Purchase of this document is not possible at this time because the store is temporarily disabled. Continue searching

Three essays on the fertility preferences of rural Ghanaian women: A longitudinal perspective

ProQuest Dissertations and Theses, 2009
Dissertation
Author: Ivy A Kodzi
Abstract:
Attitudinal data are expected to be useful for the design and implementation of population policies and family planning programs. The reason for including questions on fertility attitudes and expectations in fertility surveys is the belief that such attitudes affect future fertility behavior. Despite the recognition that fertility preference data may provide useful insights for understanding future fertility behavior, most studies in sub-Saharan Africa rely on cross-sectional surveys which do not allow for examining the dynamic aspects of fertility preferences. As a result, research on the formation and stability of fertility preferences has received little attention. There is also little empirical work along the dominant line of research which investigates the predictive dimensions of fertility preferences. This dissertation focused on understanding the stability of individual stated preferences over time, the determinants of changes in preferences and the prediction of fertility outcomes by relating changes in stated preferences to subsequent pregnancy outcomes. This dissertation extends the body of literature on fertility preferences generally by its focus on the dynamic aspects of fertility preferences, and particularly for sub-Saharan Africa. In a 5-year prospective longitudinal study of a sample of Ghanaian women, I find that approximately 20 percent of the women would change their fertility preferences from one interview to the next. Women who had attained or exceeded their ideal family size showed considerable stability in the preference to stop childbearing over multiple interviews. The desire to stop childbearing was mainly driven by their perceptions of their husbands fertility desires and the attainment of the normative family size of four children. Prior stated preferences were also strongly predictive of subsequent pregnancy occurrence.

TABLE OF CONTENTS

LIST OF TABLES

................................ ................................ ................................ .......

vi

ACKNOWLEDGEMENTS

................................ ................................ .........................

viii

CHAPTER ONE: INTRODUCTION

................................ ................................ ..........

1

References

9

CHAPTER TWO: GENERAL DESCRIPTION OF DATA

................................ .......

12

2.1

Introduction

12

2.2

Community Characteristics

12

2.2.1

Timing of Int erviews

14

2.2.3

Retention Rates and Sample Characteristics

15

2.3

Social and Economic Context of Study

15

2.3.1

Economic Resources

15

2 .3.2

Educational Attainment

17

2.3.3

Employment Status

18

2.4

General Reproductive Context

18

2.4.1

Marriage

18

2.4.2

Childbearing Exper ience

20

2.5

Fertility Preferences

23

2.6

References

25

CHAPTER THREE: THE TIME DYNAMICS OF INDIVIDUAL FERTILITY PREFERENCES AMONG RURAL GHANAIAN WOMEN

.............................

42

Abstract

42

3.1

Introduction

43

3.2

Fertility Preference Dynamics

46

3.3

Methods

51

3.3.1 Study setting and sampl e description

51

3.3.2 Variables

52

3.3.3 Analytical Strategy

54

3.3.4 Explaining Fertility Preferences dynamics through Latent Class Analysis

56

3.3.5 Model estimation and selection

59

3.4

Results

62

3.5

Summary and conclusion

68

3.6 References

70

v

CHAPT ER FOUR: DETERMINANTS OF THE DESIRE TO STOP CHILDBEARING AMONG WOMEN IN SOUTHERN GHANA: PARITY PROGRESSION, PARTNER EFFECTS, AND SITUATIONAL FACTORS

....

83

4.1 Introduction

84

4.2

Conceptual Framework

86

4.2.1

Previous Reproductive Experience

88

4.2.2

Partner Preferences

89

4.2.3

Material Conditions

92

4.2.4

Normative Expectations and

Social Pressure

94

4.3

Hypotheses

96

4.4

Methods

97

4.4.1 Empirical Model

99

4.4.2 Data Description

101

4 .4.3 Variables

102

4.5

Results

105

4.6

Summary and Conclusions

109

4.7 References

11 3

CHAPTER FIVE: THE USE OF FERTILITY PREFERENCES TO PREDICT SUBSEQUENT FERTILITY: EVIDENCE FROM GHANA

.............................

124

5.1 Introduction

125

5.2 Literature Review

127

5.3 Data and Methods

134

5.3.1 The Model

135

5.4 Results

138

5.5 Summary and conclusion

141

5.6 References

144

CHAPTER SIX: SUMMARY AND CONCLUSIONS

................................ ..............

151

6.1 Introduction

151

6.2 Key Findings

152

6.3 Limitations and Directions for Further Research

154

6.4 Policy implications

156

vi

LIST OF TABLES

Table 2.1: Percent distribution of selected characteristics of communities

......................

26

Table 2.2: Percent distribution of women by selected background characteristic and round of survey ................................ ................................ ................................ ............................

27

Table 2.3: Percent of households possessing items or facilities by round of survey

........

28

Table 2.4: Percentage of households having amenities and durable assets by wealth quartiles in round 3

................................ ................................ ................................ ...........

29

Table 2.5: Percent distribution of women by highest level of schooling attained and English literacy, according to age and round 3 wealth quartiles

................................ ......

30

Table 2.6: Percent distribution of working st atus according to selected background characteristics

................................ ................................ ................................ ....................

31

Table 2.7: Percent distribution of women by marital status and round of survey

............

32

Table 2.8: Percent distribution of woman - rounds according

to marital status and background characteristics

................................ ................................ ................................

33

Table 2.9: Percent distribution of woman - round by number of living children, mean number of living children and mean number of children ever born (CEB)

......................

34

Table 2.10: Percent distribution of duration of birth intervals (months) of children ever born according to selected background characteristics

................................ .....................

35

Table 2.11: Median age at first birth according to selected backgro und characteristics at round 7 and percent distribution of women aged 20 - 49

................................ ...................

36

Table 2.12: Number of births by all women during survey period by round of survey and age

................................ ................................ ................................ ................................ .....

37

Table 2.13: Age - specific

(per 1000 women) and total fertility rates (TFR) of all women 2000 - 2003 and GDHS 2003 ................................ ................................ ..............................

38

Table 2.14: Percent distribution of women by ideal number of children, according to the number of living children ................................ ................................ ................................ ..

39

Table 2.15: Mean ideal number of children by selected background characteristics

.......

40

Table 2.16: Percent distribution of woman - rounds by desire for children, according to the number of living children ................................ ................................ ................................ ..

41

Table 3.1: Percent distribution of fertility preferences by round of survey

......................

75

Table 3.2: Transition matrix for consecutive rounds and percent distribution of woman - rounds by transition type between consecuti ve rounds of the survey

...............................

76

Table 3.3: Subsequent fertility preferences of women who wanted no more children in a previous round by attainment of ideal family size in the previous round and the occurrence of a birth before the next round

................................ ................................ ......

77

Table 3.4: Subsequent fertility preferences of women who wanted more children in a previous round by attainment of ideal family size in the previous round, the occurrence of pregnancy or birth before the next rou nd and desired timing of pregnancy

.....................

78

Table 3.5: Sequential comparison of goodness - of fit statistics of latent class models without missing values

................................ ................................ ................................ ......

79

Table 3.6: Sequential comparison of goodness - o f fit statistics of latent class models with missing values

................................ ................................ ................................ ...................

80

Table 3.7: Conditional probability of fertility preference given cluster membership

......

81

Table 3.8: First and last rounds background ch aracteristics by cluster membership

........

82

vii

Table 4.1: Means and standard deviations of explanatory variables for the overall sample, sample of women wanting no additional children and sample used for estimation

.......

119

Table 4.2: Odds - ratios for fixed effects logit regression of the determinants of the desire to stop childbearing among Ghanaian women, 1998 - 2003 –

main effects models

........

120

Table 4.3: Odds - ratios for fixed effects

logit regression of the determinants of the desire to stop childbearing among Ghanaian women, 1998 - 2003 -

reverse variable inclusion

121

Table 4.4: Odds - ratios for fixed effects logit regression of the determinants of the des ire to stop childbearing among Ghanaian women, 1998 - 2003 -

model with interactions

...

122

Table 5.1: Descriptive statistics of variables included in the regression model

.............

147

Table 5.2: Random Effects

Logistic Regression results showing monthly odds ratios of pregnancy for married women in 1998

................................ ................................ ...........

148

Table 5.3: Fixed Effects Logistic Regression results showing monthly odds ratios of pregnancy for married women in 1998

................................ ................................ ...........

150

viii

ACKNOWLEDGEMENTS

I wish to thank Professor David Johnson for his inspiring hope, expert guidance, positive criticism and readiness to help in getting this dissertation completed. I would also like to express my deepest appreciation to Professor John Casterline for the use of the dataset, and support during the entire program. Professor Gordon De Jong was always prepared with practical advice on problems –

I remember how kind you have been. My sincere gratitude goes to my

other committee members -

Professors David Shapiro and Michael Massoglia -

for their invaluable input to my work.

Professor John McCarthy deserves special mention for his unique sense of duty, and his genuine concern about my progress. I am grateful to Professor Duane Alwin for serving on my committee. I am also delighted to acknowledge Professor Francis Dodoo who was always ready to help. I am very grateful to Dr. Peter Aglobitse for his role in gathering the data and for allowing me to use the dataset.

A special thank you goes to Dr. Laurie Scheuble for her support and advice.

Many thanks go to my friends, colleagues and extended family who constantly prayed, and encouraged –

Charles and Elizabeth Nderitu, Aunty Pearl and Uncle Fred, Grandma and Grand pa Cape Coast. Finally, I would like to thank my closest friends and family. I deeply appreciate your unfailing confidence in me. To my dear father who did not live to see the end of this work, my mother, and siblings –

sincere thanks to you. To Nene, Papa a and Sika -

my best friend, son and daughter, I cannot thank you more for your sacrifice.

1

CHAPTER ONE

INTRODUCTION

The concept of “fertility preferences” theoretically captures the extent to which human agency or intentional behavior affect s the reproduction process. Fertility preference measures can broadly be defined as measures that seek to capture some dimension of an individual’s attitude or motivation to influence fertility outcomes. Fertility preference data are routinely collected in

demographic studies and are used for various purposes. Typical uses for such data include estimating completed fertility for couples, and extending cohort fertility in aggregate - level forecasts. In developing countries, fertility preference measures are u sually used to estimate the demand for fertility control, and the level of unmet need for contraception (Casterline et al. 2003; Westoff 2001).

Despite the usefulness of the concept of fertility preferences, understanding how fertility attitudes impact f ertility behavior through onetime survey interviews is inherently

challenging .

Part of the challenge stems from the fact that stated preferences are naturally associated with uncertainty and potentially subject to revisions in the short - term (Manski 1990;

Monnier 1989). For example, Manski (1990) argues that measures of “intention” to have a child should be regarded as a subjective probability that the event would occur, rather than taken with certainty. The degree of uncertainty associated with preference s depends on how well respondents are able to factor in exogenous (sometimes unanticipated) factors which occur over time and which impact both preferences and future behavior. Factors such as fecundity impairments, partner disagreements,

2

contraceptive fai lure, economic conditions, health conditions, and marital instability may change stated preferences.

Fertility preferences change over time not only because of factors related to this intrinsic uncertainty, but also due to certain developmental processes

associated with the reproductive life course. Over the reproductive life course, the preference for more children is expected to evolve based on either age - related factors associated with decreased fecundity or on the attainment of individual or culturall y - defined reproductive milestones. For example, if any fertility - limiting calculus is being employed, then when we measure an individual’s fertility preferences over their entire reproductive life span, changes in preferences should generally move from wan ting a child to wanting no more children, with “unsure” being a middle ground where genuine uncertainty can be captured, all else equal. If parity - dependent fertility limitation is employed, it would be expected that people would remain in the state of wan ting a child, in keeping with their desired family size until they have attained that number of children or are physiologically constrained (Lee 1980). Those who want no more children at any point in time are also likely to remain in that state if preferen ces are strongly - held. When a sample is examined over some time, however, what one observes is a mixture of change processes. Some changes in observed responses could be as a result of the kind of life course progression explained

above; or a reflection of

the effects of short - term constraining factors; or due perhaps to individuals’ psychological propensity to change. Some aspects of change would be developmental, others would be temporary and/or circumstantial, or pertain to

3

individual psychological dispo sitions to change and lastly, observed change could result from measurement error.

Recent fertility studies on childbearing intentions (notably Schoen et al. 1999) have emphasized that the degree of individual preference certainty and stability play an important role in determining the likelihood of future implementation. These finding s

confirm what Westoff and Ryder (1977) and Morgan (1982) called attention to decades ago. Westoff and Ryder (1977) pointed out two important issues to be considered when e xamining preference - behavior correspondence, namely, the tendency to change one’s mind between surveys and the probability of not fulfilling stated preferences. When preferences are strongly - held, they are less susceptible to changing life situations and t end to remain relatively stable over time. The likelihood of eventually fulfilling stated preferences is affected by the occurrence of expected and unexpected factors and the passage of time. To understand the relationship between stated preferences and fu ture childbearing, it is important to use a framework that adequately delineates the impact of not only situational factors, but also reproductive and psychological dispositions on both preferences and behavior. Therefore, the theoretical framework for thi s dissertation consists of a conceptual plan which draws from theories of planned behavior and of fertility decision - making. Specifically, the natural versus parity - specific (Henry 1961) and the sequential theories of fertility decision - making (Namboodiri

1983) provide the basis for analyses regarding the determinants of preferences, and for explaining how fertility preference may change over time. Furthermore, the dominant theory of reasoned action and planned behavior (Azjen and Fishbien 1980) posits tha t “intentions” mediate

4

between attitudes and social norms, and actual behavior. Fertility preference, which technically may not be regarded as a measure of intentions (intentions is a measure which is considered more goal - oriented), is not given a central

causal role in the theory of planned behavior. However, recent modifications of this theory within the social psychological literature provide a way to link preferences and fertility behavior. Perugini and Bagozzi (2001) proposed a modification to the the ory of reasoned action and planned behavior, which made desires (preferences) an intermediate variable antecedent to intentions in the path model predicting planned behavior.

Empirically, the nature of the dynamics of fertility preferences over time, the

relative importance of intervening events or situations and reproductive life cycle factors in explaining preference formation and stability, and the correspondence between stated preferences and future behavior are all different dimensions that deserve s pecial attention in the study of fertility attitudes. Over the past several decades, the literature has focused almost exclusively on the debate on the predictive validity of reproductive preferences. While some studies have examined the predictive validit y of statements about intentions or preference for additional children on future childbearing outcomes (Schoen et al. 1999; Williams et al. 1999; Thomson 1997; Westoff 1990; Bankole 1995; De Silva 1991; Foreit and Suh 1980), others have focused on the rela tionship between desired family size and actual fertility (Hagewen and Morgan 2005; Chang et al. 1981). In sub - Saharan Africa, however, there is very little systematic research on the dynamic aspects of fertility preferences –

even the widely - researched as pect of preference - behavior consistency has

5

received little attention in the region due to the scarcity of longitudinal data. This dissertation seeks to fill this gap.

This dissertation is a collection of three essays which focus on the dynamics of ferti lity preferences among rural women in Ghana. Specifically, I investigate whether women’s fertility preferences are stable over time and the extent to which preferences change or

are revis ed

as women’s reproductive life circumstances change . I also investig ate the relative impact of women’s reproductive history, spousal preferences and other situational factors in preference formation, specifically the preference to stop childbearing. Finally, I examine the degree to which stated preferences

correspond with actual future behavior, in terms of subsequent pregnancy. I examine the above questions through analy zing an eight - wave prospective longitudinal dataset collected in six communities in southern Ghana

from 1998 to 2004. At each round

of the survey, women o f reproductive age were asked the fertility

preference /desire question

that is considered to be most predictive of future behavior and typically fielded in Demographic and Health Surveys (Hermalin et al. 1979, Westoff and Ryder 1977; Ghana Statistical Serv ice and Macro International Inc. 1999; Ghana Statistical Service, Noguchi Memorial Institute for Medical Research, and ORC Macro. 2004) : “W ould you like to have a(another) child with your husband/partner, or would you prefer not to have any more children w ith him?”

Birth, marriage, contraceptive histories, and other household and attitudinal data were also collected. The advantage of using this dataset is that it provides several rounds of data and has a rich of set of variables to directly test the influen ce of factors impacting fertility preferences

and behavior at the individual level

over the period. There are just a

6

few previous empirical studies that have examined fertility preferences longitudinally in sub - Saharan Africa , (notably Bankole 1995; Bankol e and Westoff 1998 )

but none of these contain as many as eight rounds of data over a period as long as five years. As such, this research will provide

a distinctive contribution to research on fertility preferences

in sub - Saharan Africa .

The next chapter provides a brief overview of the dataset used to set the general context of the analyses. The chapter primarily describes the sample, and the reproductive, economic and social characteristics of the women being studied in this dissertation. Apart from the fertility preference questions, there were also a number of attitudinal questions regarding the benefits and costs of another child, the cost and benefits of children generally, recent and future household economic prospects, recent and future personal hea lth perceptions, and contraception, among others. The respondents were also asked whether their sexual partners wanted a child or another child and when their partners wanted the next child. Round - by - round marriage, contraceptive use and birth histories we re also recorded in the study. Demographic background and household data were also collected. These data allow for modeling changes in situational factors that affect both preference and behavior by the inclusion of time - varying variables.

The analyses i n this study begin in chapter 3 with a description of the nature of change or stability in fertility preferences over the multiple occasions of measurement. I

describe typical patterns of change (or stability) in individual fertility preferences over the f ive - year period

using Latent Class Analysis. In principle, a woman can have all her children

7

within five years; however given that our sample contains women at different reproductive life stages, over a 5 - year period we cannot observe the whole development al evolution of preferences in a single woman -

we only observe a snapshot of her reproductive span. It can be reasonably assumed that in West Africa, where completed fertility is high and birth intervals are exceptionally long, the active reproductive lif e - span is also relatively long (Bledsoe 1998), in which case observing women over a 5 - year period may mostly capture short - run changes. Given this assumption, it can be expected that fertility preferences would fairly concur over repeated measures or would

change in tandem with other reproductive variables to

establish reasonable stability. Thus, the main focus of the essay is to investigate whether change patterns were meaningfully consistent with women’s reproductive life circumstances.

In Chapter 4, th e question regarding how reproductive life cycle events and situational factors affect changes in fertility preferences is addressed. Using the fixed - effects logit regression technique, I

model the impact of reproductive life cycle events and outcomes, spo usal interactions, health experiences, perceptions of past and future household economic conditions, and interactions with friends, relatives and family planning workers, on the woman’s preference to stop childbearing. Here again, the eight waves of data provide a strong advantage in capturing the variation within a woman’s fertility preferences over time.

Thus far, only the time dynamics and the determinants of fertility preferences have been considered. Chapter 5 tackles the question regarding the ext ent to which prior stated

8

preferences predict subsequent fertility. This study is directed at understanding the predictive validity of fertility preferences. Prospective studies on the predictive validity of fertility preferences or intentions have confirm ed a modest association between prior preferences and future fertility (examples are Schoen et al. 1999; Williams et al. 1999; Thomson 1997; Razzaque

2000 ; Chang and Tey 1994; Freedman and Hermalin 1975; Nair and Chow 1980; De Silva1991; Foreit and Suh 198 0; Bankole 1995). These studies typically used two rounds of data, and assumed that fertility attitudes were fixed over the study period. With the multi - wave panel data, I examine whether changes

in prior fertility preferences

affect fertility, controlling

for changes in demographic life cycle and socioeconomic factors. The final chapter presents an overview of the results of the studies in the dissertation, including policy implications and areas for further research.

9

References

Ajzen, Icek and Mart in Fishbein. 1980. Understanding

Attitudes and Predicting Social B ehavior . Englewood Cliffs, NJ: Prentice - Hall.

Bankole, Akinrinola and Charles Westoff. 1998. The C onsistency and V alidity of R eproductive A ttitudes: E vidence from Morocco. Journal of Biosoc ial Science . 30:439 - 455.

Bankole, Akinrinola. 1995. “Desired fertility and fertility behaviour among the Yoruba of Nigeria: A study of couple preferences and subsequent fertility,” Population Studies

49(2): 317 - 328.

Bledsoe

Caroline; Fatoumatta Banja;

Allan Hill. 1998. “ Reproductive Mishaps and Western Contraception: An African Challenge to Fertility Theory. ”

Population and Development Review

24 (1):15 - 57.

Casterline, J ohn

B., El - Zanaty Fatma, El - Zeini Laila 2003. “ Unmet need and unintended fertility:

longitudinal evidence from Upper Egypt. ”

International Family Planning Perspectives

29(4):158 - 166.

Chang, Poo Tan and Nai Peng Tey. 1994. “ Do fertility Intentions Predict Subsequent Behavior? Evidence from Peninsular Malaysia. ”

Studies in Family Planning

25(4): 222 - 231.

Curtis Sian L. and Charles Westoff. 1996. “ Intention to use contraceptives and subsequent contraceptive behavior in Morocco. ”

Studies in Family Planning

27(5):239 - 250.

De Silva, W. Indralal. 1991. “Consistency between reproductive pref erences and behavior: The Sri Lankan experience,” Studies in Family Planning

22(3): 188 - 197.

Foreit, K aren G. and M.H. Suh. 1980. “The effect of reproductive intentions on subsequent fertility among low - parity Korean women, 1971 - 76.” Studies in Family

Planning

11(3): 91 - 104.

Freedman, Ronald, Albert I. Hermalin, and M.C. Chang. 1975. “Do statements about desired family size predict fertility? The case of Taiwan, 1967 - 1970.” Demography

12(3): 407 - 416.

Ghana Statistical Service (GSS), Noguchi Memori al Institute for Medical Research (NMIMR), and ORC Macro. 2004. Ghana Demographic and Health Survey 2003. Calverton, Maryland: GSS, NMIMR and ORC Macro.

10

Ghana Statistical Service and Macro International Inc. (MI). 1999. Ghana Demographic and Health Survey

1998. Calverton, Maryland: GSS and MI.

Hagewen, Kellie J. and S. Philip Morgan. 2005. “Intended and ideal family size in the United States.” Population and Development Review

31(3): 507 - 527.

Henry, Louis. 1961. “ Some data on natural fertility. ”

Euge nics Quarterly

8(2):81 - 91.

Hermalin, A., R. Freedman, Te - Hsiung Sun, and Ming - Cheng Chang, 1979. “ Do Intentions Predict Feritilty? The Experience in Taiwan 1967 - 74. ”

Studies in Family Planning

10(3): 75 - 95.

Lee, R onald D. 1980. “ Aiming at a Moving Target : Period fertility and Changing Reproductive Goals. ”

Population Studies

43:205 - 226.

Manski, Charles. 1990. “ The Use of Intention Data to Predict Behavior: A best - case Analysis. ”

Journal of the American Statistical Association . 85 (412): 934 - 940.

Monnier,

A. (1989). “Fertility Intentions and Actual Behaviour. A Longitudinal Study: 1974, 1976, 1979.” Population: An English Selection

44: 237 - 259.

Morgan, Philip S.1982. “ Parity - specific Fertility Intentions and Uncertainty: The United States, 1970 - 1976. ”

Dem ography

19(3): 315 - 334.

Nair, N. K., L. P. Chow. 1980. “ Fertility Intentions and Behavior: Some Findings from Taiwan. ”

Studies in Family Planning

11(7/8): 255 - 263.

Namboodri, K. 1983. Sequential fertility decision making and the life course, in Balatao R .A. and Lee R. (eds.), Determinants of fertility in developing countries, vol. 2, New York: Academic Press.

Perugini, Marco and Richard P. Bagozzi. 2001. “The role of desires and anticipated emotions in goal - directed behaviour: broadening and deepening t he theory of planned behavior. ”

British Journal of Social Psychology

40: 79 - 98.

Razzaque, Abdur. 2000. “Preference for children and subsequent birth: Evidence from Matlab, Bangladesh . ” Genus

LV1(3 - 4): 209 - 221.

Schoen, Robert, Nan Marie Astone, Young J . Kim, Constance A. Nathanson and Jason M. Fields. 1999. “Do fertility intentions affect fertility behavior?” Journal of Marriage and the Family

61(3): 790 - 799.

Thomson, Elizabeth. 1997. “Couple childbearing desires, intentions, and births. ” Demograph y

34(3): 343 - 354.

11

Westoff ,

C harles

F. 2001. Unmet need at the end of the century, DHS Comparative Reports, Calverton, MD, USA: ORC Macro. No. 1.

Westoff, C harles

F. and N. B. Ryder (1977). “The Predictive Validity of Reproductive Intentions.” Demography

14: 431 - 453.

Westoff, Charles F. 1990. “Reproductive intentions and fertility rates.” Family Planning Perspectives

16(3): 84 - 96.

Williams, Lindy, Joyce Abma and Linda J. Piccinino. 1999. “ The correspondence Between intention to avoid childbearing and

subsequent fertility: a prospective analysis. ”

Family Planning Perspectives . 31(5):220 - 227.

12

CHAPTER TWO

GENERAL DESCRIPTION OF DATA

2.1

Introduction

While the essays in this dissertation focus on the dynamics of fertility preferences, it is helpful to fir st describe the general context of the data. This chapter provides an overview of the sample, and the reproductive, economic and social characteristics of the women being studied in this dissertation. In contrast to subsequent chapters where descriptive st atistics are related to the particular question on hand, the descriptive findings presented here are general in scope.

2.2

Community Characteristics

The data for this dissertation

come from a

prospective

longitudinal study conducted in six communities in sou thern Ghana between 1998 and 200 4 . These towns and villages are located in the Central, Western and Greater Accra R egions of Ghana. The communities were purposively selected to reflect some diversity by way of ethnicity, religion, ecology and socioeconomic

condition (Agyeman and Casterline 2003). Eight rounds of reproductive and household surveys were conducted among women who were between 15 - 50 years at the onset of the study. The partners of those who were married or in stable sexual unions were also inte rviewed. The analyses reported here focus only on the sample of women -

a total of 1219 in the first round. Two hundred and nine women were added

13

to the sample in the second round. Some of these women were originally scheduled to be interviewed in round on e; others were added to refresh the sample . They were asked all applicable round one questions retrospectively as well as the second round questions. There were 1205 women by the last round of the survey. For confidentiality reasons, we designate the name s of the communities by A, B, C, D ,

E and F.

Apart

from com m u nity B, most of the communities surveyed are predominantly farming communities (see Table 2.1). Members of communities D and E are mainly engaged in subsistence farming, while those in C and F

grow cash crops. Community B, a market town, is closest to the national capital and is the most urbanized. In terms of relative wealth distribution, women surveyed in community B were mostly (87 percent) in the top twenty percent of the entire sample. The

Full document contains 168 pages
Abstract: Attitudinal data are expected to be useful for the design and implementation of population policies and family planning programs. The reason for including questions on fertility attitudes and expectations in fertility surveys is the belief that such attitudes affect future fertility behavior. Despite the recognition that fertility preference data may provide useful insights for understanding future fertility behavior, most studies in sub-Saharan Africa rely on cross-sectional surveys which do not allow for examining the dynamic aspects of fertility preferences. As a result, research on the formation and stability of fertility preferences has received little attention. There is also little empirical work along the dominant line of research which investigates the predictive dimensions of fertility preferences. This dissertation focused on understanding the stability of individual stated preferences over time, the determinants of changes in preferences and the prediction of fertility outcomes by relating changes in stated preferences to subsequent pregnancy outcomes. This dissertation extends the body of literature on fertility preferences generally by its focus on the dynamic aspects of fertility preferences, and particularly for sub-Saharan Africa. In a 5-year prospective longitudinal study of a sample of Ghanaian women, I find that approximately 20 percent of the women would change their fertility preferences from one interview to the next. Women who had attained or exceeded their ideal family size showed considerable stability in the preference to stop childbearing over multiple interviews. The desire to stop childbearing was mainly driven by their perceptions of their husbands fertility desires and the attainment of the normative family size of four children. Prior stated preferences were also strongly predictive of subsequent pregnancy occurrence.