THE FERTILITY BEHAVIOUR OF MUSLIMS OF SRI LANKA
By
SHAHUL HAMEED.HASBULLAH
B.A. (Hons.), University of Peradeniya, Sri Lanka, 1975
A THESIS SUBMITTED IN PARTIAL FULFILMENT OF
THE REQUIREMENTS FOR THE DEGREE OF
MASTER OF ARTS
In
THE FACULTY OF GRADUATE STUDIES
Department of Geography
We accept this thesis as conforming
to the required standard
THE UNIVERSITY OF BRITISH COLUMBIA
October 1984
(c) Shahul Hameed Hasbullah , 1984 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission.
Department of -g-va "TOO /^Y» Vv^
The University of British Columbia 1956 Main Mall Vancouver, Canada V6T 1Y3
Date - ) o - ,
DE-6 (.3/81) i i
ABSTRACT
Although fertility rate is comparatively higher among
Muslims than other religious groups in Sri Lanka, the reason for this high fertility rate has, not been fully investigated.
This study attempts to determine what factors influence Muslim fertility using individual household data. The data are derived from a systematic interview of 323 household heads in eleven
Muslim settlements in Sri Lanka in 1981.
Using crosstabulation and path analysis, the study ..found that socio-economic and demographic factors such as female level of schooling, female age at marriage, socio-economic status, and child mortality are closely associated with the fertility variation of the surveyed population. A woman with little education is more likely to marry early, is engaged in household activities, and may lose many children, all of which results in high fertility. The path analysis further suggests that low female schooling directly and indirectly through age at marriage "causes" high fertility of the surveyed population.
The female level of schooling is found to be low among Muslims; therefore fertility is comparatively higher. However, the youn• ger women have shown comparatively lower fertility due to recent educational changes.
The low aspiration level for education among this group is mainly due to the bitter experience of the colonial educa• tion system which lasted until 1948 and the comparative lack of educational opportunities in the present. Increasing female education and the improving socio-economic status of the female iii population may reduce the present high fertility level of this community in the future. iv
TABLE OF CONTENTS
Page
Abstract ' ii-iii
Table of Contents iv-vii
List of Tables viii-xii
List of Figures xiii-xiv
Acknowledgements xv
Chapter One: Introduction 1- 9
1.1 Socio-Economic and Historical Background 2- 9
Chapter Two: Theories of Fertility Decline and a 10- 33
Conceptual Framework for the Study
of Fertility Behaviour of the Muslims
of Sri Lanka
2.1 A Review of Theories of Fertility Decline 10- 20
2.1.1 Malthusian and Neo-Malthusian 11- 12
Perspectives
2.1.2 The Theory of Demographic 12- 17
Transition
2.1.3 Recent Approaches to Fertility 17- 20
Decline
2.2 A Conceptual Framework for the Fertility 20- 33
Behaviour of the Muslims of Sri Lanka
2.2.1 Religion and Fertility 23- 26
2.2.2 Female Status and Fertility 26- 29
2.2.3 Socio-Economic Status, Family 29- 30
Form and Fertility 2.2.4 Infant and Child Mortality 30- 31
and Fertility
2.2.5 An Analytical Framework for 31- 33
the Fertility Behaviour of
Muslims of Sri Lanka
Chapter Three: The Recent Fertility Trends in 34- 56
Sri Lanka and of Sri Lankan Muslims
3.1 Recent Fertility Trends in Sri Lanka 34- 41
3.1.1 Population Growth 34- 35
3.1.2 Recent Fertility Trends in 35- 38
Sri Lanka
3.1.3 The Determinants of Fertility 39- 41
Decline
3.2 Fertility Trends Among Muslims 41- 56
3.2.1 Muslim Population Growth 42- 44
3.2.2 Fertility Trends Among Muslims 44- 50
3.2.3 Determinants of Muslim Fertility 50- 56
Chapter Four: Methodology 57- 64
Chapter Five: Study Area 65- 86
5.1 Muslim Settlement and their Religio- 65- 69
Cultural Enviroment
5.2 Surveyed Muslim Settlements 69- 72
5.3 Surveyed Households 72- 78
5.3.1 Head of Household 73- 75
5.3.2 Family Members 75- 76
5.3.3 Socio-Economic Status 76- 78
5.4 The Determinants of Fertility 78- 83 VI
5.5 Variables and their Measurements 83- 86
Chapter Six: Variation in Fertility by Geographic 87-106
and Socio-Economic Factors
6.1 Variation in Geographic Locations 87- 91
6.2 Education and Fertility 91- 94
6.3 Female Age at Marriage and Fertility 94- 96
6.4 Mean Birth Interval and Fertility 96-100
6.5 Child Mortality and Fertility 100-102
6.6 The Findings from Pearson and Partial 102-104
Correlation
6.7 The Summary of Findings 104-106
Chapter Seven: Multivariate Analysis of Fertility 107-121
7.1 A Predictive Model of Fertility 113-114
7.2 Fertility Behaviour of Women Aged 49 114-118
Years and Above
7.3 Fertility Behaviour of Women Aged 48 118-121
Years and Below
7.4 Summary 121
Chapter Eight: Conclusion 122-127
8.1 General Characteristics 122-123
8.2 Major Findings 123-124
8.3 Findings from Multivariate Analysis 125-127
Bibliography 128-137
Appendix 1: Muslim Settlements in Sri Lanka by 138-144
Province and District
Appendix 2: Questionaire: Fertility Survey -- 145-152
Muslims of Sri Lanka vii
Appendix 3: Selected Socio-Economic Character- 153-155
istics of Eleven Surveyed Muslim
Settlements
Appendix 4: Specific Forms in which Variables 156
are Coded and Analysed in the Path
Analysi s viii
LIST OF TABLES
Page
Table 1.1 The Level of Education by Ethinc 8
Groups
Table 3.1 Fertility Indicators (Sri Lanka) 36
1953-1974
Table 3.2 Percentage of Women Currently 39
Married in 1953, 1971 and 1981
Table 3.3 Determinants of Fertility in 42
Sri Lanka
Table 3.4 Population Growth in Sri Lanka: 43
Muslim Growth in Perspective
Table 3.5 Percentage Change in Population 47
by Religious and Ethnic Groups,
1911-1981
Table 3.6 Crude Birth Rates and Child-women 48
Ratio by Ethnic Groups 1946-1977
Table 3.7 Percentage of Change Between 1960 50
/65 and 1970/75 in Age-specific
Fertility Rates by Ethnicity and
Religion
Table 3.8 Five-Year Average Crude Death Rates, 52
1911-1975
Table 3.9 Infant Mortality by Ethnic Groups, 53
1910-1964
Table 3.10 Percentage of Female Level of 54
Schooling by Ethnic Gruops ix
Table 3.11 Mean Number of Live Births per 56
Ever-Married Women Aged 45-49 by
Socio-Economic Status and Ethnic
Groups
Table 4.1 Selected Sample Muslim Settlements 61
Along with other Background
Information
Table 5.1 Number of Surveyed Households and 73
Surveyed Population According to
the Settlements
Table 5.2 Socio-Economic Status of the 79
Surveyed Households
Table 5.3 The Number of Total Pregnancies 79
of Surveyed Population
Table 5.4 The Wife's Level of Schooling 81
Table 5.5 The Wife's Workforce Participation 84
Table 5.6 Infant and Child Mortality 84
Table 5.7 Variables and Measurements Used in 85
the Analysis
Table 6.1 Mean Number of Live Births per Ever- 89
Married Women by Women's Age and
Geographical Location
Table 6.2 Mean Number of Live Births per Ever- 89- 90
married Women by Women's Age at
Marriage, Level of Schooling, Socio•
economic Status and Geographical
Locat ion Table 6.3 Mean Number of Live Births per Ever- 90
Married Women by Wife's Level of
Schooling and Sub-cultural Regions
Table 6.4 Mean Number of Live Births per Ever- 92
married Women by Women's Present Age,
Age at Marriage, and Level of
Schooling
Table 6.5 Mean Number of Live Births per Ever- 93
married Women by Social Group and Level
of Education
Table 6.6 Mean Number of Live Birth per Ever- 94
married Women by Wife's Workforce
Participation, Husband's Occupation
and Level of Schooling
Table 6.7 Mean Number of Live Birth per Ever- 95
married Women by Women's Present Age,
Mean Birth Interval and Age at
Marr iage
Table 6.8 Mean Number of Live Births per Ever- 97
married Women by Socio-Economic Status
and Age at Marriage
Table 6.9 Mean Number of Live Births per Ever- 97
married Women by Women's Present Age
and Mean Birth Interval
Table 6.10 Mean Number of Live Births per Ever- 98
married Women by Background Variables
and Mean Birth Interval xi
Table 6.11 Mean Number of Live Births per Ever- 100
married Women by Percentage of Male
Births, Child Mortality, and Mean
Birth Interval
Table 6.12 Mean Number of Live Births per Ever- 101
married Women by Present Age, Age at
Marriage, Social Status and Child
Mortality
Table 6.13 The Relationship Between Selected 103
Background Variables and Number
of Live Births
Table 6.14 The Relationship Between Selected 103-104
Background Variables and Number
of Live Births When Controlled for
Women's Present Age
Table 6.15 Mean Number of Births per Ever- 104
married Surveyed (Women) Population
Table 6.16 Specific Influence of Certain 106
Variables on Fertility of Muslims
of Sri Lanka
Table 7.1 Path Coefficients for Model Predicting 112
Fertility Behaviour of Muslim Women in
Sri Lanka (323 cases)
Table 7.2 Path Coefficients for Model Predicting 117
Fertility Behaviour of Muslim Women-
49 Years and Above (67 cases) xii
Table 7.3 Path Coefficients for Model Predicting 120
Fertility Behaviour of Muslim Women-
48 Years and Below (256 cases) x i i i
LIST OF FIGURES
Page
Figure 1.1 Distribution of Muslim 3
Populat ion
Figure 2.1 Schematic Diagram Indicating 32
the Relationship of Selected
Variables in Fertility
Behaviour
Figure 3.1 Age-Specific Fertility Rates- 37
1953, 1963, 1971 and 1974
Figure 3.2 Age-Specific Marital Fertility 38
Rates- 1953, 1963, 1971 and
1 975
Figure 3.3 Trends in Population by Reli• 45
gious Groups
Figure 3.4 Trends in Muslim Population 46
Figure 3.5 Trends in Birth and Death Rates 51
(Muslims)- 1900-1981
Figure 4.1 The Location of Muslim Settle• 59 ments
Figure 4.2 Settlements Surveyed 60
Figure 5.1 Cultural Regionalism and Muslim 67
Distribution
Figure 5.2 Socio-Economic and Cultural 71
Characteristics of the
Surveyed Muslim Settlements xiv
Path Diagram of Children Ever-
Born and Other Socio-Economic and Demographic Characteristics
Path Diagram of Children Ever-
Born and Socio-Economic and
Demographic Characteristics
(Ever-Married Women)- 1981
Path Diagram of Children Ever-
Born and Socio-Economic and
Demographic Characteristics
(Ever-Married Women Aged 49 and Above)- 1981
Path Diagram of Children Ever-
Born and Socio-Economic and
Demographic Characteristics
(Ever-Married Women Aged 48 and Below)- 1981 XV
ACKNOWLEDGEMENTS
I would like to thank my supervisor, Professor J. S.
Duncan and my advisors, Professors B. M. Morrison (Asian Stu• dies), N. E. Waxier (Health Care and Epidomology), and T. G.
McGee (Geography) for their assistance and encouragement during
the preparation of this thesis. I would also like to acknow•
ledge the help of the members of the Muslim Majlis of the
University of Perideniya (1980/1981), Messers. Zanheer, Munou-
ver, Mohamed Ali, Hameed, Majeed, Ahamedlebbi, Yakoob, Lathiff,
and Noufal and Ms Azeema and Naleefa for their help and co•
operation during the field survey. I am also indebted to Mr.
Rajesh Chandra for his helpful discussions and suggestions. I
would further like to acknowledge the help provided by Mr.
Naoki Kanai, Mrs. Vivian Howard, Dharma Chandra, and Ms Sandy
Pan. Finally, I would like to thank numerous friends in Sri
Lanka and at the University of British Columbia for their help
and encouragement during the preparation of this thesis. 1
CHAPTER 1
INTRODUCTION
This study is an analysis of the fertility behaviour of
Sri Lankan Muslims, whose fertility rates are higher than the other religious and ethnic groups in Sri Lanka. It uses syste• matic interviews of 323 household heads in eleven Muslim set• tlements in Sri Lanka. The study focuses attention on the
socio-economic and demographic characteristics which influences directly and indirectly their fertility.
This study has the following objectives : (1) To study
the fertility differential of the surveyed population by socio• economic and geographical factors; (2) To identify the determi• nants of fertility; and (3) To find out which combination of
factors is responsible for high fertility among Sri Lankan
Muslims.
This thesis is composed of eight chapters. Chapter 1 provides a historical and socio-economic background of Sri
Lankan Muslims. The literature on fertility decline is review•
ed in Chapter 2, which also proposes a tentative analytical model for the study. Chapter 3 presents the trends of recent
fertility decline in Sri Lanka in general and of Muslims in Sri
Lanka in particular using available national statistics. Chap•
ter 4 provides information about the sampling design, intervie•
wing, and the organization and the analysis of the collected
data.
Chapter 5 provides information on socio-economic and 2 demographic characteristics of the surveyed population. This chapter provides a detailed description of the measurements of selected variables which are used in the present study. Chapter
6 presents the analysis of the variation in Muslim fertility by geographical and socio-economic factors and summarizes the major findings of cross tabulation and correlation analyses.
The findings of the path analysis are reported in Chapter 7 where the determinants of fertility are identified. Finally
Chapter 8 integrates all the findings of the previous chapters and presents the conclusion of the present study.
1.1 Socio-Economic and Historical Background
Muslims are the followers of the Islamic faith. Islam
means surrendering to God's will. The profession of the faith
(Shahada) of Islam is : "there is no god but god and Mohammed
(peace upon him) is the last prophet (Encyclopaedia Britanica,
1973: 666-7). The Sri Lankan Muslims belong to the Sunni divi•
sion of Islam.
Muslims are the second largest minority group in Sri
Lanka, and according to the 1981 census, constitute 7.6 percent
of the total Sri Lankan population. Figure 1 shows the geograp•
hical distribution of the Muslim population. They are thinly
spread throughout the island and are to be found equally in
areas where the majority community, the Sinhalese, are settled
as well where the principal minority community, the Tamils are
concentrated. The majority of the Muslims in both areas speak
Tamil.
The geographical distribution of the Muslim population 3
SOURCE: Population Census, 1981. 4 is reflected in their economic activities as well. The majority of the Muslims in the dry zone provinces (Eastern, Northern, and North Central) are mainly engaged in rice cultivation. The economic activities of the Muslims in the rest of the country vary from saleswork to small and large scale business activi• ties. According to the World Fertility Survey (1975), one of the single largest economic activities of the Muslims is sales- work which comprise nearly 26 percent of the total workforce.
Saleswork refers to work as a labourer in large and small scale private business firms. Economically, salesworkers are poor and dependent on unreliable employment. The majority of them live in urban and semi-urban settlements. A very small percentage of
Muslims are engaged in medium and large scale business activity and they are wealthy urban dwellers.
Ethnically, Sri Lankan Muslims are divided into three groups : the Sri Lankan Moors, the Indian Moors and the Malays.
The Sri Lankan Moors constitute nearly 93 percent of the total
Muslim population. They are the decendants of long established
Muslim migrants from the Arabian peninsular. The term "Moor" was mistakenly given to the Muslims by the Portuguese (Ragha- wan, 1964 :19) ; however, the term has assumed a traditional significance from its long usage. The Indian Moors are relati• vely recent migrants from South India who came mainly as estate labourers and urban entrepreneurs during the late nineteenth and the early twentieth century. They constitute nearly 3 percent of the total Muslim population. The Malays, who com• prise nearly 4 percent of the total Muslin population, trace 5
their ancestry to South East Asia, particularly, Java, Sumatra and Malacca (Hussainmia, 1982). They settled in the country during the period of Dutch rule around the eigtheenth century.
Very little is known about the history of Muslims (Sri
Lankan Moors) because of the lack of research in this area
(Mauroof, 1974 and Samaraweera, 1978). According to some sour•
ces (Johnstones, 1876), Muslims are the decendants of the Arab
traders who established eight settlements along the west coast
of Sri Lanka at the beginning of the eighteenth century. During
the early years, the Muslims numbered a few thousand (Gunawar-
dena, 1959-60 :83) and for obvious reasons, they were predomi•
nantly males (Arasaratnam, 1964 :119). They married among the
local people — both the Sinhalese and the Tamils. The preval•
ence of the Tamil language among Muslims is due to the supposed
importance of Tamil in South-east Asian trade between' the
twelvth and fourteenth centuries and the extensive practice of
inter-marriage that took place with Tamil converts (Arasarat•
nam, 1964 :120-21). An alternative view traces the immediate
ancestry of Sri Lankan Muslims to earlier Arab settlement in
Tamil speaking Kayalpattanam in South India (Ramanathan, 1887).
The Muslims integrated themselves with the local people
and participated in the socio-economic and political develop•
ment of the country. They conducted external trade and thereby
brought Sri Lanka closer to the known world. They also extended
internal trade and developed transportation and communication
between trading centres. They provided armed security to the
Sinhalese king and protected the sovereignty of the country 6 against the foreign aggressors. (Arasartnam, 1964 :119). There is no evidence of political domination by Muslims of the non-
Muslim native people in Sri Lanka (Mauroof, 1974 :68).
Colonial rule under the Portuguese, the Dutch, and the
British, for nearly 450 years had a great impact both socio- economically and religiously of the Muslim population. The
Portuguese viewed the Muslims as their religious and economic enemies and therefore restricted their activities in both of these realms. These events led to a large Muslim migration
(beginning of 17th century) to the Central highlands and East coast which at that time were controlled by the Kandyan king.
The Muslims who settled in the Kandyan areas soon extended their peddling trade all over the Kandyan kingdom and assisted the king in developing his trade through the ports that were still left in his hand. Others who chose to settle in the east coast of the island became farmers and established a rural tradition (Arasaratnam, 1964 :122).
During the Dutch and the British periods (1658-1948), the religious and cultural identities of the Muslims were further threatened. In the nineteenth century,the Muslims like other religious groups, too faced the challenge of Protestant
Christianity, but resisted completely, albeit at the expense of social and economic advancement (De Silva, 1974 :101).
English education was a primary vehicle for conversion to Protestanism under British rule (Samaraweera, 1978 :470) because education was largely Christian in content (De Silva,
1974 :101). English education was also an essential prerequiste 7
for entry into highly valued and prestigious employment. The
Muslims, unlike other religious groups, did not take advantage
of the new educational structure established by the colonial
government and the missionaries even though this meant sacrifi•
cing material benefits that english education brought (De Sil-
va, 1974 :101; Samaraweera, 1978 :470). Instead, the Muslims
fulfilled their essential educational needs in the Madrasas,
the traditional Muslim educational institution, which provided
only rudimentary instruction (Azeez, 1971). As a result, they
became one of the least educationally advanced group in the
country.
The Buddhists and the Hindus saw the need to create
their own educational institutions which would provide modern
education not only to counter Christian missionary activities
but also to provide their people with access to the channels of
mobility opened up by the British. To some extent, the Bud•
dhists and the Hindus succeeded in their efforts but Muslims
were not as successful (Samaraweera, 1978 :476) and the majori•
ty were left without any development in education.
The situation remained unchanged and the gap in educa•
tion and other social development widened between Muslims and
other religious groups until around 1960 (Jennings, 1944; Tam-
biah, 1955). However, some concessions in education had been
made in recent years to Muslims partly in recognition of the
fact that they trailed the other ethnic and religious groups in
Sri Lanka in education. In these efforts, Muslim schools had
been upgraded and brought under a separate administration under 8 the Ministry of Education. Muslims have also been attracted to higher education with the provision of employment as school teachers. These and other attempts have, to some extent, im• proved their level of schooling and the impact could be seen in improving socio-economic status and demographic patterns in recent years. However, as Table 1.1 shows, Muslims are still well behind in education compared to other religious and ethnic groups.
TABLE 1.1
THE LEVEL OF EDUCATION BY ETHNIC GROUPS
Ethnic % of Pop. % of Illiteracy % of Univer• % of Group (1981 ) (1981) sity Enrolment Medical (1981 ) Doctors (1973)
Sinhalese 74 11.6 79.0 57.4 Sri Lankan Tamil 12.6 13.4 17.1 36.7 Sri Lankan Moor 7.1 20.7 3.3 2.3 Indian Tamil 5.6 33. 1 - - Others 0.4 8.5 0.6 3.7
SOURCE: For population, Census of Population and Housing, Sri Lanka, 1981; for illiteracy, Census of Population and Housing (10 percent sample), Sri Lanka, 1981; for univer• sity enrolment, Annual Reports of University Grants Commission, 1981; for doctors, Census of Doctors, 1973.
The majority of Muslim children, both males and females, leave school before reaching secondary schools. The illiteracy rate among Muslims is still much higher than among many other religious and ethnic groups. Very few Muslim males or females enroll in higher education (Table 1.1). The enrolment of Muslim 9 students in science faculties is even lower. The low level of aspiration to and enrolment in higher education among Muslims is due to a lack of educational facilities in Muslim schools and to the general socio-economic conditions of the Muslim population.
Low educational level and related socio-economic under• development of Muslim society is reflected in their fertility behaviour as well. Many recent studies have found that the high fertility rate of Muslims is strongly correlated with low level of schooling (Fernando, 1982), high infant mortality, low fe• male age at marriage (World Fertility Survey, 1975) and low socio-economic status (Hanna and Nadarajah 1975). This study will explore the fertility patterns of Muslims and will identi• fy which combination of factors "causes' high fertility among the Muslim population. 10
CHAPTER 2
THEORIES OF FERTILITY DECLINE AND A CONCEPTUAL FRAMEWORK FOR THE STUDY OF FERTILITY BEHAVIOUR OF THE MUSLIMS OF SRI LANKA
The explanation of fertility decline in both developed and developing countries has undergone fundamental changes in recent decades. Until very recently, the conventional popula• tion theories were much simpler than those of today because the available facts were few; as a result, it was possible to define simple coherent theories to explain the population phe•
nomenon. In recent years, however, new data on both the histo•
rical demographic experience of the West and the recent ferti•
lity trends in the developing countries have brought into
question the conventional theories. In particular, the theories
have failed to explain under which circumtances the fertility
of a population falls from a high to a low level. In recent
years, various disciplines have responded to the new challenge
to explain fertility behaviour. However, so far no common
ground has been reached by these attempts. The purpose of this
chapter is, therefore, first to review the major theories of
fertility decline and second, to formulate an analytical frame•
work for the study of fertility behaviour of the Muslims of Sri
Lanka.
2.1 A Review of Theories of Fertility Decline
The main theories of fertility decline include Malthus
and other historical thinkers, the theory of demographic
transition, and recent macro and micro approaches (Burch, 11
1980; Srivastava, 1980; Keyfitz, 1972).
2.1.1 Malthusian and Neo-Malthusian Perspectives
The Malthusian perspective derives from Thomas Robert
Malthus' essay on population published in 1798 and its subse• quent editions over more than thirty years. Malthus proposed that while a population tends to increase in a geometric or exponential fashion, its subsistance can at best increase in an arithemetic fashion. A growing population, sooner or later, will exceed the capacity of its resources to sustain it (Mal• thus, 1798). Thus Malthus argued that fertility will be checked by what he called " moral restraint" - the postponement of marriage combined with "virtuous" behaviour on the part of unmarried women (Malthus, 1830:153). In effect, he proposed
that fertility be lowered by reducing the number of adults of child bearing age living in sexual union. Low fertility caused
by the reduced proportion of married women can be called "Mal•
thusian low fertility". On the other hand, neo-Malthusians have
been arguing in recent decades that low fertility can be achi•
eved by modern birth control methods (Ehrlich, 1971 and Hardin,
1968).
Malthusian and neo-Malthusian perspectives have been
tested in different circumtances (Coale, 1965; Kleinman, 1980).
For example, in 1965, Ansly J. Coale analysed the 19th century
fertility behaviour of European countries using special ferti•
lity indices. In his study, he did not find any evidence of
Malthusian fertility decline. Instead, he found that the reduc- 1 2
tion in the proportion of marriages had varied from country to country and from time to time. In addition, the development of
late marriages and abstinence from marriages, Coale suggested, may have originated from the higher standard of living and
special prevalence of the "stem-family" in 19th century Western
Europe, Coale (1965: 15) concluded that
since [birth control methods] have long been known in every society, the part they played in fertility reduction implies the importance of changed atti• tudes or motives rather than the discovery of new devices or techniques.
discounting thus the significance of neo-Malthusian theories to
explain low fertility.
Attempts to form an alternative to Malthusian and neo-
Malthusian perspectives have resulted in the development of
several prominent population theories. Among them, the theory
of demographic transition is one of the important attempts in
modern times to synthesize present, past and future human
fertility decline.
2.1.2 The Theory of Demographic Transition
The theory of demographic transition describes the chan•
ges that take place in birth and death rates as a population
passes from traditional or pre-modern, social and economic
conditions to an urbanized and industrialized modern society.
The theory was based on two observations: first that fertility
and mortality are high in traditional societies and second that
every modern society has passed from high to low rates of
fertility and mortality . In addition, the phenomenon of demog•
raphic transition was thought to occur in three stages: (1) a 1 3
balance stage of high potential growth when birth and death rates are high; (2) a transitional stage of rapid growth when the death rate is low but the birth rate continues at a high level; and (3) a new balance stage when birth and death rates are low, with the potential for a declining population (McNama- ra, 1982:146).
About two hundred years of demographic experience of the
West European countries has resulted in this three stage model of demographic transition. Thus, Stolnitz stated that " demog• raphic transition ranks among the most sweeping and best docu• mented historical trends of modern time" (Stolnitz, 1964:31).
Moreover, the theory also strongly suggests that the process which took place in Europe will also take place in developing countries (Teillbaum, 1975:420). There have been many attempts
(Trewartha, 1969; Clark, 1965 and others) in recent decades to apply and prove the transition theory in the developing coun• tries where a sudden demographic change has occurred during the
last thirty years.
2.1.2.1. The Relevance of Transition Theory for West European Countries
Recent research findings (Coale, 1973 and 1979; Walle and Knodel, 1977; Loschky and Wilcox, 1974) have raised seri• ous questions regarding the validity of transition theory and are forcing a critical re-examination of the assumption on which the theory is based and the conceptual approach to which
it has given rise. First of all, the recent observation and analysis of the fertility behaviour of 19th century Europe 1 4
where the transitional theory based its universal principle of fertility behaviour has raised many fundamental questions rega• rding the validity of the theory. The findings of this re• examination of European historical fertility can be summarized into three major conclusions.
First, contrary to the transition theory, the new fin• dings indicate that there was a wide variation in fertility and mortality levels between regions and countries of 19th century
Europe. For example, fertility varied considerably from provin• ce to province and from country to country (Coale, 1973:62-63).
In addition, fertility began to decline prior to or simulta• neously with the decline in mortality (Knodal, 1979). Secondly, the timing of the fertility transition was confined to a rela• tively short period and only loosely related to supposed under• lying socio-economic factors. For example, in some areas, such as parts of France, fertility began to decline before the spread of industrialization. In addition, certain socio-econo• mic factors were associated with (but not necessarily caused by) demographic changes in some countries but in others these relationships were not apparent at all (Teitelbaum, 1975:421-
22). Finally, often socio-economic factors seemed less closely related to regional fertility patterns than linguistic, cultu• ral, and ethnic differences (Coale, 1979). Thus, Walle and
Knodal (1977:53) concluded that "if conventional statements of transition theory are inappropriate for the European experien• ces, on which they profess to be based,they are even less likely to be useful in predicting or explaining future transi- 1 5
t ions".
2.1.2.2 The Relavance of Transition Theory to
Developing Countries
The term "developing countries" refers to non-industrial low and middle income countries (International Encyclopedia of
Population, 1982: 147). Three geographical areas -- Latin Ame• rica, Africa and Asia (excluding Japan and U.S.S.R — contain• ing more than 70 percent of the total world population fall into the above category (Clark, 1970: 7-24). After the long colonial experience, many countries have recently received their political independence. They experienced a marked change in death rate after their independence which declined from 23 to 13-14 per thousand population (Mauldin, 1981:72) while fer• tility remained high, resulting in a sudden population growth.
The results were that more than half of the total population was under the age of eighteen. Subsequently, other socio• economic problems have been exacerbated by the sudden popula• tion growth. There have been many attempts to ease the popula• tion problems by both individual governments and international organizations. Family planning programmes have been activated to attempt to achieve the "optimum population". The failure of these activities has been blamed on the religious and cultural attitudes of the people. However, even in the absence of econo• mic development, the fertility rates in recent years in many developing countries appear to be falling at a much faster rate than that experienced by West Europe during the demograp• hic transition (Kleinman, 1980:190 and Coale, 1983). 1 6
The application of modern techniques of demographic and social measurements provided better opportunities to validate the transition theory. The evidence indicates that the transi• tion theory does not apply in developing countries. The reason is that the demographic evolution of the developing countries
is fundamentally different from the experience of the developed nations upon which the theory is based. For example, one of the preconditions for mortality decline, according to the theory,
is favourable economic and social development (Peterson,
1969:576 and Stolnitz, 1964:31). The decline in mortality of
European countries was gradual and was generally related to the
socio-economic forces of development and industrialization
(Walle and Knodal, 1980:24-26). Contrary to this, in many developing countries, mortality declines have been more drama•
tic and, in most instances, have been associated with high and even rising birth rates (Arriaga, 1970:218-22). Moreover, the
rapid decline in mortality has resulted largely from imported
technologies.
On the other hand, the strength of the transition theory
lies in the fact that with sufficient modernization, fertility
and mortality will change in a predictable manner (Internatio•
nal Encyclopaedia of Population 1982). The process of moderni•
zation, however, the theory insists, is through industrializa•
tion and urbanization. The recent evidence from the developing
countries, on the other hand, indicates that the fertility
decline has been achieved in many rural agrarian countries in
the Third World without those above mentioned preconditions For 17
example, it has been reported that fertility has declined by 42 percent in China in recent years (Coale, 1983 :830); Kerala and
Sri Lanka have managed to lower their birth rate within a relatively short period of time (Wright, 1982 and Ratcliffe,
1978) ; and finally, Bali, which is also rural , poor and low in literacy and with health conditions far worse than those of
China, Kerela, and Sri Lanka, has achieved considerable ferti• lity decline in recent years (Freedman, 1979 :12).
Finally, the theory claims that fertility is high in poor, transitional societies because of their "attitudes, be• liefs, traditions and irrationalities" (Caldwell, 1982:119).
Thus, the theory claims that it is rational to have fewer children, and that this rationality comes from industrializa• tion and urbanization. In contrast to the arguments, recent micro-level fertility behaviour studies (Caldwell, 1982) in
rural areas of many developing countries indicate that rural agrarian societies are economically and socially rational in
their own fertility behaviour. Thus, Caldwell (1982: 120) con• cluded that the demographic transition theory has failed to explain the reality in the developing countries because "the criteria employed are highly ethnocentric and are laden with
Western values".
2.1.3 Recent Approaches to Fertility Decline
The challenges from the failure of conventional theo•
ries have been responded to by two major approaches in fertili•
ty behaviour studies: macro and micro-level approaches (Bird- 18 sail, 1982:241).
2.1.3.1 The Macro-Level Fertility Behaviour Approach
A seminal contribution to the analysis of the economic- consequences of population growth was the work of Ansley J.
Coale and Edgar M. Hoover in 1958, in which they constructed a mathematical model of the economy of India. The main conclusion of this study was that national economic growth is impaired if the fertility level greatly exceeds the mortality level. Maul- ding and Berelson (1978) have claimed that certain indicators of advanced developments are correlated with low rates of fertility, high rates of literacy, high per capita consumption of energy, high rates of urbanization, and low rates of infant mortality. Thus, they suggested that government support for a family planning programme, which is part of the result of the socio-economic progress in developing countries, can increase the influence of such progress in reducing fertility.
Finally, the threshold hypothesis which was initiated by the United Nations in 1965, directed the attention of research to the prime importance of the changing conditions that led to fertility decline at a point identified as the "threshold".
This hypothesis is an attempt to link the demographic transi• tion with social change and economic development. According to this hypothesis, fertility is initially high and improving economic and social conditions are likely to have little if any effect on it until a certain economic and social level is reached; but once that level is achieved, fertility is likely to decline and to continue downward until it is again establis- 19
hed on a much lower level (Sri Karan, 1982).
2.1.3.2 The Micro-Level Fertility Behaviour Approach
The macro-level approach has been criticized as too general and limited to explain or predict population growth with any degree of specificity. Thus, it has been suggested that the fertility analysis should deal with individuals or households as units of analysis as well as with nations, states, countries, or cities (Burch, 1980:2). Firstly, micro-economists argue that the changes in household fertility as a function of changes in the families' economic situation are attendent upon socio-economic development. They treat the child as both a produced and a consumer good. They say, for example that
fertility is the result of rational economic choice within the household and because children are assumed to be non-inferior goods, increase income increases the demand for children for the parents. Child services are produced in the household, though inputs of parent's time goods brought in the market such housing, formal education and health services. Children may also be an investment, short term if they work during their childhood, long term if they support parents in old age (Birdsall, 1982: 246).
Secondly, Freedman (1962), Hawthorne (1970), Davis and
Black (1956) and others have advanced theoretical frameworks for analysing fertility decisions and behaviour which are esse• ntially sociological in character. The sociological approach focuses on the group, an extended family or clan, a social class, or society as a whole (Robinson and Harbison, 1980:221).
Caldwell (1982) , on the other hand, develops a new "general theory" of fertility which stresses the essential role played 20
within the family by inter-generational "net-wealth flows" and the sharp differences in pre-transition and post transition demographic regions.
Finally, the focus of the anthropologist has been the analysis of kinship, marriage patterns and the systematic con• sideration of the impact of social structural patterns on fertility behaviour. Psychologists, on the other hand, focus on individual needs, some biological or innate, some socially conditioned by interactions with other individuals (Robinson and Harbison, 1980:212-221). Recently, an attempt has been made by Robinson and Harbison (1980) in order to integrate the different micro-level fertility analysis approaches.
I am convinced that micro studies which focus on the household, attempting to find causes of fertility variation, offer considerable hope for understanding the dynamics of Third world populations and deserves further encouragement. The next
section discusses such a general framework for a study of
fertility among Muslims in Sri Lanka.
2.2 A Conceptual Framework for the Fertility Behaviour of the Muslims of Sri Lanka
The level of fertility is comparatively high in many developing countries. The conventional theories of fertility decline have failed in their explanations because they believe
that high fertility results from irrational attitudes of pea•
sants in developing countries. It has been suggested in recent years that there is no particular advantage to the parents in
reducing the number of children since children are economi- 21
cally valuable and there is a net transfer of resources from children to parents (Caldwell, 1982a). It has further been suggested that if the time comes when children are disadvanta• geous fertility will fall (Mamdani, 1976 and Caldwell, 1982).
The Sri Lankan Muslims, a high fertility group, have been taken as a case study in this research to analyse the fertility mechanisms. Religious beliefs have been suggested as the reason for high fertility among this group. However, the socio-economic characteristics of the population have not been considered. Thus, a conceptual framework which includes both aspects will be discussed in this section.
It has been recognized that religion is one of the more influential institutions in society and as such is one of the more important social characteristics in differentiating human behaviour (Ford and Jong, 1970 : 212). Thus the fertility beha• viour of religious groups has been given attention in many studies (Freedman, 1959; Westoff and Ryder, 1969 and others).
Most of these studies have been based on three major hypothe• ses .
The first hypothsis, that of "particularized theology", states that the impact of religion on fertility behaviour and attitudes varies according to the particular church doctrine or
religious ideology on birth control, contraceptive usage, and norms of family size (Goldscheider, 1971:272). Secondly the
"characteristics proportion" hypothesis states that the reli• gious differentials in fertility are essentially the result of differences in demographic, social and economic attitudes of 22
the members of the religious groups (Goldscheider, 1971:272).
Thirdly, the "minority group status" hypothesis views religious
fertility differentials within the larger context of a socie• ty's fertility behaviour and social organization. The hypothe•
sis is used to explain not only religious differences in ferti•
lity but also ethnic and racial fertility differences (Bean and
Frisbie, 1978:5). However, recently a new hypothesis, "the
interaction hypothesis", has been proposed by Chamie (1981) for
a comprehensive study of fertility behaviour. He says that
there is no single constant effect on fertility that may be attributed to membership in a particular religious group. Religious fertility differentials will depend on the interaction of the socio-economic levels of the religious groups and the local orientation (by which we mean the current moral attitudes of the religious community) of those groups towards procreation and fertility control (Chaime, 1982:9).
Davis and Black (1956) and Freedman (1967) have sugges•
ted broad determinants of fertility in their analytical models.
According to those models, fertility is assumed to be directly
affected by the level of intermediate variables such as age at
marriage, use of contraception, and infant mortality. The so•
cial and economic structures and environmental factors of a
population are assumed to affect fertility through impact on
the intermediate variables. According to Freedman's (1967)
model, religion forms a part of the social and economic struc•
ture. He further suggests that the effect of religious affilia•
tions operates via norms about family size and intermediate
variables. 23
2.2.1. Religion and Fertility
2.2.1.1 Religiosity, Religion and Fertility
Religiosity has been found to have a strong positive
relationship with fertility in all major religions (Hartman,
1984; Chaudhury, 1976; Westoff, 1973 and 1977. On the other hand, differentials in fertility according to religion have
been reported in many studies (Freedman, 1959; Goldberg, 1967;
and others). In the West, religious affiliations have also been
found to have a significant effect on fertility. For example,
in Europe, Canada, the United States, South Africa, Australia
and New Zealand, studies have shown that Catholics have a
higher fertility rate than non-Catholics (Jones and Norman,
1968). In the developing countries, studies have shown higher
fertility for Muslims in comparison with their non-Muslim nei•
ghbours (Driver, 1963; Kirk, 1968; and others). In Sri Lanka,
recent studies on fertility (Abbeyaratne, 1967 and Fernando,
1982) and reports on population (census reports, vital statis•
tics and World Fertility Survey, 1975) all have clearly indi•
cated higher fertility patterns among Muslims. Thus, many re•
searchers have concluded that Islam is a pro-natalist religion
and therefore the followers of Islam, Muslims, are pro-natalist
in their attitudes towards fertility (Kirk, 1965).
However, the findings of high fertility among Muslims
are not all uni-directional. There are other studies which
clearly indicate that either no differences exist between Mus•
lims and non-Muslims in fertility levels or that those differe•
nces have been due to other factors. For example, Yankey (1961) 24 noted a similar fertility level for Muslims and Christians in rural areas of Lebanon; Rizk (1963) also found this to be the case in rural Egypt; Dandekar and Dandekar did not find any
Hindu-Muslim differences in India when socio-economic variables were controlled. Interestingly, the lowest fertility rates in
Sri Lanka have been found among Sri Lankan Malay Muslims in
recent vital statistics reports. Thus it would be rather hasty
to conclude that religion and high fertility are more closely
related among Muslims than among other religious groups.
2.2.1.2 Islam and Fertility
It has been suggested that Islam's demographic atti•
tudes promote high fertility among its followers (Kirk, 1965).
Others claim that the above allegations are based on a misinte•
rpretation of Islamic doctrine and practices. For example,
regarding birth control, Chaudhury says that
Moreover, Islam does not expressly forbid the voluntary restriction of birth : there is no provision in it for central supreme authority corresponding to the Papacy which proscribes birth control. Instead, the authority for Muslims is the Quran which makes no such prohibition [Hoque, 1974], nor does it prohibit any method of contraception, although it does not proscribe abortion, there are clear authoritative state• ments made by experts in Islamic law that would permit the practice of birth control (1982:120).
In addition, the religion of Islam is often considered
to be pro-natal in character because it insists that children
are among the richest blessings of god. On the other hand, it
is also written in the Holy Quran and Hathid (Traditional
sayings of the Prophet) that Islam very strongly insists on the
responsibilities of parents for their children's well being. 25
Therefore, to understand the fertility behaviour of Muslims,
El-Hamansy (1972) feels that greater attention should be given to the impact of the belief system on the people's behaviour level and less to Islamic theology. She stresses that there are conditions in Muslim societies other than religion that encou• rage procreation. Finally, Omran (1973) suggests that in order to understand the high fertility of Muslims, greater emphasis should be placed on the existing conditions in their countries rather than on the doctrine of Islam.
2.2.1.3 Muslims as a Religio-ethnic and Cultural
Group and Fertility
Muslims have been found to constitute a religio- ethnic group in many parts of the world regardless of their previous racial, linguistic, and cultural orientation (Israeli,
1981; Mines, 1981; Tessler, 1981; Wright, 1982 ; Nagata, 1981 and Ulack, 1970). The introduction of Islam and the religious practices among the followers have weakened their previous ethnic indentities (Weekes, 1969) because Islam is not only a set of beliefs and practices but essentially a way of life
(Israeli, 1981: 3). As a result, Muslims constitute, in some
important aspects, a new religio-ethnic group. For example, Sri
Lankan Muslims who share a common society with other ethnic groups (Sinhalese, Tamils and others) identify themselves and are recognized by others as a different ethnic group.
The Muslim culture in many parts of the world resisted
Westernization in order to preserve its own identity (Rahman,
1982). As Freedman (1963:4) states: 26
When many members of society face a recurrent common prob• lem with important social consequences, they tend to deve• lop a normative solution for it. this solution becomes part of the culture and the society . . .
There is evidence which suggest why Muslim societies in developing countries are so underdeveloped with respect to
education compared to their non-Muslim neighbours. For example,
Sri Lankan Muslims have been isolated from the rest of the
society with respect to socio-economic development, including
education, for many centuries during the colonial period and,
to some extent, even at present because of the fear of reli•
gious conversion and cultural destruction. Therefore, they have
adopted a set of cultural behaviour which, in fact, still
influences their fertility behaviour.
2.2.2 Female Status and Fertility
The status of women and their role in the community and
family decision making, including the timing and number of
births and the choice of contraceptives, have an important
impact on the fertility behaviour of any society (Chaudhury,
1982: 71).
2.2.2 Female Education and Fertility
Education is considered to be one of the most important
variables affecting fertility behaviour and is believed to be
the single most important variable accounting for a large
reduction in fertility in those countries that have have alrea•
dy experienced fertility decline (Coale, 1965; Cochrane, 1979).
Increasing female education, therefore, affects the fertility 27 by:
(i) providing opportunities for personal advancement and awareness of social mobility, new outlook, the freedom from tradition, the values and patterns of behaviour and developing rationalism. Education meets some of the basic psychological needs of women as the need for creativity, a desire to acquire knowledge, and a desire to obtain freedom from close familial controls; (ii) it reduces [should be increases] the age at marriage and increases the probabi• lity of non-marriage; (iii) it reduces the desired family size by fostering a high standard of living for a couple and their children, and stimulates a woman's interest and involvement with activities outside the home, particularly by employment; (iv) it exposes women to knowledge, atti• tudes, and practices favourable to birth control; (v) it allows greater female participation in family decision making; (vi) it increases the chances of survival of infants (Chaudhury, 1982: 84-85).
In a cross-national correlation analysis, it was found that the level of fertility increases consistently with declin•
ing literacy of the female population (U.N.O., 1953). A number of studies have found an inverse relationship between female education and fertility level in European countries (Riser,
1971). Moreover, Sri Lanka is one of the few countries in the developing world that has managed to reduce its fertility level considerably in a relatively short period of time solely by
increasing education alone.
2.2.2.2. Female Age at Marriage and Fertility
Female age at marriage and the proportion of married women have been considered as two of the important factors affecting fertility behaviour since the work of Malthus. Davis and Black (1956), on the other hand, used female age at mar•
riage as an intermediate variable in their model. Early mar•
riage for females means commencement of reproductive life at an early age. Since the majority of births are within marriage 28
in most developing countries, increasing the female age at marriage means decreasing the active reproductive age span and consequently a smaller family. A recent detailed investigation of the historical experience of West European countries has concluded that the first demographic transition took place when
Europe adopted patterns of delayed marriage (Coale, 1973).
Increased female age at marriage led to a considerable fertili• ty decline in many developing countries. Tien (1970) in China and Cho and Retheford (1974) in West Malaysia have found that rising female age at marriage resulted in a decline in fertili• ty in those countries.
2.2.2.3 Female Workforce Participation and Fertility
The occupation of women determines the degrees of their participation in the community decision making which in turn affects their ability to participate in decisions concerning fertility. Gainful employment of women outside the home is found to be inversely related to fertility in many parts of the world. The relationship between fertility and occupation of married women may be explained in the following terms:
(i) the opportunity cost of an additional child becomes high when the prospective mother is gain• fully employed away from home; (ii) women engaged in non-familial activities are likely to develop non-familial interests and their desire for additional children may therefore tend to be relatively low ; (iii) when the status of women depends more on their profession than their fertility, maternity becomes unimportant in their scheme of life (Chatterjee, 1979: 37).
However, the relationship between workforce participation and fertility is very complex and the empirical evidence is highly 29
inconclusive (Mueller, 1982).
2.2.3 Socio-Economic Status, Family Form and Fertility
The influence of socio-economic status variables in explaining fertility behaviour has been important in fertility studies. Historically, studies conducted in many Western coun• tries have shown that there has been an inverse relationship between socio-economic status and fertility during the period of urbanization and industrialization. The social status of the household is identified with the level of income, occupation, and education.
With regard to the influence of household income on fertility, Freedman (1959: 138) states that:
There are several possible factors which may have forced- the higher income groups to have fewer children. One impor• tant case may have been (the first) that children repre• sent an economic burden impeding social mobility. Another possible reason is the location of the couples with higher income in large cities where fertility is usually low. Still another factor may be the less familistic orientation of the economically advantaged class. It must not be sup• posed that they have been uninterested in their children, but they may have had other cultural interest possibly resulting from their better education that may have reduced the importance of large families.
Many studies in developing countries have shown a strong negative relationship between income and fertility levels. On the other hand, some studies do not show any definite relation• ship between the two (Driver, 1963; Agarwala, 1970; Repetto,
1979). However, in Sri Lanka, the level of income has been
found to be a significant indicator of fertility differentials
(Abehyaratne, 1974; W.F.S., 1975).
The occupation of the household, on the one hand, may be 30
considered an even more crucial measure of socio-economic sta•
tus than income level because its endeavour is the most domi• nant single influence on a man's life. Regarding the relation•
ships between occupation and fertility, empirical studies in different Western countries point out sharp fertility differen•
ces between white collar and blue collar workers. The above
relationship is evidenced in the developing countries as well.
For example, the findings of the World Fertility Survey (1975)
in Sri Lanka clearly indicate relatively low fertility patterns
among professional, technical and managerial workers as com•
pared to farmers, fishermen and unskilled workers. Finally, it
is assumed that the fertility level of a society is influenced
by its dominant family structure. It is usually hypothesized
that the nuclear family and household structure promote lower
fertility (Bebarta, 1979). Some empirical investigations sup•
port the above argument (Palmore, 1972), while others have
found little or no relationship between the two variables
(Agarwala, 1970).
2.2.4. Infant and Child Mortality and Fertility
High infant and child mortality is considered as one of
the determining factors in the adoption of small family size
norms in developing countries because successful reproduction
requires high fertility to offset high mortality (Scrimshaw,
1978). Thus, many societies in the developing countries have
adopted an early age at marriage, universal marriage, frequent
child-bearing and high family size norms in order to maintain 31 their demographic balance (Chaudhury, 1982: 130). It has been suggested that if mortality declines, the above pro-natalist norms will disappear. Schultz(1974) found a clear statistical relationship between mortality and fertility in many developing countries. Moreover, several studies at the micro level of analysis have also found higher fertility among women who lost children compared with those whose children survived and con• cluded that the observed differentials in fertility were due to a desire to replace the children (Wyon and Gordon, 1971; Harri• ngton, 1971).
2.2.5 An Analytical Framework for the Fertility Behaviour of Muslims in Sri Lanka
The schematic representation of a model for the fertili• ty study of Muslims of Sri Lanka is shown in diagram 2.1. The variables in the model are arranged according to the causal sequence which were observed during the field survey. The model works backward from the measure of fertility through interme• diate or intervening factors to much wider socio-economic fac• tors to geographical and cultural factors which ultimately affect fertility. The model suggests that cultural, geographi• cal, and socio-economic factors affect fertility through a number of intermediate factors.
The arrangement of this model is based on the following logic. Firstly, fertility is directly influenced by a number of intermediate or intervening factors. For example, increasing child mortality increases fertility; likewise, increasing fe• male age at marriage and her workforce participation decreases FIGURE 2.1: SCHEMATIC DIAGRAM INDICATING THE RELATIONSHIP OF SELECTED VARIABLES IN FERTILITY BEHAVIOR
CULTURAL FACTORS
T r i \
ECONOMIC
FERTILITY
FACTORS
SOURCE: Field Survey, 1981. 33 the total number of pregnancies. Secondly, intermediate or
intervening variables are mediated by socio-economic characte• ristics of the household. For example, the level of the house• hold income as economic factor may influence fertility via a number of intermediate variables. Likewise, types of family, social norms (social factors) may also influence fertility through intermediate factors. Fourthly, socio-economic factors are mainly determined by geographical locations of the house• holds. Finally, cultural factors, such as ethnic characteris•
tics, religious beliefs and attitudes, influence fertility directly and also through a number of other factors.
2.2.6.1 Hypotheses
The following hypotheses will be tested in the analysis
of fertility behaviour of Muslims of Sri Lanka.
(i) Low female education is positively correlated with low
female age at marriage.
(ii) Low female age at marriage is positively correlated with
high fertility.
(iii) Low female employment participation is positively corre•
lated with high fertility.
(iv) Income is a significant determinant of fertility.
(v) Differences in fertility behaviour occur among different
socio-economic groups.
(vi) There is significant variation in fertility between diffe•
rent cultural regions.
(vii) High infant and child mortality are positively correlated
with a high number of pregnancies. 34
CHAPTER 3
THE RECENT FERTILITY TRENDS IN SRI LANKA AND OF SRI LANKAN MUSLIMS
Fertility has declined significantly in Sri Lanka in recent decades. Studies have indicated that the country has been experiencing a demographic transition from high to low birth and death rates (Fernando, 1976; Alame and Cleland,
1981). For example, birth and death rates have declined by 30 and 50 percent within the last two decades. The considerable reduction in fertility and mortality, however, is not common to all ethnic and religious groups. For example, demographically, the Sri Lankan Muslims have behaved differently compared to other religious groups. Therefore, the purpose of the following chapter is to compare the recent fertility trends of Sri Lanka
Muslims with other religious and ethnic groups.
3.1 Recent Fertility Trends in Sri Lanka
3.1.1 Population Growth
The population of Sri Lanka has increased fourfold since the beginning of the century, from 3.6 million in 1901 to 14.9 million in 1981. Much of the population increase occurred during the 25-year period between 1946 and 1971. During this
25-year period, the total population doubled from 6.0 million to 12.0 million. More than 95 percent of the total population increase during this period was due to natural increase rather than migration. For example , the death rate declined by 62 percent from 20.2 deaths per thousand population in 1946 to
7.7 deaths in 1971. The birth rate, on the other hand, fluct- 35
uated until 1960; since then , it has declined steadily from 36 births per thousand population to 30 births in 1971 resulting in a 20 percent decrease. Fertility has further declined bet• ween 1971 and 1981 as well (Table 3.1). The significance of the fertility decline will be examined in the following section with the help of selected fertility indices.
3.1.2 Recent Fertility Trends in Sri Lanka
Three fertility indicators, crude birth rate (number of births per thousand population), general fertility rate (number of births per thousand women aged 15-49), and total fertility rate (number of births per woman'), show that fertility has been declining since 1953 (Table 3.1). The general fertility rate has shown about 25 percent decline from 185 births in
1953 to 139 in 1971. A similar reduction in crude birth rate is also evident between 1953 and 1971. In addition, the total
fertility rate has declined by two births from 5.32 births in
1953 to 3.35 in 1974. The fertility reduction in recent years has been reflected by the decreased average annual growth rate which decreased from 2.8 percent in 1963 to 1.8 percent in 1981
(see Table 3.4).
Figures 3.1 and 3.2 illustrate the age-specific fertili•
ty (number of births per thousand women in each age group) and age-specific marital fertility rates (number of births per
thousand married women in each age group) which indicate the
recent fertility decline more specifically by women's age co•
horts. The information of age-specific and age-specific marital 36
TABLE 3.1
FERTILITY INDICATORS (SRI LANKA) 1953-1974
Years Crude Birth General Total Rate (per 1,000 Fertility Fertility total population) Rate(per 1,000 Rate(per women aged 15-49) women)
1953 37.7 185 5.32 1963 34.4 165 5.04 1971 30.1 138.7 4.16 1974 27.5 121.2 3.35
SOURCE: * Data from Registrar-General, Reports on Vital Statistics and the Population censuses.
** Sri Lanka Fertility Survey, 1978: 7. fertility rates for 1953, 1963 and 1971 are based on census reports and the information for 1975 is based on the World
Fertility Survey. Firstly, age-specific fertility rates have shown a clear fertility decline across the ages (Figure 3.1).
In addition, the young age cohorts have shown a clearer ferti• lity decline than the elders. For example a pronounced fall of
42 percent was witnessed in the 15-19 age group while a mode• rate decline of 29 percent was observed in the 20-24 age group in the census information between 1953 and 1971. The World
Fertility Survey has recorded further declines of age-specific
fertility rate in 1975. Therefore, the available information
indicates . that the women in younger age cohorts have been undergoing rapid demographic changes in recent years. Further• more, the age-specific marital fertility rates have strengthen•
ed the above argument as well (Figure 3.2). 37
FIGURE 3.1: AGE-SPECIFIC FERTILITY RATES - 1953, 1963, 1971 AND 1975
20-24
SOURCE: Data for 1953, 1963, 1971 are from Population Census Reports and 1975 is from the World Fertility Survey. 3 8 •*
FIGURE 3.2: AGE-SPECIFIC MARITAL FERTILITY RATES - 1953, 1963, 1971 AND 1975
15-19 20-24
SOURCE: Data for 1953, 1963, 1971 are from Population Census Reports and 1975 is from the World Fertility Survey. 39
3.1.3. The Determinants of Fertility Decline
The recent fertility decline in Sri Lanka has been attributed to several interrelated demographic and social fac• tors. Firstly, the most significant contributing factor in fertility decline has been the changes in the proportion of marriage in different age cohorts in recent years (Wright, 1970 and Fernando, 1974 and 1979). An examination of currently married women by age groups (Table 3.2) is clearly evidence of a decreasing proportion of female marriages among the women aged 25 and below. For example, between 1953 and 1981, the
TABLE 3.2
PERCENTAGE OF WOMEN CURRENTLY MARRIED IN 1953, 1971 AND 1981
Age of Women Census Year
1953 1963 1971 1981
15-19 23.7 14.8 10.4 10.1 20-24 65.7 57.6 45.9 43.8 25-29 84.4 81.0 73.4 68.2 30-34 87.8 88.6 85.9 81.1 35-39 86.5 89.8 89.0 85.7 40-44 80.7 86.1 86.9 85.8 45-49 73.8 81.6 83.5 83.8
SOURCE: Population Censuses, 1953-1981. percentage of currently married women in the 15-19 age group has declined by 57 percent while the 20-24 and 25-29 age groups have declined by 33 and 19 percent respectively. In addition, female age at marriage has shown a rapid increase by nearly
four years from 20.9 years in 1953 to 25.1 years in 1981. 40
Therefore, female marriage postponement played an important role in fertility between 1953 and 1963 (Wright, 1968) and in the period 1963-1971 (Fernando, 1976).
Secondly, the ever-increasing female level of schooling has directly and indirectly influenced Sri Lankan fertility.
For example, approximately 83 percent of the women in 1981 compared with 91 percent of men are literate (Sri Lanka Popula• tion and Housing Census, 1981). There has been a nearly 50 percent increase in the female literacy rate between 1946 and
1981. In addition, the rates of female school enrolment have
increased rapidly during this period. For example, of the total population, women represent 44 percent of those with primary education, 43 percent of those with secondary education, and 37 percent of those with higher education in 1971 (Sri Lanka
Census of Population and Housing, 1971).
Several studies have investigated the relation of female education and fertility (W.F.S., 1975; Fernando, 1974, 1979).
All have found a strong negative relation between female educa•
tion and fertility. For example, Fernando (1982), in his recent
study of the relationship between fertility and selected back•
ground variables in Sri Lanka using 1953, 1963, and 1971 census
data, has found that the literacy of women aged 20-24 years is
significantly related to fertility.
Finally, as indicated earlier, the age-specific marital
fertility rate has changed over time between 1953 and 1974
(Figure 3.2). The recent decline in marital fertility was
partly a response to increasing use of contraceptives (W.F.S., 41
1978: 9). For example, the new acceptors of family planning have increased nearly 100 percent between 1967 and 1974. Howe• ver, the World Fertility Survey (1975) reported that the use of modern contraception was low when compared either with the level of knowledge or with the level of desire to stop child bearing. Another recent survey on variation in life quality in
Sri Lanka also found similar results (Waxier, Morrison et al.,
1984). In other words, despite the knowledge of family planning and the desire to stop bearing children, very few people use modern contraception to regulate their fertility behaviour.
Fernando (1982) found that female literacy and infant mortality
contributed to the variation in fertility but those factors
remained unchanged between 1953 and 1971. Therefore, he sugges•
ted that family planning activities may have had little impact
on fertility.
Table 3.3 summarizes the determinants of fertility in
recent years. As Table 3.3 indicates, the total fertility rate
has declined by nearly two births between 1953 and 1975. The
female literacy and age at marriage have increased by 54 per•
cent and four years respectively. In conclusion, rapid decline
in the proportion of marriage and ever increasing literacy
rates may have contributed much to the recent fertility decline
in Sri Lanka (Fernando, 1982: 65).
3.2 Fertility Trends Among Muslims
A notable variation in fertility exists among religious
and ethnic groups in Sri Lanka. The Muslims have been quoted as
a " a high fertility group" in many recent studies (Abehayarat- 42
TABLE 3.3
DETERMINANTS OF FERTILITY IN SRI LANKA
Year Female Female Age Proportion Total Ferti- Literacy at Marriage of Marriage lity (Percent) (Years) (Females) (Per Woman) (Percent)
1953 53.6 20.9 24 5.3 1963 63.2 22.1 15 5.0 1971 70.7 23.1 10.3 4.23 1981 82.4 24.4 10.1 3.35
SOURCE: Data from Population Census, 1953 to 1981 except for the total fertility for 1981, which is from Sri Lanka Fertility Survey, 1978b: 7. ne, 1967; W.F.S., 1978; Fernando, 1980, 1982). Sri Lankan
Moors, who constitute nearly 95 percent of the total Muslim
population, were also given the same title. Comparatively high
fertility among Muslims was explained in many studies as due to
their religious beliefs and attitudes. For instance, Fernando
concluded , in his recent fertility analysis of five Sri Lankan
Moor communities, that the religious conservatism of the Muslim
population has influenced their fertility (Fernando, 1982). The
purpose of the following section is to identify some of the
socio-economic and demographic factors which influence the
fertility behaviour of the Muslim population by using available
national statistics.
3.2.1 Muslim Population Growth
The growth of the Sri Lanka Muslim population has been a
reflection of national population growth (see Table 3.4). The
growth of the Muslim population has not been uniform throughout 43
TABLE 3.4
POPULATION GROWTH IN SRI LANKA--MUSLIM GROWTH IN PERSPECTIVE
Census Muslim Population Increase (%) Average Annual Year Growth Rate (%)
Number Increase (%) Musiim Total Musiim Total
_ 1871 171543 _ _ 1881 197775 262333 15.3 15.0 1 .4 1 .4 1891 211995 1 4220 7.2 8.7 0.7 • 0.9 1901 246118 34123 16.1 21 .9 1 .5 1 .7 1911 283631 37513 15.2 12.3 1 .4 1 .4 1 921 302532 18901 6.6 9.6 0.6 0.9 1 946 436556 134150 44.4 48.0 1 .6 1 .7 1953 541506 104950 24.0 21.6 1 .4 1 .5 1 963 724043 182537 33.7 30.7 3.1 2.8 1 971 909941 185898 25.7 19.9 2.8 2.7 1981 1134556 224615 24.7 16.8 2.8 2.2
SOURCE: Computed from various census reports, 1871-1981. this century as noted in the national population growth ear• lier. For example the percentage growth of the total Muslim population in the seventy-five year period between 1871 and
1946 was more or less the same as during the period 1946-1971
(Table 3.4). This means that the Muslim population has grown very slowly with an average of 1.2 percent annual growth; during the second period (twenty-five years), however, it grew at a much faster rate with an average annual rate of 2.4 per• cent. In addition, a moderate growth was observed between 1971 and 1981. The rate of Muslim population growth, however, has reversed since 1953 compared with the rate of national popula• tion growth. For example, the growth rate of Muslim population was lower than the national rate before 1963 and higher after 44
1963 (Table 3.4).
The population growth has varied among religious and ethnic groups as well. As Table 3.5 indicates, the growth of the Muslim population, when compared with other religious and ethnic groups, has varied. For example, the percentage growth of Muslim population was lower than for Buddhist and higher
than for other religious groups between 1911 and 1946. On the other hand, the growth rate of the Muslim population was higher
than for all other religious groups but slightly higher than
the Buddhists during the period 1946-1971 and 1971-1981. Figure
3.3 illustrates the population growth of religious groups bet•
ween 1911 and 1981. The Hindu and Christian populations have a
comparatively lower growth rate since their populations have
been subject to outmigration. The population growth of ethnic
groups indicates (Table 3.5) that the Moor population has grown
slower than the Sri Lankan Tamil population during 1971 and
1981. Finally, the population growth has varied among the
Muslims as well. For example Figure 3.4 illustrates that the
Sri Lankan Moor population has been growing steadily compared
with the Indian Moors and Malays.
3.2.2 Fertility Trends Among Muslims
In the following section, the variation in fertility
among religious groups will be analysed by using available
fertility indicators. The available socio-economic and demogra•
phic information is, in many cases, limited to ethnic groups
only. Therefore, the Sri Lankan Moors, who constitute nearly 95 45 46
FIGURE 3.4: TRENDS IN MUSLIM POPULATION
100,000
10.000
1961 1971 1981 1921 1931 1941 1951 YEAR
SOURCE: Population Census Reports, 1911 - 1981, 47
TABLE 3.5
PERCENTAGE CHANGE IN POPULATION BY RELIGIOUS AND ETHNIC GROUPS, 1911-1981
Group 1911-1946 1946-1.971 1971-1981
Religious Group
Buddhist 74 100 20 Hindu 41 41 03 Christian 47 65 12 Muslim 54 108 25
Ethnic Group
Sinhalese 70 100 20 Sri Lankan Tamil 40 93 32 Sri Lankan Moors 60 121 28
SOURCE: Calculated from census reports. percent of the total Muslim population, will be represented by the Sri Lankan Muslim population as a whole.
Much has been written about Sri Lankan fertility in
recent years. The pattern of Muslim fertility is nearly identi• cal to the national fertility trends in many respects. For example, crude birth rate for Muslims has fluctuated along with the death rate until the middle of this century (Figure 3.5).
The improvement in mortality and/ or morbidity contribute to
increases in Sri Lankan fertility after 1946 (Lambard, 1980:
291). Likewise, the fertility of the Muslim population has also
increased during this period. For example, crude birth rates and child-women ratios (Table 3.6) have shown an increase and
remained high until 1963. In addition, Lambard noted an in• crease in marital fertility rate during the same period among 48
TABLE 3.6
CRUDE BIRTH RATES AND CHILD-WOMEN RATION BY ETHNIC GROUPS 1946-1977
Ethnic Group Census Year Percent Decline
1946 1953 1963 1971 1977 1963-77
Crude Birth Rate
Sinhalese 38.7 41 .0 34.5 29. 9 27 .3 -20.9 Sri Lanka Tamils 35.6 39.2 37.6 31 . 8 28 .3 -24.9 Indian Tamils 41.2 33.0 28.3 25. 7 30 .6 + 8. 1 Sri Lanka Moors42.7 42.7 42.9 39. 0 32 .6 -24.0 Malays 42.3 39.6 34.8 27. 2 20 .8 -40.0
All 37.4 38.7 34.4 30. 4 27 .5 -18.4 * Child-Women Ratio Percent Declini 1963-7
Sinhalese 596 720 741 598 -19 Sri Lankan Tamils 510 633 774 609 -21 Indian Tamils 647 697 617 537 -13 Sri Lanka Moors623 803 941 767 -19 Burgers 468 592 550 428 -22
All 594 715 749 603 -19
Note: * Population aged 0-4 year per 1,000 women aged 15-44.
SOURCE: Computed from Census Reports and Registrar General's Report on vital statistics.
Sri Lankan Moors and possibly among Indian Tamils (Lambard,
1980: 301). He states that the reason for the increase is "...
improvement in health and/ or level of nutrition, or because of
the breakdown of social norms which had acted previously to 49 constrain fertility ..." (1980: 302).
Fernando (1982: 421) using crude birth rates and women- child ratio noted that Muslim fertility rates fell more slowly
than that of the other ethnic groups between 1953 and 1971.
Table 3.6 shows, however, a different picture. Muslim fertility has declined faster than the national rate and that of many
other religious and ethnic groups in recent years.
As noted earlier, Muslim fertility declined since 1963.
Muslim fertility has shown a considerable decline between 1963
and 1981 (Table 3.6). For example, crude birth rate is evidence
of a 24 percent decline during this period which was one of the
largest declines (Table 3.6). In addition, the child-women
ratio indicates more than a moderate decline between the 1963
and 1971 period. Above all, the crude birth rate of Malay
Muslims has. shown a rapid decline and has remained at a low
level when compared to all the other ethnic groups.
Finally, the age-specific marital fertility rate has
also declined in all age groups among the Muslim population
(Table 3.7). Therefore, from the available information, it
could be suggested that the Sri Lankan Muslims are not an
exception to the present national trends of rapidly decreasing
fertility. To understand why some (Muslim) behave differently
in their child-bearing activities, greater emphasis should be
placed on the existing socio-economic and political condition
of this group. 50
TABLE 3.7
PERCENTAGE OF CHANGE BETWEEN 1960/65 AND 1970/75 IN AGE-SPECIFIC FERTILITY RATES BY ETHNICITY AND RELIGION
Age of Women Ethnicity and Religion at Birth
Sinhalese Tamil Hindu Moor Muslim Christian Buddhist and Others
15-19 4.5 1 .4 -6.8 20-24 -6.2 8.4 -0.9 -14.3 25-29 -17.4 -10.7 -15.1 0.3 30-34 -16.6 -8.4 -16.3 -20.6 35-39 -37.5 -27. 1 -20.8 -26.8
SOURCE: Alam and Cleland, 1981: 15.
NOTE: Women on estates have been excluded.
3.2.3 Determinants of Muslim Fertility
First of all, mortality as a socio-economic indicator may have had an effect on the fertility behaviour of the Sri
Lankan Muslims throughout this century. As noted earlier (Fi• gure 3.5), the crude birth rates of the Muslim population had fluctuated along with crude death rates during the first half of this century. Malaria and other epidemics had directly and indirectly influenced Sri Lankan fertility during this period
(Sarkar, 1957). The mortality at the national level has dec• lined since 1946 due to malaria eradication, improvements in medical services, improved nutrition, and a better standard of living (Meegama, 1967; Gray, 1974; Collumbine, 1950; Fernando,
1982). The Muslim mortality level, however, remained comparati•
vely high until recently when it began to decline in line with FIGURE 3.5: TRENDS IN. BIRTH AND DEATH RATES (MUSLIMS) - 1900-1981
A : : BIRTH RATE
DEATH RATE
19j0 1940 1950 1960 1970 I960 1910 1920 YEAR. SOURCE: Registrar-General's Reports on Vital Statistics, 1900-1981. 52
TABLE 3.8
FIVE-YEAR AVERAGE CRUDE DEATH RATES, 1911-1975
Years Sri Lanka Moors Sri Lanka (Muslims)
1911-1915 29.8 30.6 1916-1920 31.9 29.7 1921-1925 31 .9 27.8 1926-1930 28.5 24.3 1931-1935 27.8 24.6 1936-1940 24.2 21.4 1941-1945 23.2 20.4 1946-1950 18.3 14.6 1951-1955 13.7 11.1 1956-1960 11.7 9.5 1961-1965 9.7 9.0 1966-1970 9.3 8.8 1971-1975 8.3 8.2
SOURCE: Calculated from Registrar-General's Reports on Vital Statistics, 1911-1975.
the national population (Table 3.8).
Table 3.8 clearly indicates that the crude death rate of
the Muslim population has been higher than the national level
throughout this century. Furthermore, a comparison of the Regi•
strar General's reports on age-specific mortality rate indi• cates similar findings.
The mortality rates of the Muslim population have been higher than those of other religious and ethnic groups. An
examination of crude death rates between 1900 and 1977 for the different ethnic groups indicates that Muslims (Moors) have
experienced higher mortality than many other ethnic groups
throughout this century. A similar attempt has also been made
to compare the age-specific mortality rates of ethnic groups 53
between 1946 and 1971. It has been found that comparatively higher rates of infant, maternal, and old-age mortality exist among Muslims both male and female. Table 3.9 indicates, for instance, that a comparatively higher rate of mortality existed among Muslims throughout this century. According to recent studies (Meegama, 1980: 37; Waxier, et al, 1984). The socio• economic, environmental, and political factors have been found to influence infant mortality. In addition, many studies, both inside and outside of Sri Lanka, have found a strong positive relation between mortality level and the level of fertility
(Ramakumar, 1972; Sarma, 1972; Preston, 1978; Fernando, 1982;
W.F.S., 1975). Therefore, Sri Lankan Muslim fertility has pro• bably been influenced more by the high level of mortality than in the case of the other ethnic groups.
TABLE 3.9
INFANT MORTALITY BY ETHNIC GROUPS, 1910-1964
Ethnic Group 1910-1912 1945-1947 1962-1964
Sinhalese 190 124 49.2 Tamil 229 130 52.0a Moor (Muslim) 237 143 62.0b All 202 127 55.1
SOURCE: Sarkar, 1957: 143. ESCAP, 1976: 143.
aSri Lanka Tamils only.
^Sri Lanka Moors only.
One of the great achievements of independent Sri Lanka has been the ever-increasing rate of female education (Jayawee- 54
ra, 1979). The increasing female level of education among Sri
Lankan Muslims has been slower than at the national level and remains lower than is the case among the other ethnic and religious groups. As Table 3.10 indicates, more than 82 percent of the Muslim women have had less than primary education while only 4 percent were found in higher education. The average education level is much lower among Muslims than among other ethnic groups with the exception of Indian Tamils. My recent survey on socio-economic characteristics of Muslim university students at Perideniya University and Dumbara campus indicates that less than one quarter of Muslim students are female. Since the greater proportion of Muslim women do not reach high level of education, their fertility has been directly and indirectly influenced by their education level. The findings of Fernando, from his recent analysis of fertility in Sri Lanka strongly supports this (Fernando, 1982).
TABLE 3.10
PERCENTAGE OF FEMALE LEVEL OF SCHOOLING BY ETHNIC GROUPS
Ethnic Group No School Primary Secondary Tertiary
Sinhalese 1.8 53.0 29.6 14.7 Sri Lankan Tamil 1.2 57.0 25.8 14.9 Indian Tamil 0.7 92.0 5.1 1.0 Sri Lanka Moor 2.2 80.0 13.6 3.9
SOURCE: Computed from Weekes-Vagliani, 1980: Table 3.15.
As noted earlier, female age at marriage has been in• creasing in recent decades in Sri Lanka. In addition, the 55
fertility increases steadily with female age at marriage
(W.F.S., 1975). Many studies have found that female age at marriage is negatively correlated with fertility (Abehayaratne,
1967; W.F.S., 1975). Muslims, however, marry comparatively early. An examination of common, Kandyan, and Muslim marriage records between 1950 and 1977 indicates that the average female age at marriage under Muslim law has been lower than twenty years, while the age of women in common and Kandy marriages was on average more than twenty years (Registrar-General's Reports,
1950-1977) .
In addition, the World Fertility Survey indicates that
85 percent of the ever-married Muslim women had married between
fifteen and nineteen years of age, while 47.7, 64.8, and 68.9
percent of Sinhalese, Sri Lankan Tamils, and Indian Tamils
respectively in the same age groups. Therefore the first report
of the World Fertility Survey concluded that high fertility of
Muslims was caused mainly by their practice of early marriage
(W.F.S., 1978).
Finally, workforce participation by women may have also
strongly influenced Muslim fertility. The same report notes
that more than 82 percent of the total ever-married Muslim
women have not engaged in any gainful employment—the highest
rate of non-participation of any ethnic group. Hanna and Nada-
raja's findings on the impact of workforce participation and
the level of education on fertility have been summarized in
Table 3.11. As it shows, fertility decreases with women's
socio-economic development, and Muslims are no exception to 56
TABLE 3.11
MEAN NUMBER OF LIVE BIRTHS PER EVER-MARRIED WOMAN AGED 4 5-49 BY SOCIO-ECONOMIC STATUS AND ETHNIC GROUPS
Soc io-Economic Ethnic Group Status
Sinhalese Sri Lanka Indian Sri Lanka Tamil Tamil Moor
Educated & 4.15 5.24 3.50 4.10 Economically Act ive Educated & Eco- 5.18 5.26 4.87 4.29 nomically Inactive Uneducated & Eco- 6.83 6.34 6.02 6.67 nomically Active Uneducated & Eco- 6.57 6.13 6.29 6.93 nomically In• active
SOURCE: Hanna and Nadrajah, 1976. this.
In summary, fertility rates among the Muslim population has started to decline later than among the other ethnic and religious groups, and the rate of population increase remains comparatively high. Available evidence suggests that socio• economic and political factors may have influenced Muslim fert•
ility. The available national data are not adequate to find out exactly what combination of factors are influencing Muslim
fertility. A survey of the kind undertaken for this study
fulfills a gap in data, and will enable a comprehensive analy•
sis of the fertility behaviour of Muslims in Sri Lanka. 57
CHAPTER 4
METHODOLOGY
It was found that national level statistics were not adequate for an in depth study of the fertility behaviour of ethnic and religious groups in Sri Lanka, because they do not provide individual level information on fertility. Therefore, a sample household survey which represents the Sri Lankan Muslim population was conducted. The methods which were used in the survey will be examined in this chapter.
The survey was conducted out of the Geography Department of the University of Peradeniya, during the second half of
1981. The survey was conducted with the assistance of the
Muslim student unions (Muslim Majlis) at both Peradeniya Uni• versity and Dumbara Campus. A number of undergraduate students from the above associations voluntarily assisted in every stage of this survey.
The purpose of this survey was to collect fertility and other relevant socio-economic background information from ran• domly selected sample households. Eleven administrative distri• cts which numerically, geographically, and culturally represent
The Sri Lankan Muslim population were selected in order to choose eleven Muslim settlements from them. For each district the Muslim population along with that population as a percen• tage of the total Sri Lankan Muslim population are given in
Table 4.1.
Since the Muslim settlements in many parts of the coun- 58
try are small and located within the same administrative boun• daries as other religious and ethnic groups, it was not possib•
le to separate out Muslim settlements and their population
charateristics from those of the other communities using the
available national statistics. Moreover, the settlements were
not listed categorically by religious or ethnic groups in
administrative statistics. Therefore, a list of prominent Mus•
lim settlements and their locations in Sri Lanka was prepared.
It was possible to identify and locate nearly 500 prominent
Muslim settlements. The location of those settlements is illu•
strated in Figure 4.1 and their names are given in Appendix 1.
One settlement in each of the eleven districts was
randomly selected by using the prepared Muslim settlement list.
The names of the selected settlements are given and their
location is shown in Figure 4.2. On the average forty house•
hold units were randomly selected from each of the settlements
in order to represent the larger proportion of the total Muslim
population. In some cases, a list of Muslim households, which
was provided by the local Gramasevaka, was used to select the
households while in other cases, the voters list was used. In
other settlements where both household and voters list were not
available a list of names used for local mosque administration
was used.
A questionnaire which consisted of detailed questions on
fertility was prepared as a means to collect the household
information. The questionnaire was pre-tested in a nearby sam•
ple settlement and then was modified for data collection. The 59 SOURCE: Field Survey, 1981. 61
TABLE 4.1
SELECTED SAMPLE MUSLIM SETTLEMENTS ALONG WITH OTHER BACKGROUND INFORMATION
Name of District Total Dist. Proportion Cultural Settlement Musiim to the Region Populat ion Total (1981 ) Musiim Population
Erukkalam- piddy Mannar 44,003 2.4 North-West Puttalam Puttalam 50,243 4.3 North-West Topoor Trincomalee 75,761 6.8 East-coast Kathankuddy Batticaloa 79,662 6.9 East-coast Nintavur Ampara i 161,754 13.9 East-coast Kekunugolla Kurunegala 64,213 5.4 Central & South Pangollamada Kandy 125,646 11.5 II Hemmathagama Kegalle 36,548 3.3 II Wanahapuwa Nuwara Eliya 15,791 1 . 1 I! Weiigama Matara 16,853 1 .6 II Muthuwela Colombo 168,956 20.6 Colombo
Total 839,430 77.8
SOURCE: Field Survey, 1981.
full text of the modified questionnaire is given in Appendix 2.
The modified questionnaire consisted of eight major parts. The contents of those major parts are as follows : (1) location and
identification of households; (2) general characteristics of household members e.g. age, sex, marital status,etc.; (3) sour• ce and amount of household income ; (4) type of house; (5)
health condition of household members and type of treatment ;
(6) social activities of household members; (7) information on
fertility ; (8) migration. The questionnaire was originally
written in the Tamil language which almost all the respondents 62
were able to read and understand.
The survey was conducted with the assistance of selected
Muslim undergraduate students. Among the undergraduate students who voluntarily offered their assistance in the survey, 16 of them , 12 males and 4 females, were selected to assist in the household data collection. One student or in some cases, two students were assigned to each of the eleven settlements. In most cases, the students who were assigned to a particular settlement were from the settlements in which the survey was being conducted. The selected students were given detailed instructions regarding the techniques of conducting a household survey. In addition, they all participated in the pilot survey.
The survey in each settlement was initiated by me and continued by student assistants. During the first visit, an attempt was made to establish a friendly relationship with the local people in each settlement. For example, we accepted an invitation from the local school to deliver a public speech to its students and we made use of this opportunity to convey to the local people our purpose for coming to the village. In addition, we also met local religious and political leaders who subsequently gave valuable information regarding the socio-economic and histori• cal character of the settlement. We also surveyed the geograp• hical and the socio-economic condition of the settlement as well.
When we completed the background survey, we began the household survey using a random sample selection. Nearly ten households in each settlement were interviewed during my first 63
visit to the settlement. In order to assure that the student assistants understood the interviewing techniques properly, I accompanied them to every household during the first visit.
Most of the necessary information was recorded by interviewing the head of the household who was usually the husband. The information on fertility was directly collected from the female member. The interviewees were cordial and co-operative in al• most all cases. One of the main reasons for the successful interviews was the presence of a student assistant from the same settlement who were in many instances blood relations of the interviewees.
The survey was completed as planned, however, some prob• lems which were encountered during the interviews are worth• while mentioning. Some of the respondents were not able to give accurate information on their date of birth, source and amount of income, etc. In addition, old female respondents could not remember some information on early age pregnancies. Moreover, some mothers were not happy to discuss their children who died during pregnancies or later. Those and other problems, however, were solved by asking both direct and indirect questions on these matters. One problem which was not solved in the field was the information on the use and the desire to use modern contraception methods. Many female respondents were reluctant to discuss those matters in the presence of male interviewers.
However, in those settlements, where female assistants were used, we were able to get reliable information on family plan• ning methods. 64
Each questionaire was reviewed in the field after the interview. However, when the entire survey was completed, ques• tionnaires were once again re-checked and among them the incom• pleted questionnaires were excluded. Finally, 323 examples of household information (questionnaires) from eleven settlements were selected for the analysis. Those selected questionnaires were manually tabulated at the University of Peradeniya with the assistance of undergraduate students. This manually tabu• lated household information was coded according to the SPSS instruction and made available for further computer analysis at the University of British Columbia in Canada.
The following steps were taken in the computer analysis.
The study began by running a series of frequency distribution tables for 323 cases. It was possible to identify the general characteristics of surveyed population by using frequency dist• ribution tables. Secondly, the analysis of cross tabulation between selected background variables and fertility was done to compare the bivariate relation. The findings of cross tabula• tion have been compared and confirmed with Pearson correlation coefficient analysis. Thirdly, the study moved to multivariate analysis; in this section, multivariate cross tabulation and
Partial correlation coefficients have been done to determine selected background variables which were significantly corre• lated with fertility. Finally, stepwise multiple regression which provides beta coefficient for path analytical model, was done. The strength of each of the five selected background variables to fertility was determined by using path analysis. 65
CHAPTER 5
STUDY AREA
The study area comprises eleven Sri Lankan Muslim set• tlements. The word 'settlements' refers here to a group of households that live together and this includes villages, gra- masevaka divisions, and urban wards.
Freestone (1974) categorized human settlements into six groups according to geographical location and economic fun• ction. The function of any settlement is not only economic but also socio-behavioural. The members of a settlement must colla• borate with one another to achieve their goals and to satisfy their needs. In this process, the members of society are ex• posed to several social values, aspirations and behaviours within an existing society. In this regard, Sri Lankan Muslim settlements are unique in character and in function as well.
For example, although the Muslim settlements are spread across various cultural and sub-cultural environments, they remain distinct in that they provide a religio-cultural environment to their members.
5.1 Muslim Settlements and Their Religio-Cultural Envi ronment
Sri Lanka is a highly sophisticated and complex society with respect to the richness of its cultures, and religious and social organizations (Yalman,1967:3). There are two distinct ethnic and cultural groups—the Sinhalese and the Tamils--who form the basis of the cultural diversity of this society. The 66
Sinhalese and the Tamils claimed descendancy from distinct racial stocks, from the Aryans and the Dravidians respectively
(Samaraweera, 1977: 86). In addition, the Sinhalese embraced
Buddhism and the Tamils largely retained the Hindu faith. The
Sinhalese community is divisible into low-country and Kandyan sub-cultures (Fernando, 1979: 5). The Tamils, on the other hand, are divided into Sri Lankan and Indian Tamils; the Sri
Lankan Tamils are further identified as Northers or Easterners.
All these divisions of Sinhalese and Tamils are in turn subdi• vided into a variety of castes (Yalman, 1967: 12). Thus, Ryan
(1950) categorized Sri Lanka into seven major regions on the basis of cultural ethnic criteria and patterns of community organization (Ryan, 1950).
The Sri Lankan Muslims, who share a common society with other ethno-cultural groups (the Sinhalese and the Tamils), identify themselves and are recognized by others as a distinct religio-cultural (ethnic) group. At the same time, for obvious reasons, there are variations in cultural behaviours among the
Muslims themselves according to whether they live in a Sinha• lese or Tamil region.
I have divided Sri Lanka into four major cultural re• gions (Figure 5.1) according to the dominant culture (the
Sinhalese and the Tamils) and the distinct occupational and cultural behaviour of the Muslims themselves in those respect•
ive regions. The first two regions, the North West and the East coast, are predominantly Tamil in culture; the other two are predominantly Sinhalese in culture. In the North West region, 67
SOURCE: Field Survey, 1981. 68 the Muslims have largely been influenced by South Indian Tamil-
Muslim culture. On the East coast, Muslim culture is much closer to the dominant Sri Lankan Tamil culture of this region
(McGilvray, 1974). Nearly 70 percent of the total Muslim popu• lation lives in the Sinhalese cultural areas. To some extent, the Sinhalese culture also has influenced the Sri Lankan Isla• mic culture. In short, the culture of the Sri Lankan Muslims is the combination of Islamic as well as the local (mainly Tamil and to some extent Sinhalese) cultures. However, Muslims have modified many aspects of the Tamil and Sinhalese cultures in light of the teachings of Islam, and made them their own. This has been evidenced in many other non-Islamic countries as well
(Israeli, 1979; Ahmed, 1979; Mines, 1981).
A Muslim settlement regardless of its number of inhabi• tants provides a religious and cultural environment for its members. A necessary element of a Muslim settlment is a mosque, a centre for prayer, whose interaction among the members of the community takes place several times per day. The number of families who belong to a mosque varies from five to more than five thousand. In other words, Muslims prefer to live in an established Muslim environment, or failing that, to create such an environment wherever they live, regardless of their numbers.
In addition, a number of religious institutions, such as Arabic schools, hathi court etc, play a more important role within the community than do their counterparts among the other ethnic groups in Sri Lanka. For example, the mosque administrators and
trustees, usually representatives of large families or clans, 69 make informal laws and regulations not only concerning the religious aspects of the community, but also the social and economic aspects as well. The administrators implement these rules and regulations whenever necessary. Therefore, it has been noticed that the formal and informal government institu• tions, such as village councils, rural development boards, etc., are less effective or less important than the religious institutions in many Muslim settlements in Sri Lanka. In addi• tion, the 1886 Mohammedian Marriage Registration Ordinance facilitated their own marriage court (hathi court) to solve marriage disputes according to Islamic law.
5.2 Surveyed Settlements
Figure 3.2 illustrates the location of the surveyed
Muslim settlements in Sri Lanka. The eleven surveyed settle• ments represent eleven administrative districts which are spread throughout the country. The majority of the surveyed settlements (six) are located in the central and Southern
Sinhalese cultural areas while three other settlements are located in the East coast and two in the North and the North
West regions. Colombo district has been represented by Muthuwe- la which is one of the suburbs of Colombo city. Geographically, as Figure 3.2 indicates, the surveyed settlements are distri• buted in wet and dry zones and in hilly and coastal areas.
Economically, four out of eleven settlements are engaged in agriculture. Among them, Nintavur and Topoor are engaged in rice cultivation while Kekunugolla in coconut and Wanahapuwa in a mixture of tea plantation employment and vegetable gardening. 70
On the other hand, Weligama is a business community; most of its inhabitants are engaged in some sort of business activities which range from large scale retail business to sales workers.
In addition, although the other non-agricultural settlements
(see Figure 5.2) are not categorized as business settlements, a proportion of the workforce in those settlements are engaged in unspecified business activities. At the same time, in Puttalam and in Erukkalanpiddy, a small porportion of the total workfor• ce is engaged in fishing while in Pangollamada and in Henatha- gama some are engaged in small scale agricultural activities.
The economic activities of Muthuwella are greatly influenced by
Colombo city. In almost all the settlements, some young house• holds headed by school teachers and clerical employees have been found. Seven out of the eleven settlements are considered to be rural according to the local government definition.
However, since the economy of some of those settlements is based on non-primary economic activities and as these settle• ments are located in close proximity to urban centres, they have been categorized as semi-urban centres in this study
(Figure 5.2)
Most of the surveyed settlements have been long establi• shed. For example, Puttalam and Weligama are known as some of the earliest Muslim settlements on this island, while Pangolla mada, Hemathagama, Kathankuddy , and Erukkalampiddy have served as centres for Muslim traders for centuries. Likewise, the history of the East coast agricultural settlements (including
Nintavur and Topoor) go back to the historical migration 71
FIGURE 5.2
SOCIO-ECONOMIC AND CULTURAL CHARACTERISTICS OF THE SURVEYED MUSLIM SETTLEMENTS
Cultural Area Settlement Economic Characteristics
Sinhalese Area Kekunugolla Agricultural & Rural Wanahapuwa Agricultural & Rural Muthuwela Non-agricultural & Urban and Semi-urban Weiigama Non-agricultural & Urban and Semi-urban Hemathagama Non-agricultural & Urban and Semi-urban Pangollamada Non-agruclutural & Urban and Semi-urban
Tamil Area Nintavur Agricultural & Rural Topoor Agricultural & Rural Kathankuddy Non-agricultural & Urban and Semi-urban Puttalam Non-agricultural & Urban and Semi-urban Erukkalampiddy Non-agricultural & Urban and Semi-urban
SOURCE: Field Survey, 1981.
NOTES: The settlements of Muthuwella, Weligama, Kathan• kuddy, and Puttalam are urban centres above the rank of town council. Further, we should note that Puttalam is located in the Sinhalese cultural area.
of the Muslim population from the West coast after its persecu•
tion by the Portuguese colonial power during the 16th century.
These communities have traditionally maintained friendly rela•
tions with Sinhalese and Tamil settlements. The Muslims are
still called as 'Marakalayo' by Sinhalese and 'Sonakar' by
Tamils signifying their historical role as friendly traders.
The surveyed Muslim settlements are representative of
the cultural and economic characteristics of the other Muslim 72
settlements and in their respective cultural regions and admin• istrative districts. For example, the surveyed East coast set- tlements--Nintavur, Topoor and Kathankuddy--are representative both culturally and economically. Weligama, in the South coast of Sri Lanka, not only represents the nature of business acti• vity of the majority of the Muslim settlements in the Matara district but, to some extent, also represents some of the other neighbouring districts such as Galle and Kalutura as well.
Likewise, a similar pattern has been identified among other surveyed settlements. However, interestingly, those surveyed settlements also vary individually in cultural behaviour, in economic activities and in other characteristics since they are spread over a large geographical area. For example, in the East coast, Topoor and Nintavur are rural and agricultural villages while Kathankuddy is urban and engaged in mostly non-agricultu• ral economic activities.
5.3 Surveyed Households
Table 5.1 gives the total number of households and the total population of each surveyed settlement. On average thirty households from each of the eleven surveyed settlements were interviewed; giving a total of 323 households out of a popula• tion of 2098 of the eleven settlements. All the households are
Muslim. However, nearly ninety percent of the total surveyed population are Sri Lankan Moors while the rest are the decen- dents of Indian Moors who have mainly been found in the Wanaha-
puwa estate settlement. In addition, the male ratio for 73
TABLE 5.1
NUMBER OF SURVEYED HOUSEHOLDS AND SURVEYED POPULATION ACCORDING TO THE SETTLEMENTS
Settlement Surveyed Households Surveyed Population
Topoor 30 193 Kathankuddy 28 1 72 Nintavur 30 1 58 Erukkalampiddy 30 1 66 Puttalam 30 1 92 Muthuwella 28 208 Weligama 27 1 97 Kekunagolla 32 1 99 Hemathagama 27 208 Pangollamada 31 199 Wanahapuwa 30 206
Total 323 2098
SOURCE: Field Survey, 1981.
the total population is more or less the national average. The age structure varies from settlement to settlement; however,
48.3 percent of the total population is below twenty years of age. Although the surveyed population is homogeneous in its
religious affiliation , the socio-economic and demographic cha-
racteristics vary greatly among settlements and within the
settlements themselves.
5.3.1 Head of Household
The person who has more say in the family decision making while, in many cases, also earning the larger proportion
of the total family income, is considered the head of the
household. According to the survey, 88 percent of household
heads were husbands or fathers. Mothers and sons (usually elder 74 sons) constituted about 8 and 4 percent of the total households respectively.
The role of the husband, economically and otherwise, is very important in the household decision making. The characte• ristics of the husband as the head of the family varies from household to household. Firstly, the age distribution varies from around twenty-five years to more than sixty years. Howe• ver, the variation in the age correlates with the family types.
For example, the age of the husband in the nuclear family is less than that in joint family households. Secondly, the educa• tion level also varies according to age, with the young being more highly educated than their elders. However, the average educational level for husbands is much lower than the national average for all ethnic groups. For example, more than half of the total heads of the households (husbands) have had only between one and five years of education in this survey. Howe• ver, the educational level for the husbands varies from settle• ment to settlement. In addition, the employment patterns of husbands have been greatly influenced by their educational level. About 18 percent have been categorized as professionals; they are mostly school teachers.
A female as the head of the household is not unusual in this society. However, in most cases, the circumstances in which the female becomes the head of the household is either after the death of the husband or a separation from the hus• band. In both cases, she would be the main income earner and
decision maker in the family. The study indicates that many 75 widows lead their families with some sort of wage employment.
On the other hand, if the elder son is an adult at the death of the father, the son leads the family economically although the
family decision making, to some extent, remains in the.hands of the mother.
5.3.2 Family Members
The survey indicates that the average family size is
generally high among Sri Lankan Muslims. Family members include
spouses, children, and other family and non-family members who
live in the household. The average family size is 6.2 members
for the 323 surveyed households. More than 64 percent of the
total households have five or more members while another 11
percent have nine or more members, although the family size
varies from settlement to settlement (Appendix 3). The settle•
ments which are located in the Central and Southern cultural
regions have a larger family size than the others. The majority
of the family members are unmarried children. However, in some
cases, household's parents, grandparents, in-laws, widowed
sisters, unmarried brothers and sisters and other non-family
members are also included in the household.
The role of women in this society is much more important
than it has been described in some other studies. The marriage
tradition is strong, as evidenced by the fact that no separated
family has been found in this study. Although divorce can be
easily obtained, its frequency is very low among all sections
of the survey population. A widow or divorcee can remarry after
completing her period of Iddat, i.e., four months after the 76
death of or divorce from her husband. Widow remarriage is quite prevalent and is encouraged by all Muslims. On the other hand, the incidence of polygamous marriage among the surveyed house• holds is almost nil. People react very unfavourably to a person who marries while his first wife is still living.
Children are considered as assets by the parents, as their role in the family economic activity is very important.
As noted earlier, the majority of the surveyed households are either dependent on primary economic activities (agriculture or fishing) or some sort of business activities for their liveli• hood. In both sectors, children, regardless of their age, contribute their service to the family economy. It has been noted during the field survey that the children of all ages drop out of school temporarily or permanently in order to assist their parents in agricultural or business activities.
Thus male children are considered more important than female children for economic reasons. In addition, parents prefer comparatively higher education for male children than for fe• male children.
5.3.3 Socio-Economic Status
The social stratification of many South Asian societies is based on the caste system (Dumont, 1980; Srinivas, 1976), and - Sri Lanka is no exception to this (Fernando, 1979:29).
However, the traditional caste system has not been found among
Sri Lankan Muslims. Therefore, the social stratification of the
surveyed Muslim households will be analysed according to the 77
income level and socio-economic status of the individual house• hold.
The survey provides comprehensive information on the household's total income, source of income, and the recipient of the income. Firstly, the household's total income varies significantly. For example, about 52 percent of the total households received between 1,000 to 10,000 rupees per year.
Another 25 percent received between 11,000 to 20,000 rupees.
Secondly, the source of income for a number of surveyed house• holds is based mainly on the head of household's main income.
However, in agricultural settlements, some additional income had been received from the sale of cattle, gardening, and land
leasing, while in the business sectors income from vehicles and
rent. Finally, family members other than the head of household,
wife and children (especially male children), contribute a
significant amount of service to the family economy specially
in primary sectors and business activities.
The surveyed Muslim households are divided into four
major social groups (Table 5.2). The criteria for the classifi•
cation is based on the socio-economic standard of households,
including their level of income. As Table 5.2 indicates, about
40 percent of the total households are in the lower socio•
economic class while only 7 percent are in the upper socio•
economic class. In addition, a considerable proportion of hou•
seholds (38.4 percent) has been found to be in the lower middle
class. On the other hand, the percentage of social groups in
each settlement varies greatly, and the settlements or house- 78 holds engaged in non-primary activities are economically bet• ter-off than the others . For example, Weligama, a business settlement, is comparatively richer than many agricultural settlements, such as Wanahapuwa, and Kakunugolla. Moreover, the survey information on housing conditions, usage of pipe water and electicity indicate that social stratification varies grea• tly within the settlement itself (Appendix 3).
5.4 The Determinants of Fertility
Fertility is the number of actual live births to a woman during her child-bearing years. Fecundity, on the other hand,
is the capacity to have a number of births (Jones, 1974 : 111).
The fertility behaviour of a particular society is affected by a wide variety of psychological, socio-economic, and physiolo• gical factors. Although fertility is an individual behaviour,
the behaviour of individuals is guided by the norms of the
group to which they belong. Therefore, in this section, rele• vant socio-economic and demographic variables which influence
the fertility behaviour of Muslims of Sri Lanka will be iden•
tified from the surveyed data for the purpose of further analy•
sis.
The national statistics on fertility in Sri Lanka indi•
cate a comparatively higher rate for Sri Lankan Muslims (Chap•
ter 3). The survey data indicate that the mean live birth per
ever married women is 5.1 among Muslims, while the national
average is (W.F.S., 1978). In addition, as Table 5.3 indi•
cates, about 25 percent of the total households have had seven
or more pregnancies until the date of survey. However, these 79
TABLE 5.2
SOCIO-ECONOMIC STATUS OF THE SURVEYED HOUSEHOLDS
Socio-Economic Number of Percentages Status Households
Lower 1 32 40.9 Lower middle 1 24 38.4 Upper middle 45 13.9 Upper 22 6.8
Total 323 100.00
SOURCE: Field Survey, 1981. are variations in fertility level among the surveyed settle• ments. For example, Weligama, with an average of 6.63 pregnan• cies, has the highest rate while Kathankuddy, with 3.68 pregna• ncies has the lowest (Appendix 3). In addition, the pregnancy rates vary significantly among educational, occupational, and social groups as well.
TABLE 5.3
THE NUMBER OF TOTAL PREGNANCIES OF SURVEYED POPULATION
Number of Number of Percent Frequency Households
1-2 60 18.2 3-4 89 26.8 5-6 90 29.8 7-8 48 14.5 9-10 24 7.2 11 and more 1 2 3.6
SOURCE: Field Survey, 1981. 80
The status of women has been found to be closely related
to fertility behaviour in many societies in developing coun•
tries (United Nations, 1975). In addition, it has been stated
that the prevailing low status of Muslim women has been the cause of high fertility among this community in many countries
(Youssef, 1978; Allman, 1978). Moreover, the equal educational
opportunities, economic independence, and social participation
of women in a society all directly or indirectly influence the
fertility behaviour.
While Sri Lanka has been known for equal education
opportunities for males and females for many decades, the
educational level for the Sri Lankan Muslim males and females
is comparatively low (see Chapter 3). A low level of aspiration
for higher education for both male and female children among
Muslims is due to religious resistance during the last colonial
period (see Chapter 1). A number of Muslims still believe that
educating their children in non-Muslim schools would lead to
the socio-religious and cultural destruction of their society.
Although this attitude has been changing in recent years, to
some extent, it still remains strong with regard to females.
Not only are females discouraged from attending non-Muslim
schools but also co-educational schools. Thus, the survey indi•
cates that one half of the total women who were interviewed for
fertility information were illiterate (the ability to read the
Holy Qur'an is not considered as a sign of literacy). In addi•
tion, as Table 5.4 indicates, the majority of the women have
had less than grade five education. However, this has been 81 changing in recent years. For example, in the survey, the educational level for children who completed schooling does not show any significant differences between males and females. On the other hand, as expected, the rates of literacy and the level of education for wives vary among the surveyed settle• ment. For example, some settlements (Erukkalampiddy, Puttalam, and Muthuwela) have the higest literacy rates (more than 90 percent) while some others (Weligama, Pangollamada and Ninta- vur) have very low levels--less than 30 percent (Appendix 3).
Importantly, the settlements which had higher literacy rates and educational levels for females also had separate high schools for girls.
The female workforce participation is very low among Sri
Lankan Muslims. For example, as Table 5.5 indicates, only about
10 percent of the total women for whom fertility, data were recorded for fertility analysis are reported to be working
TABLE 5.4
THE WIFE'S LEVEL OF SCHOOLING
Level of Schooling Number of Cases Percent
No schooling 54 16.7 Grade 1-5 121 37.5 Grade 6-9 95 29.4 Grade 10-11 52 16.1 More than grade 11 01 0.3
Total 323 100.00
SOURCE: Field Survey, 1981. 82 outside of their homes. On the other hand, more than 70 percent of them claimed that they were not even participating in their household economic activities such as agriculture, business etc. When compared with the World Fertility Survey General
Report (1978), the workforce participation of Muslim women in the survey is very low. This is due to the low educational opportunities for women and importantly the distant location and different environment of the workforce. For example, wo• men's workforce participation in both agricultural and business activities were avoided because the working places, paddy fields and business firms, were located far away from the
Muslim settlements which usually were surrounded by non-Muslim settlements. Therefore, women's work participation outside the home has not been socially esteemed. However, in an extreme economic situation, the women do work outside or inside the home in order to increase total family income. A good example is Wanahapuwa which is one of the poorest settlements in the survey where more than 70 percent of the women work outside the home.
Since the female children of Muslim households terminate their education early and they do not effectively contribute to the household economy, their marriage is arranged by their parents at an early age. For example, the average age at mar• riage for the surveyed female population is 19. Moreover, the early age at marriage for females is prevalent among most of the surveyed settlements (Appendix 3). According to this, near•
ly three quarters of the women surveyed married before twenty 83 years of age. In addition, about 40 percent married before seventeen years of age. The national statistics on marriage age indicate that on average Muslim women marry four years earlier than does the average Sri Lankan woman.
Two other variables which were examined among the sur• veyed households are infant and child mortality. Infant morta• lity refers to the death of a child within the first twelve months of its life; child mortality, on the other hand, is death within the first five years. Both measurements not only predict the fertility trends but also indicate the socio• economic conditions of the surveyed household. Table 5.6 summa• rizes the data on infant and child mortality for the 323 house• holds. According to Table 5.6, 67 out of the 323 households have lost from 1 to 100 percent of the total pregnancies.
Moreover, a similar pattern has been found among most of the surveyed settlements but for Puttalam and Hemathagama infant mortality rates were .001 per live births. Finally, as previou• sly reported, several socio-economic and demographic variables such as husband's education level, family economic status, average birth interval, etc. have also been found to be impor• tant factors in this study.
5.5 Variables and Their Measurements
Finally, twenty six socio-economic and demographic va• riables were selected for further analysis of the fertility behaviour of Muslims. Table 5.7 presents the names and the measurements of the selected variables.
Among the above 26 variables, the number of pregnancies 84
TABLE 5.5
THE WIFE'S WORKFORCE PARTICIPATION
Type of Work Number of Cases Percent
Housekeeping only 227 70.3 Housekeeping & helping to household economy 50 15.5 Housekeeping & small income earning work at home 1 1 3.4 Labourer 24 7.4 Own work 6 1.8 Goverment employee 5 1 .6
Total 323 100.00
SOURCE: Field Survey, 1981.
TABLE 5.6
INFANT AND CHILD MORTALITY
Percentage of Percentage of Percentage of Death Infant Death Child Death
None 79.8 74.4 1 -24 11.7 14.8 25-29 5.4 7.2 50-74 0.6 0.9 75 and more 2.4 2.7
Total 100.00 100.00
SOURCE: Field Survey, 1981. of a women whether within the present marriage or previous marriage is considered as the dependent variable. The pregnan• cies not resulting in live birth are not included while the
infants that die after birth are included in this measurement. 85
The other socio-economic and demographic variables are consi• dered as background variables. Among them, the wife's level of schooling, her age at marriage, socio-economic status, and child mortality and considered as intervening variables or intermediate variables.
TABLE 5.7
VARIABLES AND MEASUREMENTS USED IN THE ANALYSIS
Variable Measurement
Number Name
1 Family Size Exact Number of Family Members 2 Wife's Literacy Dichotomized: (read, write and (1) Illiterate (2) Literate understand at least one language) 3 Husband's Level of Level (Grade) of School Completed Schooling Altogether Seven Categories 4 Wife's Level of Scholling 5 Husband's Occupation Rankings of Occupation Based on Status and Education Seven Categories 6 Wife's Occupation Rankings of Occupational Acti• vities Based on Day to Day Work Seven Categories 7 Number of Rooms Exact Number of Rooms in the House 8 Water for Drinking Four Categories 9 Energy for Lighting Dichotomized: (a) Kerosene (b) Electricity 10 Type of Toilet Rankings Based on the Standard of Health 11 Wife's Reading Dichotomized: (1) No (2) Yes Habits 12 Wife's Income Proportion of the Total Family Income 13 Total Family Exact Income in Rupees Per Year Income 14 Socio-Economic Rankings Based on the Socio- Status Economic Standard Six Categories 86
TABLE 5.7 (Continued)
15 Wife's Age at Exact Age at Marriage Marriage 16 Wife's Present Age Exact Age at the Time of Survey 17 Literacy Level for Proportion to the Total Family Male Family Members Literacy 18 Literacy Level for Proportion to Total Family Female family Members Literacy Rate 19 Literacy Level for Adult Family Members " " 20 Percentage of Male Births Percentage of Male Births to Total Births of a Mother 21 Infant Mortality Percentage of Infant Death (Aged Below One) 22 Child Death Percentage of Child Death (Aged Below 5) 23 Mean Average Birth Exact Interval in Years Interval 24 Number of Pregnancies Exact Number 25 Urban-Rural Based on Administrative Definition 26 Cultural Region Four Categories
SOURCE: Field Survey, 1981. 87
CHAPTER 6
VARIATION IN FERTILITY BY GEOGRAPHIC AND SOCIO• ECONOMIC FACTORS
This chapter will briefly examine the variation in fer• tility among 323 surveyed Muslim households. In addition, the examination of such differences will be analysed by using cross-tabulation between fertility rate and background varia• bles. The information on fertility is based on the ever-married women (the wife of each household). The background information, on the other hand, is mainly taken from the selected variable list (Table 5.7). The present age of women will be controlled when examining the relation between background variables and fertility because of its strong correlation with fertility e.g. young women will have fewer children. Other background variab• les will also be used as controlled variables whenever necessa• ry and possible. The explanation of the findings from the cross-tabulation will be accompanied by the observations which were made during the field survey as well.
6.1 Variation in Geographical Locations
One of the significant fertility variations in Sri Lanka has been by geographical location. Historically, in Sri Lanka, fertility has varied by provinces, districts, ecological zones and urban-rural places (Abehayaratne, 1967; W.F.S., 1978).
Therefore, the variation of Sri Lankan Muslim fertility as well has been analysed for geographical location such as urban-rural places, Sinhalese-Tamil cultural and sub-cultural regions, 88 economic characteristics etc. In the end, it has been found that rural-urban and cultural regions are significantly corre• lated with the fertility variation among the surveyed popula• tion.
Table 6.1 presents the mean number of live births per ever-married women for various age groups. The variation in fertility is evident for rural-urban places and cultural re• gions when controlled for wife's present age. That is, in general, Muslims in rural areas and in Sinhalese areas have higher fertility than urban and Tamil area Muslims. In addi• tion, an inverse relation of geographical location to fertility is evident across the age cohorts as well. However, the women aged 45 years and over in urban centres and 35-45 years in
Tamil cultural regions have been found to be exceptions to the above . inverse relation of geographical location to fertility.
It is obvious that the importance of geographical location is less strong among the older cohorts.
The association of geographical location to fertility when controlling for such variables as wife's age at marriage, level of schooling, and social status is examined in Table 6.2.
It is evident that the geographical location undoubtedly in• fluences the fertility behaviour of the surveyed population even when other background variables are controlled. When con• trolled for other variables as well, a similar inverse relation of geographical location to fertility has been found. However, the cultural variation is stronger than rural-urban differences in many examples. In other words, the surveyed population that 89
TABLE 6.1
MEAN NUMBER OF LIVE BIRTHS PER EVER-MARRIED WOMAN BY WOMEN'S AGE AND GEOGRAPHICAL LOCATIONS
Geographical Location
Rural-Urban Cultural Area
Age Group Rural Urban Sinhalese Tamil
< 24 2.6(10) 2.1(15) 2.8(6) 2.1(19) 25-34 4.1(61) 3.4(31) 4.0(41 ) 3.7(51) 35-44 6.3(64) 4.9(26) 5.7(49) 6. 1(41) > 45 5.4(75) 7.3(75) 6.4(79) 5.5(37)
SOURCE: Field Survey, 1981.
TABLE 6.2
MEAN NUMBER OF LIVE BIRTHS PER EVER-MARRIED WOMEN BY WOMEN'S AGE AT MARRIAGE, LEVEL OF SCHOOLING,SOCIO• ECONOMIC STATUS AND GEOGRAPHICAL LOCATION
Geographical Location
Rural-Urban Cultural Area
Rural Urban Sinhalese Tamil
Marriage Age
< 17 5.8(80) 5.2(41 ) 6.4(49) 5.0(72) 18-20 4.9(87) 3.8(21 ) 5.0(62) 4.2(46) 21-23 4.8(32) 7.3(20) 6.4(34) 4.7(18) > 24 4.2(11) 3.9(31) 4.2(30) 3.6(12)
Wife's Level of Schooling
< 5 5.4(29) 5.9(46) 6.0(104) 4.8(71) 6-9 5.0(56) 4.8(39) 5.4(47) 4.5(48) > 9 4.2(25) 3.6(28) 3.5(24) 4.2(29) 90
TABLE 6.2 (Continued)
Soc io-Economic Status
Lower 5.3(98) 4.7(34) 5.2(83) 5.1(49) Lower Middle 4.9(86) 5.6(38) 5.7(65) 4.5(59) Upper Middle 5.9(20) 4.2(25) 6.4(14) 4.3(31 ) Upper 4.2(6) 5.3(16) 5.9(13) 3.7(9)
SOURCE: Field Survey, 1981. lives in the Tamil cultural areas has had a comparatively low fertility than those in Sinhalese areas when controlled for several background variables.
The low fertility trend in Tamil cultural area has been further investigated by sub-cultural regions and one of many examples is reported in Table 6.3. A similar trend has been noted between two of the above variables even when controlled for the background variables. Therefore, it is evident that
TABLE 6.3
MEAN NUMBER OF LIVE BIRTHS PER EVER-MARRIED WOMAN BY WIFE'S LEVEL OF SCHOOLING AND SUB-CULTURAL REGIONS
Level of Education Sub-Cultural Region
Central & South North West East Coast
< 5 6.0(104) 5.4(11) 4.8(60) 6-9 5.4(47) 4.8(25) 4.2(23) > 9 3.6(24) 4.4(24) 3.4(5)
SOURCE: Field Survey, 1981.
East coast settlements have had comparatively lower fertility 91
than any other sub-cultural area. What then, are the factors
influencing the fertility variation among cultural regions? The
living environment and minority feeling of the surveyed popula•
tion in different cultural regions might influence their ferti•
lity behaviour. For example, Muslims have been found to cons•
titute the majority in a number of administrative districts,
A.G.A. divisions, and Gramasevaka divisions in Eastern province
(East coast region) and to some extent, in Mannar and in Putta•
lam districts (North West region) as well. In contrast, Muslim
settlements are scattered in pockets throughout the Sinhalese
area. Therefore, it could be speculated that concern for reli•
gio-ethnic identity has positively influenced the fertility
behaviour of the surveyed population who live in Sinhalese
cultural area more than it has for those who live in the Tamil
cultural area. However, the regional (cultural) variation in
fertility among Sri Lankan Muslim needs further investigation
for a firm conclusion.
6.2 Education and Fertility
Historically, of all characteristics, education probably
has had the most consistent relationship to fertility. The more
highly educated people have tended to have fewer children, a
phenomenon that can be observed in almost all societies (Chap•
ter 2). Therefore, the association of wife's level of schooling
to fertility, when controlled for wife's present age and age at
marriage, is examined in Table 6.4. An inverse relation of
level of schooling to fertility is clearly evident when either 92
TABLE 6.4
MEAN NUMBER OF LIVE BIRTHS PER EVER-MARRIED WOMAN BY WOMEN'S PRESENT AGE, AGE AT MARRIAGE, AND LEVEL OF SCHOOLING
Level of Schooling
Below Grade 5 Grades 6-9 Grades 10 & More
Age Group
< 24 2.3(13) 2.5(8) 1.8(4) 25-34 4. 1 (38) 3.6(27) 3.6(27) 34-44 6.6(39) 5.7(33) 4.7(18) > 44 6.2(85) 6.0(27) 4.2(4)
Marriage Age
<17 5.7(77) 5.6(32) 4.8(12) 18-20 4.8(54) 4.8(35) 3.9(19) 21-23 7.8(26) 4.4(12) 3.1(14) >23 3.7(17) 4.2(16) 4.0(8)
SOURCE: Field Survey, 1981. wife's present age and age at marriage is controlled. However,
fertility has its greatest decline with grade 10 and more education. In addition, a similar inverse relation of level of
schooling to fertility has been found when controlling for husband's and children's education as well.
The association of female education with fertility has been re-examined when controlled for the social status of the
surveyed households and is presented in Table 6.5. Table 6.5
indicates that the influence of female education on fertility
is much more significant other than for member of the lowest
social group. Thus, the association of female education with
fertility still remains stronger though women's present age is 93
not taken into account. Table 6.6 presents the relationships between the level of education and fertility when controlled for husband's occupation and wife's workforce participation.
Firstly, as Table 6.6 illustrates, the effect of female educa• tion seem weakest among the women who are the wives of labour• ers and sales workers and/ or are economically poor while others have shown comparatively a stronger influence of female education on fertility. Likewise, as shown in the Table 6.6, the influence of education is much stronger among the women who have worked outside their home. Therefore, in the above analy• sis (as illustrated in the Tables) when another variable is controlled, female education remains a strong predictor of fertility.
TABLE 6.5
MEAN NUMBER OF LIVE BIRTHS PER EVER-MARRIED WOMAN BY SOCIAL GROUP AND LEVEL OF EDUCATION
Soc io-Economic Level of Schooling Status
Below Grade 5 Grades 6-9 Grade 10 & More
Lower 5.1(80) 5.5(35) 4.6(17) Lower Middle 6.0(66) 4.5(33) 3.7(25) Upper Middle 5.7(18) 4.7(20) 3.6(7) Upper 5.6(11) 5.3(7) 3.0(4)
SOURCE: Field Survey, 1981. 94
TABLE 6.6
MEAN NUMBER OF LIVE BIRTHS PER EVER-MARRIED WOMAN BY WIFE'S WORKFORCE PARTICIPATION, HUSBAND'S OCCUPATION, AND LEVEL OF SCHOOLING
Level of Schooling
Below Grade 5 Grades 6-9 Grade 10 & More
Workforce (wife's)
Housekeeping only 5. 2(118) 4.,8(71 ) 4.,3(38 ) Housekeeping & helping household economy 6. 1(31) 5.,2(13 ) 3..8(6 ) Housekeeping & small income earning work 7. 4(5) 3.,0(3 ) 3,.0(3 ) Labourer 5. 3(20) 6.,0(3 ) 3..0(3 ) Own work & govern• ment employee 8. 0(1 ) 5..6(5 ) 1 .8(5, )
Husband's occupation
Labourer 4. 5(21 ) 5..3(4 ) 4,.6(5 ) Farmer 6. 1(24) 4,.9(10 ) 2,.0(3 ) Salesworkers 4. 8(25) 5..3(9 ) 4,.3(4 ) Business 7. 2(22) 4,.9(16 ) 3,.8(9 ) Government employee 6. 2(8) 5,.6(14 ) 3,.5(11 ) Others 5. 5(66) 4,.8(38 ) 4,.3(20 )
SOURCE: Field Survey, 1981.
6.3 Female Age at Marriage and Fertility
Several studies indicates that increasing female age at
marriage has strongly influenced recent fertility decline in
Sri Lanka (Wright, 1970; Fernando,1979). Table 6.7 presents the
association of age at marriage and mean birth interval with
fertility for various age groups. As Table 6.7 indicates, an
inverse relation is evident across the age groups. In other
words, women who married early have had a comparatively higher 95
number of live births than the women who married later. There•
fore, age of women is an important factor in fertility, indepe•
ndent of the age of women. However, the women aged twenty five
years and over who married between twenty one to twenty three
years show an unusually high fertility compared to the women
who married between the ages of eighteen and twenty years. In
addition, as Table 6.7 indicates, the same age group unusually
'has the highest mean number of live births with shorter birth
interval. Both results suggests that a group of women who
married in their early twenties has comparatively higher ferti•
lity with a shorter birth interval.
TABLE 6.7
MEAN NUMBER OF LIVE BIRTHS PER EVER-MARRIED WOMAN BY WOMEN'S PRESENT AGE, MEAN BIRTH INTERVAL AND AGE AT MARRIAGE
Age at Marriage
<17 18-20 21-23 24 and more
Age Groups
<24 2.6(15) 2.2(6) 1.3(4) 25-34 4.6(31) 3.6(33) 3.7(20) 2.1(8) 35-44 6.9(39) 5.4(30) 6.1(8) 4.0(13) >44 6.3(36) 5.4(39) 8.6(20) 4.7(21 )
Birth Interval
<2 years 5.9(59) 4.7(46) 6.9(29) 4.0(22) >2 years 5.3(62) 4.9(62) 4.3(23) 4.0(20)
SOURCE: Field Survey, 1981. 96
The wife's level of schooling has been found to be
important with relation to fertility in the previous analysis.
In this section the effects of female age at marriage on
fertility still remain stronger even when controlled for female
level of schooling. Likewise, a similar strong relation has been found between fertility and age at marriage when control•
led for child mortality (Tables not shown). The women experien• cing child deaths have shown a stronger relation and have had a comparatively larger number of live births than the women who have not had this experience.
Interestingly, the relationship between age at marriage and fertility is not clear among the different socio-economic
groups. For example, as Table 6.8 indicates, the effects of age at marriage on fertility is weakened, or even disappears when
socio-economic status is controlled.
6.4 Mean Birth Interval and Fertility
The mean birth interval is related to fertility when
controlled for several background variables. In this first
instance, the influence of the wife's present age has been
removed in the relation between birth interval and fertility
and the results are presented in Table 6.9. It evident that an
inverse relation of mean birth interval to fertility exists
across the age groups. However, when controlled for other
background variables, different and interesting trends between
birth interval and fertility has been noted.
Table 6.10 presents the association of mean birth inter- 97
TABLE 6.8
MEAN NUMBER OF LIVE BIRTHS PER EVER-MARRIED WOMAN BY SOCIO-ECONOMIC STATUS AND AGE AT MARRIAGE
Socio-Economic Status Age at Marriage
< 1 7 18-20 21 -23 >23
Lower 5. 9(49) 4.8(48) 5. 1(19) 4.2(16) Lower Middle 5. 4(41 ) 4.5(49) 6. 7(23) 3.6(11) Upper Middle 5. 3(22) 5.4(7) 5. 0(7) 3.8(9) Upper 5. 7(9) 4.3(4) 5. 0(3) 4.5(6)
SOURCE: Field Survey, 1981.
TABLE 6.9
MEAN NUMBER OF LIVE BIRTHS PER EVER-MARRIED WOMAN BY WOMEN'S PRESENT AGE AND MEAN BIRTH INTERVAL
Age Group Birth Interval
Below Two Years Above Two Years
< 24 2.8(12) 1.9(13) 25-34 4.2(54) 3.4(28) 35-44 6.0(45) 5.8(45) > 44 7.2(45) 5.7(67)
SOURCE: Field Survey, 1981. 98
TABLE 6.10
MEAN NUMBER OF LIVE BIRTHS PER EVER-MARRIED WOMAN BY BACKGROUND VARIABLES AND MEAN BIRTH INTERVAL
Background Birth Interval Var iable
Below Two Years Above Two Years
Social Status
Lower 5.3(66) 5.0(66) Lower Middle 5.6(64) 4.6(60) Upper Middle 5.7(16) 4.5(29) Upper 5.6(10) 4.5(12)
Husband's Occupation
Business 6.5(30) 4.6(17) Farmer 6.1(19) 4.9(18) Labourer 4.9(8) 4.6(22) Government Employee 4.8(19) 5.4(15) Sales Workers 4.8(18) 5.0(20)
Wife's Economic Activity
Housekeeping 5.6(136) 4.7(152) Working Outside 4.4(20) 5.6(15)
SOURCE: Field Survey, 1981 val with fertility when controlled for social class, husband's occupation and wife's economic activities. Firstly, in Table
6.10, an inverse relation is evident when controlled for social status. However, no relationships have been found vertically
(among social classes) between the two variables. On the other, hand, an interesting finding has been noted in Table 6.10 when controlled for husband's occupation. The wives of farmers and businessmen have experienced a large number of pregnancies 99 with short birth intervals. In contrast, wives of government servants and sales workers have spaced their birth intervals but still have a large number of pregnancies. Likewise, the women who work outside their homes have experienced a large number of pregnancies with long birth intervals.
These results suggest that the wives of the government
servants and sales workers and women who work outside their homes probably use birth control methods to space their child-
bearing but still have many children. The reason for this desire to have a large number of children may be because of the
large family size norm which still, to some extent, prevails among all section of this community. Therefore, even the people who are comparatively educated and economically active
while spacing their child-bearing have achieved large number of
live births which they think is an ideal family size for them.
Finally, Table 6.11 presents the relationship between mean birth interval and fertility when controlled for the
percentage of male births and child mortality. As Table 6.11
indicates, the sex distinction of children may be one of the
better predictors of fertility since the parents who had not
had an equal number of male births have had higher fertility
with short birth intervals. However, the effect of spacing on
fertility disappears or weakens when percentage of male births
is taken into account. Finally, the Table also indicates that
the birth interval still remains a predictor when controlled
for the percentage of child mortality. 1 00
TABLE 6.11
MEAN NUMBER OF LIVE BIRTHS PER EVER-MARRIED WOMAN BY PERCENTAGE OF MALE BIRTHS, CHILD MORTALITY AND MEAN BIRTH INTERVAL
Birth Interval
Below Two Years Above Two Years
Percentage of Male Live Births
< 50 6.4(69) 4.5(59) > 50 4.7(71 ) 4.9(108)
Percentage of Child Mortality
1-1 00 6.9(40) 5.0(116) None 6.4(38) 4.3(129)
SOURCE: Field Survey, 1981.
6.5 Child Mortality and Fertility
Table 6.12 illustrates the mean number of live births per ever-married woman and the percentage of child mortality when controlled for women's age groups, age at marriage, and
the level of schooling. The controlled variables previously have been found to influence fertility. Therefore, in this
section, those variables have been controlled in relation bet• ween child mortality and fertility. A positive relation is
evident in this analysis. In other words, the women who had not
had a child death had fewer live births than those who have had
one or more children die even when other important variables
are controlled. It should also be noted that comparatively more 101
TABLE 6.12
MEAN NUMBER OF LIVE BIRTHS PER EVER-MARRIED WOMEN BY PRESENT AGE, AGE AT MARRAIGE, SOCIAL STATUS AND CHILD MORTALITY
Percentage of Child Mortality
1-100 None
Age Group
< 24 4.0(1) 2.2(24) 25-34 5.3(14) 3.6(78) 35-44 6.6(25) 5.6(65) >44 7.3(38) 5.5(78)
Marriage Age
< 17 7.7(32) . 4.8(87) 18-20 6.1(27) 4.2(81) 21-22 6.0(8) 5.7(44) > 23 4.9(9) 3.8(83)
Wife's Level of Schooling
< 5 6.4(45) 5.2(130) 6-9 7.0(25) 4.2(70) > 9 5.5(8) 3.6(43)
SOURCE: Field Survey, 1981.
educated women who married at a later age have been found to
have low child mortality and low fertility when compared to
other women or vice versa.
The influence of child mortality on fertility has also
been analysed when controlled for the socio-economic indicators
of households such as social status, types of drinking water,
the availability of latrines etc. A similar positive relation
of child mortality to fertility is evident in those circumstan- 1 02 ces as well. Finally, the evidence, when controlled for house• hold's occupation and wife's workforce participation, has stre• ngthened the argument that the child mortality strongly influ• ences the fertility behaviour of the surveyed population.
6.6 The Findings from Pearson and Partial Correlat ions
The relation between background variables and fertility has been analysed by using the Pearson correlation coefficient analysis. The significant background variables and their 'r' and 'p' values along with the total number of cases are report• ed in Table 6.13. As Table 6.13 indicates, the wife's level of schooling, age at marriage and child mortality have been found to be significantly related to fertility than other background variables. However, the manner in which fertility varies with the age of women is well known and it is r = -.23 in this sample. Therefore, the impact of present age of women on ferti• lity has been investigated by using partial correlation analy• sis. Partial correlation coefficient analysis enables us to remove the effect of the controlled variable, the present age of women, from the relationships between background variables and fertility without manipulating the row data. Thus, the findings of partial correlation analysis are presented in Table
6.14. Interestingly, as Table 6.14 indicates, the wife's age at marriage, her workforce participation and child mortality re• mained significantly related to fertility while some other variables have become weaker than the pearson correlation ana- 103
TABLE 6.13
THE RELATIONSHIP BETWEEN SELECTED BACKGROUND VARIABLES AND NUMBER OF LIVE BIRTHS
Variable Pearson Correlation
r p n
Wife's level of schooling -.11 .02 323 Wife's literacy level -.07 .09 323 Husband's level of schooling -.06 NS 309 Wife's workforce participation NS NS 312 Husband's occupation NS NS 312 Wife's age at marriage -.06 . 1 3 323 Wife's present age -.23 .000 323 Infant mortality .07 .8 323 Child mortality . 1 1 .01 323
SOURCE: Field Survey, 1981.
TABLE 6.14
THE RELATIONSHIP BETWEEN SELECTED BACKGROUND VARIABLES AND NUMBER OF LIVE BIRTHS WHEN CONTROLLED FOR WOMEN'S PRESENT AGE
Variables Partial Correlation
r p n
Wife's level of schooling NS NS 323 Wife's 1iteracy level NS NS 323 Husband's level of schooling NS NS 323 Wife's workforce participation -.20 .01 312 Husband's occupat ion NS NS 312 1 04
TABLE 6.14 (Continued)
Wife's age at marriage -.23 .00 323 Wife's present age - - Infant mortality NS NS 323 Child mortality .19 .01 323
SOURCE: Field Survey, 1981.
lysis.
6. 7 The Summary of Findings
Firstly, among the geographical factors, cultural re•
gions have shown a clear variation in fertility. The fertility
rate is comparatively low among the surveyed population who
live in Tamil cultural area than Sinhalese (Table 6.15). Secon•
dly, the female level of schooling has been found to be of the
stronger predictors of fertility even when controlled for a
number of background variables. Thirdly, the effect of increa•
sing female age at marriage is evident in many analysis. Howe
TABLE 6.15
MEAN NUMBER OF LIVE BIRTHS PER EVER-MARRIED (WOMEN) SURVEYED POPULATION
Population Categories Number of Live Births
Total population 5.1 Rural 5.2 Urban 5.0 Sinhalese cultural region 5.5 Tamil cultural region 4.6
SOURCE: Field Survey, 1981. 1 05 ver, a group of women who married in their twenties have shown higher fertility than other women in this study (Table 6.16).
Fourthly, the variation in fertility among socio-econo• mic groups is not clear in many analyses. The relation of socio-economic status to fertility weakens or disappears in many instances. As a result, a large number of pregnancies is commonly found among many socio-economic and occupational groups. Among them, the wives of farmers and businessmen have achieved comparatively large number of pregnancies with a short birth interval while the wives of government servants and sales workers have also shown a similar trend with longer birth interval. Finally, the fertility is found to be influenced by sex preference and child mortality as well.
Table 6.16 summarizes the specific influence of certain variables on fertility. Among the several investigated variab• les, wife's present age, level of schooling and child mortality have strong relation even when controlled for other background variables. 1 06
TABLE 6.16
SPECIFIC INFLUENCE OF CERTAIN VARIABLES ON FERTILITY OF MUSLIMS OF SRI LANKA
Variables for Other Variables Category in Category Which Influ- Whose Influence Which Ferti- in Which ence is is Controlled lity is Fertility Measured Lower is Highest
Wife's level Present age, Grade 10 and Grade 5 & of schooling age at marriage, above below social class, wife's workforce participation, and husband's occu• pation . Wife's Age at Present age, Age 24 & Ages 21- Marriage birth interval, above 23 years and social class Wife's Work• Wife's level Government Housekeep• force part• of schooling, employees ing women icipation husband's occu• pation, and social status Husband's Social class, Grade 10 & Grade 5 & Level of husband's occu• above below Schooling pation, and wife's workforce parti• cipation Husband's Husband's level Labourer Business• Occupation of schooling & men social class Social Class Husband's & Upper middle Lower wife's occu• class class pation, husband's and wife's school• ing Cultural Area Present age, age East coast Central & at marriage, & (Tamil) Southern wife's level of (Sinhalese) schooling
SOURCE: Field Survey, 1981. 107
CHAPTER 7
MULTIVARIATE ANALYSIS OF FERTILITY
The purpose of this chapter is to analyse the fertility behaviour of Sri Lankan Muslims in a multivariate framework using the 323 surveyed households. Path analysis has been
selected to measure the causal relationship of background va•
riables to fertility. The analysis is based on five independent variables shown to be significantly correlated to fertility.
The number of children born to ever-married women will be considered as the dependent variable. The analysis of the 323 households is further specified by examining different age
groups of women in order to compare the causal patterns of
fertility among the sub-population groups.
The analysis of cross tabulation and correlation in the
previous chapter conveyed substantial information on the ferti•
lity mechanism of the Muslims of Sri Lanka. However, those
findings have failed to indicate the causal relationships and
the strength between background variables and fertility. In
fact, fertility is determined by interrelated independent va•
riables. Therefore, it is necessary to determine how and what
combination of factors "causes" fertility patterns. The path
model, therefore, enables us to hypothesize a complex causal
structure, draw a visual display of a model and calculate the
strength of the various relationships to fertility (Asher,
1976; Kendall and O'Muiracheartaigh, 1977).
The first step in the process of building a path model 108
is the selection of a set of significant background variables
in relation to fertility. An examination of the relationships
of fertility to socio-economic, demographic characteristics in previous chapters has shown that five background variables
have stronger statistical relations to fertility. Those varia•
bles are wife's present age, her age at marriage, her level of
schooling, her socio-economic status and child mortality. These
five variables have been selected because they explain a consi•
derable part of the variation in fertility. Therefore, it
should be possible to examine the fertility behaviour of the
Muslims of Sri Lanka with the help of the above selected varia•
bles along with adequate sample size. Specific forms in which
the selected variables are coded and analysed are given in
Appendix 4. However, the analysis of fertility is more complex
and probably operates through a number of other intermediate
variables which were not included in this analysis. Therefore,
the findings of the path analysis here should be interpreted
with caution.
After identifying the dependent and independent varia•
bles in the path model, the second step is to build a causal
model that includes the temporal and causal ordering of varia•
bles and hypothesizes the effects of independent variables with
intervening and dependent variables. Diagram 7.1 shows the
causal ordering of the selected variables along with predicted
positive and negative relationships. In this model, the child•
ren ever born are treated as the dependent variable and five
other background variables are treated as either explanatory or FIGURE 7.1: PATH DIAGRAM OF CHILDREN EVER-BORN AND OTHER SOCIO-ECONOMIC AND DEMOGRAPHIC CHARACTERISTICS
SOURCE: Field Survey, 1981. 1 10
intervening variables. In addition, as indicated in Diagram
7.1, the causal order of variables is that child mortality is prior to the children ever born, social status is prior to child mortality and so on. The wife's present age is treated as an explanatory variable, though the purpose of its inclusion is
largely that of control.
The direction of expected causation is indicated by
arrows in Diagram 7.1. Therefore, several directions of causa•
tion could be hypothesized along the arrows. The determinants
of fertility are either through age at marriage, education and
present age or via child mortality to other variables etc. For
example, it could be hypothesized that women with little educa•
tion may marry early, may have more children die, and may
replace them resulting in high fertility. Likewise, other va•
riables, as well, are hypothesized to affect children ever born
with a varying degree of importance.
Finally, the following statistical steps were taken in
the proceeding path analysis. Firstly, the stepwise multiple
regression analysis is used with the path model.
Secondly, the beta coefficients which are comparable to stand•
ard regression coefficients are used as path coefficients.
2 . Finally, R , which represents the the proportion of variance explained in multiple regression analysis, is used in the
2
calculation of residual paths 1-R (Nie, et al, 1975: 391-92).
A path diagram which includes path coefficients for 323
surveyed women is shown in Figure 7.2. In addition, Table 7.1
presents further statistical information for the same analysis. FIGURE 7.2: PATH DIAGRAM OF CHILDREN EVER-BORN AND SOCIO-ECONOMIC AND DEMOGRAPHIC CHARACTERISTICS (EVER-MARRIED WOMEN) - 1981
SOURCE: Field Survey, 1981. 1 1 2
TABLE 7.1
PATH COEFFICIENTS FOR MODEL PREDICTING FERTILITY BEHAVIOUR OF MUSLIM WOMEN IN SRI LANKA (323 CASES)
Variable Total Causal Non- Pairs Covariance Relation Causal
Direct Indirect Total
Children ever born, .12 .06 -- .06 .06 child mortality Children ever born, soc io-economic status -.02 .01 -.01 0 -.01 Children ever born, age at marriage -.06 -.08 -.01 -.09 .03 Children ever born, level of schooling -.14 -.07 -.02 -.08 .06 Children ever born, present age .23 .21 .02 .23 0
Child mortality, soc io-economic status -.14 -.13 -- -.13 -.01 Child mortality, age at marriage -.07 -.09 — -.09 -.03 Child mortality, level of schooling -.09 0 -.03 -.03 -.05 Child mortality, present age .21 .22 -.02 .21 0
Soc io-economic status, level of schooling .10 .10 — .10 0 Soc io-economic, present age — — - . 0 3 -.03 —
Age at marriage, level of schooling .13 .19 — .19 -.06 Age at marriage, present age . 1 6 .21 -.05 . 1 6 0
Level of schooling, present age -.27 -.27 -.27 0
SOURCE: Field Survey, 1981. 1 1 3
7.1 A Predictive Model of Fertility Behaviour of Muslims of Sri Lanka
The findings of the path anaysis indicate that less than
10 percent of the total variance of fertility can be explained by those variables which means that 90 percent of the variance in children ever born remains unexplained by those variables.
The reason for low variance may be because the variables which were included in the analysis were selective; therefore, unex• plained factors remain higher or as several earlier studies indicate, the capacity of variance to explain is much lower in micro-data analysis than in aggregated data analysis (Balakris- hnan, 1979: 216).
However, the beta coefficients (direct causal effects) indicate that most of the variance in fertility was explained directly by the wife's present age (.21). In fact, the highest association of wife's present age is expected because young married women would have had fewer births that is, have incomp• lete families, than older women. Among the other variables, social status does not show any relation either by direct causal effects (.01) or total causal effects (.00) with ferti• lity. However, three other variables, wife's age at marriage, her level of schooling, and child mortality, whose causal effects range from .06 to .09, might explain fertility.
Among the three other variables, child mortality is somewhat weaker but has a direct causal effect (-.06) on ferti• lity. Secondly, wife's age at marriage is sronger in its direct causal effect (.08) on fertility; however, much of its effect is not through child mortality which means that young marriages 1 1 4
have more children but not necessarily due to replacement.
Finally, wife's level of schooling is shown to be stronger in its direct causal relation (-.07) as well as its indirect causal relation through age at marriage (-.02) on fertility.
The later causal path suggests that more educated women have fewer children by delaying their marriages or vice versa.
Therefore, fertility is indirectly caused by wife's level of schooling which is mediated by her age at marriage.
However, on the whole, the model does not provide a strong causal relation of background variables with fertility since beta coefficients are relatively low. In other words, the overall explanatory power of the model is weak. The weakness of the model may be due to the fact that all the age cohorts were taken together in an analysis. Sometimes, the casual effects of each variable with fertility may differ for different age cohorts if the age cohorts were analysed separately. Those findings might lead to a better understanding of fertility behaviour of a sub-population group in the surveyed population.
Therefore, the following section will apply path analysis sepa• rately for two age cohorts.
7.2 Fertility Behaviour of Women Aged Forty-Nine Years and Above
The first attempt was to analyse the fertility behaviour of women aged forty-nine years and above who already had comp• leted their child bearing. Such an analysis is important be• cause the effect of women's age, which explained much of the variance in fertility, no longer exists in these age cohorts. 11 5
The findings of the analysis are presented in Figure 7.3 and in
Table 7.2. Interestingly, the unexplained variance in fertility decreased to 86 percent in this analysis.
In this analysis two factors have shown a strong direct causal effect on fertility: wife's age at marriage (-.35) and
socio-economic status (.35). None of the factors show any
indirect causal effect on fertility through other variables. It means that fertility which is determined directly by wife's age
at marriage and socio-economic status is not through child mortality (see Figure 7.3 and Table 7.2). For example, high
fertility among early married women is not necessarily to
replace dead children. However, child mortality, itself, is
shown to have a somewhat stronger causal effect (.25) on ferti•
lity. This suggests that the women who have had more child
deaths replaced them with additional births which means high
fertility.
Wife's level of schooling shows the highest total causal
effect (.42) on fertility in this analysis. Much of its effects
were not direct but through wife's age at marriage. This causal
relationship suggests that educated women marry late and as a
result of that, end up with fewer children. This finding is
similar to the previous path analysis but the causal relation
of wife's level of schooling with fertility through age at
marriage is much stronger and clearer in this analysis.
The direct causal effects of wife's present age on
fertility have been reduced as expected but the relationship
has changed from positive to negative (-.17) compared to the FIGURE 7.3; PATH DIAGRAM OF CHILDREN EVER-BORN AND SOCIO-ECONOMIC AND DEMOGRAPHIC CHARACTERISTICS (EVER-MARRIED WOMEN AGED 49 AND ABOVE) - 1981
SOURCE: Field Survey, 1981. 1 1 7
TABLE 7.2
PATH COEFFICIENTS FOR MODEL PREDICTING FERTILITY BEHAVIOUR OF MUSLIM WOMEN, 4 9 YRS AND ABOVE (67 CASES)
Variable Total Causal Non- Pairs Covariance Relation Causal
Direct Indirect Total
Children ever born, child mortality .24 .25 .25 -.01 Children ever born, soc io-economic status .24 .35 -.02 .33 -.08 Children ever born, age at marriage -.28 -.35 -.04 -.38 .10 Children ever born, level of schooling .09 .07 .35 .42 -.33 Children ever born, present age -.12 -.17 .07 -.10 -.02
Child mortality, soc io-economic status -.12 -.07 — -.07 -.05 Child mortality, age at marriage -.18 -.14 — -.14 -.03 Child mortality, level of schooling -.05 -.03 0 -.03 -.02 Child mortality, present age . 1 5 . 1 3 .02 . 1 5 0
Soc io-economic status, level of schooling .03 .03 — .03 0 Soc io-economic, present age — — 0 0 0
Age at marriage, level of schooling 0 -.02 — -.02 .02 Age at marriage, present age -.12 -.12 0 -.12 0
Level of schooling, present age -.15 -.15 -.15 0
SOURCE: Field Survey, 1981. 1 18 previous analysis. This means that fertility of women aged forty-nine years and above decreases along with increasing wife's present age. The changed direction in this age group may be because of increased nutritional and health conditions of relatively younger women would have incresed the child bearing capacity compared to elderly women. In addition, the changed direction (from negative to positive) of socio-economic status with fertility is probably because the present measurement may not be too relevant for those women who had completed their child bearing long ago.
7.3 Fertility Behaviour of Women Aged Forty-Eight Years and Below
The third attempt in path analysis was to find out which combination of factors influence the fertility behaviour of women aged forty-eight years and below. The younger age cohorts are also important because the recent socio-economic and demo• graphic changes may be more prevalent among this age group than among the older cohorts. The findings of the path analysis are presented in Table 7.3 and in Figure 7.4.
Wife's present age has emerged as a strong predictor of fertility in this analysis because of the connection between fertility increase and increasing women age among the younger women (forty-eight years and below). Of the four other back• ground variables, child mortality (total causal effect=.03) and socio-economic status (total causal effect".04) are not good explanations of fertility of this group of women. Instead, for this group, wife's age at marriage is a somewhat weaker varia- FIGURE 7.4: PATH DIAGRAM OF CHILDREN EVER-BORN AND SOCIO-ECONOMIC AND DEMOGRAPHIC CHARACTERISTICS (EVER-MARRIED WOMEN AGED 48 AND BELOW) - 1981
SOURCE: Field Survey, 1981. 1 20
TABLE 7.3
PATH COEFFICIENTS FOR MODEL PREDICTING FERTILITY BEHAVIOUR OF MUSLIM WOMEN, 48 YRS AND BELOW (2 56 CASES)
Variable Total Causal Non- Pairs Covariance Relation Causal
Direct Indirect Total
Children ever born, child mortality .09 .03 — .03 .07 Children ever born, soc io-economic status -.07 -.03 C -.04 -.03 Children ever born, age at marriage -.04 -.06 -.01 -.06 .02 Children ever born, level of schooling -.14 -.08 -.01 -.09 -.06 Children ever born, present age .31 .30 .01 .31 0
Child mortality, soc io-economic status -.16 -.15 -- -.15 -.01 Child mortality, age at marriage -.05 -.07 -- -.07 .02 Child mortality, level of schooling -.07 0 -.04 -.04 -.03 Child mortality, present age .18 .18 0 .17 -.01
Soc io-economic status, level of schooling . 14 . 1 4 -- . 1 4 0 Soc io-economic, present age -.02 -.02
Age at marriage, level of schooling .22 .24 -- .24 -.02 Age at marriage, present age . 12 . 1 5 -.04 . 1 2 0
Level of schooling, present age -. 16 -.16 — -.16 0
SOURCE: Field Survey, 1981. 121 ble (total causal effect=-.06) and wife's level of schooling is the better predictor (total causal effect=-.09) of fertility.
Fertility patterns among women aged forty-eight years and below is different from the older women (aged forty-nine years and above). The wife's level of schooling and her age at marriage have become disconnected in their effects on the fertility behaviour of younger women. This means that with more education and more information about modern contraceptive tech• niques, attitudes towards having children have changed. As a result, younger, educated women who marry early have fewer children.
7.4 Summary
Fertility of the surveyed population (323 cases) is mainly determined by the wife's level of schooling which is mediated by her age at marriage. The finding suggests that women with little education marry early and during their longer child-bearing period achieve comparatively larger number of pregnancies. The trend remains unchanged and becomes stronger among the women aged forty-nine years and above who complete their child bearing. The trend, however, is different for younger women aged forty-eight years and below whose fertility
is mainly determined by two disconnected variables: the level of schooling and age at marriage. 1 22
CHAPTER 8
CONCLUSION
The final chapter will summarize the findings of this study and integrate then to give an overall picture of the different aspects of Muslim fertility behaviour. The conclu• sion will also outline some of the policy implications of my research.
The study is based on 323 Muslim households randomly chosen from eleven Muslim settlements in Sri Lanka. Twenty-six variables were used in cross tabulation and pearson correlation analysis. Among them, the six most highly correlated variables were used in the path analysis.
8.1 General Characteristics
The socio-economic and the demographic characteristics of the surveyed population vary within settlement and between settlements. Nearly 45 percent of the total households are poor and nearly two thirds are below the lower middle class level.
The majority of the population is engaged either in agriculture in rural areas or in saleswork in urban areas. The level of schooling among the adult population is very low; nearly 50 percent of the male (husband) and nearly 55 percent of female
(wife) heads of the households were only educated up to grade five. However, the level of schooling among young males and females is comparatively high. Female workforce participation
is very low, only about 10 percent of the total female house• hold heads being engaged in gainful employment outside their 123
home. Females marry early; their average age at marriage is
18.5 years. The mean number of live births per ever-married women is 5.1 and the average birth interval is less than two years. One out of four households has experienced up to 100 percent child mortality.
8.2 Major Findings
The wife's level of schooling has been shown to have a consistent inverse relationship with fertility. The more educa• tion women receive, the lower the number of pregnancies. How• ever, a major reduction in the total number of pregnancies occurs if the wife receives grade ten education or more. It means that the longer women stay in school, the higher the age at marriage. Education also increases the knowledge of contra• ception which directly and indirectly influence fertility. The inverse relation of education to fertility remains when contro• lled for other variables as well. Another important factor contributing to the high fertility of Muslims was the wife's age at marriage. It is strongly negatively correlated with fertility even when controlled for age of women, socio-economic status, and education. A considerable reduction in fertility is noted when women marry at twenty-two years of age. Likewise, women who are engaged in gainful employment outside their home have shown low fertility.
High fertility with a shorter birth interval is common among little educated, early married housewives who comprise the majority of the 323 cases in my sample. Therefore, until 1 24 the socio-economic and educational standards of females are improved, fertility level will likely remain high.
One of the important factors contributing to high ferti• lity seems to be the prevailing high child mortality rate.
Child mortality is strongly related to poor economic situa• tions, illiteracy, early age at marriage and short birth inter• val. As long as a high child mortality rate and the subsequant fear of child death prevail among Muslims, the scope of ferti• lity reduction will remain limited. In addition, the desire for male children also seems to be influencing fertility. Male children are only as a source of income right from childhood.
Fertility is higher among the rural poor, farmers, and urban businessmen whose economic orientation requires more labourers, preferably from the family. On the other hand, fertility is lower among the urban poor, salesworkers, and government employees who, with their limited income, feel that rearing many children is a heavy burden.
Fertility also varies between the two major cultural regions: the Sinhalese and the Tamil areas. The fertility is higher among Muslims who live in the Sinhalese areas and lower in the Tamil areas. Concern about ethnic identity and minority feeling may have led to higher fertility in the Sinhalese areas compared with Tamil areas. However, I strongly feel that this question needs further investigation in order to draw any firm conclusions. 1 25
8.3 Findings from Multivariate Analysis
Since many of the socio-economic and demographic vari• ables are highly correlated, the path analysis was employed to determine the strength and direction of the background vari• ables with fertility. The wife's present age, her level of schooling, age at marriage, socio-economic status, child morta• lity and children ever born are used in the path analysis.
The findings of the path analysis indicate that the fertility of the 323 surveyed women is mainly determined by their level of schooling which is, in fact, mediated by their age at marriage. The above trend is much stronger among the older women who have completed their child bearing (aged forty- nine years and above). Traditional fertility patterns have changed among younger women (aged forty-eight and below). Al• though fertility is mainly determined by female level of schoo• ling and her age at marriage, the direction of influence has changed among younger women. It means that both factors influe• nce fertility directly without any mediation. Therefore, the new trend among younger women indicates that if the female level of schooling increased, fertility will eventually decline regardless of increasing age at marriage.
The fertility of the Muslim population is mainly determ• ined by the female level of schooling. Female level of school• ing however, is directly and indirectly (through age at mar• riage and other factors) influencing fertility. For example, a woman with little education normally marries early, is engaged in household activities, may lose many children which results 1 26 in high fertility. The female level of schooling is found to be low among Muslims; therefore, fertility is comparatively high.
However, the younger women have shown a decline in fertility due to recent educational changes.
The low education level is common among all sections
(age, sex, geographical areas) of the Muslims population. Sta• tistically, as noted earlier, the educational achievements of
Muslim is far behind the other communities except Indian (Es• tate) Tamils. This has mainly been due to the resistance again• st the cultural and religious influence of the colonial power.
Until very recently, educational opportunities were available only up to the primary level in many Muslim settlements. How• ever, this has, to some extent, changed as a result of educa• tional reforms undertaken since 1965. During this period, Mus• lim schools were up-graded and males and particularly females were attracted to higher education with the provision of emplo• yment opportunities as school teachers. This and other socio• economic changes have resulted in changing demographic patterns among the younger generation.
Therefore, a significant reduction in fertility of Mus• lims is possible if the socio-economic conditions of females are further improved. Female education must be given first priority in this regard. The educational facilities in Muslim
settlements could be further improved to facilitate higher education for the male and the female population. The vocation• al education for females would be another way to provide educa• tion as well as employment opportunities for them. Muslim 1 27
organizations such as the Renassan Movement and Young Muslim
Men and Women Association must focus their attention on female
vocational education at the settlement level. 1 28
BIBLIOGRAPHY
Abhayaratne, 0. E. R. and C. H. S. Jayawardena. Fertility Trends in Ceylon. Colombo: Colombo Apothecaries Limited, 1967.
Agarwala, S. N. "Effect of a Rise in Female Age at Marriage on Birth Rate in India." In Proceedings of the World Popu• lation Conference - 1965. Vol 2. New York: United Nations, 1965.
. A Demographic Study of Six Urbanized Villages. Bombay: Asia Publishing House, 1970.
Ahmad, I. "Introduction." In Family, Kinship and Marriage Among Muslims in India. Ed. I. Ahmad. New Delhi: Manohar, 1 976.
Alam, I. and J. G. Cleland. "Illustrative Analysis: Recent Fertility Trends in Sri Lanka." World Fertility Survey: Scientific Reports. No. 25. London: WFS Publication, 1981 .
Arasaratnam, S. Ceylon. New Jersey: Prentice-Hall, 1964.
Arriaga, E. Mortality Decline and Its Demographic Effects in Latin America. Berkeley: Institute of International Studies, University of California, 1970.
Asher, H. B. Causal Modeling. Quantitative Application in the Social Science Series (Sage University Paper), No. 3. Beverly Hill: Sage Publications, 1976.
Azeez, A. M. A. "The Muslim Tradition." In Education in Ceylon: From Sixth Century B.C. to the Present Century. Vol. Ill, Chapter 95. Colombo: Ministry of Education and Cultural Affairs, 1969.
Bebarta, P. C. "Family Structure and Fertility." In Proceedings of the Fifty-first and Fifty-second Sessions of the Indian Science Congress. Part 3. New Delhi: Indian Science Congress, 1964.
Birdsall, N. "Analytical Approaches to the Relationship of Population Growth and Development." Population and Development Review, 3 (March and June 1977), 63-102.
Burch, T. K. "Decision-Making Theories in Demography: An Introduction." In Demographic Behaviour: Interdiscipli• nary Perspectives on Decision-making. Ed. Thomas K. Burch. Colorado: Westview Press, 1980. 1 29
Caldwell, J. C. "Foreward." In Social Aspects of Fertility: with Special References to Developing Countries. Ed. R.H. Chaudhury. New Delhi: Vikas Publishing House Pvt. Ltd., 1982a.
. Theory of Fertility Decline. London: Academic Press, 1982b.
., et al. "The Causes of Demographic Change in Rural South India: A Micro Approach." Population and Develop• ment Review, 8 (December 1982c), 689-727.
Chamie, J. Religion and Fertility. London: Cambridge Univer• sity Press , 1981.
Chaudhury, R. H. Social Aspects of Fertility: With Special Reference to Developing Countries. New Delhi: Vikas Publishing House Pvt. Ltd., 1982.
Chatterjee, P. K. Aspects Of Human Fertility: A Socio-Economic Study. Kalyani: University of Kalyani, 1979.
Cho, L. and R. D. Retherford. "Comparative Analysis of Recent Fertility Trends in East Asia." In International Popu• lation Conference, Liege, 1973. Vol. 2. Liege: Inter• national Union for the Scientific Study of Population, 1 974.
Clark, J. H. Population Geography. 2nd Ed. Oxford: Pergamon Press, 1965.
.Population Geography and the Developing Countries. Oxford: Pergamon Press, 1970.
Coale, A. J. "Factors Association with the Development of Low Fertility: An Historic Summary." In Proceeding of the World Population Conference, 1965. United Nations Depart• ment of Economic and Social Affairs, Vol. 2 (1965).
. "The Demographic Transition Reconsidered." In Inter- national Population Conference, Liege, 1973. Liege: International Union for the Scientific Study of Popu• lation, Vol. 3 (1973), 53-71.
. "Recent Trends in Fertility in Less Developed Countr• ies." Science, 221 (August 1983), 823-832.
. and E. M. Hoover. Population Growth and Economic Deve• lopment in Low-Income Countries: A Case Study of India's Prospects. Princeton: Princeton University Press, 1958.
., et al. Human Fertility in Russia Since the 19th Century. Princeton: Princeton University Press, 1979. 1 30
Cochrane, S. H. Fertility and Education: What do we Really Know? Baltimore: John Hopkins University Press, 1979.
Cullumbine, H. "An Analysis of the Vital Statistics of Ceylon." Ceylon Journal of Medical Science, 7 (December 1950), 91-272.
Dandekar, V. M. and K. Dandekar. Survey of Fertility and Mor• tality in Poona District. Poona: Gokhale Institute of Politics and Economics, 1953.
Davis, K and J. Blake. "Social Structure and Fertility: An Analytic Framework." Economic Development and Cultural Change, 4 (1956), 211-235.
De Silva, K. M. "Hinduism and Islam in Post-Independence Sri Lanka." The Ceylon Journal of Historical and Social Studies, 4 (January-December, 1974), 98-103.
Driver, E. D. Differential Fertility in Central India. New Jersey: Princeton University Press, 1963.
Dumont, L. Homo Hierarchicus: The Caste System and Its Impli• cations . Chicago: Chicago University Press, 1980.
Economic and Social Commission for Asia and the Pacific. Population of Sri Lanka. Bangkok: United Nations, 1976.
Ehrlich, P. R. Population Bomb. 2nd Ed. New York: Sierra Club/ Ballantine Books, 1971.
El-Hamamsy, L. S. "Belief Systems and Family Planning in Peasant Societies." In Are Our Descendants Doomed? Ed. H. Brown and E. Hutching, Jr. New York: Viking Press, 1972.
Enlyclopaedia of Britanica. I slam. Vol. 2. Chicago: Encyclo• paedia Britanica Inc., 1973.
Fernando, D. F. S. "Recent Fertility Decline in Ceylon." Population Studies, 26 (November 1972), 445-453.
. "A Note on Differential Fertility in Sri Lanka." Demography, 11 (August 1974), 441-456.
. "Fertility Trends in Sri Lanka and Future Prospects." Journal of Biosocial Science, 8 (January 1976), 35-43.
. "Nuptiality, Education, Infant Mortality and Fertility in Sri Lanka." Journal of Biosocial Science, 11 (April 1979), 133-140.
. "The Continuing Fertility Decline in Sri Lanka." Journal of Biosocial'Science, 12 (January 1980), 51-60. 131
. "Social and Demographic Determinants of Fertility in Sri Lanka." Journal of Biosocial Science, 14 (January, 1982), 59-67.
. "The High Fertility of Sri Lanka Moor Concentrations." Journal of Biosocial Sceince, 14 (October 1982), 421-429.
Fernando, T. "Introduction." In Modern Sri Lanka: A Society in Transit ion. Ed. Tissa Fernando and Robert N. Kearney. Syracuse: Maxwell School of Citizenship and Public Affairs, 1979.
Ford, T. R. and G. F. De Jong. "Religious Institutions." In Social Demography. Ed. T. R. Ford and G. F. De Jong. London: Prentice-Hall, Inc., 1970.
Freedman, R. "Norms for Family Size in Underdeveloped Areas." In Population and Society. Ed. Charles B. Nam. Boston: Houghton Mufflin, 1968.
. The Sociology of Human Fertility. New York: John Wiley and Sons., 1975.
. The Sociology of Human Fertility: An Annotated Biblio• graphy . New York: Irwington, Publishers, 1975.
. "Theories of Fertility Decline: A Reappraisal."Social Force, 58 (September 1979), 1-17.
. et al. Family Planning and Sterility and Population Growth. New York: McGraw Hill, 1959.
Freestone, C. The South-east Asian Village: A Geographic, Social and Economic Study. London: George Philip & Sons., 1974.
Goldscheider, C. Population, Modernization and Social Structure. Boston: Little Brown and Company, 1971.
Gunawardena, K. W. "The History of the Muslims in Ceylon,. Before the British Occupation." University Muslim Majlis, 9 (1959-60), 82-92.
Gray, R. H. "The Decline of Mortality in Ceylon and the Demographic Effects of Malaria Control." Population Studies, 28 (July 1974), 205-227.
Hanna, B. and T. Nadarajah. "Some Aspects of Fertility Differentials in Sri Lanka." In Population Problems in Sri Lanka. Colombo; Demographic Training and Reseasrch Unit (University of Colombo), 1967.
Hardin, G. "The Tragedy of the Commons." Science, 162 (1968), 1243-1248. 1 32
Harrington, J. A. "The Effects of High Infant and Child Mor• tality on Fertility: The West Africa Case." Concerned Demography, 3 (1971), 22-23.
Hawthorne, G. Sociology of Fertility. London: Colliner- Macmillan Ltd., 1970.
Hussainmia, B. A. "Sri Lankan Malays." The University of Muslim Majlis (Peradeniya), 1982.
International Encyclopaedia of Population. Ed. J. A. Ross. New York: The Free Press, 1982.
Israeli, R. Muslims in China: A Society in Cultural Confront• ation. Scandinavian Institute of Asian Studies Monograph Series, No. 29. Bangkok: Craftsman Press Ltd., 1978.
Jennings, W. I. Sir. "Race Religion and Economic Opportunities in the University of Celyon." University of Ceylon Review, 2 (November 1944), 1-13.
Johnston, A. Sir. "A Cufic Inscription Found in Ceylon; with a Translation by the Rev. Samuel Lee." In A Criticism of Mr. Ramathan's Ethnology of the 'Moors' of Ceylon. By, I. L. M. A. Azeez. Colombo: Moor's Islamic Cultural Home Inc. , 1957.
Jones, G. and D. Norman. "Roman Catholic Fertility and Family Planning: A Comparative Review of the Research Litera• ture." Studies in Family Planning, 34 (October 1968), 1-27.
Kendall, M. G. and C. A. 0'muircheartaigh. "Path Analysis and Model Building." World Fertility Survey:Technical Bulletins. London: WFS Publication, 1977.
Keyfitz, N. "Population Theory and Doctrine: A Historical Survey." In Readings in Population. Ed. W. Peterson. New York: The Macmillan Company, 1972.
Kirk, D. "Factors Affecting Muslim Natality." In Family Plan• ning and Population Programs: A Review of World Develop• ment . Ed. B. Berelson et al. Ann Arbor: The University of Chicago Press, 1965.
Kiser, C. "Unresolved Issues in Research on Fertility in Latin America." Milbank Memorial Fund Quarterly, 49 (1971), 379-388.
Kleinman, D. S. Human Adaptation and Population Growth: A Non- Malthusian Perspective. Allanheld: Osman & Co., 1980. 1 33
Langford, C. M. "Fertility Change in Sri Lanka Since the War: An Analysis of the Experience of Different Districts." Population Studies, 35 (February 1981), 285-305.
Luschky, D. J. and W. C. Wilax. "Demographic Transition : A Forcing Model." Demography, 2 (May 1974), 214-225.
Mauroof, M. "Aspects of Religion, Economy and Society among the Muslims of Ceylon." In Muslim Community of South Asia: Culture and Society. Ed. T. N. Madan. New Delhi: Vikas Publishing House Pvt. Ltd., 1976.
Mauldin, P. W. "Patterns of Fertility Decline in Developing Countries 1950-75." In Fertility Decline in the Less- Developed Countries. Ed. N. Eberstadt. New York: Praeger Publishers, 1981.
Mauldin, W. P. and B. Berelson. "Conditions of Fertility Decline in Developing Countries, 1965-75." Studies in Family Planning, 9 (May 1978).
Mayer, A. C. Caste and Kinship in Central India. Berkeley: University of California Press, 1966.
McGilvray, D. B. "Sexual Power and Fertility in Sri Lanka: Batticaloa Tamils and Moors." In Ethnography of Fertility and Birth.Ed. C. P. MacCormark. London: Academic Press, 1982.
McNamara, R. "Demographic Transition Theory." In International Encyclopedia of Population. Ed. J. A. Ross. New York: The Free Press, 1982.
Meegama, S. A. "Malaria Eradication and Its Effects on Morta• lity Decline: A Discussion." Population Studies, 21 (November 1967), 207-237.
Mines, M. "Labbai." In Muslim People: A World Ethnographic Survey. Ed. Richard V. Weekes. Westport: Greenland Press, 1978.
Mueller, E. "Economic Cost and Value of Children: Conceptual• ization and Measurement." In The Satisfactions and Costs of Children. Ed. J. T. Fawcett. Honolulu: East-West Centre Press, 1972.
. "Economic Motives for Family Limitation: A Study Con• ducted in Taiwan." Population Studies, 27 (November, 1972), 383-403.
Multhus, T. R. "A Summary View of the Principle of Popu• lation." In Introduction to Malthus. Ed. D. V. Glass. London: Frank Cass and Co., 1959. 1 34
. An Essay on the Principles of Population. 7th Ed. London: Reeves and Turner, 1872.
Mumdani, M. The Myth of Population Control: Family, Caste and Class in an Indian Village. New York: Monthly Review Press, 1972.
Nagata, J. "In Defence of Ethnic Boundaries: The Changing Myths and Charters of Malay Identity." In Ethnic Change. Ed. Charles F. Keyes. Seattle : University of Washington, 1 973.
Nie, N. H. SPSS : Statistical Package for the Social Sciences. 2nd Ed. New York: McGraw-Hill Book Company, 1975.
Omran, A. R. "Islam and Fertility Control." In Egypt: Popu• lation Problems and Prospects. Ed. A. R. Omran. Chapel Hill: Carolina Population Center, 1973.
Palmore, J. A. "Population Change, Conjugal Status and the Family." In Population Aspects of Social Development. Asian Population Studies Series No. II. Bangkok: ECAFE, 1 972.
Petersen, W. Population. 2nd Ed. New York: Macmillan Company, 1969.
Preston, S. H. "Introduction." In The Effects of Infant and Child Mortality on Fertility. Ed. Samuel H. Preston. New York: Academic Press, 1978.
Raghawan, M. D. India in Ceylonese History, Society and Culture. New York: Asia Publishing House, 1964.
Rahman, F. Islam and Modernity. Chicago: The University of Chicago Press, 1982.
Ramakumar, R. "Effects of Changing Mortality and Fertility on the Population of Kerala." In Studies on Mortality in India. Ed. K. E. Vaidyanathan. Gandhigram: The Gandhi- gram Institute of Rural Health and Family Planning, 1972.
Ramanathan, Hon. P. "On the Ethnology of Moors of Ceylon." Journal of Royal Asiatic Society (Ceylon Branch), 10 (1888), 234-262.
Ratcliffe, J. "Social Justice and the Demographic Transition: A Lesson From India's Kerala State." International Journal of Health Services, 8 (1978), 123-144.
Repetto, R. Economic Equality and Fertility in Developing Countries for the Future. Washington D. C: The Johns Hopkins University Press, 1979. 1 35
Rizk, H. "Social and Psychological Factors Affecting Fertility in the United Arab Republic." Marriage and Family Living, 25 (1963), 69-73.
Robinson, W. C. and S. F. Harbison. "Toward a Unified Theory of Fertility." In Demographic Behaviour: Interdiscipli• nary Perspectives on Decision-Making. Ed. T. K. Burch. Colorado: Westview Press, Inc., 1980.
Ryan, B. "Socio-Cultural Regions of Ceylon." Rural Sociology, 15 (1950), 3-18.
Samaraweera, V. "The Evaluation of a Plural Society." In Sri Lanka: A Survey. Ed. K. M. De Silva. London: C. Hurst and Company, 1977.
. "Some Sociological Aspects of the Muslim Revivalsim in Sri Lanka." Social Compass, XXV (1978), 465-475.
Sarkar, N. K. The Demography of Ceylon. Colombo: Ceylon Gov't. Press, 1957.
Sarma, R. S. S. "Effects of Declining Mortality on Fertility." In Studies on Mortality in India. Ed. K. E. Vaidyanathan. Gandhigram: The Gandhigram Institute of Rural Health & Family Planning, 1972.
Scrimshaw, S. C. M. "Infant Mortality and Behaviour in Regularity of Female Size." Population and Development Review, 4 (September 1978), 383-404.
Shrivastava, S. C. Studies in Demography. Meerut: Jai Prakash Nath & Co., 1980.
Srikantan, K. S. "Threshold Hypothesis." In International Encyclopedia of Population. Ed. J. A. Ross. New York: The Free Press, 1982.
Sri Lanka. Annual Reports of University Grants Commi• ssion, 1979-1981. Colombo: Ministry of Higher Education, 1979-1981.
. Census of Doctors. Sub-Study "C" of National Health Manpower. Colombo: Ministry of Health, 1973.
. Department of Census and Statistics. Census of Population. (Various Years, 1871-1981). Colombo: Depart• ment of Census and Statistics, 1871-1981.
. Department of Registrar General. Reports of Regist• rar General on Vital Statistics. (Various Years, 1900- 1980). Colombo: Government Press, 1900-1980. 1 36
Srinivas, M. N. The Remembered Village. Berkeley: University of California Press, 1976.
Stolnitz, G. J. "The Demographic Transition: From High to Low Birth Rates and Death Rates." In Population: The Vital Revolution. New York: Anchor Book, 1964.
Tambiah, S. J. "Ethnic Representation in Ceylon's Higher Administrative Services, 1870-1946." University of Ceylon Review, (Peradeniya), 13 (April-July 1955), 113-134.
Teitelbaum, M. S. "Relevance of Demographic Transition Theory For Developing Countries." Science, 188 (May 1975), 420-425.
Tessler, M. A. "Ethnic Change and Non-assimilation Minority Status: Jews in Tunisia and Morocco and Arabs in Israel." In Ethnic Change. Ed. C. F. Keyes. Seattle: University of Washington Press, 1981.
Tien, Y. H. "Marital Moratorium and Fertility in China." Population Studies, 24 (November 1970), 307-323.
Trewartha, G. T. A Geography of Population: World Patterns. New York: Wiley, 1969.
United Nation, Department of Economic and Social Affairs. Population Bulletin of the United Nations. No. 7. New York: United Nation, 1965.
. The Determinants and Consequences of Population Trends: A Summary of the Findings of Studies on the Relationship Between Population Changes and Economic and Social Condition. Population Studies, No. 51. New York: United Nation, 1972.
Walle, V. De E. and J. Knodel. "Demographic Transition and Fertility Decline." International Union for the Sci• entific Study of Population, Sydney. 21-25 August 1967. Sydney Conference..
. "Europe's Fertility Transition: New Evidence and Lessons for Today's Developing World." Population Bulletin, 34 (February 1983).
Waxier, N. E., et al. "Infant Mortality in Sri Lankan House• holds: A Causal Model." Social Science and Medicine, Forthcoming.
Weekes-Vagliani, W. "Case Study 3 : Sri Lanka." In Women in Development: at the Right Time for the Right Reasons. Ed. Winifred Weekes-Vagliani. Paris: OCED Publication, 1980. 1 37
Weekes, R. V. "Introduction." In Muslim People: A World Ethnographic Survey. Ed. Richard V. Weekes. Westport: Greenweood Press, 1978.
Westoff, C. and N. B. Ryder. "Recent Trends in Attitudes Toward Fertility Control and the Practice of Contra• ception in the U. S." In Fertility and Family Planning. Ed. B. Samuel et al. Ann Arbor: University Of Michigan Press, 1969.
World Fertility Survey, 1975. Sri Lanka First Report. Colombo: Department of Census and Statistics, Ministry of Plan Implementation, 1978a.
. The Sri Lanka Fertility Survey, 1975: A Summary of Findings. Voorburg: WFS Publication, 1978b.
Wright, N. H. "Recent Fertility Changes in Ceylon and Pros• pects for the National Family Planning Programme." Demography, 5 (February 1968), 745-756.
. "The Relationship of Demographic Factors and Marital Fertility to the Recent Fertility Decline." Studies in Family Planning. 59 (November 1970), 17-20.
Wright, T. P. Jr. "The Movement of Convert Harijans to Islam in South India," The Muslim World, LXXII, (September 1982), 239-244.
Wyon J. and J. B. Gordon. The Khanna Study: Population Prob• lems in the Rural Punjab. Cambridge: Harvard University Press, 1971.
Yalman, N. Under the Bo Tree: Studies in Caste, Kinship, and Marriage in the Interior of Ceylon. Berkeley: University of California Press, 1967.
Yankey, D. Fertility Differencess in a Modernizing Country: A Survey of Labanese Couples. Princeton: Princeton Uni• versity Press, 1961. 1 38
APPENDIX 1
MUSLIM SETTLEMENTS IN SRI LANKA BY PROVINCES AND DISTRICTS
1.Western Province
1.1 Colombo District
Settlement
1. Kotte 2. Slave Island 3. Kollupitiya 4. Bambalapitiya 5. Jawatta 6. Wellawatte 7. Karuwakaadu 8. Dematagoda . 9. Maiigawatte 10. Maradana 1 1 . Hulftsdorf 12. Kotahena 13. Kochchi kade 14. Mattakkuliya 15. Muthuwela 16. Hunupit iya 17. Mahara 18. Nawala 19. Nugegoda 20. Dehiwela 21 . Mount Lavinia 22. Ratmalana 23. Moratuwa 24. Padukka 25. Galagedara 26. Hanwella 27. Malwana 28. Poogoda 29. Pettah 30. Panchi kawatta 31 . Modara 32. Grandpass 33. Avi ssawella 34. Makola 35. Madampitiya 36. Wattala
1.2 Gampaha District
Settlement
1. Kadawata 2. Gampaha 3. Minuwangoda 4. Negombo 5. Mirigama 6. Attanagalla 7. Nittambuwa 8. Pasyala 9. Kala Eliya 10. Nambuluwa 11. Algama 12. Thihariya 13. Katuowita 14. Veyangoda
1.3 Kalutara District
Settlement
1. Egodauyana 2. Panadura 3. Kalutara 4. Maggona 5. Beruwala 6. Dorga-town 7. Welipanne 8. Alutgama 9. Palanda 10. Horana 11. Neboda 12. Matugama
2.Central Province
2 . 1 Matale District
Settlement
1. Mandandawela 2. Gongahawela 3. Oyapahala 4. M.C.Road 5. Walakumbura 6. Ukuwela 7. Manamboda 8. Marokona 9. Parakawela 10. Maperiya 11. Kuriwela 12. Rythalawela 1 39
APPENDIX 1 (Continued)
13. Kumbirangoda 14. Ulpothapitiya 15. Aluvihara 16. Gonamada 17. Madipola 18. Nikakola 19. Kaudupalalla 20. Mahawela 21. Rattota 22. Dambulla 23. Galewela 24. Naula 25. Matale 26. Alawatugoda
Settlement
1. Hanguranketa 2. Ramboda 3. Thalawakale 4. Nanuoya 5. Nuwara-eliya 6. Ragala 7. Udapussellawa 8. Kandapola
2.3 Kandy District
Settlement
1. Watadeniya 2. Walamboda 3. Kurukuttala 4. Bambaradeniya 5. Buweli kada 6. Daulagala 7. Hendeniya 8. Meewaladeniya 9. Pethiyagoda 10. Dehipagoda 1 1 . Elamaldeniya 12. Geli-oya 13. Kalugamuwa 14. Elpitiya 15. Kahatapi t iya 16. Illawatura 17. Andiyagodawatte 18. Ahindamale 19. Gampolagala 20. Mariyawatta 21 . Apakotuwa 22. Ulapane 23. Keerapane 24. Getambe 25. Mahaiyawa 26. Katugastota 27. Mawilmada 28. Watapuluwa 29. Wattegedara 30. Madawala 31 . Akurana 32. Bulugahatenne 33. Malwanahinne 34. Kurugoda 35. Pangullamada 36. Neeralla 37. Kurundugahaela 38. Telambugahawatte39. Udalagama 40. Murutalawa 41 . Dehianga 42. Yahalatenna 43. Malwathugoda 44. I1iukwatta 45. Amunugama 46. Menikhinna 47. Tennakumbura 48. Hapugastalawa 49. Salem-Bridge 50. Palangoda 51 . Kotmale 52. Hatton 53. Dickoya 54. Maskeliya
3. Southern Province
3.1 Galle Distrist
Settlement
1. Palapitiya 2. Panapitiya 3. Dunduwa 4. Gintota 5. Kaluwella 6. Thundalla 7. Kotte 8. Makuluwa 9. Talapitiya 10. Katugoda 11. Milinduwa 12. Hirimbure 13. Nawinna 14. Alutgama 15. Elpitiya 16. Talagaswella 17. Urugama 18. Nagoda 19. Galle 1 40
APPENDIX 1 (Continued)
3.2 Matara District
Settlement
1. Matara 2. Gandara 3. Kirinde 4. Godapitiya 5. Weligama 6. Meeulla 7. Yakkasmuni 8. Horagoda 9. Dickwella 10. Deniyaya 11. Akuressa 12. Hakmana
Settlement
1. Hakmana 2. Periatta 3.Tangalle 4. Nelugama 5. Hambantota
4. Northern Province
4.1 Mannar District
Settlement
1. Vidatal Tivu 2. Per iamadu 3. Aandankulam 4. Ueilankulam 5. Arippu 6. Pandarawelli 7. Murunkan 8. Puthuvelli 9. Vappankullam 10. Potkerney 1 1 . Ahathimur ippu 12. Tampatta- Muthali kkatu 13. Chilawathurai 14. Koolankullam 15. Marichikaddy 16. Palai kuli 17. Kondachi 18. Karadi kuli 19. Mannar town 20. Erukkalampiddy 21 . Karusal 22. Puthukudiyi rruppu 23. Tarapuram 24. Talaimannar
4.2 Vavuniya District
Settlement
1. Poonthottam 2. Pattanichivoor 3. Pulliankulam 4. Sooduwanthapulavu 5. Salanpaikulam 6. Pawatkulam 7. Neriakulam 8. Chettikulam 9. Aandiyapuli- ankulam 10. Mankulam 11. Kakaikulam
4.3 Mullaitivu District
Settlement
1. Mullaitivu 2. Mulliawalai 3. Mankulam 141
APPENDIX 1 (Continued)
4 . 4 Jaffna District
Settlement
1. Jaffna town 2. Chawakachcheri 3. Kilinochchi (Zonahar Teru)
5. Eastern Province
5.1 Trincomalee District
Settlement
1. Akkarachennai 2. Ana ichennai 3. Iqbal Street (Vannar Chennai] 4. Kullatadi Theru 5. Palai Topu 6. Allainahar (Jaya Street) 7. Zonakar Theru 8. Pulmoddai 9. Mullipothanai (96) 10. Waloru 1 1 . Sinna Kinniya 12. Periya Kinniya 13. Kuricha Keney 14. Kuddi Kachi 15. Munai Chenai 16. Kachaikodi Tivu 17. Ni lavelli 18. Tirincomalee 19. Kinniya 20. Mutur 21 . Periya Palam 22. Kathalai 23. Pankulam 24. Topoor
.2 Batticaloa District
Settlement
1 . Batticaloa town 2. Poonochimunai 3. Nochimunai (Kottaimunai) 4. Kathankuddy 5. Kangeyan Odai 6. Palamuna i 7. Meera Odai 8. Valaichenai 9. Ottamawadi 10. Kalichai 1 1 . Eravur 12. Meerakeney 13. Urugamam 14. Koppaveli
.3 Amparai District
Settlement
1 . Maruthamunai 2. Kalmunai 3. Sainthamaruthu 4. Kalaitivu 5. Nintavur 6. Oluvil 7. Palamunai 8. Attalachchena i 9. Akkaraipattu 10. Pottuvil 1 1 . Ullai 12. Paradichenai 13. Malkampitti 14. J Block East 15. Udanka 16. Samanthurai 17. Karuwattukal 18. Veerachollai 19. Chorikalmunai 20. Chavalakadai 21 . Navithanveli 22. 13th Colony 23. Natpittimunai 24. Ma jeedpuram 25. Varipathanzan 26. Isaikamam 142
APPENDIX 1 (Continued)
6. North-Western Province
6.1 Puttalam District
Settlement
1. Karampai 2. Thethapilai 3. Chett ichenai 4. Taluwa 5. Navatkadu 6. Maapoor 7. Marakali 8. Panaiadi 9. Karavankudi1 (Poolanchenai) 10. Nurai Chelai 1 1 . Alankuda 12. Andankerny 13. Ethalai 14. Pallivasal Thurai15. Palakuda 16. Talaimillu 17. Chottupittu Vadi 18. Kandal Kuda 19. Musal Pitti 20. Pallivasal Thurai21. Kurinchi Pitti 22. Kadaiyamottai 23. Nalantli swa 24. Ganamula 25. Perukku Wattan 26. Sameeragama 27. Ratpitti 28. Mathurankuli 29. Palavi 30. Puttalam 31 . Vattakandal 32. Karaitivu 33. Eluvankulam 34. Chilaw 35. Anamaduwa 36. Saravana
Settlement
1. Kurunegala town 2. Theliyakona 3. Weharapanatha 4. Malkaduwawa 5. Wehara 6. Aswedduma 7. Nagalamuna 8. Mallawapitiya 9. Thalgodapitiya 10. Thoraya 1 1 . Wallawa 12. Mawatagama 13. Talahattala 14. Telambugala 15. Tedilianga 16. Mudundua 17. Kumbalanga 18. Rambukkana 19. Banagamuwa 20. Moragolla 21 . Thalgaspitiya 22. Paragagahadeniya 23. Pannala 24. Pottuhara 25. Kurumada 26. Patmeekola 27. Alakoladeniya 28. Kadi rawalawa 29. Meeragampit iya 30. Senagedara 31 . Walpitigedara 32. Pendaligoda 33. Polgahayaya 34. Narammala 35. Plogahawela 36. Welikumburu- gedara 37. Kuliyapitiya 38. Yayawatta 39. Mummanna 40. Dambadeniya 41 . Ecapodagama 42. Pampenna 43. Kadahapola 44. Horambawa 45. Kurivikotuwa 46. Kekuwugolla 47. Siyambalagas- 48. Arikiyala kotuwa 49. Pandawa 50. Oraliatta 51 . War iyapola 52. Timbirigaskotuwa 53. Walpaluwa 54. Matto 55. Nikaweratiya 56. Namuwawa 57. Kiripokuna 58. Rambawa 59. Apakagama 60. Galgamuwa 61 . Kudawewa 62. Pulnewa 63. Atawarala 64. Walpaluwa 65. Madapokuna 1 43
APPENDIX 1 (Continued)
7 North-Central Province
7.1 Anuradhapura District
Settlement
1. Kekirawa 2. Ganewalpola 3. Madatugama 4. Horapola 5. Nidigama 6. Kalawewa 7. Negampaha 8. Palaluwewa 9. Kalagiyagama 10. Maradankadawela 1 1 . Nachchaduwa 12. Galenbind unuwewa 13. Horovpathana 14. Kahatagasdigi- 1 iya 15. Neliwagama 16. Kadukeliyawa 17. Kadiyawa 18. Endagala (Track 7) 19. Medawachchiya 20. Kadupuliyankulam 21 . Udumbugala 22. Mahakiripawa 23. Wilwewa 24. Muki r iyawa 25. Kirigollawa 26. Gambirigaswewa 27. Ni kawewa 28. Anuradhapura 29. Nochchiyagama
7.2 Polonnaruwa District
Settlement
1. Habarana 2. Hingurakgoda 3. Nugagahathamana 4. Polonnaruwa 5. Elindapathana 6. Sumkawila 7. Tampala 8. Onegama 9. Putoor 10. Diulana 1 1 . Porawewa 12. Kalela 13. Rahmath town 14. Kaduruwela 15. Muslim colony 16. Manikkampitiya
8 . Province of Uva
. 1 Badulla District
Settlement
1. Mahiyangana 2. Badulla 3. Luwugala 4. Medagama 5. Badulupitiya 6. Haliela 7. Kottegoda 8. Palugama 9. Passara 10. Ella 1 1 . Weiimada 12. Bandarawela 13. Haputale
.2 Moneragala District
Settlement
1. Malagastalawa 2. Bibile 3. Medagama 4. Modagama 5. Pakkinikawela 6. Badalkumbura 7. Aluthpotha 8. Wellawaya 9. Thamanwella 1 44
APPENDIX 1 (Coninued)
9. Province of Sabaragamuwa
9.1 Keqalle District
Settlement
1. Warakapola 2. Thummaldeniya 3. Nangalla 4. Uduhupura 5. Nelundeniya 6. Kegalle 7. Siyamalapitiya 8. Rambukkana 9. Pattampit iya 10. Kiringudeniya 1 1 . Hinguloya 12. Uyanwatta 13. Delgahagoda 14. Ganetenna 15. Nagulgama 16. Wilpola 17. Dippitiya 18. Hemmathagama 19. Madulbowa 20. Dumbuluwewa 21 . Kottegoda 22. Palliporvai 23. Aranayaka 24. Ruwanwella 25. Yatiyantota 26. Dehiowita 27. Ambepussa
.2 Ratnapura District
Settlement
1 . Malwana 2. Thiruwanaketiya 3. Hedandola 4. Ratnapura town 5. Pombagala 6. Kedangama 7. Nivitigala 8. Karavita 9. Uda Karavita 10. Hapugasdena 1 1 . Kotamulla 12. Welandara 13. Kahawatte 14. Nambuluwa Estate 15. Hawpe State 16. Nilgama 17. Opanayake 18. Palmadulla 19. Kuruwita- 20. Pusselle 21 . Hemingport Hikgashena Estate 22. Asgangala- 23. Kurulmulla 24. Thalawitiya- Wilegoda Paduwata 25. Moragalla 26. Rakwana 27. Balangoda 28. Eheliyagoda town 29. Embilipitiya 30. Madampe 1 45
APPENDIX 2
QUESTIONNAIRE: FERTILITY SURVEY--MUSLIMS OF SRI LANKA
1. LOCATION AND IDENTIFICATION OF HOUSEHOLD
1.1 District:
1.2 D.R.O. Division:
1.3 G.S. Division:
1.4 Name of Village:
1.5 Name of Household:
1.6 Address:
1.8 Household Number:
1.9 Name of Investigator:
1.10 Date of Investigation:
2. GENERAL CHARACTERISTICS OF HOUSHOLD MEMBERS
No. Relation- Sex Age Marital Literacy Level ship to M/F Date of Birth Status Yes/No of Ed- Household Age Yr Month (Over 12 uca- Yrs) tion (Grade)
2.1 2.2 2.3 2.4 2.5 2.6 2.7 1 46
APPENDIX 2 (Continued)
2.GENERAL CHARACTERISTICS OF HOUSEHOLD MEMBERS (Continued)
Known Type of For Employed Only Employment Other Activity Status Languages (for per• sons 5 yrs and over)
Pr inc ipal Type of Place Occupation Economic of or kind of Activity Work work
2.8 2.9 2.10 2.11 2.12 2.13 1 47
APPENDIX 2 (Continued)
3.INCOME
No. Income Wage, Profit Income Income Household Salar ies f rom f rom from and others and related business rent agricu• (relation• receipts or farm dividends, lture ship to (including interest, household) self empl• etc . oyment )
3.1 3.2 3.3 3.4 3.5 3.6
Person Other Income Total Income and like other Food,goods Goods Others income and other and services services from provided employer from household
3.7 3.8 3.9 3.10 3.11 1 48
APPENDIX 2 (Continued)
4.HOUSING CONDITION FACILITIES
4.1 Particulars of the house:
4.2 Tenure of accommodation:
4.3 Number of rooms (excluding bathroom, toilets and garage, including kitchen):
4.4 Amenities-- Source of Water
Source of Pipe Well Tank River or Water Water Stream
Private Common Private Common
Drinking
Cooking
Bathing and washing
4.5 Toilet facilities:
4.6 Main Source of Energy
Electricity Gas Kerosene Fire- Other wood
Cooking
Lighting 1 49
APPENDIX 2 (Continued)
5.HEALTH
5.1 How many members of household, if any suffered from any illness during the past two weeks:
5.2 Whether treatment was taken: Yes / No
Nature of Illness
5.3 If yes, indicate source of treatment:
5.4 Is there any handicap in your family: (spec i fy)
6.GENERAL
6.1 Does any member of your household own any of the following: 1 50
APPENDIX 2 (Continued)
6.2 Does any member of household actively participate in any society: (Specify)
6.3 Does any member of household subscribe regularly to a newspaper, magazine, journal or periodical: (Specify)
7.FERTILITY (FOR MARRIED WOMEN):
7.1 Age at marriage:
7.2 Present age:
7.3 Was there a live birth during the last twelve months :
7.4 Is she pregnant now: If yes, when will be the delivery:
7.5 Is this her first marriage: If no, give the details of other marriages:
7.6 Pregnancy Details:
Pregnan• Outcome Date Mother's If Live- Is Child If cy Order of Preg• Age at Born , Name Alive(A) Dead nancy Yr. Mth.Time of and Sex or Dead When, Delivery M / F (D) Age & A / D Cause of Death
2 151
APPENDIX 2 (Continued)
3
4
5
6
7
8
9
7.7 Contraceptive knowledge and use:
8. MIGRATION
8.1 Are you living here from your birth: If no, give details:
8.2 In the last 5 years, did any former member of your household leave your home to live elsewhere on a fairly permanent basis?
If yes, give details: 1 52
APPENDIX 2 (Continued)
8.3 For the last one year, has any one of your household left your house or temporarily lived out of the village for more than one month, give details: 1 53
APPENDIX 3
3.1 SELECTED SOCIO-ECONOMIC CHARACTERISTICS OF 11 SURVEYED MUSLIM SETTLEMENTS
Settle- Socio-Economic Economic Characteristics ment Status(%)
Lower Middle Upper % of Houses % of House % of with no hold using House- Latrine Electricity hold for using Pipe Water for Drinking
Light- Cook• ing ing
Topoor 37 50 1 3 33 1 3 0 0 Kathan- kuddy 19 70 1 1 48 71 20 1 1 Nintavur 33 67 0 40 60 0 0 Erukkal- ampiddy 52 44 04 80 0 0 1 00 Puttalam 34 63 03 1 0 67 0 1 00 Muthuwela 42 48 10 0 68 25 1 00 Weligama 31 35 34 0 59 0 41 Kekunu- golla 72 28 0 75 0 0 06 Hemmatha- gama 36 64 0 27 22 3 07 Pangoll- amada 39 57 04 0 29 0 51 Wanaha- puwa 70 30 0 27 0 0 10 1 54
APPENDIX 3 (Continued)
3.2 SELECTED OCCUPATIONAL AND EDUCATIONAL CHARACTERISTICS OF 11 SURVEYED MUSLIM SETTLEMENTS
Settle- Occupational Characteristics Educational ment Characteristics
Husband's Occupation Wife's Percentage Workforce of Partici- Literacy pation (%)
Primary Profess- Other Male Female ional
Topoor 70 27 03 1 0 73 52 KathanKuddy 1 9 09 72 07 87 71 Nintavur 57 1 0 33 0 60 48 Erukkalan- piddy 1 3 30 57 1 0 96 92 Puttalam 10 30 60 0 90 86 Muthuwela 0 30 70 1 7 92 76 Weligama 0 0 1 00 0 90 51 Kekunugolla 57 18 25 20 62 51 Hemmathagama 0 19 81 03 95 74 Pangollamada 30 03 67 0 87 48 Wanahapuwa 71 06 23 70 67 54 1 55
APPENDIX 3 (Continued)
3.3 SELECTED DEMOGRAPHIC CHARACTERISTICS OF 11 SURVEYED MUSLIM SETTLEMENTS
Settle- Total Percent of Average Mean Average Infant ments Average Population Female No. of Birth Deaths of Below Age Age at Live ' Interval per House- 20 years Marriage Birth (in Live hold (in per years) Birth years) Female House• hold
Topoor 6.4 55 1 7 5.4 2.3 .045 Kathankuddy 5.6 55 1 5 3.7 1 .8 .054 Nintavur 5.6 46 18 4.4 2.0 .076 Erukkalam- piddy 6.9 61 19 5.9 2.2 .068 Puttalam 5.2 48 20 3.5 1 .9 .001 Muthuwela 6.0 38 20 6.0 1 .9 .046 Weiigama 9.0 35 25 6.6 2.1 .051 Kekunugolla 6.1 54 1 7 5. 1 2.4 .05 Hemmatha- gama 6.4 47 1 9 4.8 2.2 .001 Pangolla- mada 6.1 44 18 5.2 2.4 .034 Wanahapuwa 6.8 53 20 5.3 2.1 .013 1 56
APPENDIX 4
SPECIFIC FORMS IN WHICH VARIABLES ARE CODED AND ANALYSED IN THE PATH ANALYSIS
Var iables Forms of Measurement
Children Ever Born Number of live births of female household head
Child Mortality Percentage of live births of female household head that ended in death before 5 years of age
Household's Socio- Economic Status 1 -- Lower (socio-economic class ) 2 -- Lower Middle (socio-ecomonic class) 3 -- Upper Middle (socio-ecomonic class) 4 — Upper (socio-economic class)
Wife's Age at Marr iage Actual age at marriage
Wi fe's Level of School i.ng 1 -- Grade 5 and below 2 -- Grade 6 to 9 3 — Grade 10 and above
Wife's Present Age Actual age at the time of survey