i.e. ht1 = Pr (yti =1/ysi = 0, s Logit (hti) = log (hti ⁄ 1 - hti) = α (t) + β' Xti, Where the covariates Xti, may be fixed or time-varying and α(t) is some function of t, which is the baseline logit-hazard. The form of α (t) is approximately linear, and the linear function fitted is of the form: α (t) = α0 + α1t, including t as an explanatory variable in the model. This parameterisation leads to a piecewise-constant hazards model where the hazard is assumed constant within each category of the independent variables. A coefficient β is interpreted as the effect of a 1-unit change in a covariate x on the log-odds of an event in interval t. The exponent of β is interpreted as the multiplicative effect of x on the odds or an odds ratio. In each model I adjusted for inflated standard errors following multiple records for a single adolescent by applying the Huber-White clustering correction with individual identification number as a clustering variable. All discrete-time hazard models were estimated using STATA statistical package. For the second phase of the analysis, I employ the binary logistic regression models to examine the predictors of condom use at first premarital sex, perception of risk 45 of sexually transmitted infections and being sexually active before marriage. The binary logistic model developed for this purpose is written as: Log (Pi / (1- Pi) = β0 + β1X1 +β2X2 +β3X3 +β4X4…..+ βpXp + E1 where Pi is the probability of any of the defined events of interest for an individual i and (1 - P is the compliment). β0 is the base log odd of the reference categories of Xs and βj is the estimate of log odds of each outcome due to the net effect of a given category of the explanatory variables Xj. E is the error term associated with regression. All reported binary logistic regression models were estimated using the STATA statistical package. These series of analyses identify how adverse first intercourse is among those who initiated premarital sex, as well as test the proposition that understanding the etiology of sexual debut will inform our understanding of other sexually-related risk behaviors. The operational definition of predictor variables, perceived direction of associations with premarital sexual behavior, identified mechanisms and relevant citations are summarized in Table 1. 46 Tables 1 Operational definition of independent variables, direction of effects, suggested mechanisms and citations. Variables Operational definition R/Ship Mechanisms Citation +/_ /Reasons Independent Variables Migration Status Rural Non-Migrants=0, _ Presence of a sexual Lurie, 2004; Urban Non-Migrants=1, _ partner provides Chirwa, 1997; Rural-Rural Migrants=2, + uxorial services, Caldwell et al. Rural-Urban migrants=3, + separation engenders 1997; Lurie et Urban-Urban Migrants=4, + other sexual partners, al. 1997; Urban-Rural Migrants=5. + separation from Brockerhoff social control & Biddlecom, structures, 1999; Conservative rural Smith, 1999. origins to sexually permissive urban Individual level destinations. Controls Gender Male=0, _ Few hazard, benefits Mensch et al. Female = 1. + (status), Rape, early 2001;Konde- maturity, sex with Lule et al.1997 older partners -age ; Luke,2003; asymmetry Smith,2004a. Years exposed to risk 10-12=0 _ Exposure time to Meekers, 13-14=1 - opposite sex, 1994; Slap et 15-16=2 +/- Biological and al. 2003; 17-18=3 + Social maturity Kiragu & 19-20=4 + Zabin, 1993. 21-22=5 + 23-24=6 + Educational attainment No Schooling=0, +/_ Exposure to enabling Feyisetan&Peb Primary education=1, +/_ views and ley, 1989 ; Secondary=2, +/_ environment, erosion Kaufman et al. Higher =3. +/- of traditional 2002. strictures, locus of sexual harassment, supervised time, security from predators. Status/Nature of Unemployed=0, +/_ Dependents/under Bledsoe & Employment Formal employment=1, + parental control Cohen, 1993 ; Agric/Self employment=2, _ Independent, out of NISER, 1997. Manual workers=3. + reach of supervision, own resources to finance sexual exploration. 47 Table 1 Operational Definition of Independent Variables Continued Variables Operational definition R/Ship Mechanisms /Reasons Citation +/_ Religious affiliation Catholic=0, + Different levels of Feyisetan Muslims=1, _ emphasis on premarital &Pebley, 1989; Protestant/Others=2. + chastity; Religiosity Meekers, 1994; than affiliation Slap et al. 2003; Gupta, 2000; Kiragu & Zabin, 1993; Zulu et al. Household level 2002. Predictors Living Child of Head=0 _ Under supervision/ Feyisetan & Arrangements Adolescent as Head=1, + subjection Pebley, 1989; Relative of Head=2 _ Independent not Isiugo-Abanihe, Accountable Under 1993; Isiugo- supervision/accountable Abanihe& Oyediran, 2004; Hogan & Kitagawa, 1985. Wealth Status of Poorest=0, + Premarital sex as a Kiragu & Zabin, household Poorer=1, + survival strategy of 1993: UN, 1994. Middle=2, +/_ poor females, perhaps Richer=3, +/_ leisure for the rich, or Richest=4. +/_ protected from sexual exploitation Community level Place of current Urban =0, + Neighborhood Gupta, 2000. residence Rural = 1. _ influences differ but evidences inconclusive. Place of childhood Urban =0, + Environment during the Gupta, 2000. residence Rural = 1. _ formative years are significant Ethnicity Hausa/Fulani/Kanuri=0 _ Cultural values, gender Gage, 1998; Igbo=1, + norms differ in the Hollos & Leis, Niger Delta tribes=2, + employment of positive 1989; Middle Belt tribes=3, +/_ and negative sanctions, Bledsoe & Yoruba=4, + economic and social Cohen, 1993; Others. +/_ support differ Smith, 2004a. Index of Media Lowest=0 _ High levels of media Feyisetan & Exposure Low=1 _ exposure represent Pebley, 1989; High=2 + levels of exposure to Isiugo-Abanihe, Highest=3 + westernization and 1993; Puri & modernization, and Cleland, 2006. sexual stimuli from Western entertainment and liberal sexual values. It may also imply independent views and values 48 Limitations of the study Pooling of data often poses drawbacks particularly in terms of compatibility of survey design and responses. However in this study, both surveys have compatible questionnaire designs and therefore comparable responses. A virtue of repeated surveys is that they can be pooled to increase sample size so long as the effect of the study variable is relatively invariant over the period (Firebaugh, 1997). In this case, the pooled surveys are conducted very close to each other, within the first four year term of the new Nigerian democratic government (1999-2003). Also, the primary interest of the study is in assessing the relationship between adolescent migration and premarital reproductive behavior outcomes, rather than in drawing inferences about how these outcomes vary over time. Notwithstanding, the multivariate analysis will control for variations related to the year of survey. One important limitation is that associated with limited migration data in the Demographic and Health Surveys. It is important to recognize that following how migration status was derived from a series of indirect questions, the results of this study may be sensitive to alternative definitions of migration status that could be derived from direct and more detailed data sets on migration histories. It is important to note that the migration variable defined represents lifetime migration status. There are possibilities of previous migration experiences but the limitation of data did not enable a clarification of this situation. Also since the survey did not collect information on the name of previous or childhood place of residence or origin, it is important to recognize the possibility of misclassification of migrants’ place of origin. Chattopadhyay et al. (2006) indicated the tendency in Ghana for migrants to identify their places of origin as the nearest town 49 rather than their specific village, and the possibility for migrants to classify their places of origin as urban, if following their migration, the rural classification of the place changed. While the extent to which this is true concerning Nigeria is not known, the possibility of reclassification is very high following the creation of new local governments areas in Nigeria between 1976 and 1996 ( 776 Local Government Areas were created over the period). However, since the subject of this study is adolescents, with likely very short migration histories, misclassification of origin among them should be low, as the likelihood of reclassification increases with time and the tendency is higher for long term migrants and those who migrated as children (Chattopadhyay et al. 2006). There is evidence that sexual activity, particularly sex outside marriage, is normally underreported by women, and even more so by adolescents (Mensch et al. 2001). A comparative examination of data from adolescent sexuality surveys and the Demographic and Health Surveys (DHS) from three Latin American countries show that DHS data greatly understate adolescent sexual activity (Mensch et al. 1998). The proportion of women who report premarital sex in the DHS ranges from three to five percent in the three countries. However the survey by the Centers for Disease Control reported 15-16 percent of adolescent women having premarital sex (Mensch et al. 1998). In a survey in Lagos, Nigeria, Adegbola and Babatola (1999) confirm the reluctance of many respondents to disclose details of their sexual activities to survey interviewers. Under these circumstances, the likelihood of under-reporting the activity of interest is very plausible. However, since the goal of this study does not include establishing absolute levels of premarital sexual behaviors and given that there is no basis to think that under-reporting if it exists, varies systematically across settings, the study will compare 50 the relative occurrence of premarital sexual behavior across all migration categories and other covariates as reported. One key limitation of the data relates to the disproportionate representation of adolescent males relative to females in the sample. Most DHS data including the Nigeria DHS used for this study include fewer unmarried adolescent men and the sample sizes are generally too small to permit rigorous analysis of male-female differences in multivariate statistical modeling (NCRIM, 2005). The 7,799 study sample for this study includes only 20 per cent adolescent males. This limits the level of gender-related comparative analysis that is possible in the study despite the fact that the male sample selection is random and representative. Defining migration status at the time of premarital sexual initiation highlights a key limitation of using cross sectional data in studying time-variant issues. Using information on age at first sexual intercourse, current migration status, current age and migration duration, 541 migrants who initiated sex before their current move are identified and were excluded in the analysis of the relationship between migration and sexual debut. While it can be plausibly argued that adolescents that migrated are generally selected by their similar characteristics, they could not still be included among migrants to avoid overestimating the impact of migration. Also they could not be included as non-migrants at the time of first sex, because of the potential of other moves before the most current move, and the data did not provide information on whether they had moved earlier before current move. Statistical tests indicate that this exclusion has no adverse effect on the estimation outcomes. For other time-variant covariates, the lack of time-series data is also a key limitation and in recognition, the outcomes of the analysis 51 will be treated with caution, limiting interpretation of results of time-variant predictors to correlational associations rather than drawing causal inferences. Despite the limitations identified and the lack of data with longitudinal designs in Nigeria, the use of DHS survey data in this study represents a pragmatic way of getting some handle on issues of national research, policy and program importance, which otherwise will not have been done. Also the national representativeness of the data remains its invaluable strength, which enhances the generalizabilty of its findings to other settings in developing countries beyond Nigeria. CHAPTER 4 DATA ANALYSIS: INTERNAL MIGRATION AND PREMARITAL SEXUAL INITIATION 4.0 Introduction This chapter presents the distribution of the characteristics of the study population, and the bivariate relationships between migration status, other adolescent characteristics and premarital sexual initiation. Following the transformation of data into person-years, using the event history framework, the Kaplan-Meier survival curves and the discrete-time hazard regression models predicting premarital sexual initiation conclude the chapter. 4.1. An overview of the characteristics of the study population The primary outcome of interest in this study is adolescent transition into first premarital intercourse. Table 2 column 1 indicates that 38.8% of all adolescents aged 15-24 had initiated premarital sex at the time of the survey. Despite the attention and concerns focused on migration and its consequences, the data indicates high level of immobility, with over 60% of all adolescents currently residing in their places of birth. It is important to observe that urban-urban and rural-rural streams of migration are quantitatively larger than the much focused on rural-urban migration. This distribution is consistent with findings on migration patterns observed for sub-Saharan Africa generally and Nigeria in particular (Chattopadhyay et al. 2006; Lacey, 1985; Mberu, 2005; Oucho, 1998). Approximately 80% of the entire sample is female. The proportion female underscores a key criticism of the Demographic and Health Surveys, which generally 52 53 sample and collect less information on reproductive behavior of unmarried adolescent males, often with sample sizes too small to permit many useful analyses (NRCIM, 2005: 202). Part of the explanation for this gender imbalance in sampling is the simplistic assumption that women are more relevant on issues relating to fertility and sexual behaviors. The mean age of the study population at the time of the survey is 19.02 years. For the entire sample, 55% are Christians, 44% are Muslims, and others are traditionalists. Ethnic origin of the adolescents reflects the dominance of the Hausa/Fulani/Kanuri, the Igbo and the Yoruba in the general population. Minorities of central and southern Nigeria make up 40% of the sample. The study population is composed of a significant percentage of adolescents with no education (25%). However, the sample also indicates high level of literacy with an impressive 49.5% having attained secondary education. Notwithstanding, the low levels of higher educational attainment is consistent with observations that listed sub-Saharan Africa with the former Soviet Asia and South America, as world regions where percentage of tertiary education enrollments have dramatically declined in recent years (NCRIM, 2005). The largest proportion (67%) of adolescents is unemployed, and among those employed, more are in formal than in agriculture or self employment. This few adolescents employed in the agricultural sector may not reflect the efficiency of the sector as in developed societies, where lower participation reflects better factor utilization, but instead may be an indicator of the unattractiveness of the agricultural sector for young adults, which among other factors engender migration out of rural origins to urban non-agricultural sectors (Mberu, 2006). 54 Table 2 Univariate characteristics of the study population and bivariate association with premarital sexual initiation Variable All Sample Premarital Sex Status Total (N=7799) Initiated Premarital sex Premarital sex abstainers (n) Premarital sexual status Abstainers from premarital sex 61.2 - 61.2 4771 Initiated premarital sex 38.8 38.8 - 3028 Migration Status*** Urban-Urban migrants 13.6 45.7 54.3 1064 Urban-Rural migrants 3.7 49.8 50.2 289 Rural-Urban migrants 10.6 45.5 54.5 830 Rural-Rural migrants 11.8 40.6 59.4 917 Urban non-migrants 20.5 34.4 65.6 1597 Rural non-migrants 39.8 35.4 64.6 3102 Gender Female 79.9 38.7 61.3 6230 Male 20.1 39.3 60.7 1569 Current age*** 15-16 22.1 14.8 85.2 1720 17-18 23.7 31.8 68.2 1847 19-20 25.7 45.6 54.4 2005 21-22 15.3 54.6 45.1 1197 23-24 13.2 59.6 40.4 1030 Religion *** Muslims 43.7 32.7 67.3 3407 Catholics 16.4 42.9 57.1 1279 Protestants/Other Christians 39.0 44.1 55.9 3040 Traditional/Others 0.9 34.2 65.8 73 Place of current residence Rural 62.2 38.1 61.1 4849 Urban 37.8 40.0 60.0 2950 Childhood place residence*** Rural 45.5 38.0 62.0 3551 Urban 49.8 40.4 59.6 3882 Outside Nigeria 4.7 30.9 69.1 366 Ethnic origin*** Hausa/Fulani/Kanuri 26.1 28.8 71.2 2032 Igbo 17.9 38.5 61.5 1394 Niger-Delta groups1 10.9 52.4 47.6 848 Middle-Belt groups2 9.4 46.0 54.0 731 Yoruba 16.5 44.0 56.0 1288 Others 19.3 37.1 62.9 1506 Education Attained*** No education 25.3 32.3 67.7 1974 Primary level education 20.7 37.9 62.1 1618 Secondary education 49.5 40.7 59.3 3863 Tertiary education 4.4 59.3 40.7 344 Nature of employment*** Unemployed 67.1 33.7 66.3 5236 Formal employment 23.0 53.3 46.7 1794 Agriculture/Self employment 9.9 40.1 59.9 769 Living arrangements*** Adolescent head of household 4.3 62.5 37.5 336 Adolescent as spouse of head 28.7 45.2 54.8 2236 Adolescent as child of head 48.3 33.2 66.8 3768 Adolescent as relative of head 18.7 38.0 62.0 1459 Household wealth index Poorest 20.1 38.4 61.6 1571 Poor 20.7 38.0 62.0 1614 Middle 19.5 37.7 62.3 1519 Rich 20.2 40.3 59.7 1575 Richest 19.5 39.6 60.4 1520 Index of media exposure*** Lowest exposure 25.1 36.0 64.0 1958 Medium exposure 50.6 38.3 61.7 3943 Highest exposure 24.3 42.8 57.2 1898 1The Niger-delta is composed of the Urhobo/Isoko/Edo/Itshekiri/Annang/Efik/Ijaw/Ogoni/Ibibio of the South-South. 2The Middle-belt group is an amalgam of the Tiv/Igala/Idoma/Nupe/Kamberi/Gwari/Ibira of central Nigeria. χ2 one-tailed test ***p<.001 **p<.01 *p<.05 55 In terms of living arrangements, 48.3% of adolescents live with their parents, 18.7% live with relatives. Of the remainder, 4.3% and 28.7% live independently and with spouses respectively. The household wealth index is evenly distributed in all five wealth status categories and the index of media exposure has an inverted U-shaped distribution. While 25% of adolescents are in the least exposed category, an equivalent proportion is in the highest exposed category, and 50.6% indicates medium-level exposure to the mass media. 4.2. Migration status, adolescent characteristics, and premarital sexual initiation In this subsection, I present the bivariate relationships between adolescent migration status, other adolescent characteristics, and premarital sexual debut. The outcomes are summarized in the complimentary columns of Table 2. Table 2 indicates significant relationships between migration status and initiation of premarital sex. The highest proportion of migrants who have initiated premarital sex are urban-rural migrants, but urban-urban and rural-urban migrants show similar levels of premarital sexual debut. Rural-rural migrants indicate the lowest levels of premarital sexual initiation among migrants, though not as low as urban and rural non-migrants. The emerging picture here suggests that migrants are generally more likely to initiate premarital intercourse. However, it seems that migrants’ exposure to the urban environment may be a key predictor of premarital sexual engagement. The mean age of premarital sexual debut for all adolescents is 16.3 years. This breaks down to 16.2 and 16.5 years for male and female youths respectively. The outcome for girls represents a 9 months decrease in age at sexual debut observed for the 56 same cohort in 1999 (Isiugo-Abanihe and Oyediran, 2004). Consistent with the biological and social exposure hypothesis, the proportion of those who initiated premarital intercourse consistently rises with age. On gender and premarital sexual debut, the result suggests no significant difference between the sexes. While this outcome contradicts the emphasis on female exposure to risks associated with earlier sexual debut following age asymmetry of sexual partners, it strengthens the observation of increasing risk of exposure of male youths to similar sexual risks often associated with female adolescents. Religion is significantly associated with premarital sexual debut, with Muslims and traditionalists exhibiting lowest levels of premarital sexual initiation relative to Christians of all denominations. While there is no significant difference between premarital sexual initiation and current place of residence, place of childhood residence is significantly associated with premarital sexual initiation. Adolescents who resided in urban areas before age 12 reported higher levels of initiation of premarital sex than those who resided in rural areas. One profound outcome observed in the data is the wide variation in premarital sexual initiation for adolescents from different ethnic origins, which ranges from 28.8% for the Hausa/Fulani/Kanuri adolescents to 52.4% for those from ethnic groups in the Niger-Delta. For other groupings the outcomes are: 38.5% for the Igbo, 44% for the Yoruba, and 46% for adolescents from the Middle-Belt region. The relationship between educational attainment and premarital sexual debut is consistent with expectations and existing literature, which suggests that the highly educated are more likely than the less educated to initiate premarital sexual intercourse. While 32.3% of adolescents with no education initiated premarital sex, the corresponding 57 proportion for those with tertiary education is 59.3%. Youths with primary and secondary education fall between the two extremes. Status and nature of employment significantly relate to premarital sexual debut. Youths with formal employment (a status that confers most physical and financial independence from parents and family) indicate the highest level of premarital sexual initiation. Conversely, being unemployed (a status suggesting higher propensity of dependence on parents or guardians) shows the lowest levels of premarital sexual initiation. In terms of living arrangements, adolescents who live as heads of households and those who indicate that they are living with spouses at the time of the survey report the highest level of premarital sexual debut, while those living with their parents report the least level of premarital sexual initiation. This may be speaking to the behavior controlling roles of authority figures generally identified in other settings. Adolescents with highest media exposure reported the highest level of premarital sexual initiation. The relationship between premarital sexual initiation and household wealth suggests a shallow u-shaped outcome, with youths from households in the highest and lowest socioeconomic categories showing higher propensities to premarital sexual debut than those from the middle wealth ranks. However, the relationship between premarital sex and household wealth is not statistically significant. In accounting for censoring in the data, the cross-sectional data set is restructured following an event-history framework into person-year format. The Kaplan-Meier survival curves were fitted for sexual initiation by all independent variables. The fitted survival curves showing the statistically significant relationships between categories of 58 independent variables and premarital sexual initiation are presented in the appendix. The results generally reiterate the cross-sectional bivariate outcomes. While the bivariate cross-sectional analysis and the Kaplan-Meier survival curves provide an overview of the repercussions of migration and other adolescent characteristics on premarital sexual initiation, these relationships are further refined, accounting for confounders, by the fitting of the discrete-time hazard regression models, presented in the subsection that follows. 4.3 Discrete-Time Hazard Multivariate Analyses: Premarital Sexual Initiation Table 3 presents the estimated odds ratios for the covariates included in the model. Model 1 includes only migrant status as a predictor of premarital sexual initiation. Gender, years of exposure to the risk of premarital sex and religion (key time-invariant factors at the individual level) are included in the Model 2 while Model 3 controls for the effects ethnic origin (a key time-invariant social identity covariate), place of childhood residence, and year of survey. The full model includes time-variant individual characteristics: educational attainment and household levels variables such as household living arrangements, wealth status, and the index of media exposure. Model 1 shows a strong relationship between migration status and initiation of premarital intercourse. Urban-urban migrants are 1.29 times as likely as rural non- migrants to initiate premarital sex. The corresponding odds ratios for other migrant streams relative to the reference category are: urban-rural 1.43; rural-urban 1.48; and rural-rural 1.14. Urban non-migrants are not statistically different from their rural counterparts in premarital sexual debut. In Models 2 and 3, controlling for individual and 59 Table 3 Discrete-time hazard models predicting the odds of premarital sexual initiation among Nigerian adolescents Variables Model 1 Model 2 Model 3 Model 4 Odds ratio RSE Odds ratio RSE Odds ratio RSE Odds ratio RSE Migration Status Rural non-migrants 1.00 - 1.00 - 1.00 - 1.00 - Urban-Urban migrants 1.29*** .07 1.09 .07 1.09 .09 0.99 .08 Urban-Rural migrants 1.43*** .12 1.23* .12 1.21† .13 1.00 .11 Rural-Urban migrants 1.48*** .08 1.40*** .10 1.45*** .13 1.35*** .12 Rural-Rural migrants 1.14* .07 1.22** .09 1.24** .10 1.07 .09 Urban non-migrants 0.96 .05 0.89* .05 0.90 .07 0.93 .07 Gender Male 1.00 - 1.00 - 1.00 - Female 0.96 .05 0.97 .05 0.87** .05 Years of exposure to risk 10-12 ( 1-3 years) 1.00 - 1.00 - 1.00 - 13-14 (4- 5 years) 4.09*** .42 4.07*** .42 4.11*** .43 15-16 (6-7 years) 12.29*** 1.17 12.11*** 1.16 12.44*** 1.19 17-18 (8-9 years) 22.02*** 2.11 21.66*** 2.08 22.28*** 2.14 19-20 (10-11 years) 27.30*** 2.83 27.46*** 2.85 28.12*** 2.93 21-22 (12-13 years) 13.70*** 2.12 14.03*** 2.18 14.28*** 2.23 23-24 (14-15 years) 11.29*** 3.12 11.79*** 3.28 12.14*** 3.36 Religion Muslims 1.00 - 1.00 - 1.00 - Catholics 1.28*** .08 1.14† .09 1.26** .10 Protestants/Other Christians 1.40*** .07 1.06 .06 1.16* .07 Traditional/Others 1.34 .32 1.06 .26 1.12 .28 Ethnic origin Hausa/Fulani/Kanuri 1.00 - 1.00 - Igbo 1.37** .15 1.63*** .18 Niger-Delta groups 2.64*** .28 3.07*** .33 Middle-Belt groups 1.92*** .20 2.19*** .23 Yoruba 1.94*** .17 2.18*** .20 Others 1.57*** .15 1.76*** .16 Childhood place of residence Rural 1.00 - 1.00 - Urban 0.97 .07 0.96 .07 Year of survey 2003 1.00 - 1.00 - 1999 0.87** .04 0.82** .05 Education Attained No education 1.00 - Primary level education 0.91 .08 Secondary education 0.97 .08 Tertiary education 0.97 .11 Status/Type of employment Unemployed 1.00 - Formal employment 1.41*** .07 Agriculture/Self employment 0.99 .08 Living arrangement Adolescent child of head 1.00 - Adolescent head of household 1.53*** .13 Adolescent as spouse of head 1.74*** .11 Adolescent relative of head 1.15* .07 Household wealth index Poorest 1.00 - Poor 0.94 .07 Middle 0.83* .06 Rich 0.90 .06 Richest 0.88† .07 Index of media exposure Lowest exposure 1.00 - Medium exposure 0.95 .06 Highest exposure 1.07 .10 Number of adolescents= 7,796 Number of person years=59, 089 Wald Chi-Square (d.f.) 79.4*** 1781.1*** 1858.8*** 2108.0*** (5) (15) (22) (36) † p<.10, *p<.05, **p<.01, ***p<.001 60 community-level time-invariant covariates respectively, the association between adolescent migration status and premarital sexual initiation is slightly altered for all migration streams. In Model 4, however, the statistically significant variation between all migration categories and premarital sexual debut disappeared except for the rural-urban category, which remains consistently significant in all models. On gender and premarital sexual initiation, the results suggest that female adolescents are marginally less likely to initiate premarital sex than males. In Model 4, where the outcome is statistically significant, female adolescents are 13% less likely to initiate premarital sex than males. Consistent with the proposition that biological and social maturity increases the risk of exposure to premarital sexual activities, adolescent years of exposure to the risk of premarital sex significantly predicts premarital sexual initiation. One noteworthy outcome here, however, is that the independent effects of years of exposure to the risk of premarital sex is not monotonic, rather it represents an inverted U-shaped effect that peaks at ages 19-20 and continually decreases until the end of the observation period. This outcome suggests that the propensity to initiate premarital sex increases until a certain age and those who survive up to that threshold have lower propensity to premarital sexual initiation. However, this does not mean that they are primary abstainers as their propensity for premarital sexual initiation remains significantly higher than the reference group. In fact, those exposed to the risk of premarital sex for 14-15 years are 12.14 times as likely as those exposed for 2-3 years to initiate premarital sex in the final model. 61 In line with expectations, religious affiliation is significantly associated with premarital sexual initiation. Catholic and Protestant Christians are significantly more likely to initiate premarital sex than Muslims. However, while this outcome remained significant in Model 4, this association is greatly weakened in subsequent models, particularly when ethnic origin was controlled for. In general, this finding is consistent with strict sexual behavioral norms generally associated with Islamic cultures and enclaves relative to other religious groups. One of the most interesting outcomes of this analysis is the marked variation in the association between ethnic origin and premarital sexual initiation. Relative to the Hausa/ Fulani/ Kanuri of the core north of Nigeria, Igbo youths are 1.37 times more likely to initiate premarital intercourse. For other groups, the corresponding odds ratios are: the Urhobo/Isoko/Edo/Itshekiri/Annang/Efik/Ijaw/Ogoni/Ibibio of the Niger Delta, 2.64; the Tiv/Igala/Idoma/Nupe/Kamberi/Gwari/Ibira of Central Nigeria, also called the Middle Belt, 1.92; the Yoruba of Western Nigeria, 1.94; and “Others”, 1.57. These outcomes identify youths from the Niger-Delta as generally the most likely to engage in premarital sexual initiation in Nigeria. That these effects remain very strong after controlling for all defined covariates including education and household wealth status in Model 4 may suggest the important independent role of socio-cultural context in our understanding of sexual norms and behavior in Nigeria. The continued potency of ethnicity in defining identity, interaction and behavioral boundaries in the socio- economic and political spheres of modern Nigeria makes the result even more persuasive. Generalized conclusions that neglect nuances of cultures even within a given country 62 may be overlooking or masking within-group variations necessary for pointed understanding of behavior dynamics within a state. Adolescent childhood residence in urban or rural areas shows no significant independent effect on premarital sexual initiation. The increasing fusion of urban and rural spaces in many parts of Nigeria in recent years linked to the diffusion of behavioral norms through globalization, the mass media and migration circulation (Chukwuezi, 2001; Smith, 1999; Mberu, 2005), may be related to this outcome. These set of studies have shown that urbanization in Africa is experienced not in terms of a deep or strict rural/urban divide, but characterized by continuing commitment of migrants to their rural communities. The fluidity of migration circulation between urban and rural areas in Nigeria led Gugler (1971, 1991) to describe rural migrants to urban Nigeria as those “living in a dual system.” Lastly for time-invariant covariates, there is an 18% increase on the incidence of premarital sexual initiation between 1999 and 2003 in the full model. This outcome is consistent with observations of increased premarital sexual activities in recent years. The level of education attained is negatively associated with the likelihood of premarital intercourse but the outcome is not statistically significant. This outcome may be counter intuitive following expectations spelt out in the literature. However, that time spent in school also implies time that allows adolescent physical and social maturity and therefore pressure for sexual exploration, may have been captured by the age of exposure variable controlled for above. Notwithstanding, the outcome may be consistent with the role incompatible hypothesis, which states that adolescents engaged in worthy goals for successful transition to adulthood are less distracted by incompatible risky behaviors. 63 The status/nature of adolescent employment is significantly associated with premarital sexual debut. Employment in the formal sector, which confers the highest level of physical and financial independence on youths, indicates significant association with premarital sexual initiation. Adolescents employed in the formal sector are 1.41 times as likely as the unemployed to initiate premarital intercourse. Consistent with the proposition that adolescent transition to independent living increases the risk of exposure to premarital sex, adolescents who are heads of own households are 1.53 times as likely as those who live under the authority of their parents to initiate premarital sexual relationships in the full model. This outcome contrasts with those who live under their relatives, who are 1.15 times as likely as those living with parents to initiate premarital sex. These estimates strongly reinforce the behavioral conformity roles of authority figures in the lives of young people. One outcome worthy of note relates to adolescents who are currently living as partners of household heads, but whose living arrangement at the time of premarital sex could not be determined because of data limitation. This group is 1.74 times as likely as those living with their parents to have initiated premarital sex. Their increased probability may be related to their early independence or relative autonomy as those who live independent of parents and relatives. However, this outcome may be related to the positive correlation found between the propensity for union formation and premarital sexual activities in the region. Researchers in Nigeria and Ethiopia found that premarital sex may be a means to consolidate a relationship that promises future marriage or engaged as a prelude to marriage (Smith, 2004; Lindstrom et al. 2005). 64 The association between household wealth and premarital sexual initiation is generally weak and marginally significant. Relative to adolescents from the poorest households, those from middle-wealth families are 17% less likely to initiate premarital sex. Similarly those from the richest households are 12% less likely to initiate premarital sex relative to the reference group. Finally the level of media exposure is not significantly associated with premarital sexual initiation. Generally higher exposure to the mass media is expected to be associated with higher propensity for premarital sexual initiation, and the result obtained points in that direction. Following other findings suggesting that girls from poor backgrounds use sex as an economic survival strategy (Luke, 2003; Fatusi, 2004; Okonkwo et al. 2005) and the expectation that ethnicity affects males and females differently, interaction terms were introduced into the analysis to verify the differential effects of ethnicity and household wealth on gender and their association with premarital sexual initiation. The estimates of the relevant variables of the two models that included interaction terms for ethnic origin and household wealth and being female respectively are presented in Table 4. Model 1 examines the gender variation related to ethnic origin, while Model 2 examines that related to household wealth status. Model 1 indicates a significant gender variation in premarital sexual initiation. With Hausa/Fulani/Kanuri males as the reference group, male adolescents from all other groups are over 300% more likely to initiate premarital sex than the reference group. The Yoruba and Niger Delta males are the most likely to initiate premarital intercourse, with odds ratios of 4.99 and 4.36 respectively. 65 On the other hand, female adolescents from all ethnic groupings are significantly less likely to initiate premarital sex than the reference group. The most distinctive aspect of the result is that Hausa/Fulani/Kanuri female adolescents are 2.08 times as likely to initiate premarital sex as the Hausa/Fulani/Kanuri males, identifying them as the most vulnerable to premarital sexual initiation among Nigerian female adolescents. The intriguing aspect of the vulnerability of the Hausa/Fulani/Kanuri females is the fact that male adolescents from the group are the least likely to initiate premarital sex among all adolescent males in the study. This scenario suggests that premarital sexual activities of these female youths are likely undertaken with males outside adolescents’ cohorts. This outcome confirms in an empirically significant way the asymmetry of sexual partners in Northern Nigeria. Following the feudal and patriarchal structure of these groups, the result at least suggests that adolescent girls may be having sex with adults outside their age cohorts confirming the age asymmetry of sexual partners found in most of the region (Luke, 2003). While my data did not allow a direct test of this proposition, other studies in countries of sub-Saharan Africa suggest higher HIV/AIDS seroprevalence for women than men for similar reasons. In an ongoing multi-country study in SSA, Pongou (2007) proposed that higher female vulnerability to HIV/AIDS may be explained in relation to the influence of men with means, whose sexual activities result in the embeddedness of most of the women in the sexual network and therefore exposing them to infection. Consequently, men of lesser means who are eliminated from the network are spared from infection. However, the veracity and consistency of this association will need further test on other adequate data from the region in order to arrive at a more assured and certain conclusion. 66 Table 4 Odds Ratios for the Effects of Ethnicity and Household SES Interaction with Gender on the Hazard of Premarital Sexual Initiation among Nigerian adolescents. Variables Model 1 Model 2 Odds ratio RSE Odds ratio RSE Interaction Model 1: Gender x Ethnic Origin1 Male (reference) 1.00 - Female 2.08*** .35 Ethnic origin Hausa/Fulani/Kanuri 1.00 - Igbo 3.58*** .68 Niger-Delta groups 4.36*** .87 Middle-Belt groups 3.99*** .80 Yoruba 4.99*** .87 Others 3.24*** .58 Ethnic origin x Female Igbo x Female 0.32*** .06 Niger-Delta groups x Female 0.56*** .12 Middle-Belt groups x Female 0.41*** .09 Yoruba x Female 0.30*** .06 Others x Female 0.41*** .08 Interaction Model 2: Gender x Household SES Male (reference) 1.00 - Female 1.11 .16 Household wealth index Poorest 1.00 - Poor 0.92 .17 Middle 1.10 .18 Rich 1.18 .17 Richest 1.15 .19 Gender x Household wealth index Poor x Female 1.01 .20 Middle x Female 0.71† .13 Rich x Female 0.69* .11 Richest x Female 0.72† .13 Number of adolescents= 7,796 Number of person years=59, 089 Wald Chi-Square (d.f.) 2168*** 2129*** (41) (40) 1Interaction models include all variables from main effects models in Table 5. † p<.10, *p<.05, **p<.01, ***p<.001 Next to the high probability of premarital sexual engagement of the Hausa- Fulani/Kanuri females are female adolescents from the Niger-Delta of Nigeria with odds ratio of 1.06 (2.08 X 0.56). The odds are lower for other female groups particularly the Yoruba and the Igbo with odds of 0.62 (2.08 X 0.30) and 0.67 (2.98 X 0.32) respectively. The interaction between gender and adolescent household wealth status yielded only marginally significant differentials. However, the direction of the outcome suggests that adolescent females from the poorest households are more likely to early premarital 67 sexual initiation relative to male adolescents from poorest households. Further, the result suggests that middle rich and richest household backgrounds give some marginal but significant protection for female adolescents by reducing their propensity to early premarital sexual initiation. However, for male adolescents, rich household wealth status does not seem to have the same outcome. Though the result is not statistically significant, the estimates suggest an opposite experience for adolescent males from middle, rich and richest family backgrounds. 4.4 Summary This chapter primarily addresses the prevalence of premarital sexual initiation among migrant and non-migrant adolescents, accounting for the net effects of confounding individual, household and community level variables. The results indicate a significant association between migration status and adolescent premarital sexual initiation. The most important finding relates to rural-urban migrants, who show persistent and significant likelihood of premarital sexual initiation in all models. This outcome lends support to the observation that migration may expose youths from highly conservative rural environments to more sexually permissive urban areas. It may also be a pointer to the level of independence or autonomy of migrant adolescents relative to their non- migrant peers in matters relating to premarital sexual engagement. Consequently, programmatic attempts to control adolescent sexual behavior and its consequences, may need to direct attention at issues relating to why youths move out of rural origins and seek for ways and means to provide alternative economic survival opportunities than that pursued through massive rural out-migration of youths to urban destinations. 68 One important dimension of the result is evidence that migration alone is not a sufficient explanation for adolescent premarital sexual involvement; rather personal, household and community level characteristics mediate this association, supporting a multi-factor approach in the study of premarital sexual behavior. In general, adolescent gender, years of exposure to the risk of premarital sexual initiation, religious affiliation, ethnic origin, employment in the formal sector, and early transition to independent living are strong independent predictors in varying degrees. On years of exposure to the risk of premarital sex, the finding is consistent with the proposition that biological and social maturity increases the risk of exposure to premarital sexual activities. However, a particularly noteworthy finding is that years of exposure to the risk of premarital sex does not have a monotonic effect, rather peaks at ages 19-20 and continually decreases to lower levels until the end of the observation period. While the odds of premarital sexual initiation did not decrease to levels as low as it is for those exposed for only few years, the outcome does suggest that adolescents who successfully abstain from premarital sex may have lower propensities for initiation after a certain threshold. In a way, they may have lived long enough to may have developed some defense mechanisms and/or other means of navigating their relationships beyond sexual engagement. Put differently, they may be able to exercise more agency against pressures to initiate premarital sex. From the perspective of protecting and preparing the next generation for successful transition to adulthood, identifying this group and showcasing their experiences may be an important step in promoting postponement of premarital sexual engagement until certain ages are attained, by which time firm groundwork for successful transition to adulthood would have been achieved. 69 Further, transition to independent living exposes adolescents to significant risks of premarital sexual initiation. Unlike youths who live under the supervisory authorities of parents and relatives, adolescents who are heads of households and those who live with partners are 53% more likely to initiate premarital intercourse than the reference group. A similar result is obtained for formal employment- a livelihood opportunity that gives adolescents physical and financial independence from their significant others. A key outcome relates to the cultural dimension of premarital sexual behavior. Cultural identity factors (religion and ethnicity) stood among the strongest independent predictors of premarital sexual initiation. Interaction terms on the gender dimension of ethnic origin and household’s wealth status yielded intriguing results, which call for particular attention on the role of social and cultural contexts in understanding adolescent reproductive behavior within a heterogeneous society like Nigeria. The general high propensity of male adolescents to early premarital sexual initiation underscores the need to pay attention to the male gender in policy and programs. The particular higher likelihood of Hausa/Fulani/Kanuri female youths, seem to suggest that adult men beyond their age cohorts may be responsible for their premarital sexual engagement. These findings highlight the simplicity of one-dimensional explanations that neglect the complexity of contextual nuances between groups within a single society. CHAPTER 5 DATA ANALYSIS: CONDOM USE AT PREMARITAL SEXUAL INITIATION 5.0 Introduction Contraceptive use at first intercourse is an important indicator of adverse outcomes of first intercourse and is related to teens’ propensity to use contraception consistently and regularly thereafter (Abma and Mcgill, 2007; Shafii et al. 2004). The use of condom is one of the most effective risk-reducing strategies available to young people who have initiated intercourse. Consistent and correct use of condoms can prevent 90-95% of HIV infections among adolescents that cannot abstain, and part of the success of the ABC program in HIV reduction in Uganda was linked to condom use (Singh et al. 2003). In this chapter, I examine reported condom use at sexual initiation among Nigerian adolescents and identify significant predictors. As indicated above, the relevant question on condom use at premarital sexual initiation was asked only in the 2003 survey and data on 1,295 adolescents who have initiated premarital intercourse and about whom I have information on all relevant defined predictors are analyzed. 5.1 Condom use at premarital sexual initiation: Descriptive analysis Table 5 indicates that for those who initiated premarital sex, only 15.8% used condom at sexual initiation. This outcome suggests that condom use at first sex is lower in Nigeria relative to some African countries like Lesotho (34% for men and 59% for women), and Uganda (43% for men), Ghana (18% for men and 27% for women) (Tumwesigye et al, 2005). The table shows significant bivariate association between migration status and the 70 71 Table 5 Characteristics of adolescents who used condom at sexual initiation (N= 1,295) Variable Yes No Ever used condom 15.8 84.2 Migration Status*** Urban-Urban migrants 21.1 78.9 Urban-Rural migrants 13.8 86.3 Rural-Urban migrants 19.0 81.0 Rural-Rural migrants 6.8 93.3 Urban non-migrants 22.3 77.7 Rural non-migrants 11.1 88.9 Age at first intercourse*** 10-14 5.7 94.3 15-16 12.9 87.2 17-18 19.0 81.0 19+ 23.9 76.1 Gender*** Female 13.8 86.2 Male 21.1 78.9 Religion Muslims 13.7 86.3 Catholics 15.9 84.1 Protestants/Other Christians 16.8 83.2 Traditional/Others 22.2 77.8 Childhood residence** Rural 8.8 91.2 Urban 21.6 78.4 Ethnic origin*** Hausa/Fulani/Kanuri 1.9 98.1 Igbo 19.6 80.4 Niger-Delta groups 13.0 87.0 Middle-Belt groups 14.6 85.4 Yoruba 34.6 65.4 Others 6.9 93.1 Education Attained*** No education 1.2 98.8 Primary level education 8.2 91.8 Secondary education 17.5 82.5 Tertiary education 42.7 57.3 Status/Type of employment*** Unemployed 19.1 80.9 Formal employment 13.3 86.7 Agriculture/Self employment 6.3 93.7 Living arrangements*** Adolescent head of household 24.0 76.0 Spouse of head of household 3.1 96.9 Adolescent as child of head 20.6 79.4 Adolescent as relative of head 13.7 86.0 Household wealth index*** Poorest 9.8 90.2 Poor 9.0 91.0 Middle 12.5 87.5 Rich 19.2 80.8 Richest 23.4 76.6 Index of media exposure*** Lowest exposure 9.6 90.4 Medium exposure 20.6 79.4 Highest exposure 14.3 85.7 χ2 (One-tailed test) ***p<.001 **p<.01 *p<.05 use of condom at premarital first intercourse. Higher proportions of urban-urban, rural- urban and urban non-migrants indicate condom-use at premarital sexual initiation than 72 urban-rural, rural-rural and rural non-migrants. These outcomes point to some positive relationships between living in the urban environment and condom use at premarital sexual initiation. Both migrants to urban destinations from rural and urban areas and urban non-migrants indicate higher propensity for condom use at sexual initiation than rural destination migrants and rural non-migrants. These results are consistent with the principle that rural-urban migrants adapt to the new economic, social and cultural environment at the places of destination, resulting in changes in behavior. However, that migrants to rural destinations are less likely to use condom at sexual initiation suggests that the result may be speaking more to the role of urban areas in access to condoms than that of migration status. This is reinforced by the outcome that youths who lived in urban than rural areas before age 12, indicate higher proportion of use of condoms at premarital sexual initiation than others. Further multivariate analyses in the subsequent section will help to refine these outcomes by accounting for confounding variables. The proportion of adolescents who use condom at premarital first intercourse is lowest among those who initiated premarital sex in the earliest years. Conversely use is highest among those who initiated sex after age 18, underscoring a strong age effect. Also higher proportion of males than females, Christians than Muslims used condom at premarital sexual initiation. The result shows large variation in condom use at premarital first intercourse among ethnic groups. The proportion ranges from 1.9% among the Hausa/Fulani/Kanuri; to 19.6 % among the Igbo; and 34.6% among the Yoruba. For time-variant covariates, higher levels of educational attainment increases condom use at first intercourse, with the highest proportion of use among those who 73 attained tertiary education. Further, adolescents who live independent of their parents or relatives, those with more exposure to the media, and those from the richest households indicate higher condom use at premarital sexual initiation than others. The Chi-square (χ2) statistical tests of association between adolescent individual, household and community characteristics and condom use at premarital first sex mostly yield significant results. Apart from religious affiliation, which indicated no statistically significant association, all other independent variables examined in Table 5 are significantly associated with condom use at first sex. However, these outcomes are further refined in the subsequent multivariate models to account for bias that may be related to confounding variables. 5.2 Condom use at premarital sexual initiation: Multivariate models I further estimate multivariate logistic regressions of the likelihood of condom use at first premarital sex and present the results in Table 6. Four binary logistic models are estimated. Models 1 through 3 successively include groups of time-invariant predictors, while Model 4 adds all the time-variant covariates. As in the bivariate analysis, Model 1 confirms that urban-urban, rural-urban and urban non-migrants are strongly associated with condom use at premarital sexual initiation. In Model 2, with the introduction of time-invariant personal characteristics of adolescents, the advantage of urban resident migrant and non-migrant youths remain significant. In Model 3 however, with controls for ethnicity and place of childhood residence before age 12, the migration dummies became non-significant. These results underscore the strong role of the urban environment and ethnic background in protective reproductive behavior. 74 Table 6 Estimates predicting condom use at premarital sexual initiation among Nigerian adolescents (N=1,295) Variables Model 1 Model 2 Model 3 Model 4 Odds ratio SE Odds ratio SE Odds ratio SE Odds ratio SE Migration Status Rural non-migrants 1.00 - 1.00 - 1.00 - 1.00 - Urban-Urban migrants 2.14*** .48 1.91** .44 0.78 .23 0.78 .24 Urban-Rural migrants 1.28 .46 1.09 .40 1.04 .41 0.93 .39 Rural-Urban migrants 1.88* .50 1.80* .49 0.88 .29 1.10 .38 Rural-Rural migrants 0.58 .20 0.59 .21 0.80 .30 0.88 .35 Urban non-migrants 2.30*** .49 2.07*** .45 1.10 .29 0.90 .25 Gender Male 1.00 - 1.00 - 1.00 - Female 0.62*** .10 0.58** .10 0.76 .15 Age at first premarital sex 10-14 1.00 - 1.00 - 1.00 - 15-16 2.40** .79 2.19* .74 2.06* .71 17-18 3.47*** 1.1 2.73** .89 2.38** .80 19+ 4.20*** 1.4 3.33*** 1.12 2.58** .91 Religion Muslims 1.00 - 1.00 - Catholics 1.02 .30 0.79 .25 Protestants/Other Christians 1.12 .26 0.88 .22 Traditional/Others 2.11 1.8 3.18 3.0 Ethnic origin Hausa/Fulani/Kanuri 1.00 - 1.00 - Igbo 10.91*** 7.1 5.57* 3.9 Niger-Delta groups 7.19** 4.8 3.66† 2.6 Middle-Belt groups 7.49** 4.9 4.27* 3.0 Yoruba 17.85*** 11.0 9.90*** 6.6 Others 3.27† 2.1 1.85 1.3 Childhood place of residence Rural 1.00 - 1.00 - Urban 2.09** .59 1.61 .48 Education Attained No education 1.00 - Primary level education 1.87 1.6 Secondary education 3.02 2.5 Tertiary education 6.04* 5.1 Status/Type of employment Unemployed 1.00 - Formal employment 0.70† .14 Agriculture/Self employment 0.53 .21 Living arrangement Adolescent child of head 1.00 - Adolescent head of household 1.92* .60 Adolescent spouse of head 0.26** .10 Adolescent relative of head 0.65* .14 Household wealth index Poorest 1.00 - Poor 1.14 .41 Middle 1.61 .48 Rich 2.25** .51 Richest Index of media exposure 1.00 - Lowest exposure 1.02 .21 Medium exposure 0.83 .35 Highest exposure LR Chi2 (d. f) 33.49 (5) 70.76 (9) 155.22 (18) 223.72(32) † p<.10 *p<.05, **p<.01, ***p<.001 75 The strong association between the Yoruba and condom use at premarital sexual initiation may be located in their high exposure to the urban environment as urban-urban and urban non-migrants (34. 6% and 35% of Yoruba youths in the sub-sample are urban- urban and urban non-migrants respectively). The same can also be said of the Igbo in a lesser dimension and as rural-urban migrants. This analysis indicates that unprotected premarital sex is gendered. Female adolescents are 38% and 42% less likely to use condom at premarital sexual debut than male adolescents in Models 2 and 3 respectively. In the full Model 4 there remains a 24% female disadvantage in protective premarital sexual debut relative to male adolescents, though the outcome is no longer statistically significant. This finding is consistent with the conclusion drawn from a four African country study (Ghana, Burkina Faso, Malawi and Uganda), which suggests that girls are likely to have first sex because they were forced or that the partner insisted (APHRC, 2007). Under such circumstances, the ability to negotiate condom use seems remote. An interesting dimension of the result suggests that sexual initiation among young men is more likely to involve the use of condom and may therefore be safer. Consistent with findings in other settings, adolescents who initiated sex in later years are more likely to do so using condom than those who initiated sex at younger years. This outcome is consistent in all models, pointing to the likelihood of protective premarital sexual behavior as young people mature in age and perhaps able to negotiate for safe sexual encounters. An intriguing aspect of the result is the significant variation in condom use at premarital sexual debut that exists between adolescents from different ethnic groups. In 76 Model 3 the Igbo are 10.9 times and the Yoruba are 17.8 times as likely as the Hausa/Fulani/Kanuri to use condom at premarital first sex. Groups in the Niger-Delta are 7.2 times and those in the Middle-Belt are 7.5 times as likely to use condom as the reference group. While the strength of these estimates is reduced in the full model, the magnitude of the variations remains consistent and significant. These outcomes find support in findings in other parts of Africa, and reiterate the suggestion that social and cultural contexts are primary determinants of adolescent sexual behaviors (APHRC, 2007). Residence in an urban area before age 12 is a significant independent predictor of condom use at premarital sexual initiation. In Model 3, adolescents resident in an urban area before age 12 are 2.09 times as likely to use condom at premarital sexual initiation as those that reside in rural areas for the same period. In the full model, the direction of the association remained at 39% advantage, though the estimate is no longer statistically significant. This outcome is consistent with the adaptation principle of the social environment theories, which finds plausible explanation in the advantages offered by the urban environment for adolescent access to knowledge and condoms relative to rural areas. Among the time-variant covariates (characteristics measured at the time of the survey) education, living arrangement and household wealth status indicate significant association with adolescent use of condom at premarital sexual debut. As level of education increases, the magnitude of the association with condom use at premarital sexual initiation monotonically increases. In the full model, adolescents with tertiary 77 education are 6.04 times as likely as those with no education to use condom at premarital sexual initiation. Adolescents who live independently are more likely to initiate premarital sex using condoms than those who live with their parents. This outcome is consistent with the proposition that parental presence may undermine young people’s ability to exercise their knowledge in protecting themselves. The finding may be related to the suggestion that young people may be ashamed to seek sex-related information or buy condoms and use same when they are under their parents' supervision for fear of reprimand. The danger here is that they end up initiating unprotected premarital intercourse. Relative to adolescents living with their parents, those who initiated premarital sex but living in unions at the time of the survey are 74% less likely to have used condom at premarital sexual debut. This outcome suggests that early union formation may be a strong indicator of unprotected premarital sexual vulnerability. Similarly, adolescents who live with relatives are also vulnerable to unprotected initiation of premarital intercourse. The problem concerning this group is that their sexual behavior has rarely attracted focus, despite the high level of child fostering and apprenticeship prevalent in African societies generally and Nigeria in particular (Isiugo-Abanihe, 1983; Chukwuezi, 2001). Finally, adolescents from richest households are 2.24 times as likely as those from poorest households to use condom at premarital sexual initiation. Those from households whose SES are in between fall within the continuum. Again this may be connected to the fact that the use of condom indicates some level of exercising choice among participants and youths from the poorest families particularly females may lack financial capacity to exercise such independence. 78 5.3 Summary The low level of condom use at premarital sexual initiation (15.8%) is comparable to low prevalence among youths in most countries of sub-Saharan Africa; for example 15% in Tanzania, Malawi, and Ethiopia (Zlidar et al. 2003; William et al. 1999); and 15.2% in Angola (Ndola et al. 2005). However, the prevalence rate is lower compared to Lesotho (34% for men and 59% for women), Uganda (43% for men), and Ghana (18% for men and 27% for women) (Tumwesigye et al. 2005). The result underscores the potentially high levels of vulnerability to sexually transmitted infections (STIs) and unintended pregnancy among sexually initiated Nigerian youths. Against the backdrop of the proposition that contraceptive use at first intercourse predicts contraceptive use subsequently and consistently (Abma and Mcgill, 2007; Shafii et al. 2004), the result points to the enormity of challenge that must be surmounted in any effort at protecting the next generation of Nigerians from the consequences of unprotected premarital sexuality. The high levels of use reported for Uganda suggests the potential of focused campaigns to boost use (Finger, 1998). The result identifies urban-urban and rural-urban migrants, together with urban non-migrants, as more likely to use condoms at premarital sexual initiation relative to rural non-migrants. This outcome generally suggests a key role for present contact with the urban environment as a migrant or a non-migrant. The explanation may be located in the role of the urban area in access to information and condom than in rural areas. This conclusion is supported by the result that residence in an urban area before age 12 is significantly associated with condom uses at premarital sexual initiation. While rural- urban and urban-urban migrants are more likely to use condom at premarital sexual 79 initiation relative to rural non-migrants, these outcomes however, were no longer statistically significant in Model 3, which controls for the net effects of adolescent age, childhood place of residence and ethnic origin. The result indicates a strong age effect consistent with literature on condom use, reiterating the increasing protective value of sexual initiation at an older age. As the age of premarital sexual initiation increases, the odds of condom use at premarital sexual initiation increases. This suggests that empowering young people to abstain from and delay sexual initiation as a choice can make a significant difference in promoting protective behavior for STIs and successful transition to adulthood. The association between gender and condom use at premarital sexual initiation is supported by recent findings in other African countries, and consistent with the explanation that young women are often unprepared for their first sexual encounters due to coercion, insistence of their often older partners, and powerlessness in transactional sexual encounters. Therefore, better quality policies and programs will need to address female disadvantages to access and use of protective services as condom. The strong ethnic effect observed is very intriguing and finds support in the theoretical and empirical findings of the cultural diffusion theory generally employed in explaining the demographic transition, particularly the acceptance of contraceptive innovation and fertility decline in Europe. Proximate areas with similar socioeconomic conditions but dissimilar cultures were identified to have entered the fertility transition period at different times, whereas areas differing at the level of socioeconomic development but with similar cultures entered the transition at similar times (Knodel and van de Walle, 1979). It is therefore instructive that significant variation in condom use at 80 first premarital sex among Nigerian youths is located within ethnic boundaries in this study. These findings, which are consistent with findings in other parts of Africa reiterates the suggestion that social and cultural contexts are primary determinants of adolescent sexual behavior. In particular, opposition to sex education and condom use campaigns vary along ethno-religious lines in the country, underscoring the need for systematic strategies for different groups. The findings that adolescents with higher education are more likely than those with no education to use condom at premarital sexual initiation is consistent with other findings and underscores the opportunities provided by education for promoting awareness and empowerment, thereby promoting protected sexual practices. That youth from the richest backgrounds are more likely than those from the poorest households to use condom at premarital sexual initiation may be related to the cost of obtaining condoms. Researchers have strongly linked condom use in sub-Saharan Africa to access, and poverty constitutes a primary hindrance to accessibility. Finally, adolescents that live independent of parents and relatives are significantly more likely to use condom at sexual debut than otherwise. This finding advances more evidence of the hindering roles of parents and relatives in accessing and utilizing sexual- related information and services even when they are available. Sex remains part of the sacred norms in many African cultures and young people who live with their parents or relatives may be discouraged from accessing sex-related services such as condoms for fear of being identified as sexually promiscuous or being reprimanded if discovered. It is for similar reasons that health care providers are identified as unwelcoming to young people, consequently hindering them from seeking sexual and reproductive health 81 information and services in the region. The key problem here is the inadequacy of parental controls to successfully hinder premarital sexual engagement in the long run, often leading to clandestine premarital sexual activities fraught with dangers. The challenge for policy and programs will therefore be how to reconcile parental and relatives’ control of adolescent sexual behavior, with the need to create opportunities for access to protective services for youths that need protection. CHAPTER 6 DATA ANALYSIS: PERCEPTION OF RISK OF SEXUALLY TRANSMITTED INFECTIONS 6.0 Introduction Subjective perception of one’s own risk of sexually transmitted infection has been shown to be an important correlate of whether an individual adopts risk-reduction strategies (Kohler et al. 2007). In sub-Saharan Africa, despite the widespread knowledge about sexually transmitted infections, particularly in the context of HIV/AIDS epidemic, individual subjective perception of risk of STIs have been found to vary substantially, ranging from those who perceive themselves at great risk to those who perceive no risk at all. A more problematic dimension of perception in the region is the finding that significant proportion of men and women who considered themselves at no risk or at small risk of contracting HIV were actually at moderate or high risk. In the context of adolescent premarital sexual activities, inaccurate perception of risk becomes even more problematic when sexually initiated and active adolescents perceive themselves at no risk of STIs, particularly following the significant association between individual correct assessment of risk and the prevalence of condom use (Prata et al. 2006). Consequently, understanding the determinants of individual perception of risk of infection is identified as an essential step in the prevention of STIs particularly HIV/AIDS (Kohler et al. 2007). In this chapter, I examine risk perception of STIs among adolescents who have initiated premarital sex. The analysis specifically focuses on the characteristics of adolescents who perceive no risk of infections (the incidence of risk denial) relative to 82 83 those who admit some level of risk of STIs. The particular focus on premarital sexually initiated youths is critical since inaccurate risk assessment among the group is a direct indicator of the current level of risk associated with premarital sexual engagement among sexually engaged adolescents. I use data on 1,308 adolescents who have initiated premarital sex and have no missing information on HIV risk perception. 6.1 Perception of risk of sexually transmitted infection: Descriptive analyses The bivariate outcome in Table 7 shows that 45.7% of adolescents who have initiated premarital sex perceive themselves at no risk of sexually transmitted infections. This group, which fits a sexually high-risk adolescent group, is consistent with the disconnect found between what youths know about HIV/AIDS and what they do to protect themselves. In assessing their individual characteristics relative to those who admitted their risks, Table 7 shows that rural-urban migrants relative to other migrant streams are most likely to deny the risk of infections. Generally, adolescents with some level of education relative to those with no education, females compared to males, and Muslims than Christians, perceive themselves at no risk of sexually transmitted infections. Similarly, adolescents who currently live with their partners, those in formal employment, with highest media exposure and those from poorest households are most likely to indicate that they are at no risk of STIs despite premarital sexual initiation. Current place of residence shows no significant relationship with risk perception, but place of childhood residence prior to age 12 indicates a significant outcome, with urban residents more likely to indicate perception of no risk of infection. An important bivariate outcome is the association between year of study and perception of risk. While 58% of 84 Table 7 Perception of risk of STIs among the premarital sexually initiated adolescents (N= 1,308) Variable At Some Risk of Infection At No Risk of Infection Perception of STD Risk 55.3 45.7 Migration Status† Urban-Urban migrants 56.4 43.6 Urban-Rural migrants 69.3 30.7 Rural-Urban migrants 47.8 52.2 Rural-Rural migrants 58.0 42.0 Urban non-migrants 55.0 45.0 Rural non-migrants 56.3 43.7 Gender*** Female 51.0 49.0 Male 66.3 33.7 Current age 15-16 46.6 53.4 17-18 54.3 45.7 19-20 59.0 41.0 21-22 55.1 44.9 23-24 55.2 44.8 Religion*** Muslims 46.9 53.1 Catholics 67.5 32.5 Protestants/Other Christians 53.9 46.1 Traditional/Others 70.0 30.0 Ethnic origin*** Hausa/Fulani/Kanuri 69.2 30.8 Igbo 70.4 29.6 Niger-Delta groups 50.4 49.6 Middle-Belt groups 55.3 44.7 Yoruba 35.6 64.4 Others 65.4 34.6 Childhood residence* Rural 59.4 40.6 Urban 52.6 47.4 Year of survey*** 1999 41.6 58.4 2003 100 00.0 Education Attained* No education 62.7 37.3 Primary level education 48.3 51.7 Secondary education 57.3 42.7 Tertiary education 52.5 47.5 Type of employment*** Unemployed 57.5 42.5 Formal employment 48.5 51.5 Agriculture/Self employment 67.0 33.0 Living arrangements** Adolescent head of household 58.5 41.5 Adolescent as spouse of head 44.8 55.2 Adolescent as child of head 55.9 44.1 Adolescent as relative of head 61.3 38.7 Household wealth index** Poorest 49.8 50.2 Poor 47.0 53.0 Middle 59.8 40.2 Rich 59.2 40.8 Richest 58.9 41.1 Index of media exposure*** Low exposure 70.2 29.2 High exposure 40.4 59.6 χ2 ***p<.001 **p<.01 *p<.05 85 adolescents surveyed in 1999 perceived no risk of infection after premarital sexual initiation, all adolescents surveyed in 2003 acknowledged some level of risk of STIs. This may be an indicator of successful enlightenment campaigns over the period and/or the maturity of the HIV/AIDS epidemic to the extent that its ubiquity can no longer be denied. Finally, higher proportions of the Yoruba and adolescents from the Niger-Delta perceive no risk of STI/AIDS infections. While these associations are statistically significant, they are only suggestive as they may be likely biased by confounding variables. To account for potential bias, the independent association between risk perception and adolescent characteristics are refined in the multivariate logistic regression models summarized in Table 8. 6.2 Perception of risk of sexually transmitted infections: Multivariate analyses The results in Table 8 indicate that urban-rural migrants relative to rural non-migrants are consistently more likely to admit some risk of STI/AIDS infection in all models. For rural-urban migrants however, the outcome is in the opposite direction as they are significantly more likely to deny any risk of STI/AIDS infection in Models 1 – 3. This positive association with risk denial remains in the full model (rural-urban migrants are 47% more likely than rural non-migrants to perceive no risk of infection) though the relationship was no longer statistically significant. The disturbing implications of risk denial among rural-urban migrants relates to the finding in chapter 4 that rural-urban migrants are the most likely to initiate early premarital intercourse. However, the perception of no risk may objectively reflect the finding that rural-urban migrants are 86 more likely to use a condom at premarital sexual initiation than rural non-migrants as found in chapter 5. In all models, gender difference is significant, with females more likely to perceive no risk of infection. Again, this perception may reflect their objective assessment of their risks following their lower likelihood of initiating premarital sex relative to male youths. However, it may also be the logical extension of the tendency of female adolescents to under-report their sexual activities. Indicating perception of some STI/AIDS infection risk may be counter productive in maintaining a respectful image, which may be important in navigating societal expectations of a respectable young woman and the marriage market (Smith, 2003, 2004a). Adolescents in the older age categories are less likely to deny exposure to risk of sexually transmitted infection relative to those who are younger. This outcome may be linked to the cumulative effect of exposure time, which may have afforded older adolescents opportunities to experience more sexual partnerships, increased likelihood of knowing some other infected persons, and adequate time to adequately assess their risks. Relative to Muslims, Christians of all denominations are significantly more likely to admit some risk of sexually transmitted infections. This outcome again may be related to an objective assessment of risk as Christians from previous analysis, are more likely to initiate premarital sex than Muslims. Therefore Muslim perception of no risk of STIs reinforces the suggestion that groups who see themselves as less likely to engage in premarital sexual activities should objectively perceive low or no risks. However, it is plausible that adolescents may more likely deny exposure to risk for sexually transmitted 87 1 Table 8 Odds ratios predicting perception of no risk of STIs among premarital sexually initiated Nigerian adolescents (N=1,308) Variables Model 1 Model 2 Model 3 Model 4 Odds ratio SE Odds ratio SE Odds ratio SE Odds ratio SE Migration Status Rural non-migrants 1.00 - 1.00 - 1.00 - 1.00 - Urban-Urban migrants 1.00 .16 0.93 .16 0.86 .20 0.76 .19 Urban-Rural migrants 0.57† .17 0.54* .16 0.55† .18 0.53† .19 Rural-Urban migrants 1.41* .23 1.33† .22 1.60* .38 1.47 .37 Rural-Rural migrants 0.93 .20 0.88 .19 0.93 .21 0.85 .21 Urban non-migrants 1.05 .17 1.12 .18 1.07 .23 0.95 .22 Gender Male 1.00 - 1.00 - 1.00 - Female 1.90*** .25 2.17*** .30 1.73*** .28 Current age 15-16 1.00 - 1.00 - 1.00 - 17-18 0.75 .18 0.73 .18 0.75 .20 19-20 0.60* .14 0.57* .14 0.55* .14 21-22 0.74 .17 0.70 .17 0.63† .17 23-24 0.73 .17 0.70 .17 0.59* .16 Religion Muslims 1.00 - 1.00 - Catholics 0.55* .12 0.61* .14 Protestants/Other Christians 0.68* .12 0.73† .14 Traditional/Others 0.35 .26 0.41 .31 Ethnic origin Hausa/Fulani/Kanuri 1.00 - 1.00 - Igbo 1.43 .50 1.53 .61 Niger-Delta groups 3.14*** 1.1 2.86** 1.1 Middle-Belt groups 2.58** .88 2.77** 1.1 Yoruba 5.65*** 1.8 4.27*** 1.6 Others 1.61 .54 1.62 .61 Childhood residence Rural 1.00 - 1.00 - Urban 0.78 .16 0.75 .17 Education Attained No education 1.00 - Primary level education 1.19 .40 Secondary education 0.86 .28 Tertiary education 1.00 .39 Status/Type of employment Unemployed 1.00 - Formal employment 1.37* .20 Agriculture/Self employment 0.83 .21 Living arrangements Adolescent as child of head 1.00 - Adolescent head of household 0.84 .21 Adolescent as spouse of head 1.47* .29 Adolescent as relative of head 0.77 .12 Household wealth index Poorest 1.00 - Poor 0.81 .17 Middle 0.52** .12 Rich 0.59** .12 Richest 0.49*** .11 Index of media exposure Low exposure 1.00 - Higher exposure 3.73*** .52 LR chi2 (d. f) 11.0 (5) 41.4(10) 152.5 (19) 274.2 (32) No of observations N=1308 † p<.10, *p<.05, **p<.01, ***p<.001 1Year of survey is not included in the models since its predictive value for the outcome of interest is fully determined in the bivariate analysis above, as all adolescents surveyed in 2003 admitted that they are at some level of risk of sexually transmitted infections, indicating perhaps the undeniable ubiquity of HIV/AIDS and/or the success of education campaigns over the two survey years. 88 infections if such denial synchronizes with the normative expectations of their religious authorities. This perspective is reinforced by findings that religious interpretation of HIV/AIDS in Nigeria frequently promotes a moral understanding of risk. Accordingly, the implicit immoral construction of HIV/AIDS infection among religious leaders and followers alike exacerbate the extent to which people must hide sexual relationships and deny risk (Smith, 2004a). Finally, while the Igbo exhibit no significant difference in perception of risks compared to the Hausa/Fulani/Kanuri, significant differences exist between the reference category and youths from other ethnic groups. Model 4 shows that Yoruba adolescents are 4.27 times as likely as the Hausa/Fulani/Kanuri to deny risk of exposure to sexually transmitted infections. Further, youths from the Niger-Delta are 2.86 times, and those from the Middle-Belt are 2.77 times as likely as the Hausa/Fulani/Kanuri to perceive no risk of infection. While these outcomes may reflect the objective assessment of the risk of each group relative to another, they generally reflect the problematic disconnect between what people know about HIV/AIDS and what they actually do. The “state of denial” associated with each group, is discordant with their actual premarital sexual behavior. Earlier part of this analysis in chapter 4 indicates that each of these ethnic groups is significantly more likely to initiate early premarital sex relative to the reference group. This outcome continues to underscore the high level of risk from STI/AIDS to which sexually active youths may be subjected in each group. Among the time-variant covariates, educational attainment is not significantly associated with risk perception in the full multivariate model. This outcome together with the bivariate result in Table 7 runs contrary to the expectation that higher education 89 should expose adolescents to scientific knowledge that will rationally inform less risky sexual perception and choices, particularly if they have initiated premarital sex. Other results indicating the significant likelihood of denial of risk of STIs include adolescents employed in the formal sector relative to the unemployed; and those with the highest exposure to the mass media relative to the lowest exposed. These counter-intuitive outcomes suggest that those who are more likely to know the routes of STI/AIDS infections are more likely to deny their vulnerability to the consequences of their premarital sexual engagement. However, it may also mean that these groups are able to access protective services that hinder sexually transmitted infections and therefore perceive themselves at no risk of STIs, but the low level of use of condom suggests a remote possibility for this line of thought. Adolescents living with a spouse at the time of the survey are 47% more likely to deny exposure to the risk of STI/AIDS infection despite having initiated premarital sex. Since the outcome here reflects perception at the time of the survey, the estimate may not have any direct association with premarital sexual behavior. However, it may speak directly to the belief of these adolescents that marriage protects sexual activities and a demonstration of faith in the fidelity of their sexual partners. It may also be part of building a respectful image, as admitting perception of some risk may be indicative of lack of trust in a spouse or a self report of infidelity. A recent study among adolescent aged 15-24 in South West Nigeria identified expectation of mutual fidelity as a determinant of risk perception. Adedimeji et al. (2007) found that adolescents who live with spouses attributed their perception of no risk of STIs to having only one partner and trusting them. However, earlier studies point out that actual risk may be elevated for such 90 women because of the sexual behavior of their partners as evidence has shown that male infidelity/multiple partnerships is tolerated in most African societies (Caldwell et al. 1993), together with the context of high levels of sexual networking in Nigeria (Orubuloye et al. 1991; Messersmith et al. 2000). Finally, Model 4 indicates significant variation in perception of risk among youths from households with varying levels of socioeconomic status. Adolescents from the poorest households are more likely than those from the middle, rich and richest household categories to perceive no risk of sexually transmitted infection. Understanding the mechanisms that fully drive these perceptions require further studies beyond this analysis. However, it is important to underscore the glaring discordance between risk perception and corresponding behaviors identified in previous results. In particular, groups that are more likely to initiate premarital intercourse in chapter 4 are not significantly different in risk perception; and some are significantly more likely to deny any risk of sexually transmitted infections. These discordant outcomes may be pointers that existing perception of risks are based on faulty premises, or are outright misperception. The emanating challenge for policies and programs therefore is identifying and understanding ways to accurately assess both personalized and generalized risks, in order to achieve alignment of perception with actual sexual behavior. 6.3 Summary Understanding how men and women perceive their risk of HIV/AIDS infection is linked to understanding responses to the HIV/AIDS epidemic. This chapter of my dissertation 91 examines risk perception of STIs among premarital sexually initiated adolescents. The results indicate that 45.7% of sexually initiated adolescents perceive themselves as having no risk contracting an STI. These results are consistent with previous findings that knowledge and perceptions are not generally translated to protective sexual engagement among Nigerian youths (Adedimeji et al. 2007; Smith, 2003, 2004a). The multivariate analysis shows that rural-urban migrants, who have the highest propensity to initiate early premarital sex, are also more likely to deny risk of STIs. This discordant perception of risk relative to actual sexual behavior is also found among youths in formal employment relative to the unemployed, the Yoruba, Middle-Belt and Niger-Delta youths relative to the Hausa/Fulani/Kanuri. Further, Igbo youths who are 63% more likely to initiate premarital sex relative to the Hausa/Fulani/Kanuri, indicate no significant variation in perception of risk of infection relative to the group. Understanding discordance between perception and behavior has engaged public health and social science researchers, and HIV/AIDS literature has identified significant insights. In a study of migrant Igbo youths, Smith (2003) linked the disparity between behavior and perception of risk to “conception of risk in ethical and moral terms and the complex intertwining of collectively-shared moralities with individual assessments of ethics of personal behavior.” More elaborately, Smith’s research reveals that young Nigerian migrants do not perceive significant personal risk of HIV infection because they construct the risk of AIDS in ethical and moral terms, projecting immorality and danger onto imaginary others. Thus young people interpret the epidemic without internalizing their own risks (Smith, 2003). In another study in South Western Nigeria indicating some 92 rationalizations, Adedimeji et al. (2007) suggest that so long as sexual relationships do not occur within institutionalized prostitution, risks are severely underestimated. For adolescents living with spouses, their perception of no risk of STIs may indicate their confidence in the faithfulness of their partners. This becomes problematic for behavior change and control of infection if their partners did not maintain sexual fidelity, which is likely in the context of high levels of sexual networking among men and women in Nigeria (Orubuloye et al. 1991; Messersmith et al. 2000). In summary, the findings from this study suggest that adolescent risk perception of sexually transmitted infections flows from at least two distinct mechanisms. On the one hand is the perception that emanates from objective assessment of risks. This relates to adolescents whose risk perception is congruent with their premarital sexual behavior. This group of adolescents has initiated premarital sex and subsequently perceives some risk of contracting sexually transmitted infections. Among youths that objectively assessed their risks include those who initiated premarital sex, but used condom to protect themselves. It is reasonable to expect the 15 percent of sexually initiated adolescents, who indicated protective behavior at sexual initiation, to logically perceive no risk of sexually transmitted infections. On the other hand, there are perceptions that fit the problematic discordant model, emanating perhaps from youth rationalizations. This includes rural-urban migrant youths, who are the most likely to initiate premarital sex and yet are the most likely to deny risk of contracting sexually transmitted infections. There are also youths in formal employment, those from all other ethnic groups except the Hausa/Fulani/Kanuri, who are more likely to initiate premarital intercourse, yet deny risk of STIs. 93 Against the backdrop of inaccurate assessment of risk for men and women, and the importance of the behavioral approach in a successful effort to limit HIV/AIDS transmission (Ndola et al. 2006), the results of this analysis provides an important nationally representative evidence base relevant to intervention policies and programs that will seek to enhance young people’s ability to correctly assess their sexual risks, adopt sexually protective behavior and achieve a successful transition to adulthood. In the light of the propositions of rationalistic theories of human behavior that behavioral change in response to HIV/AIDS should in part be the consequence of a sense of personal vulnerability to infection (Toroitich-Ruto, 2001) key interventions in Nigeria need to include an aggressive educational approach to educate adolescents on the misperceptions about STIs including HIV/AIDS. CHAPTER 7 DATA ANALYSIS: ACTIVE PREMARITAL SEXUAL BEHAVIOR 7.0 Introduction In further assessment of the level of risk associated with premarital sexual behavior, I examine in this chapter the role of migration and other predictors in transition to active premarital sexual engagement among sexually initiated adolescents. Sexually active youths are at immediate risk of becoming pregnant and/or of acquiring a sexually transmitted infection and the younger among them are much less likely than older youths to use contraception (Child Trends Data Bank, 2006). A great deal of public health attention has been focused on developing programs designed to diminish sexual risk behaviors associated with sexually transmitted infections and pregnancy among sexually active youths. Accompanying this attention has been considerable controversy, particularly in the United States, around what might be the most effective interventions (Loewenson et al. 2004). On the one hand, the message that abstinence is the only absolute protection against these dangers is gaining acceptance and has shown some positive effect upon adolescents’ attitudes and sexual behavior, together with much political and financial support (Thomas, 2000). On the other hand, a corpus of research indicates that there has been little evidence supporting the efficacy of such programs (Kirby, 2001). While evaluating abstinence-only programs using behavioral indicators and long-term follow-up will help provide evidence for or against this debate in the United States, little is yet known on active premarital sexual behavior among youths and the role of migration and other predictors in developing countries. In the African context 94 95 however, the importance of addressing active sexual engagement among unmarried adolescents follows the consensus in the region that promoting positive behavior, particularly delaying sexual engagement, is a key to addressing significant hindrances to adolescent successful transition to adulthood, including stemming the spread of sexually transmitted infections and HIV (Caldwell, 2000; UNAIDS, 2000; Cleland and Ferry, 1995; van der Straten et al. 1995; Konde-Lule, 1991). As a research effort in this direction, the primary question addressed here is: Do adolescents who initiated premarital sex transit into being sexually active and what are the predictors? The comparative group is adolescents who have initiated premarital sex but have abstained from sex for at least three months prior to the survey, otherwise termed secondary abstinence. Following the lack of time-series data on sexual behavior from the time of sexual initiation, only sexual behavior of adolescents at the time of survey are modeled. Being sexually active is defined as having at least one incident of premarital sex in the three months preceding the survey. It is important to point out that secondary abstinence (a prolonged period without sexual activity among those who have already been sexually active) is defined by a group of researchers as abstinence for a period of up to one year (Brown et al. 2004; Abdool-Karim and Abdool-Karim, 2005: 273). However, the three months definition is adopted for this study, following other more functional and flexible perspectives (CDC, 2004; Child Trends Data Bank, 2006; Cleland et al. 2003; The Colorado Health Foundation, 2007). A total of 1,662 adolescents who have initiated premarital sex and have complete information relating to the subjects under consideration are examined. 96 7.1 Active premarital sexual behavior: Descriptive analyses The descriptive statistics in Table 9 shows that among adolescents examined, 64% were sexually active three months preceding the survey. Once premarital sex is initiated, the bivariate result and Chi-square test of association indicate that adolescent transition to active premarital sex is not significantly associated with adolescent migration status. A similar insignificant association is found for gender. Further, the results indicate that as age increases, the proportion of sexually active adolescents gradually but consistently increase. There is a significant ethnic variation in active premarital sexual engagement. The highest proportions of adolescents that are sexually active are from the Middle-Belt (73.9%) and the Niger-Delta (68.2%), while the Igbo and the Hausa/Fulani/Kanuri have the lowest proportions: 57.7% and 56.9% respectively. There is a statistically significant difference in being sexually active between the two survey years, with 71% being sexually active in 1999, while the corresponding proportion in 2003 is 60%. This continues to suggest, as in chapter 6, a positive behavior change in adolescent sexual activity in a four year period. Adolescents with the highest educational attainment show highest propensities to being sexually active relative to other educational attainment categories. There is an 11% difference between adolescents with the highest education and those with no education. While 76% adolescents who have transited into independent living are sexually active, the corresponding proportion for those who live with parents and relatives are 66% and 58% respectively. As expected, following the sexually stimulating mass media, youths with the highest media exposure indicate the highest propensity for premarital sexual 97 Table 9 Characteristics of sexually active never married adolescents three months prior to survey (N= 1,662) Variable Sexually active Secondary abstainers Active premarital sex 64.6 35.4 Migration Status Urban-Urban migrants 64.1 35.9 Urban-Rural migrants 59.3 40.7 Rural-Urban migrants 62.2 37.8 Rural-Rural migrants 58.7 41.3 Urban non-migrants 68.7 31.3 Rural non-migrants 65.3 34.7 Gender Male 64.2 35.8 Female 64.8 35.2 Current age** 15-16 60.7 39.3 17-18 63.1 36.9 19-20 65.1 34.9 21-22 65.6 34.4 23-24 66.8 33.2 Religion Muslims 68.5 34.2 Catholics 62.9 37.1 Protestants/Other Christians 68.5 31.5 Traditional/Others 53.9 46.1 Ethnic origin** Hausa/Fulani/Kanuri 56.9 43.1 Igbo 57.7 42.3 Niger-Delta groups 68.2 31.8 Middle-Belt groups 73.9 26.1 Yoruba 64.6 35.4 Others 65.5 34.5 Childhood residence Rural 62.8 37.2 Urban 65.9 34.1 Year*** 1999 70.8 29.2 2003 59.8 40.2 Education Attained* No education 60.0 40.0 Primary level education 57.6 42.6 Secondary education 65.6 34.4 Tertiary education 71.3 28.7 Status/Type of employment Unemployed 64.7 35.3 Formal employment 63.5 36.5 Agriculture/Self employment 68.4 31.6 Living arrangements*** Adolescent head of household 75.9 24.1 Adolescent as child of head 65.9 34.1 Adolescent as relative of head 58.1 41.9 Household wealth index Poorest 67.5 32.5 Poor 60.6 39.4 Middle 63.3 36.7 Rich 68.1 31.9 Richest 62.2 37.8 Index of media exposure*** Lowest exposure 60.0 40.0 Medium exposure 61.2 38.8 Highest exposure 72.5 27.5 χ2 ***p<.001 **p<.01 *p<.05 activity. While 72% of the highest exposed are sexually active the corresponding proportions for the lowest and medium exposed are 60% and 61% respectively. 98 7.2 Active premarital sexual life: Multivariate models The bivariate associations above are further examined using multivariate logistic regression models. In isolating the independent roles of covariates on being sexually active, four multivariate logistic regression models are executed and summarized in Table 10. Model 1 examined the role of migration status and Model 2 introduced time-invariant individual characteristics of adolescents. Model 3 included time-invariant community level variables and Model 4, which is the full model, controlled for all time-variant covariates. Consistent with the bivariate results, all four models show no significant association between migration status and being sexually active. A test of joint significance of all migration categories confirmed the no significant association. The result in Model 2 indicates that sexual activity increases with the chronological age of adolescents, and while this increase is very gradual, it is not statistically significant. Again a joint test of significance of all the age variable categories confirmed this outcome. In Model 3, cultural identity factors of religion and ethnic origin, as well as year of survey are significant sources of variation in being sexually active. These variations, though weak, hold true in the full Model 4. In the final model, Christian youths are generally less likely to be sexually active than Muslim youths, and this propensity is significant for Protestants Christians. This finding is intriguing to the extent that Muslim youths are less likely than Christian youths to initiate premarital sex, following strict control norms about youth sexual engagement. However, once this strict control is overcome, Muslim youths seem more likely to engage in a somewhat accelerated pace of premarital sexual engagement. This interpretation draws support from 99 perspectives on how migration affects fertility in high fertility populations. Demographers have documented the fertility reducing effect of disruptions due to migration, following spousal separation, which reduces the frequency of intercourse, and together with the stress related to the move, reduces fertility in the short run (Hervitz, 1985; Brockerhoff and Yang, 1994; Potter and Kobrin, 1982). However, fertility decline due to disruption is expected to be only temporary and a more normal or even somewhat accelerated pace of fertility is expected as migrants may engage in catching up fertility behavior to compensate for the disruption effect. Following this logic, it becomes plausible to suggest that strict control of adolescent sexual behavior among Muslims may act as a disruption to sexual engagement and therefore effective only in the short term. However, whenever they are able to navigate out of the structural barriers, their propensity to premarital sexual engagement becomes elevated in a somewhat “catching up” premarital sexual behavior. While being cautious with this analogy since further more direct analysis is needed for any certain conclusion, the result suggests that structural and mechanically imposed controls of adolescent sexual engagement may be counter-productive in the long run and ineffective overall. Further, that Protestant Christians are significantly less likely to be sexually active offers support to earlier findings identifying the particular influence of the increasingly popular and dominant evangelical and Pentecostal identity and cultural style as important for how young people navigate premarital sexual relationships in Nigeria. Smith (2004b) generally draws our attention to how religious interpretations of HIV/AIDS and moral judgments about sexual relationships seem clearly to inhibit sexually active young migrants from adequately protecting themselves. 100 Table 10 Logistic regression estimates predicting active premarital sexual life among never married Nigerian adolescents Variables Model 1 Model 2 Model 3 Model 4 Odds ratio SE Odds ratio SE Odds ratio SE Odds ratio SE Migration Status Rural non-migrants 1.00 - 1.00 - 1.00 - 1.00 - Urban-Urban migrants 0.95 .14 0.93 .14 0.97 .20 0.90 .19 Urban-Rural migrants 0.77 .19 0.77 .19 0.91 .23 0.92 .24 Rural-Urban migrants 0.87 .14 0.89 .14 0.83 .18 0.81 .18 Rural-Rural migrants 0.76 .15 0.76 .15 0.83 .17 0.87 .19 Urban non-migrants 1.16 .17 1.15 .17 1.17 .20 1.09 .20 Gender Male 1.00 - 1.00 - 1.00 - Female 1.05 .12 1.02 .12 1.10 .13 Current age 15-16 1.00 - 1.00 - 1.00 - 17-18 1.09 .22 1.13 .23 1.09 .22 19-20 1.18 .23 1.30 .25 1.18 .24 21-22 1.22 .25 1.32 .27 1.19 .25 23-24 1.28 .27 1.35 .29 1.15 .26 Religion Muslims 1.00 - 1.00 - Catholics 1.04 .21 0.92 .19 Protestants/Other Christians 0.76† .12 0.70* .12 Traditional/Others 0.52 .30 0.51 .31 Ethnic origin Hausa/Fulani/Kanuri 1.00 - 1.00 - Igbo 1.15 .40 1.15 .41 Niger-Delta groups 1.92† .67 1.90† .68 Middle-Belt groups 2.33* .81 2.31* .83 Yoruba 1.37 .44 1.33 .45 Others 1.83† .61 1.82† .63 Childhood residence Rural 1.00 - 1.00 - Urban 1.16 .20 1.17 .21 Year 2003 1.00 - 1.00 - 1999 1.70*** .19 1.51** .23 Education Attained No education 1.00 - Primary level education 1.03 .34 Secondary education 1.53 .49 Tertiary education 1.80 .68 Status/Type of employment Unemployed 1.00 - Formal employment 1.02 .13 Agriculture/Self employment 1.47† .31 Living arrangements Adolescent as child of head 1.00 - Adolescent head of household 1.54† .35 Adolescent as relative of head 0.78* .10 Household wealth index Poorest 1.00 - Poor 0.74 .14 Middle 0.88 .17 Rich 1.05 .19 Richest 0.77 .14 Index of media exposure Lowest exposure 1.00 - Medium exposure 0.93 .14 Highest exposure 1.19 .26 LR chi2 (d. f) No of observations N=1662 6.26 (5) 8.31 (10) 58.32 (20) 92.06 (33) † p<.10, *p<.05, **p<.01, ***p<.001 101 However, Smith (2004b) also identified how the fervent Christianity of the ‘born again’ young migrants created, or at least solidified, for some an important arena for individual choice, namely: refraining from premarital sexual intercourse entirely or ‘secondary abstinence’ for people who were sexually active. While this finding may not finally prove that fervent brand of Christianity is increasing the prevalence of premarital secondary abstinence its significance in this analysis provides further reason to examine the sexually protective potentials of the phenomenon. Following the logic expressed by Smith (2004b), the proportion of the population that entirely abstain from premarital sex or renewed abstinence and the possibility that some young people do seem to heed strongly promoted religious messages cannot be ignored. Among the ethnic groups, two outstanding variations relate to adolescents from the Niger-Delta and those from the Middle Belt, who are 1.90 and 2.3 times respectively, as likely as the Hausa/Fulani/Kanuri to be sexually active in the full model. This outcome taken together with the results in chapter 4, confirm that adolescents from the Niger Delta and the Middle Belt are not only more likely to initiate early premarital sex, but are also more likely to be sexually active relative to the Hausa/Fulani/Kanuri. A peculiar characteristic common to the two groupings relates to the complex amalgamation of heterogeneous ethnic minorities in the two regions in Nigeria. Apart from economic and political marginalization, these are the flash points of past and present religious and economic conflicts in Nigeria. Whether the conflict of cultural norms and values between groups and/or the various political and economic conflicts in these regions are related to active adolescent premarital sexual behavior cannot be categorically assumed here and further investigation is needed. 102 The results also show a significant reduction in the odds of being sexually active among adolescents between 1999 and 2003. In Model 3 the odds was 70% higher in 1999 but that reduced to 51% in Model 4. This outcome suggests an increasing level of secondary abstinence over the two periods of the survey, which would suggest behavioral change in response to the maturity of the AIDS epidemic and/or HIV prevention messages (Abdool-Karim and Abdool-Karim, 2005: 273). This result is consistent with the findings on perception of risk over the period in chapter 6, and also with findings in the United States reporting decline in teen sexual activity in recent years (Rector et al. 2003). However, the overall proportion for Nigeria (65%) remains high. Among time-variant covariates, adolescent who are in agriculture/self employment are 46% more likely to be sexually active than those who are unemployed and those who live independently are 54% more likely to be sexually active relative to those who live with their parents. Though these outcomes are marginally significant, they are consistent with findings in previous chapters indicating the role of independent living in facilitating adolescent premarital sexual exploration and experience. For policy and programmatic purposes, the findings here continue to illustrate the complexity of understanding the influences of premarital sexual behavior, the inadequacy of imposing simplistic explanations and the limitations of one-dimensional intervention strategies that ignore these nuances. 103 7.3 Summary This analysis identified few significant predictors of being sexually active, but reiterates the heterogeneity of youths, their diverse vulnerabilities and the inadequacy of one- dimensional explanations and intervention strategies. Muslim youths are less likely to initiate early premarital sex, but once initiated, they are more likely to be sexually active relative to Christian youths, particularly Protestants. Drawing inference from the relationship between migration disruption and fertility in high fertility societies, which tends to lower fertility in the short run but may lead to elevated fertility in the long run, as migrants engage in a ‘catching up’ fertility behavior to compensate for the time lost to migration disruption (Hervitz, 1985; Brockerhoff and Yang, 1994; Potter and Kobrin, 1982), it is logical to suggest that strict premarital sexual control of Muslims youths may have achieved delayed sexual debut in the short run, but fails to limit active premarital sexual activity once premarital sex is initiated. This outcome underscores the limitations of structurally imposed sexual controls in a society where premarital sexuality is common for other youths. It also cautions that successful control of adolescents sexual behavior in one dimension in the short run does not guarantee success in same and other dimensions in the long run. That Christian youth particularly Protestants are less likely to be sexually active after sexual debut finds support in previous findings that a fervent brand of Protestantism in Nigeria has sexual protective potential that may no longer be ignored. Smith (2004b) found that the increasingly popular and dominant evangelical and Pentecostal identity and cultural style, is significantly linked to both primary and secondary premarital sexual abstinence in Nigeria. Evidence from successful AIDS control in Uganda and Senegal, 104 that delaying sexual debut and partner reduction may be major players in reducing HIV transmission, together with significant religious virtues of abstinence and the moral underpinnings of HIV/AIDS in Nigeria (Smith, 2004b), suggest that it may be worthwhile to further explore how fervent Christianity through premarital sexual abstinence and fidelity is important in combating HIV/AIDS. This perspective is relevant for Nigeria following observations that public health programs are yet to harness the positive roles of religious beliefs and institutions. Further, the results continue to highlight the salience of cultural identities such as ethnicity. The consistent variation in the outcome for different groups points to the complexity of social contexts that exist within a heterogeneous society as Nigeria, and reiterates the point that conclusions about sexual behavior for Nigeria or sub-Saharan Africa will be too simplistic if they fail to account for these contextual nuances. Though the results suggest a 51% decrease in the odds of being sexually active between 1999 and 2003, the overall proportion (65%) of sexually active youths remains very high. Following the lengthening period of years spent unmarried in recent years and in many cases sexually active, (NRCIM, 2005), and the linkages between sexually active teenagers and emotional health problems including depression and suicide in the United States (Rector et al. 2003), addressing the determinants of being sexually active and reproductive health interventions in the form of education and services will remain a public health policy and program challenge in sub-Saharan Africa in many years to come. CHAPTER 8 SUMMARY AND CONCLUSION 8.1. Summary of study Following theoretical and empirical research that links geographical mobility to social disruptions and behavior change for migrants and non-migrants, this dissertation examines the linkages between internal migration and premarital reproductive behavior of young people in Nigeria. In particular, the project primarily focuses on adolescent migration status and its association with early initiation of premarital intercourse. Among adolescents who initiated premarital sex other related risky premarital sexual perceptions and behaviors such as non-use of condom at premarital sexual initiation, perception of no risk for sexually transmitted infections, and transition to active premarital sexual engagement are further examined. The significance of understanding the linkages between migration and reproductive behavior in Nigeria flows from an unprecedented growth of young people (63% of the 140 million population is under 25 years), their high rate of migration to urban and economic centers in search of livelihood opportunities in the context of increasing poverty, the fast pace of socio-cultural and economic changes unleashed by globalization, and the recent policy and program attention being focused on adolescents in the wake of the devastating effects of the HIV/AIDS epidemic among young people in sub-Saharan Africa (NRCIM, 2005; Isiugo-Abanihe and Oyediran, 2004; UN-HABITAT, 2003; APHRC, 2002; Brockerhoff and Brenan, 1998). 105 106 Using data from the 1999 and 2003 Nigeria Demographic and Health Surveys, the study seeks to (1) examine the linkages between adolescent migration status and premarital sexual debut; (2) assess the extent to which adolescent premarital sexual debut varies in urban and rural areas, among men and women, and Nigeria’s heterogeneous religious and ethnic groups; (3) identify other individual, family and community variables that significantly predicts premarital sexual initiation; and (4) examine whether the predictors of premarital sexual debut also predict other premarital sexual perceptions and behaviors: adolescent use or non-use of condom at premarital sexual debut, perception of risk of sexually transmitted infections and active premarital sexual behavior. The study systematically reviews available theoretical and empirical work that relates migration to reproductive behavior and develops a comprehensive multi-factor study framework that incorporates individual, household, and national level factors. To examine the proposed hypotheses on the association between migration, other predictors and adolescent premarital sexual behavior, the analysis of data employs univariate, bivariate and multivariate statistical tools using the STATA statistical software. For early premarital sexual debut, the data was transformed into an event history framework to account for right censoring in the data, and to obtain unbiased probability estimates of the timing of premarital sexual initiation. The Kaplan-Meier procedure is used to generate estimates of the survival curves, showing the ages at which a given percentage of adolescents initiate premarital sex. To examine the role of migration on premarital first sex, net of other relevant factors, discrete-time hazard regression models, which are appropriate models for binary responses, were specified and estimated. Following multiple data entries for a single adolescent, the analysis applies the Huber- 107 White clustering correction of standard errors with individual identification number as a clustering variable. For adolescents that have initiated premarital sex, I employ the simple binary logistic regression models to examine the predictors of condom use at first premarital intercourse, perception of risk of sexually transmitted infection, and active premarital sexual engagement. 8.2 Key findings and implications for policy, programs and theory One important finding of the study is the overall level of premarital sexual activity in the study population. Accounting for exposure time and probability of premarital sex for all adolescents, the Kaplan-Meier survival curve (showing levels of “survival” at every point of the observation period), shows that at 18 years1 about 30% of the sample have not initiated premarital sex and at 24 years of age, which is the end of the observation period in this study, about 20% of youths remain primary abstainers. While previous studies focus on the increasing prevalence of adolescent sexual activities and legitimately so, the evidence that significant proportions of youths are primary abstainers is worth emphasizing as the first line of adolescent protection from the fallouts of premarital sexual engagement. While there may be merits of arguments against ‘abstinence only’ programs, it is worth emphasizing that the significant level of abstinence observed in this analysis needs to be examined more closely, sustained and improved upon by appropriate abstinence promotion policies and programs. 1 International conventions established 18 years as the legal age of consent to a sexual union (see The Universal Declaration of Human Rights (1984); The Convention on the Elimination of All Forms of Discrimination Against Women (1979); The convention on the Rights of the Child(1989); and The African Charter on the Rights and welfare of the Child (1990). 108 Closely related to primary abstinence is the level of secondary abstinence observed among those who have initiated premarital sex. While 64.4% of those who initiated premarital sex are sexually active, the remaining 35.6% are secondary abstainers. Related to the above is the finding that odds of being sexually active decreased by 51% from 1999 to 2003. Though the overall level of premarital sexual activities remain high, these findings are indicators of positive behavior in recent years that lends support to the consensus in the region that promoting positive behavior, particularly delaying sexual debut, is a key to addressing significant hindrances to adolescent successful transition to adulthood, particularly those stemming from the spread of sexually transmitted diseases and HIV (Caldwell, 2000; UNAIDS, 2000; Cleland and Ferry, 1995; van der Straten et al. 1995; Konde-Lule, 1991). One of the most important findings from this study is the significant association between migration status and premarital sexual initiation. However, contrary to the hypothesis at the onset of the study, this association is only significant for rural-urban migration status, after controlling for the influence of other predictor variables. This independent outcome for rural-urban migrants lends strong support to the notion that exposure to a more sexually lenient urban environment different from the normatively traditional rural origins may increase the propensity of rural-urban migrants to premarital engagement in sex. Consequent upon this outcome, it may be important for policy initiatives for controlling HIV/AIDS, particularly related adolescent premarital sexual behavior, to address migration push factors in rural areas to discourage youth out- migration in the first place. Such policies must among other goals promote livelihood 109 opportunities in rural origins to discourage the overall rural-urban migration in search of livelihood in urban destinations. The study finds that the role of migration for premarital sexual initiation is however, weak for other migrant streams. Contrary to the hypothetical expectation at the onset of my dissertation and what obtains in developed countries of North America and Europe, the weak outcome may be related to the fact that adolescent migration in Nigeria does not seem to involve a significant loss of social capital. In the study sample, 66% of migrant youths and 75% of non-migrants live with their parents or relatives at the time of the survey. Consequently, while migrant youths are more likely to live independent of parents and relatives compared to non-migrants, the overall level of independent living among both groups is substantially low. This outcome reinforces previous findings in Senegal, Mali, Togo, Ethiopia, Nigeria and other African countries, which generally conclude that in seeking to diversify their risk and improve their odds of survival, migration is undertaken as a family survival strategy in the region (Casacchia et al. 2001; Gugler and Ludwar-Ene, 1995; Lockwood, 1990; Trager, 1995). In Nigeria in particular, migration for marriage or joining spouses and for business apprenticeship under a relation or a master are significant reasons for young peoples’ migration (Chukwuezi, 2001; Watts, I984; Trager, 1995; Hollos, 1991). In this context therefore, the weak association between other migration streams and premarital sexual initiation may be related to the likelihood of the presence of authority figures (parents and relatives) in the adolescent migrant and non-migrant households. Consequently, it is plausible to expect that as long as migration remains family-based, and not sine-qua-non to adolescent independent 110 living, the loss of social capital occasioned by such movements will be minimal and the independent effects of migration on premarital sexual engagement will also be marginal. Further, that migration does not generally result into substantial loss of social capital as in this study, underscores the weakness of theoretical models based in developed countries, where migration involves significant loss of social capital, in adequately explaining behavioral trajectories, social processes and their implications in developing countries such as sub-Saharan Africa. The findings show that controlling migration and adolescent living arrangements per se may not be the sole panacea for promoting positive premarital sexual behavior. Consistent with stated hypothesis at the onset of the study, other individual and household characteristics such as years of exposure to the risk of premarital sex, religious affiliation, employment in the formal sector and ethnic origin are important predictors. Among the key findings is the consistent variation in adolescent premarital sexual behavior associated with cultural identity factors of religion and ethnicity. While this is expected, the empirical support for the two outcomes is partial in relation to the stated hypothesis that guided this analysis. On religion, while Christians are more likely to initiate premarital sex, Muslim youths are more likely to remain sexually active once initiated and deny any risk of sexually transmitted infections. On ethnicity, there is a strong independent effect in the full model showing that Niger-Delta youths are 300% more likely to initiate premarital sex relative to the Hausa/Fulani/Kanuri. For the Yoruba, Middle-Belt and Igbo, relative to the Hausa/Fulani/Kanuri youths, the corresponding likelihood estimates are 219%, 218% and 163% respectively. On condom use at premarital sexual initiation, ethnic variations is so 111 great that the Yoruba, the Igbo, and the Middle-Belt youths are approximately 1000% , 550% and 400% respectively, more likely than the Hausa/Fulani/Kanuri to use condom after controlling for the effects of all covariates. The ethnic effect is also strong in perception of risk of STIs. However, the results indicate high level of risk denial among sexually initiated Yoruba, Niger-Delta and Middle-Belt youths, when compared to the Hausa/Fulani/Kanuri. These outcomes underscore the complexity of factors that must be sifted through in particular settings in attempts to understand adolescent reproductive behavior even within a particular society. One important implication of these great variations is the simplicity of studies that generalize about sexuality in sub-Saharan African societies without paying attention to the cultural nuances of different groups, which are significantly different in their influences, as this study have shown. Some aspects of the findings throw up challenges to some pre-existing focus on sexual behavior in the region. For instance, the hypothetical proposition that young women have higher propensity to premarital sexual engagement was not supported by this analysis. In chapter 4, female adolescents are 13% less likely to initiate premarital sex than males and the interaction terms on gender and ethnicity indicates that being female reduces the likelihood of premarital sex in all ethnic groupings except among the Hausa/Fulani/Kanuri ethnic groups, where female adolescents are 2.08 times as likely as their male counterparts to initiate premarital sex. Despite the peculiar case of female adolescents among the Hausa/Fulani/Kanuri, the result speaks more strongly to the higher propensity of young men to premarital sexual initiation. This outcome directs our attention to the need for gender balance in research aimed at understanding adolescent premarital sexual behavior and on programs to achieve their protection from the 112 consequences of premarital sexuality. This position finds support in the increasing recognition of the need to involve men in sexual and reproductive health initiatives and to understand their needs, perceptions and motivations in reproductive health matters (Isiugo-Abanihe, 2003; Bledsoe and Cohen, 1993). Beyond perspectives that boys have relatively few hazards and many potential benefits, including gratification and social prestige from engaging in early premarital sexual experience (Mensch et al. 2001), recent studies have identified the vulnerability of adolescent men to some of the problems faced by young women. Such problems identified include sexual coercion by older women and men, unwanted sexual touch, being made to touch someone sexually, having penetrative sex and being “rented” as prostitutes in the mines by older men with evidence coming from a spectrum of developing countries including South Africa, Nigeria, and Cameroon (Ganju et al. 2004; Jejeebhoy, S.J. and S. Bott. 2003; Ajuwon, A., 2003). The high propensity of early sexual initiation among Hausa/Fulani/Kanuri female adolescents relative to their males merits a closer attention, as it underscores the complexity of forces in operation among different groups in culturally heterogeneous societies like Nigeria. While this analysis did not specifically trace their sexual partners, it is logical to suggest that the premarital sexual experience of Hausa/Fulani/Kanuri female youths may be linked to the sexual activities of adult males beyond the age cohorts of their male counterparts. Consequently, while most female adolescents are embedded in the sexual network, their male counterparts are less likely to be in the network. However, further analysis beyond this study will be needed to arrive at a more certain conclusion of the mechanisms in operation here. 113 The finding of very marginal negative association between educational attainment and premarital sexual initiation contrary to the study’s stated hypothesis, throws up some questions worthy of consideration. The consensus in general theoretical literature and particular empirical findings in Nigeria that informed the earlier hypothesis is that cumulative years of education and being enrolled in school increases the risk of experiencing first sex, and that the typical Nigerian higher education environment offers opportunities for high levels of sexual networking, and lenient sexual lifestyles (Castro and Juarez, 1995; Jejeebhoy, 1995; Fatusi, 2004; Okonkwo et al. 2005). However, the outcome here seems to place caution on such sweeping conclusions, lending support to the primary proposition of the role incompatible hypothesis that adolescents focused on attaining higher education are less likely to gravitate towards risky sexual behaviors. Further, the result raises questions about what the education variable is actually measuring. Adolescents in school are interacting with complex social variables, including school curricula, peers, teachers, and school regulations. They are simultaneously experiencing biological maturity and sexually stimulating influences from the mass media and the global culture, all of which are sensitive to duration of exposure and adolescent age. In this study, following the control of the number of years of adolescent exposure to the risk of premarital sex, the role of education was drastically reduced. This outcome highlights the need for future studies to closely focus on collecting contextually relevant school variables that will directly capture the mechanisms generally clustered under educational attainment. The results from the analyses of the predictors of condom use at premarital sexual initiation, perception of risk of sexually transmitted infection, and being sexually active 114 collectively quantify the high levels of risks that premarital sexual engagement entailed among Nigerian youths. The high-risk context is indicated by the finding that only 15.8% of premarital sexually initiated youths used condom at sexual initiation. Following the proposition that contraceptive use at first sex is an indicator of future and consistent use, the low level of condom use at premarital sexual initiation in this study highlights the enormity of the challenge of achieving protective premarital sexual behavior among youths in Nigeria. Further, with 64% of adolescents who initiated premarital intercourse sexually active, and 45% perceive themselves at no risk of sexually transmitted infections, the motivation for protective premarital sexual behavior seems remote. Aspects of the findings support the preliminary hypothesis at the onset of the study that the predictors of premarital sexual debut also predict other related premarital sexual behaviors (French and Dishion, 2003). However, what is also important is that key aspects of the results yielded contrary and behaviorally discordant outcomes. On the part of consistent results, female adolescents are less likely to initiate premarital sex, less likely to use condom at premarital sexual initiation and less likely to perceive themselves at risk of sexually transmitted infections than young men. These outcomes can be understood in the context of people living in their objective realities. Youths who experience less sex, will objectively perceive lower risks of sexually transmitted infections and less likely to use condoms. Again, older adolescents are more likely to initiate premarital sex, use condom and admit exposure to the risk of sexually transmitted infections. In the same light and consistent with stated hypothesis at the beginning of the study, adolescents from medium, rich and richest households are less likely to initiate premarital sex relative to those from the poorest households, more likely 115 to use condom at sexual initiation, and more likely to admit their exposure to the risk of sexually transmitted infections. Further, adolescents living independent of parents are more likely than those living with parents and relatives to initiate premarital sex, use condom at premarital sexual initiation and be sexually active. Also, ethnic origin consistently predicts in the same direction premarital sexual initiation, condom use at sexual debut, and being sexually active. To the extent that these outcomes are logically consistent, they provide evidence in support of the proposition that other adolescent premarital risk behaviors can be understood using predictors of sexual debut (French and Dishion, 2003). However, the findings include other sets of compelling results to the contrary. For instance, all youths from ethnic groups positive for premarital sexual debut, less likely to use condom and sexually active, deny any risk of sexually transmitted infections. Other discordant results include rural-urban migrants who are more likely than rural non-migrants to initiate premarital sex, but not as strong in using condom at sexual initiation, and yet are more likely to deny any exposure to the risk of sexually transmitted infections. Again, Christian youths more likely to initiate premarital sex than Muslims, show no significant variation in condom use at such initiation. While Muslim youths are less likely to initiate early premarital sexual intercourse, they are more likely to be sexually active and less likely to admit the risk of exposure to sexually transmitted infections. Further, living with parents compared to living independently significantly reduces the likelihood of early premarital sexual initiation, but it hinders the use of condom at premarital sexual initiation and predicts the denial of risk of exposure to sexually transmitted infections. Finally, adolescents in formal employment are more 116 likely than the unemployed to initiate premarital sex, less likely to use condom at premarital sexual initiation and yet more likely to deny the risk for sexually transmitted infections. The inconsistencies in these outcomes point to the fluidity of premarital sexual attitudes and behaviors, the danger of rigidity and the need for dynamism in policy and programs in pursuing interventions. While explanations are proffered for outcomes throughout the discussion of the results, some particular discordant outcomes and explanations relevant for intervention policies and programs are worth noting here. That Muslim youths are less likely to initiate early premarital sex, relative to Christians is consistent with strict sexual control of young people associated with Islamic frontiers in Nigeria. However, that Muslim youths are more likely to be sexually active after premarital sexual initiation, suggest that strict premarital sexual control of Muslims youths may have achieved “disruption” in sexual debut in the short run, but fails to limit active premarital sexual engagement in the long run. The outcome underscores the limitations of structurally imposed sexual controls and cautions that successful control of adolescents sexual behavior in one dimension does not guarantee success in other dimensions. Another inconsistent finding relates to Christian youths, particularly Protestants, who are more likely than Muslims to initiate early premarital sex but less likely to be sexually active after sexual debut. This outcome may be related to previous findings among mostly migrant Christian youths, indicating that fervent brand of Protestantism in Nigeria have sexual protective potentials that may no longer be ignored. Smith (2004b) found that the increasing evangelical and Pentecostal revival and the lifestyle associated with it is significantly linked to both primary and secondary premarital sexual abstinence. 117 This finding is relevant for Nigeria following observations that public health programs are yet to harness the positive roles of religious beliefs and institutions in combating HIV/AIDS, particularly as it relates to promoting behaviors around sexual abstinence and fidelity. One intriguing finding relates to the association between the presence of parents and relatives in a household and condom use at premarital sexual debut. This result supported by other findings, suggest that parents and relatives hinder adolescents in accessing and utilizing sexual-related information and services. Young people who live with their parents or relatives may be discouraged from accessing sex-related services such as condoms for fear of being identified as sexually promiscuous or being reprimanded if discovered. It is also for similar reasons that health care providers are identified as unwelcoming to young people, consequently hindering them from seeking sexual and reproductive information and services in the region. In the circumstances of increased youth sexuality and high dangers of STIs including HIV/AIDS, these barriers need to be frontally addressed by policies and programs. In sum, the concordant and discordant outcomes speak to the heterogeneity of youths, the diverse predictors of their vulnerabilities, and the complexity of social contexts that exist within groups in one country. These results underscore the overriding need to account for the contextual nuances that are likely to influence premarital sexual behaviors and the relative currency of studying each problem behavior separately and detailedly. However, it is important to balance the significance of culture-context with some underlying motivations to risky sexual behavior that are not culture-bound. Of particular importance here is the strong and positive age effect for not only premarital 118 sexual initiation but also condom use at such sexual debut. This factor, which relates primarily to biological and perhaps social maturity, is relevant following evidence from various parts of the developing world that age at puberty has been falling slowly over decades for both boys and girls due to primarily improvements in nutrition (Pasquet et al. 1999; El-Tawila et al. 1999). It is equally important following the changing context of sexual debut, with a greater likelihood now than in the past that first sex will be experienced prior to marriage in many countries following delays in the age of marriage (NRCIM, 2005) as more youths enroll in schools and seek livelihood opportunities in the dwindling or stagnant formal economies of developing countries. Finally, the strength of the findings of this study relates to how its key findings are consistent with key theoretical perspectives in the literature and empirical findings in other developing countries. The key strength of the study lies in its representativeness of all sections of Nigeria, affording us an opportunity to simultaneously examine premarital sexual behavior among all Nigeria’s heterogeneous adolescent groups. This strength is relevant for comprehensive national policy initiatives and enhanced generalizability of findings to other developing countries contexts. The significant association between rural-urban migration and early premarital sexual initiation is consistent with theoretical and empirical findings in other settings regarding the role of migration for elevated probability of risky sexual behavior. Accordingly, general strategies to bring about behavioral change among youths, promote and encourage healthier sexual relationships, and accomplish complete physical, mental and social well-being of the reproductive process in Nigeria and other settings will benefit from closer consideration of the key findings of this study. 119 8.3 Future Research Understanding contextual variations and the mechanisms that produce observed outcomes matter among heterogeneous groups within the country. Qualitative research into these variations and mechanisms remain part of my future research agenda in seeking to deepen our understanding of adolescent premarital sexual behavior. A key part of my future agenda is to seek ways and means of integrating aspects of these findings into policy and program goals to protect the next generation from diseases and death. One key limitation of the study is the lack of a nationally representative data set that directly measures migration histories and simultaneously incorporates sexual behavior variables. The importance of examining the linkages between the two variables, as highlighted in both the theoretical and empirical aspects of this study, places the collection of nationwide data that appropriately incorporates these key issues at the top of the agenda for further research on the subject. Such data collection enterprises need to assume a longitudinal framework or at least collect detailed migration histories and include deliberate time-markers for retrospective examination of time-variant events. The focus on adolescent women in sexual behavior research requires considerable re-focus in the light of the results of this study, and my future research will continue to seek to create gender balance in this enquiry. In the same light, further understanding will be served if attempts are made to seek further clarifications on measuring the educational attainment variable as identified above. Despite my attempt to control for many aspects of the individual, family and community level variables, I recognize the point made by Fletcher (2007) that in empirical work on adolescent choices, omitted variables are ubiquitous. Because data in 120 this case mostly focus around ages 15-24, preadolescent factors cannot be taken into account. This reiterates the importance of future work on other data sets on the subject. The body of evidence from the results of this study suggests that understanding primary and secondary premarital sexual abstinence should be a worthwhile future research venture that need not be discountenanced despite current high incidence of sexual activity. Exploring how to exploit the potential positive role of religious groups to promote positive sexual-related behavior such as primary and secondary abstinence and fidelity will be a worthwhile aspect of future work in this area in Nigeria. The results hold the potential of providing evidence that will inform multifactor and holistic policies and programs in dealing with the challenge of youth premarital sexuality. In particular the challenge of specific educational policies and programs that will address misperception of risks of STIs should be one of the central points of this agenda. Also, is the need to create an adolescent friendly environment that will simultaneously integrate positive reproductive behavior socialization and control, and provision of reproductive health services supported at home by parents and at work by service providers. Finally other policies already outlined above including promoting livelihood opportunities for young people in their rural origins to weaken rural push- factors that engender adolescent rural-urban migration are important dimensions of the next steps. 121 References Abdool Karim, S.S. and Abdool Karim, Q. ed (2005). HIV/AIDS in South Africa. Cambridge: Cambridge University Press. Abma, J. and McGill, B. (2007). Teenagers’ use of contraceptives at first intercourse: Long-term trends in use, correlates, and predictors for males and females. Paper presented at the Population Association of America meeting, New York. Adedimeji, A.A. (2005). Beyond knowledge and behavior change: The social-structural context of HIV/AIDS risk perceptions and protective behavior among young urban slum inhabitants in Nigeria. Boston, Massachusetts: Department of Population and International Health Harvard School of Public Health Adegbola, O., and Babatola, O. (1999). Premarital and extramarital sex in Lagos, Nigeria. In I.O. Orubuleye, J.C. Caldwell and J.P.M. Ntozi (Editors), The Continuing HIV/AIDS Epidemic in Africa: Responses and Coping Strategies. Canberra, Australia: Health Transition Center, Australian National University. Adepoju, A. (1983). Selected Studies on Dynamics, Patterns and Consequences of Migration: Medium-Sized Towns in Nigeria. Paris, UNESCO, Vol. IV. Pp 7-8. ______. (1986). Rural Migration in Nigeria. Ile Ife, Nigeria: Department of Demography and Social Statistics, University of Ife, Ile Ife. Aja-Nwachuku, I. (2004). Abortion among Non-School Youths in Nigeria. Paper presented at the 2004 Annual Meeting of the Population Association of America, Boston, Massachussets, U.S.A. Ajuwon, A.J., W. McFarland, E.S. Hudes, . Adedapo, T. Okikiolu, and P.Lurie. (2002). 122 HIV risk-related behavior, sexual coercion, and implications for prevention strategies among female apprentice tailors in Ibadan, Nigeria. AIDS and Behavior, 6 (3): 229-235. Amazigo, U., Silva, N., Kaufman, J., and Obikeze, D. (1997). Sexual Activity and Contraceptive Knowledge and Use among In-School Adolescents in Nigeria. International Family Planning Perspectives, 23(1): 28-33. Andersson, J.A. (2001). Mobile Workers, Urban Employment and ‘Rural’ Identities: Rural-Urban Networks of Buhera Migrants, Zimbabwe, In: M. Dedruijn, R. Van Dijk and Dick Foeken (Eds.), Mobile Africa: Changing Patterns of Movement in Africa and Beyond. Lieden, the Netherlands: Brill. APHRC (African Population and Health Research Center). (2007). Young people and HIV/AIDS in Africa: Not providing information is proving “deadly’ and costly. Nairobi: African Population and Health Research Center ______. (2002). Population and health dynamics in Nairobi informal settlements, Nairobi, Kenya: African Population and Health Research Center. Arowujolu, A.O., Ilesanmi, A.O., Roberts, O.A., Okunola, M.A. (2002). Sexuality, contraceptive choice, and AIDS awareness among Nigeria undergraduates. African journal of Reproductive Health, 6: 60-70. Auerbach, J.D., Wypijewska, and K.H. Brodie (eds). (1994). AIDS and Behavior: An Integrated Approach. Washington DC: National Academy Press. Aziken, M.E., P.I. Okonta and A.B. Ande. (2003). Knowledge and Perceptions of Emergency Contraception Among Female Nigerian Undergraduates. International Family Planning Perspectives, 29(2): 84-87. 123 Bandura, A. (1989). Perceived self-efficacy in the exercise of control over AIDS infection. In V.M. Mayes, G.W. Albee and S.F. Schneider (Eds), Primary Prevention of AIDS: Psychological Approaches (pp. 128-141. London: Sage Publications. Bankole, A., B. Oye-Adeniran, S. Singh, I.F. Adewole, D. Wulf, and R. Hussain. (2006). Unwanted Pregnancy and Induced Abortion in Nigeria: Causes and Consequences. Guttmacher Institute. Bartle, P.F.W. (1981). Cyclical Migration and Extended Community: A West African Example, in: R.B.Mandal (Ed.), Frontiers in Migration Analysis. New Delhi: Concept Publishing Co. Beck .S., Cole., B., and Hammond, J., (1991). Religious heritage and Premarital Sex. Journal of the Scientific Study of Religion, 30, 173-180. Blanc, A.K. and A.A. Way. (1998). Sexual Behavior and Contraceptive Knowledge and Use among Adolescents in Developing Countries. Studies in Family Planning, 29(2): 106-116. Bledsoe, C. and Cohen, B. (1993). Social Dynamics of Adolescent Fertility in Sub- Saharan Africa.Washington DC: National Academies Press, Blum, R.W. (2002). Mother’s Influence on Teen Sex: Connections That Promote Postponing Sexual Intercourse. Minneapolis: Center for Adolescent Health and Development, University of Minnesota. Bongaarts, J. and Cohen, B. (1998). Introduction and Overview: Adolescent Reproductive Behavior in the Developing World. Studies in Family Planning, 29(2): 99-105. 124 Brockerhoff, M. and Biddlecom, A. (1999).Migration, Sexual Behavior and Risks of HIV in Kenya. International Migration Review, 33(4): 833-856. Brockerhoff, M. and Brenan, E. (1998). The poverty of cities in developing countries. Population and Development Review, 24 (1):75-114. Brockerhoff, M. and Yang, X. (1994). Impact of migration on fertility in sub-Saharan Africa. Social Biology, 4(1): 19-43. Brown, L., A.H. Karim, P. Hutchinson, S. Agha, K. Maclntyre, and R. Magani. (2004). Church and Clubs: Influence on secondary abstinence in KwaZulu-Natal, South Africa. http://paa2004.princeton.edu/downloads.asp? Caldwell, J.C. (2000). Rethinking the Africans AIDS epidemic. Population and Development Review, 26, 117-135. Caldwell, J.C. , J.K. Anarfi, and P. Caldwell. (1997). The Social Context of AIDS in sub-Saharan Africa. Population and Development Review, 15(2):185-234. Caldwell, J.C. and Caldwell, P. (1993). The nature and limits of the sub-Saharan AIDS epidemic: Evidence from Geographic and other patterns. Population and Development Review, 19, 817-848. Caldwell JC, Santow G, Orubuloye IO, Caldwell P, Anarfi J, editors (1993). Sexual networking and HIV/AIDS in West Africa. Health Transition Review, 3(Supplement): 1-191. Casacchia, O., M. Crisci and C. Reynaud. (2001). Chapter 3: Internal Migration in Ethiopia.In: Migration and Urbanization in Ethiopia, with Special Reference to Addis Ababa. www.irpps.cnr.it/etiopia/sito/progetto3.htm Castro, M and Juarez, F. (1995). The impact of women education on fertility in Latin 125 America: Seraching for Explanations. International Family Planning Perspectives 21 (2): 52-80. CDC. (2006). Physical Dating Violence Among High School Students, United States, 2003. Morbidity and Mortality Weekly Report, 55(19): 532-535. Chattopadhyay, A., M. J. White and C. Debpuur. (2006). Migrant fertility in Ghana: Selection versus adaptation and disruption as causal mechanisms. Population Studies, 60(2): 1-15. Child Trends Data Bank (2006). Sexually Active Teens. www.childtrendsdatabank.org Chirwa, W. C. (1997). Migrant labour, sexual networking and multi-partnered sex in Malawi. Health Transition Review, Supplement 3 to Vol. 7: 5-15. Chukwuezi, B. (2001). Through Thick and Thin: Igbo Rural-Urban Circularity, Identity and Investment, Journal of Contemporary African Studies. Clayton, R.R. and Bokemeir, J.L. (1980). Premarital Sex in the Seventies. Journal of Marriage and the Family, 42: 34-50. Cleland, J. and Ferry, B.(1995). Sexual behavior and AIDS in the developing world. London: World Health Organization/Taylor and Francis. Cleland, J. et al. (2006). Sexual abstinence, contraception, and condom use by young African women: a secondary analysis of survey data. The Lancet, 368: 1788-1793. Coleman, J.S. (1990). Foundations of Social Theory. Cambridge, Massachussets: The Belknap Press of Havard University Press. Davidson, J.K. and Leslie, G. (1977). Premarital Sexual Intercourse: An Application of Axiomatic Theory Construction. Journal of Marriage and Family, 39: 15-25. Decosas, J. and V. Pedneault. (1992). Women and AIDS in Africa: Demographic 126 Implications for Health Promotion. Health Policy and Planning, 7(3): 227-233. Ekpunobi, C.I. (2003). Migration into Abuja environmental effects. Daily Trust. Abuja, Nigeria. Elder, G. H., M. K. Johnson, and R. Crosnoe. 2003. The Emergence and Development of Life Course Theory. In J.T. Mortimer and M.J. Shanahan (eds.) Handbook of the Life Course. New York: Kluwer Academic/Plenium Publishers. Elias , C. and L. Heise. (1993). The development of Microbicides: a new method of HIV prevention for women. Population council Programs Division Working Paper No. 6. New York: The Population Council. El-Tawila, S., Ibrahim, B., El-Gibaly, O., El Sahn, F., and Lee, S. (1999). Transition to adulthood: A national survey of Egyptian adolescents. Cairo, Egypt: Population Council. Evans, J. (1987). Introduction: Migration and Health. International Migration Review, 21(3): v-xiv Family Care International. (1995). Commitment to Sexual and Reproductive Health and Rights for All: Framework for Action. New York: Family Care International. Fatusi, A.O. (2004). Study of African universities’ response to HIV/AIDS: the Nigerian universities. Report submitted to the Association of African University, Ghana, June, 2004. Federal Ministry of Health. (2001). A Technical Report on the 2001 National HIV/Syphilis Sentinel Survey among Pregnant Women Attending Antenatal Clinics in Nigeria. Abuja: Department of Public Health, National AIDS/STDs Control Programme, Federal Ministry of Health. 127 Feyisetan, B. and Pebley, A.R. (1989). Premarital sexuality in Urban Nigeria. Studies in Family Planning, Vol. 20, No. 6: 343-354. Fichter, M. M., Elton, M., Diallina, M., and Koptagel, I. (1988). Mental Illness in Greek and Turkish adolescents. European Archives Psychiatry and Neurological Sciences, 237: 125-134. Filmer D. and Pritchett, L. (1998). Estimating wealth effects without income or expenditure data-or tears: Educational enrolment in India. World Bank policy Research Working Paper No.1994. Washington DC: The World Bank. _____. (1999). The effects of Household Wealth on Educational Attainment: Evidence From 35 Countries. Population and Development Review 25(1): 85-120. Finger, W.R. (1998). Condom Use is Increasing. Network, 18(3): 20-23. Firebaugh, G. (1997). Analyzing repeated surveys. Sage University Paper Series on Quantitative Applications in the Social Sciences, No. 07-115. Thousand Oaks, California: Sage Publications. Fix, M. and Zimmerman, W. (1997). Immigrant Families and Public Policy: A Deepening Divide. in Immigration and the Family: Research and policy on US Immigrants, Booth, A., Crouter. A.C., and Landale, N., editors. Mahwah, New Jersey: Lawrence Earlbaum Associates. Fletcher, J.M. (2007). Social Multipliers in High School Students’ Sexual Initiation Decisions. Demography, 44(2): 373-388. Fontanet, A., and Piot, P. (1994). State of our knowledge: the epidemiology of HIV/AIDS. Health Transition Review, 4 Suppl:11-22 French, D.C. and Dishion, T.J. (2003). Predictors of Early Initiation of sexual Intercourse 128 among High-Risk adolescents. The Journal of Early Adolescence, Vol. 23, No. 3, 295-315. Gage, A.J. (1998). Sexual Activity and Contraceptive Use: The Components of the Decisionmaking Process. Studies in Family Planning, Vol. 29, No.2:154-166. Gage-Brandon, A. and Meekers, D. (1993). Sex, Contraception and Childbearing Before Marriage in Sub-Saharan Africa. International Family Planning Perspectives, 19(1): 14-18+33. Ganju, Deepika, Shireen Jejeebhoy, Vijaya Nidadavoluand, K.G. Santhya, William Finger, ShyamThapa, Iqbal Shah and InaWarriner.2004. Sexual Coercion: Young men’s experiences as victims and perpetrators. www.popcouncil.org. Geschiere P, and Gugler, J. (1998). The Urban-Rural Connections: Changing Issues of Belonging and Identification, Africa 68(3): 309-319. Goldstein, S. and A. Goldstein. (1983). Migration and Fertility in Peninsular Malaysia: An Analysis Using Life History Data. Santa Monica: Rand Corporation. Gugler, J. (1971). Life in a Dual System: Eastern Nigerians in Towns 1961. Cahiers d Etudes Africaines, 11:400-421. Gugler, J. (1991). Life in a Dual System Revisited: Urban-rural Ties in Enugu, Nigeria, 1961-87. World Development19 (5): 399-409. Gugler, J and G. Ludwar-Ene, 1995. Gender and Migration in Africa South of the Sahara, In: Baker, J. and Aina, T.A. (Eds.), The Migration Experience in Africa. Uppsala, Sweden: Nordiska Africainstitutet. Gupta, N. (2000). Sexual Initiation and Contraceptive Use among Adolescent Women in 129 Northeast Brazil. Studies in Family Planning, 31(3): 228-238. Gurmu, E., Sidney Goldstein and Alice Goldstein. (2000). Migration, Gender and Health Survey in Five Regions of Ethiopia: 1998. (A United Nation Training and Research Project on the Interrelations of Migration and Economic Change, Women’s Status, Reproduction and Health). Addis Ababa: Addis Ababa University and Providence: Brown University. Gwatkin, D.R., S. Rustein, K. Johnson, R.P. Pande, and A.Wagstaff. (2000). Socioeconomic Differences in Health, Nutrition, and Population in Kenya. Washington, DC: HNP/Poverty Thematic Group, The World Bank. Hanna, J.W. and Hanna, J.L. (1981). Urban Dynamics in Black Africa: An Interdisciplinary Approach. New York: Aldine Publishing Company. Hervitz, H.M. (1985). Selectivity, adaptation or disruption? A comparison of alternative hypotheses on the effects of migration on fertility: the case of Brazil. International Migration Review, 19(2): 293-317. Hogan, D.P. and E.M.Kitagawa. (1985). The Impact of Social Status, Family Structure and Neighborhood on the Fertility of Black Adolescents. American Journal of Sociology, 90: 825-855. Hollos M., and Leis, P.E. (1989). Becoming Nigerian in Ijo society, New Brunswick [N.J.] New Brunswick [N.J.]: Rutgers University Press. Hugo, G.J. (1993). Migrant Women in Developing Countries. In Internal Migration of Women in developing Countries. New York: United Nations. Hunt, C.W. (1989). Migrant Labor and Sexually Transmitted Disease: AIDS in Africa. Journal of Health and Social Behavior, 30: 353-373. 130 Isiugo-Abanihe, U.C. (1993). Sexual behaviour and Exposure to the Risk of AIDS in Nigeria. Faculty of Social Sciences Lecture Series No. 8, Ibadan, Nigeria: University of Ibadan ______. (2003). Male Role and Responsibility in Fertility and Reproductive Health in Nigeria. Lagos: Ababa Press Ltd. ______and Oyediran, K.A. (2004). Household Socioeconomic Status and Sexual Behaviour among Nigerian Female Youth. African Population Studies, 19(1): 81-98. Janz, N.K. and Becker, M.H. (1984). The Health Belief Model: A Decade Later. Health Education Quarterly, 11(1):1-47. Jejeebhoy, S.J. (1995). Women’s Education, Autonomy, and Reproductive behavior. Experience from Developing Countries. Oxford: Clarendon Press. Jejeebhoy, S.J. and S.Bott. (2003). Non-consensual sexual experience of young people: A review of the evidence from developing countries. Population Council Working Paper No. 16. Kaufman, C.E., Clark, S., Manzini, N., May, J. (2002). How Community Structures of Time and Opportunity Shape Adolescent Sexual Behavior in South Africa. Working Papers No.159, New York: The Population Council Kiragu, K and Zabin, L.S. (1993). The Correlates of Premarital Sexual Activity Among School-Age Adolescents in Kenya. International Family Planning Perspectives, 19(3): 92-97+109. Kirby, D. (2001). Emerging Answers: Research Findings on Programs to Reduce Teen Pregnancy. Washington, DC: The National Campaign to Prevent Teen Pregnancy. 131 Kohler, H-P., Behram, J.R., and Watkins, S.C. (2007). Social networks and HIV/AIDS risk perceptions. Demography, 44(1): 1-33. Konde-Lule, J.K. (1991). The effects of urbanization on the spread of AIDS in Africa. African Urban Quarterly, 6(1 &2): 13-18. Konde-Lule, J.K., Sewankambo, N., Morris, M. (1997). Adolescent sexual networking and HIV transmission in rural Uganda. Health Transition Review, 7 (Supplement): 89-100. Knodel, J., van de Walle, E. (1979). Lessons from the past: Policy implications of historical fertility studies. Population and Development Review 5 (2), 217-245. Krieger, N., D.R. Williams, and N.E. Moss (1997). Measuring Social Class in Us Public Health Research: Concepts, Methodologies, and Guidelines. Annual Review of Public Health, 18: 341-378. Lacey, L. (1985). Inter-urban flows of population and occupational skills to three cities in Nigeria. International Migration Review, 19(4): 686 – 707. Landsdale, N. S. and Oropesa, R.S. (2001). Migration, Social Support and Perinatal Health: An Origin-Destination Analysis of Puerto Rican Women. Journal of Health and Social Behavior, 42: 166-183. Langstone, J. (1989). AIDS: The facts. Canada: Little Brown and Company. Larson, J. (1989). Social Context of Human Immunodeficiency Virus Transmission in Africa: Historical and Cultural Bases of East and central African Sexual Relations. Review of Infectious Diseases, 11(5): 716-731. Lee, B.S. and S. Faber. (1984). Fertility adaptation by rural-urban migrants in developing countries: the case of Korea. Population Studies, 38(1): 141-156. 132 Lindstrom, D.P., G-E. Kiros, and D.P. Hogan (2005). First intercourse, marriage, and first birth among educated Ethiopian Women: An African exception? Paper presented at the 2005 annual meeting of the Population Association of America, Philadelphia, March 31-April 2. Liska, A.E. (1981). Perspectives on deviance. Englewood Cliffs, NJ: Prentice Hall Lockwood ,V. S. (1990). Development and Return MIgration to Rural Polynesia. International Migration Review 24(2): 347-371. Loewenson, P.R., Ireland, M. and Resnick, M.D. (2004). Primary and secondary abstinence in high school. Journal of Adolescent Health, 34(3):209-15. Luke, N. (2003). Age and Economic Asymmetries in the Sexual Relationships of Adolescent Girls in sub-Saharan Africa. Studies in Family Planning, 34(2): 67-86. Lurie, M. N. (2004). Migration and the Spread of HIV/AIDS in South Africa. Paper presented at the 2nd Wits-Brown-Colorado Colloquium on Emerging Population Issues, March 29-30. Providence: Brown University. Lurie, M.N., A. Harrison, D. Wilkinson, and S. Karim. (1997). Circular Migration and Sexual Networking in Kwazulu/Natal: Implications for the Spread of HIV and Other Sexually Transmitted Diseases. Health Transition Review Supplement 3, 7: 17-27. Makinwa, P.K. (1981). Internal Migration and Development in Nigeria: Lessons From Bendel State. Ibadan: Heineman Educational Books (Nig) Ltd. Makinwa-Adebusoye, P. (1992). Sexual Behaviour, Reproductive Knowledge and Contraceptive Use among Urban Youth in Nigeria. International Family Planning Perspectives, 18(2): 66 - 70. 133 Massey D. S.; J. Arango; G. Hugo; A.Kouaouci; A. Pellegrino; J.E. Taylor (1994) An Evaluation of International Migration Theory: The North American Experience. Population and Development Review, 20(4):699-751. Mberu, B.U. (2007). Household Structure and Living Conditions in Nigeria. Journal of Marriage and the Family, 69 (May 2007): 513–527. ______. (2006). Internal Migration and Household Living Conditions in Ethiopia. Demographic Research, 14(21): 509-540. ______. (2005). Who moves and who stays? Rural out-migration in Nigeria. Journal of Population Research, 22(2): 14 1- 161 McCaulay, A.P. C. Salter, K. kiragu, and J.Senderowitz. (1995). Meeting the needs of Young adults. Population Reports Series. Journal of Family Planning Programmmes, 41:1-43. McCollum, A. (1990). The trauma of moving. Beverly Hills, Carlifornia: Sage. Meekers, D. (1994). Sexual Initiation and Premarital Childbearing in Sub-Saharan Africa. Population Studies, 48(1): 47-64. Mensch, B.S., Clark, W.H., Lloyd, C.B., Erulkar, A.S. (2001). Premarital Sex, Schoolgirl Pregnancy and School Quality in Rural Kenya. Studies in Family Planning, 32(4): 285-301. Mensch, B.S., Bruce, J., and Greene, M.E. (1998). The uncharted passage: Girls’ adolescence in the developingg world. New York: Population Council. Messersmith, L.J., T.T. Kane, A.I. Odebiyi, and A.A. Adewuyi, (2000). Who is at risk? Men’s STD experience and condom use in Southwest Nigeria. Studies in Family Planning, 31(3): 203-216. 134 Miller, B.C. and Moore, K.A. (1990). Adolescent Sexual Behavior, Pregnancy, and Parenting: Research through the 1980s. Journal of marriage and Family, 52(4):1025-44 Montgomery, M., M. Gragnolati, K. Burke, and E. Paredes. (2000). Measuring Living Standards with Proxy Variables. Demography, 37(2): 155-174. Moreno, L. (1994). Residential Mobility and Contraceptive Use in Northeast Brazil. DHS Working Paper, No 9. Calverton, MD: Macro International Inc. Mugisha, F., Jacqueline Arinaitwe-Mugisha and Bilhah Hagembe. (2003). Alcohol, substance and drug use among urban slum adolescents in Nairobi, Kenya. Cities, 20(4): 231-240. Mugusha, F. and Zulu, E.M. (2004). The influence of alcohol, drugs and substance abuse on sexual relationships and perception of risk to HIV infection among adolescents in the informal settlements of Nairobi. Journal of Youth Studies, 7(3): 279-293. National Intelligence Council. (2002). The Next Wave of HIV/AIDS: Nigeria, Ethiopia, Russia, India, and China. Washington, D.C. National Intelligence Committee. National Population Commission and ORC/Macro (2003) Nigeria Demographic and Health Survey, 2003.Calverton, Maryland: National Population Commission and ORC/Macro. Ndola, P., L. Morris, E. Mazive, F. Validnia and M. Stehr. (2006). Relationship between HIV Risk Perception and Condom Use: Evidence from Population-Based Survey in Mozambique. International Family Planning Perspectives, 32(4): 193-200. Ndola, P., F. Vahidnia and A. Fraser, (2005). Gender and Relationship Differences in 135 Condom Use Among 15-24 Years Olds in Angola. International Family Planning Perspectives, 31(4): 192-199. NRCIM (National Research Council and Institute of Medicine).(2005). Growing Up Global: The Changing Transitions To Adulthood In Developing Countries. Lloyd, C.B. (Editor).Washington, DC: The National Academies Press. NISER (Nigeria Institute for Social and Economic Research). (1997). Nigeria Migration Survey, 1993. Ibadan, Nigeria: Nigeria Institute for Social and Economic Research. Okonkwo, P.I., Fatusi, A.O., and Ilika, A.L. (2005) Perceptions of peers’ behaviour regarding sexual health decision making among female undergraduates in Anambra State, Nigeria. African Health Sciences, 5(2): 107-113. Oladele, O.O. and William, R.B. (1994). Rural-urban mobility in southwestern Nigeria: implications for HIV/AIDS transmission from urban to rural communities. Health Education Research, 9(4): 507-518. Onifade, I. (1999). Unmet reproductive health needs of adolescents. In The Continuing HIV/AIDS Epidemic in Africa: Responses and Coping Strategies, I.O. Orubuleye, J.C. Caldwell and J.P.M. Ntozi (Editors). Canberra, Australia: Health Transition Center, Australian National University. Oropesa, R.S. and Gorman, B. K. (2000). Ethnicity, Immigration and Beliefs about ‘Ties That Binds.’ in Ties that Bind: Perspectives on Marriage and Cohabitation, Waite, L.J., Bachrach, C., Hindin, M., Thompson, E., and Thornton, A., Edited. Hawthorne, Carlifornia: Aldine de Gruiter. Orubuloye, I. O., J.C.Caldwell, and P. Caldwell (1991). Sexual networking in the Ekiti 136 District of Nigeria. Studies in Family Planning, 22 (2): 61-73. Otoide, V.O., F. Orosanye, and F. Okonofua. (2001). Why Nigerian adolescents seek abortion rather than contraception: Evidence from focus-group discussions. International Family Planning Perspectives, 27(2): 77-81 Oucho, J.O. (1998). Recent Internal Migration Processes in Sub-Saharan Africa: Determinants, Consequences and Data Adequacy Issues. In R.E. Bilsborrow (ed.), Migration, Urbanization and Development: New Directions and Issues, New York: UNFPA and Kluwer Academic Publishers. Pasquet, P., Biyong, A.M. and Rikong-Adie, H. (1999). Age at Menarche and urbanization in Cameroon: Current status and secular trends. Annals of Human Biology, 26(1): 89-97. Peil, M. (1985). Changing Structures: A Democratic Comparison. Contemporary Urban Research 10(2): 76-91. Peterson, W. (1958). A General Typology of Migration. American Sociological Review, 23(2): 256-265. Pongou, R. (2007). Ethnic Identity, Strategic Cross-Ethnic Sexual Networks formation, and HIV/AIDS in Africa. Department of Economics (Mimeograph), Brown University. Population and Development Review. (2007). Report of Nigeria’s National Population Commission on the 2006 Census. Population and Development Review, 33 (1): 206-210. Population Reference Bureau (PRB) and Center for Population Options. (1994). A special focus on reproductive health. PRB and CPO Fact Sheets. 137 Portes, A. (1998). Social Capital: Its Origins and Applications in Modern Sociology. Annual Review of Sociology, 24: 1-24. Potter, R.G. and Kobrin, F. E. (1982). Some effects of spouse separation on fertility. Demography, 19: 75-95. Prata, N., Morris, L., Mazive, E., Vahidnia, F., and Stehr, M. (2006). Relationship between HIV risk perception and condom use: Evidence from a population-based survey in Mozambique. International Family Planning Perspectives, 32 (4): 192- 200. Puri, M. and J. Cleland. (2006). Sexual behavior and perceived risk of HIV/AIDS among young migrant factory workers in Nepal. Journal of Adolescent Health, 38: 237- 246. Quinn, T.C. (1994). Population Migration and the Spread of Types 1 and 2 Human Immunodeficiency Virus. Proceedings of the National Academy of Sciences of the United States, 91(7): 2407-2414. Rambault, R.G. (1997). Ties That Bind: Immigration and Immigrant families in the United States. In Ties that Bind: Perspectives on Marriage and Cohabitation, Waite, L.J., Bachrach, C., Hindin, M., Thompson, E., and Thornton, A., Edited. Hawthorne, Carlifornia: Aldine de Gruiter. Rambault, R.G. and Weeks, J.R. (1996). Unraveling a Public Health Enigma: Why Do Immigrants Experience Superior Perinatal Health Outcomes? Research in the Sociology of Health Care, 13B: 337-391. Rector, E. R., K.A. Johnson, and L.R. Noyes. (2003). Sexually active teenagers are more 138 likely to be depressed and to attempt suicide. Center for Data Analysis Report, #03-04. www.heritage.org Reiss, I., and Milller, B. (1979). Heterosexual permissiveness: A theoretical analysis. In W.Burr, R. Hill, F.I. Nye, and I. Reiss (Eds.), Contemporary theories about the family. New York: Free Press. Rosenstock, I., Strecher, V., and Becker, M. (1994). The Health Belief Model and HIV risk behavior change. In R.J. DiClemente, and J.L. Peterson (Eds), Preventing AIDS: Theories and Methods of Behavioral Interventions (pp.5-24). New York: Plenum Press. Rosenzweig, M. and Wolpin, K. (1988). Migration Selectivity and the Effects of Public Programs. Journal of Public Economics, 37: 265-289. Sampson, R. and Laub, J. (1993). Crime in the making. Cambridge, Massachusetts: Harvard University Press Shafii T, Stovel K, Davis R, Holmes K. (2004). Is Condom Use Habit Forming? Condom Use at Sexual Debut and Subsequent Condom Use. Sexually Transmitted Diseases 31(6): 366-372. Shanahan, M.J. (2000). Pathways To Adulthood in Changing Societies: Variability and Mechanisms in Life Course Perspective. Annual Reviews of Sociology, 26: 667-92. Sherraden, M. and R.E. Barrera. (1997). Family Support and Birth Outcomes among Second Generation Mexican Immigrants. Social Service Review, December: 607- 633. Singh, S., Darroch, J.E., and Bankole, A. (2003). A, B and C in Uganda: The roles of 139 abstinence, monogamy and condom use in HIV decline. Occasional Report No. 9. New York: The Alan Guttmacher Institute Slap, G.B., Lucy Lot, C.A.Daniyam, T.M. zink and P.A. Succop. (2003). Sexual Behavior of Adolescents in Nigeria: Cross Sectional Survey of Secondary Schools Students. BMJ; 326: 15-18. Smith, D.J. (1999). Having People: Fertility. Family and Modernity in Igbo Speaking Nigeria, PhD Dissertation. Dept of Anthropology, Emory University, Atlanta, Georgia. ______. (2000). ‘These girls today na war-o’: premarital sexuality and modern identity in southeastern Nigeria. Africa Today, 47(3/4): 98-120. ______. (2003). Imagining HIV/AIDS: Morality and perceptions of personal risk in Nigeria in Nigeria. Medical Anthropology, 22(4): 343-372. ______. (2004a). Premarital sex, Procreation, and HIV Risk in Nigeria. Studies in Family Planning, 35(4): 223-235. ______. (2004b). Youth, sin, sex in Nigeria: Christianity and HIV/AIDS-related beliefs and behavior among rural-urban migrants. Culture, Health and Sexuality, 6(5): 425-437. Stack, S. (1994). The Effect of Geographic Mobility on Premarital Sex. Journal of Marriage and Family, vol.56 (1): 204-208. The Colorado Health Foundation. (2007). The Colorado Health Report Card. The Colorado Health Foundation. Thomas, M.N. (2000). Abstinence-based programs for prevention of adolescent pregnancies: A review. Journal of Adolescent Health, 26: 5-17 140 Toroitich-Ruto, C. (2001). The Effect of HIV/AIDS on Sexual Behavior of Young People in Kenya. African Population Studies. 12: 2. 39-50. Trager, L. 1995. Women Migrants and Urban-Rural Linkages in Southwestern Nigeria, In: Baker, J. and Aina, T.A. (Eds.), The Migration Experience in Africa. Uppsala, Sweden: Nordiska Africainstitutet. Tumwesigye, N.M., Ingham,R., and Holmws, D. (2005) Condom use at first and most recent sex among unmarried people aged 15-24 in Uganada: A case of kabale and Mukono districts. Centre for AIDS Research (CAR) University of Southampton. United Nations. (1994). AIDS and the Demography of Africa. New York: United Nations United Nations. (2003). World Urbanization Prospects. The 2003 Revision. Geneva: United Nations Population Division. UNFPA. (2003). UNFPA State of the World Population: HIV/AIDS and Adolescents. New York: UNFPA. UN-HABITAT. (2003). The Challenge of Slums: Global Report on Human Settlements. London and Sterling: Earthscan. United Nations. (2004). World Urbanization Prospects, The 2003 Revision: Data Tables and Highlights. Department of Economic and Social Affairs, Population division, New York: United Nations. UNAIDS. (2000). Report on the global HIV/AIDS epidemic. Geneva Switzerland: UNAIDS. Vega, W. and Amaro, H. (1994) Latino Outlook: Good Health, Uncertain Prognosis. Annual Review of public Health, 15: 39-67. 141 van der Straten, A. King, R. Grinstead, O., Serufilira, A. and Allen, S. (1995). Couple communication, sexual coercion, and HIV risk reduction in Kigali, Rwanda. AIDS, 9, 935-944. Vold, G. and Bernard, T. (1986). Theoretical Criminology. New York: Oxford University Press. Watts, S.J. (1984). Marriage Migration, A Neglected Form of Ling-Term Mobility: A Case Study from Ilorin, Nigeria. International Migration Review 7(4):682-698. William, K.A., and C.S. Alexander (1999). Determinants of Condom Use to prevent HIV Infection Among Youth in Ghana. Journal of Adolescent Health, 24: 63-72 World Bank. (1996). Nigeria: Poverty in the Midst of Plenty: The Challenge of Growth with Inclusion. A World Bank Assessment. Washington DC, World Bank. World Bank. (1999). World Development Indicators, 1999. Washington, DC.World ank. Zambrana, R.E., N.C.Scrimshaw, C. Dunkel-Schetter. (1997). Prenatal Health Behaviors and Psychological Risk Factors in Pregnant Women of Mexican Origin: The Role of Acculturation. American Journal of Public Health, 87: 1022-1026. Zhou, M. (1997). Growing Up American: The Challenge Confronting Immigrant Children and Children of Immigrants. Annual Review of Sociology, 23:63-95. Zlidar, V.M. et al, (2003). New survey findings: the reproductive revolutions continues. Population Reports, Series M, No 17. Zulu, E.M., F. Nii-Amoo Dodoo and Alex Chika Ezeh. (2002). Sexual risk-taking in the slums of Nairobi, Kenya, 1993-98. Population Studies, 56: 311-323. 142 Appendices Kaplan-Meier survival curves for initiation of premarital intercourse by adolescent characteristics. Fig. 2 Kaplan-Meier survival estimates, by migration status 1.00 0.75 0.50 0.25 0.00 0 5 10 15 20 25 Analysis time Urban-urban migrant Urban-rural migrant Rural-urban migrant Rural-rural migrant Urban non-migrant Rural non-migrant 143 Fig. 3 Kaplan-Meier survival estimates, by gender 1.00 0.75 0.50 0.25 0.00 0 5 10 15 20 25 analysis time female male Fig. 4 Kaplan-Meier survival estimates, by age cohorts 1.00 0.75 0.50 0.25 0.00 0 5 10 15 20 25 Analysis time 15-16 17-18 19-20 21-22 23-24 144 Fig. 5 Kaplan-Meier survival estimates, by religion 1.00 0.75 0.50 0.25 0.00 0 5 10 15 20 25 Analysis time Catholic Protestant/Other Christians Muslims Traditional/Other adherents Fig. 6 Kaplan-Meier survival estimates, by childhood place of residence 1.00 0.75 0.50 0.25 0.00 0 5 10 15 20 25 Analysis time Urban Rural 145 Fig. 7 Kaplan-Meier survival estimates, by Current place of residence 1.00 0.75 0.50 0.25 0.00 0 5 10 15 20 25 Analysis time urban rural Fig. 8 Kaplan-Meier survival estimates, by Ethnic groups 1.00 0.75 0.50 0.25 0.00 0 5 10 15 20 25 Analysis time Hausa/Fulani/kanuri Igbo Niger delta groups Middle Beltt Groups Yoruba Other ethnic groups 146 Fig. 9 Kaplan-Meier survival estimates, by education level 1.00 0.75 0.50 0.25 0.00 0 5 10 15 20 25 Analysis time No education Primary Secondary Higher Fig. 10 Kaplan-Meier survival estimates, by media exposure index 1.00 0.75 0.50 0.25 0.00 0 5 10 15 20 25 Analysis time Lowest exposure Medium exposure Highest exposure 147 Kaplan-Meier survival estimates by Living arrangement 1.00 0.75 0.50 0.25 0.00 0 5 10 15 20 25 Analysis time Adolescent as head Adolescent spouse to head Adolescent child of head Adolescent relative of head Kaplan-Meier survival estimates, by Household SES 1.00 0.75 0.50 0.25 0.00 0 5 10 15 20 25 Analysis time Poorest Poor Middle Rich Richest 148 Kaplan-Meier survival estimates, by status/nature of employment 1.00 0.75 0.50 0.25 0.00 0 5 10 15 20 25 Analysis time Unemployed Formal employment Agriculture self employment/domestic