lllllllllllllllllllllllllllllllllllllllllllllllllllllllllllll 3 1293 01686 5044 This is to certify that the dissertation entitled DIRECT PRIVATE COSTS AND ACCESS TO SECONDARY SCHOOLING IN BURUNDI presented by Marie Mayoya has been accepted towards fulfillment of the requirements for Educational Ph.D. degreein Administration Major professor (a Date mm} /3/ //// fl MSUiJ an Affirmative Action/Equal Opportunity Institution 0-12771 O LIBRARY Michigan State University PLACE IN RETURN BOX to remove this checkout from your record. TO AVOID FINES return on or before date due. DATE DUE MTE DUE DATE DUE 1M W14 DIRECT PRIVATE COSTS AND ACCESS TO SECONDARY SCHOOLING IN BURUNDI BY Marie Mayoya A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Educational Administration 1997 DIRECT PRIVATE COSTS AND ACCESS TO SECONDARY SCHOOLING IN BURUNDI BY Marie Kayoya Secondary school is highly valued in Burundi, but it is a great financial burden on the parents of children in both day or boarding school. School fees are determined by the government, but other costs (school supplies, uniforms, and shoes, and transportation) have not previously been analyzed. The research investigated the direct private costs of secondary school and the determinants of access to secondary schooling in 1993. A multisite (Gitega, Karuzi, and in Bujumbura, the capital city), stratified sample of 197 families, including families with and without children in secondary school was used. The families included 1,161 children with 518 children enrolled in school and 635 children not in school in 1993. The mean of the direct private cost of secondary schooling per child was 26,256 FBu for boarding and 20,725 FBu for day schools. It varied according to school, gender of the student, family background, and location. It was higher for female students in both types of school. Family expenditure for schooling increased with parents education, higher paying occupation, income, and assets. Mothers with more education had more children in day school. The economic burden of the direct private cost of secondary schooling was 10 t of family income in boarding school and 3% of family income per child in day school. That economic burden was 23% for the lowest quintile income families for children in boarding school. It was 1% of the total family income for the highest quintile income. Farmer parents and fathers with no formal schooling did not have children in day schools. Family assets and education of the father were the main determinants of access to secondary schooling in the entire sample. 111 the urban area, family assets and fathers' education were the determinants of access to secondary schooling. In the rural area, predicted burden of secondary schooling, sixth grade repetition, children's chores, and parents' attitude to secondary school, were the determinants of access to secondary schooling. Educated members of the extended family played a role in access to schooling. Increased education of parents would improve access to secondary schooling. Analysis should made of the financial, institutional, and management dimensions involved in expanding compulsory education from the sixth grade to lower secondary school. A policy of scholarships for poor rural children and girls from poor families should be explored to help equalize educational opportunity. Copyright by Marie Mayoya 1997 DEDICATED TO MY PARENTS. MY HUSBAND, AND CHILDREN To my Dad, Mr. Mayoya Tharcisse, and my Mom, Mrs. Renata Ngenzebuhoro, for your prayers, values, and loving guidance. To my Mom, for choosing to send me to school instead of helping you with chores, staying up late to prepare my lunch after the others had gone to sleep, and getting up early to warm my breakfast. You warmed my heart every day. To my Dad, for teaching me the values of education and showing me the path to better schools and a better life. For your challenge to go farther than you were able to in school. To both of you, for your love of children. And to my husband Mevin Ndarusigiye, and children, Diane, Eric, and Claire with a promise to love you as much as I was loved. ("Urukundo ukunze umwana, arubikira uwiwe"). ACKNOWLEDGEMENT My heartfelt gratitude goes to: The Rockefeller Foundation which sponsored my data collecting. Without their generous financial support, the project would not have been possible. My dissertation co-directors Dr. Jack Schwille and Dr. Mun Tsang, for their unfailing support during the shaping of the dissertation proposal, data collection and throughout my graduate studies. I very much appreciated the visit by Dr. Jack Schwille during the data collection in Burundi and the funding by the Rockefeller Foundation that made that visit possible. Dr. Schwille helped in many ways to improve the quality of the data collected. I am equally grateful, as well, to my academic advisor, Dr. Mun Tsang, for his interest in the study his encouragement, and his commitment to working with me to help me reach my goal. I am indebted to the other two members of my dissertation committee, Dr. Gary Sykes and Dr. David Wiley, who have been most supportive, understanding despite their busy schedules. I am also really grateful to Dr. Richard Houang, and Mr. Jean Bosco Sibomana, for statistical help and comments and advice, to Mr. Ernest Habonimana, Mr. Nestor Riragendanwa, and Ndayishimiye Jean for their help in conducting interview. I am also thankful to all those who contributed to my training at Michigan State University, especially the College of Education. My appreciation as well to Ms. Martha Green, Mrs. Maggie Schuesler, for editing my dissertation. My thanks to: The government of Burundi for financing the early years of my graduate studies. The Centre Uhiversitaire pour la Recherche et le Developpement vi Economique (CURDES) dans La Faculty d’ Economie, a 1’ University du Burundi, for hosting me during the data collection in Burundi. The Ministries of Education and Interior for the permission to conduct this study. The Communal administrators and "chefs de zones", who facilitated contact with the population in rural Bujumbura, Gitega, Karuzi, and Mairie de Bujumbura. I shall never forget the warm welcome, the kind attention and the enthusiastic interest of my interviewees. The motivation they showed in answering my questions surely kept mine alive until the end of the dissertation. My heartfelt sympathy goes to those who saw secondary education as an unreachable dream for their children and who perceived it as only for "others". Maybe, one day, it will be theirs, too. I sincerely acknowledge the many informants who helped me locate the families with children in secondary school on their collines without them this project would have been impossible. I owe a large debt of gratitude to my family: To my parents who sent me to school, financed my studies, and taught me during primary schooling; you provided me with the foundation for doctoral studies. To my brothers and sisters for all your sacrifices. To my husband, Mevin Ndarusigiye, and children, Diane, Juste-Eric, and Claire M. Noelle, for your unconditional love, prayers, and support during my doctoral studies. To the family of Joseph and Magdalena Rudagaza and their children, especially Gregoire and Rose. My sincere thanks to all my friends everywhere for their continuing personal support. vii TABLE OF CONTENT LIST OF TABLES .............................................. LIST OF FIGURES . ........................................... LIST OF MAP ................................................. CHAPTER I: INTRODUCTION 1.1. Problem Statement OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO 1.2. Physical Characteristics and Population .......... 1.3. Educational System .............................. I.3.1. Origins of Unequal Access to Education ........................... 1.3.2. Formal Education .................... 1.3.2.1. Primary School ................ 1.3.2.1.1. Double Shifts ............ 1.3.2.1.2. Automatic Promotion ..... 1.3.2.1.3. Concours National ....... 1.3.2.1.4. Grade Repetition ........ 1.3.2.1.5. Para-Primary - Yaga Mukama ........ 1.3.2.2. Secondary School .............. 1.4. Direct Private Costs of Secondary Education ..... 1.5. Proposed Research .............................. 1.6. Organization of the Dissertation ................ CHAPTER II. LITERATURE REVIEW AND CONCEPTUAL FRAMEWORK II.1. Literature Review ............................... 11.1.1. Private Cost of Education ............ 11.1.2. Access to Schooling ................. 11.1.2.1. Rural Versus Urban Settings .......... 11.1.2.2. Girls in Rural Areas ................ 11.1.2.3. Family Background ................... 11.1.2.4. Dropout 11.1.2.5. Extended Family ..................... viii 11 11 12 14 16 18 22 23 25 25 30 3O 32 33 36 37 II.2. II.3. 11.1.2.6. Grade Repetition .................... Conceptual Framework ............................ II. II. II. 2.1. Human Capital Theory ................. 2.2. Status Attainment ................... 2.3 Accesss to Secondary Schooling in Burundi ....................................... Merit of the Study .............................. CHAPTER III. METHODOLOGY .................................... 111. 1. Pilot Study ................................... III.2. III.1. III.1. III.2. III.2. III.2. 1. Choice of the Site and Families ..... 2. Results of the Pilot Study .......... III.1.2.1. The questionnaire ............. III.1.2.2. Sampling Procedure ............ III.1.2.3. Number of Interviews per Day... Sampling for the Main Study ............... 1. Selection of Provinces .............. 2. Selection of Communes ............... III.2.2.1. Province of Karuzi ............. III.2.2.2. Province of Gitega ............. III.2.2.3. Bujumbura ...................... III.2.2.4. Problems of Data Collection 3. Sampling Families in the Selected Sites 111.2.3.1. Counting Families with Secondary School children 111.2.3.2. Buhiga ........................ 111.2.3.2.1. Sampling of Families with Children in Secondary School 111.2.3.2.2. Sampling of Families without Children in Secondary school III.2.3.3. In Gitega Commune ............. III.2.3.3.1. Selection of Zones ix 40 41 42 43 44 48 51 51 51 53 53 S4 56 S7 57 60 61 62 63 63 64 65 66 66 67 68 68 III.3. III.4. III.2.3.3.2. Sampling of Families with Children in Secondary School III.2.3.3.3. Sampling of Families without Children in Secondary school 111.2.3.4. Bujumbura ..................... III.2.3.4.1. Sampling of Families with Children in Secondary School III.2.3.4.2. Sampling of Families without Children in Secondary school Limitations of the Sample ................. Data Analysis ............................. III.4.1. III. III III. III III.4.2. III.4.3. III.4.4. III Weighting the Data .................. 4.1.1. Nyakabiga ..................... .4.1.2. Kamenge ....................... 4.1.3. Gitega ........................ .4.1.4. Buhiga ........................ Family as a Unit of Analysis ........ The Child as a Unit of Analysis ..... Analysis of the Cost of Education .4.4.1. Creating Data Files ........... III.4.4.1.1. File of all the Children .......... 111.4.4.1.2. File of Children Enrolled in Post Primary ........... 111.4.4.1.3. File of Children Enrolled in Secondary School ............ 111.4.4.1.4. File of Children Fully Supported by the Parents ........... 111.4.4.2. Breakdown of Entire Sample III.4.5. of Children by Age and Gender .................. Analysis of Direct Private Cost ..... 69 7O 71 71 72 75 77 77 79 8O 8O 82 87 88 89 89 89 90 91 92 94 98 III.4.5.1. Cost of Secondary Schooling per Child per Year ..... 111.4.5.2. Economic Burden of Secondary Education on the Parents 111.5. Access to Secondary Education .................. 111.5.1. Multiple Regression Analysis ............. 111.5.1.1. Independent Variables at the Family Level ................... 111.5.1.1.1. Father Education (FATHEDUC) ........ a) Father's Education Across the Regions ............ 111.5.1.1.2. Mother Education (MOTHEDUC) ........ a) Mother's Education Across Regions ................. 111.5.1.1.3. Main Occupation of the Father (FATHEROC): ............ 111.5.1.1.4. MOTHEROC: Mother's Occupation 111.5.1.1.5. INCOME: Total Yearly Nominal Monetary Income of the Family ...... 111.5.1.1.6. ASSETS: Total Assets of the Family ....... 111.5.1.1.7. New Variables in the family file ...................... 111.5.6. Parents Values and Beliefs about Secondary Schooling ........................... CHAPTER IV. RESULTS ON DIRECT PRIVATE COST OF SECONDARY EDUCATION IN BURUNDI IV.1. Type of Costs, Objectives and Sample ........... IV.2. Direct Private Cost of Secondary Schooling ..... 1V.2.1. Types of Public School ............. 1V.2.2. Costs Across Provinces ............. 1V.2.3. Gender Differences ... ............. IV.2.4. Parental Education .................. IV.2.4.1. Father’s Education (FATHEDUC) 1V.2.4.2. Mother's Education ............ xi 98 99 100 100 100 101 102 104 105 106 106 108 109 114 118 120 124 124 126 128 129 129 132 IV.2.S. Direct Private Cost and Parental Occupation ......................... IV.2.6. Family Wealth ...................... IV.2.7. Total Family Expenditure ........... IV.2.8. Direct Private Costs by Grade Level IV.2.9. Direct Private Costs and Number of Children in Secondary School ........ IV.2.10. Public vs Private Schools ........... 1V.3. Total Family Spending on Secondary Education ................................. 1V.3.1. Fathers' Education (FATHEDUC) ....... 1V.3.2. Mothers' Education (MOTHEDUC) ....... 1V.3.3. Fathers' Occupation ................. 1V.3.4. Mothers' Occupation ................. 1V.3.S. Income Quintile (INCOME) ............ 1V.3.6. Assets Quintiles (ASSETS) ........... 1V.3.7. Expenditure Quintile (EXPENSES) ..... 1V.3.8. Number of Children in Secondary School (INSEC) ...................... IV.4. Total Family Spending on Education .............. IV.5. Economic Burden of Direct Private Cost of Public Secondary Schooling on the Family 1V.5.1. Burden of DPC per Child ............. 1V.5.2. Burden of Total DPC for Public Secondary School per Family ......... 1V.5.3. Burden of Total Family Spending' on Education ........................ 1V.6. Correlation Among Variables .................... 1V.7. Summary of the Chapter .......................... CHAPTER V. DETERMINANTS OF ACCESS TO SECONDARY SCHOOLING IN BURUNDI v.1. Overview ............................................ v.2. Characteristics of Children in Secondary V.2. School ....................................... Characteristics of Helped Children ........... xii 133 137 139 141 145 147 148 149 150 150 151 151 151 152 152 154 157 157 162 162 165 169 171 172 172 V.2 2. Characteristics of Helpers .................... V.2.3. Characteristics of Children fully Supported by Parents ........................ v.3. Comparing Characteristics of Families With and Without Children in Secondary School ....................................... V.3.1. Matrix of Correlatio ......................... V.3.2. Parents Education ........................... V.3.2.1. MOTHEDUC ............................. V.3.2.2. FATHEDUC ............................ v.3 2 3 Family Income (INCOME) ............... V.3.2.4. Family Expenditure (EXPENSES) ........ V.3.2.S. Family Assets (ASSET) ................ V.3.2.6. Existance of Educated Relative in the Family ....................... V.3.2 7 Predicted Burden (PREDBURD) .......... V.3.2.8. Families with children Who Repeated Sixth Grade (REP) ........... V.3.2.9. Highest Number of Time for Sixth Grade Repetition (TIME6) ................... V.3.2.10. Chores ............................. V.3.2.1l. Areas (AREAS) ...................... v.4. Determinants of Access to Secondary Schooling ........ v.4.1. Predicted Burden Equation .................. v.4.2. Use of Logistical Regression .............. v.4.3. Logistic Regressions:Family as unit of. Analysis .................................... V.4.3.1. Entire Sample ..................... V.4.3.2. Urban Sub-Sample .................. v.4.3.3. The Rural Sub-sample (Gitega and Karuzi) ............... v.4.4. Logistic Regressions: Child As Unit of Analysis ....................... V.4.4.1. Entire Sample .................... v.4.4.2. Urban Subsample .................. V.4.4.3. Rural Sub-Sample ................. xiii 174 174 175 175 177 177 179 179 179 180 180 180 181 181 182 182 183 183 185 186 187 189 191 193 194 197 199 v.4.5. Discussion of Findings on Access to Secondary Schooling .................... CHAPTER VI. CONCLUSION V1.1. Summary of Findings ................................... v1.1.1. Direct Private Costs and Economic Burden ...... v1.1.2. Predicted Economic Burden ..................... v1.1.3. Access to Secondary Education .................. VI.1.3.1. Determinants of Access to Secondary Schooling ............ V1.2. Policy Implications ................................. v1.2.1. Economic Burden ............................. v1.2.2. Access to Secondary ......................... VI.2.2.1. The National Test .................... v1.2.2.2. Family Income ........................ v1.2.2.3. Gender Difference in Access .......... VI.2.2.4. Household Chores: Indirect Private Costs of Schooling ........... V1.3. Future Policy Analysis and Research .................. v1.3.1. Policy of Expansion of Lower Secondary Schoolin ........................... v1.3.2. Parents' Attitude Toward Schooling ............ v1.3.3. Replication of the Study ...................... APPENDICES A. TABLES: NUMBER OF CHILDREN IN SECONDARY SCHOOL ..... i ...... B. INTERVIEW QUESTIONNAIRE (IN ENGLISH) ..................... C. INTERVIEW QUESTIONNAIRE (IN KIRUNDI) ..................... D. LETTERS OF PERMISSION TO CONDUCT THE STUDY ............... BIBLIOGRAPHY ................................................ xiv 201 204 204 208 208 209 211 212 213 214 216 216 217 218 218 218 219 220 222 233 246 248 Table Table Table Table Table Table Table Table Table Table Table Table Table Table Table Table Table Table Table Table Table 10 11 12 13 14 15 LIST OF TABLES Gross Enrollment of Primary and Secondary Schools by Province, Burundi, 1991-92 .................... Number of Children Who Passed the Concours National by Sex and Region, Burundi, 1988-92 .............. Annual School Fees per Child and Percentage Increase for both Boarding and Day Secondary .............. Burden of School Fees with Respect to Family Income (%), Burundi, 1982-83 - 1991—93 School Years (Selected Years) ................................ Population Density, Number of Primary and Secondary Schools by Province, Burundi, School Year 1986-87. Summary of the Sampling .......................... Summary of Proportions ........................... Population and Sample Description By Commun ...... Proportion of the Sample and Weight by Commune and Families with and without Secondary School Students Student Distribution by Province and by Gender (FULSUP) School Expenses Fully Covered by Parents Student distribution by Grade Level (GRADE) and by Province in the Sample .................... Child Gender (GENDER) by Province (PROV). ....... Crosstabulation of Age and Grade Level of the Children Still in School, Burundi, 1992-93 ................ Highest Grade Completed for Children Who Had Already Left School in 1992-93 ................................ 16 17 18 19 20 21 FATHEDUC and MOTHEDUC in the Whole Sample and per Region. ..................................... Cross Tabulation of INSEC (Number of Children With Secondary Experience) and Father' Education (FATHEDUC) Cross Tabulation of INSEC (Number of Children in Secondary School) and the Level of Education of the Mother (MOTHEDUC) ............................... FATHEROC and MOTHEROC in the Whole Sample and per Region .......................................... Average Family Income by Source in the Sampled Areas, Burundi, 1992-93 ................................ Items of the Family Assets (ASSET), Burundi, 1993... XV 13 20 21 74 83 83 84 91 92 92 93 95 96 103 104 106 107 109 112 Table Table Table Table Table Table Table Table Table Table Table Table Table Table Table Table Table Table Table Table Table 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 Distribution of TIMES ........................... Distribution of Families with Number of Children Never Enrolled in Primary School, Burundi, 1992-93 Opinion About Secondary Schooling Direct Cost (OPWILPAY) Burundi, 1992-93 School Year ......... Opinion on the usefulness of secondary school (USEFSCHL) in Burundi, 1993 ..................... Annual (YCOST) and Itemized Direct Private Costs (FBu and percentage) in Both Private and Public school, Burundi, 1992-93 ........................ Direct Private costs of Public Secondary Boarding Schools by Region, Burundi, 1992-93 ............. Direct Private Costs of Secondary Education (in FBu and percentages) by Gender and Type of School, Burundi, 1992-93 ................................ Costs of Secondary Boarding School (FBu, percentages) by Fathers’ level of Education, Burundi, 1992-93 level Costs of Secondary Day School by Fathers' of Education, Burundi, 1992-93 level Costs of Secondary Boarding School by Mothers' of Education, Burundi, 1992-93 Costs of Secondary Day School by Mothers' Level of Education, Burundi, 1992-93 .................. Costs of Secondary Boarding School by Fathers’ Occupation, Burundi, 1992-93 .................... Costs of Secondary Boarding School by Mothers' Occupation Burundi, 1992-93 ..................... Costs of Secondary Day School by Fathers' Occupation, Burundi, 1992-93 ................................ Costs of Secondary Day School by Mothers' Occupation, Burundi, 1992-93 ................................. Costs of Secondary boarding School by Income, Burundi, 1992-93 ......................................... Costs of Secondary Day School Education by Income, Burundi, 1992-93 ................................ Costs of Secondary Boarding School by Assets, Burundi, 1992-93 ................................ Costs of Secondary Day School by Asset, Burundi, 1992-93 ......................................... Costs of Secondary Boarding School by Total Families' Expenditure, Burundi, 1992-93 ................... Costs of Secondary Day School by Total Family xvi 116 118 119 119 125 128 129 131 131 132 133 134 135 136 137 138 138 139 139 140 Table Table Table Table Table Table Table Table Table Table Table Table Table Table Table Table Table Table Table Table Table Table Table 43 44 45 46 47 48 49 50 50 51 51 52 53 54 55 56 57 58 59 60 61 62 63 Expenditure, Burundi, 1992-93 Direct Private Costs of Boarding School per Grade Level (GRADNOW), Burundi, 1992-93 School Year ......... Direct Private Cost of Day School per Grade Level, Burundi, 1992-93 School Year .................... Direct Private Cost per Cycle of Secondary School per Type of school, Burundi, 1992-93 ...... DPC of Secondary Boarding School and Number of Children in Secondary School per Family ...... DPC of Secondary Day School and Number of Children in Secondary School per Family ................... Total Cost of Secondary School per Family (TYCOST) Total Family Spending on Education (STUD). Burundi, 1992-93 Burden of the Direct Private Cost of Public Secondary Schooling per Child (DPC) (cont'd) Total Burden of Secondary School (TSECBURD, TSECEXP) and Total Burden of Education (BURDED, EDUCEXP) per Family (cont'd) ............................................ Correlations Among the Variables of Secondary Boarding School Correlations Among the Variables of Secondary Day School Correlation Matrix of Factors and Access Variables Summary Table Section v.3. T-test Mean-comparison Distribution of Dummy Variables Coding .......... Variables in the Equation of Burden ....... y ...... Equation for Access to Secondary Schooling (Entire Sample and family as Uhit of Analysis) Equation for Access to Secondary Schooling (Urban Sub-Sample and family as Unit of Analysis) Equation for Access to Secondary Schooling in Burundi (Rural Sub-Sample and Family as Unit of Analysis) Equation for Access to Secondary Schooling (Entire sample and Child as Uhit of Analysis) ...... Equation for Access to Secondary Schooling (Urban Sub-Sample and Child as Unit of Analysis) Equation for Access to Secondary Schooling (Rural Sub-Sample and Child as Unit of Analysis) xvii 141 142 143 143 146 146 153 156 160 161 163 164 167 168 178 182 184 184 187 189 191 195 197 199 LIST OF FIGURES Figure 2.1. Framework of Access to Secondary Schooling in Burundi ...................................... 47 Figure 5.1. Determinants of access to secondary schooling in Burundi, Entire Sample and Family as Unit of Analysis ..................................... 188 Figure 5.2. Determinants of Access to Secondary Schooling in Burundi, (Urban Subsample and Family as Unit of Analysis) .................................... 190 Figure 5.3. Determinants of access to secondary schooling in Burundi, (Rural Subsample and Family as Unit of Analysis) .................................... 192 Figure 5.4. Determinants of access to secondary schooling in Burundi, (Entire sample Child as Unit of Analysis) .................................... 196 Figure 5.5. Determinants of access to secondary schooling in Burundi, Urban Sub-sample and Child as Unit of Analysis ..................................... 198 Figure 5.6. Determinants of access to secondary schooling in Burundi, Rural Sub-sample and Child as Unit of Analysis ..................................... 200 xviii LIST 0? MAP Map 1 Administrative Divisions of Burundi .......... xix CHAPTER I. INTRODUCTION This chapter introduces problems of access to secondary education in Burundi. In particular, it addresses issues raised by the structure, evolution, and expansion of primary schools, and other factors affecting access to secondary schooling. It also presents background information on the economy and population of the country, its ethnic composition, and a brief description of both the formal and non-formal educational system. Finally, this chapter formulates the objectives and presents the organization of the dissertation. 1.1. Problem Statement Secondary school education in Burundi brings both joy and burden to the family and the community. On the one hand, education is seen as a way out of poverty. Children who pass the secondary school entrance examination (the gggggu;§_flgtiggal, which is administered in sixth grade, last grade of primary school) and go to secondary school, expect a better life upon completion of their studies. Parents are also relieved for many reasons when their children graduate from secondary school. First, they no longer bear the cost of education. Second, some parents expect to receive money from their children. Third, and in 2 particular, the children who graduate usually finance the secondary education of siblings. . Finally, these same children often motivate, morally and financially, other children in the neighborhood, and in the extended family, to seek higher levels of education. On the other hand, secondary school education is a burden to the parents because they must pay for it, whether it takes place in day school or in boarding school. In addition to the school fees, which are determined by the government, there are other costs, such as school supplies, uniforms, and shoes (the amount varying according to the parents’ ability to pay)‘. In some cases, parents are willing to send their children to secondary school but cannot afford its cost. They rely upon extended family and friends for help. Consequently, secondary school education involves the community. 1 Private costs of education, as opposed to institutional costs of schooling, include direct private costs of education, household contributions to school and indirect private costs of education (Tsang, 1994). Tsang (1994) defines direct private costs of education as expenditures by jparents on their children’s schoolingu IHousehold contributions to school are contributions, in cash or in kind, from families to the school, or to school personnel. Indirect private costs of education refer to the economic value of the forgone opportunities of schooling (e.g., child’ s labor in family production and in performing other domestic chores). Institutional costs.are public expenditures on education, which are supported by school fees and household contributions to school. 3 I. 2. Burundi: Physical Characteristics and Population Burundi is a small landlocked, overpopulated, and poor country in Eastern-Central Africa. It is bordered by Rwanda in the north, Tanzania in the south and in the east, and Zaire in the west. The total area comprises about 27,384 square kilometers (or about 12,000 square miles). In 1995, Burundi was divided into 15 provinces and 114 communes’, (See Map 1). Despite its very small size, equivalent to the State of Maryland in the United States of America, Burundi had a total population of six million people in 1990. This population was projected to reach 9.1 million people in 2010, and 13 million people in 2025 (Population Reference Bureau, 1994). This population constitutes a major challenge to the capacity of the educational system of the country to provide education to all. Physical, financial and human resources cannot increase fast enough to match the population growth rate and support its human development needs. Nearly 94% of the population of Burundi lives in the rural areas; the urban population represents six percent of the total population. Likewise, about 90 percent of the 2 The limits of the province of Mwaro, which was created in February 1996, were not known yet during the period of time this dissertation was being written. 4 population of Burundi are involved in subsistence agriculture. Consequently, as arable land becomes divided into smaller plots by inheritance, families with many children, but limited resources, may not be able to afford the cost of secondary education. Hence, the subsistence economy hinders access to secondary schooling. The population of Burundi is comprised of three ethnic groups: Bahutu (Hutu 85%), Batutsi (Tutsi 14%), and the Batwa (Twa 1%). The Hutu were traditionally associated with farming, whereas the Tutsi were involved in livestock activities; the Twa lived by gathering and hunting types of economy. All three ethnic groups speak Kirundi which is the native language. They live mingled in all regions of the country (Weinsten, 1974). These three ethnic groups have had unequal access to formal education during the post-independence era. The Tutsi population was dominant in the formal school system and harvested more of the fruits of Western education, in terms of occupational, economic, and political power. Although ethnic inequalities in schooling are hard to document, there are registered regional imbalances in access to primary and secondary education (Ndimira, 1995; Ndimurukundo, 1995). Table 1, below, indicates a very disproportionate access to formal schooling. These disproportion is highly skewed towards the provinces of Bururi, Muramvya, and Bujumbura city, which have the highest 5 Map 1. Administrative Divisions of Burundi I j l 293! I" “i TANZANIE ru- % 2:20 40km - V(‘ . ~34 Wan. ; .7; $3133} ‘ * "3°. ...-m . Chi-wan . _m*m LII“ mm o W-mam __- .. (h HML'WW'MM ' m1990. Source: Bidou et al., 1991, p. 10 6 Table 1 Gross Enrollment of Primary and Secondary Schools by Province, Burundi, 1991-92. Province Gross enrollment ratio 1991-92 Primary Secondary Bururi 98 28.0 Muramvya 91 8.0 Bujumbura (1) 78 17.0 Bujumbura (2) NA 4.9 Makamba 75 3.8 Gitega 75 5.0 Cankuzo 70 3.0 Ruyigi 63 3.0 Rutana 63 6.5 Bubanza 61 2.5 Ngozi 60 2.2 Kayanza 59 2.6 Karuzi 54 2.0 Cibitoke 53 2.0 Muyinga 49 2.0 Kirundo 46 2.5 Note: (1) urban (2) rural Source: Ndimira, 1995. 7 concentration of Tutsi pOpulation (Ndimurukundo, 1995; Ndimira, 1995). The dissertation will focus on three different areas of Burundi. The first is the capital city where most of the people are employed in the modern sector of the economy. The second is located in the center of the country, where Gitega, the second largest city of the country is located. The third is located in the province of Karuzi, which is mostly rural. 1.3. The Educational System I.3.1. Origins of Unequal Access to Education This study is an attempt to document disparities which have long seemed to characterize Burundian education. Modern education opened its door first to the wealthy families: royal families, wealthy Tutsi and Hutu. The Twa rarely participated in modern education. Families who owned land (land and cattle were the traditional symbols of wealth) were most likely to send children to school. Those who went through school were able to get out of this subsistence economy and into the wage market. Their offspring became educated, and capable of helping their children motivationally, as well as financially. The other traditionally poor families were, and are, still trying to 8 get at least one child admitted to, and graduated from, secondary school. As secondary schooling is expensive, its access is related to the economic system of the country. Historically, the Tutsi, who have been in power since independence, controlled the educational system and increased unequal access to secondary education by introducing a highly selective testing system. This led to unequal access to the benefits that formal education confers, such as high status jobs and political power (Weinsten, 1974). The following is a short description of the education system and its most pertinent features. 1.3.2. Formal Education. The formal education system is divided into three different levels: primary school, secondary school and the university. There is no public preschool in Burundi. All preschool are private and limited to urban areas only (Rwehera, 1994). 1.3.2.1. Primary School Primary school consists of six classes: grade one through grade six. The ministry of education provides books for all subjects for teachers and students. This ensures 9 that the schools provide the same instruction to all enrolled children in the country. The system provides a schedule that is supposed to be followed for each day of the week by grade level, from first to sixth. In the early eighties a number of measures were introduced with the objective of increasing the capacity of primary schools. These measures included, most importantly, the double shifts and automatic promotion. 1.3.2.1.1. Double Shifts In order to expand schooling to all children, with the same facilities available to all, the system of double shifts was introduced in 1982-83. Double shift is a system of using the same teacher and the same classroom for two groups of students during the same day. This practice can, therefore, double the capacity of the system. For a week, one group attends school in the morning, while the other attends in the afternoon. The groups switch the following week. This system provides a half day of school instead of the whole day students received before the introduction of double shifts. Each group was targeted to consist of a maximum of 50 students. Double shifts have greatly increased access to primary education, especially for girls. Prior to the introduction of the double shift in 1982, the primary gross enrollment 10 rate was 29%. It rose to 72.5% in 1991-92 (Rwehera, 1994). Female literacy increased from 10 percent to 40 percent, or a 300 percent increase, over the period 1970—1990. Male literacy increased from 29 percent to 61 percent during the same period, or a 110 percent increase (World Resources, 1994). However, it can be argued that the double shift system makes children from the rural areas less competitive in the Concours National. With double shifts, the children from poor families spend less time in school and on academic subjects. Once they arrive home, these children have to perform household chores instead of studyingJ. In contrast, the children in the city, and those whose families’ have house workers, can use that time to review their school lessons and read other academic materials. The more day time student spend at home, the shorter the time they are exposed to the precious books which they do not find in their own homes. This situation, it appears, may increase the chances of the advantaged children while reducing those of the disadvantaged ones. Hence, inequalities of opportunity may be perpetuated through the economic status of family and the system of double shifts. The advantages for children in 3 Domestic chores are performed mainly by children, especially girls, when parents are poor and/or located in the rural areas. As a result, the level of academic achievement of these children is expected to be lower than that of boys or children who do not have domestic chores. 11 this context are synonymous with the advantages of time and resources in their homes. 1.3.2.1.2. Automatic Promotion The system of double shift in Burundi was introduced simultaneously with the system of automatic promotion. In order to create room for the new generation of first graders, automatic promotion meant that every child enrolled in first grade would be admitted to the next grade at the end of the school year, even with a minimum of achievement. The system of automatic promotion has offered the seeming opportunity for children of the poor to complete primary school. However, automatic promotion did not empower many children to compete for entrance into secondary schools. The limited time to study at school due to double shift, and domestic chores after school, have prevented many children from having the quality time necessary to prepare for the secondary school entrance examination. 1.3.2.1.3. Concours National To pass from sixth grade to secondary one must pass the Co N ' a . The ancours Natignal is a highly selective national exam which is used to select students who go from primary school to secondary school. So far, less 12 than 10 percentfiof the sixth graders are allowed to pass this exam because there are very limited places in secondary schools in comparison to the number of children finishing primary school (Schwille et al., 1991). This exam was introduced in sixth grade in 1973. Overall, however, the number of children entering secondary schooling increased, specifically for females, up to 1990. There was an overall increase of 4% in secondary school enrollment. During the following school year 1990- 91), the increase was 6%. Thereafter, there was a decrease of 4% in secondary school enrollment during 1991-92. At the regional level, enrollment in secondary school followed the national enrollment pattern, except in the capital city of Bujumbura, where it has been consistently increasing. The number of children who passed the angourg National in the sampled areas and the entire country are summarized in Table 2 below. 1.3.2.1.4. Grade Repetition Primary grade repetition is one of the main consequences of limited access to secondary education in Burundi. Children repeat grades so that they can catch up with their competitive peers and be able to have a stronger foundation in later grades. Consequently, these able students are expected and encouraged to pass the Concours 13 «z «z «z «z «z 42 «z .42 .22 .«z «z «z s 2.8 42 «2 Sb «2 «z omm «z «2 sm «2 «z mmmH <2 «2 <2 sea «2 dz mm Em mm mm. as mm c. mmmm mmsm 8: «mm Has m3 omm m: «ms mm mm om Hmms «z «2 <2 m . m m «z «z 4.2 42 «z m A . m «z «z s mmmm Hmmm moom mmm mmm Hos mmm mos m3 mm mm mm ommH «z 42 «z m . m A 42 «z . a H «z «z m . o H «z .«z .1 mm: mmsm mmms «ms m3 Gm in s: mmH m3 3 m: mmms «z «z «z m . m H «z «z s . m «z «z m . m «z «z m memo 83 some mom mom mmm Sm mm m: 3H mm m: 83 ESE. m 2 .H. m z e m 2 .H. m 2 use.» newumz musnesnsm mmmumw musumx >Omm mm-mmms .Accsssm .soflmom pom xmm >n ammoflumz msdoooou mnu Ummmmm 033 cmupafinu mo HmQESZ m canoe 14 National the following year. As a result, unlike in many countries, repeaters achieve substantially higher test scores than non-repeaters (Schwille et al., 1991). Grade repetition is considered a positive sign for later success in school in Burundi. It is not everyone who is entitled to repeat grades. Children do not usually make the decision to repeat a grade; their parents or immediate relatives do. In this situation, parents who can perceive the benefits of repeating will make their offsprings repeat even in lower grades to boost their later achievement. In short, as less than 10% of children go to secondary school, the remaining 90% of the children enrolled in sixth grade undertake many strategies to maximize their chances to pass the concours National. These strategies include: * repeat sixth grade; * return to lower grades, especially the fifth grade, * change school and go to a lower grade Those who get discouraged drop out of school without repeating or repeat only once. As such, the sixth grade is the bottleneck to access to secondary school. 1.3.2.1.5. Para-Primary - Yaga Mukama Not all children attend formal schools, nor do all who enroll in those schools stay or finish them. Yaga Mukama schools, Catholic bible schools that end at the primary 15 level, provide an alternative. These informal elementary schools have been frequented by adults and children who would never have enrolled in formal school, or who were primary school dropouts. Yaga Mukama schools are free and meet only twice a week. They, therefore, allow students to carry out other activities, such as domestic chores and off-farm activities. Along with religious education and reading, Yaga Mukama schools offer some practical farming lessons. It was only recently that writing was introduced in the Yaga Mukama curriculum. During the 1992-93 school year, some school-age children were enrolled in Yaga Mukama rather than the formal primary schools of the Commune of Gitega and Buhiga. Yaga Mukama alone represented 6% of the school age children in the commune of Buhiga and nearly 18% for the commune of Gitaramuka (République du Burundi, 1992). In the Commune of Gitega, children attending Yaga Mukama were not enrolled in the regular primary school. In contrast, Yaga Mukama schools in the capital city have been adapted to the schedule of the children attending regular primary school (verbal communication with a Yaga Mukama teacher in Bujumbura, 1993). All the children attending Yaga Mukama in Bujumbura were enrolled in the regular primary school. They were attending primary school one shift and still participated in the Yaga Mukama twice a 16 week. If they were in the morning shift, they would attend Yaga Mukama school in the afternoon. The system of double shift has therefore facilitated children participating in Yaga Mukama schools in Bujumbura. 1.3.2.2. Secondary School In 1986-87 school year, 86% of secondary schools in Burundi were public and 14% private. Most of the private schools were in Bujumbura and 51% of private schools students were foreign; these schools served only 8% of all secondary students. By 1991-92, private schools comprised 9% of the total population of students. Public secondary schools totaled 91% of the students (Burundi, Ministére de l’Education Primaire et Secondaire, 1992). Secondary schooling is subdivided into two cycles: Cycles Q’Oriegtatigg, which provide nationwide standardized curriculum, and gyg1g_§upéri§ur whose curriculum varies according to the student’s fields of specialization. Cygle d’ Orientatign consists of grades seven to grade ten. It is considered the general basis of secondary school. After the gygle d’Orientatign, students are subject to a national test. The results of this test serve to track students into general secondary schools and technical secondary schools. The students who have high scores on this test attend general secondary schools. This cycle used 17 to be called Cycle Inférieur. ngle Supérieur is divided into two different schools: general secondary schools and technical secondary schools. General secondary schools consists of grade 11 through grade 13. It includes three sections: Scientific A, Scientific B, the Humanities. Technical secondary schools consist of grade 11 to grade 14. Teacher training Schools: some of the secondary schools go from 11 to 14, i.e., Lycées Pédagogigueg. At present, they consist of four years divided into two cycles of two years. The first two years are called Egglg de Formation deg Instituteurg (EFI). Graduates from these Egglg_gg W have to teach in primary school for at least two years before starting the last two years, which complete the cycle of Lycées Pedagogigues. After secondary school, there is another test which is meant to track students to the University of Burundi. All students take this exam during the last year of secondary school. Secondary schools were exclusively boarding schools until the 1982-83 school year, when day schools started. Boarding schools provide facilities for students to stay during the school year. However, students are required to purchase supplies such as extra bed-sheets, and blankets. Most of these supplies would not be needed if the students were attending day schools. 18 Day schools were established in 1982-83 to lessen the burden of boarding secondary schooling costs to both the government and parents. In day schools, students go to school every morning and go back home to eat and sleep after school each day. Initially, students who passed the national secondary school entrance examination, Concours National, were sent to nearby secondary schools. However, the objective of building day schools and having day students in all the communes of the country encountered many problems, such as lack of transportation, electricity, running water, and qualified teachers. The number of available day students in a given region was not large enough to justify the cost of these day schools. Consequently, day schools remain only in Bujumbura, the capital city. Moreover, as a result of this policy, day school students from poor households are highly disadvantaged by the lack of facilities such as electricity and water in their homes, which are available to day school children from middle and upper class families, and to students in boarding school. 1.4. Direct Private Costs of Secondary Education Direct private costs of secondary schooling are the expenditures by parents on a child’s secondary education. 19 To date, school fees are the only part of the total direct private cost of secondary schooling in Burundi that receive public attention. In boarding school, school fees have been increasing since the 1973-74. In day schools, school fees have increased since 1985. In addition, there has not been any government scholarship or assistance of any kind to students from poor families in Burundi. The direct private costs of secondary schooling continue to be a major concern of the government and the population of Burundi. The greater part of the direct private costs of secondary education is unknown to parents and to the government, because the government regulates and monitors only the school fees. The burden of the other direct private costs on the family, such as school supplies, transportation and boarding accessories, has long gone unreported in the literature on education in Burundi. Thus, the government and the general population do not have a clear idea of the total direct private cost of secondary schooling in Burundi. Until the child passes the Concours Natignal, some parents stay ignorant of the magnitude of those costs and find themselves unprepared for them. Annual school fees per child have been increasing for both boarding and day secondary schools. Overall, these fees increased by 800 percent for boarding secondary schools from 1973-74 to 1992-93 (or 20 years), that is, from 1,000 FBu to 9,000 FBu. They increased by 350 percent for day 20 schools from 1983-84 to 1992-93 (i.e. from 1,500 FBu to 4,500 FBu) school years. The amounts of annual school fees and corresponding percentage increases for both boarding and day secondary schools are summarized in Table 3 below. Table 3 Annual School Fees per Child and Percentage Increase for both Boarding and Day Secondary Schools, Burundi, 1973-74 - 1992-93 Schools Years. School year School Fees and % Increase over Previous Year Boarding % * Day % * 1973-74 - 1000 NA NA NA 1980-81 - 1982-83 2000 100 NA NA 1983-84 - 1984-85 3000 50 1000 NA 1985-86 - 1986-87 4500 50 1500 50 1987-88 - 1989-90 6000 33 3000 50 1990-91 - 1992-93 9000 50 4500 50 Note: * % increase over previous year. Source: - République Burundi: Ministere de l’Education Nationale, 1988 - République Burundi: Ministére de l’Education Nationale, 1993. - République Burundi: Ministers de l’Education Nationale, 1993. ‘ While school fees for secondary school increased, family income (GNP per capita) in Burundi decreased. For example, GNP per capita decreased by 22 percent from 1991 to 1993. It went from $218 in 1991 (World Resources Institute et a1. 1994) to $210 in 1992 and $170 in 1993 (Bernarek, 1994). The GNP per capita should be even lower in subsequent years because of the civil unrest that started in 21 October 1993 in Burundi. As a percentage of household income, school fees constituted a major burden for families. For instance, school fees for secondary schools represented 30% of the GNP per capita for boarding secondary school, and 10% for day school in 1991-92 school year. The results are summarized in Table 4 below. Table 4 Burden of School Fees with Respect to Family Income (%), Burundi, 1982-83 - 1991-93 School Years (Selected Years). School Year Boarding school Day School 1982/83 7% 3.5% 1983/84 10% 5.0% 1985/86 15% NA 1987/88 20% NA 1991/92 30% 10.0% Source: - Mayoya, 1989; — République Burundi: Ministers de l’Education Nationale, 1988; - République Burundi: Ministere de l’Education Nationale, 1991; - République Burundi: Ministere de l’Education Nationale, 1993. Hence, access to secondary school became increasingly difficult for poor families as they became poorer and poorer over time. Informal primary schools became an alternative to them. 22 1.5. Proposed Research This dissertation is a study of the direct private costs of secondary schooling and the impact of such costs on access to secondary schooling in Burundi. It estimates the magnitude of these costs and their economic burden on Burundian families. It seeks to find out whether such costs are a barrier to access to secondary schooling in Burundi. The specific research questions are: * How much are Burundian parents spending on the direct private costs of secondary schooling? * How do direct private costs of secondary education vary by family background, type of school (day or boarding), and sex of the student? * Which are the families who can afford to send children to secondary school and fully support them financially? * What is the magnitude of the economic burden of the direct private costs of secondary education on parents and how does this burden vary with family background? * Does the economic burden of direct private costs of secondary education affect access to secondary education? In addition, the following questions deal with access to secondary schooling: * Who are the peOple send children to secondary school? * To whom are direct private costs of secondary education 23 a barrier to school access? * Are families with children in secondary school different from those which do not have children in secondary school? * What are the effects of parents’ education, wealth and occupation on children’s access to secondary school? * Does grade repetition affect access to secondary education? This research will bridge the information gap existing between the already known government contributions and the still unknown contributions of families to secondary education. It informs policy about the portion of the direct private costs of secondary schooling which is not regulated by the government. It uncovers the magnitude of the direct private costs of education by family background. It points to the impact of grade repetition, the extended family, parents' education, parents’ occupation, family income and assets, and location to access to secondary schooling. 1.6. Organization of the Dissertation The rest of this dissertation has five chapters. Chapter Two presents the review of the literature and the conceptual framework. Chapter Three describes the methodology used, including a discussion of methods of data 24 collection and preparation for data analysis. Chapter Four presents the results of the analysis of the direct private costs of secondary schooling. Chapter Five presents the results of the analysis of access to secondary schooling. Chapter Six is the conclusion; it summarizes the major issues of the study, addresses policy implications, and indicates futurey research needs. CHAPTER II. LITERATURE REVIEW AND CONCEPTUAL FRAMEWORK This chapter reviews studies on private costs of education and on access to schooling. It also presents the conceptual framework of access to secondary schooling in Burundi. Finally, it provides the justification for this study. 11.1. Literature Review The literature review is subdivided in two parts: the first part deals with the private costs of education. The second part explores factors influencing access to secondary schooling. 11.1.1. Private Cost of Education Private costs of education are made of direct private cost and indirect private cost of education. Direct private costs are the monetary contributions by families to their children’s schooling. These include family expenditures on items such as tuition and school fees, textbooks, supplementary study guides, writing supplies, uniform, school bag, transportation, and boarding school costs (Tsang 1994). Indirect private cost are measured by the economic 25 26 value of the foregone opportunity of schooling. The foregone opportunity of the time the child would have contributed to the family production or other household chores, if s/he was not enrolled in school (Tsang, 1994). Until recently, relatively little empirical research has been done on the private costs of education in developing countries. Evidence of the impact of private costs of education on families was reported in research studies primarily in terms of access and equity (Lockheed, 1979; Waweru, 1982; Robinson et al., 1985; Wolff, 1985; Anderson 1988; and Mankha, 1990). Much of the literature referred to poverty, remoteness from schools, gender, family composition, birth order, race, ethnicity, religion, handicaps, needs for special education, and children in continual migration (Anderson, 1988). Previous studies tend to focus on public, and not private, educational expenditures. Most of the published studies on private costs of education in developing countries were conducted in the late 19803 and early 19903 by Tan (1985), Tsang (1988; 1990), and others. For example, Tan (1985) concentrated on the direct private costs of secondary schooling in Tanzania, and Tsang (1988) focused on the direct private costs of primary schooling in Pakistan, and the direct and indirect private costs of primary schooling in Thailand (Tsang and Kidchanapanish, 1992). 27 Empirical studies show that private costs constituted a significant burden on low-income households (Tan, 1985; Tsang, 1988a; Tsang, 1992). In Malaysia, direct costs of all levels of education accounted for 18 percent of family income for low income households, while it represented only six percent of income for wealthy households. In China, these costs represented 11 percent of households' income in the lower-secondary level and 19 percent in the upper- secondary level (Tsang, 1994). Thus, private costs can be burdensome on families, especially those from poor backgrounds. Empirical research also points out that private resources for education were very substantial compared to public resources for education (Tilak, 1985; Tsang, 1992). In Thailand for instance, total private resources amounted to 28 percent of institutional cost in government primary school in 1987 (Tsang, 1994). A 1988 World Bank study of education in two provinces of the People’s Republic of China found that the direct private cost equaled about 70-75 percent of the total institutional expenditure at the primary level and 50-70 percent of the public institutional expenditure for secondary general education (Tsang,1994). It is erroneous to think that tuition is the only private expenditure for education. In Thailand for example, there was an extensive list of non-tuition costs, including students’ uniforms, school bags, textbooks, writing supplies 28 (pencils, rulers, notebooks, erasers, color pencils and pens), transportation, school fees (for lunch program and other school activities), shoes, and sports wear (Tsang & Kidchanapanish, 1992) Moreover, empirical studies found significant disparities in private costs in relation to family income and wealth. For example, even though poorer families in Thailand spent much less on education, private resources necessary to finance education accounted for a much higher proportion of household income of these families than richer families (Tsang & Kidchanapanish, 1992). In general, families with higher cash income, accompanied by other forms of wealth, and more educated parents, allocated more private resources to schooling. Many studies suggested that costs of schooling limit enrollment for specific groups, such as the rural poor and females. In India, poor families attributed their failure to send children to school to high direct private costs (Tilak, 1985; Tsang, 1994). Solutions to reduce the burden of cost of schooling to parents included free education and/or subsidies for specific groups, especially during the period of economic downturn (Noor 1981; Tilak and Varghesa, 1985; Kelly, 1986; Anderson, 1988; Tsang, 1994). Finally, gender was another important factor in the direct private costs of education in developing countries. For instance, female students in Tanzania were found to 29 require significantly higher expenditures for schooling than male students for both government schools and private schools (Tan, 1985). In Thailand, the costs of education was higher for girls than for boys in government schools and the costs were the inverse in private schools. Furthermore, girls had higher costs in urban areas but lower costs in rural areas (Tsang, 1994). Thus, gender disparities could be compounded by disparities due to poverty and remoteness (Smith & Cheung, 1981; Stafilios-Rothschild, 1982; Kelly, 1986; Robinson et al., 1986; Tsang, 1994; and Okwach & Wamahiu, 1995). Consequently, in the study of private costs of education, special emphasis should be placed on gender issues since girls have been under-represented in many developing countries in secondary education. In two districts studied in Kenya, for example, girls made of 41% of the students in forms 1-4 in secondary school, but they represented 61% of dropouts as a results of financial constraints (Okwach & Wamahiu, 1995). Another study by Zamberia (1996) found that although girls' enrollment increased by 517% from 1971 to 1990 (or a 26% annual increase), as opposed to 260% increase for boys, the proportion of girls to boys enrollment increased only by 40% (or 2% annual increase) in the same period of time (Zamberia, 1996). In Burundi, girls constituted 36% of all students enrolled in secondary school (République du Burundi, 1993). 30 II.1.2. Access to Schooling Research has indicated that there are regional differences in access to schooling ( Heyneman, 1978; Niles, 1981; Anderson, 1988; Lockheed, Fuller & Nyirongo, 1988; Ndimurukundo, 1995; Ndimira, 1995). It was also found that family socio-economic status could be an exceedingly important, sometimes the single most important, factor in determining access to school. However, since there are many differences between countries, each country has to be put in its special context. Therefore, country-specific research is needed. 11.1.2.1. Rural Versus Urban Settings Many countries of the developing world are characterized by unequal regional economic development and regional disparities in educational participation. Regions which are remote from urban areas lack information and infrastructure and experience more poverty. Inequitable resource distribution across regions constitutes, therefore, a limiting factor to poor families in accessing schooling (Fuller, 1985). Poverty, illiteracy of the parents and the remoteness of the region from educational infrastructure are among the barriers to school participation. In addition, social class factors, such as parents’ occupation, have an 31 negative impact on school access, especially in the rural areas (Lockheed, Fuller, & Nyirongo, 1988; Harbison & Hanushek, 1992). In Thailand and Pakistan, private financing of education was inequitable and contributed significantly to inequalities in educational opportunities for students from different family backgrounds or regions (Tsang & Kidchanapanish, 1992; Tsang et al., 1990). Urban, higher income, wealthier, more educated, and professional or managerial households spent more than rural, lower income, less wealthy, less educated and agricultural households (Tsang, 1994). Income was a key factor that affects access to schooling in Kenya, where 64% of students in rural areas dropped out of school because of a lack of funding by their parents (Okwach et al., 1995). School participation in rural Uganda was 10%, whereas it was 90% in the capital city (Heyneman, 1978). In Malawi, many children who attended schools located in the rural areas had parents in the skilled occupations and had electricity and running water. Some poor families in rural areas never enrolled their children in school. (Lockheed, Fuller & Nyirongo, 1988). Likewise, in British Guyana, children of white collar workers had more access to secondary schooling than children of farmers or blue collar workers (Bacchus, 1966). In rural Brazil, efforts were made since early 19703 to increase 32 access to schooling by making primary schooling free and compulsory between the ages of seven and fourteen (Harbison & Hanushek, 1992). Moreover, the rural areas of the country had limited access to secondary schooling. Enrollment in secondary schools in the rural northeast of brazil for example was 3.3 percent in 1982 while enrollment in the urban northeast was 15.2 percent (Harbison, & Hanushek, 1992). In Burundi, in addition to the above factors affecting access to schooling, ethnicity constitutes another barrier. In areas of high concentrations of Tutsi, there was higher gross school enrollment in primary school and higher access to secondary school (Ndimurukundo, 1995; Ndimira, 1995)1. Earlier studies on access to secondary schooling in Burundi indicated that children in remote areas had limited access to secondary school (République du Burundi, 1986) 11.1.2.2. Girls in Rural Areas Many studies done in Africa, Asia, the Middle East, and Latin America show that girls are more likely than boys to drop out of school before completing their primary cycle, especially in rural areas. Illiteracy is experienced mostly among poor and socially disadvantaged women. While adult 1This issue of ethnicity is not developed in this dissertation because of its political volatility. 33 women are responsible for child care and cooking, their daughters bear a parallel burden by helping the family fetch water and look for fire wood (Stromquist, 1990). In rural areas, child labor affects mostly girls, which makes their school completion more problematic. The demand for child labor in the rural areas could lead parents to withdraw their children from school (Psacharopoulos and Woodhall, 1985; Tsang & Kidchanapanish, 1992). This explains why fewer rural girls from disadvantaged families go on to secondary schools. Remoteness from school affects more girls than boys. This factor is never reported in national statistics which may not reveal inequalities of opportunity based upon gender in rural and urban areas. In Thailand, for example, studies at the national level failed to show sex differences in achievement for the total sample. The disparity between rural and urban sex differences showed that in urban schools girls outperform boys by 1.2 points, while in the rural schools, boys outperformed girls by 1.0. The better performance of urban girls covered up the lower performance of rural girls (Lockheed, Fuller & Nyirongo, 1988). 11.1.2.3. Family Background Studies in developed countries have confirmed the influence of family background on children’s educational 34 attainment (Coleman, 1966; Jenks 1972; Lockheed et al., 1988; Lupton, 1983; Behrman, 1989). In France, the elitist educational system has been recognized as central in perpetuating the interests of the dominant groups (Millot, 1981). In developing countries, the few studies conducted in the 1970’s, such as in Uganda (Heyneman, 1978; Currie, 1978) concluded that family background has little to do with academic achievement in these countries. However, the sample used in these studies was highly homogeneous, as the target population was pre-selected. The results of Heyneman’s study were based on a limited variability in student achievement. In countries with highly competitive systems of education, as was Uganda during the time of the study, children in the seventh grade are a handful of gifted children that have survived every kind of selection test, and the high primary school dropout rate. These very intelligent children made it difficult to find a discriminating examination; they all pass or fail it together. Thus, in the district where only the top 10% of the pupils were in the seventh grade, Heyneman (1978) could not reasonable have expected that there would be a difference in the socio-economic background among these students. Much remains unexamined about those 90% who did not make it to the seventh grade. Who were they? Thus, the 10% of the children in Kampala came from very advantaged 35 families, educated officials, civil servants, and some prosperous business people. This explains why Heyneman did not find differences within the schools in his sample (Heyneman, 1978). Studies in the early 1980’s, however, support the consensus of those conducted in developed countries. They contend that family background plays a major role in the child’s educational access and retention through school. Studies in several developing countries have now illustrated the crucial, unequivocal, role of the family background. Findings show that impoverished children do not succeed as well as those from wealthy families; many children from poor families do not even enroll in primary school, many of those who enroll do not survive to finish (Cooksey, 1981; Niles, 1981; Fuller, 1985; Lockheed et al., 1988; Anderson, 1988; Mankha, 1990; Nzamutuma, 1992). In Botswana, traditional rulers’ children enjoyed all the education the colonial powers offered, whereas those of the common people remained herders and laborers as a tribute to the same rulers (Mankha, 1986). In Burundi, children of traditional rulers were the first to access secondary education, while children of poor families had little access to the educational system. Most rural schools are located in areas where the children share similar socio-economic background. Children from poor families are significantly less involved in 36 education in Pakistan and other developing countries (Niles, 1981; Tsang, 1991). Many of the studies illustrate the importance of parents’ education, occupation, income and status in their society (Niles, 1981; Cooksey 1981; Fuller, 1985; Lockheed, Fuller & Nyirongo, 1988; Mankha, 1990). Empirical studies indicated that children from lower socio- economic backgrounds are confronted with many educational inequalities, and face stronger pressure to stay out of school (Tsang, 1994). 11.1.2.4. Dropout Dropout rates differ substantially by region. In Uganda, for example, the drop out rate was 10% in the capital city and 90% in Karamoja, a region far away from the capital city (Heyneman, 1978). In Burundi, many children enrolled in schools in poor neighborhoods lack basic necessities to be successful; school buildings in poor rural neighborhoods lack the benefits of a better learning environment. In many Sub-Saharan countries, if there were three first grade classrooms, there would be two second grade classrooms and only one third grade classroom. At the end of each school year a sizeable number of children were sent home for two main reasons: (a) they did not perform competitively at school, (b) there was not enough room to accommodate everybody in the next grade since there were 37 usually fewer classrooms in the upper grades of elementary school (République du Burundi, 1987). In Burundi in 1987, there were 1,321 first grade classrooms and only 941 sixth grade classrooms (Republique du Burundi, 1987). During the same school year, there were 114,125 pupils in first grade and only 45,037 pupils enrolled in sixth grade in Burundi. About 61% of those enrolled in first grade did not reach the sixth grade. Children who reach the seventh grade are most probably from high socio-economic family background since the poorest children drop out for different reasons during the school year. Some children drop out even during the first term of the school year, a period which corresponds to the collection of school fees (République du Burundi, 1987). Therefore, students in grade seven are not representative of the population. 11.1.2.5. Extended Family Although recent studies have started to address the question of private costs of education in developing countries, they have failed to address the role of the extended family in financing the education of the children from poor backgroudnd. As a Burundian saying puts it: "Umwana si uwumwe; umwana ni uwo umuryango" (a child is not for the nuclear family only; a child belongs to the extended 38 family). Educated and wealthy relatives financially support children in the extended family in Burundi. Very few studies about socio-economic background have investigated the role of the extended family (Niles, 1981; Lanzas & Kingston, 1981 ; Nzamutuma, 1992). Niles’ study was limited to the impact of grand-parents status in Sri Lanka. Lanzas & Kingston (1981) emphasized the educational environment of children who moved from their nuclear families to homes of the extended family members, because schools were distant from their primary family home. Consequently, Lanzas & Kingston (1981) seems to suggest that in the case of Zaire, the extended family, more than the family status, played an important role in the education of children Yet in Zaire, sometimes out of economic necessity and sometimes by custom, many parents have little to do with their children’s education .Within the various webs of extended family ties, a number of relatives besides the parents often have the greatest socio- economic impact on a student’s education... The word "father" or "mother" in the Zairian languages refers to any person of a certain age and sex to whom respect is owed." (Lanzas & Kingston, 1981). Likewise, some children born in areas remote from schools, in Burundi, left their parents and went to live with relatives, mainly in the urban areas, so that they could increase their chances to access secondary schooling. The involvement of the extended family also expanded the chilsdren’s school participation in Rwanda (Nzamutuma, 1992). 39 There are no known institutions which provide students or parents with scholarships or monetary loans for attending secondary schools in Burundi. Having educated and wealthy relatives may be an advantage, especially for children from, rural families. These relatives not only can serve as financial providers but also as role models for this marginalized population as Windham (1991) calls it. As for the poor students without educated relatives, they are the ones who most need the benefit of policies designed to rescue them from marginalization in education. Thus, "Poverty is a source of multiple disadvantages. The children of the poor are more likely to suffer from nutritional and health problems, to grow up in environments that fail to support intellectual stimulation, and to have inferior school resources" (Windham, 1991). It is, therefore, important to consider the role of the extended family in accessing schooling in many parts of the world, especially in Africa. Relatives' involvement in financing the education of children can be an important factor in access to schooling. Children who have an educated extended family may have better access to schooling than those who have an uneducated extended family. As such, the extended family may play a major role in shaping the perception of the world by the child, and may impact the way the student perceives formal education. Hence, the neglect, exclusion, or ignorance of the role of the extended family 40 may have lead researchers to misleading conclusions about access to schooling. 11.1.2.6. Grade Repetition The meaning of grade repetition in developing countries such as Burundi, is different from that in developed counties because the educational systems are different. In the United States, for example, children living in families with incomes below the poverty line are nearly twice as likely to be retained in a grade as children in non-poverty stricken families (Bianchi, 1984; Natriello, McDill, & Pallas, 1990). In addition, low income children in the US were twice as likely to drop out of school (Natriello, McDill, & Pallas, 1990). In countries where education is not compulsory, and access to upper levels of education not granted at the end of the academic year, the situation may be different. Empirical research showed that Burundi had one of the world’s most selective secondary school systems, and that grade repetition was high because of the difficulty of obtaining access to secondary school (Schwille et al., 1991). Studies on grade repetition in Burundi indicated that repeaters achieved substantially more than non repeaters. Children repeated grades so that they could catch up with their competitive peers and be able to have a - - - — - — ..—-— — . .— ..—..——_-—-— ~v~—-——e“ - ”a“ v‘d-«n - a. 41 stronger foundation in later grades (Schwille et al., 1991). Repetition, in Burundi, is an advantage, a positive reinforcer, another chance to succeed in the entrance examination for secondary school. Repetition gives the repeater an edge over the younger cohort by providing the student the opportunity to be seen as the ultimate candidate to pass, which confers upon the teacher the strength to teach more, so that someone in his/her class will pass and confer upon him/her recognition as a good teacher. This is another motivation for the repeater because s/he gets more attention from the teacher. II.2. Conceptual Framework This study draws from human capital theory and the status attainment models. These theories have been applied to both developed and developing countries and have reached similar conclusions about the significance of returns from investment in education. Human capital theory and status attainment models are applied to the situation of Burundi in order to explore the conditions of access to secondary schooling. 42 11.2.1. Human Capital Theory Human capital refers to the skills that one acquires inside or outside the school that enhances one productive capacity. According to human capital theory, education can increase an individual's productivity and earnings (Schultz, 1971; Becker, 1964). Expenditure on education is a form of investment, since education has economic returns. Individuals and families make decisions on investment in education by considering both the costs and returns of additional schooling. If the private rate of return from schooling is higher than those of alternative activities, an individual will invest in more schooling. Similarly, parents will invest less in the education of the children if the costs of education go up and make the rate of return from education less attractive compared to alternatives. Likewise, a government should consider investing more in schooling if the rate of return of schooling to society is higher than that of alternative social investment activities. Empirical studies have found that the returns to education are quite high (Psacharopoulos, 1994). Education has an intergenerational value. When today’s students reach adulthood, their children will gain by virtue of the informal education received at home. Better educated parents are more likely to raise children who recognize the value of education in terms of job opportunities and — —| . _ -—r—‘_~—v M's-1" f“; 'm-L .‘a - 43 cultural opportunities. Consequently, the social value of educated parents, especially women, is not zero even if they never enter the wage labor force to utilize the skills developed in school (Weisbrod, 1971). 11.2.2. Status Attainment Prior to 1980 in developing countries, school characteristics were considered by some researchers to be the most important determinants of student achievement and status attainment (Heyneman & Loxley, 1983). After 1980 family background was found to be one of the major factors influencing children’s achievement and educational attainment (Niles, 1981; Fuller, 1986; Lockheed, Fuller & Nyirongo, 1988). Family background and the ability of children to adjust to and incorporate psychological factors into the learning environment were among the major contributing factors to status attainment and access to schooling. Social class origin of children was also found to be a powerful factor explaining school achievement (Carnoy and Levin 1985; Jenks et al., 1972; Kohn, 1959; Clement, 1975). Family background determines in large part the probability that children will enroll in, attend, repeat classes, and complete various levels of education (Anderson, 1988; Cheng, 1986; Lockheed, Fuller & Nyirongo, 1988). Children start from their family socio-economic status 44 and achieve their own socio-economic status, in part, via schooling. Their status, however, will also depend upon structural factors like historical period, regional environment, psychological factors (expectations, motivation and aspiration) and ability to learn (Fujita, 1978). Several factors stemming from family background, influence educational and occupational attainment of children. A child’s status attainment depends in large part on his/her parent’s occupation and/or education (Sewell & Hauser, 1975; Carnoy and Levin, 1985). Therefore, socio-economic inequalities lead to inequality of educational opportunity (Clement, 1975; Shea, 1976). 11.2.3. Access to Secondary Schooling in Burundi Apart from the study of unequal distribution of schools across regions (Ndimira, 1995), little empirical research has been conducted to document unequal access-to secondary schooling in Burundi. This study explores characteristics of families whose children had, or have, access to secondary schooling in Burundi. Consequently, family background, which determines resources available to students, may be one of the most important variables in determining access to secondary schooling. Other variables consist of cultural factors related to gender and the obligation of children and specifically female to perform domestic chores. They also 45 include the direct private cost of secondary schooling and education policies (e.g. grade repetition, national exams, collective promotion, double shift, school fees, etc.). Family background variables consist of fathers’ and mother’s education and occupation, family income, assets and family expenditure, and extended family. Extended family is included because relatives (wealthy or educated) often facilitate schooling of children in the extended family. Support by relatives includes, but is not limited to, money, room and board, and transportation. Children in Burundi are traditionally involved in domestic chores, such as fetching water, cooking, farming, and babysitting younger siblings. Nonetheless, while many urban families and educated parents hire workers to perform these chores, children from rural families must usually perform domestic chores before sunset because there is no electricity and they do not hire workers. Consequently, domestic chores conflict with homework and study time of children in rural areas. As a result, domestic chores stemming from lack of facilities and financial resources can create unequal access to secondary school between children from families who can hire workers and/or have electricity and families located in areas which do not have these benefits. The division of labor between men and women in Burundi, is also a relevant factor, in that women are in charge of 46 domestic chores, while men are involved in activities that are seasonal and require muscular strength such as clearing land, drainage, fencing, and building houses. Domestic chores, however, are done on a daily basis and are performed for extended hours. Girls are traditionally required to help their mothers in performing domestic chores in Burundi. Hence, since girls spend more time on domestic chores than boys, gender may be a factor that affects unequal access to secondary schooling in Burundi. Furthermore, parents pay for education of their children in Burundi. In many instances, the burden of the cost of secondary schooling on the family income may be prohibitive for poor families. Those who can not afford the cost of secondary schooling, and whose extended family is unable to financially support these children may not have access to secondary schooling. Thus, the burden of the cost of secondary schooling can be a limiting factor in accessing secondary schooling in Burundi. Finally, education policies in Burundi may hinder access to secondary schooling. Since the cost of education limits access to secondary schooling, one can expect many cases of dropout. In addition, given the value Burundian families put on education and the selective system of education (i.e., Concours National), grade repetition may be a key variable in accessing secondary schooling. The potential factors influencing access to secondary schooling .1 m m3 R llllllllllll 1T llllll Lu will r m E lllllllllllll j iiiii ill n S u 1% S E III E w G W O iiiiiiii . n .Il lllllllll .. lllll II. m H fl .1 C. w s (IIIHHHWII II] II lllll J .m .m fins 1%]!1 |_I < < < m .m R m E i lllllll r lllllllllllll o C lllll 1 llllllll . h m o m 1|; 4 i m m mm mm m .11. 2 Y & N S r llllllllllllll III mm IIIIII I e r 3mm um. IIIIII J mxs I. n y r .m I I T I < < S A T M m m RMA mm -III. smmm m we 7 .1 u m U m (”In E l..|..li u 0 N 4 F S m m m m lllllll .ill a M S > > E m m .m m o ”HHHHHHI mmm IIIIIIIIIIIIIII I: m l mm d u m lllll D P C G T m e m m lllllll IIIIL L fi 2 C B S > > w E .1 A . N E (I L lllllllllll r m C r Y W O C C (lil llllllllllll 1 lllll II C a f L I I m S trawl: . . m n m .. m D A. i m llllllllllllllllllll u .m F m U “No A m S o E D C T E villi llllllllllllllllll > > > e W [Rama-WWW lllll I. IIIIIIIIIIIIIIIIIIIIIIIIIIIII L I— r m lllllllllll r lllllllllllllllllllllllllllll a r .4 Iii] F S M i . m s m m N m 1 1m T o.-- u m mums um r F V F m C llllllllllllllllllllllllll u e I B m. IIIIII w m n .1 E P .1 F iiiiiiiiiiiiii [ill 48 11.3. Merit of the Study Previous empirical studies of private costs do not link costs directly to achievement or access to schooling. Studies on both private costs of schooling and family background often do not assess the impact of relatives, or the extended family members’ education, on the students’ achievement. Only the study in Rwanda (Nzamutuma, 1992) included the education of extended family members, and found that it influenced student achievement positively. The study in Zaire (Lanza, 1981) did show the link between the extended family and student achievement, but did not clearly show who among the extended family was included. Therefore, this study not only examines the influence of the extended family on access to secondary schooling, but also influence of repetition and the burden of private costs on schol access. This study includes marginalized people who do not have children in secondary schools, as well as those who do not stay to finish primary schooling. This study also explicitly investigates which students repeat grades in primary school, which do not, and how this impacts access to secondary schooling. Lastly, contrasting rural and urban students’ access to secondary schooling will shed some light on the causes of the limited numbers of rural girls, as well as which rural girls participate. 49 A review of previous studies on direct private costs and access to secondary schooling indicated that only households with children in secondary school were included in the analysis. In addition, most of the previous studies collected information on students only, and mainly in the school setting. Previous studies focused also on either rural or urban areas. Likewise, little analysis was done on the impact of grade repetition on access to secondary school. Unlike other studies, this study includes both households with children in secondary school and those without children in secondary school. Data were collected from both urban and rural families, and across regions. Detailed information on gender differences was collected for those from rural areas and those from urban areas. The information on analysis of grade repetition is more complete, as it includes age, the age of starting school, frequency of repetition, and cases of success resulting from grade repetition. In addition, data were collected on the number of repetitions and the highest grade completed by the children who dropped out of school. This study also includes children who never enrolled in formal school. The role of extended family in accessing secondary school is documented. The information on educated relatives includes the relationship with the family, the place of residence, and their level of education. This 50 study goes beyond the unidentified relative or guardian. CHAPTER I I I . METHODOLOGY This chapter presents the research design which consists of a survey in a multisite case study. It discusses the pilot study, methods for sampling, data collection, limitations of the dissertation, and preparation for data analysis. III.1. Pilot Study 111.1.1. Choice of the Site and Families A pilot study was conducted in the commune Isale in the province of Rural Bujumbura for three reasons. First, the pilot study served to check the adequacy of the questionnaire in terms of understandability and completeness. The ultimate goal of the pilot study was to make necessary modifications of the questionnaire to be used during intensive interviews in three selected areas. Second, the pilot study consisted of finding efficient procedures to use in data collection, given the unstable political environment of the country at that time. Third, the commune of Isale was appropriate because it is not located in the area selected for the study. The commune of Isale is located in the high altitude area which includes mountains with steep slopes. As a 51 52 result, for reasons of accessibility and safety, it was necessary to find informants who were knowledgeable about the region. The choice of informants was facilitated by primary school teachers of the central primary school in Isale. The informants helped in identifying families with secondary school children in the commune. The collines‘, in which the interviews were conducted, were selected on the basis of the information provided by the informants concerning the accessibility of the collines and safety. The informants were a mix of headpersonsziof each colline and young adults living in that colline. A census of families with children in secondary school was made for ten collines that were safe to visit. With the help of the informants from the colline, families with children in secondary school were counted. Thereafter, a list of collines was drawn up giving the number of such families on each colline. A random sub-sample of four collines was drawn from the ten collines. Interviews, using the pilot questionnaire, were conducted on the four collines. 1 A colline is the smallest administrative subdivision of a commune. 2 A headperson of a colline is a male person who is democratically elected by the inhabitants of that colline. He is a spokesperson at different levels of the commune. He represents the interests of the colline. He is engaged in all the aspects of life of the colline, such as the political, social, cultural, legal and economic activities. — . . .. ..-. .__ ”—B-‘H- .._.-.—~,-~ "v” < ' ' "I 53 A pilot study of 30 families was originally intended, including 20 families with children in secondary school and 10 families without children in secondary school. After 22 interviews (12 families with children in secondary school and 10 families without children in secondary school), there was a lack of the following important groups for the study: (1) day students in secondary school; (2) families which pay the full cost of secondary school for their children; (3) families which support other students who are not their children in secondary school; and (4) non-farmers. As a result, additional pilot study interviews in Bujumbura, the capital city, were necessary to include all these categories. A purposive sample of eight families was made to include the groups mentioned above. The interviews took place on one avenue in Nyakabiga, one of the urban neighborhoods primarily composed of middle class families. III.1.2. Results of the Pilot Study III.1.2.1 Questionnaire The pilot study helped refine the way the questionnaire was to be used. The interviews conducted for the pilot study provided an opportunity to determine which questions each respondent needed to answer. Thus, respondents did not need to be asked questions which did not apply to their 54 children’s situation. As an illustration, a person who did not have children in secondary school was not questioned about the cost of secondary schooling. No major changes were made in the questionnaire as a result of the pilot study because it was considered satisfactory for interviews. However, a few modifications were made especially for questions about child preference (see Appendix 1). The interviews indicated that many families had no preference for a specific gender. Thereby, an answer identifying no preference for either gender was added. The place of birth for both the respondent and the spouse was removed because the respondent whose spouse was absent did not know the place of birth of that spouse. Likewise, the interviews showed that it was difficult to know the level of education of people in the army. III.1.2.2. Sampling Procedure The original design of the study assumed that there existed a register of all the families which had children in secondary school at the commune level. This used to be the case. The pilot study indicated that it was not the case in the Isale commune. There was no longer any trace of the information that used to be collected. Even the old registers were no longer retrievable. The parish church did not keep the register either. Hence reliance on the 55 informants at the colline level became an efficient and necessary way to get accurate information on families which had secondary school children. Due to the political environment, which had created a climate of suspicion, the names of families with children in secondary school could not be given. Once families with children in secondary school to be sampled on each colline were identified, a systematic sampling was carried out. All the families of the colline were assigned a number from the nearest family to the farthest from the place we were located. The nearest family with children in secondary school was assigned number one. Each family was represented by a number. That number was written on a piece of paper. All these pieces of paper were folded, put together and mixed. Then, the children who were with us or other passers-by drew one piece of paper until the required number of families with children in secondary school to be interviewed was reached. Thereafter, the informant led me to the family for the interviews. As a result, the role of the informant was necessary throughout data collection. Living near the families with children in secondary school were families without children in secondary school. These families without children in secondary school were targeted in order to investigate the characteristics which differentiated them from those who had children in secondary school. By living in the proximity of families with 56 children in secondary school, families without children in secondary school had incentives to send children to secondary school. However, a colline is so populated and small that there should be no difference between families which lived near those with secondary school and those which lived a little farther from them. Therefore, there is no known bias built in targeting families without children in secondary school living near those with children in secondary school. Once the interview with the family with children in secondary school was finished, the interviewee helped me establish a list of families with children in the age group of secondary school, but who were not themselves enrolled in school. These families were assigned a number following a clockwise direction. These numbers were written on small pieces of paper. The papers were mixed and one piece was picked by one of the people or the interviewee. Then, the randomly selected family without children in secondary school was interviewed. III.1.2.3. Number of Interviews per Day It was originally planned that four to five families would be interviewed per day. However, the pilot study showed that only two to three interviews per day were possible because of the following: 57 * walking distance between families which were randomly selected for the interviews; * climatic conditions (i.e. rainy season) which made travel difficult; * certain types of questions made the interview last longer (i.e., questions related to income, family expenses, family assets, and the level of education of educated relatives); * waiting time for the respondent to be ready for the interview since each one was involved in daily activities like farming, domestic chores, and business; * non-availability of the respondent as a result of party meetings, farming far from home, social gatherings etc. III.2. Sampling for the Main Study 111.2.1. Selection of Provinces Before randomly selecting any provinces, the following provinces on the periphery of the country were excluded from the sample because of violence and concern for safety: Bubanza, Cankuzo, Cibitoke, Kayanza, Kirundo, Makamba, Muyinga, and Ngozi (see map 1). Since 1991, there has been turmoil in all these provinces. These provinces all border Rwanda, Zaire or Tanzania. Rwanda was at war since 1990. These peripheral provinces were believed to be affected by 58 PALIPEHUTU (Party for the Liberation of the Hutu People) attacks from Tanzania and Rwanda. Consequently, it was dangerous to go there. As a result only seven provinces constituted the final sampling space. The criteria used to stratify the seven provinces were: * population density; * number of primary schools; and * number of secondary schools. These criteria were thought to be influencing secondary school participation. On the basis of these criteria, three strata of provinces were formed. The first stratum represented provinces with a high density population (150 or more inhabitants per square kilometer) and a high concentration of both primary schools (100 schools or more) and secondary schools (five schools or more). The provinces which met these criteria were Muramvya, Gitega, and Bururi (Map 1). Within this first stratum, the province of Gitega was randomly selected. The second stratum included provinces with low population density (less than 150 inhabitants per square kilometer) and a low concentration of both primary schools (less than 100 schools) and secondary schools (less than five). Thus, Karuzi was randomly selected from four provinces in the low density stratum. The strata are summarized in Table 5 below. 59 'Ifsalale 5 Population Density, Number of Primary and Secondary Schools by Province, Burundi, School Year 1986-87. E?z:c:vince population density # primary schools # secondary number of people / schools square km PIj_§;h Density (First) stratum In Terms of Schools batizramvya 287 121 8 Gitega 286 111 15 E3Lizruri 160 160 8 IS<>vv Density (second) stratum In Terms of Schools Rural Buj umbura 3 0 0 82 4 Karuzi 207 65 2 Ruyigi 109 76 3 Fititana 102 54 0 53<3urce: - Annuaire Statistiques, 1989 - Statistiques Scolaires, 1986-1987 - Census of population, 1990 The capital city, Bujumbura, was a stratum by itself. IBaijumbura had the highest density of population (2,700 ixnhabitants per square kilometer). It also included the flighest number of both primary and secondary schools. As a (:apital city, it comprised the highest number of educated Ipeople and almost all the socio-economic categories of the population, including those underrepresented in other sites <3f interest to the study. 60 III.2.2. Selection of Communes There was a purposive selection of communes in the provinces of Gitega and Karuzi. A total of two communes vvesxe selected. There was one commune per province. The easelected commune included secondary schools and an urban (Beenter because the results of the pilot study indicated that Eiri entirely rural commune without secondary schools would Ilcbt have enough variety to answer questions related to the jLsssues of day/boarding schools, occupation and income mq oomuo cam mm< mo COwumHSQMDmmOHU "4H mHQmB 96 mN N N H HVH N H H H HN mm H. H H H H 0N HV m H ON mm m H MN H N H mH mm H. vN m H N mH NN N H HVH N N H hH OH N m N mH HH 0 m H H mH OH m w H «H v N H H MH H H NH o .. HH N H H OH H H m N N m w v H H. m m m MH MH m m w v OH OH m m m N OH OH H coon 44809 H0>0G >H§ «H MH NH HH 0H m m H. w m e m N H Hoonom ou soon Ho>os souoHHno pom Hoosom mCH>mmH mucuon owansHm spasm 0mm mmummmH 3H Hoosom uumq Hommqu pom 0:3 :oHoHHHU MON pouonEou oomuo umoanm "mH mHnme 97 .oHo mums» we was memo 0:0 HHCOHmv nmsosnu he moms you Hoonom DmoHlonz ConoHaso mo mono 02 "H4. ”oumz mmm HON mH m mH OH OH m HH m m bNN mm mH mH m mHfl BOB iH+H Nv N u H H Hfi m H H H 0v n H . mm u H H mm H H hm H mm mm mm mm OO‘O’thQ‘MHWN H H Nm V' H I H H H H m <' <' m m N H H N H Hm om v H H H e H l‘ N H N mN H H mN H 5N NP MH mmvo H H Ch HHNHHH H O N fi‘HLflHNNH H mN m m vH N H mN vN l‘fl' MN 0 H H H H 1‘ NH H N mH m H N MN H NN H co co H H H H H 0H NH e N H H N N N N In M 0" H .... H HU.DGOUV mH OHDMB 98 III.4.5. Analysis of Direct Private Costs 111.4.5.1. Cost of Secondary Schooling per Child per Year The annual total cost of secondary schooling per child (YCOST) constitutes the total amount of money spent on secondary schooling per child by the family. The data included in this file (COST.SYS) allows complete computation of the direct private cost of secondary schooling in Burundi. These variables were collected as SCFEES (school fees), PNEED (personal needs), CLOTH (uniforms), SHOES (shoes), BLANK (blankets); and BSHEET (bed sheets), MISC (miscellaneous items). The variable YCOST is the total amount of money spent on secondary school per child per year per family. The following costs are computed: * average annual cost of secondary education (boarding and day schools); * average annual cost of education of secondary boarding schools; * average annual cost of secondary boarding education for boys; * average annual cost of secondary boarding education for girls; * average annual cost of education in day schools; * average annual cost of secondary day school for boys; 99 * average annual cost of secondary day school for girls; * variation of the cost of education by region and family background. III.4.5.2. Economic Burden of Secondary Education on Parents The total direct private cost of secondary school as a proportion of the net household income (and as a proportion of total family expenditure) indicates the burden of the cost of secondary schooling on parents. It is assumed that the family is the unit of analysis. The burden of the schooling cost of one child on the family’s income will be analyzed and compared in terms of family background. This analysis will identify characteristics of families which can support one or more children in secondary school. This indicates the affordability of secondary education by parents, and the number of children they are financially able to support. 100 111.5. Access to Secondary Education 111.5.1. Multivariate Analysis A multivariate analysis is used to analyze the influence of economic burden of direct private cost, family characteristics, and other factors related to access to secondary education. The dependent variable is access to secondary school. Access is measured by the presence of a secondary student in the family (SECOSCHL), or by whether or not a child of secondary-school age is in secondary school. Separate equations are estimated for the entire sample, the rural sub-sample, and the urban sub-sample. The independent variables affecting this access are defined in the next sections. 111.5.1.1. Independent Variables at the Family Level. Based on the literature review, independent variables, important in terms of the way they influence the child's access to education, are as follows: * FATHEDUC: the level of education of the father of the children; * MOTHEDUC: the level of education of the mother of the children; * FATHEROC: the main occupation of the father; 101 * MOTHEROC: the mother’s occupation; * INCOME: total monetary income of the family, * ASSETS: the index of wealth of the family; * FAMILED: the presence of educated relatives in the extended family of the child; * PROV: the location of the family either in the rural or urban areas (rural- urban: Bujumbura versus Gitega and Karuzi; * BURDEN: the economic burden of the direct private costs of secondary schooling; and * T1MES6: the number of times children repeat in the last grade of primary schooling. 111.5.1.1.1 Father Education (FATHEDUC) The father’s education is expected to play an important role in differentiating between children’s access to secondary education. There were more educated men than women in the age group of the sample. Women had a lower level of education or no formal education at all. The variable FATHEDUC was created with SPSS. During the data collection, this information was collected either as the respondent’s level of education or it was located in the respondent’s spouse’s level of education if the respondent was a female. 102 The variable FATHEDUC is a result of information obtained via a combination of two variables: RESPLVED (level of education of the respondent) and RESSPLED (level of education of the respondent’s spouse). The first set of data consisted of all male respondents (sex = 1). The information came from the variable RESPLVED (level of education of the respondent). The command select "if ASEX, respondent’s sex = 1" was used. The data was saved as FATHEDUC with SPSS. The second set of data was obtained from the variable RESSPLED (level of education of the respondent’s spouse ). The command "select if ASEX, respondent’s sex = 2" was used. The set of data was saved as FATHEDUC. The two sets of data were merged and the data are all located in the same variable FATHEDUC. a) Father’s Education Across the Regions. Some regions have had better access to education than others. To show that FATHEDUC is a meaningful variable, it will be of interest to compare the father’s education for the three different regions of the study as indicated in Table 16 below. It was expected that urban areas would have more educated fathers than would rural areas. The farther we go from Bujumbura, the lower will be the levels of education of the father. Table 16 shows the comparison of fathers’ education and mothers’ education. 103 Table 16. FATHEDUC and MOTHEDUC in the whole sanple and per region. Level Gi tega Karuzi Bu j urbura Total Sanple :Lication FATHER MOTHER FATHER MOTHER FATHER MOTHER FATHER MOTHER (48) (56) (42) (50) (52) (85) (142) (191) x X X Z X X X X No school 60.4 64.3 52.4 84.0 19.2 37.6 43.1 57.8 low Prim 10.4 10.7 19.0 10.0 1.9 7.1 9.9 8.8 Upper Prim 18.8 16.1 16.7 2.0 19.2 27.1 16.8 16.1 Lou Sec 10.4 8.9 11.9 4.0 59.6 28.2 30.2 17.1 Total 100 100 100 100 100 100 100 100 It was expected that regions with more educated fathers would have more children with secondary school experience. A cross tabulation of father’s education and the proportion of children with secondary school experience would indicate how regions vary in terms of education of the father and its consequences for the children’s experience with secondary school (see the table below?‘. This is as an exploratory analysis to examine the covariation in father’s edugation and number of family’s children in secondary school (INSEC) before finalizing the more complex analysis. The variable INSEC (number of children enrolled in secondary school per family) was created with SPSS. The variable INSEC (child in secondary school coded 1 yes and 0 for no) was created from GRADENOW. Table 17 below indicates the total number of children who were still in secondary school per family by the 6 The same cross tabulation of INSEC are shown for all the variables which are thought to influence access to secondary schooling such as parents’ education, occupation, income, expenditure, and assets. 104 fathers’ level of education (FATHEDUC). It shows the percentage of fathers who had no child, one, two and more children in secondary school. This included fathers who have never been to school, those who had completed the lower level of primary school, the upper primary school, and those who have been to secondary school) Table 17. Cross tabulation of INSEC (number of children with secondary experience) and father’s education (FATHEDUC) FATHEDUC INSEC NONE 1 2 2< in t in t in t in t 100% (142) No SCHOOLING 69.0 29.0 1.6 0.0 100 (62) Low PRIMARY 57.1 42.9 0.0 0.0 100 (51) UPPER PRIMARY 46.0 29.2 0.0 4.0 100 (13) SECONDARY & UP 7.0 42.0 14.0 39.5 100 (6) III.5.1.1.2. MOTHER EDUCATION (MOTHEDUC) This variable was created in the same way that the variable FATHEDUC was created. A combination of two sets of variables, RESPLVED, RESSPLED resulted in the variable MOTHEDUC. - RESPLVED (level of education of the respondent); - RESSPLED (level of education of the respondent's spouse). The first set of data consisted of all the female respondents (sex = 2) in the variable RESPLVED. These data were selected with the command "select if respondent' sex is 2. They were lOS saved as MOTHEDUC. The second set of data of MOTHEDUC were located in the respondent’s spouse level of education RESSPLED and the respondent was a male (sex = 1). These data were selected using the command "select if sex = 1" and were saved as MOTHEDUC. The data of MOTHEDUC selected from the RESSPLED were merged with the data of MOTHEDUC selected from the RESPLVED. They are all located in the same variable MOTHEDUC. a) Mother Education Across Regions The sample may contain a smaller number of mothers who have had access to secondary school themselves because previous data show that female literacy is lower than male's in Burundi. Since the number of educated mothers is expected to be small, this means that only a few privileged children will have educated mothers. This privilege is intended to be observable in terms of the number of children who have had secondary school experience among the children from 12 to 25 as indicated in Table below. The number of children who have secondary school experience indicates the possible impact of the level of education of the mother to children's secondary schooling. 106 Table 18. 7 Cross tabulation of INSEC (number of children in sec) and the level of education of the Mother (MOTHEDUC) MOTHEDUC INSEC NONE 1 2 2< in % in % in t in % Total% N NO SCHOOLING 52.0 43.6 2.6 1.8 100 (75) LOW PRIMARY 64.7 29.4 5.9 0.0 100 (82) UPPER PRIMARY 16.0 52.2 19.3 12.9 100 (18) SECONDARY & UP 6.0 39.4 18.2 36.4 100 (16) III.5.1.1.3. Main Occupation of the Father (FATHEROC). Data for this variable were located in the respondent occupation (RESPOCC) when the respondent is male and the rest were located in the spouse occupation (SPOSEOCC) when the respondents were female. For the data located in the respondent’s occupation (RESPOCC) they were selected (if sex = 1) and copied under the new variable father occupation (FATHEROC). When the respondent was a female, the data located in the respondent’s spouse occupation (SPOSEOCC) were selected (if the respondent's sex = 2) and copied under the new variable father occupation (FATHEROC). The two sets of variables formed the variable FATHEROC. See Table 20 of comparison of FATHEROC and MOTHEROC below. 7 Tables of INSEC and the other family variables (FATHEROC, MOTHEROC, INCOME, .ASSET, and. IEXPENSES) are included. in appendix.... 107 III.5.1.1.4. Mother’s Occupation (MOTHEROC) Likewise, the data for this variable are located in the RESPOCC when the respondents were female and the rest are located in the spouse occupation (SPOSEOCC) when the respondents were male. For the data located in the respondent’s occupation (RESPOC) they were selected (if respondent’s sex = 2) and copied under the new name MOTHEROC. When the respondent was a male, the data located in the respondent’s spouse occupation (SPOSEOCC) were selected (if the respondent’s sex = 1) and copied under the new name MOTHEROC. The two sets of variables made a single set of variables of MOTHEROC. Table 19. FATHEROC and MOTHEROC in the Whole Sample and per Region. Level of Gitega Karuzi Bujumbura Total Sample education FATHE MOTHE FATHER MOTHER FATHER MOTHER FATHER MOTHER ROC ROC CC CC CC CC CC CC (46) (SS) (41) (50 (45) (81) (132) (186) t t % % t % '% % FARMER 54.3 90.9 75.6 90.0 2.2 2.5 45.8 52.9 GEN LAB 23.9 5.5 7.3 2.0 26.7 46.9 16.6 20.8 TECH 17.5 0 O 4.9 4 O 17.8 16.1 12.1 6.7 BUSINES 0.0 0.0 2.4 0.0 4.4 19.8 2.5 8.0 PROFES 4.3 3.6 9.8 4.0 48.9 19.8 22.9 11.6 Total 100 100 100 100 100 100 100 100 108 III.5.1.1.5 INCOME: Total Yearly Nominal Monetary Income of the Family This variable was computed from the summation of 12 variables: AGPROD: sold agricultural produce; LIVEST: sold livestock ; PROPTY: sold property; SALAR: salary of the respondent; SPSAL: salary of the spouse; SOCSEC: social security; GIFTS: received gifts; HIREDLAB: money from respondent hired labor; FRMLAB: money from the family members' hired labor; BASSOON: money from a business; RENTHSE: money from rented houses; MISCEL: Money from miscellaneous sources. Table 20 indicates the distribution of income by source in the areas sampled. 109 Table 20 Average Family Income by Source in The Sampled Areas, Burundi, 1992-93. Variable Mean (FBu) All Gitega Karuzi Bujumbura (N=186) (n=54) (n=48) (n=84) AGPROD 26663.66 53028.52 35872.92 4452.38 LIVEST 5358.06 1722.22 4762.50 8035.71 PROPTY 607.53 .00 1729.17 357.14 SALAR 82378.49 41870.37 26258.33 140488.10 SPSAL 78161.29 16111.11 14500.00 154428.57 SOCSEC 5570.73 689.26 433.33 11644.48 GIFTS 15338.71 11074.07 1875.00 25773.81 HIRELAB 1215.59 2400.00 260.42 1000.00 FARMLAB 20867.74 622.22 187.50 45700.00 BASSOON 260302.15 23859.26 11312.50 554580.95 RENTHSE 108941.94 5333.33 6250.00 234228.57 MISCEL 2443.22 2117.37 35.42 4028.57 III.5.1.1.6. ASSETS: Total Assets of the Family. Data for this variable were created from a weighted average of the following variables: - RADIO: radio (assigned the value: 1); - NEWS: newspaper (assigned the value: 1. This is not an expensive asset but it distinguishes between those who consider it an item worthy of spenditure and can afford it; * CAMERA: camera: value 1; * LIGHT: electricity in the house: value 3; * RURHSE: house location (1= rural; 2 = urban); * THATCHED: thatched house (value: 1); 3); * MOTO: motorcycle (value * TELE: television (value 3); 110 * CAR: car (value = 5); * FRIDGE: refrigerator (value 3); * RUNWAT: running water in the house (value = 3); * IRONHSE: house with iron sheet roof (value 2); * CEMENT: cemented house (value = 3); * BIKE: bike (value = 2); * COW: cows (values: 1 cow 1; 2-5 cows =2; more than 5 =3); * SHEEP: sheep (values: 1); * GOAT: goat (values: 1-10 1; more than 10 goats = 2); * PIGS: pig ( values 1-5 pigs = 1; more than 5 = 2); * CHICKEN: up to 10 chickens = 1; more than 10 = 2); and * COFFEE: coffee trees (values: up to 200 trees = 1; 201-400 =2; 401-600 = 3; more than 600 = 4). Each of these values was assigned in comparison to its importance in monetary equivalent value. The value of 1 was assigned for an asset valued up to 3000 FBu (around $10) or those which are common equipment. The higher the cost of the asset was, the higher its value. These values were added up and the total value per family was called ASSETS. A high value of ASSETS for a family indicates that the family had more aggregate wealth than a family with a lower value. The value of family home (HOUSE) was constituted by all the items characteristics of each family house. These items are: rural house or urban house, thatched or iron roof, cement, electricity, and running water. The highest value of 14 represented an urban house with all the items except rural and 111 thatched roof. The lowest value of two indicated a rural, and thatched roof house. The distribution of the items included in ASSETS is summarized in Table 21 below. 112 Table 21. Items of the Family Assets (ASSET), Burundi, 1993. Items PROV 1 PROV 2 PROV 3 ALL Number % Number % Number % Number % Radio 42 72.4 23 54 61 71.8 126 65.3 Newsp. 15 25.9 7 14 28 32.9 50 29.9 VCR 0 0.0 0 0 6 7.1 6 3.1 Camera 2 3.4 0 O 4 4.7 6 3.1 Light 5 8.6 0 o 42 49.4 47 24.4 Urbhse 8 13.8 2 4 17 90.6 87 45.1 Rurhse 50 86.2 48 96 1 1.2 99 51.3 Thach. 9 15.5 20 40 1 1.2 30 15.5 Moto 2 3.4 0 0 5 5.9 7 3.6 Tele 2 3.4 1 2 31 36.5 34 17.6 Car 0 0.0 0 0 11 12.9 11 5.7 Fridge 1 1.7 O 0 19 22.4 20 10.4 Runwat 3 5.2 1 2 42 49.4 46 23.8 Ironhs 46 79.3 30 6O 69 87.2 145 75.1 Cement 26 44.8 13 26 62 72.9 101 52.3 Bike 29 50.0 18 36 19 22.4 66 34.2 Cow 1 5 8.6 3 6 - - - - Cow 2 1 1.7 1 2 2 2.4 4 2.1 Cow 3 0 0.0 3 6 4 4.7 7 3.6 Sheep 5 8.6 7 14 3 3.5 15 7.8 Goat 1 21 36.2 18 36 3 3.5 42 ‘ 21.8 Goat 2 O 0.0 6 12 2 2.4 8 4.1 Pig 1 9 15.5 1 2 0 0.0 10 5.2 Pig 2 7 12.1 0 0 0 0.0 7 3.6 Chickl 0 0.0 2 4 0 0.0 2 2.1 Chick2 1 1.7 0 0 3 3.5 4 2.1 Coffel 33 56.9 19 38 1 1.2 53 27.5 Coffe2 6 10.3 14 28 2 2.4 22 11.4 Coffe3 4 6.9 6 12 1 1.2 11 5.7 Coffe4 3 5.2 0 O 0 0.0 3 1.6 Table 21 cont'd Housel l 1. 2 9 15 3 22 37 4 1 1 S 1 1 6 15 25 7 4 6 8 0 6 9 O O. 10 2 3 11 O 0. 14 3 S. COW‘OQQWU’IQ .h 113 16 01401000 as 38 18. .h \l 44. OHQWU ONHbU‘ImNm 35 39 17 34 11 41 21. OUTQO‘I—‘QNHUT O1 114 III.5.1.1.7. New Variables in the Family File. a) Predicted Burden of the Direct Private Cost (PREDBURD) The predicted burden of the private cost of secondary schooling (PREDBURD) on the family was computed for all the families in the sample. These families included those with and without children in secondary school, for families which fully financed their children in secondary school as well as those which did not. This variable was estimated by using the burden of the direct private cost of secondary schooling of one child (BURDEN). This burden was obtained from children whose families fully financed their secondary schooling. b) Real Values of the Burden (RBURDEN) A new variable RBURDEN was created using the true values of BURDEN for families which were fully financing their children (73 families) and predicted values of burden PREDBURD for all other families. 115 c) Number of Repetition in Sixth Grade (TIME6) This variable was collected at the child level. Some families had children who attended sixth grade and did not repeat this grade. Among these children who did not repeat sixth grade, some were admitted to seventh grade and others were not admitted to secondary school. Among these children some were still in school, while others had already dropped out of school. Regardless of their condition at the time of the data collection, for all children, the number of times they repeated sixth grade was collected. The highest number of times a child repeated sixth grade in each family was entered to represent the effort of the family to get access to secondary school. The range of this variable was between 0 and 6. Zero was entered for a family where none of the children repeated sixth grade. The distribution of the highest number of times a child repeated the sixth grade is shown in Table 22 below. The following steps were used to obtain TIME6 variable: first, compute the number of children age 12 or more; second, compute the number of children of secondary school age (12-28 years); third, identify the child with highest number of sixth grade repetitions per family; fourth, drop all other cases (children); fifth, retain the variable (TIME6) of the child with the highest number of sixth grade repetitions to represent the family and; sixth, add the TIME6 for that child to the family file. 116 Table 22 Distribution of TIME6 TIME6 Frequency % 0 34 17.6 1 43 22.2 2 54 28.0 3 38 19.7 4 16 8.3 5 3 1.6 6 5 2.6 Total 193 100 d) Repetition in Sixth Grade Per Family (REP) There were families whose children never repeated sixth grade. Regardless of their situation, the variable REP was created. It was coded 0 or 1. The value of zero was assigned to families (34 or 17.6%) with no child who repeated sixth grade. The value of one was assigned to families (159 or 82.4%) which had at least one child who repeated sixth grade. 117 e) Families With Children Who Never Enrolled in Formal School (BEENSCH). The variable BEENSCH (total number of children who never enrolled in school in the family) was computed for each family in the family file. During the data collection, this variable was collected for each child. It is coded O or 1 in the child file. The code zero means that the child never enrolled in primary school. The code 1 means that the child enrolled in primary school. In the family file, the total number of children who never enrolled in school per family is coded 0 through 6. Families with all their children enrolled in school represented 77.7%, whereas there were 22.3% of families who had children, in the age-group of secondary school, who never enrolled in school. The distribution of families with the number of their children who never enrolled in school is presented in the Table 23 below. 118 Table 23 Distribution of Families with Number of Children Never Enrolled in Primary School, Burundi, 1992-93 BEENSCH Frequency % 0 150 77.7 1 16 8.7 2 9 4.7 3 11 5.7 4 5 2.6 5 2 1.0 6 5 2.6 Total 193 100 III.5.8. Parents’ Values and Beliefs About Secondary Schooling Nearly 92% of the parents interviewed strongly agreed that secondary schooling in Burundi was expensive. Five percent of the parents agreed that secondary schooling was somewhat expensive and three percent did not consider secondary schooling as expensive at all. The results on the parents’ perception about the direct private costs are shown in Table 24 below. 119 Table 24 Opinion About Secondary Schooling Direct Cost (OPWILPAY) Burundi, 1992-93 School Year Opinion Frequency % Not expensive at all 5 2.6 A little bit expensive 9 4.7 Expensive 47 24.5 Very expensive 131 68.2 Total 192 100 Despite the perception of high direct private cost of secondary schooling, 98% of the parents strongly agreed that secondary education was useful. Only 2% of the parents found education either not useful at all or were not convinced of its usefulness. Respondents included 48% male and 52% female. The results related to the usefulness of secondary education are summarized in Table 25 below. Table 25: Opinion on the usefulness of secondary school (USEFSCHL) in Burundi, 1993 Opinion Frequency % Not at all useful 1 .5 Somewhat useful 3 1.6 Useful 28 14.5 Very useful 161 83.4 Total 193 100 The major reasons given for sending children to secondary school were that educated people had a better life, helped their parents, and earned more than those who did not have a secondary education. CHAPTER IV. DIRECT PRIVATE COST OF SECONDARY EDUCATION IN BURUNDI IV.1. Type of Cost, Objectives, and Sample Direct private costs of secondary education in Burundi consist of parental spending on school fees, personal needs, school uniforms, shoes, blankets, bed sheets and miscellaneous items related to secondary education. This research estimates the annual spending on secondary education by families who had children in secondary school for the school year 1992-93. The amounts are in Burundi francs (FBu). In 1992-93, there were 254 FBu per one US dollar. The following describes each variable item embodied in the direct private cost in 1992—93. School fees consisted of a fixed charge by the government of Burundi per individual student in secondary school. School fees were charged according to the type of school. These fees were paid at the beginning of the school year or in three installments. Each installment was due at the beginning of the term. School fees were regulated by the government in public schools only. Private schools were not subject to any school fee regulations. Therefore, each private school charged its own school fees. 120 121 Personal needs included school supplies, transportation and personal care. Personal needs varied from student to student because expense requirements for each type of school were different. Transportation cost depended on the distance between home and school; there could be large variations in such expenses. Expenses for personal care depended on the gender and wealth of the family/relatives. School uniforms were compulsory in every school in Burundi, beginning with primary for private and public schools. In addition, apart from public primary schools, each secondary school selected its own school uniform’s colors. Consequently, any student who changed schools had to buy new uniforms. Spending on school uniforms varied according to the type of fabric selected and the number of uniforms parents chose to provide for the student. Shoes were included in the analysis because they were considered a luxury. Most people in rural areas in Burundi do not wear shoes on a daily basis. As such, most students wore shoes primarily for school purposes. They probably would not wear or own shoes if they were not in school. Bedclothes consisted of blankets and bed sheets.‘ At least one blanket and a set of bed sheets were required in every boarding secondary school. Buying a blanket or bed sheets for each child in secondary boarding school constituted an extra cost to the family because children who do not go to secondary school share one bed (e.g. two or 122 more children in one bed); and do not require individual sets of bedclothes. Miscellaneous consisted of money used for school related materials purchased by the student. Some parents provided only money to the student. These parents made their children responsible for managing the funds they provided to obtain all the required school materials. In such case, the students had to buy everything they needed for schooling. In some cases, children were given all the school supplies and some pocket money for their personal needs. As such, miscellaneous included some personal needs items, and in some cases all of the other items if parents gave all the money to the children. Annual expenditures by the parents on school fees, personal needs, school uniforms, shoes, blankets, bed sheets, and miscellaneous items are represented by the variables SCFEES, PNEED, CLOTH, SHOES, BLANK, BSHEET, and MISC. The sum of all these costs constituted the total annual direct private cost of secondary school (YCOST). Thus, YCOST = SCFEES + PNEED + CLOTH + SHOES + BLANK + BSHEET + MISC . The analysis of the direct private costs of secondary education in this chapter is aimed at answering the following questions: * How much money did parents spend on secondary schooling per child in Burundi during the school 123 year 1992-1993? * How did the direct private cost of secondary schooling vary by family background, type of school (day or boarding), gender of the student and region? * What was the magnitude of economic burden of these costs on parents? Not all secondary students in the sample were included in answering questions related to the direct private cost of secondary education. Out of the 210 secondary school students in the sample, 74 percent were fully supported financially by their parents. The analysis, in this chapter, focuses on students who were fully sponsored by their parents because parents knew the direct private cost of their children's secondary schooling. These parents constituted 65% of all families with children in secondary school. Among the secondary students fully supported by their parents, 51% were boarding students and 49% were day students. There were 32 public school boarding students (43% of all public boarding students) from Gitega (PROV 1); 22 students (29%) from Karuzi (PROV 2); and only 21 students (28%) from Bujumbura. There were 63 day students (97%) from Bujumbura, one student (1.5%) from Gitega and another one (1.5%) from Karuzi. The remaining 26 percent of the students were helped by relatives or friends. Secondary students helped by 124 relatives or family friends were not included in this analysis because their parents did not necessarily know the amount of money that their relatives or friends spent on the secondary schooling of their children. IV.2. Direct Private Cost of Secondary Schooling This section estimates the direct private costs of secondary education per student in Burundi. The analysis considers differences in the type of school, region, gender, education and occupation of the parents, family wealth, grade level of the students and the number of children in secondary school in the family. In the Tables 26- 47 below, the number of students in each sub-sample is indicated at the top of the table in parentheses. The unit of analysis is the student. IV.2.1. Types of Public School The direct private costs of public day schools and public boarding schools are shown in Table 21. (Direct private costs of private schools, also shown in Table 26, will be analyzed later in the chapter.) The total direct private cost was 26,256 FBu for boarding school students and 20,725 FBu for public day school students. The total cost was higher for boarding 125 schools because of higher school fees and boarding related expenses. However, day school students had higher spending on personal needs, clothing, and shoes. Most students attending day school lived in Bujumbura, where transportation costs were high. These students had to ride buses to and from school at least three times a day (some students walked home after school). Moreover, life style necessities in the capital city required expenditures that might not be necessary in rural areas where most public boarding schools were located. Consequently, day public schools were more expensive than public boarding schools. Table 26: Annual (YCOST) and Itemized Direct Private Costs (FBu and percentage) in Both Private and Public school, Burundi, 1992-93 Type of Private school Public Schools cost (16) Day (73) Boarding (70) Amount % Amount % Amount % SCFEES 19,825 40.3 4,012 19.4 9,000 34.3 PNEED 12,381 25.2 9,885 47.7 7,284 28.0 CLOTH 7,482 15.2 4,260 20.6 3,958 15.0 SHOES 3,247 6.6 2,504 12.1 1,983 7.6 BLANK 1,164 2.4 0 0.0 1,161 4.4 BSHEET 1,373 2.8 54 0.2 1,357 5.2 MISC 3,669 7.4 0 0.0 1,513 5.8 YCOST 49,142 100 20,725 100 26,256 100 126 IV.2.2 Costs Across Provinces The comparison of costs across provinces was possible only for public boarding schools because there were very few public day schools in the rural areas of Burundi. Only two cases of day students in the commune of Gitega and one case in the commune of Buhiga were observed. Almost all day school students were found in Bujumbura, the capital city. Table 27 presents direct private costs of public boarding schools in the three provinces of the study. School fees were standard in all public secondary schools; they represented a major part of the total direct private cost of secondary school education in all three provinces. The other costs differed among the provinces. Expenses in the Karuzi province, which was more rural, were mainly high for personal needs, clothing and miscellaneous items. Since the province of Karuzi was located in a remote area, parents had to make sure that their children had the necessary school materials to prevent students from running out of such materials before the end of the term. It might be more costly to replace lost materials or replenish the stock of school materials. Thus, students attending secondary school in Karuzi were in charge of managing the money given by their parents to buy needed school materials and clothes. These three items alone, i.e. personal needs, miscellaneous and 127 clothing, represented 55% of the total direct private costs in Karuzi. In the province of Gitega, most of the expenses were directed to school fees and personal needs. These items represented 61.3%, (41% for school fees, and 20.3% for personal needs), of the total direct private cost. The personal needs item, in Gitega, was the lowest of the three provinces because students were walking to and from school, thus saving on transportation expenses. Being the second largest city in Burundi, Gitega also necessitated a lifestyle which is closer to that of Bujumbura. Similarly, school fees and personal needs were the most expensive items in Bujumbura. These items represented 60% (29.7% for school fees, and 30.3% for personal needs) of the total direct private costs. 128 Table 27: Direct Private costs of Public Secondary Boarding Schools by Region, Burundi, 1992-93 Type Public Schools gist PROV GITEGA PROV KARUZI BUJUMBURA (32) (22) (21) Amount % Amount % Amount % SCFEES 9,000 41.0 9,000 35.3 9,000 29.8 PNEED 4,459 20.3 7,843 30.5 9,205 30.4 CLOTH 2,931 13.3 3,750 14.6 5,090 16.8 SHOES 2,294 10.4 1,570 .0 1,786 .0 BLANK 1,390 6.3 445 1.7 1,719 .6 BSHEET 1,412 6.4 563 .2 2,152 .0 MISC 500 2.3 2,545 10.0 1,333 .4 YCOST 21,986 100 25,716 100 30,285 100 IV.2.3 Gender Differences In both types of public school, direct private costs of secondary school were higher for female than for male students (see Table 28). These costs were 6.9% higher for boarding schools and 6.3% for day schools. In boarding schools, female students tended to require more money on almost all items except shoes. In day schools, female students needed more money for personal needs, clothing, and shoes, whereas, male students spent more on school fees. 129 Table 28 Direct Private Costs of Secondary Education (in FBu and percentages) by Gender and Type of School, Burundi, 1992-93 Type of Public Schools cost Day Boarding Male Female Male Female (32) (41) (31) (39) Amount % Amount Amount % Amount % SCFEES 4,102 20.5 3,958 18.6 9,000 35.6 9,000 33.3 PNEED 9,632 48.1 10,086 47.4 6,830 27.0 7,648 28.3 CLOTH 3,825 19.1 4,606 21.6 3,669 14.5 4,190 15.5 SHOES 2,465 12.3 2,535 11.9 2,145 8.5 1,854 6.9 BLANK 0 0.0 0 0.0 1,011 4.0 1,281 4.7 BSHEET 0 0.0 97 0.5 1,174 4.6 1,503 5.6 MISC 0 0.0 0 0.0 1,463 5.8 1,552 5.7 YCOST 20,024 100 21,282 100 25,292 100 27,028 100 IV.2.4. Parental Education The education level of fathers and mothers was related to the direct private costs of secondary school. Parents with lower levels of education generally paid less for the secondary education of their children. In all cases school fees, personal needs, clothing and shoes were the major components of the direct private cost of secondary schooling. IV.2.4.1 Fathers' Education (FATHEDUC) The results of the analysis of the costs data with respect to fathers' education are summarized in Table 29 and 30 below. Fathers with no formal schooling spent the least 130 on secondary schooling and their children were almost exclusively in public boarding schools. Fathers who reached upper primary school education tended to spend more on the secondary schooling of their children in both day and boarding schools. In secondary boarding school, direct private cost increased with the fathers' level of education except for the fathers with secondary schooling experience. Compared to fathers who never attended formal school, fathers with a lower primary level of education spent 16.5% more, those with an upper primary level of education spent 28.6% more, and those with at least some secondary schooling spent 16.8% more. In secondary day school, direct private cost increased with the fathers' level of education except for the fathers with secondary schooling experience. Compared to fathers who never attended formal school, fathers who had lower primary spent only 2.8% more, those with an upper primary spent 46.3% more, and those with at least some secondary schooling spent only 2.4% more. 131 Table 29: Costs of Secondary Boarding School (FBu, percentages) by Fathers' level of Education, Burundi, 1992-93 Type of Fathers' Level of Education (FATHEDUC) COSt No Schooling Lower Prim Upper Prim Secondary & Up (17) (17) (22) (9) Amount % Amount % Amount % Amount % SCFEES 9,000 40.0 9,000 34.0 9,000 31.0 9,000 34.0 PNEED 5,832 26.0 7,038 27.0 9,909 34.0 5,298 20.0 CLOTH 2,687 12.0 3,159 12.0 4,708 16.0 4,825 18.0 SHOES 1,308 6.0 2,053 8.0 1,962 7.0 2,864 11.0 BLANK 718 3.0 709 3.0 1,511 5.0 1,815 7.0 BSHEET 730 3.0 1,057 4.0 1,894 7.0 1,758 7.0 MISC 2,267 10.0 3,256 12.0 0 0.0 762 3.0 YCOST 22,542 100 26,272 100 28,984 100 26,322 100 Table 30: Costs of Secondary Day School by Fathers' level of Education, Burundi, 1992-93 Type Fathers’ Level of Education (FATHEDUC) gist No Schooling Lower Prim Upper Prim Secondary & (2) (7) (39) Up (24) Amount % Amount % Amount % Amount % SCFEES 3,750 23.0 4,367 26.0 4,114 17.0 3,786 22.5 PNEED 8,200 50.0 6,703 40.0 11,980 50.0 7,552 45.0 CLOTH 3,500 21.0 3,254 19.0 5,117 21.2 3,219 19.0 SHOES 975 6.0 2,562 15.0 2,726 11.4 2,262 13.5 BLANK 0 0.0 0 0.0 0 0.0 0 0.0 BSHEET 0 0.0 0 0.0 100 0.4 0 0.0 MISC O 0.0 0 0.0 0 0.0 0 0.0 YCOST 16,425 100 16,886 100 24,037 100 16,819 100 132 IV.2.4.2 Mothers' Education In boarding schools, the more educated the mothers were, the higher the expenditures were for the secondary education of their children (see Table 31). However, in day schools, mothers with only lower primary schooling spent the most on their children's schooling. In addition, the more educated the mothers were, the more likely their children were to attend day schools. Less educated mothers tended to have children in public boarding schools. The main explanation of this phenomenon is that more educated mothers lived in urban areas where day schools were located, while less educated mothers lived in the rural areas. Table 31: Costs of Secondary Boarding School by Mothers' level of Education, Burundi, 1992-93 Type Mothers' Level of Education (MOTHEDUC) gist No Schooling Lower Prim Upper Prim Secondary & Up (28) (26) (15) (2) Amount % Amount % Amount 1: Amount % SCFEES 9,000 38.0 9,000 33.4 9,000 31.0 9,000 30.0 PNEED 6,581 27.5 6,538 24.3 10,010 35.0 5,761 19.0 CLOTH 3,277 14.0 4,139 15.3 4,858 17.0 4,261 14.0 SHOES 1,647 7.0 2,186 8.0 2,139 7.0 2,946 10.0 BLANK 901 4.0 1,359 5.0 1,371 5.0 554 1.0 BSHEET 940 4.0 1,524 6.0 1,490 5.0 4,261 14.0 MISC 1,559 6.5 2,203 8.0 0 0.0 3,693 12.0 YCOST 23,905 100 26,949 100 28,868 100 30,476 100 133 Table 32: Costs of Secondary Day School by Mothers' Level of Education, Burundi, 1992-93 Type Mothers' Level of Education (MOTHEDUC) gist No School Lower Prim Upper Prim Secondary & (S) (18) (21) Up (29) Amount % Amount % Amount %' Amount % SCFEES 3,000 17.8 4,125 18.0 4,049 19.0 4,096 21.0 PNEED 7,500 44.4 11,325 49.0 10,069 47.3 9,235 47.4 CLOTH 4,625 27.4 4,769 21.0 4,957 23.3 3,392 17.4 SHOES 1,750 10.4 2,822 12.0 2,221 10.4 2,627 13.5 BLANK 0 0.0 0 0.0 0 0.0 0 0.0 BSHEET 0 0.0 0 0.0 0 0.0 134 0.7 MISC 0 0.0 O 0.0 O 0.0 0 0.0 YCOST 16,875 100 23,041 100 21,296 100 19,484 100 IV.2.5. Direct Private Cost and Parental Occupation Analysis of the direct private costs of secondary education by parents’ occupational groups indicated that such costs varied with parents’ occupation. School fees, personal needs, clothing and shoes were the major items in the direct private cost of secondary education. Regarding secondary boarding schools, in term of fathers' occupation, general—laborers spent the least on secondary schooling for their children, followed, in ascending order, by farmers, technicians, businessmen, and professionals (See Tables 33). Professionals spent 29.8% more than general-laborers, 28.3% more than farmers, 13.1% more than technicians, and 7.5 % more than businessmen. 134 In terms of mothers' occupation, farmers also spent the least amount on the secondary schooling of their children, followed by technicians, general-laborers, professionals, and business women. Business women spent 34.3% more than farmer-mothers, 20.4 % more than technician-mothers, 14.8% more than general-labor mothers, and 14.3% more than professional mothers (see Table 34). Table 33: Costs of Secondary Boarding School by Fathers’ Occupation, Burundi, 1992-93 Type Fathers' Occupation (FATHEROC) gist Farmer Gen Labor Business Technic Profess (17) (3) (14) (5) (16) Amount X Amount X Amomt X Amount X Amomt X SCFEES 9,000 40.3 9,000 41.0 9,000 31.0 9,000 33.0 9,000 28.9 PNEED 6,454 28.9 4,980 22.7 7,820 27.0 3,883 14.3 1,105 35.5 CLOTH 3,097 13.9 2,820 12.8 4,892 17.0 907 3.4 5,119 16.4 SHOES 1,197 5.4 2,330 10.6 2,464 8.6 883 3.3 3,102 10.0 BLANK 521 2.3 1,400 6.4 2,189 7.6 0 0.0 1,181 3.8 BSHEET 588 2.6 1,440 6.5 2,529 8.8 0 0.0 1,681 5.4 MISC 1,469 6.6 0 0.0 0 0.0 12,392 46.0 0 0.0 YCOST 22,327 100 21,870 100 28,794 100 27,065 100 31,141 100 13S Table 34: Costs of Secondary Boarding School by Mothers' Occupation, Burmdi, 1992-93 Type of Mothers’ Occupation (MOTHEROC) cost Farmer Gen Labor Business Technic Profess (35) (4) (10) (5) (15) Amomt X Amomt 7: Amount X Amount 7: Amomt x SCFEES 9,000 40.0 9,000 30.8 9,000 26.2 9,000 33.0 9,000 30.6 PNEED 5,511 24.4 8,182 28.0 7,792 22.7 10,425 38.2 10,189 34.7 CLOTH 3,042 13.5 5,264 18.0 5,650 16.5 4,000 14.7 4,856 16.5 SHOES 1,975 8.8 2,771 9.5 1,873 5.5 1,375 5.0 2,254 7.7 BLANK 729 3.2 1,763 6.0 1,921 5.6 2,000 7.3 1,283 4.3 BSHEET 862 3.8 2,237 7.7 2,541 7.4 500 1.8 1,817 6.2 MISC 1,430 6.3 0 0.0 5,523 16.1 0 0.0 0 0.0 YCOST 22,548 100 29,217 100 34,301 100 27,300 100 29,398 100 Regarding secondary day schools, there were no day schools in the rural areas in Burundi. Therefore, farmers did not have children in day school. Only families from urban areas, i.e., Bujumbura, were concerned. The interview results are shown in Tables 35 & 36 below. In comparison to other groups, technician fathers spent the least on secondary schooling of their children. They were followed by general-labor fathers, businessmen, and professionals. Professional fathers spent 28% more than technician fathers; 22.1 % more than general-labor fathers, and 1.3% more than businessmen. General-labor mothers spent the highest amount for day school because they lived far from day schools (e.g., Kamenge), which were located in downtown Bujumbura. Thus, personal needs (PNEED), which include transportation, constituted a major item in the direct cost of secondary day 136 school. General-labor mothers spent 31.3% more than technician mothers; 26.8 % more than professional mothers; and 25.4% more than business mothers. Table 35: Costs of Secondary Day School by Fathers' Occupation, Burundi, 1992-93 Type of Fathers' Occupation (FATHEROC) cost Farmer (0) Gen Labor Business Technic (6) Profess (3) (9) (48) Amount % Amount % Amount % Amount % Amount % SCFEES na an 4,500 27.0 3.593 17.0 3.600 23.0 4,151 19.2 PNEED na na 7.667 45.0 8.812 41.0 5.060 33.0 10.688 49.5 CLOTH na na 3.167 19.0 5.145 24.0 4.500 29.0 4.219 19.5 SHOES na na 1.500 9.0 3.780 18.0 2,400 15.0 2.474 11.4 BLANK na na 0 0.0 0 0.0 0 0.0 0 0.0 BSHEET na na 0 0.0 0 0.0 0 0.0 81 0.4 MISC na na 0 0.0 0 0.0 0 0.0 0 0.0 YCOST na na 16.833 100 21.329 100 15,560 100 21.614 100 137 Table 36: Costs of Secondary Day School by Mothers’ Occupation, Burundi, 1992-93 Type of Mothers' Occupation (MOTHEROC) cost Farmer Gen Labor Business Technic Profess (0) (8) (21) (6) (35) Amount % Amount % Amount % Amount % Amount % SCFEES na na 4.500 16.5 4.184 20.7 3.300 18.0 4.016 20.0 PNEED na m 15,743 57.9 8.226 40.5 8.500 45.4 9,906 50.0 CLOTH na na 5.314 19.5 4.963 24.4 4.900 26.0 3.410 17.0 SHOES na m 1.664 6.1 2.926 14.4 2.000 10.6 2.477 12.4 BLANK na na 0 0.0 0 0.0 0 0.0 0 0.0 BSHEET na na 0 0.0 0 0.0 0 0.0 113 0.6 MISC na na 0 0.0 0 0.0 0 0.0 0 0.0 YCOST na na 27.221 100 20.300 100 18.700 100 19.923 100 IV.2.6. Family Wealth Family wealth (cash income and assets) was also related to the direct private cost of secondary education. The results are presented in Tables 37 through 40. The more income and assets a family had, the more it spent on secondary education. Families in the first and second income quintiles did not send their children to day schools. These quintiles included poor families, mainly farmers and general-labor families. They lived in the rural areas and sent their children to boarding school. The third, fourth, and fifth income quintiles represented families mainly from professional, and business backgrounds. Technicians were mainly represented in the third quintile. Their children attended day school. 138 Table 37: Costs of Secondary boarding School by Income, Burundi, 1992-93 Type Family's Income Quintile 2:3. 1 (15) 2 (14) 3 (21) 4 (5) 5 (15) Amount % Amount % Amount % Amount % Amount 96 SCFEES 9.000 40.8 9.000 41.3 9.000 33.2 9.000 32.5 9.000 27.0 PNEED 5.435 24.7 5.634 25.9 8.199 30.3 6.486 24.0 9.777 29.3 CLOTH 3.036 13.8 2.972 13.7 4.273 15.8 4.604 17.0 5.192 15.5 811088 707 3.2 1,793 8.2 3.385 12.6 2.710 10.0 1.231 3.7 BLANK 468 2.1 708 3.2 1.186 4.3 1.520 5.5 2.134 6.4 BSHEET 693 3.1 1.055 4.9 1.011 3.8 3.045 11.0 2.231 6.6 MISC 2.722 12.3 597 2.8 0 0.0 0 0.0 3.846 11.5 YCOST 22.063 100 21.760 100 27.054 100 27,635 100 33.431 100 Table 38: Costs of Secondary Day School Education by Income, Burundi, 1992-93 Type Family's Income Quintile if... 1 (0) 2 (1) 3 (6) 4 (28) 5 (37) Amount % Amount % Amount % Amount % Amount % SCFEES na na na na 3537 20.0 4.500 19.7 3.727 19.0 PNEED na na na na 7903 44.1 11.880 51.8 8.694 44.4 CLOTH na na na na 4409 24.6 4.080 17.9 4.394 22.5 SHOES na na n8 na 2036 11.3 2.312 10.0 2.776 14.1 BLANK na na na na 0 0.0 0 0.0 0 0.0 BSHEET na na na na 0 0.0 140 0.6 0 0.0 MISC na na na na 0 0.0 0 0.0 0 0.0 YCOST na na na na 1.7885 100 22.912 100 19.590 100 139 Table 39: Costs of Secondary Boarding School by Assets, Burundi, 1992-93 Type Families' Assets Quintile of cost 1 (10) 2 (10) 3 (10) 4 (20) 5 (15) Amount % Amount % Amount % Amount % Amount % SCFEES 9.000 42.5 9.000 41.0 9.000 38.0 9.000 27.5 9.000 30.4 PNEED 5.984 28.2 4.739 21.4 5.739 24.0 10.221 31.3 9.149 31.0 CLOTH 2.409 11.3 3.512 16.0 3.584 15.0 4.376 13.4 5.468 18.5 SHOES 1.172 5.5 1.153 5.2 2.535 10.7 3.063 9.4 1.218 4.1 BLANK 172 0.8 1.143 5.0 1.103 4.7 761 2.3 2.381 8.0 BSHEET 366 1.7 1.200 5.4 1.092 4.5 1.553 4.7 2.341 8.0 MISC 2.063 10.0 1.332 6.0 743 3.1 3.733 11.4 0 0.0 YCOST 21.166 100 22.080 100 23.797 100 32.707 100 29.557 100 Table 40: Costs of Secondary Day School by Asset, Burundi, 1992-93 Type Families' Asset Quintile if... 1 (0) 2 (0) 3 (7) 4 (31) 5 (34) Amount % Amount % Amount % Amount % Amount % SCFEES na na na na 4.250 15.1 4.087 21.7 3.900 18.4 PNEED na n8 n2 n8 16.733 59.4 9.122 48.4 9.353 44.2 CLOTH na n na na 5.500 19.5 3.118 16.5 5.113 24.2 SHOES na na na na 1.658 6.0 2.410 12.8 2.787 13.2 BLANK na na na na 0 0.0 0 0.0 0 0.0 BSHEET na mu m m 0 0.0 127 0.6 0 0.0 MISC na na na [18 0 0.0 0 0.0 0 0.0 YCOST na na na na 28.142 100 18.865 100 21.153 100 IV.2.7. Total Family Expenditure Direct private costs also varied with the total family expenditure in 1992-93. The results are presented in both Table 41 and 42 below. In boarding schools, the first expenditure 140 quintile spent less on their children's schooling than any other quintile, followed by the fourth, the second, third, and fifth expenditure quintiles. Families in the fifth expenditure quintile spent 45.8% more than those in the first quintile, 34.7% more than the second quintile, 30.7% more than third quintile, and 35% more than the fourth quintile families. In day school, the fourth quintile families spent 24.6% more than the third quintile. The families in the fourth and fifth expenditure quintiles spent almost the same amount. School fees represented the most expensive item for quintiles one through four in boarding school, whereas personal needs were the major expense for quintile five in boarding school and for all families who had children in day school. Clothing was the third most expensive item in both types of schools. Table 41: Costs of Secondary Boarding School by Total Families’ Expenditure, Burundi, 1992-93 Type of Families' Expenditure Quintiles coat 1 (14) 2 (14) 3 (16) 4 (12) s (14) Amomt X Amomt X Amomt X Amotnt X Ammt X SCFEES 9,000 44.7 9,000 37.1 9,000 35.0 9,000 37.2 9,000 24.2 PNEED 5,139 25.5 5,688 23.5 4,732 18,3 6,135 25.4 15,188 41.0 CLOTH 2,277 11.3 4,524 18.7 2,928 11.4 3,885 16.1 6,400 17.2 SHOES 641 3.2 1,655 6.8 2,851 11.0 1,936 8.0 2,759 7.4 BLANK 332 1.6 1,222 5.0 959 3.7 1,446 6.0 1,947 5.2 BSHEET 347 1.7 1,554 6.4 1,337 5.2 1,769 7.3 1,864 5.0 HISC 2,409 12.0 613 2.5 3,967 15.4 0 0.0 0 0.0 vcosr 20,144 100 24,256 100 25,775 100 24,171 100 37,167 100 141 Table 42: Costs of Secondary Day School by Total Family Expenditure, Burundi, 1992-93 Type of Family’s Expenditure Quintile cost 1 (0) 2 (0) 3 (7) 4 (32) 5 (33) Amount % Amount % Amount % Amount % Amount % SCFEES na na na na 4.250 26.5 4.364 20.5 3.621 17.0 PNEED na 11 na na 7.133 44.5 10.597 50.0 9.745 46.0 CLOTH na n na na 3.117 19.4 4.293 20.0 4.490 21.0 SHOES na na na na 1.533 9.6 1.891 9.0 3.362 16.0 BLANK n2 na na na 0 0.0 0 0.0 0 0.0 BSHEET na n8 n8 n8 0 0.0 122 0.5 0 0.0 mmc 1a m, an 1m 0 00 0 00 0 00 YCOST na na na na 16.033 100 21.259 100 21.218 100 IV.2.8 Direct Private Costs by Grade Level The cost of the first cycle of secondary education (grades seven through ten) was lower than that of the second cycle of secondary education (grades eleven through thirteen) for both day and boarding schools. The results are summarized in Table 43 through 45. In boarding school, the eighth grade was the most expensive in the first cycle whereas the 12th grade was the most expensive grade in the second cycle. In day school, however, the seventh grade was the most expensive grade in the first cycle and the 11th grade in second cycle. The ninth grade was the cheapest grade of all day and boarding school. School fees were the most expensive item in boarding school, while personal needs item was the most expensive item in day school. On the average, school fees represented 33% of the 142 direct private cost in boarding school and 19% in day school. Personal needs represented 28% in boarding school and 48% in day school. Table 43: Direct Private Costs of Boarding School per Grade Level (GRADNOW), Burundi, 1992-93 School Year. Type of GRADE Cost 7(7) 8 (14) 9 (13) 10(12) 11(7) 12(5) 13(11) 9,000 9,000 9,000 9,000 9,000 9,000 9000 SCFEES 37% 35% 42% 35% 37% 26% 28% 4,940 6,118 6,463 7,605 7,879 9,520 9633 PNEED 20.3% 24% 31% 29% 32% 28% 30% 4,891 4,073 2,492 4,172 3,131 5,321 4677 CLOTH 20.2% 16% 12% 16% 13% 15% 15% 1,989 2,111 1,055 2,643 1,900 1,002 2764 SHOES 20.2% 16% 12% 16% 13% 15% 15% 1,513 845 1,128 1,102 667 1,611 1573 BLANK 6.2% 3% 5% 4% 2.7% 5% 5% 1,968 1,011 878 1,440 943 2,436 1671 BSHEET 8.1% 4% 4% 6% 3.8% 7% 5% 0 2,738 295 0 918 5,670 2560 MISC 0% 10% 1% 0% 3.8% 16% 8% 24,301 25,896 21,311 25,962 24,438 34,560 31878 YCOST 100% 100% 100% 100% 100% 100% 100% 143 Table 44: Direct Private Cost of Day School per Grade Level, Burundi, 1992-93 School Year. Type of GRADE Cost 7 (12) 8 (17) 9 (10) 10 (15) 11 (7) 12 (9) 13 (3) SCFEES 4,227 4,300 3,524 3,923 3,500 4,312 4,000 19.7% 20.7% 20% 19% 14.5% 19% 23% PNEED 9,373 10,120 7,832 10,000 12,167 11,250 7,767 43.6% 48.7% 45% 49% 50.5% 50% 46% CLOTH 4,664 4,460 3,684 4,038 5,250 4,163 2,667 21.7% 21.5% 21% 20% 22% 18% 16% SHOES 3,223 1,887 2,466 2,369 2,583 2,875 2,500 15.0% 9.1% 14% 12% 11% 13% 15% BLANK 0 0 0 0 0 0 0 0% 0% 0% 0% 0% 0% 0% 0 0 0 0 583 0 0 BSHEET 0% 0% 0% 0% 2% 0% 0% 0 0 0 O 0 0 0 MISC 0% 0% 0% 0% 0% 0% 0% 21,487 20,767 17,506 20,331 24,083 22,600 16,934 YCOST 100% 100% 100% 100% 100% 100% 100% Table 45: Direct Private Cost per Cycle of secondary school per Type of school, Burundi, 1992-93 Type Type of school of cost Day School Boarding school first cycle second cycle first cycle second cycle Amount % Amount % Amount % Amount % SCFEES 4,040 20.0 3,971 17.9 9,000 36.9 9,000 30.0 PNEED 9,502 46.9 10,959 49.5 6,420 26.3 9,058 30.0 CLOTH 4,252 21.1 4,282 19.0 3,773 15.4 4,338 14.4 SHOES 2,432 12.0 2,706 12.2 1,928 7.9 2,096 6.9 BLANK 0 0.0 0 0.0 1,095 4.4 1,297 4.3 BSHEET 0 0.0 206 0.9 1,231 5.0 1,615 5.3 MISC 0 0.0 0 0.0 912 3.7 2,746 9.1 YCOST 20,226 100 22,124 100 24,359 100 30,150 100 144 The seventh grade was expected to be the most expensive grade of the first cycle because of necessary expenditures for initial school supplies, such as bedclothes, school uniforms, table supplies, and a suitcase. Analysis of survey results showed that the seventh grade in boarding schools did not seem to be as expensive as was expected. This was because there were no seventh grade students from Bujumbura in boarding school in this study. Also, some of the school equipment used in seventh grade was kept for more than one or two school years. Parents were required to replace only what was worn out, lost, broken, or outgrown (i.e., uniforms and shoes). For this reason, the 9th grade was the cheapest in terms of direct private costs of secondary schooling. The 8th grade was more expensive than the 7th grade, probably due to the increase in the number of courses (three courses were added in the eighth grade). Furthermore, students changed schools in eighth grade, requiring new equipment, such as school uniforms. After the tenth grade, students Changed schools as a result of a national test which was used for tracking purposes. Consequently, the eleventh grade’s direct private costs were higher than those in the first cycle. The last two grades of boarding school were the most expensive. The twelfth grade was found to be unusually expensive for the following three possible explanations: * there were more girls in this grade and the cost of 145 girls' education was found to be more expensive than boys’; * these were the graduation years (12th grade e.g., EFI = Ecole de Formation des Instituteurs; 13th grade: graduation of the regular humanities); * many of the students were from Bujumbura and paid higher than the average cost of other regions. 0 IV.2.9. Direct Private Costs and Number of Children in Secondary School Some parents had more than one Child enrolled in secondary school. Tables 46 and 47 show how direct private costs per child varied with the number of children a family had in secondary school. For boarding school, direct private costs per student increased with the number of students that a family had in secondary school. The average spending on secondary boarding school for parents having only one student in secondary school was 19.6% less than those with two, 26.5% less than those with three, and 40.3% less than those with more than three students. In day school, it cost relatively less to send one Child to secondary school. Families with one child spent 34.6% less than those with two Children; 18% less than those with three; and 37.1% less than those with more than three children in day secondary school. Table 46 DPC of Secondary Boarding School and Number of Children in 146 Secondary School per Family Type Number of Children in Boarding School gist INSEC 1 (23) INSEC 2 (21) INSEC 3 (9) INSEC>3 (17) Amount % Amount % Amount % Amount % SCFEES 9,000 41.0 9,000 34.1 9,000 32.2 9,000 29.3 PNEED 5,280 23.9 7,214 27.4 8,463 30.3 9,468 30.5 CLOTH 3,088 14.0 4,079 15.5 5,159 18.5 4,344 14.0 SHOES 1,265 5.7 2,635 10.0 1,168 4.2 2,609 8.4 BLANK 521 2.3 1,422 5.4 1,861 6.7 1,334 4.3 BSHEET 756 3.4 2,025 7.6 2,256 8.1 867 2.8 MISC 2,143 9.7 0 0.0 0 0.0 3,334 10.7 YCOST 22,053 100 26,375 100 27,908 100 30,957 100 Table 47. DPC of Secondary Day School and Number of Children in Secondary School per Family. Type Number of Children in Day School gist INSEC 1 (3) IESEC 2 INSEC 3 (16) INSEC>3 (44) Amount % Amount % Amount % Amount % SCFEES 4,000 25.2 4,125 19.3 3,823 20.5 4,077 18.6 PNEED 7,400 46.7 9,112 42.7 8,628 46.1 10,705 49.2 CLOTH 3,233 20.4 4,350 20.4 4,314 23.1 4,300 20.0 SHOES 1,217 7.7 3,750 17.6 1,931 10.3 2,562 11.8 BLANK 0 0.0 0 0.0 0 0.0 0 0.0 BSHEET 0 0.0 0 0.0 0 0.0 90 0.4 MISC 0 0.0 0 0.0 0 0.0 0 0.0 YCOST 15,850 100 21,337 100 18,696 100 21,734 100 Only families with more income could afford to have more than one child in secondary school. The analysis of the direct private cost showed that the more income parents had, the more they spent on the schooling of the individual child. Consequently, income played a major role in 147 determining the amount of the direct private cost of schooling and the number of children a family was able to finance. As income increased, not only did parents send more Children to secondary school, but they also spent more per child. IV.2.10. Public versus Private Schools The average direct private costs in private schools were higher than those in public schools mainly because school fees constituted the major source of income for these schools (see Table 26). Direct private costs in private schools were 87% higher than those for public boarding schools and 137% higher than those for public day schools. School fees constituted 40% of the direct private cost of private schools. These were used to pay the teachers and cover all the expenses of running the school. In public schools, the school fees were never used to pay the teachers' salaries. They were used for maintenance of school building and to purchase food for students. Therefore, the school fees of private schools were 394.1% higher than those of public day school, 120.3% higher than those of boarding school. The major items of direct private costs, found in both day and boarding schools, (i.e., personal needs, clothing, shoes), were higher in private school. Personal needs were 148 70% higher than those of public boarding school, 25.3% higher than those of day school. Clothing costs were 89% higher than those in public boarding and 75.6% higher than those in public day schools. Shoes were 63.7% higher than those in public boarding schools and 29.7% higher than those in public day schools. Expenditure on these items generally increased with the income of the family. Thus, analysis of the background of students in private schools shows that their families were wealthy and able to finance private school costs. Indeed, almost all the private students came from advantaged family backgrounds (the 4th and 5th quintile income). These findings, which indicate that direct private costs for private secondary schools were higher than those for public secondary schools, support findings from other developing countries (Tsang and Kidchanapanish 1992; Tsang, 1994). IV.3. Total Family Spending on Secondary Education While section IV.2 presents the direct private cost of secondary education per student, this section computes the total family spending on secondary education. Family spending on secondary schooling increased with the parents’ level of education, the family income, total family expenditures, family assets, and with the number of children 149 in secondary school. The results are presented in Table 48. TYCOST, total family spending on secondary education is in Burundi francs (Fbu). The unit of analysis is the family. IV.3.1. Fathers’ Education (FATHEDUC) The total family spending on secondary education increased with the education of the father. In families where the fathers had no schooling, the total cost of secondary schooling was 23,696 Fbu. The total cost of secondary schooling increased by 51% when fathers had at least a lower primary education, by 271% when fathers had a upper elementary education, by 161% when fathers had at least some secondary education. These increases were related to the number of children enrolled in secondary school in those families. For example, fathers with an upper-primary schooling spent the highest amount (87,961 FBu). This amount was able to pay for at least three Children in secondary school, whereas the fathers with no schooling had barely enough to spend on one student. Fathers with the highest level of education spent relatively less on schooling than the fathers with an upper primary school education because their children were mostly enrolled in day schools. These schools were cheaper than boarding schools. 150 IV.3.2. Mothers’ Education (MOTHEDUC) In families where the mothers had no schooling, the total cost of secondary schooling was 31,498 FBu. The total spending on the cost of secondary schooling increased by 95% when mothers had at least a lower primary education, by 149% where mothers had a upper elementary education, by 112% when mothers had at least some secondary education. Mothers with the highest level of education also spent relatively less than the mothers with upper primary school because their children were also enrolled in day school. Since these families lived near these schools, they were spending less on transportation. This reduced the cost of day secondary schooling on these families. IV.3.3. Fathers' Occupation (FATHEROC) In families where the fathers were farmers, the total spending on secondary schooling was 20,329 FBu. The total spending on secondary schooling increased by 92% where the fathers were general-laborers, by 161% when fathers were qualified technicians, by 295% where the fathers were businessmen, and by 312.5% when fathers were professionals. 151 IV.3.4. Mothers’ Occupation (MOTHEROC) In families where the mothers were farmers, the total spending on secondary schooling was 26,344 FBu. The total spending on secondary schooling increased by 97% when the mothers were general-laborers, by 156% when mothers were qualified technicians, by 172% when mothers were professionals, and by 279% when the mothers were in business. The business women spent the highest amount on schooling because they were wealthier. IV.3.5. Income Quintile (INCOME) The total family spending on secondary schooling systematically increased with the increase in income of the family. The total expenditure for secondary schooling was 20,532 FBu for families in the first income quintile. It increased by 31.6% in the second quintile, by 142% in the third quintile; 250% for the fourth quintile, and by 376% for parents in the fifth quintile. IV.3.6. Assets Quintiles (ASSETS) The total family spending on secondary schooling systematically increased with the increase in assets Of the family. Families in the first assets quintile spent an 152 average of 19,148 FBu as the total cost of secondary schooling. This expenditure increased by 40% for the second quintile, by 99% for the third quintile, by 237% for the fourth quintile, and by 406% for the parents in the fifth quintile. IV.3.7. Expenditure Quintile (EXPENSES) The total family spending on secondary schooling systematically increased with the increase in the total family expenditure of the family. On average, families in the first expenditure quintile spent a total of 19,428 FBu on secondary schooling. This expenditure increased by 61% for the second quintile, by 112% for the third quintile, by 254% for the fourth quintile, and by 408% for the fifth quintile. IV.3.8. Number of Children in Secondary School (INSEC) The total direct private costs of secondary school (TYCOST) increased with the number of children in secondary school. The families with only one child in secondary school spent an average of 22,283 FBu on the total cost of secondary schooling. The total cost of secondary school increased by 135% for families with two children in secondary school. 153 Table 48 Total cost of Secondary School per family (TYCOST). FATHEDUC 1 2 3 4 5 TYCOST 23,696 35,768 87,961 NA 61,769 MOTHEDUC 1 2 3 4 5 TYCOST 31,498 61,538 78,281 NA 66,844 FATHEROC 0 1 2 3 4 TYCOST 20,329 38,993 80,240 46,843 83,871 MOTHEROC 0 1 2 3 4 TYCOST 26,344 51,934 99,865 67,480 71,674 INCOME 1 2 3 4 5 TYCOST 20,532 27,011 49,656 71,839 97,813 ASSET 1 2 3 4 5 TYCOST 19,148 26,858 38,066 64,555 96,932 EXPENSES 1 2 3 4 5 TYCOST 19,428 31,207 41,208 68,811 98,620 INSEC {NSEC INSEC 2 INSEC 3 INSEC >3 TYCOST 22,283 52,466 66,012 113,751 Note: - FATHEDUC and MOTHEDUC: 1: no schooling; 2: Lower Primary; 3: upper primary; 5 - FATHEROC, and MOTHEROC: businessmen/women; 3= qualified Technicians, 4= Professionals - INCOME, ASSET, second quintile; 3 = fifth quintile. quintile; 5 Farmer; at least some secondary 1=general-laborer; 2 and EXPENSES: 1 = first quintile; 2 = third quintile; 4 = fourth - INSEC: INSEC1= the family has only one child in secondary school INSEC2 = the family has two children in secondary school INSEC3 = the family has three Children in secondary school INSEC>3 = the family has more than three Children in secondary school 154 The increase was 196% for families with three children in secondary school, and 410% for families with more than three children in secondary school. IV.4 Total. Family Spending on Education The finding of the total family spending on all levels of education (STUD) are presented in Table 49. The unit of analysis is the family. STUD increased with the increase in parental education, income, assets, expenditure, and the number of children in secondary school per family. It varied with parental occupation. The increase in family spending on education with reference to total family expenditure on secondary schooling was due to the additional Children enrolled in primary school. The total family spending on education (STUD) can be compared with the total family spending on secondary education. As the direct private costs of primary schooling were definitely cheaper than those of secondary school, the total family spending on secondary education was Closer to the total spending on education. For example, the total family spending on education with respect to total family spending on secondary schooling increased by 24.4% for the fathers with no schooling, by 14% for the fathers with a lower primary education, by 4.3% for fathers with an upper primary schooling, and 14% for fathers with at least some 155 secondary education. The highest increase was for fathers with no schooling, probably because they felt the burden of primary schooling more than the other categories. With respect to income, total spending on education, when compared to the total spending on secondary schooling, increased by 22% for the first quintile, by 5% for the second quintile, by 6% for the third quintile, by 10.2% for the fourth quintile, and by 7% for the fifth quintile. However, these findings on the total family spending on education must be taken with caution as the study focused on the direct private costs of secondary schooling. The direct private costs of primary schooling were not collected for individual children. Therefore, the targeted families might have had fewer children in primary school and may also have concentrated on the direct private cost of secondary school because they are more burdensome. 156 Tables 49 Total Family Spending on Education (STUD), Burundi, 1992-93 FATHEDUC 1 2 3 4 S STUD 28,104 40,810 91,778 NA 70,885 MOTHEDUC 1 2 3 4 5 STUD 35,067 64,229 85,211 NA 75,574 FATHEROC 0 l 2 3 4 STUD 24,814 42,253 83,358 60,077 91,835 MOTHEROC 0 1 2 3 4 STUD 29,048 62,798 99,135 75,976 81,110 INCOME 1 2 3 4 5 STUD 25,016 28,457 52,919 79,197 104,665 ASSET 1 2 3 4 5 STUD 22,280 29,543 40,553 74,247 101,402 EXPENSES 1 2 3 4 5 STUD 20,960 30,589 47,819 75,520 107,076 INSEC INSEC l INSEC 2 INSEC 3 INSEC >3 STUD 26,271 53,940 74,528 122,067 Note: - FATHEDUC and MOTHEDUC: 1= no schooling; 2: Lower Primary; 3= upper primary; 5 = at least some secondary. - FATHEROC, and MOTHEROC: 0 = Farmer; 1=general-1aborer; 2 businessmen/women; 3: qualified Technicians, 4= Professionals. - INCOME, ASSET, and EXPENSES: 1 = first quintile; 2 = second quintile; 3 = third quintile; 4 = fourth quintile; 5 = fifth quintile. - INSEC: INSEC1= the family has only one Child in secondary school INSEC2 = the family has two children in secondary school INSEC3 = the family has three children in secondary school INSEC>3 = the family has more than three children in secondary school 157 IV.5. Economic Burden of Direct Private Cost of Public Secondary Schooling on the Family The burden of the direct private cost of public secondary schooling was computed using two measures. The first is computed with respect to the income (YCBURD = YCOST /INCOME *100). The second is computed with respect to the family expenditure (YCEXP =YCOST/EXPENSE*100). The average burden of the direct private cost of secondary education on the family income (YCBURD) and on expenses (YCEXP) are summarized in Table 50 below. The unit of analysis is the family. IV.5.1. Burden of DPC per Child Overall, the average direct private cost of secondary schooling per Child (YCOST) in Burundi represented 9.67% of total family income (YCBURD) in boarding school, and 1.53% of total family income (YCBURD) in day school. It constituted 9.51% of family expenditure (YCEXP) in boarding school, and 2.96% of family expenditure (YCEXP) in day schools. The burden was higher for boarding school than for day schools because the families with children in boarding schools are poorer than those with Children in day schools. Parents with more education, higher paying occupations (professionals), wealthier (third through fifth quintiles 158 income, assets) had Children enrolled in day schools. Therefore, the burden of direct private costs in day schools is relatively much lower and is subject to less variation. The burden of direct private cost with respect to income (YCBURD) was sightly higher than the burden with respect to expenditure (YCEXP) for families in boarding school because they spent relatively more than their income. Thus, most of these families (mostly rural ones) were in debt at the end of the year. The burden of direct private cost with respect to income (YCBURD) was lower than the burden with respect to expenditure (YCEXP) for families in day school because they spent relatively less than their income. Thus, these families (mostly urban ones) were able to have savings at the end of the year. In boarding school, the burden for male students was higher than that for female because most female students were from wealthier families whereas male students came from all the income categories, especially lower income families. It represented 10.9% of income for males, which was above the average, and 8.7% for female, which was below average. They were about the same in YCEXP. In day schools, the burden was about equal for both male and female students because they came from wealthy families. The burden of the direct private cost of secondary schooling to the parents generally decreased as income, assets, expenditure, and parental education increased. In 159 boarding schools, the heaviest burden was borne by families in the first income quintile (23.21%), the first expenditure quintile (20.47%), and the first assets quintile (18.02%); and by families whose fathers (17.13%) and mothers (14.99%) did not have any formal schooling, and families whose parents were farmers. The same pattern could be observed for secondary day schools. The heaviest burden was borne by fathers with less education (1.95% for fathers with lower primary schooling), less income (7.33% for the second income quintile), less assets (3.55% for the third asset quintile), and less expenditure (2% for the third expenditure quintile). The burden of secondary schooling was higher for students in lower secondary education than in upper secondary education. 160 Table 50. Burden of the Direct Private Cost of public Secondary Schooling per Child (DPC) Independent Type of Public Secondary School Variables Boarding School Day School YCBURD YCEXP YCBURD YCEXP All 9.67 9.51 1.53 2.91 Female 8.68 9.43 1.46 2.98 Male 10.90 9.61 1.62 2.82 FATHEDUC 17.13 14.38 NA NA 8.82 9.99 1.95 5.73 5.56 1.54 .14 .98 9.11 8.43 1.07 13.99 11.22 4 2 2.05 MOTHEDUC .30 1.97 3 2 2 8.77 10.50 .00 .61 3.07 4.20 .31 .48 H +4 R) H NA NA .44 .93 MOTHEROC 14.87 12.67 NA NA 9.76 9.66 3.75 5.6 4.93 8.41 0.75 2.09 1 2 3 4 l 2 3 4 0 1 2 3 1.60 2.77 1.27 1.95 4 0 1 2 3 4 1 2 3 4 5 2.85 4.57 1.41 2.88 FATHEROC 19.98 16.07 NA NA 9.17 9.36 1.80 3.63 6.35 6.36 1.27 2.20 5.43 9.13 0.75 2.50 2.75 4.93 1.32 2.72 Income 23.21 17.94 NA NA 12.70 11.18 7.33 9.07 5.53 7.01 3.26 4.62 2.53 5.42 1.89 3.65 1.18 4.22 0.80 1.87 Table 50 (cont’d) Asset 1 U‘IAUN Expense 1 GRADNOW \DGJQUTBUN 10 11 12 13 7-10 11-13 Note: GRADNOW a 18. .83 16 11. 4. 1. 20. 15. 6. 2. 2. 15 12 12. .64 9. 4. 3. 11. 5. Grade level 7 02 55 43 42 47 52 35 84 38 .41 .54 36 43 13 65 66 57 13 15. 10. 17. 12 14. 11. .60 .22 .23 .71 10 161 .40 98 55 .18 .05 98 .61 .72 .33 .06 70 22 .01 .80 .87 NA NA 3.55 1.48 1.10 2 Z V V .00 .86 .91 .91 .60 .30 .44 .24 .56 .35 .58 .41 H H H H H H H H H o H N NA NA 5.29 3.19 2.10 5% .39 .37 .73 .47 .11 .28 .49 .47 .50 .77 .88 .00 WNNWNNNWWHWUI 162 IV.S.2. Burden of Total DPC for Public Secondary School per Family The burden of total family spending for public secondary schooling as a percentage of the total family income (TSECBURD) and as a percentage of the total family expenditure (TSECEXP) are presented in Table 51. The burden of secondary schooling per family was mostly borne by the families in the first and second quintiles of income, expenses, assets. The highest burden was mainly borne by families who had only one child in secondary school. Likewise, the highest burden was borne by parents without formal schooling, and farmers. IV.S.3. Burden of Total Family Spending on Education The burden of total family spending on education as a percentage of the total family income (BURDED) and as a percentage of the total family expenditure (EDUCEXP) are presented in Table 51 above. The burden of the total family spending on education was highly related to the total burden of secondary education on the family. Again, these findings may be influenced by the overall focus of the study, which is the direct private costs of and access to secondary education. It was mostly borne by families in the first and second quintiles of income, expenses, and assets. The 163 Tables 51 Total burden of secondary school (TSECBURD, TSECEXP) and Total Burden of Education (BURDED, EDUCEXP) per family INDEPENDENT TSECBURD TSECEXP BURDED EDUCEXP VARIABLE ALL 11.79 13.05 13.34 14.21 DAY 4.82 8.32 5.49 9.43 BOARD 14.63 14.97 16.52 16.15 FATHEDUC l 17.65 15.40 21.83 18.32 2 09.93 12.22 11.03 13.36 3 8.07 12.24 8.28 12.53 4 9.69 11.40 10.32 11.98 MOTHEDUC 1 15.25 12.67 17.48 14.11 2 12.43 15.96 13.87 16.87 3 6.03 10.20 6.72 11.24 4 6.08 10.82 6.32 11.68 FATHEROC 1 21.66 17.25 26.68 20.29 2 7.98 10.70 6.72 10.13 3 10.19 11.42 11.15 12.09 4 6.36 12.23 10.17 15.94 5 5.61 11.16 6.17 12.25 MOTHEROC 1 18.33 15.77 21.54 17.42 2 9.62 12.88 11.17 15.70 3 6.78 13.81 5.82 12.85 4 3.26 5.39 4.03 6.67 5 5.49 10.21 6.29 11.57 INCOME 1 22.91 17.65 28.19 20.61 2 14.85 13.38 15.43 13.73 3 9.85 13.17 9.83 12.97 4 5.81 11.65 6.41 13.26 5 3.26 8.93 3.57 9.59 164 Table 51 (cont’d) EXPENSES 1 21.21 18.53 23.19 20.03 2 17.81 15.18 20.35 15.67 3 8.73 13.88 10.11 15.63 4 6.41 11.21 7.09 12.42 5 4.31 7.48 4.89 8.17 ASSET 1 16.57 13.45 18.89 15.11 2 20.41 19.24 22.30 21.15 3 14.33 14.04 16.69 14.76 4 7.08 12.57 7.73 14.10 5 4.37 8.75 4.61 9.12 INSEC 1 15.25 13.35 18.19 15.30 2 10.94 12.65 10.67 12.34 3 4.92 9.19 5.51 10.35 >3 6.75 14.09 7.49 15.11 165 highest burden was mainly borne by families who had only one Child in secondary school. Likewise, the highest burden was borne by parents without formal schooling, and farmers. Hence, families with more children in secondary school were able to finance them and still bear a lower burden than families with only one child. IV.6. Correlation Among Variables For those in secondary boarding schools, the results are shown in Table 52. All the variables about direct private cost of education (YCOST, TYCOST and STUD) were highly and positively correlated with one another at .001 level of significance. They were also significantly and negatively correlated with the variables of burden (YCBURD, YCEXP); YCOST was negatively but not significantly correlated with YCEXP. The variables of family wealth (INCOME, ASSET and EXPENSES) were highly and positively correlated with the variables related to direct private cost of secondary boarding schools. Family wealth variables, however, were highly but negatively correlated with the variables related to the burden of education. The variables about parents education (FATHEDUC and MOTHEDUC) were not Correlated with neither the variables about direct private cost nor those about the burden of education (YCOST, TYCOST, YCBURD and YCEXP). 166 For secondary day school, the results are shown in Table 53. All the variables about direct private cost of secondary education (YCOST), the total cost of education per family (TYCOST) and the total cost of education (STUD) for secondary day schools were highly and positively correlated with one another at .001 level of significance. On the contrary, variables related to family background (FATHEDUC, MOTHEDUC, INCOME, ASSET, and EXPENSES) were not significantly correlated with the variables about the direct private costs. The variables about family wealth, however, were highly and negatively correlated with the variables of the burden (YCBURD, YCEXP). The burden variables were not correlated among themselves. For both types of secondary schools, the variables about family wealth were negatively correlated with the variables related to the burden of education. The wealthier the family the lower their burden of secondary education. In other words, the more family income, family expenditure, and family assets, the lower the burden of education on the family. 167 Hows 4 m¢ umwmmv HO 2 “0902 oooo.H 4Hm4m. «mavm. 4momv. pmmo.- .Hmmm.- 4omms.- 4>mmm. 4s>om. «same. mmmzmmxw oooo.H 4mm4m. 4oam4. maeo.- «omam.- 4o4os.- «vene. .mmmn. 4mvmm. swans oooo.H 44mH4. mmao.- 44Hm>.- 4mvmm.- 46mph. 44nmm. 44Hom. msoozH oooo.a momo.- maua.- mmmm.- mmvm. Hmmm. mmam. Oonmmeoz oooo.a 4mmo.- pamo.- 4mmo.- oopo.- oawo. Ooommssm oooo.H 4mm4m. 4mmHm.- 4mmmm.- m¢mm.- mxmo» coco.“ 44mmm.- 4mmmm.- «ammm.- omomuw oooo.H «ommm. 466mm. spam oooo.H «muse. smoosa oooo.H amour wmmzmzxm .53 5.502. Donn—Fox gem—HE}. me0> 959.; 03.5 .Goot. .500» 3.555; Hoosum mswpumom humpsouum no mudguaum> onu m:o&4 msoflumauuuoo mm OHQMH 168 HO. A9 4 0H ummeU MO 2 “0002 oooo.H «mmmn. «Ndmm. wmmm. ibmom. amflbm.1 immmm.1 HHmv. mmmv. momm. mmmzmmxm oooo.H ammmm. mmme. anhv. %N¢bm.1 Naom.1 mmon. mHmN. mmmN. Bflmfld oooo.H moon. «mamm. «mmom.1 «mmmm.1 mmmm. mmmfi. Hmmm. mZOUZH oooo.H immmh. momm.1 Obbm.1 Hmmm. mhmH. mNoo. 000M380: oooo.H immem.1 comwm.1 NHNM. vaN. vmmo.1 UDQflZBdh oooo.H ammmm. Hmwd.1 MFOH.1 mood. QXMUM oooo.H ¢hmm.1 ommm.1 mo¢o. DNDNUN oooo.H immbm. tmdbh. DDBW oooo.H imbmm. BmOUNB oooo.H HmOu» mmmZm—LXM Fm? ”1002— UDDm—FFOZ saw—:2". mXMUr DMDGU> GE :00: 50> gg:> Hoonum >09 Humocoumm mo moanmfium> on» msoe< mGOAumeuHOU mm manna 169 IV.7. Summary of the Chapter Secondary boarding schools were more expensive than secondary day schools. The direct private costs of secondary schooling were higher for boarding school than for day school. These costs varied by region, parents' education, parents’ occupation, family wealth, the number of Children in secondary school, as well as gender of the student. The direct private costs increased with the income, asset, and education of the parents. Urban families, which included more educated and wealthy parents, spent more than rural families on individual children in boarding school. They also were able to choose to pay for either day or boarding schooling because secondary day schools were available only in urban areas. Most urban families sent children to secondary day schools. Poor families, mostly in rural areas, sent their Children mainly to secondary boarding schools. Direct private costs increased with an increase level of education of both parents. It also increased with higher paying occupations for either parent, and family wealth. Direct private costs were highest in the highest income and expenditure categories. Poor families, mostly farmers, enrolled a very limited number of Children in secondary schools because of limited income. 170 Direct private cost of female students in both secondary boarding and day schools was higher than that of male students. In addition to the required supplies for every student, female students required more expenditures on specificly female items. The burden of the direct private cost decreased with increased family wealth and parents’ level of education. The burden of day secondary school was lower than that of boarding schools. Despite the higher direct private cost for female students, its burden was lower than that of male students. This is an indication that female students came from relatively wealthier families and/or educated parents. Furthermore, poor families, especially farmers, whose children represented 11.5% of all the children financed by their parents, bore the heaviest burden (20%) of secondary schooling. The total direct private costs of education on the family increased with the number of Children in secondary school, while the number of children in secondary school increased with family wealth and parents education. Despite that, poor families with only one child in secondary school bore the highest burden of the total direct private costs of schooling. CHAPTER V. DETERMINANTS OF ACCESS TO SECONDARY SCHOOLING IN BURUNDI V.1. Overview This Chapter attempts to answer the following questions: * Which families had access to secondary school in Burundi in 1993? * What factors affect the economic burden of the direct private cost of secondary schooling in Burundi? * What factors affect access to secondary schooling in Burundi? To answer these questions, the Chapter begins with a profile of children still enrolled in secondary school in 1993. These Children are made up of those helped by extended family members and those fully financed by their parents. The characteristics of the helpers in the extended family are described. The chapter then presents a mean- comparison of independent variables to determine if families without children in secondary school in 1993 (GROUP1) and families with children in secondary school in the same year (GROUP2) were significantly different. A linear multiple regression analysis used to identify the determinants of the economic burden of the direct private cost of secondary schooling on the family is subsequently provided. Finally, 171 172 this chapter presents the results of a logistic regression analysis used to determine the factors predicting access to secondary schooling. This analysis is conducted at both the family and the child levels. V.2. Characteristics of Children in Secondary School Not all children attending secondary school were financed by their parents. Seventy four percent of all Children in secondary school were financed by their parents. Twenty six percent of them were financed by relatives and/or friends. As such, it was important to look at the characteristics of the helpers (those who helped in financing the Children in secondary school), and the role of relatives in motivating children to go to secondary schooling. In addition, it was necessary to present the characteristics of children who were financially supported by their parents and those who were not prior to analyzing the characteristics of families. V.2.1. Characteristics of Helped Children In 1993, children who were helped by relatives and friends had the following characteristics: * 44% of the children were male and 56% of them were 173 female * 91% of the children never repeated a lower grade of primary school (grades 1-3). * 78% of the children repeated upper grades of primary school (4-6). * 93% of the children repeated the sixth grade. Only seven percent acceded to secondary school without repeating the sixth grade. * 61% of the Children repeated sixth grade two or three times. * 65% of the children did not repeat a grade in secondary school. * 35% of the Children repeated a grade in secondary school. * 61% of helped Children lived in Bujumbura, the capital City. Moreover, there was a positive correlation between the opportunity to go to secondary school and the proximity to urban areas. In Karuzi, which was a rural province, children dropped out of primary school after the first repetition in sixth grade. In the rural Gitega area, which was next to the urban center, most children repeated the sixth grade three times. Children in Bujumbura also repeated the sixth grade three times. Among the children financed by the extended family or friends, very few passed the Concours National before they 174 repeated the sixth grade. None of these children passed it in Karuzi. Only 6% of them did in Bujumbura and 13% of them did in Gitega. V.2.2. Characteristics of Helpers Direct relatives (e.g., brothers, sisters, uncles, aunts and cousins) from both the mother’s and the father’s sides represented 96% of the people who financed secondary school Children in the extended family. The remaining 4% of helpers were friends of the family. In addition, about 74% of the relatives who financially helped the children had at least a secondary school education. Furthermore, 84 % of the helpers lived in urban areas, while 16% of them lived in rural areas. Nearly 72% of the helpers paid the full cost of secondary schooling of the children they supported. V.2.3. Characteristics of Children fully Supported by Parents In 1993, the Children who were fully supported by their parents, had the following characteristics: * 40 % of the children were male and 60 % were female. * 93% of the children never repeated in lower primary grades. 175 * 78% of the children repeated in higher primary grade. * 81% of them repeated sixth grade. * Most children wrote the Concours National at least three times in both rural and urban areas. * Very few Children passed the Concours National the first time, (i.e., 4% in Karuzi, 17% in Gitega, and 24% in Bujumbura). Among the children who were supported by their parents, 15% changed schools. Among the 15% who changed schools, 46% went to other rural schools and 54% went to urban schools. V.3. Comparing Characteristics of Families With and Without Children in Secondary School V.3.1. Matrix of Correlation In this section, the family is the unit of analysis. A correlation matrix was established to indicate the relationship between the variables. These correlations were obtained from weighted data. The following relationships were Observed (at the .05 level of significance): there was a high91and positive correlation between having a child in secondary school (SECOSCHL) and family assets (ASSET), and 1 There is high correlation if the value of the coefficient of correlation is at least 0.5 (see Table 54). 176 the father’s level of education (FATHEDUC), and amoderate correlation with the mothers’ level of education (MOTHEDUC). There was a positive correlation between having Children in secondary school (SECOSCHL) and the urban origin (AREAS). Having children in secondary school (SECOSCHL) and the predicted burden (PREDBURD) were moderately but negatively correlated. Apart from income (INCOME), presence of educated member in the extended family (FAMILED), highest number of times a child repeated in the sixth grade (TIME6) children' s Chores (CHORES), predicted values of the burden (PREDBURD), highly and negatively correlated with all other variables. The predicted burden of secondary schooling (PREDBURD) was highly but negatively correlated with family assets (ASSET), level Of education of the father (FATHEDUC), level of eucation of the mother (MOTHEDUC), and urban origin (AREAS); it was moderately but negatively correlated with family expenditure (EXPENSES), family income (INCOME), and access to secondary schooling (SECOSCHL). FATHEDUC and MOTHEDUC (parents’ education) were highly and positively correlated with family assets (ASSET), and urban origin of the family (AREAS). Family assets (ASSET) was positively correlated with income, fathers’ education, mothers' education, family expenditure; it was highly but negatively correlated with PREDBURD; moderately correlated with access to secondary schooling (SECOSCHL), and urban/rural origin (AREAS). TIME6 was moderately and positively correlated 177 with FAMILED. CHORES was weakly, significantly but negatively correlated with FATHEDUC AREAS. EXPENSES was highly correlated with family income (INCOME), and family assets (ASSET); it was moderately but significantly correlated with fathers' level of education (FATHEDUC), and mothers’ level of education (MOTHEDUC). The correlations between variables are presented in Table 54 below. The rest of this section shows the results of the comparison of the characteristics of families without children in secondary school and those with children in secondary school. All the independent variables used in this Chapter were included in the comparison based on the t- test. The level of statistical significance was set at .05. V.3.2. Parents’ Education V.3.2.1. MOTHEDUC The two groups of families were found to be significantly different in terms of mothers’ level of education. The pooled T- value was -5.59. Therefore, mothers who had children in secondary school had more formal schooling than those without children in secondary school. 178 coco; SON. .38.- ca? gomf oak? 85.4.1 Emf 30¢: 2.3.. 3:. not. cacao? 9.39m: 88 .H 3.8. .063. ¢~m~. 1 .39.... 1 nose. 1 m-p . 1 N¢¢H . 1 mono. 0mg. 33 . M¢~_. . 1 53%: coco... o¢_.~. 9.18.1 93o. 3n... Hoop. NOON. mg... H03. 33.. 83. out: coco... ..o~n. ~msn. wmmm. 03¢. 0¢o~ . open. 0.8.. . 1 0%.... . 1 58¢. Jaumouum coco. H 08¢. Owen. 000m. -m~ . mm~_. . h¢m~ . 1 H~o~1 1 02.3. m8 coco. H mem< 958me rumwm: 8.2:. .58on m PREDBURD . MOTHOCi MOTHOCZ MOTHOCS 193 which indicated that the more sixth grade repetition, the greater the access to secondary schooling. Therefore, the likelihood Of accessing secondary school for families from the rural area, was determined by the predicted burden of the direct private cost of secondary school, which involved mothers' education (MOTHEDUC), father’s occupation (FATHOCl, FATHOCZ, FATHOC3, FATHOC4), and mothers' occupation (MOTHOCl, MOTHOC2, and MOTHOC3), and the highest number of sixth grade repetition in the family (TIME6). Access to secondary schooling in the urban area is presented in the following Figure 5.3. V.4.4. Logistic Regressions: Child As Unit of Analysis GRADFIN is the highest grade level a child had finished before leaving school, and GRADNOW gives the grade level in which the Child was enrolled at the time of the survey. The mean grade level for children still in school was 9.63 and the standard deviation was 1.83 (N = 216). The mean finished grade level for the children who had left school was 6.52 and the standard deviation was 3.03 (N = 300). Based on the information from GRADFIN and GRADNOW, the access to school variable, ACCESS, was defined. For ACCESS, a value of 1 was assigned to children who had finished, or were enrolled in, the seventh grade; and a value of 0 was 194 assigned to children who left school before reaching grade seven. ACCESS was used as a dependent variable in a logistic regression analysis. The results of the analysis are presented for the entire sample, and for the urban and rural subsamples. A stepwise variable selection was used to add or omit variables which had a Wald-value smaller than 2 and which were not statistically significant at .05 level of significance. The final equation includes only statistically significant variables. Except for the variable, BOY defined as the family preference to send a boy into secondary school, which was statistically significant only for the rural sub-sample, the cultural values about secondary schooling were not significant in this analysis. V.4.4.1. Entire Sample The statistically significant variables in the final equation were the family assets (ASSET), the existence of educated relatives in the extended family (FAMILED), the level of education of the father (FATHEDUC), Children chores (CHORES), the number of times a child repeated sixth grade (TIME6), whether or not the child repeated in the last three grades of primary school (REPHIGH), and whether or not the child repeated in the lower three grades of primary school (REPLOW). The results of the logistic regression are 195 summarized in Table 64 below. Table 61. Equation for Access to Secondary Schooling (Entire Sample and Child as Unit of Analysis) Variable B S.E. Wald Siggi R ASSET .1421 .0271 27.4786 .0000 .1912 CHORES - 6434 .2788 5.3258 .0210 -.0691 FAMILED .8279 .2682 9.5309 .0020 .1040 FATHEDUC .4622 .0950 23.6950 .0000 .1765 REPHIGH 1.0336 .3104 11.0844 .0009 .1142 REPLOW -1.4272 .5403 6.9772 .0083 -.0845 TIME6 .5764 .1380 17.4418 .0000 .1489 Constant -3 7315 .4421 71.2471 .0000 The variables ASSET, FAMILED, FATHEDUC, REPHIGH, and TIME6 had a positive sign. This sign indicated that increased family assets, increased presence of educated relative in the extended family, higher level of education of the father, higher repetition of the upper grades of primary school, higher number of sixth grade repetition, increased access to secondary schooling. The variables REPLOW and CHORES had a negative sign, which indicated that more repeating the lower three grades of primary school, and having to do more Chores at home before or after school decreased access to secondary schooling. Access to secondary schooling in the entire sample using the child as the unit of analysis is presented in the following Figure 5.4. 196 Figure 5.4 Determinants of access to secondary schooling in Burundi, (Entire sample Child as Unit of Analysis) \ V.4.4.2. Urban Subsample 197 The statistically significant variables in the final equation were the existence of educated relatives in the extended family (FAMILED), level of education of the father (FATHEDUC), education of the mother (MOTHEDUC), Child repeated sixth grade (TIME6), a child repeated fifth grade (TIMES). logistic regression are summarized in Table 65 below. the family assets (ASSET), the the level of the number of times a and the number of times The results of the Table 62. Equation for Access to Secondary Schooling (Urban Subsample and Child as Unit of Analysis) Variable B S.E. Wald Sig R MOTHEDUC .4892 .1885 6.7346 .0095 .1171 ASSET .1123 .0324 12.0312 .0005 .1704 FATHEDUC .4639 .1479 9.8345 .0017 .1506 FAMILED 2.1720 .4408 24.2804 .0000 .2540 TIME6 .8097 .2279 12.6242 .0004 .1754 TIMES -.9898 .3414 8.4068 .0037 -.1362 Constant -5.3336 .7831 46.3925 .0000 The variables MOTHEDUC, ASSET, FATHEDUC, FAMILED, TIME6 had a positive sign. level of education of the mother, more family assets, and This sign indicated that higher level of education of the father, higher increased presence of educated relative in the extended family, higher number of 198 Figure 5.5 Determinants of access to secondary schooling in Burundi. Urban Sub-sample and Child as Unit of Analysis MOTHEDUC ASSET \ ( W 199 sixth grade repetition, increased access to secondary schooling. The variable TIMES had a negative sign, which indicated that increase in the number of fifth grade repetition decreased access to secondary schooling. Figure 5.5. presents the determinants of access to secondary schooling at the urban subsample. V.4.4.3. Rural Subsample The statistically significant variables in the final equation were the family income (INCOME), the number of times a child repeated sixth grade (TIME6), the children's chores (CHORES), whether or not the child repeated in the last three grades of primary school (REPHIGH), and whether or not the family preferred to send a boy to secondary schooling (BOY). The results of the logistic regression are summarized in Table 66 below. Table 63 Equation for Access to Secondary Schooling (Rural Subsample and Child as Unit of Analysis) Variable B S.E. Wald Sig R INCOME 9.23E-06 1.794E-06 26.4668 .0000 .2850 TIMES .5970 .1783 11.2127 .0008 .1749 CHORES -1.4100 .5531 6.4992 .0108 -.1222 REPHIGH 1.1250 .4328 6.7586 .0093 .1257 BOY -.9251 .3912 5.5930 .0180 -.1092 Constant -1.4216 .6182 5.2889 .0215 200 Figure 5.6 Determinants of access to secondary schooling in Burundi, Rural Sub-sample and Child as Unit Of Analysis / 201 The sign of the variables CHORES and BOY was negative which indicated that doing Chores before or after school, as well as the family’s preference to send a boy to secondary school decreased the Children access to secondary schooling. The sign of the other variables was positive which indicated that access to secondary school would increase with higher family income, more repetition in sixth grade, and more repetition in the last three grades of primary school. Figure 5.6. presents the determinants of access to secondary schooling at the rural subsample. V.4.5. Discussion of Findings on Access to Secondary Schooling The most important factors in determining access to secondary schooling in Burundi, in 1993, were the level of the education of the father, the level of education of the mother, family assets, family income, the predicted burden of secondary schooling, and the number of times of sixth grade repetition. Among these factors, the level of education of the father was consistently significant in predicting access to secondary schooling when the family was used as the unit of analysis, and when the child was used as the unit of analysis, except in the rural area. However, income, which was statistically significant in the rural area, was highly correlated with the level Of education of 202 the father. The findings of this Chapter showed the role of parents' education in predicting access to secondary schooling. The level of education of the father was highly correlated with family income, family assets, family expenditure, and the level of education of the mother. In general, significant factors in determining access to secondary schooling in the urban area dominated in predicting access to secondary schooling in the entire sample because the urban area provided more opportunities to schooling. Thus, family assets and level of education of the father were statistically significant in both the entire sample and the urban area, and in both family and child as unit of analysis. The cultural factors (preference to send a boy to secondary school), direct private costs (predicted burden), and indirect private costs of primary schooling (Children's Chores) inhibited access to secondary schooling only to children from the rural area. In the rural area subsample, the predicted burden of secondary schooling, the number of times of repeating the sixth grade (family as unit of analysis), as well as the parents’ preference to send a boy to secondary school (child as the unit of analysis), which were not significant in the urban subsample nor in the entire sample, were the only determinants of access to secondary schooling. Increase in the fathers' education and the mothers' education and the 203 parents' occupation reduced the predicted burden, and therefore, increased access to schooling. Rural farmers, with less income and without any level of education, had limited access to secondary schooling. All the other occupations had more advantaged in acceding secondary schooling. When the family was the unit of analysis, the nuclear family background factors (father' education, and mothers’ education) were the statistically significant determinants of access either directly (entire sample, and urban area), or indirectly (through the predicted burden of secondary schooling in the rural area). In addition to family background determinants, when the Child was the unit of analysis, sixth grade repetition significantly increased access to secondary schooling; the presence of educated relatives increased access to children at the entire sample and in the urban area. Increased level of education of the parents would counteract many of the problem created by the parents' cultural values of education, indirect private costs of schooling, and limited income. CHAPTER VI . CONCLUSION This chapter presents a summary of the findings on the direct private cost of, and access to, secondary schooling in Burundi. The findings were presented using both the family and the child as the unit of analysis. They were also presented at the entire sample level, as well as at the urban and rural sub-sample levels. These findings include the financing of secondary schooling, the economic burden of the direct private cost of secondary schooling, predicted burden and the determinants of access to secondary schooling in 1993. The chapter also suggests policies to improve access to secondary schooling. Finally, further research needs are recommended. VI.1. Summary of Findings VI.1.1. Direct Private Costs and Economic Burden The estimate of the direct private cost of secondary schooling in Burundi was based on information on children fully supported by their parents. The annual direct private cost of secondary schooling per child in Burundi varied according to the type of secondary school, gender of the student, family background, and location of families. The total direct private cost of secondary schooling was 26,256 204 205 FBu for secondary boarding school and 20,725 FBu in secondary day school. The main items for both boarding and day secondary schools were school fees and personal needs, which included transportation, school supplies and personal care. School fees constituted 33% of the total direct private cost in boarding secondary school, whereas they were 19.4% in day secondary school in 1993. School fees have increased rapidly overtime. For example, they increased by 800% for secondary boarding schools during the period of 1973 to 1993, and by 350% for secondary day schools during 1983 to 1993. During the period of 1982-1993, the burden of school fees per child with respect to family income went from 7.0% to 30% for boarding secondary school and from 3.5% to 10% for day secondary school. Personal needs in secondary day school, especially transportation, accounted for 48% of the total direct private cost, while they represented 28% in secondary boarding school. Shoes and Cloths, including school uniforms, were respectively 32.7% and 22.6% of the total direct private cost in day and boarding secondary schools. Bedding items and miscellaneous expenses (i.e., pocket money) accounted for an additional 15.6% of the total direct private cost in boarding secondary school. The total direct private cost was higher for female students than for male students in both day and boarding 206 schools. In secondary day school, female students spent 6.9% more than male students. In secondary boarding school, female students spent 6.3% more than male students. This suggested that female students came mainly from wealthier and more educated parents, living especially in the urban areas. Male students came from all family backgrounds. In addition, there was a positive correlation between the education of the parents and the direct private cost of secondary schooling in Burundi. The more educated the parents were, the more was spent on secondary schooling. Furthermore, the less educated the parents (especially fathers) were, the more likely it was that secondary boarding school became the only alternative available for the children. On the contrary, the more educated the mother was, the more likely Children would attend day secondary school because most educated mothers lived in the urban areas. Parents’ occupation also positively influenced the direct private cost of secondary schooling in Burundi. The more high paying the occupation of the parents, the more parents spent on secondary schooling. Professional and business parents spent more than technicians, general labor and farmers. The possession of assets, and total family income, were the resources used to pay for secondary schooling in Burundi. The more assets a family had, the more money it 207 spent on secondary schooling. As such, the more income a family had, the more Children could be sent to secondary schooling. It cost 135% more to send a second Child to secondary school, and 196% more to send a third child to secondary school. Almost all low income families had only one child in secondary. The main source of income was agricultural production in the rural provinces (i.e., Gitega and Karuzi), whereas salaries, business and rent were the major sources of income in the urban areas (i.e., Bujumbura). As income and assets increased, the burden of the direct private costs decreased. The economic burden of the direct private cost of seCondary schooling represented about 10 % of family expenditures per child in boarding secondary school, where most children of low income families went. This burden constituted three percent of family expenditures per child in day secondary school, which was attended mainly by children from wealthy urban families. Hence, the greatest burden was borne by low income families and by those who had parents, especially fathers, with no formal schooling and farmers. The economic burden of the lowest quintile income families was 23% of the total family income for Children in boarding school. It was almost one percent of the total family income for the highest quintile income. Farmer fathers and mothers and fathers with no formal schooling did 208 not have Children in day secondary schools. The wealthier the family was, the lower the burden of secondary schooling borne by parents. Consequently, low income families had fewer children in secondary school. VI.1.2. Predicted Economic Burden The economic burden of secondary schooling was predicted by the fathers’ occupation, the mothers' occupation, and the level of education of the mother. The predicted burden for all the families in the sample was computed. The results of the predicted burden were included in the logistic regression analysis of determinants of access to secondary schooling using both the family as the unit of analysis and the child as the unit of analysis. VI.1.3. Access to Secondary Education Analysis of access to secondary schooling included all the children of secondary school age, a group comprised of children who were still attending secondary schools and those who were not in school. There were 528 children of secondary school age group distributed as follows: 210 children were enrolled in secondary schools and 318 children were not in school. Among the category of children who were not in school, 52 children had access to secondary school. 209 Most of the Children in secondary school (74%) were financially supported by their parents. The remaining 26% were financed by relatives and friends. The following were the characteristics of children helped by relatives and friends: (1) 93% repeated the sixth grade; (2) 65% of these children never repeated a grade in secondary school; (3) and 61% of them lived in Bujumbura. Relatives and friends who helped these children lived in urban areas, and were educated and wealthy. Children fully supported by their parents presented the following Characteristics: (1) 83% repeated the sixth grade and (2) 15% of them had to Change schools to go to study in urban areas . VI.1.3.1. Determinants of Access To Secondary Schooling When the family was used as the unit of analysis, access to secondary schooling in Burundi in the entire sample was determined by the family assets and the level of education of the father. Access would increase with increased level of education of the father, and increased family assets. In the urban area, access to secondary school was determined by the family assets and the level of education of the father. Access would increase with increased family 210 assets and increased level of education of the father. In the rural area, access to secondary school was determined directly by the predicted burden and the highest number Of children sixth grade repetition, and indirectly by the fathers' occupation, the level of education of the mother, and the mothers’ occupation which were used to compute the predicted burden of secondary schooling. Access would decrease with increased burden and it would increase with increased number of sixth grade repetition. When the Child was used as the unit of analysis, access to secondary schooling in the entire sample was determined by the family assets, the level of education of the father, the presence of educated relatives in the extended family, grade repetition in upper-primary schooling, the number of times a Child repeated the sixth grade, and the grade repetition in the lower primary grades and the child's Chores. Access would increase with increased family assets, level of education of the father, presence of educated relatives in the extended family, grade repetition in upper- primary schooling, number of times a child repeated the sixth grade, and access would decrease with increased grade repetition in the lower primary grades and child's chores. In the urban area, access to secondary schooling was determined by the level of education of the mother, the level of education of the father, the family assets, the presence of educated relatives in the extended family, the __jfi..-_n_j :: I ' 17.3 11.71"— 211 number of times a child repeated the sixth grade, and the number of times a child repeated the fifth grade. Access would increase with increased level Of education of the mother, family assets, level of education of the father, presence of educated relatives in the extended family, and number of times a Child repeated the sixth grade; and access would decrease with increased number of times a Child repeated the fifth grade. In the rural area, access to secondary schooling was determined by the family income, the number of times a Child repeated the sixth grade, the repetition in the upper primary grades, the child's chores, and the family preference to send a boy to secondary school. Access would increase with increased family income, number of times a child repeated the sixth grade, and repetition in the upper primary grades; and access would decrease with increased child's chores, and family preference to send a boy to secondary school. VI.2. Policy Implications Policy recommendations are presented to alleviate the burden of the direct private cost of secondary schooling to the parents and, thus, improve access to secondary schooling in Burundi. 212 VI.2.1. Economic Burden The government of Burundi should formulate policies that are fair for both secondary day and boarding schools. Most expenses were made on school fees and personal needs. While transportation was a major problem to the students in the urban areas, especially Bujumbura, which comprised almost exclusively all day secondary schools, school fees and school supplies were a common problem to both types of secondary schools. Therefore, any policy to reduce the economic burden of the cost of secondary schooling should focus primarily on school fees and school supplies. Moreover, in a country where most people emphasize education as a means of status attainment, and where the community (extended family) is involved in educating Children, but where the income of most families is not high enough to sustain secondary schooling, government intervention and community involvement would be essential in improving access to secondary schooling. Expenditures on secondary schooling reflected the expected burden of secondary school education by the parents, which, in turn, indicated the level of education of the father, that of the mother, the amount of family income, and family assets parents had. The more educated the parents were, the more knowledgeable about the direct private cost of education they were. As such, the 213 government of Burundi should to target its effort at disadvantaged families, especially families from rural and poor backgrounds. In cooperation with local institutions, the government should inform the population about the usefulness of formal education for female especially in the rural areas, and the burden of the cost of education to both the government and the parents. It may provide financial assistance to children from disadvantaged backgrounds. For example, scholarship may be established especially for encouraging secondary school participation of girls from rural and farming backgrounds. The government should also start discussing alternative solutions to reduce the economic burden of the cost of secondary schooling, as well as defining the shared responsibility between the government and the community. For example, the government may want to explore ways to get the extended family more involved in sponsoring additional children. It may explore ways to get the local institutions such as Churches to set up educational funds to assist children from disadvantaged families. VI.2.2. Access to Secondary Schooling Access to secondary schooling was predicted by many factors that could be grouped into three categories: grade 214 repetition, affordability and family demand for schooling, and the type of school attended. Grade repetition was directly related to the national test. Secondary—school affordability and family demand included family income, family assets, parents’ level of education and their occupation, family values ad household chores. The attendance of either day or boarding schools indicated the location of families and the income category of the parents. VI.2.2.1. The National Test lFfiv... ..1 . One of the major obstacles to accessing secondary school was grade repetition, especially in the sixth grade, resulting from the institution of the national test at the sixth grade level. Consequently, some children had to Change schools. They moved to urban areas, where there were better educational facilities, which were conducive to a higher probability to pass the national test. Children who Changed schools had educated and/or wealthy relatives. Children whose families were poor dropped out of school because they had no Chance to try somewhere else. Subsequently, the national test at the sixth grade level brought about several problems. It created inequalities between the families who did not have educated or wealthy relatives and those who did. Moreover, while parents were willing to invest in education as a way out of poverty, the 215 national test constituted a serious constraint to their effort. Hence, some parents found school not useful because of a higher rate of drop out after the sixth grade, as a result of the national test. The children who dropped out of school at the sixth grade level were not academically equipped to apply what they learned. Instead of solving the problem of excess manpower in the subsistence sector, these Children increased the population involved in farming. As a result, both the parents and their children would not consider formal education useful. In order to achieve the objective of increased literacy among youth, the policy of the national test should be revised. The government of Burundi, while advocating collective promotion in primary school and adopting the double shift program to increase the literacy rate of Children, should explore the elimination of the national test at the Sixth grade level and administer it at the tenth grade level for tracking purposes. The children would be tracked in areas where they demonstrated more ability. This policy would reduce the gender bias, involve more children in formal education, rather than the Yaga-Mukama religious education, and lead to improved parents’ literacy. However, there is a number of related issues to examine before deciding on the national test (see section VI 3.1). Tam.— 216 VI.2.2.2. Family Income Family income was a key factor in accessing secondary schooling. The burden of the direct private cost of secondary schooling to wealthy families in Burundi was much lower than that for low income families. Income was positively associated with education or occupation of the parents. Since the rural areas of Burundi offered few income generating activities, apart from subsistence farming, the government of Burundi should diversify the economy, modernize farming and revamp small business activities. Moreover, since the results of this study suggested that spending on secondary education increased with the wealth and eduCation of the mother, income generating activities involving women should be given a priority. VI.2.2.3. Gender Differences in Access The family preference to send boys to secondary schooling was a major determinants of access to secondary schooling in the rural area. The more the parents preferred to send a boy to secondary schooling, the less access to secondary schooling for Children in the family. In conjunction with the local institution, the church, the government should inform the rural population about the 217 importance of sending girls to secondary school. Measures to encourage rural girls to attend and stay in secondary schooling may be considered because of the many benefits of educated girls to the future generations, such as financing their own Children and promoting and financing Children in the extended family, as this was shown in this study. Measures should be taken to reinforce the implementation of the government policy of universal primary education which was already in place in 1993. VI.2.2.4. Household Chores: Indirect Private Cost of Schooling Most rural children suffered from limited access to formal schooling because of indirect private costs of schooling. Many Children from the rural areas did not enroll in formal schooling so as to help the family with household chores. Measures to increase family income and level of awareness about the benefits of education may help rural parents make decisions about children schooling and limit their chores. 218 V1.3. Future Policy Analysis and Research VI.3.1. Policy of Expansion of Lower Secondary Schooling A policy of the expansion of lower secondary schooling needs to be studied in terms of implementation issues and the country’s development needs. Analysis should be made of the financial, institutional, and management dimensions involved in expanding compulsory education from the sixth grade to lower secondary school. The government should investigate the national development goals of the country and the role of increased compulsory schooling in contributing to these goals. Examples-of specific studies include the estimation of the costs of educational expansion and the costs of financial assistance to disadvantaged families, the evaluation of the financial feasibility of the expanded compulsory educational policy, the assessment of the management capacity of the education ministry, and the assessment of the employment prospects of school graduates. VI.3.2. Parents’ Attitude Toward Schooling Further research should be conducted to better understand the rural parents’ attitude toward girls 219 education. Rural children’s access to secondary schooling decreased because of the parents’ preference to send boy to secondary schooling. Factors which influence this attitude should be identified in the context of Burundi. At the same time this research should identify factors which promote rural parents support for girls education. In addition research should be conducted to investigate how to channel information to parents about educational assistance available from government,and community organizations school. VI.3.3. Replication of the Study At the national level, this study’s sampling may be improved to include the provinces and communes excluded (see section on limitations of the study) in the present study to make findings more representative at the national level. This could be done when the politiCO-social conditions allow it. Improved sampling may allow regional comparisons and improve government intervention policies. At the international level, this study may be replicated in other developing countries to allow comparative analysis of the findings. APPENDICES APPENDIX A: TABLES: NUMBER OF CHILDREN IN SECONDARY SCHOOL 220 APPENDIX A Table 64. Cross Tabulation of INSEC (Number of Children in Secondary School) and the Father's Occupation (FATHEROC) NUMBER OF CHILDREN IN SECONDARY SCHOOL IN THE TOTAL FATHEROC FAMILY NONE 1 2 2< Farmer 66.2 32.2 1.6 100 Gen Lab 54.5 40.9 4.5 100 Qual Techn 25 31.4 18.8 18.8 100 Business 0 33.3 33.3 33.3 100 Profess 3 37.6 18.8 40.6 Table 65. Cross Tabulation of INSEC (NUmber of Children in Secondary School) and Mothers' Occupation. NUMBER OF CHILDREN IN SECONDARY SCHOOL IN THE TOTAL MOTHEROC FAMILY NONE 1 2 2< Farmer 56 38 1 1 100 Gen Lab 29 63.1 2.6 5.2 100 Qual Techn 0 38.5 38.5 23 100 Business 33.3 40 13.3 13.3 100 Profess 0 45.4 13.6 41 100 III-7.11.1; 1 ' 221 Table 66. Cross Tabulation of INSEC (Number of Children in Secondary School) and Family Income. NUMBER OF CHILDREN IN SECONDARY SCHOOL IN THE TOTAL INCOME FAMILY NONE 1 2 2< 1 86 14 0 O 100 2 43.2 56.8 0 0 100 3 40.6 50 9.4 0 100 4 17.1 65.7 11.5 5.7 100 5 0 100 Table 67. Cross Tabulation of INSEC (Number of Children in Secondary School) and Family Expenses. NUMBER OF CHILDREN IN SECONDARY SCHOOL IN TOTAL EXPENSES THE FAMILY NONE 1 2 2< 1 86 14 0 0 100 2 43.2 56.8 0 0 100 3 40.6 50 9.4 0 100 4 17.1 65.7 11.5 5.7 100 APPENDIX B: INTERVIEW QUESTIONNAIRE (IN ENGLISH) 222 APPENDIX B QUESTIONNAIRE Questionnaire # : /.../.../.../ Province : ............. Commune : ............. Colline : ............. Date : ---/---/--- Mo Day Year Direct Private gogtg Ang Educatigngl Ingggglitiog: A Gag; Stud! Qf Agcool T9 Sogogdagx §chooling In Burundi The purpose of this study is to try to understand how people deal with private costs of secondary schooling in Burundi. Your name is not going to be on any document of this study, and the findings of this study will be kept confidential. I. RESIDENCE grgggg 1.1. Respondent sex male female. 1: male 2: female 1.0. 1.2. How many people live with and eat with you? 1.2. spouse brothers sons sisters daughters grandparents others (specify) 1.3. How many children do you have now? 1.3. 1.4. Do you have other children who do not live with you? 1.4. 0 a no. Go to question 2.1. 1 a yes. Go to question 1.4. 1.5. I would like to know'how many they are, where they are, and what they are doing. Place how many (Canone) school university army civil servant Other (specify) I‘m-21:. .1". 3 -‘LI 223 II. CHILDREN'S INFORMATION 2.1 Now I am going to ask questions about each of your children. Let's start with the Oldest (then I will take the second oldest, etc... and finish with the youn est. Question Coding 1 2 3 4 5 6 7 8 1. Age Age in years 2. Sex 1 = Male 2 a Female 3. Has s/he ever been 0 . No 4 Q4 enrolled to school? 1 a Yes» Q5 4. (If no on Q3) Why RNOSCHO did the child not go 9 = N/A to start school? After answering this queetion, go to 0 2.6.1 5. (If yes on 03) Is 0 = No the child still 1 = Yes attending school? 6. How old was s/he Age in when s/he started years school? If child still in school, go to Qll. 7. If s/he has stopped GRADEFIN going to school, which 99 2 N/A grade did s/he finish before s/he stopped? 8. Did s/he repeat a OaNo grade before leaving 1=Yes schooling? 8=Don't If YOlo go to Q2.2.1 remember 9. Which grade is s/he GRADNOW in currently? 10. Is s/he in a day 1 a Day or boarding school? 2 a Board 11. Is s/he studying 1 a Rural in the rural or urban 2 - Urban area? * Note: The Q 2.1.1- 2.1.9 were answered to according to eacflu child situation. These items constitute the core qustions for each Child. From the case of each Child, I know who to continue and how to follow the child: the age determines that the child is in the target age group 12-25 year. The other items determine that the Child is in or out of school and the grade. 224 2.2. Primary school Grade repetition * Note: All the children in the targeted age group (12-25) are considered as well as those who are Older than 25 but are still in secondary school. Question Coding 1 2 3 4 5 6 7 8 1. Did s/he repeat any 0 = no grade in primary school? 1 a Yes 2. Which grades did s/he repeat? grade 1 0: no l= Yes grade 2 0 a no 1 - yes grade 3 0 a no 1 :- yes grade 4 0 a no 1 . yes grade 5 0 a no 1 = yes grade 6 0 a no 1 a yes 3. How many times did s/he no.time repeat the 5th grade? 4. How many times did s/he no.time repeat the 6th grade 5. Did s/he change school 0 . no to repeat the 5th grade? 1 = yes 6. Did s/he change school 0 a no to repeat the 6th grade? 1 a yes yes/no 7. If yes, which school 1: did s/he change to (rural Rural or urban) 2: Urban 8. when changing school, 0. no did s/he need to stay with 1: yes a relative or with a friend? If the child is not in school go to Q2.6. 225 2.3. Went back to an early grade * Note: Only the children in the targeted age group (12-25) who are still in school are considered as well as those who are older than 25 but are still in secondary school. Question Coding, 1 2 3 4 5 6 7 8 1. Did s/he go back to 0 a no repeat an earlier grade? 1 =yes (from 6th to 3rd grade) 13 Y“: continue, if no, 9° to 4.4 2. From which grade did grade s/he go back? no. 3. To which grade did grade s/he go back to? no. 4. Did s/he Change 0 . no school in this process? 1 ayes Yes/No 5. If yes in which 1. school did s/he go rural (rural or urban school) 2- urban 6. Did s/he need to stay 0 a no with a relative or with 1 a a friend? yes 2.5. Time on extra-academic activities fgr primary school Children * Note: All the Children in the targeted age group (12-25) who are still in school are considered as well as those who are older than 25 but are still in secondary school. gguestion Coding' 1 2 3 4 5 6 7 8 1. Did the child have to 0 - no do chores home after 1 a yes school? If no go to 4.6 2. As I read you a list 0 = no of Chores, can you tell — yes me the kinds of chores s/he was likely to do before, after school? ...: I . babysit fetch water fetch firewood cook . work garden . hired labor sell small items . other(specify) CDQQU‘IIHWNH __ arr-.3 226 2.6. Foregone income * Note: All the children of the interviewee who are not in school are considered. Except those who are too young and are under school age. Question Coding 1 2 3 4 5 6 7 8 1. What kind of job does s/he do? 2. How much money money/m does a child who does not go to school earn money/d on average? Money earned/month Money earned /day gzxtocgild and repeat with the 2.7. Fin Cial hel for seconda school relate e ens * Note: All the children who are still in secondary school regardless of the age. Question Coding 1 2 3 4 5 6 7 8 1. Did the parent pay no a 0 all secondary school yes a 1 reiated,expenses for (If yes and the child 1. still this Chlld? in secondary school go to direct private costs 4.8.) 2.If no, what is your HELPREL relationship with the 99=N/A helper (eg. cousin, friend) 3. Does the helper live' 1=rural in the rural/urban 2=urban areas? 4. What is the relative degree / friend's level of 99=N/A education? 5. Does s/he pay part of 1: part the cost or all the 2: all expenses? 99=N/A 227 2.8. Direct private cogt of segondagy education by parents themselves * Note: All the children who are still in secondary school regardless of the age for whom the parents pay ALL the expenses of secondary schooling. Question Coding 1 2 3 4 5 6 7 8 1. How much money did you DPCYEAR need to have available per term last year for this Child including school fees, and all other expenses? first term DPCTRMI 2nd term DPCTRM2 3rd term DPCTRM3 2. When did s/he start SCHOOLY secondary school? (SCHOOL-YEAR) CHOOSE AND ANSWER Q4 OR 05 TO 3. How much did you spend HOWMU on each of these items last term? school fees clothes blanket bed sheets shoes school supplies (pen and pencil, School- bag, compass, ruler) 4. If your child bought DAYSNO. them, how much money did you give him/her? 5. Was s/he ever sent back home for not paying school fees during the term(while others stayed at school) 6. For how long did the DAY.NO Child stay home? no yes HO 7. Did the child have to 0 = on repeat a course 1 = yes "repechage" in the year s/he was sent back home ? 8. Did the Child repeat 0 a no any grade in secondary 1 . yes school? GO TO Q2.1. AND REPEAT WITH THE NEXT CHILD 228 III. DIRECT PRIVATE COST OF SECONDARY EDQCATIQN * Note: This applied to families who currently have a child or children in secondary school. 3.2. How did you find money to pay? 3.2. 1 From my savings 2 Sold cattle 3 Sold harvest (specify) 4 Sold land 5 Borrowed money IV. RELATIVE L NC WI THE FAMILY. 4.1. Are there relatives living permanently (at least for one school year) with you? 4.1. 0 a NO GO TO 04.5. 1 = YES GO ON 4.2. How many are they? 4.2. 4.3. Now I will ask you some information about each of these young relatives who live with you, one at a time. Let's start with the oldest (then I will take the second Oldest, etc... and finish with the youngest. Question Coding 1 2 3 4 5 6 7 8 1. Age Age in years 2. Sex 1 a Male 2 . Female 3. When did s/he start to age live with you? ofthe child 4. What is the reason for living with you? 5. Is s/he going to school? 0 a no 0 13 e yes/ n Y 8 If no, go to the next child. 6. If yes, which grade did s/he join when s/he came to live with you? 7. What is his/her courent grade now? 8. Did s/he repeat any grade 0 . no while s/he was with you? 1 = yes 9. How many times did s/he TIMENO repeat grades? 99-N/A 10. Which grades did s/he repeat? 11. Did s/he repeat grade 0 a no before living with you? 1 = yes 229 4.4. Financing of secondary education Question Coding 1 2 3 4 5 6 7 8 1 Do you pay for 0 a no all his/her 1 a yes schooling? yes/ no 2. If no, who else helps to pay for his/her schooling? 4. If s/he has parents, where do . they live? 5. What is the father's education? 6. What is the mother's education? 7. What is the father’s occupation 8. What is the mother's occupation? AFTER THIS CHILD GO TO THE NEXT OLDEST. After you finish with the last relative living with the family go to Q 4.4. 4 .5. Do you financially help other students besides your Children and relatives living with you? 4.5. 0 1 4. ##1##31: ONU'IIHWNH = no. GO to question 5.1. = yes. Go to next question 5.1. Who do you help? relation grade residence kind of help AsPrimary 1- City 1-financial BsSecondary 2- rural 3amotivation c-university 1 2 3 4 5 V. LD I .1. How much, on average, do you spend on: per month last year food Not apply . beverages Not apply agric. equipment agric. inputs (fertilizers...) clothing . household equipment 230 7. health and medical services 8. school expenses 9 taxes/contributions (cash & in kind) 10. transportation/travels 11 rent 12 hired labor 13. others (specify) 5.2. Do you live in your own house? YES/NO not applicable 5.2. 5.3. If no how much do you pay for rent 5.3. VI. INCOME (FINANCIAL CAPITAL) 6.1. As I read you some of the ways people get money, please tell me whether or not you get money in this way and how much money you get? 6.1. Choose all that apply 0 . no, 1 ayes yes or no AMOUNT sales of agricultural products : sales of livestock products sales of land . salary of my spouse : . my salary - social security benefits remittances (kinship relationship) as hired labor(yourse1f) mummbuNH \0 Family members hired labor 10. business (specify) 11.rents 12. other (specify) 01 N S 5' P O 33‘ O H (T 33" (B H: O H ...: O t g. (.0 p (T (0 3 to D. O '< O (3 {'3‘ II) < (B p :1 your home? 0=no 1- yes radio television newspaper car or lorry VCR refrigerator camera piped water electricity iron sheet house own house (rural/urb) cement house thatched roof bicycle motor cycle 6.2. How many of these do you have in your home? 0: none Number Number 1.cows 4.pigs 2.sheep 5.chicken 3.goats 6.coffee trees ___Jfl-"_'.‘. 231 VII. RELATIVBS’ ED CA ION 7.1. Do you have educated relatives? yes or no 7.1. 0 = no. Go to question 8.1. l = yes. Go to question 7.2. 7.2. What do they do? Where do they live? What level of education? relation Occupation residence marital status educ level rural/urban ########## Hmmqmmbwww 0 VIII. PARENTS' 8D CATION 8.1. Have you ever been to school? 8.1. 0 = no. Go to question 8.3. 1 a yes. Go to question 8.2. 8.2. What was the highest grade you completed? 8.2. .3. Has your spouse ever been to school? 8.3. = yes, go to question 8.4. = no, go to question 9.1. OHCD 8.4. What was the highest grade that your spouse completed? IX. P ' 0C TION AG! 9.1. What is your principal work ? Your title? 9.1. 9.2. What is your second work (which brings income) if you have one? 9.2. 9.3. What is the principal work of your spouse? What is his/her title? 9.4% What is your spouse's second work (whg g brings income) if s/he has one . . 9.5. Do you mind telling me how old you are (age)? 9.5. 9.6. What is your spouse's age? 9.6. “Pt- ...—n— 13.-.1". Y". ‘_ a 232 x. OPINION AND WILLINGNBSS TO PAY 10.1. Do you think secondary school education is for you (choose only one) 10.1. . not expensive at all a little bit expensive expensive . very expensive .91-LAMP ('7 5' P. 5': 10.2. How worthwhile do you it is to spend your money on secondary schooling for any child? 10.2. .not at all worthwhile .somewhat worthwhile .worthwhile .very worthwhile #00th 10. 3. If no, why? Mention all that apply. 10.3. 1. people who did not go to school earn more : 2. educated people do not financially help their parents: 3 educated people are unemployed 4. Prestige 5. others (specify) 10.4. If yes, why? Mention all that apply. 10.4. 1. educated people earn more : 2. educated people have better lifestyle : 3. educated people help parents/siblings/relatives : 4. others (specify) : 10.5. would you be willing to pay more than you are now paying for your children's secondary education? yes or no. 10.5. 0 a no go to Q10.8. and skip 010.6-7. 1 a yes Continue with 010.3. * Note: These questions (10.5, 10.6, 10.7) are intended for those who have the experience of paying for secondary school. 10.6. Why? 10.6. 10.7.If yes, how much more would you be willing to pay (choose one). 10.7. 1. less than 100? Bu per child per year 2. Between 100FBu and 1000FBu per child per year 3. Between 1000FBu and SOOOFBu per child per year 4. More than 5000 per child per year 10.8. If you could educate only one of your children would you prefer that the educated person be a boy or a girl? 10.8. 1 a girl 2 3 boy 10.9. For which reason (only one)? 10.9. APPENDIX C: INTERVIEW QUESTIONNAIRE (IN KIRUNDI) 233 APPENDIX C QUESTIONNAIRE Questionnaire # : /.../.../.../ Province : ............. Commune : ............. Colline : ............. Date : ---/---/--- Stu Of Acc II To 80 ad: School n In Burundi Igituma turiko turabaza ibi bibazo, ni kugira mumfashe gutohoza ingene amahera y'ukurungika umwana mu mashure makuru angana; hamwe n’ingene abantu babigenza kugirango abana bashobore gushika muri ari ayo mashure kandi bayaheze. Inyishu zose muzompa ntizizogukurikirana. Izina ryanyu ntirinakenewe. I. RESIDENCE 8TATU§ 1.0. Uwishura umugabo Umugore. 1- umugabo 2a umugore 1.0. 1.1. umutumba w'amavuka ------- 1.2. umutumba ubako ------- 1.3. Commune y'amavuka ------- 1.4. Commune ubamwo ------- 1.5. Umutumba w'amavuka w'umucance 1.5.1. Commune yavukiyemwo 1.6. Mutunze abantu bangahe munzu? 1.6. Umucance abahungu babavukanyi Abahungu banyu Abakobwa babavukanyi Abakobwa banyu abavyeyi banyu Abandi (mupfana iki?) 1.7. Ubu mufise abana bangahe? 1.7. 1.8. Murafise abana bataba ngaha? 1.8. 0 a oya. ja ku kibazo ca 2.1. 1 ego. Ja ku kibazo ca 1.9. 1.9. Abo bana ni bangahe, bari hehe, bakorayo iki? Mumbarire mukurikije umwana umwumwe. Place ni bangahe iyo ari icakorayo (Oanone) (egzauja) mw'ishure muri Kaminuza mu Ntwaramiheto akorera reta Ahandi (Tomora) 234 II. CHILDREN’S INFORMATION 2.1 Ubu nagira tuze turaganira tuvuga umwana umwumwe duhereye ku mukuru. Duhejeje ivyerekeye umukuru niho tuja ku wukurikira. Tuja guhereza ku mutoyi. Question Coding 1 2 3 4 5 6 7 8 1.Imyaka yiwe Age in years 2. umuhungu canke 1: Male umukobwa 2 a Female 3.Yarigeze aja 0 a No » mw’ishure ? 04 l a Yes» 05 4. Murazi icatumye RNOSCHO adatangura ishure? 9 - N/A After answering this quostion,_go to 0 2.6.1 5. K0 yatanguye, 0 . No I aracari mw'ishure? 1 a Yes 6. Yatanguye ishure Age in I l I l afise imyaka ingahe? years If child still in school. go to 011. 7. Yavuye mw'ishure GRADEFIN ahejeje umwaka wa 99 a N/A kangahe? 8. Yarigeze ahitira 0=No mw' ishure imbere 1-Yes yuko ava mw’ishure? 8=Don't remember If you, go to 02.2.1 9. Ubu ari mu mwaka GRADE.NO wa kangahe? 10. Yiga ararayo 1 a Day canke yiga ataha? 2 a Board , 11. Ari mw'ishure iri 1 - mugisagara canke mu Rural mihana? 2 a Urban 235 2.2. Primary school Grade repetition Question Coding 1 2 3 4 5 6 7 8 1. Yarahitiye umwaka 0 a no numwe akiri mu 1 a Yes mashure matoyi? 2. Yahitiye imwaka iyihe? Umwaka wa 1 0: no l= Yes umwaka wa 2 0 a no . yes umwaka wa 3 I no agyes umwaka wa 4 I no l O 1 O 1 - yes 0 1 0 l umwaka wa 5 - no I yes umwaka wa 6 - no sgyes 3. Yahitiye kangahe no.time mu mwaka wa 5? 4. Yahitiye kangahe no.time mu mwaka wa 6? 5. Aho yahitira 0 = no umwaka wa 5 yagiye 1 = yes guhitira ahandi canke yagumye mw‘ishure gyahora yigamwo? 6. Aho yahitira 0 a no umwaka wa 6, yagiye 1 . yes guhitira ahandi canke yagumye mw'ishure yahora yigamwo? 7. Aho yahindura 1= ishure yagiye muyo Rural mugisagara canke mu 2- mihana? Hehe? Urban 8. Yagiye kw' ishure, 0- no aho yiga ataha aha 1- yes muhira canke yataha ku ncuti canke umugenzi? If the child is not in school go to 02.6. 236 2.3. 2 U IRA I MU MU MWAKA YAHEJEJE XE Question Coding 1 2 3 4 5 6 7 8 1. Yarigeze asubira 0 a no inyuma nk'uko benshi 1 ayes basubira inyuma mu mashure matomato? (Ava nko mu wa 5 aja nko mu wa :f4yes, continue, if no, go to 4) ° 2. Yasvubiye inyuma ava grade mu mwaka wa kangahe ? no. 3. Asubira mu wa kangahe? grade no. 4. Aho yasubira inyuma 0 . no yarahinduye ishure? ego 1 syes canke oya 5. Aho yahindura ishure la yagiye muyo mugisagara rural canke mu mihana? Hehe? 2- urban 6. Yagiye kw' ishure, aho 0 a no yiga ataha aha muhira 1 . canke yataha ku ncuti yes canke umugenzi? 2.5. A B RA IBI RW VY' B I I DMELLEHHE§_EIQXA_ Question Coding, 1 2 3 I 4 S I 6 7 [8 1. Uyu mwana 0 . no yarategerezwa gukora 1 a yes ibikorwa vy'imuhira ’ imbere yuko agenda canke If no go to 4.6 avuye mw'ishure? 2. Muri ibi bikorwa 0 . no ngira ndagusomere ni 1 . yes ibihe ahora akora imbere yuko mu gitondo canke atashe? kurera umwana kuvoma amazi . gusenya inkwi . guteka gukora mu mirima . gukorera amahera kudandaza ibindi (tohora) mxlChU'lbbJNH 237 2.6. Foregone income Question Coding 1. Uwu mwana atari mw'ishure akora akazi gaki? 2. Akorera amafaranga money/m angahe ku kwezi Canke akorera angahe ku munsi? money/d Go to 02.1 and repeat with the next child. 2.7. Finangial help for secondapy school related expenses Question Coding 1 2 3 4 5 6 7 8 1. Muramurihira no a 0 amafaranga y'ishure yes a 1 yose yo kuja mu mashure yésfimbuyg? muramugurira (If yes and the child is still 1 f 0:93 S vyose in secondary school go to VY ls ure. direct private costs 4.8.) 2.None ko mutamurihira HELPREL vyose mupfana iki nuwo 99=N/A muntu abibafashamwo? 3. Uwo abafasha aba 1=rural mugisagara canke mu 2=urban mihana( mu kirundi? 4. Yize amashure angana degree gute? Yagarukiye mu wa 99=N/A kangahe 5. Amurihira vyose 1: part canke abafasha igice 2: all gusa? 99=N/A 238 2.8. Diregt private cost of secondary education by parents themselves Question Coding 1 2 3 4 5 6 7 8 1. Uyu mwana DPCYEAR mwakoresheje amaherera angahe kugira aje mw'ishure umwaka uheze mushizemwo ayishire nayo ibindi bintu vyose mwaguze kw’itrimestre ? Itrimestre ya 1 DPCTRMl Itrimestre ya 2 DPCTRM2 Itrimestre ya 3 DPCTRM3 2. Yatanguye amashure SCHOOLY makuru ryari? (umwaka w'ishure) CHOOSE AND ANSWER Q4 OR 05 TO 3. Wakoresheje angahe HOWMU kugira ugure ibi bintu canke ubirihe Amahera y'ishure Impuzu uburengeti amashuka ibirato Ibikoresho vy’ishure (aamakaye amakaramu, amavarisi, canke isandugu incamurongo, icompa, n'ibindi 4. Nimba umwana DAYSNO. mwamuhaye amafaranga ngo yigurire ibikorshovy'ishure mwamuhaye angahe kw’itrimestre? 5. Barigera bamwirukana ngo nuko atatanze amafaranga y’ishure abandi banyeshure basigaye bariga? no yes l-‘O Illl 6. Nimba baramwirukanye DAY.NO yasubiye mw'ishure haheze iminsi ingahe? 7. Muri uwo mwaka 0 - on yarigeze agira 1 = yes repachage? 8. yirigeze guhitira 0 a no H II muri segondaire? yes GO TO 02.1. AND REPEAT WITH THE NEXT CHILD 239 III. DIRECT PRIZATE COST OF SECONDARY EDUCATION ABANTU BAFISE ABANA BARI MU MASHURE MAKURU UYU MWAKA. 3.2. Mwaronse amafaranga yo kurungika umwana mw'ishure mu buryo ubuhe? 3.2. 1 Ayo twari twarabitse 2 twashoye inka 3 Muvy'imburwa twashoye 4 twagurishije umurima 5 Twaraguranye 3.3. Nimba mwaraguranye mwaguranye angahe? 3.3. 3.4. Mwayaguranye hehe? 3.4. 3.5. Hari ayo murasubira gusubiza? Angahe? 3.5. IV. RELATIVE LIVING WI THE FAMILY. 4.1. Hari incuti muhaye indaro mubana ngaha? 4.1. 0 . NO GO TO 04.4. 1 2 YES GO ON 4.2. Nibangahe? 4.2. 240 4.3. Ubu nagira tuze turaganira tuvuga umwana umwumwe duhereye k'umukuru. Duhejeje ivyerekeye umukuru niho tuja k’uwukurikira. Tuja guhereza kumutoyi. Question Coding 1 2 3 4 5 6 7 1. Afise imyaka Age in ingahe years 2. umuhungu canke 1 = umukobwa Male 2 a Female 3. yatanguye kubana 0 a No _ namwe ryari? 1 = {1: Yes ' 4. Ni kuki yaje kuba ; ngaha? i 5. Araja mw’ishure? 0 = no 1 a . 4yes If no, go to the next child. ; 6. Yashikiye mu i_ mwaka wa kangahe aje " kuba ngaha? 7. Ari mu mwaka wa kangahe? 8. Yarigeze ahitira 0 a no aba ngaha? 1 a yes 9. Amaze guhitira TIMENO kangahe? 99=N/A 10. Amaza guhitira mu mwaka wa kangahe? 11. Mwba muzi nimba 0 . no yarigeze guhitira 1 = imbere yuko aza kuba yes iwanyu? 241 4.4. Financing of secondary education Question Coding 1 2 3 4 5 6 7 8 1 Muramutangira 0 = no amafaranga y’ishure 1 = yes yose mukamugurira n'ibikoresho vy'ishure? 2. None ninde abaibashashmwo mukumurihira amafarangagy'ishure? 4. Nimba agifise abavyeyi baba hehe? 5. Se wiwe yagarukiye he mumashure niyaba yragiye mw'ishure? 6. Nyina wiwe yagarukiye he mw'ishure ni yaba yaragiye mw'ishure? 7. Se atunzwe nakazi akahe? 8. Nyina wewe akora uduki? AFTER THIS CHILD GO TO THE NEXT OLDEST. After finishing with the last relative living with the family, go to 0 4.4. 4.4. Uretse abana banyu hamwe nabo baba iwanyu hari abandi bana mufasha mukurihira amashure? 4.4. 0 a no. Go to question 5.1. 1 = yes. Go to next question 4.5. Mufasha bande? Ico mupfana umwaka arimwo Iyo aba Mumufashisha iki A-Primary 1: city 1=financial B=Secondary 2: rural 3=motivation C=university 33131138148383: UIIbU-JNH 242 V. HOUSEHOLD EXPENDITURES 5.1. How much, on average, do you spend on: per month last year 1.indya Not apply 2. ibinyobwa Not apply 3. ibikoresho vyo kurima 4. amase yo gutabira 5 impuzu 6 ibikoresho vyo munzu 7 kwivuza hamwe n'imiti 8. abanyeshure 9. tagisi hamwe nabo muha amahera 10.Amaafaranga ya transport 11 amafaranga y'inzu 12 kuriha abakozi 13.ibindi 5.2. Iyi nzu mubamwo niwanyu? YES/NO not applicable 5.2. 5.3. Ko atari rwanyu muyiriha angahe ku kwezi 5.3. VI. AMAEARANQ MURONXA 6.1. Ngira ndabasomere ingene abantu baronka. Maze muze murambarira ko muronka amahera muri ubwo buryo hamwe n'amahera muronka uko angana? 6.1. choose all that apply 0 = no, 1 =yes yes or no AMOUNT 1. kugurisha ivyimburwa : 2. kugurisha ibitungwa 3. kugurisha itongo 4. Umushahara wawe 5. umushahara w'umucance 6. social security benefits 7. Ayo abantu bampa) 8. Ndaca ingero 9. Abo ntunze baraja guca ingero 10. Udandaza iki? 11. Ayo bapanze ko inzu zawe 12. Ubundi buryo ntavuze(tomora) 6.2. Muri bino bintu ivyo mutunze ni ibihe? 0=no 1- yes iradiyo television ibinyamakuru umuduga canke ikamyo VCR ifirigo icuma cugufata amafoto robine y'amazi munzu umuyagankuba(amatara) inzu y'amabati inzuiri hehe (rural/urb) isima inzu y'ivyatsi ikinga ipikipiki 243 6.2. Muri ibi bintu mutunze ibihe? bingahe? 0: none Number Number 1.inka 4. ingurube 2.intama 5.inkoko 3.impene 6.ibiti vy'akawa VII. AMASHURE INCUTI ZANQ! ZIZE 7.1. Hari incuti ufise zize amashure? 7.1. 0 a no. Go to question 8.1. 1 . yes. Go to question 7.2. 7.2. mupfana iki? Bakora iki? Baba hehe? bize amashure angana iki? ico muphana akazai iyo baba arubatse Amashure E" rural/urban oya /ego ##ttttttikt Hmmqmmoth-I gPhl-Ci‘ .inu‘. VIII. HURE ABAVYEYI SIZE None wewe waragiye mw'ishure? 8.1. no. Go to question 8.3. yes. Go to question 8.2. H00 III-I 8.2. Wahejeje umwaka wa kangahe? 8.2. U Umucance wawe wewe yaragiye mw'ishure? 8.3. yes, go to question 8.4. no, go to question 9.1. GHQ I 8.4. Umucance yahejeje umwaka wa kangahe? IX. 555;; ABAYYEYI BAKQRA 9. None ko twayaze kera wewe ukora canecane akazi akahe? Titre yawe ni iyihe? (akarorero: deregiteri w’amashure matoya) 9.1. 9.2. Akandi kazi woba ukora ni akahe? 9.2. 9.3. Umucance wewe akora akazi akahe? Titre yiwwe ni iyihe? 9.3. 9.4. Akazi kandi yoba akora ni akahe? 9.4. 244 X. ICO MWIYUMVIRA XU MASHURE MAXURU 10.1. Muri bino vyiyumviro kuvyerekeye amashure makuru ni ikihe gihuye nivyo muyiyumvirako (Tora kimwe gusa). 10.1. ntazimyve noa gatoyi arazinmvye buhoro buhoro arrazimvye arazimvye cane gose QWNH 10.2. Mubona bifise akamaro gutanga amahera yo kurungika umwana mu mashure makuru? 10.2. 1.ntaco bimaze 2.birakimaze gatoyi __ 3. birakimaze rni 4.birakimaze cane gose ' 10.3. Ko wavuze ngo ntaco bimaze ni kuki( tora kimwe muri ibi ngira ngusomere. 10.3. 1. Abize ntibaronka amahera menshi nkabatize 2. Abize ntibafasha abavyeyi babo 3. Abize ntibaronka akazi 4. Ibindi mwitoreye 1.1.")! r... on .- 10.4. Ko wishuye ngo hari ico bimaze ni kuki? (Tora muri ivyo bikurikira). 10.4. abize baronka amahera menshi gusumba abatize Abize baba ho neza Abize barafasha abavyeyi harimwo akarusho. Ibindi witoreye UIIhWNH 10.5. Mwuyumvira ko mwoshaka kwongerako ku mafaranga mukoresha mukurungika abana mw'ishure? yes or no. 10.5. 0 a no go to 010.8. and skip 010.6,7. 1 a yes Continue with 010.6. 10.6. Kuki? 10.6. 10.7.Kowemeye Wumva worenzako angahe kuyo utangira umwana umwe ubu? Tora kimwe muri ibi. 10.7. ari munsi y’ amafaranga 100 ku mwana umwe ku mwaka Hagati y’amafaranga 100 n’ 1000 ku mwana umwe ku mwaka Hagati y'amafaranga 1000 n’ 5000 ku mwana umwe ku mwaka Arenga amafaranga 5000 ku mwana umwe ku mwaka hWNH 10.8. Ushobora gusomesha umwamna umwe gusa wosomesha umuhungu canke umukobwa? 10.8. 1 a umuhungu 2 . umukobwa 10.9. Kubera iki (vuga igituma kimwe gusa)? 10.9. ll. 11. l. 2. 245 XI. AGE OF RESPONDENT Urashobora kumbarira imyaka yawe? 11.1. Umucance afise imyaka ingahe? ll. .3".- ‘-.' l7?" APPENDIX D: LETTERS OF PERMISSION TO CONDUCT STUDY IN BURUNDI 2346 LETTER FROM THE MINISTRY OF INTERIOR L_JZZ__4uou_4333.. Remauaue DU mum Buiumbuml (MINISTERE DE L'INTERIEUR ET DU A Madame HAYOYA Marie )EVELOPPEAENT DES COLLECT I U! 758 LOCALES B.P. 1317 BUJUMBURA. DINNEFDUIHNHHRE Eff ILP.IMII TdL.22‘2fl2 u' tout/M'— Hedeee, £a’ Feisant suite a votre lettre du 12 avril 1993 deeendent l'eutorisetion de ‘ mener dee enquétes en milieu rural pour une etude relative eu coat direct de l'education eu Burundi dens lee Provinces de Karuzi et Gitega, ainsi qu'a la Heirie de Bujusbure, J'ei l'honneur de vous informer que je marque non accord. Les eutorites adsinistratives qui me lisent en copie peuvent vous faciliter 1e tache. Veuillez agreer. Madame. l'assurence de me consideration distinguee. LE HINISTRE DE L'INTERIEUR SEE£§-EQQB-£!EQE!!1198-5-3 ET 0" DEVELOPPEHENT - Monsieur 1e Gouverneur de DES COLLECI;V§T§§&LOCALES Province de et a ........ . Fr.n;é{; "3525 _ ': - Monsieur 1e Maire de la Ville P6 de Bujumbura a ggagggggg_ - Monsieur l'Adninistreteur Cossunel d. C: E ........... _‘<."'.' ...:zx‘rio - Monsieur 1e Directeur du CURDES - Monsieur le Chef de Zone (TOUS) 247 LETTER FROM THE RESEARCH CENTER: CURDES my Centre Universitaire de Recherche pour' le Développement Economique a: Social 0 Je soussigne, Pascal RUTAKE, Directeur du CURDES (Centre Universitaire de Recherche pour le Developpe-ent Econo-ique et Social), atteste que la nonn‘e MAYOYA Marie, Etudiante a l’Université de Michigan, effectue un travail de recherche sur le "Coat supporte par les parents pour l’enseigne-ent secondaire- de leurs entants au Burundi”. dans le cadre d’une convention entre l’Université de Michigan et le GUIDES. Le travail de recherche se (era sur base d’enquetes qui se dérouleront du 5 octohre 1992 i Aofit 1993 dana les provinces suivantes : Bujunhura rural Mairie de Bujumbura Province de Karusi Province de Gitega. La présente attestation est A faire valoir & qui de droit. Fait a Buja-burs, le 5 octobre 1992 MM:IM u.“- . F“ 2 257 223721 1m:1m8 248 BIBLIOGRAPHY Anderson, M. B. (1988). Improving Access to Schooling in the Third World: An Overview. BRIDGES Research Report Series, (1), March. Burundi, (1991). Renouveau du Burundi. Enseignement Primaire et Secondaire: Concours National Session 1991. No 552, (Samedi, Juillet 25). Bidou, J-E, et al. 1991. Geographie dp Burpngi. Hatier, Paris, France. F Bowman, M. J. (1971). Converging Concerns of Economist and Educators. Edpcation and the Economies ef Human ; gepipel. The Free Press New York. P52-67. f Burundi Ministére de 1' Education Primaire et Secondaire. (1993) : Stapiepigpee §colairee 1221-22. Burundi, Ministére de 1' Intérieur, Bureau Central de if Recensement. (1990). Réeultape Ppeviseiree Recepsemept general de le Pepulation e; de 1' Hapipatiop. Gitega, Burundi. Burundi, Ministére de l'Education National. (1988). S a ' i e laire 1986-1 7. Bujumbura, Burundi. Burundi Ministére de 1' Education Primaire et Secondaire. (1991). §te§iepigpee 2colairee 1222—20. Bujumbura, Burundi. Carnoy, M. (1982). Education for Alternative Development. WW. p-140-160- June- Carnoy, M., & H. M. Levin. (1985). geheelipg egg Werk in the Demeerepie spate. California Stanford University Press. Clement, W. (1975). h ana 'a o El't : An apelyeie pf Eeenemic Pewer. Toronto: McClelland and Stewart. Coleman, J. S. 1988. Social Capital in the Creation of Human Capital. Amepieep Jeprne; pf geeielegy. 94 895- 8120. Coleman, J. 8., & T. Hoffer. 1987. Ppplie and Erivepe High - T m o n' . New York : Basic Books. 249 Currie, J. (1975). Family Background, Academic Achievement, and Occupational Status in Uganda. In Heyneman S.P. & J. Currie. Schooling,Aeademic Performance and Qccupational Attainment in a Non- Industrialized geeiepy. Washington D.C.: University Press of America. 80-93. Fuller, B. (1986). Raieing School Quality in Developing Counpries; What inveetmeppe pepep leerning? Discussion Paper No. 2. Washington, D.C. The World Bank. Harbison, R.W., & E. A. Hanushek. (1992). 222922292 Performanee of the Boer Leseope frem gppel Nerppeeep pp Brazil. The World Bank. New York: Oxford University r Press. Herz, B., K., et al. (1991). Lepting girls Learn. I Promising Approaches in Primary and Secondary Education. Discussion Paper No.133. Washington D.C.: World Bank. Heyneman, S. P., & J. K. Currie. (1979). Seneeling, L Academic Pe f c a c a i nal At ainmen in a Non-Industrialized Seeiepy. Washington D.C. University Press of America. Heyneman, S. P. (1983). Education during a Period of Austerity: Uganda 1971-1981. Cemperapive Edpeation Revieg. 27(3) Feb. p. 403-413. Heyneman, S., D. Jamison, & X. Montenegro. (1984). Textbooks in the Philippines: Evaluation of the Pedagogical Impact of a Nationwide Investment. Ed ' v ' P ' i . 6(2), p. 139-150. James, E. (1984). Benefits and Costs of Privatized Public Services: Lessons from the Dutch Educational System. W November, p- 585-605- Jamison, D., & M. Lockheed. (1987). Participation in schooling: Determinants and learning outcomes. Ec mi v m t a l n . 35(2). 279-306. Jenks, C., et al. 1972. Ipegpality; A Reeeeeeemep; of phe Effeep pf Eemily end geheeling ip Americe, New York: Basic Books. Jenks, C., et al. (1979). Whe gepe Ahead} The W. New York: Basic. 250 Kelly, C. P. (1987). Setting State Policy on Women's Education in the Third World: Perspectives from Comparative Research. Comparetive Edueapiep Review. 24(1). Kennedy, P. (1992). A Guide to Eeepemepriee. Third Edition. Cambridge Massachusetts: The MIT Press. Kohn, M.L. (1959). Social Class and Parental Values. Ameriean Journal ef §oeiplegy. Lanzas, A., & P. W. Kingston. (1981). English Achievement in Zaire: The Influence of Family Status and Residential Disruption. Qempepepive Egueeeien Review. 25 p 431-441. Lewis-Beck, S. (1980). A ' -R ' n- An Lpeppgpepiep. Beverly Hills: Sage. Lockheed, M. E., B. Fuller, & R. Nyirongo. (1988). Femily Ba k r n a u A ' v m . Washington D.C.: The World Bank Working paper. Lockheed, M. E., & D.J. Tamison. (1979). Seme_erepmipeppe f hoo ar ici a i n in Ter ' e ' Ne . Washington DC: World Bank, September. Lockheed, M. E., & E. Hanushek. (1988). Improving Educational Efficiency in Developing Countries: What Do We Know? Compare. 18(1), p 21-38. Lockheed, M. E., S. Vail, & B. Fuller. (1986). How Textbooks Affect Achievement in Developing Countries: Evidence from Thailand. Educati2n_Exaluetign_and 221122.5nelxeis. 8(4). p 379-392- Lulat, Y. G. N. (1982). Political Constraints on Educational Reform for Development: Lessons from an African Experience. gemparepi e Egpeetien review. June, p 219-235. Lupton, 9.3. (1983). Sele2teQ_H2me_and_femilx__£agt2re_ ' ' h A 'v . Unpublished Ph.D. Dissertation. University of Wisconsin-Madison. Lynch, P.O. (1986). From Quantity to Quality the Decline of Access to Schooling in Developing Countries in the 80's. Paper presented at the American Educational Research Association, San Francisco, April. Mankha, C.B. (1990). Edueetigp epd Seeial Q1aes _n—“\ -". . 'P’Ul. e ". 251 Representation in Botewene; A gage §pudy. Paper presented at the AERA in Boston, April, p. 1—23. Manley, M. K. (1963). Mental Ability in Jamaica: An explanation of the Performance of Children in the Jamaican Common Entrance Examination". §oeiel and Econemie Spudiee. No.12 (March), p. 51-72 Mayoya, M. (1989). The Private Coet pf Seeondepy Schogl Educatiep. Paper presented at African Educational Research Symposium in Ohio: Ohio University. Dec. p. 1-10. Mingat, A., & G. Psacharopoulos. (1985). Edpeegion Coepe Em" and Eineneing ip Afriee; geme Feepe epg Eeeeiple Lines ' Q£_A£L12E- Washington, D.C.: Education and Training Department, World Bank Report EDT13. Mwenene, M., et al. (1984). Education and Occupational , Attainment from Generation to Generation: The Case f study of Zaire. m 'v d ° v' w. ; February, p 52-68. Natriello, G., E. McDill, A. M. Pallas. (1990). Scheoling W. New York: Teachers College Press. Ndimira, P. F. (1995) Le Systéme Educatif et le Processus Démocratique au Burundi. In Les ri o iti Buppngi e; ep Rwenge (1222-1223). 139-157. Ndimurukundo, N. (1995). Scolarisation des Elites et Renforcement de la Conscience Ethnique. In Lee Criees WWI 125- 137. Ndarusigiye. M. (1982) W Amélierée ep ge 1' Aete;lie ep Milieu Rprel de Gihepa. Bachelor of Science Thesis. Bujumbura. Ndarusigiye. M. (1990) . W Beheyie;_in_fleppp§i. MSU Plan B paper for the Degree of Master's, 13. Niles, F. (1981). Social Class and Academic Achievement: A Third World Reinterpretation. m 'v E 'on gm. 25, p 419-430. Nkinyangi, J.A. (1982). Access to Primary Education in Kenya: The Contradictions of Public Policy". m 'v ' n v' w. June, p 180-199. 252 Noor, A. (1981). Steps Toward Implementing e Policy for Universal Basic Education. Washington, D.C.: World Bank, Education and Basic Human Needs, Working Paper No. 450, April. Neumann, P., & M. Cunningham. (1982). Mexico's Free Textbooks: Nationalism and Urgency to Education. Staff Working paper No. 541, Washington, D.C.: World Bank. Norusis, M. J. (1992). P a ' m s ' uid Version 5. SPSS Inc. Chicago, Illinois. Norusis, M. J. (1994). SPS§ Advaneed Spatistice 6.1. SPSS Inc. Chicago, Illinois. Nzamutuma, I. ( 1992). Family Background and Achievement: A Consideration of Financial, Human, and Social Capital Factors in the Context of Rwanda. Paper presented at the Annual Meeting of the Comparative and International Society, Annapolis, Maryland, March. Nzamutuma, I. ( 1992). Family Beekgppppd and Sppdepp Aehievement: A Coneidepation of Finaneial, Hpmen, and Seeiel Qapipal Fecpere in the Qonpez; pf Rwande. Ph. D. Dissertation. Michigan State University. Okwach, A. & S. Wamahiu. (1995). H us h l oBas Factors a d h o P rt fr m ggieeipg_§ppyeye.l Abridged Research Report 6. Academy Science Publishers. Nairobi, Kenya. Post, D. (1993). Educational Atainment and the Role of the 'State in Hong Kong. Cempepetiye Edpeetien Review. 26 (3) . August. Psacharopoulos, G. (1994). Returns to Investment in education: A Global Update. Wepld Develepmenp. 22(9): 1325 -1343. Psacharopoulos, G. & Woodhall, M. (1985). Edpeetien for Qeyelepmepp. New York: Oxford University Press, chap III. Population Reference Bureau. (1994). World Population Data Sheet (Broadside). Ram, R. (1982). Public Subsidization of Schooling and Inequality of Educational Access. gempepepiye Educatign_8exieg. 26(1). p 36- 47 Feb- “...l'hl'n ..‘ \' .I 9“. 253 Robinson, J. (1992). Social Status and the Academic Success in South Korea. Comparaeive Edueapiop geview. November. 516-530. Robinson, W. M., et al. (1986). Third Annual Repep; ef the Study of Q§AID contributions to the Egyptian Beeie Edueatiep Ppegrem. (Vol. II). Washington, DC: Creative Associates. Safilios-Rothschild, C. (1982). Adolescent Urban Girls in Developing Countries. Washington D.C.: International center for research on Women, Oct. Schultz, T. W. (1989). Investing in People: Schooling in low income countries. Eeepepiee pf Edpeapiop Review. 5 8(3), p 219- 223. i Schultz, T. W. (1981). Inv tin i T m' s of Pepulaeiep Qpalipy. Berkeley: Univesity of California Press. Schultz, T. W. (1972). Human Capital: Policy Issues and pf- Research Opportunities. prep_3eeeppeee. NBER. New York, Columbia University Press, p 1-84. Schultz, T. W. (1971). v m ' m Ca ' . New York: The Free Press. Chapter 3. Schwille, J., et al. (1990). I r eti ' A wa Waepefipl? New Depa end Unanewered Questiop. Paper presented at The Comparative and International Education Society, Anaheim, California (March). Sendazirasa, F. (1990). A propos de la Réforme de la Réforme. Le_gepepyeee_dp_§ppppdi. No.3282. Dimanche (September 2). Sewell, W. H. & R. M. Houser. (1975). Edgeepiepy 0 -9: '0! -90 - °=' 1; 17‘ ‘ ‘9 'n ‘ 34.. gepeep. New York. Academic. Sewell, W. H. & R. M. Houser. (1972). Causes and Consequences of Higher Education. Model of Status Attainment Process. m ' l Eeenemiee. 54, (5), Dec. p 851-61. Shea, B. M. (1976). Schooling and Its Antecedents: Substantive and Methodological Issues in the Status Attainment Process _exieu.2f_Edusatignal_Be§ear2h Fall, 46(4), 463- 526. 254 Smith, P.C., & P.P.L. Cheung. (1981). Social Origins and Sex-Differential Schooling in the Philippines. gempapetive Edueepiop Review. 25(1), p 28-42, Feb. Strong, F.M. (1989). Ending Hunger Through Sustainable Development. Jour a n i n D v 1 m n . Tan, J. (1985b). The Private Direct Cost of Secondary Schools in Tanzania. Ippernatippei gournal of Edueetienal deveiopment. 5(1), p 1-10. Thurow, L. (1969). Pove n D' 'm'n ion. W ' n, D. .: ' . ashingto C Brookings 5.. Thurow, L. (1975). gepezepipg_1pegpelipy. New York: Basic I Books. Tilak, J.B.G., & M.V. Varghesa. (1985). Discriminatory E pricing in Education. Occasional Paper No. 8. New Delhi: National Institute of Educational Planning and Administration. '1. p ‘1 Tilak, J. (1985). Ana 1 C o u a ' n in Ind'a. National Institute of Educational Planning and Administration (India), Occasional Paper No. 10. (ERIC Document Reproduction Service No. ED 261 946). Thobani, M. (1984). Charging User Fees for social Services: Education in Malawi. gemperegive Educapion Review. P 379-402. August. Tsang, M.C. (1988). Cost Analysis for Educational Policymaking: A Review of Cost Studies in Education in Developing Countries. BRIDGES Research Report Series, (3), October. Tsang, M. (1991). Resource Mobilization to Primary Education in Pakistan: An Exploration of Policy Options. College of Education, Michigan State University. Tsang, M., & S. Kidchanapanish. (1992). Private Resources and the Quality of Primary Education in Thailand. Internat1Qnal_J2urnal_gf_Educa119nal_8esearsh. P 179- 198. Tsang, M. (1994). Private and Public Costs of Schooling in Developing Countries. In Husen, T. & Postlethwaite, N. (eds )- Internat19nal_Ensxslsped1a_gf_fiducatign. Second Edition. Pergamon Press. P 4702-4708. 255 Vulliamy, G., K. Lewin, & D. Stephens. (1990). Doing Educational Research in Developing Countriee: Qualipapive §pra§egies. New York: The Falmer Press. Waweru, J. M. (1982). Socio-Economic background as an Influence Factor in Pupils' Achievement in Primary Schools in Embu District, Kenya. Afpigan Studiee in Curriculum Develepment and Evaluation. (69) African Curriculum Organization. Nairobi: Kenya Institute of Education. Weinsten, W. (1976). Politieal Confliet and Ethnic S r i s: A c e t f uru 1. New York: Syracus, Syracus State University (Maxwell School of Citizenship and Public Affairs), p1-13. ‘54- Weis, J. (1979). Education and the Reproduction of Inequality: The Case of Ghana. m a iv d i n Review. P 41-51, February. li‘rfl <:ms.u.:_ er .W’l’ Weisbrod, B. A. (1971). Investing in Human Capital. In Ed 1 a th Ec nomi o Huma Ca i . New York, The Free Press. 68-88. (f= Windham, D. (1984). Financing Educational Development. gemperapive Edueegiep Beview. p. 660-661.l Windham, D. M. (1991). The Role of Basic Education in Promoting Development: Aggregate Effects and Marginalized Population. In D. Chapman, & H. Walberg. Strapegiee in Enhancing Edueetional Ereducpivity. JAI Press, forthcoming, chapter 5. Wolff, L. (1985). gepppelling phe geepe pf Eddeepien in Eaeeepn Afgiee; A Review pf Dape, Iseuee end Eelicies. Washington, D.C.: World Bank, Staff Working paper No. 702. Wolfe, B., & J. Behrman. (1984). Who Is Schooled in Developing Countries? The Roles of Income, Parental Schooling, Sex, Residence and Family Size. Eeenomice pf Eddeapien Review. 3 (3), p 231-2245. The World Bank. (1985). Education Strategies for Sub-Saharan Africa. Draft Report. Washington, D.C. Ziarati, S. (1989). An __alxae_Queratignelle_du_Sxetene Edd§e§i£_§d;dpdei§. (UNESCO Project BDI/86/010). Bujumbura: UNESCO. ) el \II )I. \ll‘l’t lll‘lll‘ “Ill .I’Ale I“é’£{.éhl “V...