Md. Mosfequr Rahman¹, Md. Aminul Hoque²
1. Department of Population Science and Human Resource Development, University of Rajshahi, Rajshahi-6205, Bangladesh
2. Department of Statistics, University of Rajshahi, Rajshahi-6205, Bangladesh

Correspondence to:
Md. Aminul Hoque, Ph.D., Associate Professor,
Department of Statistics,
University of Rajshahi,
Rajshahi-6205, Bangladesh
Mobile: +88-1914254017
Fax: +88-0721-750064-42241 (Ext).
Email: mdaminulh@gmail.com

 

ABSTRACT Adolescent reproductive behavior is a much publicized concern among both the developed and developing nations and recently it has become a major topic of demographic research. Considering its importance, an attempt has been made in this study to investigate the reproductive behavior of married adolescents in some selected areas of Rajshahi district, Bangladesh. The study is based on data collected under the project of UNFPA entitled "Strengthening the Department of Population Science and Human Resource Development". The study indicates that the mean age at first birth for adolescent mothers is 16.34 year and on average, each married women aged 10-19 has 0.65 births. It is also observed from the result of MCA that the respondent's education, husband's education, husband's occupation and place of residence appears as the most important factor determining the mean number of children ever born. In particular, it is found that the mean number of children ever born is higher for adolescents who were from the lower household asset index (1.09) and the rural adolescents (0.95) than the other classifications. Fertility preferences or birth expectation or desire for more children is found to be higher among adolescent mothers. The outputs of the study demonstrate various policy implications that can improve the reproductive behavior of married adolescents. Key Words: Adolescent, reproductive behavior, age at marriage, children ever born, fertility preferences.

Fertility refers to an actual reproductive performance of a woman or group of women, that is, fertility is the frequency of childbearing among the women. Fertility behavior is an important indicator for understanding the trend of population dynamics of any region as well as country. Fertility in Bangladesh is high even by the standards of developing nations. In recent decades, adolescent childbearing has emerged as an issue of increasing concern throughout the developing and the developed world (Jones, 1997; Shaikh, 1997; Islam and Mahmud, 1996). Over the past four decades the developed and developing worlds have been witness to important changes in reproductive behavior among their adult and adolescent populations. Accompanied by a higher level of schooling, better health care, increased urbanization, and greater exposure to modern forms of mass communication, fertility has dropped rapidly in many regions. There is a growing awareness that early childbearing is a health risk for both mother and the child. Also, it usually terminates a girl's educational career, threatening her future economic prospects, earning capacity and overall well-being (United Nations, 1995). However, wide variations in reproductive behavior persist at the national and sub-national levels, and across social groups.

While research and analysis have been conducted on the causes and consequences of such differential behavior among adults, until recently adolescents have received relatively little attention. The factors that influence adolescents to behave similarly, or differently, than their older counterparts remain less well understood.

Adolescent childbearing has significant ramification at the personal, societal and global level. At the personal level, child bearing at an early age can shape and alter the entire future life of an adolescent girl. From the perspective of societies and government, adolescent pregnancy and childbearing have a strong and unwelcome association with low levels of educational achievement for young women, which in turn may have a negative impact on their position in and potential contribution to society (Islam, 1999). Usually, in both developed and developing countries, the rates of population growth are more rapid when women have their first child before they are in their twenties (Senderowitz and Paxman, 1995; Mazur, 1997).

The period of adolescence encompasses the transition from childhood to adulthood during the second decade of life. It is one of the most crucial periods in an individual's life, because during adolescence many key social, economical, biological and demographical events occur that set the stage for adult life.

Although the socio-economic consequences for an adolescent of having a baby will depend on her particular culture, familial and community setting, the physical or health consequences for the mother and her child are more universally recognized as problematic (Buvinic and Kurz, 1998; Acsadi and Johnson-Acsadi, 1986). As adolescent pregnancies occur before a young woman has reached full biological, physical and emotional maturity, they face a number of problems which include anaemia, retardation of foetal growth, premature birth and complications of labor. Pregnancy of a girl who is still growing means an increase in nutritional requirements, not only for growth of the foetus but also for the mother herself (Friedman, 1985). Teenage mothers have a higher incidence of low birth-weight babies, who are associated with birth injuries, serious childhood illness and mental and physical disabilities (Islam et al., 1995). Children born to teenage mothers are also at higher risk of infant and child mortality (Mahmud and Islam, 1999).

The age below which the physical risks of childbearing are considered to be significant varies depending on general health conditions and on access to good prenatal care. In societies where anaemia and malnutrition are common and where access to health care is poor, childbearing of teenagers involves enormous health risks. However, in societies with good nutritional levels and widespread access to high quality prenatal care, the physical risk of having a child during adolescence may not be considered quite so seroius (Makinson, 1985). The severity of the social and personal consequences of adolescent childbearing is also likely to be greater the younger the mother is at the time she gives birth.

Child birth before the age of 20 is more dangerous to mother and infant than it is for older women. In addition to the social and economic consequences, early fertility often jeopardizes the life and health of both the mother and the child. Pregnancy during adolescence poses an increased risk of maternal and infant morbidity and mortality resulting in an increase in cumulative fertility and restricts the opportunity for socioeconomic advancement.

Here we examine the fertility and fertility preference of the adolescents in Bangladesh. On the basis of the related questions, we try to study in brief the reproductive behavior of them. We also try to focus on the young adults aged 20-29 years to have a comparative study of fertility performance of the adolescents.

The data of this study was collected under the project of UNFPA entitled "Strengthening the Department of Population Science and Human Resource Development" of University of Rajshahi. The pattern of data was collected in three main sections namely, fertility, mortality and migration along with socio-economic characteristics of the respondents. These data were collected from three residential areas, which are rural, urban and sub-urban areas of Rajshahi district. We collected information from 6000 ever-married women by interview method from the rural, urban and sub-urban areas. From the total collected information of 6000 ever-married women, we found 426 married adolescents (6.39%). All the information was taken by purposive sampling method. The data of this study was collected in June 2004.

Multiple Classification Analysis (MCA)
In 1934, Yates developed the multiple classification analysis and it was later elaborated on by Anderson and Bancroft in 1952. In 1963, the computerized MCA program was prepared by a group of researchers at the Survey Research Center at the University of Michigan. Since then, the MCA program has been widely used in social science research. It is a technique for examining the interrelationship between several predictor variables and one dependent variable in the context of an additive model.

Unlike simpler forms of other multivariate methods, MCA can handle predictors with no better than nominal measurements and interrelationships of any form among the predictor variables or between a predictor and dependent variable. It is however essential that the dependent variable should be interval-scale variable without extreme or a dichotomous variable with frequencies which are not extremely unequal. Technically, the MCA prediction model can be described as having the overall mean as its constant term and main effects on a series of additive coefficients for the category. The additivity assumption implies that differences according to one predictor are the same for all values of the other predictors included in the model.

There are two effects in MCA: gross/unadjusted effect and net/adjusted effect. The coefficients, which are estimated by solving the normal equation systems, are called the adjusted or net effects of the predictors. These effects measure those of the predictors alone after taking into account the effects of all other predictors. If there is no interrelation among the predictors, the adjusted and unadjusted effects of the predictors will be the same. The unadjusted, eta-square (²) coefficients is a correlation ratio, which explains how well the predictor variable explains the variation in the dependent variables and is usually estimated by solving the normal equation with only one predictor.

This unadjusted coefficient indicates the proportion of variance explained by a single predictor alone. Similarly, the beta-square (ß²) coefficient indicates the proportion of variation explained by the other predictor variables. The beta coefficient is compared to the partial correlation coefficient in multiple regressions. Besides the adjusted and unadjusted effects, there are several computed statistics, which reveal the closeness of the relationship between the predictors and the dependent variable (Yates, 1934). For instance, the statistics R² measures the amount of variation about the mean explained by the predictor variables.

In statistical terms, the MCA model specifies that a coefficient be assigned to each category of each predictor, and that each individual's score on the dependent variable be treated as the sum of the coefficients assigned to categories characterizing that individual, plus the average for all cases, plus an error term.

Y_ij...n = a_i + b_j +..................+ e_ij...n

Y_ij...n = The score on the dependent variable on individual n who falls in category i of predictor A, category j of predictor B, etc
= Grand mean of the dependent variable
a_i = The effect of the membership in the ith category of predictor A
b_j = The effect of the membership in the jth category of predictor B
e_ij...n = Error term for this individuals

1 Age at First Birth
The ages at which women start and stop childbearing are important demographic determinants of fertility. The higher median age at first birth and a lower median age at last birth are indicators of lower fertility. Age at first birth may also affect child spacing by affecting the risk of pregnancy. That is, those having their first birth at young ages when fecundity is likely to be high may experience more rapid fertility than those having their first birth at later ages when fecundity is declining.

Table 1.1 presents the percent distribution of women by age at first birth according to current age. For women age 20 and over, the median age at first birth is presented in the last column of the table. Childbearing begins early in Bangladesh, with the large majority of women becoming mothers before they reach the age 20. The median age at first birth is between 18 and 19. The data shows that the median age at first birth has increased slightly from around 18 for older women to around 19 for women in their 20s. This slight change to later age at first birth is reflected in the smaller proportion of younger women whose first births occurred before age 15.

Comparison with data from other sources confirm that the age at which women in Bangladesh have their first child has increased steadily over time, in line with increases in age at marriage, with the exception of the past few years. For example, in 1975, the median age at first birth among women age 20-24 was 16.8; in 1989, it had risen to 18.0 and by 1996-97, to 18.44 (Huq and Cleland, 1990). The mean age at first birth among adolescent women (age <20) is 16.34 and the mean age at first birth among young adult women (age 20-29) is 18.14.

2 Mean Number of Children Ever Born and Mean Number of Living Children
The number of children a woman has ever born is a cohort of fertility measurement. Because it reflects the past, it provides a somewhat different picture of fertility levels, trends, and differentials than do period measures of fertility such as CBR and the TFR. It is obvious that fertility is directly proportional to current age. That is, for women of higher ages, number of children ever born and number of living children will be high as compared to women of younger ages.

Table 1.2 presents the percentage distribution of adolescent and young adult mothers by number of children ever born. For the age group 10-19, 41.3% of them have no children and 53.5% have one child. The proportion of adolescents decreases as the number of children ever born increases. For young adults, only 6.1% of women have no children 32.2% have one child and 35.6% have two children. The mean number of children ever born to the adolescent women is 0.65. The corresponding figure for young adult and overall married women is 1.61 and 2.35 respectively.

Table 1.3 shows the percentage distribution of adolescent and young adult mothers by number of living children. For the adolescent mothers about 43.7% have no children and 52.1% have one child. The corresponding figure for young adults is 6.6% and 34.3% respectively.

About 35.6% young adults have two children. The average number of living children for adolescents is 0.61 and for young adult is 1.54. For overall women, it is 2.23.

3 Determinants of Children Ever Born: MCA
Table 1.4 presents the mean number of children ever born by selected socio-economic characteristics. The result indicates that the proportions of variance explained by MCA is not very high for adolescent and young adult women (Multiple R²;=0.33 and Multiple R=0.57, for adolescent; Multiple R²;=0.29 and Multiple R=0.54, for young adult). The low value of R²; may be due to some intercorrelations among the predictor variables considered here or there may be some other factors, which may affect the mean number of children ever born. Of all the variables respondent's education, husband's education, husband's occupation, place of residence appears as the important determinants of children ever born. Types of family, household asset index, religion, current working status, bank account and property owned play a relatively less important role on children ever born.

It is often observed that in developing societies that a husband's occupation is closely related with social status. Among the selected factors husband's occupation is the most effective one and shows the strongest association (²;=0.25) with children ever born for adolescent women. The effect of husband's occupational level remains high even after adjusting for the effect of all other predictors in the model (ß²=0.23). Adolescent women whose husbands are laborers (0.91) and other category (1.00) tend to have a higher fertility than the average followed by farmer (0.59), servicemen (0.55) and businessmen (0.51). For young adult women the same pattern is followed.

Among adolescents, while higher levels of education are associated with lower probability of giving birth, the direction of causality is less clear. Findings indicate that educational attainment has another strong association
(²=0.16) with mean number of children ever born. The effect of educational levels remain high even after adjusting for the effect of all other predictors in the model (ß²=0.14). It is important to note that highly educated women have been found to have lower fertility than the illiterates. The mean number of children ever born is 0.71 for adolescent women who are illiterate and 0.57 for highly educated women, that is, adolescent women of 11 or more years of education.

For young adult women respondent's education is found to be the strongest association (²=0.32) with mean number of children ever born and it remains very high after adjusting for all other factors in the model (ß²=0.31). We also see that like adolescent women, higher educated young adult women have lower fertility than the illiterate ones. The mean number of children ever born is 2.00 for illiterate young adult women and 1.23 for higher educated young adult women.

Husband's education seems to be a less effective factor than women's education in explaining the variation in mean number of children ever born among adolescent women (²=0.15; ß²=0.11). For adolescent women of higher education the mean number of children ever born is 0.56 while for illiterate husband's it is 0.73. Though there is virtually very low significant difference in mean number of children ever born by husband's educational level except for unadjusted mean for higher education, husband's education plays another strong association (²=0.249) with mean number of children ever born for young adult women, but again the effect of husband's educational level is very low after adjusting for the effect of all other predictors in the model (ß²=0.071).

Another socioeconomic variable that emerges from the literature as an important influence on fertility behavior is place of residence. Fertility levels are expected to be lower in urban areas than in rural.

We find that place of residence has a strong effect on mean number of children ever born (²=0.25) for adolescent women. Adolescent women who are currently living in the rural area have a higher mean number of children ever born (0.95) than their counterparts in urban (0.54) and sub-urban (0.61) areas. The effect of place of residence remains strong (ß²=0.23) when other socio-economic factors are controlled.

Although young adult women from rural areas have higher fertility than their urban and sub-urban counterparts, place of residence becomes less important (²=0.07; ß²=0.058) when other socio-economic variables are controlled.

Types of family shows a moderate effect on children ever born (²=0.13) for adolescent women, but its effects are low after adjusting for the effect of all other predictors in the model (ß²=0.082). The mean number of children ever born for adolescents of a nuclear family is little advanced than the adolescents of combined or other families. This may be the reason why adolescents of a combined family are more conscious about family planning and get guidance from other older members of the family. For young adult women types of family also shows moderate effect on children ever born (²=0.124 and ß²=0.111). The mean number of children ever born for young adult women of a nuclear family is (1.33) which is less than the young adult women of combined and other families (1.64).

The household asset index shows weak strength in explaining variation in mean number of children ever born
(²=0.087), but the effect of household asset index increases after adjusting for the effect of all other predictors in the model (ß²=0.126). The mean number of children ever born for the lower class adolescent women (1.09) is much higher than the adolescent women of the upper class (0.61). For young adult women household asset index has a moderate effect on children ever born (²=0.116), but its effect becomes low after adjusting for the effect of all other predictors in the model (ß²=0.067). There is virtually no significant difference in mean number of children ever born by household asset index except for unadjusted mean for upper class young adult women.

Among adolescent women's property owned (²=0.008 and ß²=0.013) shows the least effect on children ever born. Adolescent women who have any property of their own have lower fertility than adolescent women who have no property of their own. For young adult women property owned (²=0.021) also appears as a less important predictor of children ever born. This factor also becomes insignificant (ß²=0.013) when adjusted for other predictors considered in the model.

4 Fertility Preference
Information on the fertility preferences provides a measure of the overall attitudes of society towards childbearing and the general course of future fertility. The interpretation of survey data on fertility preferences is often difficult, since it is understood that respondents' reported preferences are, in sense, hypothetical and thus subject to change and rationalization. Still, the utility of information on the desire for children to anticipate changes in actual fertility behavior, has been demonstrated in a wide range of contexts. The fertility preferences among the adolescents and young adults are discussed in detail here.

4.1 Desire for More Children
The desire for more children lends some insight into the process of changing family size norms. Desire of having one, two or more live born children or birth expectation bears a significant value in fertility study and projection. Adding the number of additional children desired to a woman's actual number of living children gives a surrogate measure of prevailing individual family size norms. Family size norms may have a programmatic value since the decision to adopt contraception is likely to be, in part, influenced by individual family size norms.

Table 1.5 shows that the percentage distribution of adolescent and young adult mothers having desire for more children. All currently married women were asked whether or not they want to have additional children and if so, how many more they want to have. About 89.0% of adolescents want to have another child and only a small proportion (11.0%) of them said no more children. On the contrary, the corresponding figures for young adults are 48.9% and 50.1%.

Overall, for all the ever married women the corresponding figures are 69.7% and 30.3% respectively. Thus we see that, adolescent mothers are keener to increase their family size as compared to their older counterparts, which consequently affect fertility to be higher.

To have a clear idea about future fertility preference, we make analysis by controlling the current number of living children, which is shown in Table 1.6. Desire to have more children is closely related to the number of living children. A woman is more likely to desire more children, if she has fewer living children. The proportion desiring more children is 82.8% among adolescent women who had one living child, while it is 79.9% among the young adults. It declined steadily to 33.3% among those adolescents who had three or more living children, while for the young adults it declined to 13.8%.

As expected, the proportion of currently married women who want to stop childbearing rises with the number of living children. Thus it is evident from the findings that, the percentage desiring more children according to the number of living children, is higher among the adolescent mothers than younger adults.

4.2 Opinion About Ideal Family Size
In this study all ever married women were asked the question: "how many children should be contained in a family and how many of them are male and how many of them are female?" Table 1.7 shows the results of this question for the adolescents and young adults. Among them 92.5% adolescents stated that they prefer two children, 5.9% prefer one child and only 1.6% adolescents prefer three and more children. Among young adults 91.6% prefer two children, 3.9% prefer one child and 4.5% prefer three and more children. So, it can be said that both adolescents and young adults are not likely to increase their family size. It may be due to the fact that all of them are more aware of family planning. Overall about 89.2 percent women prefer two children in their life. Again we also see that, 1.4% of adolescent want no male children and 4.5% wants no female children. While among young adults 0.9% want no male children and 3.3% want no female children. So we can say that adolescents prefer less female children than young adults, that is male sex preference among adolescents is higher than young adults.

A strong preference for sons has been found to be pervasive in Indian society, affecting both attitudes and behavior with respect to children (Arnold et. Al., 1998; Arnold, 1996; Basu, 1989).

4.3 Expected Gap of Next Child
In this study, opinions of the ever-married women were sought on the matter of the expected gap of next child. That is how long they think they should wait before having another child. Data are analyzed by controlling current age and age at marriage and the results are presented in Table 1.8.

The table shows that, among women whose current age is less than 20, exactly 25.1% preferred a delay of 3-4 years before the next child, followed by 62.8% with a delay of 5-6 years, 8.0% with a delay of 7 years and over, and only 4.1% with a delay of 0-2 years. While among young adults the corresponding figure is 23.1%, 61.3%, 9.4% and 6.3% respectively. The overall average gap of next child is 4.90 years among the adolescents while 4.85 years for the young adults. Thus it is seen that adolescent mothers prefer a slightly higher gap of next child than their older counterparts.

When we control age at marriage, we see a slightly different situation. About 23.8% women preferred a delay of 3-4 years before the next child followed by 61.3% with a delay of 5-6 years, 5.6% with a delay of 0-2 years and 9.3% within 7 years and more, among women whose age at marriage is less than 20 years. The corresponding figures for women whose age at marriage is 20-29 years is 30.6%, 56.6%, 6.6% and 6.2% respectively. The average gap of next child for women of age at marriage less than 20 years is 4.86 and for women of age at marriage 20-29 years is 4.57.

Finally, it can be said that both adolescents and young adults are found to be more conscious about their birth spacing which may indicate that fertility control has been a common practice among them.

NA: Not applicable

 A: Omitted because less than 50 percent of the women in the age group x to x+4 have had a birth by age x.

Adolescence is usually too young an age to become a parent. The international community and most governments view adolescent childbearing as undesirable because of its negative consequences, and increasingly parents and adolescents themselves share this view (Maina, 1995; Senderowitz, 1995). The International Conference on Population and Development held at Cairo in 1994 also placed importance on reducing the level of childbearing among adolescents (United Nations, 1994). In addition, the socio-economic advancement of teenage mothers, in the areas of educational attainment and accessibility to job opportunities, may be curtailed. According to a recent study, Bangladesh has the highest rate of adolescent childbearing among Asian countries; the country's characteristics in this regard are similar to sub-Saharan African countries (Singh, 1998).

It has been observed from the findings of the study that mean age at first birth among adolescent women are very low, 16.34 years and the corresponding figure for young adult is 18.14. The increase in the average age at first birth among young adult women may be due to the fact that they are found to be more educated having higher age at marriage. Demographically, the early childbearing leads to large complicated families and significantly shorter time periods between generations, with concomitant dramatic increase in population growth rates. Therefore emphasis on delaying first marriage and first birth may be an important element in population control program in Bangladesh.

We also observed from the present study of children ever born that about half (53.5%) of the adolescent mothers have given birth to a child, and 4.5% adolescent mothers have two births. The corresponding figures for young adult mothers are 32.2% and 35.6%. The mean number of children ever born to adolescent and young adult mothers is 0.65 and 1.61 respectively. The mean number of living children for adolescent mothers is 0.61 and the corresponding figure for young adult mothers is 1.54.

Among the selected variables, place of residence, husband's occupation, respondent's education and husband's education appeared to be the most significant determinants of children ever born for adolescent women. For young adult women respondent's education is the most important determinant of children ever born. Other important variables are husband's education and husband's occupation. The main fact is that educated women marry later and have lower fertility within marriage. Educated and engaged in an occupation where education is necessary, husband's wives have found to have fewer children. This may be because educated husbands generally married educated women and husband's decision making power is still strong enough in every section of our society, especially in the family and in having children.

Fertility preferences or desire for more children is higher among the adolescents than their young adult counterparts (80.9% Vs 49.9%). Not surprisingly, the desire for additional children drops progressively as the number of living children increases for both adolescents and young adults. For adolescent women the opinion about mean ideal family size is 1.96 and the corresponding figure for young adults is 2.01, indicating that adolescent women are now more conscious about the negative effect of more children in their family as well as in the country. Sex preferences for males also reduce among both adolescents and young adult married women. Most of the adolescent and young adult women gave their opinion about expected gap of next child as about five years.

The findings of our research hold implications for policy that could be useful in devising ways to solve adolescent's early childbearing and thus bring about a further reduction in fertility in Bangladesh. Appropriate policy and programmatic measures should be undertaken immediately to reduce the incidence of early childbearing that can have negative health, social and economic consequences, including the curtailment of education and job prospects of young mothers. Policy makers and planners should consider and pay more attention to the following points.

Because early marriage and childbearing are associated with less education and lower future income of young mothers, programs that keep girls in school should be promoted. The attainment of higher level of education by young women can be expected to yield a greater use of reproductive health services and better employment prospects.

In order to reduce the rate of early childbearing, adolescents, their parents and community should be made more aware of the negative health, social and economic consequences of early marriage and early childbearing. Such awareness could be created through social mobilization and information, education and communication campaign.

Awareness must be created through the public media as well as through the community leaders so that age at marriage for females does not come below the legal age (i.e. 18 years) and to prevent child marriage and discourage adolescent pregnancies ensuring its execution. Early childbearing should be discouraged by more publicity in the mass media, health centers or community indicating its effects on the delivery related complications at the time of delivery.

Expanded girls' social participation, in schooling and economic opportunities, understanding that these are basic entitlements and they are a framework for adolescents' reproductive behavior.
Recognizing that a large proportion of adolescents are already wives and mothers, who need support and investment at least as much as do their unmarried peers.

There is a need to extend the interval between marriage and first birth, thus delaying the timing of the first birth through the effective use of family planning methods. There is evidence that, in most developing countries, adolescencts face difficulty in obtaining family planning methods owing to a lack of knowledge and limited access to family planning services (Blanc and Way, 1998). This situation suggests the need for a more concerted family planning program and efforts should be focused specifically on newly married adolescent couples to keep the optimized fertility level for Bangladesh.

When family planning programs both private and public were initiated in the 1960's in Bangladesh, older women became the principal beneficiaries, finding modern contraceptive methods more acceptable than the traditional methods. Thus, fertility among older women has fallen more acceptably than fertility among younger women.

Anxiety and depression appear to be the main diagnoses presenting to psychiatrists in private practice. This is an important observation as anxiety and mood disorders can be effectively treated if detected early.

Acsadi, George T.F. and Gwendolyn Johonson-Acsadi 1986. Optimum Conditions for Childbearing (London, International Planned Parenthood Federation).

Anderson, R.L. and T. A. Bancroft 1952. Statistical Theory in Research. New York, McGraw Hill.

Buvinic, Mayra and Kathleen Kurz 1998. "Prospects for young mothers and their children: a review of the evidence on consequences of adolescent childbearing in developing countries". Paper presented at the Committee on Population, National Research Council, National Academy of Sciences Workshop on Adolescent Reproduction in Developing Countries, Washington DC, 24-25 March.

Benjamin R. H. and Mutambi 2004. Inference on the Likelihood of HIV Transmission among Adolescents/Teenagers (13-19 years) and Young Adult (20-29 years) in Pennsylvania. Biannual Update to Epidemiologic Profile of HIV/AIDS in Pennsylvania- June 2004.

Friedman, H.L. 1985. The health of adolescents and youth: a global overview. World Health Statistics Quarterly 38(3):256-266.

Huq, M.N. and J. Cleland 1990. Bangladesh Fertility Survey, 1989, Main report (Dhaka: National Institute of Population Research and Trainning).

Islam, M.M. 1999. Adolescent Childbearing in Bangladesh. Asia-Pacific Population Journal 14(3):73-87.

Islam, M.M. and Mamun Mahmud 1996. "Marriage pattern and some issues related to adolescent marriage in Bangladesh" Asia-Pacific Population Journal 11(3): 27-42.

Islam, M.N. and M. M. Islam and H.K.M. Yousuf 1995. Fertility and Reproductive Health Status of Married Adolescents in Rural Bangladesh (Dhaka, Population Development and Evaluation Unit, Ministry of Planning).

Jones, Gavin W. 1997. Population dynamics and their impact on adolescents in the ESCAP region. Asia-Pacific Population Journal 12(3): 3-30.

Mahmud, Mamun and M. Mazharul Islam 1999. Health status of adolescent mothers and their child: evidence from Bangladesh Fertility Survey 1989. Social Science Review 16(1).

Makinson, Carolyn 1985. The heath consequences of teenage fertility. Family Planning Perspectives 17(3): 132-139.

Mazur, Laurie Ann 1997. High Stakes: The United States, Global Population and Our Common Future (New York, The Rockefeller Foundation).

Senderowitz, Judith 1995. Adolescent health: reassessing the passage of adulthood. World Bank Discussion Papers No. 272, Washington DC, World Bank.

Shaikh, Kashem 1997. Recent changes in marriage patterns in rural Bangladesh. Asia-Pacific Population Journal 12(3): 31-48.

Sign, Susheela 1998. Adolescent childbearing in developing countries: a global review. Studies in Family Planning 29(2): 117-136.

United Nations 1989. Adolescents Reproductive Behavior. Vol. 1. Evidence from Developed Countries, Population Studies, No. 110.

United Nations 1995. The World's Women, 1995: Trends and Statistics (New York, United Nations).
Yates, F. 1934. The Analysis of Variance with Unequal Numbers in the Different Classes. Journal of American Statistical Association, 29:51-66.