Ngalinda, Innocent: Age at First Birth, Fertility, and Contraception in Tanzania

81

Chapter 4. Age at First Birth

4.1. Introduction

This chapter will examine levels and the proximate determinants of age at first birth as well as socio-economic variables or independent variables that operate through proximate determinants to influence age at first birth in Tanzania. Lastly the extent to which age at first birth influences fertility in Tanzania will be examined.

The onset of child bearing is an important demographic indicator as seen in the introductory chapter. However, postponement of first births as a result of a higher age at marriage would play an important role in the overall fertility decline of any country provided that age at first marriage is synonymous with age of entry into sexual relations. Women’s age at first birth vary considerably among developing countries and regions of the world. For example, among the 41 countries that participated in the World Fertility Survey (WFS), women’s average age at first birth ranged from about 17 years in Bangladesh to more than 25 years in Korea and Sri Lanka (Pebley, 1981). The study further found that age at first birth has remained relatively stable for almost 30 years according to cohort analysis in most of the countries that participated in the survey, although there are exceptions.

One of these exceptions is South Korea where the mean age at first birth was found to have risen from 20.7 to 27.1 in 30 years. The key factor for this rise was associated with an increase in the age at marriage, which rose from 17.8 in 1940 to 23.3 in 1970. The increasing age at marriage was due to an expansion of university education for men and women as well as a rapid economic growth.

In a recent comparative study of 37 countries, which participated in DHS from 1986 to 1992, Westoff et al. (1994) found that in sub-Saharan Africa the median age at first birth for women aged 25 to 49 ranged from 18.1 years in Niger to 21.5 years in Rwanda. In the majority of countries, it is below age 20.


82

Child bearing at an early age entails substantial health risks for both the mother and the child. Complications of pregnancy at an early age include first and third trimester bleeding, severe anemia, preeclampsia and toxemia, prolonged or obstructed labour, cephalopelvic disproportion, prematurity, low birth weight, stillbirth, high prenatal and infant mortality, and high maternal mortality (Casterline and Trussell, 1980; Cherlin and Riley, 1986; Lowe, 1977). Furthermore, early motherhood tends to impede the pursuit of other life options that might compete with child bearing (Gyepi-Garbrah, 1985c). Due to the fact that adolescent child bearing hinders a mother’s educational attainment, it often results in reducing economic opportunity for the mother and the household as a whole (Rao and Balakrishnan, 1988).

4.2. Age at first birth in Tanzania

The 1996 TDHS<13> data indicate that Tanzanian women have a low age at first birth. More than half of the interviewed population of women (50.8 percent) become mothers before they reach age 20, with a high proportion of them reporting their first birth between ages 15 and 17 ( Table 9 ). In a comparative study of DHS data, Westoff et al. (1994) found a non-uniformity of the proportion of women having birth by age 20 among sub-Saharan countries. For example the study found that between age 20 to 24 the percentage ranges from 25 in Rwanda to 75 percent in Niger. In the 1996 TDHS the percentage was 52.3 among women age 20-24 which can be considered consistent with other sub-Saharan countries.


83

Table 9: Percentage distribution of women by age at first birth according to current age in Tanzania

Current age

Never

Ever

N

Had given birth

<15

15-17

18-19

20-21

22-24

25+

N

15-19

79.2

20.8

1,732

5.3

59.6

35.6

360

20-24

25.6

74.4

1,676

4.3

29.7

36.3

23.4

6.3

1,247

25-29

8.2

91.8

1,440

4.6

26.3

32.8

20.7

11.9

3.8

1,322

30-34

3.8

96.2

1,117

7.1

28.4

26.0

17.7

14.4

6.5

1,075

35-39

3.7

96.3

888

8.5

32.8

23.5

16.6

11.1

7.4

855

40-44

1.8

98.2

679

10.5

34.0

25.4

13.7

9.3

7.2

667

45-49

1.9

98.1

585

11.8

28.0

22.8

12.7

13.5

11.3

574

15-49

24.8

75.2

8,117

6.9

31.3

29.4

17.4

10.2

4.8

6,104

Source: calculated from 1996 TDHS

Table 9 shows that almost 7 percent of the women reported having had their first birth before they were 15 years of age, with a few cases reporting ages as low as 10 years ( Figure 11 ). This finding has serious health implications since young mothers suffer more from health problems associated with pregnancy and childbirth than older women as explained above. This situation might also render such mothers infertile in the future or worse still lead to death. Fortunately, the percentage of women who became mothers before age 15 has declined in the past 30 years by more than 50 percent from 11.8 (45-49) to 4.6 (25-29) percent. But surprising enough, the number of women giving birth over age 24 has decreased dramatically in the same period from 11.3 percent to 3.8 percent. The same trend has been observed in the 1991/92 TDHS (see Table 10 ).

Table 10: Percentage distribution of women, who ever gave birth, by age at first birth according to current age, 1991/92 TDHS

Current age

<15

15-17

18-19

20-21

22-24

25+

N

15-19

3.9

59.8

36.3

507

20-24

5.3

32.1

37.7

19.4

5.5

1,424

25-29

6.0

27.9

26.4

20.5

13.4

5.8

1,497

30-34

9.6

37.1

21.8

15.2

10.9

5.4

1,115

35-39

8.4

37.6

22.4

13.9

10.3

7.4

980

40-44

10.1

35.4

21.7

14.0

11.2

7.6

695

45-49

10.2

32.7

19.9

15.1

10.8

11.2

667

15-49

7.5

35.2

27.1

15.8

9.5

5.1

6,885

Source: calculated from 1996 TDHS


84

Figure 11: Frequency distribution of age at first birth in single years

Source: calculated from 1996 TDHS

According to figure 4.1, which shows the frequency distribution of age at first birth in single years, it is evident that the majority of women in Tanzania start child bearing at an early age. Most of these first births are concentrated between ages 16 and 19 with only a few first births (3.6 percent) occurring after age 25. The figure reveals that child bearing starts as early as age 10 and as late as age 39. However, only few cases are found at these two extremes. These measures are crude, and possibly lead to false estimates due to the fact that age at first birth is the duration between the mother’s birth and the birth of her first offspring, therefore the process might be incomplete at the time of the survey. This factor and others as explained in Chapter 3, make the use of life table a necessity in this study. Age group 15-19 will not be discussed here as Chapter 7 is about adolescents in particular due to the fact that adolescents have different characteristics and consequences in comparison to adults.

Table 11 shows the adjusted mean age at first birth to be 18.4 years. It is possible that the mean could have been lower if it were not for the fact that a few women reported


85

having had their first birth at ages as late as 39 years. The table further shows that there has been an increasing trend in mean age at first birth among Tanzanian women. For example, women currently aged 25-29 have a mean age at first birth of 18.4 years whereas women aged 40-44 have a mean age of only 17.4 years. With modernisation, more women now go to school and stay in the education system longer than in the past. More years in school means postponing first birth. Few of the older women attended school and those who did, never stayed for many years in the education system. Marriage used to be at an early age, and child bearing started almost immediately afterwards, as many of them were not engaged in any wage employment.

Table 11: Mean age at first birth by current age

Age group

T10

T25

T50

T75

Trimean

Spread

N

15-19

14.8

15.8

16.9

17.9

16.8

1.4

360

20-24

15.3

16.8

17.8

19.8

18.1

2.2

1,247

25-29

15.4

17.0

18.0

20.5

18.4

2.8

1,322

30-34

14.7

16.5

17.8

20.8

18.3

3.5

1,075

35-39

14.4

16.2

17.4

20.6

17.9

3.7

855

40-44

13.9

15.8

17.0

20.0

17.4

3.8

667

45-49

14.0

16.0

17.4

21.2

18.0

4.4

574

15-49

15.0

16.5

18.3

20.3

18.4

3.3

6,104

Source: calculated from 1996 TDHS

The trimean for women in age group 45-49 is higher than expected, namely at 18.0 years. It is possible that women in this group reported the age of their first birth inaccurately. This might have caused the inflation in their mean age at first birth. In general, by the age of 15 years ( Table 11 ), 10 percent of women in Tanzania have had their first live birth. About a quarter of those, who ever gave birth to a child, have done so by the time they were 16.5 years old. At 18 years of age, only 10 percent have never experienced a live birth. The spread of 3.3 means that most women in Tanzania have their age at first births around the mean that is about three and a quarter year below or above 18.3 years. According to this spread, most Tanzanian women have their first birth between 15 and 21.6 years of age. The spread of 3.3 indicates a high fertility for a country with broad bottom age structure. In general, there seem to be marginal differences in the mean age at first birth between older women, who had their first child many years ago and younger women who had their first birth recently.


86

Let us now examine the proportions of ever giving birth according to current age so as to have a better understanding of age at first birth in Tanzania. Table 12 below shows the proportions of women, who ever gave birth.

Table 12: Proportions of women, who ever gave birth, by current age

Current age

Never

Ever

N

15-19

0.79

0.21

1,732

20-24

0.26

0.74

1,676

25-29

0.08

0.92

1,440

30-34

0.04

0.96

1,117

35-39

0.04

0.96

888

40-44

0.02

0.98

679

45-49

0.02

0.98

585

15-49

0.25

0.75

8,117

Source: calculated from 1996 TDHS

According to Table 12 , 75 percent of the women had given birth to at least one child at the time of the survey. The remaining 25 percent reported that they never have given birth. The proportion ever giving birth increases with age across the cohorts from 20 percent in age group 15-19 to 96 percent in age group 30-34 until it reaches the maximum of 98 percent in the age group 45-49.

Although women in age group 45-49 have had the longest exposure to the risk of childbirth, the proportion ever giving birth is not 1.0 as expected. This might be a result of the prevalence of infertility and sterility owing to poorer nutrition and a lack of good medical care during their prime reproductive ages. For example, a woman could have primary sterility due to a venereal disease but because of ignorance and inadequate medical care, she might become barren. With modernisation and women’s education increases, this might not occur to the same extent anymore. Also owing to better nutrition, hygiene and medical care, sub-fecund women nowadays have chances of giving birth.


87

Table 13: Logistic regression odds ratio predicting the relative risk that a woman had a live birth before age 15

Variable

Odds ratio

LiteracyLiterateSemiliterateIlliterate

.31***

.76*

RC

ReligionMoslemsCatholicsProtestantsOthers

RC

.76**

.73**

.72*

Place of ResidenceUrbanRural

1.17

RC

Age15-1920+

1.47**

RC

***

p<.001 (highly significant)

**

p<.01 (significant)

*

p<.05 (less significant)

p>.05 (insignificant)

Source: calculated from 1996 TDHS

Table 13 shows results of a logistic regression model estimating the effect of background variables on the likelihood that a woman had a live birth before attaining age 15. It seems that education is negatively related to child bearing at early ages. Literate women seem to be 3 times less likely to give birth before age 15 than illiterate women. Moslems are more likely to give birth before age 15 than other religious denominations. Other religious believers who are not Moslems are almost 1.3 less likely to have given birth before attaining age 15 than Moslem women, and adolescents are 1.5 times more likely to give birth before age 15 than adult women were at that age.


88

Table 14: Mean age at first birth by current age in sub-Saharan Africa and selected developing countries

Country

Year

T10

T25

T50

T75

Trimean

Spread

N

Benin

1996

15.3

16.8

18.7

20.8

18.8

3.3

4,201

Central African Republic

1994

14.1

16.0

17.9

20.4

18.1

3.8

4,737

Comores

1996

14.7

16.5

19.1

22.1

19.2

4.4

1,695

Côte d’Ivoire

1994

14.1

15.7

17.6

19.9

17.6

3.5

6,147

Ghana

1993

15.5

17.1

19.0

21.4

19.1

3.5

3,505

Kenya

1993

14.8

16.4

18.2

20.3

18.3

3.3

5,437

Mali

1995

14.7

16.0

17.7

20.0

17.9

3.4

7,980

Tanzania

1996

15.0

16.5

18.3

20.3

18.4

3.3

33,702

Uganda

1995

14.3

15.8

17.6

19.7

17.7

3.2

5,594

Zambia

1996

14.9

16.2

17.8

19.6

17.9

3.0

5,939

Zimbabwe

1994

15.2

16.8

18.7

20.7

18.7

3.3

4,330

Egypt

1995

15.4

17.3

19.8

22.9

20.0

4.2

13,390

Bangladesh

1996

14.0

15.1

16.7

18.8

16.8

2.9

8,543

Brazil

1996

16.2

17.9

20.4

23.5

20.6

4.4

8,405

Dominican Republic

1996

15.2

17.0

19.3

22.4

19.5

4.3

5,768

Source: calculated from DHS III

The mean age at first birth in sub-Saharan Africa ranges from 17.7 years for Uganda to 19.2 in Comores. Egypt, Brazil and the Dominican Republic which have a higher age at first birth (20, 20.6 and 19.5 years) than the selected countries in sub-Saharan Africa which conducted DHS III so far. The spread of more than 4 years can be associated with higher mean age at first birth. Countries ( Table 15 ) that have a spread of more than 4 years have higher mean age at first birth. In sub-Saharan Africa, 10 percent of women got their first child between age 14 and 15 while for Brazil the same percentage is reached by the age of 16.2. Hence in all selected countries, women start their child bearing during adolescence but the main difference is the beginning of child bearing. Tanzania is not an exception. The early age at birth found to prevail among sub-Saharan countries is consistent with the findings of previous studies which have reported unusually early age at child births, which is direct associated with high levels of adolescent fertility prevailing in sub-Saharan Africa (Gyepi-Garbrah, 1985c).


89

Table 15: Trends in age at first child bearing in sub-Saharan Africa and selected developing countries

Ages

Be-

nin

CAR*

Côte d’

Ivoire

Ghana

Kenya

Mali

Ugan-da

Zam-bia

Zim-babwe

Tanza-

nia

Egypt

Bang-ladesh

Brazil

DR**

15-19

17.0

16.2

15.9

17.1

16.7

16.2

16.3

16.6

16.8

16.9

16.9

15.7

16.2

16.2

20-24

18.5

17.6

17.1

18.3

18.1

17.3

17.7

17.8

18.4

17.8

18.7

16.9

18.6

18.1

25-29

18.9

18.3

17.8

19.3

18.6

17.9

18.1

18.2

18.9

18.0

20.1

17.1

19.8

19.3

30-34

18.8

18.5

18.1

19.3

18.4

17.9

17.8

18.0

18.7

17.8

20.3

17.0

20.9

20.0

35-39

19.3

18.8

18.3

19.8

18.3

18.5

17.8

17.7

18.9

17.4

20.4

16.7

21.3

19.7

40-44

18.9

18.0

18.5

19.3

18.4

18.4

17.9

17.9

19.2

17.0

20.7

16.4

21.4

19.4

45-49

18.8

18.7

18.6

19.6

18.7

18.5

17.8

17.7

19.0

17.4

19.6

16.7

21.9

19.7

* Central African Republic; ** Dominican Republic

Source: calculated from DHS III

In Tanzania, Kenya, Ghana, Côte d’Ivoire and the Central African Republic the mean age at first birth for women 40-44 is lower than the mean for those 45-49. This was also observed in the DHS Comparative Studies (Westoff, et al 1994). This does not necessarily indicate a trend towards lower age at first birth, but more probably indicates a bias caused by recall error. Older women appear to displace the exact timing of this event to older ages. For this reason as well as the possibility that other types of reporting errors occur, caution is warranted in the interpretation of differences across age cohorts.

Thus in examining trends over time, it is preferable to use the age group 40-44 as the starting point of the trend line. The mean age at first birth for women 40-44 ranges from 17.0 years in Tanzania to 19.3 years in Ghana. Likewise for age group 25-29 it ranges from 17.8 in Côte d’Ivoire to 19.3 years in Ghana. Trends over the time in the age at first birth can be seen by examining the differences in the mean age at birth between the cohort of women 40-44 and the cohort 25-29. The choice of these age groups is in line with earlier findings that most of child bearing in sub-Saharan countries is within adolescence.

Table 15 suggests that age at first birth has been generally at the same level for most sub-Saharan counties for the past 15 years (by examining age groups 40-44 to 25-29). Tanzania, Zambia, Uganda and Kenya show slightly increasing trends in age at first birth. Ghana and Benin show that there has been no increase or decrease in mean age at first birth. Other countries show a decline in the age at first birth in 15 years prior to the surveys in the sub-Saharan region. However, it is likely that in most cases the decline is due to reporting errors.


90

Figure 12: Trend in age at first birth in Tanzania 1991/92-1996

Source: calculated from 1991/92 and 1996 TDHSs

Figure 12 suggests a slight shift from an early age at first birth towards a slightly later age at first birth in Tanzania. This supports the earlier findings. All in all, we do not expect any meaningful trend by using the 1991/92 TDHS and 1996 TDHS, due to the fact that the period between these two is just four years.

4.3. Proximate Determinants of Age at First Birth

The major fertility variables that are of great importance in starting the patterns of family formation are the proportion remaining childless and the age at first birth for those who bear children. Since marriage is believed to be virtually universal in Tanzania and voluntary childlessness is unknown, the main proximate determinants of proportion remaining childless is the incidence of primary sterility. On the other hand, the proximate determinant of age at first birth includes:

4.3.1. Age at menarche

One of the necessary conditions for conception to take place is that the woman must have already started ovulating. Since it is difficult for a woman to know if her body has started to ovulate, except by the use of very complicated clinical tests, the occurrence can be approximated by menarche (the onset of menstrual periods). Menstruation, being a very noticeable process, lends itself as a useful survey tool for the estimation of the start of ovulation even though sometimes it precedes ovulation by several months.

Several studies have established that first marriage occurs earlier among women with early menarche (Buck and Stavraky, 1967; Kiernan, 1977; Ryder and Westoff, 1971). Other studies by Presser (1978), Zelnik (1981), and Udry (1979) have shown that women with early menarche have early first intercourse and earlier first births. The mechanisms for linking events in timing and sequence are not clearly understood. Various mechanisms have been proposed. Biological mechanisms include increased release of sex hormones at puberty leading to increased libido, and consequently too early intercourse. Those women with early puberty are more fecund than women with later puberty. This fact leads to earlier births for a given exposure to the risk of pregnancy for those women with early age at menarche.

The social processes in conjunction with pubertal hormones that lead to early development of secondary sexual characteristics, which are attractive to males, provide early opportunities for intercourse. Parents and peers provide encouragement for early intercourse or marriage to the woman with early menarche, as parents want to see their daughters married. This is due to the fact that parents would not like their daughters to have out of wedlock pregnancies, and also they would like to get their bride price as early as possible. Moreover parents are happy when their daughters are married as it signifies their maturity.


92

Mpiti and Kalule-Sabiti (1985a) in analysing data from a Lesotho Fertility Survey found that 10 percent of the interviewed women started having menstrual periods by the age of 13 years and that the mean age at menarche was 15 years. In analysing the distributions of the onset of menstruation for cohorts of women, they found that younger cohorts experienced the onset at earlier age. Gyepi-Garbrah (1985c) found a drop of age at puberty to approximately 12-14 in most areas of sub-Saharan Africa by using Fertility and Survey data. For this fact, we can conclude that the age at menarche is declining in this region of the world although a rapid decline might not be typical of subgroups with less than adequate nutrition (Muroki, 1988).

Age at menarche, though an important proximate determinant of age at first birth as it signifies the beginning of puberty, will not be analysed in this study, as the TDHS data does not provide any information on it. Further research must be on this proximate determinant of age at first birth.

4.3.2. Age at First Sexual Intercourse<14>

Even with the onset of puberty, child bearing cannot take place unless one engages in sexual activities that eventually lead to pregnancy. Information on the timing of first marriage, first intercourse, and first birth is crucial for the onset of the child bearing years. The conventional marker of the beginning of exposure to the risk of pregnancy is the date of first union. However, in some societies, sexual activity is not confined to marriage and women may bear children before the recognition of the date of first union. In such settings, the age at first sexual intercourse and date of first birth may be more appropriate indicators of the beginning of sexual exposure than the date of first union. The same is true in settings where marriages are arranged in early childhood and months or even years may elapse before the marriage is consummated. Age at first sexual


93

relationship is therefore an important indicator of the onset of child bearing. Traditionally in many African societies, having a sexual partner as a woman is associated with marriage and it is expected that sexual intercourse must strictly occur within married life (Mpiti and Kalule-Sabiti, 1985b).

Nevertheless, much literature suggests that age at first marriage is rising in many African societies while the age at menarche is believed to be declining (Rogo, 1986), and that premarital adolescent sexual activity is increasing (Cherlin and Riley, 1986). Therefore, if this change in premarital sexual behaviour is not compensated by an increase in premarital contraceptive use, then it is expected that there will be an increase in the proportion of adolescent births out of wedlock<15>, as well as an increase in the prevalence of induced abortion (Gyepi-Garbrah, 1985b).

The sexual behaviour of unmarried adolescents can be hypothesised to be a result of the breakdown of traditional social controls by elders over the sexual behaviour of adolescents (Adeokun, 1990; Ocholla-Ayayo et al., 1990). This change can be attributed at least partly to the fact that educated youths now obtain knowledge from books that can be used to challenge the wisdom of the older generation. Furthermore, an individual decision-making has become more important because individuals are no longer accountable for their behaviour directly to the elders or to the community but rather to the judges in the courts (Bauni, 1990). This hypothesis of a breakdown of traditional controls over sexuality is supported by a survey done in Kenya in which over 60 percent of the respondents reported that they believed that the rules and norms restricting premarital and extramarital sex no longer apply today (Ocholla-Ayayo et al., 1990).


94

Figure 13: Age at first intercourse by single years

Source: calculated from 1996 TDHS

An alternative theory is that unmarried teenagers use sexual relations and pregnancy to accomplish certain goals. For example, sexual relations may have economic benefits or might be a step towards marriage (Meekers, 1992). One of the problems with the existing literature on sexual behaviour is that the data often do not come from random samples but rather from selected groups, such as school children of a certain area.

However most researches have documented the consequences of adolescent and premarital pregnancies in sub-Saharan Africa (Njogu, 1991). The problems noted most often include dropping out of school (Mwateba et al., 1988), illegal abortion (Mashalaba, 1989), child abandonment (Dynowski-Smith, 1989), and high mortality among newly-born (Meekers, 1990). All in all, data from DHS provide a unique


95

opportunity to update the body of knowledge on sexual and reproductive behaviour in sub-Saharan Africa (Gage-Brandon and Meekers, 1993).

Table 16: Percentage distribution of respondents’ age at first sexual intercourse by current age

Current age

Had experience sexual intercourse
Never ever

N

At union <15 15-17 18-19 20-21 22-24 25+

N

15-19

51.7

48.3

1,724

15.1

23.2

54.2

7.7

707

20-24

7.4

92.6

1,666

20.8

13.1

38.3

19.3

7.2

1.1

1,220

25-29

1.5

98.5

1,436

22.9

15.7

37.4

14.9

5.7

2.3

1.0

1,090

30-34

0.6

99.4

1,115

24.4

14.8

37.0

14.3

6.0

1.9

1.5

837

35-39

0.2

99.8

886

31.9

15.9

34.6

11.5

3.4

1.7

0.9

602

40-44

0.0

100

678

43.2

10.3

32.7

8.1

4.0

0.6

1.0

385

45-49

0.2

99.8

587

45.1

12.8

28.0

8.5

3.9

1.2

0.5

321

15-49

13.0

87

8,092

26.7

15.2

37.9

13.3

4.8

1.4

0.7

5,162

Source: calculated from 1996 TDHS

From Table 16 it is evident that a large proportion of women in Tanzania in every age group have the first sexual experience between ages 15 and 17. After adolescence, very few women in Tanzania are still virgin as less than half of adolescents are still virgin; only 7.4 percent have not had sexual intercourse by the age of 20. It is evident that the traditional practice of experiencing the first sexual intercourse on the day of union is dying out as the percentage of women who experienced sexual intercourse at the first time in union, dramatically decreases from older to younger cohorts.

By the age of 19, more than 80 percent of women already had their first sexual experiences. By the age of 24, only 13.6 percent of women in Tanzania are expected to be inexperienced in sexual intercourse. This is an indication that women in Tanzania engage in sexual activities at an early age. This finding can be supported by the result of the 1985 Tanzania national survey on adolescent fertility, conducted among urban youth aged 12 to 24 years which showed that more than 50 percent of the females surveyed were sexually active (Mwateba et al., 1988). The number of women, who had not experienced any sexual intercourse yet, decreases with age so that almost all women at 45 years had experienced sexual intercourse at least once in their lifetime. With the low level of contraceptive practice (12.5 percent) prevailing in the country now, this would imply early child bearing and consequently a high level of fertility.


96

Table 17: Mean age of respondents at first sexual intercourse by current age

Age group

T10

T25

T50

T75

T90

Trimean

Spread

N

15-19

12.8

14.2

15.4

16.5

17.5

15.3

1.8

707

20-24

13.4

14.9

16.4

18.1

19.6

16.4

2.5

1,220

25-29

13.2

14.7

16.1

18.0

19.8

16.2

2.8

1,090

30-34

13.3

14.7

16.1

18.0

19.9

16.2

2.9

837

35-39

13.2

14.4

15.7

17.5

19.3

15.8

2.9

602

40-44

13.5

14.6

15.7

17.5

19.3

15.9

2.8

385

45-49

13.2

14.4

15.6

17.5

19.5

15.8

2.9

321

15-49

13.2

14.6

15.9

17.7

19.4

16.0

2.7

5,162

Source: calculated from 1996 TDHS

As can be observed in Table 17 age at first sexual experience in Tanzania is low across the cohorts. The trimean, which is a more sensitive measure than the mean, shows that the average ages at first sexual relationship among Tanzanian women, is 16 years. By age 13.2, 10 percent of sexually experience women have experienced sexual intercourse and by age 19 only 10 percent of women are sexually inexperienced. . This low age at first sexual experience is an indication of prevalence of pre-marital sexual activities among Tanzanian women.

The slightly increasing trend of age at first intercourse in recent years is evident as the mean age at first sexual intercourse for older women (45-49) is 15.8 years while for younger ones (20-24) it is 16.4 years. The reason for this shift was the effect of attending schools. As pregnant schoolgirls are not allowed to continue with education, they are afraid of sexual activities and the consequences that might force them to leave school. The low level of mean age at sexual intercourse for the older women is due to the fact that the older cohorts had a low age at first marriage and so sexual relationship could start as soon as they got married. On the other hand, the younger cohorts marry comparatively late and premarital sexual activity is predominant among them (premarital births will be examined later).


97

Table 18: Odds ratio predicting that a woman had sexual intercourse before age 15

Variable

Odds ratio

LiteracyLiterateSemiliterateIlliterate

.35***

.67***

RC

ReligionMoslemsCatholicsProtestantsOthers

RC

.66***

.63***

.97

Place of ResidenceUrbanRural

.74***

RC

Age15-1920+

1.93***

RC

*** p<.001 (highly significant)

Source: calculated from 1996 TDHS

Table 18 shows the effects of literacy, religion, place of residence and age on the likelihood that a sexually experienced woman had first sexual intercourse before she reached age 15. It appears that illiterate women have a greater chance to start sexual activity before turning 15. Literate women in Tanzania are almost three times less likely to start sexual intercourse before reaching age 15 compared to illiterates. On the other hand, semiliterate women are 1.5 less likely to start sexual intercourse before turning 15 than illiterate women. Therefore, literacy can delay the first sexual intercourse of a Tanzanian woman; the relationship is significant. The main reason for literate women not to engage in sexual activities might be the fear of becoming pregnant.

It seems that Moslems are more likely to start sexual activities before the age of 15 compared to believers of other religious denominations. Protestants are 1.6 times less likely to start sexual activity before attaining age 15 as well as Catholics (1.5 times) and others (1.03 times) than Moslems; the relationship is statistically significant. Women residing in urban areas are 1.4 less likely to start sexual activities before attaining age 15 than their sisters in rural areas. In general, an illiterate adolescent Moslem woman residing in a rural area is most likely to experience her sexual intercourse before reaching the age of 15 compared to others.


98

Table 19: Mean age at first sexual intercourse for sub-Saharan Africa including selected developing countries

Country

Year

T10

T25

T50

T75

T90

Tri-mean

Spread

N

Benin

1996

13.6

14.7

16.1

17.7

19.1

16.2

2.3

2,970

CAR

1994

12.9

14.1

15.4

16.8

18.4

15.4

2.2

4,954

Comores

1996

13.5

14.9

16.7

19.1

21.8

16.9

3.5

908

Côte d’Ivoire

1994

12.7

13.9

15.1

16.4

17.8

15.1

2.0

5,853

Ghana

1993

13.9

15.0

16.3

17.8

19.3

16.4

2.3

3,149

Kenya

1993

12.9

14.4

15.9

17.8

19.6

16.0

2.6

5,017

Mali

1995

13.5

14.4

15.4

16.8

18.5

15.5

2.0

3,941

Tanzania

1996

13.2

14.6

15.9

17.7

19.4

16.0

2.7

5,162

Uganda

1995

12.7

14.0

15.3

16.9

18.5

15.4

2.3

4,958

Zambia

1996

12.9

14.2

15.5

17.1

19.0

15.6

2.5

4,996

Zimbabwe

1994

14.3

15.7

17.5

19.3

21.0

17.5

2.7

4,108

Egypt*

1995

-

-

-

-

-

-

-

-

Bangladesh*

1996

-

-

-

-

-

-

-

-

Brazil

1996

14.1

15.6

17.6

20.2

23.3

17.8

3.8

7,455

DR

1996

13.4

15.0

17.0

19.7

23.1

17.2

3.9

4,506

* The question was not asked.

Source: calculated from DHS III

It is seems that in all countries sexual intercourse begins before attaining age 15 as 10 percent of women had sexual intercourse experience between age 12.7 in Uganda and age 14.3 in Zimbabwe. There might be a direct relationship between the timing of first sexual intercourse and first birth as those countries which have lower age at first sexual intercourse ( Table 19 ) posses lower age at first birth as well ( Table 14 ). Mean age at first sexual intercourse ranges between 15.1 in Côte d’Ivoire to 17.5 in Zimbabwe. The information we get from the spread might assist us to draw a conclusion of the intensity of sexual activities within a country. In countries with a spread of - say 2 years - adolescents engage more frequently in sexual activities than those having a spread of more than 3 years. Therefore, a spread of 3.5 for the Comores may be attributed to Islamic attitudes towards restrictions on premarital sexual activities. Comores is an Islamic republic.

4.3.3. Age at First Marriage

Although marriage has a definite meaning in most African cultures, it is difficult to define in general terms due to a fact that marriage is based on different concepts for different societies. In contrast to the wedding (marriage ceremony) as a discrete event in Western societies, African customary marriage is a complex institution that generally


99

proceeds by stages, most of them characterised by the performance of prescribed rites. Marriage is a process composed of several stages between the preliminary rites and the full acceptance of the couple as a social unit (Van de Walle, 1968; United Nations, 1988). In this context it is difficult to determine when a union started.

In Tanzania the type of marriage institution varies among the various ethnic groups and even within the same ethnic group there are tendencies to emphasise, elaborate or eliminate certain beliefs, practices or ceremonies. The traditional marriage always precedes either the civil or church/mosque weddings. It is usually the final stage in the marriage and cohabitation process, even though childbirth may precede the wedding ceremony. Wanyambo<16> for instance, have two types of traditional marriage. In the first type, a girl is taken to her future husband by force. Later this is followed by wedding ceremonies. However, this couple can go to the church/mosque at the later stage to formalise their marriage; that day then is the day of marriage a woman might refer to. Any births before formalising that marriage might be misquoted as premarital births although in a real sense they are not.

The second type of traditional marriage is a formal one, where parents make all necessary marriage arrangement including payment of bride price. At the end a girl is sent to the future husband. This type of marriages can also be taken to church for formalisation before a girl is sent to her future husband or at a later stage when a marriage certificate is handed out. Therefore, in this study marriage is taken in its broadest sense to embrace all forms of cohabitation. We will combine ‘living together’ with ‘currently married’ to form the married category as explained in chapter 3.

Tanzanians marry to have children, and marriage only has a meaning when a child is born and survives. Children, especially sons, are cherished as the means of cementing a marriage and perpetuating the family. Indeed, it is viewed as unusual if a child fails to come within the first year of marriage. Childless women or men are pitied or looked down upon in the society. Because of strong parental authority and traditions cum prescriptions about purity and premarital chastity, traditional Tanzanian marriages take place early and are arranged by the parents. The characteristics of agrarian economy, family organisation and socialisation structure as well as the levels of female education


100

also sustain early and universal marriage. In many societies age at first marriage, is practically synonymous with age of entry into sexual relations, where female chastity is culturally cherished. Thus it is a proximate determinant of the start of child bearing.

Figure 14: Age at first marriage in single years

Source: calculated from 1996 TDHS

Figure 14 shows that age at first marriage in Tanzania is generally low, with some women marrying as young as 6 years of age. This is common among coastal people where a man (husband to be), normally older than a wife to be, marries a young girl. Marriage is not consummated immediately, as this young girl does not start sexual intercourse with her husband immediately. She stays under the guardianship of her mother-in-law or the older wife until her first menarche. After her first menarche she is allowed to start sexual intercourse with her husband. Such incidences might deflate the age at first marriage in the TDHS data. This situation also makes it impossible to use age at first marriage as a proxy for exposure to the risk of pregnancy.


101

Table 20: Percentage distribution of women age at first marriage by current age

Married

Ages

Never

Ever

N

<15

15-17

18-19

20-21

22-24

25 +

N

15-19

74.6

25.4

1,732

15.4

66.1

18.5

442

20-24

24.5

75.5

1,676

9.7

40.8

29.4

14.8

5.3

1,268

25-29

7.4

92.6

1,440

9.0

34.2

24.9

15.8

11.1

5.0

1,308

30-34

4.5

95.5

1,118

13.4

33.4

19.7

14.0

10.2

9.3

1,085

35-39

1.7

98.3

888

18.4

36.1

17.1

11.8

8.0

8.5

879

40-44

1.5

98.5

680

19.1

38.5

18.1

10.8

5.2

8.3

662

45-49

0.8

99.2

585

22.6

30.6

17.5

12.1

8.1

9.2

577

15-49

23.3

76.7

8,120

14.0

38.1

22.0

12.8

7.6

5.6

6,221

Source: calculated from 1996 TDHS

Table 20 shows that the proportion of women who marry before age 15 has been declining across the cohorts (the percentage of those ever-married women currently aged 45-49 is 22.6 compared to 8.9 percent for women in the youngest cohort 25-29). As already mentioned, adolescence will be dealt with later in Chapter 7. However, it is expected to have fewer cases of married women who are less than 20 years of age as most of them are still in schools. All in all this is a clear indication of a rising age at first marriage in Tanzania.

Table 21: Mean age at first marriage by current age

Age groups

T10

T25

T50

T75

Trimean

Spread

N

15-19

14.0

15.0

16.2

17.3

16.2

3.8

442

20-24

14.3

15.9

17.5

19.2

17.5

2.5

1,268

25-29

14.3

16.1

18.0

20.3

18.1

3.3

1,308

30-34

13.7

15.6

17.8

20.7

18.0

4.1

1,085

35-39

13.2

15.0

17.0

20.0

17.3

4.5

879

40-44

12.9

14.9

16.9

19.5

17.0

4.7

662

45-49

13.0

14.7

17.1

20.2

17.3

5.2

577

15-49

13.7

15.3

17.3

19.6

17.4

3.8

6,221

Source: calculated from 1996 TDHS

Table 20 also depicts two general characteristics of Tanzania nuptiality, namely early, and non-universal marriage. About 25 percent of Tanzanian women aged 15-19 were married at the time of the survey. The ever-married figures increase to 76 percent and 93 percent in the 20-24 and 25-29 age groups, respectively. However, not all women are married in age group 44-49. Therefore, one cannot say that marriage is universal in Tanzania.


102

Table 21 reveals that the mean age at first marriage is 17.4 years. 10 percent of Tanzanian women marry before they are 14 years old. By the age of 20, three quarters of the women are already in marital unions. The table further reveals that there has been a rising trend in age at first marriage in Tanzania, e.g. the trimean of 18.1, 18.0, 17.3, and 17.0 years for women aged 25-29, 30-34, 35-39, and 40-44 years, respectively. The younger cohorts marry at later ages compared to the older cohorts. This increasing trend in age at first marriage is observed in T10, T25, T50, T75 and even in the trimean. The mean age at first marriage increased from 17 years in the 40-44 cohorts to 18.1 years in the 25-29 cohorts. This increasing age at first marriage could be attributed to the fact that more and more women in Tanzania today are pursuing education compared to the past. With an increasing level of education, women engage in wage employment. In addition to this, their values and attitudes are modified and these factors tend to delay marriage. The women in the oldest cohort 45-49 years seem to have overstated their age at first marriage most likely as a result of memory lapse. This resulted in an unexpectedly high age at first marriage of 17.4 years compared to 17.3 years for women at the age of 40-44.

Further computations of the TDHS data ( Figure 15 ) show that the median age at first marriage increases with the increase in the level of education of a woman. Women with no education have the lowest median age at first marriage namely 16.2 years, while those with primary school education have a median of 17.6 years. Women with secondary schooling have a median age of 9.4 years, women with an education higher than secondary schooling has a median age of 25.5 years. The low age at first marriage for women with no education could be attributed to the fact that marriage is the ultimate goal in their lives as soon as they mature. On the contrary, women with secondary education will marry later as they spend many years in school. Even after marrying, some of these women might not start child bearing immediately, as there are other important things like occupation, which might be competing with child bearing. In addition as far as most of them live in urban areas, their values have been modified by modernisation. All these factors might cause them to delay the first birth. In terms of place of residence, urban residents clearly demonstrate higher age at marriage relative to rural dwellers. Rural women enter and complete marriage at an earlier age than their urban sisters. Religion also influences age at marriage, especially among Moslems whose culture encourages earlier marriages than among Christians.


103

Figure 15: Socio-economic variables associated with age at first marriage


105

Table 22: Logistic regression coefficients predicting the relative odds that a woman marries before exact age 15

Variable

Coefficients

Odds ratio

Constant

-1.782***

LiteracyLiterateSemiliterateIlliterate

-1.486***

-.361***

RC

.23

.70

RC

ReligionMoslemCatholicProtestantOthers

RC

-.433***

-.285**

-.370**

RC

.65

.75

.69

Place of ResidenceUrbanRural

.134

RC

1.14

RC

Age15-1920+

.127

RC

1.14

RC

***

p<.001 (highly significant)

**

p<.01 (significant)

*

p<.05 (less significant)

p>.05 (insignificant)

Source: calculated from 1996 TDHS

Table 22 shows the effects of literacy, religion, place of residence, and current age on the likelihood that an ever-married woman marries before she reached age 15. After controlling for the other variables in the model, it appears less likely for literate women to have married before the age of 15 than illiterate women. Literate women are 4 times less likely to marry before age 15 than illiterates; the difference is significant. Semiliterate women are 1.4 times less likely to get married before they reach age 15 compared with illiterates. This positive effect of literacy on age at first marriage can largely be attributed to the fact that educated women tend to marry at a later age than uneducated women as they stay in school for a longer period.

Moslems seem to have a great likelihood of marrying before age 15. Catholics appear to be 1.5 less likely to marry before age 15 than Moslem women. Protestant women are 1.3 less likely to marry before age 15; the difference is statistically significant. This may be consistent with the strong emphasis on premarital virginity in Islamic society. Urban


106

residents are 1.14 times less likely to marry before the age of 15 than rural women, but the difference is not statistically significant.

Table 23: Mean age at first marriage for sub-Saharan countries and other selected countries

Country

Year

T10

T25

T50

T75

T90

T*

Spread

N

Benin

1996

14.2

15.8

17.6

19.7

22.2

17.7

3.4

4,445

CAR

1994

13.1

14.4

16.2

18.5

21.3

16.3

3.6

4,737

Comores

1996

13.1

14.9

17.2

20,1

23.8

17.4

4.3

1,862

Côte d’Ivoire

1994

13.4

15.0

16.9

19.6

22.8

17.1

4.1

5,964

Ghana

1993

14.8

16.2

18.0

20.2

22.8

18.1

3.3

3,672

Kenya

1993

14.1

15.7

17.8

20.2

22.7

17.9

3.6

5,260

Mali

1995

13.5

14.4

15.5

17.1

19.4

15.6

2.6

8,459

Tanzania

1996

13.7

15.3

17.3

19.6

22.5

17.4

3.8

6,221

Uganda

1995

13.2

14.7

16.5

18.7

21.1

16.6

3.5

5,965

Zambia

1996

13.8

15.3

16.9

19.0

21.6

17.0

3.3

5,989

Zimbabwe

1994

14.3

16.1

18.0

20.3

22.8

18.1

3.6

4,482

Egypt

1995

13.9

15.6

18.1

21.3

24.5

18.3

4.2

14,779

Bangladesh

1996

11.3

12.4

13.8

15.7

18.0

13.9

2.9

9,640

Brazil

1996

15.2

16.9

19.4

22.3

25.8

19.5

4.5

8,759

DR

1996

13.6

15.3

17.7

20.8

24.5

17.9

4.4

6,270

* Trimean

Source: calculated from DHS III

Mean age at first marriage in sub-Saharan Africa countries ranges between 15.6 in Uganda to 18.1 in Zimbabwe and Ghana. Women engage in marriage unions at an early age. Between the age of 13 to 14, 10 percent of women, who eventually got married, are already married. An interesting finding is Egypt where at the age of 13.9, 10 percent of woman have been married, but the mean age at first marriage is higher compared to other countries like Benin and Kenya. Bangladesh women seem to get married at the lowest age of all selected country. This might be attributed to the influence of Islamic faith in insisting on premarital virginity.

4.3.4. Primary Sterility

Sterility is the inability of a non-contracepting and non-lactating sexually active woman to have a live birth. The term infertility, sterility and infecundity are often used interchangeably regardless of the precise definition. However, demographic and medical definitions of these terms may differ substantially. For example in demographic terminology, primary sterility (primary infertility) is defined as the inability to bear any


107

children. Therefore according to this definition, primary sterility may arise because of the inability to conceive or the inability to carry a pregnancy to full term. One point to note is that sterility can only be considered for the non-celibate women.

In clinical studies on the other hand, infertility (sterility) is usually defined as the inability to become pregnant or to achieve fertilisation. Secondary sterility, which is highly correlated with primary infertility, is the inability to bear a child after having an earlier birth. Furthermore, the actual difference between infertility and infecundability is that infertility suggests simply a lack of demonstrated fertility rather than the physiological inability to reproduce. Infecundibility refers to inability to conceive or to bear a child after being exposed to the risk of conception for a fixed length of time.

While the epidemiological definition recommended by the World Health Organisation depends on a two-year period of exposure, the demographic measures cover a five-year period. One of the more serious problems with sterility measures that are based on the fertility of couples at the end of the child bearing period is that these measures do not reflect recent trends in sterility. The inclusion of a relatively small number of women over age 40 in the TDHS (15.6 percent of all sampled women in 1996) may influence sampling errors resulting in imprecise estimates.

Sterility might be due to the normal ageing process or to the consequences of a variety of diseases or malfunctions of the reproductive process (Vaessen, 1984). Although the causes of sterility are diverse, it is generally agreed that sexually transmitted diseases (STDs) are the major preventable cause of sterility (Farley and Besley, 1988). In particular, STDs have been implicated as the single major cause of the high level of primary and secondary sterility in sub-Saharan Africa (Caldwell and Caldwell, 1981). STDs may result in tubal obstruction, a problem that may also be caused by infections related to pregnancy and abortion (Cates et al., 1985). Infection is particularly likely after unhygienic obstetric or abortion practices.

There is no direct way of measuring primary sterility from the TDHS data. However, since most women marry and in practice there is no control to prevent a first birth, childlessness or the failure to conceive among married women may be taken as an indicator of primary sterility or very marked subfecundity. The assumption is that


108-109

marriages are stable, and abstinence is rare for married couples since marriage is an institution for producing children. And voluntary childlessness is not a common feature among Tanzanian women. The study will consider the proportion remaining childless as a fair indicator of either primary sterility or very marked sub-fecundity if the duration of marriage is more than 5 years without any live birth.

About 9 percent of the ever-married women aged 15-49 years reported that they never had a live birth. This is reduced to about 3 percent for those who had been married for at least five years. The percentages of currently married women that experienced no live birth were about 10 percent and were reduced to 2.4 percent for those who have been married for five years and more.

Table 24: Percentage of women, who remain childless by marriage duration and current age

a) Ever married women

Current

age

Duration of Marriage

0-4

5-9

10-14

15-19

20-24

25-29

30+

5+

Childless

Ever married

15-19

40.0

0.3

0.3

40.3

310

20-24

12.2

2.0

0.2

2.2

14.4

923

25-29

2.7

2.3

0.8

3.1

5.8

994

30-34

1.0

1.1

0.5

0.5

0.1

2.3

3.3

791

35-39

0.3

0.5

1.2

0.5

1.4

3.6

3.9

587

40-44

0.8

0.5

1.3

1.3

375

45-49

0.6

0.6

0.3

0.6

2.2

2.2

315

15-49

6.4

1.3

0.5

0.2

0.3

0.1

0.1

2.4

8.8

4,295

b) Currently married women

Current age

Duration of Marriage

0-4

5-9

10-14

15-19

20-24

25-29

30+

5+

Currently married

15-19

44.1

1.0

0.0

1.0

401

20-24

13.9

1.8

0.2

2.0

1,131

25-29

3.0

2.3

0.9

3.2

1,184

30-34

0.8

1.0

1.1

0.5

0.1

2.7

947

35-39

0.3

0.5

0.9

0.8

1.4

3.6

740

40-44

0.0

0.2

0.2

0.2

0.7

0.5

0.2

2.0

561

45-49

0.0

0.0

0.0

0.4

0.4

0.4

1.1

2.3

447

15-49

7.0

1.2

0.6

0.3

0.3

0.1

0.1

2.6

5,411

c) Women married for more than five years

Current age

Ever married women

Currently married women

No live birth

N

No live birth

N

15-19

26.7

15

28.6

14

20-24

4.2

526

3.2

467

25-29

3.4

1123

2.7

993

30-34

2.4

1010

1.9

896

35-39

3.3

861

2.9

729

40-44

1.8

666

1.3

556

45-49

1.7

578

1.3

445

15-49

2.9

4779

2.4

4,100

Source (a,b,c): calculated from 1996 TDHS

From Table 24 (a, b, and c) it is evident that the incidence of primary sterility is very low in Tanzania as only 2.9 percent of the ever-married women suffer this debility. It seems that this is a realistic estimate as the percentage of childless women who have been married for 25 or more years and are still currently married, amount to nearly the same figure (2.4 percent). But this figure might give a misleading estimate due to the fact that never married women were excluded in the analysis. However from the table, the incidence of primary sterility decreases with increase in duration of marriage. For example primary sterility decreases from 7 percent for married less than 5 years to 0.1 percent for those married for 30 years and more. Surprising enough, sterility tends to decrease with increasing age. For instance, 4.2 percent of ever-married women aged 20-24 were reported childless and a percentage decrease to 1.7 percent for ever married women age 45-49. This may lead us to conclude that either the choice of the period was too short to capture primary sterility hence this was secondary sterility or primary sterility is on the increase now in Tanzania. All in all, primary sterility is not very widespread.

To be precise let us use a refined method of investigating primary sterility by examining currently married women in their first union, who were married for five or more years before the survey and who are still childless.


110

Table 25: Percentage of childless currently married women in the first union

Age group

Mmarried for 5+

Regardless of duration

Married and in first union

15-19

0.6

42.0

385

20-24

0.8

14.0

1,012

25-29

1.8

4.6

995

30-34

1.7

2.4

732

35-39

2.5

2.9

522

40-44

1.2

1.2

390

45-49

0.9

0.9

320

15-49

1.4

8.9

4,356

35-49

1.7

1.9

1,232

Source: calculated from 1996 TDHS

Table 25 shows childlessness based on successively refined denominators, starting with ever-married women ( Table 24 ) and then adding in turn the conditions that the respondent was currently married at least for five years and was in her first union. The main reason for using women still in the first union, was due to the fact that the TDHS did not collect data on complete marriage histories or sexually active women. In this circumstance it is not easy to identify women who have been continuously exposed to the risk of conception for a period of five years unless they are still in their first marriage. However, measures based on currently married women may slightly underestimate sterility if couples without any children are more likely to get divorced or separated.

Table 25 confirms the earlier finding ( Table 24 ) that sterility in Tanzania is very low. It seems that 2 percent of married women in their first union for not less than five years are infertile. The relatively low level of infertility in Tanzania does not correspond with several studies, which have found a uniquely high level of sterility in sub-Saharan countries (Bongaarts et al., 1984; Caldwell and Caldwell, 1983). Farley and Besley (1988) went further to estimate primary sterility of about 10.1 percent for sub-Saharan Africa. Bongaarts et al. (1984) reported particularly high levels of sterility in most of Central Africa and East Africa of about 12-20 percent childlessness for women at the age of 45-49 years. It was previously thought that there was a core rate of sterility of about 10 percent due to genetic, anatomical, and endocrinological causes (Veevers, 1972). These findings might have exaggerated the real situation.


111

4.3.5. Contraceptive Use

Contraceptive use will be fully discussed in Chapter 6. However, it is important to discuss the use of contraceptives as a proximate determinant of age at first birth. Contraception might be practised in order to delay the first birth especially for women who marry late. In their case it would be preferable not to have premarital births as most women would like to get married before child bearing in Tanzania. In their study of induced abortion in hospital settings in Tanzania, Justesen et al. (1992) found that the proportion of unwanted pregnancies was strongly correlated with marital status, and decreases from 96 percent of single women to 31 percent of married women. They further found that the main reason for an unwanted pregnancy for never married women was being single, while married women had unwanted pregnancies due to the fact that either they already had the care of a young child or they wished to stop child bearing (Justesen et al., 1992). This confirms that women in Tanzania prefer not to have premarital births.

In this section we will try to investigate the contraceptive use in delaying first birth. However the TDHS did not collect information on contraceptive use before first birth for ever given birth. The only way is to assume that women who have already experienced a live birth had the same experience as the women who are currently contracepting in order to delay their first birth. In doing so, we further assume that marriage is an institution for generating children in a Tanzanian context, hence deliberate effort to delay first birth does not exist. In lieu of this assumption, investigation of contraceptive use among women who are not in union, but already have experienced sexual intercourse, who never had live births and or are not pregnant will give us the estimation of contraceptive use to delay first birth.


112

Table 26: Percentage distribution of non-pregnant unmarried sexually active women who use contraceptive to delay first birth by current age

Current age

Contraceptive method

Not Using

Traditional

Modern

N

15-19

86.6

4.8

8.6

291

20-24

83.6

5.3

11.2

152

25-29

76.3

15.8

7.9

38

30-34

92.3

7.7

13

35-39

100

10

40-44

100

6

45-49

100

6

15-49

85.7

5.6

8.7

516

Source: calculated from 1996 TDHS

It seems most women use modern methods rather than traditional ones to delay first birth. Table 26 shows that 8.7 percent of women use modern methods, 5.6 percent use traditional methods<17> of contraception to delay the first birth in Tanzania. However, it should be noted that the vast majority of women (86 percent) do not use any contraceptive method at all to delay their first births. In total 14 percent of women in Tanzania do something to protect first birth. In general unmarried sexually active women under 35 years of age seem to be the ones interested in delaying first birth, while older women aged 35 and over do not do anything to delay their first births. This might lead to the conclusion that women in Tanzania over 35 years are no longer hoping to get married and might not care any longer about premarital births as men in Tanzania prefer to marry young ones.

According to Table 26 , contraceptive use increases with age for unmarried women less than 30 years. However, adolescents use more modern methods than other groups do. This is an indication of the acceptance of modern methods in recent years. Let us investigate specific methods used by unmarried women to delay first birth.


113

Table 27: Percentage distribution of specific method used by current age

Age

Not using

Pill

Inj

DFJ

Condom

PA

WD

Strings<18>

N

15-19

86.6

2.7

0.7

0.3

4.8

4.5

0.3

291

20-24

84.1

2.6

7.9

5.3

152

25-29

74.4

5.1

5.1

12.8

2.6

38

30-34

92.3

7.7

13

35-39

100.0

10

40-44

100.0

6

45-49

100.0

6

15-49

85.7

2.7

0.4

0.2

5.4

5.2

0.2

0.2

516

Inj = Injection; DFJ = Diaphragm, Foam and Jelly; PA = Periodic Abstinence;
WD = Withdrawal

Source: calculated from 1996 TDHS

Table 27 shows that unmarried women mostly use condoms to protect them from premarital first birth. This can also be attributed to STD/AIDS control programs which emphasis the use of condoms (5.4 percent). The second popular contraceptive use of women in Tanzania to delay their first birth is periodic abstinence (5.2 percent). While 2.7 percent of women who want to delay their first births use pills, less than one percent use injections, and only adolescents use diaphragms as well as foams or jellies. These findings call for a review of family planning policy in Tanzania in order to accommodate adolescents and other never married women in the programme.

4.3.6. Induced Abortion and Early Pregnancy Wastage

The incidence of induced abortion and early pregnancy wastage before the first birth can delay the age at which child bearing begins. Therefore, it is one of the proximate determinants of age at first birth, too. Induced abortion is an illegal practice in Tanzania although unqualified paramedical personnel under poor hygienic conditions may secretly perform it<19>. These abortions always go unnoticed unless serious and fatal


114

complications develop and such patients are rushed to hospital where such cases must be recorded. As explained in Chapter 2, no attempt has been made in this study to assess the extent of the contribution of miscarriages, still births and deliberate abortion that occurred before the birth of the first child. This is due to the fact that no question was asked on the incidence of abortion and the pregnancy wastage and so, analysis of this variable is not possible in this study. However in this study we will use other studies to explain the magnitude and consequences of induced abortion as it contributes to infection, infertility, and mortality among young women.

Direct estimates of national levels of induced abortion are difficult to obtain except where the procedure is legal and where reporting systems function well. The WHO (1994) estimated that 3.7 million unsafe abortions are performed each year in sub-Saharan Africa (26 per 1,000 women). It is documented that unsafe abortion rates in sub-Saharan Africa are high and rising (Mashalaba, 1989; Leke, 1989; Salter et al., 1997). At the Kenyatta National Hospital in Nairobi, one ward admits 20-40 cases of incomplete abortion each day (Rogo and Nyamu, 1989).

In many countries, the majority of women seeking to terminate their pregnancy are single adolescents who wish to continue school or employment, or to wait to have a child until they can support one. In their study of 1,800 never married young people aged 14-25 in Nigeria, Nichols et al. (1986) reported that nearly half the female students and two-thirds of the nonstudents had been pregnant and nearly all had terminated their pregnancies with induced abortion. Mirembe (1994) documents that 68 percent of abortion patients at a local teaching hospital in Uganda were 15-19 years of age, and 79 percent were still in school. In a random sample of 300 women with early pregnancy loss admitted to Muhimbili Medical Centre, the teaching hospital in Dar es Salaam, Tanzania, Justesen et al. (1992) observed that 94 percent had induced abortion and that most of them were young and unmarried and their pregnancies were unwanted.

Legalising abortion might not be a perfect solution for prevention of unsafe abortions especially for first births as unsafe abortions also take place in some developing countries where abortion is legal. For example, in India abortion is legal, and yet many women seek abortions outside the formal health system because medical facilities


115

equipped to provide safe abortion are rare. Even where services are available in India, problems with confidentiality, quality, and cost prevent women from using them. In addition, many people are unaware of the fact that abortion is legal (Chandrasekhar, 1994; Henshaw, 1990; Pillai, 1993; World Bank, 1996). Of the estimated 5.3 million abortions induced in India in 1989, 4.7 million took place outside approved health care facilities and thus were potentially unsafe (Jesani and Iyer 1993). In Turkey, where abortion is legal, it must be performed or supervised by obstetrician-gynaecologists, which makes safe abortions inaccessible to most rural women (McLaurin et al., 1991). In Zambia, abortion is legal but many women and service providers are unaware of its legality. Moreover, legal safe abortion is inaccessible to most women because they must obtain the consent of three physicians (Likwa and Whittaker, 1994). Thus for every woman in Zambia obtaining a legal abortion in 1991, five sought emergency treatment for complications of unsafe induced abortions (Bradley et al., 1991).

Abortion complications account for an estimated 13 percent of all maternal deaths in Africa according to the 1994 WHO report. In some countries, hospital-based studies report much higher percentages. For example in Ethiopia, a hospital-based study estimated that abortion complications accounted for nearly 40 percent of maternal deaths (Yoseph and Kifle, 1988). In Nigeria during the 1980s at two teaching hospitals, abortion complications accounted for 20 percent and 35 percent of maternal deaths respectively (Okonofua et al., 1992). At a third hospital in Nigeria 37 percent of gynaecologic deaths were due to abortion complications (Adewole, 1992). Rogo (in Jacobson, 1990) estimates that 20 percent of maternal deaths in East and Central Africa, and as many as 54 percent in Ethiopia were due to abortion complications.

Untrained practitioners perform induced abortions or the woman induces the abortion herself. In Ethiopia, unskilled non-medical persons performed 92 percent of the illegal abortions (Ethiopia Country Paper, 1994). Leke (1989) reports that Cameroon women who do not go to traditional practitioners use local herbs and corrosive substances like potassium permanganate. According to Archibong (1991), methods used in Nigeria include sharp metal rods and hazardous chemicals. In Zambia, women are known to use twigs, drink detergents or gasoline, or take large doses of chloroquine or malariaquine (Castle et al., 1990).


116

Most common abortion complications are incomplete abortion, sepsis, haemorrhage, and intra-abdominal injury (Aggarwal, 1984; Konje et al., 1992; Ladipo, 1990; WHO, 1994). Except for intra-abdominal injury, all complications can result from either spontaneous abortion (miscarriage) or induced abortion; if left untreated, each can lead to death (Kamau, 1990). Women surviving immediate abortion complications often suffer life-long disability or face elevated risk of complications with future pregnancies (Herz and Measham, 1987; Liskin, 1992).

When tissue remains in the uterus after either miscarriage or unsafely induced abortion, women suffer ‘incomplete abortion’ - the most common abortion complication. Typical symptoms include pelvic pain, cramps or backache, persistent bleeding, and a soft, enlarged uterus (Stubblefield and Grimes, 1994; WHO, 1995). Septic abortion results when the endometrial cavity and its contents become infected (Ladipo, 1990). Usually this occurs after contaminated instruments were inserted into the cervix or when tissue remains in the uterus (Mccauley and Salter, 1995). In addition to suffering the general symptoms of incomplete abortion, women with sepsis have fever, chills, and foul-smelling vaginal discharge while bleeding might be either slight or heavy (WHO, 1995). The first signs of septic abortion usually appear a few days after the miscarriage or unsafe abortion. The infection can quickly spread from the uterus to become a generalised abdominal sepsis. High fever, difficult breathing, and low blood pressure often indicate a more extensive infection (Sweet and Gibbs, 1990).

Some techniques to induce abortion, such as sharp curettage or inserting sticks or other objects into the cervix can result in intra-abdominal injuries that cause heavy bleeding. Herbs, drugs, or caustic chemicals swallowed or placed into the vagina or cervix can cause toxic reactions and also lead to haemorrhage. The risk of post-abortion haemorrhage increases with gestational age, as well as with the use of general anaesthesia during unsafely induced abortion (Chaudhuri, 1992).

When instruments are inserted into the cervix to cause abortion, the cervix, the uterus, or other internal organs can be cut or punctured. The most common injury is perforation of the uterine wall. The ovaries, fallopian tubes, bowel, bladder, or rectum also can be damaged (WHO, 1994). Intra-abdominal injury can cause internal haemorrhage with little or no visible vaginal bleeding. Sepsis and haemorrhage resulting from spontaneous


117

abortion or unsafely induced abortion often are the most common reasons that women in developing countries seek treatment in hospital obstetric and gynaecologic wards (Ross and Frankenberg, 1993). In Kenya, for example, two hospital-based studies conducted during the 1980s found that women with post-abortion complications accounted for 60 percent of all gynaecological admissions (Aggarwal and Mati, 1982).

In addition to causing many deaths and much suffering, unsafe abortion complications consume a large portion of health-care budgets and scarce medical resources. In some areas for example, large amounts of resources, such as blood supply, are used for treating complications of unsafe abortion (Gyepi-Garbrah et al., 1985; Jacobson, 1994). It seems that the only solution for overcoming this human tragedy is to encourage the use of contraceptives rather than legalising abortion for a poor developing country like Tanzania.

4.4. Interval between First Marriage and First Birth (Premarital first Births)

Marriage though a very important landmark in fertility, does not necessarily mark the beginning of exposure to sexual activities (see discussion before). This analysis will shed light on the rate at which women start child bearing before and after first marriage. However, for the computation we assume that child bearing occurs strictly within marriage and any birth occurring before marriage will show negative interval. X represents those births which occurred before first marriage. Y represents occurrence of births within marriage but less than 9 months after first marriage, which means the conception of these births, was outside marriage. Z represents the interval between first marriage and first birth of more than 9 months - in other words, this represents women who had their first conception in marriage. X and Y represent premarital conception, or those pregnancies occurring before marriage.


118

Table 28: Percentage of ever married women who had ever given birth by current age

Current Age

X

Y

Z

Premarital

Conception

Total ever giving birth

15-19

9.2

16.9

73.8

26.1

260

20-24

12.6

19.0

68.4

31.6

1,087

25-29

18.3

18.6

63.1

36.9

1,260

30-34

20.7

18.1

61.3

38.8

1,035

35-39

16.1

15.4

68.4

31.5

843

40-44

15.0

16.7

68.3

31.7

659

45-49

13.0

19.1

68.0

32.1

571

15-49

16.1

17.8

66.1

33.9

5,715

X = first births occurred before first marriage for ever married women; Y = first births within marriage but less than 9 months after first marriage; Z = first birth of 9 months or more after the first marriage

Source: calculated from 1996 TDHS

Table 28 reveals that almost 34 percent of ever married Tanzania women had their first conception before marriage. Out of those premarital conceptions, 16 percent are among married women who reported having their first birth before marriage also known as illegitimate birth for ever married women, and 18 percent are women who conceived their first pregnancy outside marriage but were married before the first birth.

Premarital birth increases with age for ever-married women less than 35 years, whereas it decreases with increasing current age. This finding can be associated with the results of section 4.2.6 that women less than 35 years of age are the only ones who use contraceptive to delay first birth while older women by this time have had their first birth, but if not, might be interested in having a child even outside wedlock. It is interesting also to note that marital births decrease with age. This may be attributed to the chance of getting married after the adolescent period.


119

Table 29: First marriage to first birth interval (in months) by current age

Percentile

Current age

15-19

20-24

25-29

30-34

35-39

40-44

45-49

15-49

T10

5.6

4.4

4.2

4.4

4.5

4.8

4.3

4.5

T25

9.4

8.9

8.9

8.8

9.4

9.5

8.8

9.0

T50

13.0

13.0

13.0

13.4

14.3

13.6

14.0

13.4

T75

19.2

21.6

21.0

22.5

23.9

22.4

24.6

22.1

T90

27.2

33.8

35.1

37.5

42.3

36.7

43.3

36.8

Trimean

13.6

14.1

14.0

14.5

15.5

14.8

15.3

14.5

Spread

9.4

14.2

14.7

16.4

18.0

15.8

17.2

15.6

Source: calculated from 1996 TDHS

Table 29 indicates estimates derived from life table analysis of data for all ever married women in order to investigate premarital births as found in Table 28 in a refined way. Here the series of quantiles Tx represent the number of months elapsed since marriage by which x percent of women had given birth to their first child. In other words, this table represents data on interval length between the first marriage and the first birth. The average interval from first marriage to first birth is 15 months, an indication that most of first births occur in marriages. However, 10 percent of ever-married women have had their first birth by 4.5 months of their marriage and 25 percent by 9 months. This is an indication that for all women in Tanzania who had their first live births, 25 percent were premarital pregnancies. By 3 years (36.8 months), 90 percent of the ever-married women have had their first birth. In other words, only 10 percent of women have not had their first child after 3 years of their first marriage.


120

Figure 16: Mean number of children ever born to ever-married women according to interval between age at birth and marriage

Key:

X = first births occurred before first marriage for ever married women

Y = first births within marriage but less than 9 months after first marriage

Z = first birth of 9 months or more after the first marriage

Source: calculated from 1996 TDHS

Figure 16 suggests that fertility for those, who had their first live birth out of wedlock, is high compared to others. This finding supports the evidence of high rate of premarital births as already indicated earlier. This could be attributed to the early age at which sexual activities among Tanzanian women begin coupled with the low level of contraceptive prevalence to delay first birth. With increasing age at first marriage, premarital births are bound to increase if stern measures are not taken.


121

Table 30: Percentage of women who had ever given birth by current age

Age

Never married

X

Y

Z

Premarital

Conception

Premarital

Birth

Total ever giving birth

15-19

28.2

6.6

12.2

53.0

47.0

34.8

361

20-24

12.9

11.0

16.5

59.6

40.4

23.9

1247

25-29

4.7

17.5

17.7

60.1

39.9

22.2

1322

30-34

3.7

19.9

17.4

59.0

41.0

23.6

1075

35-39

1.4

15.9

15.2

67.5

32.5

17.3

855

40-44

1.3

14.8

16.5

67.4

32.6

16.2

668

45-49

0.5

12.9

19.0

67.6

32.4

13.4

574

15-49

6.4

15.0

16.7

61.9

38.1

21.4

6102

X = first births occurred before first marriage for ever married women; Y = first births within marriage but less than 9 months after first marriage; Z = first birth of 9 months or more after the first marriage;

Premarital first conception = Never married +X + Y;

Premarital first birth = Never married + X;

Legitimate first conception = Z; Legitimate first birth = Y + Z

Source: calculated from 1996 TDHS

Table 30 shows that the total of premarital first conceptions in Tanzania was almost 40 percent for women who ever gave birth in the 1996 TDHS. Out of those premarital first conceptions, almost 7 percent are never married women who reported having had a birth; 15 percent are married women who reported having their first birth before marriage; 17 percent are women who conceived their first pregnancy outside marriage but were married before the first birth. Therefore, 22 percent of women in Tanzania had premarital first birth, although premarital conception amounted to 40 percent of all conceptions.

Studies by Hobcraft and McDonald (1984) in 28 countries in Africa, Asia and America using World Fertility Survey data showed that at least 10 percent of women in most countries surveyed reported a first birth before 9 months of marriage. Our finding is within the range and supports the evidence of a high rate of premarital births and conceptions in Tanzania. This could be attributed to the early age at which sexual activities among Tanzanian women begin, coupled with the low level of contraceptive practice to delay first birth.

It seems that premarital first births are currently increasing, because premarital first births increased from 32 percent 30 years ago to 47 percent in recent years. However, the percentage of women, who became pregnant and eventually were married, also increased for the same period from 13 percent to 35 percent of all women who gave birth in Tanzania. This might be attributed to modern men’s guile in preconditioning a


122

premarital sexual relationship with a young girl to test if the intended wife is fecund. Secondly the prevailing low age at first sexual intercourse can be a cause of premarital births. Never married births have been increasing greatly for the same period from less than 1.0 percent to 28 percent of all first births in Tanzania. However, in section 4.2.6 we saw an increase in contraceptive use to avoid premarital births for unmarried women. If all efforts are directed towards the provision of contraceptives to unmarried women in Tanzania, unwanted premarital births might be avoided to some extent.

4.4.1. Covariates of Age at First Birth

Age at first birth, as already discussed, could be influenced directly by proximate determinants. However, there are also the socio-economic characteristics which act through the proximate determinants to indirectly influence the age at which child bearing begins. The socio-economic factors that will be analysed in this study include education, type of place of residence, and religion.

Although the impact of these socio-economic variables is frequently emphasised, there is no doubt that the reverse situation also might be true. For example, due to the incompatibility of child bearing with school attendance, and in many societies with wage-earning employment outside the home, the timing of child bearing can influence the education and employment experiences of young women. This section will focus on how these factors influence age at first birth. At the end of this chapter, a multiple regression analysis will be carried out to determine the contribution of each factor on age at first birth. The effect of age at first birth on fertility will be investigated in the next chapter.

4.4.2. Place of Residence

Urbanisation and its concomitant modernising effects are believed to have a depressing effect on age at first motherhood. People who migrate from rural to urban areas are mostly those with education who want to elevate their work status or look for a better life in urban areas. As such, women residing in urban areas will obviously have more education and be in the formal sector of the economy. Table 31 shows the mean age at first birth according to type of place of residence.


123

Table 31: Mean age at first birth by type of place of residence

Place of Residence

T10

T25

T50

T75

T90

Trimean

Spread

Urban

14.8

16.5

18.2

20.3

22.8

18.3

3.2

Rural

14.8

16.4

18.2

20.2

22.6

18.3

3.2

Source: calculated from 1996 TDHS

Table 31 does not show evidence of a rural-urban differential in age at first birth. The mean age at first birth according to the TDHS data is the same, 18.3 years for the urban and rural women as well. This seems to indicate that in Tanzania child bearing does not depend on place of residence. The age at which 10 percent of the urban women get their first birth is the same as their rural counterparts namely 14.8 years.

It was expected that women in rural areas would have a lower age at first birth than their urban sisters for the following reasons: Urban women are more likely to practise contraception than their rural counterparts as shown earlier. Furthermore, the values and attitudes of the educated urban women have been modified through modernisation. Rural women on the other hand still adhere to the traditional norms, values, beliefs and practises, which attach a lot of importance to marriage and reproduction. They will therefore start child bearing early so as to meet societal expectations. In Tanzania it seems that is not the case. This might be attributed to UPE (Universal Primary Education program) and Kiswahili which is an official medium for communication among all Tanzanians.

In order to account for the difference between urban areas in general and the city of Dar es Salaam in particular, it is necessary to separate one from the other. After the separation, we have different results than in Table 31 . Table 32 shows that urban women (not residents of Dar es Salaam) have a higher mean age at first birth (almost 3 months) and rural women have a lower mean age at first birth of 18.3 years. The findings reveal another unexpected result: 10 percent of Dar es Salaam women already gave birth by 14.4 years. Women in other urban areas reached the same figure at 15.2 years. 50 percent of rural resident women have given birth by the age of 18.3, urban women resident elsewhere than Dar es Salaam at the age of 18.4 and women residing in Dar es Salaam at the age of 17.9 years. This means that half of the Dar es Salaam women have their first children at a younger age than rural or other urban women. Considering 90 percent of the sampled women, we get the following results on the mean


124

age at first birth: 22.6 years (rural women), 23.2 (urban, not resident of Dar es Salaam) and 21.8 years (Dar es Salaam women) respectively. Dar es Salaam, the capital of Tanzania, has the lowest mean age at first birth. According to our assumptions, this big city was expected to have the highest mean age at first birth.

The bias might be in the TDHS data since 90 percent of Dar es Salaam women sampled were Moslems. As we concluded earlier, Moslem women are more likely to get their first child earlier than any other religious denomination. Therefore, the unexpected low age at first birth might be due to religion. Another reason might be the rural-urban migration for unmarried pregnant adolescent who are rejected by their families or young women who come to Dar es Salaam in order to find petty jobs. Since unemployment is high in the capital, a lot of them end up as prostitutes. Since contraceptives are neither available nor affordable to these young rural women, unprotected sex leads to motherhood at young ages - not to speak of STDs and HIV.

Table 32: Mean age at first birth by type of place of residence

Place of Residence

T10

T25

T50

T75

T90

Trimean

Spread

Dar es Salaam (capital)

14.4

16.1

17.9

19.9

21.8

17.9

2.9

urban areas

15.2

16.7

18.4

20.6

23.2

18.5

3.4

rural areas

15.0

16.5

18.3

20.3

22.6

18.3

3.2

Source: calculated from 1996 TDHS

4.4.3. Religion

Religion is believed to play a part in shaping the views, norms, belief attitudes and practises of the people. These will in turn govern the reproductive behaviour, and subsequently fertility. There are variations in values regarding marriage, marital dissolution and contraception among the different religions. The Table 33 show the age at first birth by religion.


125

Table 33: Mean age at first birth by religion

Religion

T10

T25

T50

T75

T90

Trimean

Spread

Moslem

14.5

16.3

18.0

19.9

22.0

18.0

3.1

Catholic

15.2

16.7

18.6

20.6

22.9

18.6

3.2

Protestant

15.2

16.8

18.4

20.4

22.9

18.5

3.2

None

14.7

16.2

18.0

20.1

23.2

18.1

3.5

Source: calculated from 1996 TDHS

From the Table 33 it is clear that religions seem to play a role in influencing the age at first birth however small. This is revealed by all the measures used, i.e. the trimean and all the percentiles, T10, T25, T50 and T75. Moslems show a lower mean age at first birth of 18 years than other religious affiliates. Although the Catholics and Protestants have different views regarding the use of contraception, Catholics seem to have higher mean age at first birth compared to Protestants. The Catholic Church does not approve of modern methods of contraception while the Protestant church is liberal on that issue. It is therefore expected that the Catholics should have a lower age at first birth. However, this is not the case in Tanzania where Catholics seem to have a higher mean age at first birth (18.6 years) compared to Protestants (18.5 years). It could therefore be interpreted to mean that the Catholics might be going against the view of their church as far as contraceptive use is concerned. As already mentioned, Moslems are known to marry earlier than Christians.

4.4.4. Education

Education, particularly women’s education, has a significant interaction effect on fertility. It is believed that lower education leads to higher fertility, and likewise early fertility means lower education. With education, new values, aspirations, and a new outlook on life as well as skills for taking advantage of new opportunities come along. Advancement in women’s education increases the potential for participation in a paid labour force in the modern sector of the economy. This in turn increases the opportunity cost of child bearing and creates trade-offs between child bearing and participation in the paid labour force. Education will keep women in the schools for longer years, thus delaying birth.


126

Table 34: Mean age at first birth by level of education

Education

T10

T25

T50

T75

T90

Trimean

Spread

None

14.2

15.8

17.6

19.9

22.8

18.5

3.6

Incomplete primary

14.3

15.9

17.6

19.5

21.5

18.2

3.0

Complete primary

16.0

17.2

18.7

20.5

22.5

19.3

2.7

Secondary +

15.9

18.1

20.2

23.0

25.7

21.1

3.8

Source: calculated from 1996 TDHS

Table 34 shows a clear relationship between the level of education and age at first birth. Those women with no education have a lower mean age at first birth compared to those with secondary and above level of education. This trend is revealed by all the percentile measures i.e. T10, T25, T50 and T75. The median age at first birth for women with no education is 17.6, while the median age for women with secondary level of education is 20.2 years. This finding supports the argument that education delays age at first birth. The reader may notice that the mean age at first birth for none and primary incomplete is at the same level. This is due to the fact that primary education in Tanzania lasts for seven years. The age to start primary school is 7 years, therefore most girls who complete primary education are 14 years old. At that age most have just attained age at menarche and few have not.

The trimean further reveals the differential in age at first birth by level of education. The mean age at first birth for women with no education is 18.5 years while that for women with secondary is 21.1 years. Women with secondary and above levels of education delay marriage as they stay for many years in schools compared with those with primary education. In Tanzania, once a girl completes primary education and fails to get a place in secondary school, she has no option but to remain in the rural area. The problem is that she has not acquired any skills that can enable her to get a job in the modern sector of the economy. This forces many of these girls into sexual relationship as a way of achieving economical security and social status for those who get married. By contrast urban primary school leavers have more opportunities to get a paid job in towns. Women who have secondary education will have higher chances of getting employment in the modern sector of the economy and will thus migrate to the urban areas. Furthermore, education modified women’s values and aspirations in life so that they are more aware of the use of contraception to delay the first birth.


127

As can be observed from the table, the spread for women with no education and those with secondary level of education is wide, at 3.6 years and 3.8 years respectively. This suggests that births to women in these two educational categories have a wider distribution. On the other hand the spread for women with primary education is small, 2.7 years suggesting that the births of these women are concentrated within a narrow range. There is a difference of almost 3 years between those women with no education and those with secondary education (trimean), suggesting that secondary education on average will delay age at first birth by 3 years, while primary education will delay age at first birth for a year.

4.5. Demographic Implications of Age at First Birth

As we have already seen, age at first birth can be influenced by a number of socio-economic and behavioural factors. It is also likely that age at first birth can influence a number of demographic phenomena such as fertility, infant and child mortality, maternal mortality, and population growth.

4.5.1. Fertility

Fertility in Tanzania is in a transition of decline (Mturi and Hinde, 1995). Mturi and Hinde (1995) associated this decline with increasing age at marriage, which will in turn mean postponement of child bearing. But this argument was proved not to be true due to the prevalence of premarital births. However, the decline of fertility could have come about as a result of improvement in the status of women, and especially as far as women’s education is concerned. In this section the mean number of children ever born will be used as a measure of fertility, and its relationship with age at first birth will be examined.


128

Table 35: Mean number of children ever born by age at first birth

Age at first birth

MNCEB

N

Spread

<15

5.9

421

2.9

15-17

4.6

1,905

2.9

18-19

3.8

1,794

2.7

20-21

3.5

1,062

2.5

22-24

3.5

625

2.3

25+

3.2

296

2.3

Total

4.1

6,102

2.8

Source: calculated from 1996 TDHS

It is evident from Table 35 that the mean number of children ever born decreases with increasing age at first birth. This is to be expected because all things being equal, women who start child bearing early are likely to have a longer period of exposure to the risk of pregnancy than those who start late. Having a longer reproductive span might result in more children. These results show that there exists a relationship between age at first birth and fertility. Further analysis on the relationship between age at first birth and fertility will be done using regression analysis in Chapter 5.

4.5.2. Infant and Child Mortality

Studies have shown that there is a relationship between age at first birth and infant and child mortality (Bicego and Boerma, 1993). Results of such studies indicate that the younger the age of the mother at the birth of the first child, the higher the chances of the child dying due to complications during childbirth. The factors that cause this effect in developing countries are not entirely understood. Some of the factors are known, such as economic hardships, malnutrition, poor medical care. However, to which extend these factors influence child mortality has to be studied yet. Certainly, there is a biological basis for the poor survival experience of first births as many first births take place before a woman has reached full physical and reproductive maturity, leading to increased perinatal risks and a more difficult delivery. Also a first time mother may be poorly prepared to handle new roles and responsibilities in her life with the baby. However, the degree of risk elevation associated with age at first motherhood varies with the level of access to, and use of, high quality antenatal and obstetrical services which are dependent on the level of community development, the economic situation of the household and the educational level of the mother. In this study infant and child mortality only will be considered for the first-born child.


129

Table 36: Childhood mortality rates (first born) by mother’s age at birth; mortality rates (deaths per 1,000)

Age at first birth

Neonatal

Post-neonatal

Infant mortality

Child mortality

Number of deaths

Number of births

<15

42.8

73.7

116.5

35.7

65

421

15-17

40.9

40.9

81.9

21.0

205

1,905

18-19

38.5

43.5

81.9

16.2

179

1,794

20-21

26.4

30.1

56.5

11.3

73

1,062

22-24

40.0

28.8

68.8

8.0

51

625

25+

30.4

30.4

60.8

30.4

29

296

Total

37.2

40.3

77.5

18.0

602

6,102

0-4 years before TDHS<20>

31.7

55.7

87.5

53.7

Source: calculated from 1996 TDHS

Table 36 shows an inverse relationship between age at first birth and under-five mortality. Neonatal mortality for first borns is 37.2 per 1,000 live first births, while neonatal mortality for all births was found to be 31.7. This means that there is excess neonatal mortality for first born. The average excess infant mortality associated with young age at birth (less than 15 years) is 39 deaths per 1,000 first births (from the average of 77.5 per 1,000 births) while women who had their first birth as adolescents have higher than average neonatal, postneonatal and infant mortality. Women who had their first birth after attaining age 20 have the correspondingly lower infant and child mortality than the average. A number of reasons could be attributed to this trend as explained earlier. Furthermore, in most cases women who start child bearing early do not seek medical attention especially for the first few pregnancies due to the shame of meeting older women in the antenatal clinics. As a result, such women face problems during childbirth, as there was no follow-up during the pregnancy. Coupled with this, their pelvis is not fully developed to accommodate the baby and it is thus likely that inflammatory diseases and other complications might develop leading to the death of the baby during or after childbirth.

Babies born to very young mothers are in most cases underweight, as the mother’s nutrition and health were not catered for during the pregnancy. It is also possible that the baby does not have enough space due to underdeveloped reproductive organs of the


130

mother. Such babies once born are likely to die unless extra care and medical attention are provided.

4.5.3. Maternal Mortality and Morbidity

Maternal deaths are defined as any death that occurs during pregnancy, childbirth or within two months after the birth or termination of a pregnancy. Therefore, the maternal mortality rate is the annual number of maternal deaths per 1,000 women aged 15-49. However, most literature prefers to express these deaths as maternal mortality ratio which is expressed per 100,000 live births. The maternal mortality ratio (MMR) can be obtained by dividing the maternal mortality rate by general fertility rate for the same reference period. Although there are no accurate data of pregnancy related mortality and morbidity, a lot of the literature shows that complications associated with early childbirth (below age 20) are worse than for births in the older age groups (20-34 years).

In a developing country like Tanzania where medical conditions are inadequate, especially in the rural areas, and high incidence of infectious disease, and poverty are prevalent, the maternal mortality is expected to be high. From this fact it is expected that maternal mortality due to the first pregnancy or birth is much higher than in subsequent pregnancies.

The information collected in the TDHS does not allow us to investigate only deaths due to first pregnancy or first birth. This is due to a fact that the only information on maternal death is based on the list of all siblings given by respondents. Siblings means all the children born to the mother of a respondent starting with the first born and whether or not each of these siblings was still alive at the time of the 1996 survey. Age and the cause of death of sisters more than 12 years was collected. In this study we will assume that adolescent maternal death was due to first births complication. Table 37 indicates results of direct estimate of maternal mortality for a period of 10 years before the survey. It is estimated that 137 women died due to pregnancy related causes. The general fertility rate (GFR) is the number of live births per 1000 women. According to Table 37 maternal deaths accounted for 27 percent of all deaths to women aged 15-49, and maternal mortality ratio was found to be 511 deaths per 100,000 live births (This figure is reached by dividing maternal mortality rate by general fertility rate and


131

multiplying by 100,000). Adolescent maternal deaths account for 27 percent of all adolescent deaths in Tanzania. Hence the adolescent maternal mortality ratio is estimated to be 408 per 100,000 live births.

Table 37: Maternal mortality by age at death

Age

Maternal deaths

Exposure

Mortality rates

Proportional of maternal deaths to female deaths

GFR

MMR

15-19

17

30,867

0.551

.266

.135

408

20-24

38

30,743

1.236

.404

.260

475

25-29

31

26,243

1.181

.287

.255

463

30-34

20

20,220

0.989

.238

.217

456

35-39

18

13,980

1.288

.257

.167

771

40-44

6

8,239

0.728

.118

.087

837

45-49

8

4,356

1.837

.267

.042

4,373

15-49

137

134,649

1.017

.274

.199

511

Source: calculated from 1996 TDHS

4.5.4. Natural Increase

The age at which women begin to have children has a close relationship to the overall growth of the population. As most births take place at a very early age as seen in this study, it is an indication of higher fertility and subsequent rising population growth. The earlier the child bearing begins the greater the number of fertile years remaining, and the greater the likelihood of a big number of completed family size.

Age at first birth cannot be the sole factor in the natural increase of the country’s population. There are other biosocial, economic and cultural factors at play. However, an increase in the average age at first birth both within and outside marriage, will in theory have a negative impact on the growth rate of the population. If age at first child bearing could be raised by increasing the contraceptive prevalence rate to delay or prevent premarital births, the improvement in lowering the population growth in Tanzania will be inevitable. This is due to the fact that for a broad based country like Tanzania late age at first birth means fewer women will have contributed for a longer duration to the natural increase. However it must be noted that late age at first birth on its own cannot reduce the fertility of a country, unless birth intervals is also checked by contraceptive use to avoid ‘the compensation factor’.


132

4.6. Regression Analysis

The analysis carried out so far shows that there are major differences between certain subgroups of the population with respect to age at first birth. However, a bivariate analysis does not determine the extent to which these differences are directly related or associated with the particular variable concerned, and the extent to which they are the effects of other intervening variables. A useful technique for testing such effects is the use of multivariate analysis, which makes it possible to estimate the net effect of each variable when variation in the other selected factors are controlled. The independent variables interrelate with each other, and their interactions can alter the effects observed in bivariate analysis.

Age of the respondent at first birth as the dependent variable was treated as a continuous variable. The unit of analysis is a woman ever given birth. We did not consider age at first marriage as a background variable in this analysis because first birth may precede first marriage and vice versa. All the independent variables have been treated as dummy variables except current age. These include religion, education, and type of place of residence. For each of these background characteristics, one category has been selected as the reference category (RC) and is omitted from the equation. This is the one to which the regression coefficients are compared. Regression equation is as follows:

Age at first birth = f (age, age at first intercourse, education, literacy, religion, place of residence, childhood place of residence).

Age at first birth = alpha + betaixi

alpha is a constant and beta1....betak are unstandardised regression coefficients for each of the respective explanatory variables xi.


133

Table 38: Regression results (unstandardised coefficients) of the relationship between age at first birth and selected variables

Variable

Name

Dummy

Unst. coef.

Std. error

Current Age

15-49

Cont.

0.047***

0.005

Age at first sexual intercourse

<15

RC

at union

I2

0.862***

0.130

15-17

I4

0.812***

0.126

18-19

I5

1.694***

0.153

20-22

I6

3.093***

0.215

22-24

I7

4.983***

0.345

25 +

I8

7.348***

0.479

Education

None

RC

Primary incomplete

E2

-0.032

0.160

Primary complete

E3

1.237***

0.186

Secondary +

E4

1.880***

0.243

Literacy

Illiterate

RC

Semiliterate

L2

-0.206

0.173

Literate

L3

-0.160

0.170

Religion

Catholic

RC

Moslem

R1

-0.353***

0.095

Protestant

R3

0.011

0.107

Residence (combined)

Rural

RC

Urban

U1

-0.211*

0.101

Place of residence

Dar es Salaam

RC

Other urban

K2

0.531**

0.200

Rural

K3

0.655***

0.181

Constant

15.016***

0.284

Adj. R2 = 0.171, Durbin-Watson = 1.918

***

p<.001 (highly significant)

**

p<.01 (significant)

*

p<.05 (less significant)

p>.05 (insignificant)

Source: calculated from 1996 TDHS

Table 38 shows the regression results estimating the effect of background variables on age at first birth. All variables included in the multiple regression analysis explain 17 percent of the variability in age at first birth. It means therefore that the available background variables have not explained much of the variation in age at first birth but they play a significant role. The Durbin-Watson was 1.918 that means there is no auto-correlation between background variables in the regression analysis. The relationship between current age and age at first birth found earlier is confirmed in the regression analysis. Current age is found to be in a positive relationship with age at first birth. This relationship is statistically significant as can be observed from the table.


134

There is also a strong relationship between age at first intercourse and age at first birth. The result of regression indicates that if the age at first sexual intercourse increases, the age at first birth also rises. Women who start their sexual intercourse between age 15 to 17 have their first birth one year later than those who start before age 15. Those who had their first sexual intercourse after adolescence give first birth more than 3 years later than those who started having sex before the age of 15; the difference is statistically significant. It is interesting to note that if a woman has her first sexual intercourse when she is older than 24 years, her age at first birth rises to more than 7 years in comparison to those who start sexual intercourse before attaining age 15. Hence it seems that discouraging adolescents from sexual intercourse may raise age at first birth in Tanzania.

As can be observed in Table 38 , age at first motherhood rises with the increase in the level of education. Primary education raises age at first motherhood by 1.2 years above those with no education; the difference is statistically significant. Women with secondary or higher education have the most significant increase in age at first birth as they delivered their first child on average 2 years later than those with no education. This confirms the hypothesis that education is positively related to age at first birth as women with higher education have a higher age at first birth. Incomplete primary education seems to be insignificant. This however can be explained by the fact that most of primary school dropouts engage in sexual activities earlier than their fellow pupils who remain at school. Pregnancy might be the main reason for dropping out.

The results of the regression analysis support the hypothesis that on average Christians give birth later than Moslems. Moslems tend to have their first birth 0.4 years before Catholics; the difference is statistically significant. These results may conflict with earlier findings that Catholics have higher fertility than other Christians. On the other hand it is possible that there is no relationship between age at first birth and lifetime fertility at all. This hypothesis will be examined in the following chapter.

The regression results show that urban women give first birth 0.2 years earlier than their rural counterparts; the difference is statistically significant. This is not an expected result since women in urban areas are more educated, more exposed to new ideas and life styles, and are thus more likely to be engaged in wage employment. Therefore they


135

are expected to start child bearing later. In addition, they are more likely to use modern methods of contraception. The modernising effects of urban areas could also change the values and attitudes as far as children are concerned. The high costs of living in urban areas could discourage women from having children.

In order to find out why the findings did not correspond to our expectations we subdivided the category ‘urban residence’ into the city of Dar es Salaam and other urban areas. It was interesting to note how the results changed after this modification. It seems that Dar es Salaam residents have lower age at first birth while other urban women gave the first birth 0.5 years later than Dar es Salaam women. Rural residents give first birth 0.7 years later than women living Dar es Salaam. Women living in rural areas in Tanzania give birth later than urban residents. The reasons why the relationship between age at first birth and type of residence are contrary to what we expected have to be further researched due to the fact that, this study is to a certain extend limited in its range of issues. Therefore, this topic has to be dealt with elsewhere.

4.7. Concluding remarks

The study has established that age at first sexual relationship, age at first marriage and contraceptive use are among the factors that influence age at which child bearing begins. In fulfilling the second objective which was to investigate the differentials in age at first birth with respect to background variables, the study has come up with the finding that education, place of residence, and religion play the greatest role in influencing age at first birth in Tanzania.

The study has revealed that the vast majority of Tanzanian women start child bearing at a relatively early age, 18.4 years on average but as early as 14.8 years for 10 percent of the women. According to the study, 50 percent of the women are mothers by the age of 18.2. Although the Tanzanian Demographic and Health survey data have revealed an increasing trend of age at first birth across cohorts, it is still relatively low by comparison with a number of developing countries although in line with other African countries South of Sahara. Illiterates and Moslems are found in this study to be more likely to start giving birth before 15 years of age. The striking results, which we cannot explain, were place of residence; it was found that rural residents have a higher mean


136

age at first birth than Dar es Salaam. Dar es Salaam has the lowest mean age at first birth.

Sexual activities among Tanzanian women start quite early. The mean age at first sexual relationship is 16.0 years with 10 percent of the women having the experience before age 14. By the age of 20, the majority of women (75.4 percent of all respondents) have had the first sexual experience. With the low level of contraceptive practice prevailing in the country (about 12.5 percent), low age at first sexual experience implies early child bearing. The data further reveal that on average the interval between first sexual intercourse and first birth is about 2.4 years in Tanzania. In sub-Sahara Africa, the interval between first sexual intercourse and first birth range from 2.3 years to 2.7 years with the exception of Zimbabwe with 1.2 years. Illiterates and Moslems residing in rural areas are more likely to experience sexual intercourse before age 15 than literate non-Moslem urban women.

Although age at first marriage in Tanzania is relatively low with a mean of 17.4 years, there seems to be an increasing trend in mean age at first marriage across the cohorts. With improvement in education, the younger cohorts stay longer in school and therefore delay marriage. The life table analysis used in this study shows that the interval between first marriage and first birth among women in Tanzania is very low, about 14.5 months on average. This therefore implies that there are no intentions to postpone the first birth after a woman has been married. About 25 percent of the women have the first birth within 9 months of marriage with the premarital births excluded from the analysis. This means illegitimacy conception is rampant in Tanzania as 25 percent of married women had their first pregnancy outside their marriage.

The rising trend in increasing age at first marriage could be attributed to several factors, namely the changing education policies that are geared towards boosting women’s education. In addition, economic changes, which are reflected by rising standards of living, and stimulation of women’s employment, are beginning to change people’s attitudes toward early marriage. However, increasing age at first marriage implies an increase in premarital births, which is now common in Tanzania, (40 percent premarital conception among all first pregnancies which resulted in live births). The marriage act of 1971 specified the age at first marriage to be not less than 15 but there is a high


137

likelihood for an illiterate and Moslem rural resident woman to get married before attaining age 15, an indication that written laws and policies are not followed in Tanzania. Hence a new approach is necessary for the problem of premarital and early births as well as early marriages.

Incidence of primary sterility among Tanzanian women is low as only a small proportion of the women currently married for over 5 years are sterile. This low level of sterility could be attributed to improved nutrition and good medical care. However, the TDHS data could not clearly show the actual situation of sterility, as there were no probing questions on sterility in the questionnaire.

The multiple regression analysis used in the study reveals the existence of differentials in age at first birth associated with the age of a woman, age at first sexual intercourse, level of education, the type of place of residence, and religion (see discussion above). There are almost two years differences in age at first birth between those women with no education and those with secondary and above level of education. Catholics tend to give birth later and Moslems give birth earlier. But surprisingly, women in rural areas seem to have higher age at first birth than their counterparts in the urban areas even after separating Dar es Salaam City residents from urban residents. However, almost 70 percent of sampled women in Dar es Salaam were Moslems as well as almost 40 percent of other urban residents. In rural areas they amount to 27 percent of the total sampled women.

According to the TDHS data, there is a relationship between age at first birth and infant and child mortality. The results indicate that the younger the age of the mother at the birth of the first child, the higher the chances that the child dies due to complications during childbirth. Early age at first birth is always associated with prematurity accompanied by low birth weight that could result in the death of an infant. Maternal mortality and morbidity is also associated with the age of the mother at first birth. Child bearing which comes too early or too late is associated with a lot of complications during pregnancy and childbirth. Some of the complications are fatal and could lead to the death of the mother.

Early age at first birth is a factor associated with high fertility. This is because the earlier the age at first birth, the greater the number of fertile years the woman will


138

spend, and the greater the likelihood of higher total fertility which will have a positive effect on population growth. In the absence of active fertility control, the total number of children women bear throughout their reproductive period is largely a function of the age at which child bearing begins. Therefore, there is an inverse relationship between age at first birth and children ever born. This will be discussed in detail in the next chapter.

Fußnoten:
<13> Unless specified TDHS in this study means the 1996 TDHS.
<14>

In fact, many ethnicities in Tanzania have different standards for men and for women, i.e. tolerating or sometimes even encouraging sexual activity among unmarried men, while restricting and harshly condemning it among unmarried women. In that way, women may be unlikely to acknowledge such behaviour. Therefore, the survey data assessing sexual activity are expected to underestimate the actual levels.

<15>

Out of wedlock birth means those births to mothers who are not married at the time of the first birth.

<16> Wanyambo is one of six ethnic groups found in the Kagera region.
<17> Modern methods include pill, IUD, injections, barriers (diaphragm, foam, jelly), condom, female sterilisation and Norplant. On the other hand, traditional methods include periodic abstinence, withdrawal, Billing method (mucus), calendar method and others.
<18>

Some women tie a special string around their waist. They believe the string to protect them from pregnancy.

<19>

Regulations that require schoolgirls to give up their schooling when they become pregnant encourage the practice of induced abortions in Tanzania. Mostly unmarried young females out of school systems undergo abortions to avoid premarital births, while married women might abort to avoid unwanted children. It is also possible that ever-married women who become pregnant from someone other than their partner or husband are in need of abortion too.

<20> Extracted from Table 67 of the 1996 TDHS report by the Bureau of Statistics, 1997 p. 98.


[Titelseite] [Abkürzungsverzeichnis] [1] [2] [3] [4] [5] [6] [7] [8] [Bibliographie] [Anhang] [Selbständigkeitserklärung]

© Die inhaltliche Zusammenstellung und Aufmachung dieser Publikation sowie die elektronische Verarbeitung sind urheberrechtlich geschützt. Jede Verwertung, die nicht ausdrücklich vom Urheberrechtsgesetz zugelassen ist, bedarf der vorherigen Zustimmung. Das gilt insbesondere für die Vervielfältigung, die Bearbeitung und Einspeicherung und Verarbeitung in elektronische Systeme.

DiDi DTD Version 1.1
a subset from ETD-ML Version 1.1
Zertifizierter Dokumentenserver
der Humboldt-Universität zu Berlin
HTML - Version erstellt am:
Mon Jan 17 19:31:05 2000