Fertility and Contraceptive Use in Bangladesh: the Role of Women's Education and Mass Media

	Md. Abdul Goni, Ph.D Associate Professor Department of Population Science and Human Resource Development University of Rajshahi, Rajshahi -6205 Bangladesh
	magoni_popsrubd@yahoo.com

ABSTRACT In this paper our aim is to examine the impact of education, both formal and informal (via media, GOs and NGOsa) education, which is believed to affect the use of contraceptives and hence, fertility levels. This paper utilized the logistic regression model of Bangladesh Demographic and Health Survey data to investigate the effects of education, and exposure to mass media, on the method of contraception and on fertility. The results show educational attainment (at least secondary) and exposure to mass media substantially increases use of contraception and consequently fertility decline. Women members of GOs/NGOs, urban residence, number of surviving male children, and employment status are other factors associated with increased use of contraception and decline in fertility. Interestingly, the effects of media and GOs/NGOs varied according to the urban-rural residences. While television was found to have a significant effect on fertility and contraceptive use between both the urban and rural women, the effect of radio and GOs/NGOs was insignificant among the rural women. However, exposure to mass media and education (except secondary) appears to have weak significant effects on fertility and strong on contraceptive use. Key Words: Fertility, Contraceptive Use, GOs/NGOs, Mass Media, Formal and Informal Education.

INTRODUCTION

As in many other developing countries, Bangladesh is a country that has achieved a substantial reduction in fertility despite little improvement in levels of material well-being, education, women's status, child survival and other factors frequently associated with demographic transition. The family planning program is credited with being the main driving force behind this reduction, while the role of social and economic change is de-emphasized (Larson and Mitra 1992; Cleland et. al. 1994). Bangladesh is internationally considered a success story in family planning (Freedman, 1995), with an increase in contraceptive prevalence rising from 8 percent to 54 percent and a decline in the total fertility rate from 6.3 to 3.3 in the three decades since independence (Mitra, et. al., 2001). Success in meeting these population goals were advanced with the help of concerted efforts of the government in conjunction with NGOs and donor organizations who ensured that free or affordable contraceptives were available in both public and private facilities throughout the country. Since the late 1980s there has been a large increase in the number of couples using family planning methods. Unfortunately, the use of family planning declined to 50 percent in 2001 BMMS. The decline in overall use is due to a decline in the use of traditional1 methods (from 10 to 6 percent). Use of modern methods has little changed since 1999-2000.

Social indicators, such as employment and electricity, have allowed women to gain in the development process. There has been an increase in the number of females in the workforce, nationally as well as in rural areas. The garment industry absorbed a good majority of female labor. Electricity also generates employment. Its impact on employment is both direct and indirect. The percentage of households with an electrical connection increased from 68 percent in 1991 to 77 percent in 1994 in urban areas and it doubled from 7 percent to 14 per cent respectively in rural areas (Khuda and others, 2000). Women in the electrified households are involved more in household level income-generation activities and depict better re-allocation of time for remunerative employment. The unemployment rate is relatively low in the electrified households; and a relatively higher share of non-agricultural employment in the electrified households indicates a modernization effect of electricity on occupation. It is interesting to note that the overall literacy rate for both males and females in the electrified households is higher, especially due to the household's access to electricity which has contributed much both in economic terms as well as in raising awareness about value of education (Barkat, et. al. 2002). This strongly suggests that access to radio and TV enhances literacy in an informal way. Also NGOs have encouraged women's participation in activities outside the formal sector. In Bangladesh most of the NGOs working at the community level have tried to increase opportunities for women through informal education (see appendix C.1) and micro-credit programs. There are so many NGOs currently working on health and family planning, as well as, on other problem areas in different regions of Bangladesh, that their activities may have encouraged people to adopt contraception in the late 1980s, precipitating a change in reproductive behavior (Goni, 2008).

There exist demographic and socio-economic variations across different regions in Bangladesh. Various studies have confirmed that during the past two decades women's status, in terms of education, employment, mobility and decision-making power, has undergone major changes. Also, there is evidence that such changes have contributed to increased contraceptive use and consequently fertility decline (Khuda and others, 1990; Khuda and Barkat, 1992). Additional changes resulting from increased access to mass media have fostered modern outlooks and attitudes, thereby lowering high-fertility norms, even among the poor. Radio ownership has increased from 25 to 31 percent. Access to television from 7 to18 percent from 1993-94 to 2000 in the rural areas; moreover, there is some access to cable networks (Mitra, et. al., 2001). The role of mass media, especially radio, is a contributing factor in popularizing the family planning movement in many of the developing countries. Villages in Bangladesh are less isolated today, having been linked to the outside world by the mass media (Cleland et. al. 1994). However, recent statistics suggest that, despite a continuing increase in contraceptive use, fertility decline in Bangladesh has stalled: Three successive Demographic and Health Surveys show that the total fertility rates were 3.4, 3.3, and 3.3 in 1991-93, 1994-96, and 1997-99, respectively and that contraceptive use has little improved (Mitra, et. al., 2001). It is believed that education and media, especially radio and television, has been the most instrumental vehicle in building awareness and spreading the knowledge of use of contraception and small family norms. Utilizing the Bangladesh Demographic and Health Survey (BDHS) data, the impact of education and the role of Radio and TV as an informal medium of education in reducing fertility and increasing contraceptive use will be examined in this chapter. We will also examine whether the effects of education and media vary among married women between urban and rural areas. This paper also tries to determine the other important factors that explain fertility behavior and contraceptive use.

The paper is organized as follows. In section 2 we discuss objectives, data and methodology; section 3 presents findings and discusses trends in fertility and current use of contraception, women's education and fertility, women's education level by some selected variables, determinants of fertility and contraceptive use; summary and conclusions are presented in section 4.

OBJECT, DATA AND METHODOLOGY

The major objective of this paper is to examine the impact of education, both formal (as captured by educational attainment) and informal (as captured by media, GOs and NGOs), on the use of contraceptives and fertility in the national and urban-rural levels. Also this paper looks at the trends in fertility levels and contraceptive use in Bangladesh from 1975 to 2000. National level data are used for the 1975-2000 period. For the multivariate analysis of the determinants of contraceptive use and fertility decline, the 1999/2000 BDHS data are used. The 1999/2000 BDHS employed a nationally representative, two-stage cluster sample design. In the 1999/2000 BDHS, a total of 10,544 eligible women (ever-married and aged 10-49 years) were interviewed. In this study, we consider only currently married women of reproductive age: 9,450, from the 1999/2000 BDHS. Among these women, 6,630 lived in rural areas, and 2,820 lived in urban areas.
Two dependent variables are used: (a) current use of contraception (yes =1; no=0), and (b) fertility preference (prefer another child =1; otherwise =0). The interesting outcome of this analysis is "fertility preference" which indicates women's attitudes toward their future fertility. Although these attitudes may not perfectly presage actual future behavior, they can demonstrate prevailing social norms about family size. Because one of the goals of the National Family Planning Program is to promote a small family norm, these preferences may provide answers about the degree to which the program is accomplishing its objectives. On a micro level, fertility preferences also reflect the degree to which women and their husbands feel that they exercise control over their reproduction. For the contraceptive use variable - current use of modern family planning methods and non-use of them as the proportion of women reported that they were using family planning methods at the time of interview. Modern methods are much preferred (43 percent of married women) over traditional methods (10 percent). Logistic regression will be used for each dichotomous dependent variable. The independent variables included are the woman's age, number of surviving children, currently pregnant women (according to the respondent reports), exposure to mass media, women's education, member of GO/NGOs, discussed about family planning method with husband, can go shopping center after marriage, religion, employment status, electricity connection, residence and geographic division. Age, age-square, and the number of surviving children are treated as continuous variables. Table 5 below shows descriptive statistics of the independent variables used in this paper.

RESULTS

1 Trends in Fertility and Current use of Contraceptive Method
The time series estimates of TFR 1 over the last 25 years beginning with the 1975 BFS indicates a decline of 48 percent in TFR, or that of 1.9 percent per year. This can mainly be accounted for by rising contraceptive use, effective immunization and reduction of child mortality. The TFR declined dramatically from 6.3 children per women in 1971-1975 to 3.3 in 1997-1999 (Figure 1). During the same period, the contraceptive prevalence rate (CPR) increased more than six-fold, from 8 per cent to 54 per cent (Figure 2). With the help of the concerted efforts of the government with NGOs in the field, the national family planning program achieved a remarkable success in a short period of time, attaining a current contraceptive prevalence rate of 54 percent in 1999-2000. The relative share of modern methods like pill and injection increased; however, the relative share of long-acting clinical methods declined. There are regional variations in contraceptive use, with Rajshahi (59%) and Khulna divisions (64%) having the highest prevalence rate and Chittagong (44%) and Sylhet (34%) divisions the lowest. TFR is higher in rural (3.54) than in urban areas (2.45) and contraceptive prevalence rate is 52% and 60% respectively. Except the age group 15-19, fertility declined substantially in all other groups, especially among women aged 35 and older (Mitra, et.al. 2001). The pace of fertility decline slowed down recently compared to the rapid decline in the late 1980s and the early 1990s. Since then it remained almost constant (Figure 1).

Figure 1: Trend in Total fertility Rates in Bangladesh, 1971-1999

Source: Bangladesh fertility survey 1975 and 1989, Contraceptive prevalence survey 1991. Bangladesh demographic and health survey 1993-94, 1996-97 and 1999-2000.

Notes: TFR 1: Total fertility rate, expressed per women. Rates are for the period 1-36 months preceding the survey. Rates for age group 45-49 may be slightly biased due to truncation.

Figure 2: Trends in current use of family planning methods among currently married women aged 10-49 years in Bangladesh.

Sources: Same as Figure 1.

2 Women's Education and Fertility
Women's education has long been recognized as a crucial factor influencing women's childbearing patterns; extensive demographic literature is devoted to examine the role of female education in promoting sustained fertility decline. There are many reasons why we would expect increased education having an impact on fertility, and many researchers have documented the close relationship between education and fertility decline. The argument made by Lucas (2003) emphasizes that the increased return in investment in education leads to fewer children. Caldwell (1982) focuses on the increased costs of raising a child and diminished expectations of the lifetime return to parents from that child - both consequences of schooling - as setting off a fertility transition.
Some other researchers have focused on the relationship between women's education and fertility decline, arguing that a woman's education reduces her desired family size, changes the relationship between her desired number of children and planned number of births, and improves her ability to achieve her desired family size (Murthi, et al. 1995). Declines in fertility and infant mortality move hand in hand, and women's education may also have an indirect effect on fertility through the role it plays in reducing infant mortality. Better-educated women are more likely to know about hygiene and nutrition, and are more likely to act on this knowledge (Caldwell, 1986). The inverse relationship between female education and fertility is often pointed out, which is usually explained by the fact that higher education leads to greater use of contraception and consequently to lower fertility. It is suggested that education may enhance women's five levels of autonomy-knowledge autonomy, i.e. decision-making autonomy, physical autonomy, economic autonomy and social autonomy that eventually affect various reproductive health behaviors (Jejeebhoy, 1995).

Notes: a The probability of dying before the first birthday. b According to the health card or mother's report. c All other women work for kind or both cash or kind. Primary education means grade 5 and secondary and higher means-grade 10 and more.
Source: Bangladesh Demographic and Health surveys 2000.

Women's education reduces desired family size by raising desired living standards and provides a greater range of general information, better understanding of the reproductive process, and access to modern and effective means of birth control. Educated women tend to start using contraception immediately after marriage or first birth, but the less educated use contraceptives only to prevent higher-parity births (Chaudhury, 1978). Thus the better educated are more likely to
(i) marry later, use contraceptives, desire a lower number of children and raise healthier children,
(ii) make better decisions for themselves and their children, and
(iii) make greater economic contributions to the household.

According to Table1, the age at first marriage is 14 with no education, compared with 20 for those who finished at least secondary education. Among uneducated women, the proportion of those who never heard or saw a family planning message on radio or TV is 75 percent and 49 percent do not use contraception; the corresponding figure for women who completed at least secondary education is 30 and 41 percent respectively. It is evident that female education can reduce economic dependence on other family members and enhance freedom of movement outside home. 64 percent of non-educated women are employed earning cash, while it is 89 percent for those who completed at least their secondary education. Better-educated women have more freedom of movement. 26 percent of the non-educated go alone to hospital for herself or for her children; compared with 33 percent for those who completed at least secondary education.

The Bangladesh Demographic and Health Survey (1999-2000) provide interesting information about actual and desired fertility rates and current contraceptive use, access to media and educational attainment of women. As expected, the educational attainment of women is strongly associated with fertility. Looking at the relationships between women's education and TFR (Table 2), it appears that TFR (2.3) of women with at least secondary education completed is substantially smaller than that of the uneducated (4.1). TFR decreases with rising levels of education. Table 2 also presents data on the same women's total wanted fertility rate (TWFR). All educational groups of women, except those with no education, want to have around replacement-level fertility (2.1) or lower. Women with primary school-level education report that they want to have 2.2 children, whereas those with at least secondary education want to have only 1.8 children. TWFR for women with no education is above replacement level. For all educational groups, TFR exceeds the TWFR. The difference is greatest for the group with no education -TFR exceeds TWFR by 1.3 children - and decreases with education to 0.5 children for the highest-level group. Such differences between actual and desired fertility suggest that by enabling women to avoid unwanted fertility through better family planning services by both government and non-government agencies, it would lead to a fertility rate around replacement level for all women except those with no education.

A quite similar picture emerges when one compares the use of contraceptives by education levels of women; only those who have completed at least secondary education are more likely to report a higher use of both modern and traditional methods than other women. It is worth mentioning here that TFR and TWFR are lower among those with secondary education. Generally contraceptive practice increases sharply with education (BDHS 2000). Those indicators in Table 1 and Table 2 suggest that educated women not only have different fertility goals, but also have their aspiration focused on reality.

Notes: 1 see figure 1; 2 Rates are based on births to women aged 15-49 in the period 1-36 months preceding the survey.
Sources: Same as Table 1.

But here it is hardly recognized that the successive two BDHS 1996-1997 and 1999-2000 (Mitra, et.al, 2001) suggested that contraceptive use has increased among women with little or no education (5%), while this figure is only 3% for those who had at least secondary education. This is really surprising. Access to media may be a factor explaining this progress among the illiterate; especially listening to the radio is more common among rural than urban women. It is that urban women are more likely to have access to radio and television than rural counter parts; and the urban-rural difference is much greater with TV watching (see Table 5). Since the 1996-1997 BDHS survey, the ownership of radios in urban areas decreased from 52 to 45 percent, while there is little or no change in rural areas at 29 percent (Mitra et. al., 2001). Listening to radio and watching TV are positively associated with educational attainment; both increase with increasing educational level of respondents.

If mass media is related to fertility preference and contraceptive use, we have Tables 3 and 4. Table 3 shows that of those who listen to radio and watch TV, 67.8 and 70.3 percent respectively do not want an additional child. Similarly 57.8 percent of those listening to radio use contraceptive methods. In the case of TV watching, 62.2 percent are users of contraceptives. Therefore, Tables 3 and 4 suggest that women listening to radio and watching TV are more likely to use contraceptives and less likely to want additional children.

Notes: Percentage of the women within brackets.

Notes: Percentage of the women within brackets.

Further investigation into such relationships with some other socio-demographic variables taken into account will be conducted in the next section. Descriptive statistics of all the independent variables used are presented in Table 5. The average age of women in the sample is around 29 years. The number of surviving male and female children was 1.24 and 1.21 respectively. The proportion educated is 29 percent at the primary level, 24 percent at the secondary level and only 6 percent at the higher level. In the case of employment only 19 percent was employed. What is suggestive in the Table is that in percentage terms mass media is more important than formal schooling: ownership of radio (32 percent) as well as access to radio listening (34 percent) is greater than the proportion of those educated at primary school; and watching TV (24 percent) is at the same level as for secondary schooling although ownership is much less frequent in this case (10 percent). It is worth noting that those actually listening to radio are a little more numerous than those who own radio sets, while the difference is much larger for TV. This suggests that mass media can serve as an informal medium of education for the un- or less educated in rural, disadvantaged situations. Moreover, membership of GO/NGOs is 24 percent. 39 percent and 29 percent of women are respectively able to discuss family planning methods with their husband and to go to a shopping center after marriage. 36 percent are reported to have electricity connection in their house. In the sample, 13 percent of women are non-Muslim, 39 were in Rajshahi-Khulna divisions and 28 percent in Chittagong-Sylhet divisions.
National averages like these often mask some important disparities between urban and rural societies. There are some noticeable urban-rural differences. Urban women are better educated than rural women. While access to radio does not greatly vary between urban and rural areas, access to TV is substantially better in urban than in rural areas. However, it is interesting to note that the opposite is true for membership of GO/NGOs: rural women are much more likely to be a member of one of such organizations than urban women.

Given the descriptive statistics of the independent/explanatory variables above, which are likely to affect women's fertility decision (whether they want another child or not) and contraceptive use among the married women, logistic regression analysis may be conducted in the next section.

3 Determinants of Fertility and Family Planning Method
When conducting logistic regression analysis, it is important to pay due attention to urban-rural differences in the relationships between the dependent and independent variables. Thus three runs are tried for the set of dependent variables: first for the whole sample of married women, the second for the urban and the third for the rural sub-samples. Tables 4 and 5 show the logistic regression estimates of odds ratios for the effects of selected background demographic and socio-economic characteristics of married women of reproductive age on fertility and on current use of family planning methods during the reference period respectively. The results are in the expected direction for most variables, so odds ratios are shown in place of regression coefficients for the easy interpretation of results.

The results in Tables 6 and 7 show, as expected, that women who do not want an additional child, or who use contraceptive methods, increase significantly with age. This effect is statistically significant for both sub-samples (but age squared does not prove to be significant in either case). Also apparent is that currently pregnant women are likely not to want more children, and that women with male children are less likely to want any more children, which is a reflection of deep-rooted son preference in Bangladeshi society, both urban and rural.

Note: "r " reference category. '*' Significant at p < 0.001; '**' Significant at p <0.01; '***' Significant at p < 0.05 and '****' Significant at p < 0.10; Others = Dhaka and Barisal

Note: "r " reference category. '*' Significant at p < 0.001; '**' Significant at p <0.01; '***' Significant at p < 0.05 and '****' Significant at p < 0.10; Others = Dhaka and Barisal.

Table 6 shows that demand for any more children declines among the educated than the un-educated, working women than housewives, TV watching women than non-watching women, and members of GO/NGOs than non-members. Two indicators of women's "autonomy" have a similar impact on fertility preference. The probability for women who can discuss family planning methods (FPM) with their husband to want an additional child is longer than otherwise, so is that for those who can go to a shopping center alone. All these effects are estimates made by controlling for electricity connection (for a detailed study on rural electrification, see appendix C.2), religion, urban/rural residence and administrative division. Table 7 tells us a similar story with respect to contraceptive use. The probability for women to use contraceptives rises with education. The educational level of the women displayed significant positive relationships with women's odds of utilizing a family planning method. Relative to women with no education, women with primary, secondary and higher education show significantly greater odds of utilizing family planning methods. Again, women who have access to TV show a higher probability of using contraceptive methods.

Moreover, a close look at the results in columns 2 and 3 (Tables 6 and 7) reveals interesting urban-rural differences, especially in relation to education, mass media and GO/NGO membership. In urban areas, both formal education (secondary and higher education) and TV watching show significant effects on fertility and contraceptive use, while radio listening and GO/NGO membership do not. In rural areas, on the other hand, formal education (primary level upwards, but except higher education on fertility), mass media (both radio listening and TV watching) and GO/NGO memberships exhibit significant effects on fertility and contraceptive use. Given low but positive correlation coefficients between primary schooling, on the one hand, and radio listening and NGO membership, on the other (but not with TV watching; see appendix table C.3), perhaps, primary education in rural areas has a complementary effect with these factors, suggesting that informal education (via mass media and NGOs) has a significant effect on fertility and contraceptive use especially among the rural less educated women. As stated in appendix C.1, some large NGOs run a variety of informal education programs providing education to members' children who are not able to avail themselves of State-run education.
One may argue that in a developing country like Bangladesh TV is a luxury and radio an everyday luxury, so that the coefficients of these variables tend to absorb an income or wealth effect. Since, fortunately, the data allow us to separate radio listening from its ownership and TV watching from its ownership, we can check if there are strong correlations between the two. And it turns out that there is no multicollinearity problem for both radio and TV. Thus, both are included in the regression equations reported in Tables 6 and 7; and the Tables indicate that the effect of radio listening in rural areas is statistically significant even when the influence of ownership as well as electricity connection is controlled for. The same can be said for TV watching in urban areas. We also checked the results by excluding electricity from the model. But the results (not shown) were almost the same with slightly different sizes of coefficient of the TV ownership dummy. Indeed, in rural Bangladesh most people listen to radio and watch TV in their neighbour's house, typically village leader's or ward commissioner's house. There is a study by Kincaid, et al (2000) on health and family planning in relation to a drama called "Shabuj Sathi'2 of Bangladesh television. They find that currently married women (4,566) who watched the drama were 18.9 percent. 57.9 percent of them did not have their own television set, and the majority of them watched it at a neighbours' house. 52.8 percent of those who watched the drama are found to have used a modern contraceptive method compared with 38.4 percent who did not watch the drama. On radio and TV advertisements of family planning products are frequently broadcasted, with a slogan like "small family is happy family" (Soto Paribar Sukhi Paribar) in a gap of the main program, which must influence women to adopt family planning services. This kind of program can increase awareness among women and encourage them to adopt a small family norm in an informal way. Thus, the impact of mass media as a medium of informal education is independent of the income/wealth effect.

SUMMARY AND CONCLUSION

Bangladesh is one of the best examples of a developing country with a strong family planning program effort, which brought about a significant fertility decline even when social and economic development was sluggish. However, this study has highlighted the significance of both formal and informal education as factors influencing contraceptive use and hence, fertility decline. Three logistic regression models have been conducted. The results of the analysis shows that women's educational attainment and exposure to mass media are indeed significantly identified as contributing factors to the spread of contraceptive use and fertility decline. Relatively speaking for rural, less educated women, it seems that exposure to mass media is somewhat more important than educational attainment. Among other factors, women's membership of GO/NGOs and women's decision making position in the household (discussion of family planning issues with husband and ability to go to shopping center alone) are also noteworthy.
Therefore, this study indicates that improvements in both formal and informal education, together with family planning services should receive priorities in policies for further reduction in fertility. Family planning services are measures to help women avoid unintended pregnancies and the abortions that sometimes follow them (Rahman and others, 2001), so are formal schooling and informal education via mass media. We find that there is a considerable amount of fertility that is in excess of desired fertility. Excess fertility is higher among women with no or little education. Informal education can play a crucial, complementary role, especially among the women with no or little education, in reducing the gap between desired and actual fertility because contraceptive use tends to increase among women with little or no formal education. GO/NGOs can also play an important role in providing opportunities for women to gain knowledge, confidence and skills to find better employment and also to promote reproductive health, including contraceptive methods. Working women are more likely to use contraceptives and have fewer children compared to non-working women (CPS 1991), so GO/NGO activities should be further encouraged across villages.

In Bangladesh, as noted earlier, the pace of fertility decline slowed down recently. Since the early 1990s TFR remained almost constant. Therefore, policy makers should carefully design family planning strategies by paying more attention to the roles that formal and informal education as well as GO/NGOs can play in order to reduce the gap between actual and desired numbers of children among less privileged families.

-------------------------

Notes: a Government Organizations (GOs), Bangladesh Rural development Board (BRDB) and Non-Government Organizations (NGOs) like Bangladesh Rural Advancement Committee (BRAC), Grameen Bank (GB), etc. their common activities are micro-credit and all other activities (such as training, awareness building, education, health) are surrounded by micro-credit activities.

¹ Periodic abstinence, withdrawal etc. are traditional methods. Also see Alan, et al. (1997).
² Kincaid, et. al (2000) studied on Shabuj Sathi (which in Bangla means "ever-green friend" ) on television drama. The objective of the drama was to increase health related knowledge and healthy behavior of men and women. The main health topics featured in the drama are family planning, safe motherhood, childhood diseases and immunization, HIV/AIDS and nutrition. They collected data through face-to-face interviews of 10,400 men and women aged 15-49 and the analysis of the TV drama was conducted on the sub sample of all 4,566 currently married women. Overall health knowledge was significantly related to watching the drama, having the strongest relationship with knowledge of HIV/AIDS, followed by knowledge of nutrition and knowledge of childhood diseases. Married women who saw the drama were more likely to visit a health or family planning service facility than women who did not, and more likely to use a modern contraceptive method.

Acknowledgement: I would acknowledge the contribution of Prof. Osamu Saito for his help and advice. I benefited from discussion at Professor Saito's seminar in the Institute of Economic Research (IER), Hitotsubashi University. I am grateful to Professor Osamu Saito for reading carefully this paper and making his valuable comments that improved essentially the content of the paper.

REFERENCES

Barkat, A., Khan. S. H., Rahman, M., Zaman, S., 2002, Economic and Social Impact Evaluation study of the Rural Electrification Program in Bangladesh. Human Development Research Centre, Dhaka, NRECA International Ltd.

Barkat, A. and Murtaza, M., M., 2003, Adolescent and Youth Reproductive Health In Bangladesh. Status, Issues, Policies, and Programs, Dhaka, Bangladesh.

Caldwell, J. C., 1982, Theory of Fertility Decline. London: Academic Press.

Caldwell, J. C., 1986, "Routes to Low Mortality in Poor Countries." Population and Development Review 12: 171-220.
Cleland, J., J. F. Philips, S. Amin and G. M. Kamal., 1994, The Determinants of reproductive Change in Bangladesh. Washington DC: World Bank.
Chaudhury, H. R., 1978, Female Status and Fertility Behaviour in a Metropolitan Urban Area of Bangladesh, Population Stodies, Vol. 32, No. 2, 262-273.

Coombs, P. H. and Ahmed, M. (1974), Attacking Rural Poverty. How non-formal education can help, Baltimore: John Hopkins University Press.

Chaudhary, S. S. (1992), "Television in Distance Education: The Indian Scene." In Indian Journal of Open Learning 1(1): 23-31.

Erica, F., and Ambrus, A.,(2006), Early Marriage and Female Schooling in Bangladesh; working Paper, Harvard University.

Freedman, R. (1995), Asia's recent fertility decline and prospects for future demographic change. Asia-Pacific Population Research Reports No. 1. Hawaii: East-West Center.

Goni, M. A. and Saito, O. (2008), Fertility and Women's Employment in Bangladesh: A Time Series Analysis.(In press). Journal of Demography India, Vol. 37(2).

Jejeebhoy, S. J. (1995), Women's education, autonomy and reproductive behavior: experience from developing countries. Oxford: Clarendon Press.

Jeffs, T. and Smith, M. K. (1996), Informal Education. Conversation, democracy and learning, Ticknall: Education Now

Khuda, B., S. Islam, R. Sultana, Sirajee, S. A. and Laila, R. A. ( 1990), Women's Savings Groups and Contraceptive Behaviour among Rural Bangladeshi Women. Dhaka University Research Corporation, Bangladesh

Khuda, B. and Barkat, A. (1992), Female Education Scholarship Program - An Impact Study. Report prepared for the World Bank, Dhaka. Dhaka University Research Corporation, Bangladesh.

Khuda, B., Nikhil, C. R. and Rahman, M. M. D. (2000), Family planning and Fertility in Bangladesh. Asia-Pacific Population Journal, Vol. 15, No.

Kincaid, L.D., Whitney, E. E. and Shajahan, M. (2000), Impact of the Shabuj Sathi Television Drama of Bangladesh: Key findings, Center for Communication Programs. Johns Hopkins University.

Larson, Ann and Mitra, S. N. (1992), Family Planning in Bangladesh: An Unlikely Success Story, International Family Planning Perspectives 18(4), 123-144.

Lucas, R. E. (2003), "The Industrial Revolution Past and Future." The Region. Federal Reserve Bank of Minneapolis. Online. Available: http://www.minneapolisfed.org/pubs/region/04-05/essay.cfm. 2003.
Mitra and Associates (1991), Bangladesh Contraceptive Prevalence Survey (CPS), Final Reports, Dhaka, Bangladesh.

___________ (2003), Bangladesh Maternal Health Services and Maternal Mortality Survey 2001. NIPORT, Dhaka, Bangladesh, ORC Macro, Calverton and Johns Hopkins University, Baltimore, Maryland, USA.

Mitra, S. N., Al-Sabir, A., Tulsi, S., and Kumar, S. (2001), Bangladesh Demographic and Health Survey, 1999-2000. Calverton, Maryland and Dhaka, Bangladesh: National Institute of Population Research and Training (NIPORT), Mitra and Associates, and Macro International Inc.

Mitra, S. N., Charles, L., and Islam, S. (1993), Bangladesh Contraceptive prevalence Survey-1991: Final report. Dhaka: Mitra and Associates.

Martin, T. C. and Juarez, F. (1995), The Impact of Women's Education on Fertility In Latin America: Searching for Explanations, International Family Planning Perspective, Vol. 21, No. 2.

Murthi, M., A. C. Guio and Dre`ze, J. (1995), "Mortality, fertility and gender biase in India: A district level analysis," Population and Development Review 21: 745-782.

Nicholas P. (2001), Mass Media Promotion of Family Planning and the Use of Modern Contraception in Ghana. Paper presented to the 24-th IUSSP general conference at Salvador, Bahia, Brazil, 18th - 24th August.

Rahman, M., Julie, D. and Razzaque, A. (2001), Do better family planning services reduce abortion in Bangladesh. The Lancet (London), vol. 358, no. 9287, pp. 1051-56.

Appendix C.1
Most of the NGOs are involved in a variety of activities: vocational training for skill development, micro-credit, adolescent family life education (AFLE, see below), sex education, reproductive health services, and legal assistance in cases of violence and abuse against women and implemented both school and out of school (Barkat and Murtaza, 2003). The Government has recognized this contribution of NGOs and there is significant collaboration between them.

Adolescent Family Life Education (AFLE) Curriculum
Adolescence: The period of adolescence; physical and mental changes during adolescence of boys and girls; importance of the adolescent period.
Reproduction and menstruation: Reproductive health; male and female reproductive organs; process of ovulation and menstruation; process of fertilization; menstrual hygiene; nutrition during menstruation.
Marriage and pregnancy: Age of marriage; age of child bearing; danger of early marriage; normal pregnancy; antenatal, natal, and postnatal care; signs of complications during pregnancy and delivery.
STIs and HIV/AIDS: Common Sexually Transmitted Infections (STIs, including a discussion of personal hygiene); common STIs; signs and symptoms of STIs; risks and transmission; complications of STIs; prevention of STIs.
Family planning and birth control: Why family planning is needed; types of contraceptives; advantages and disadvantages of contraceptives; how to use contraceptives; condoms and their advantages.
Smoking/substance abuse: Smoking-related illness; reasons for substance abuse; signs and symptoms of substance abuse; health hazards resulting from substance abuse.
Gender issues: Inequality between males and females; respect between sexes; roles of males and females in reproduction.
It is difficult to measure the impact of these kinds of adolescent reproductive health information and/or services NGOs are providing at the community level. They have been responsible in great part for increasing opportunities for women through non-formal and formal education, through delivery of reproductive health services (often on the doorstep) and by providing credit programs for women. According to their stated objectives, the majority of NGOs target the poor and disadvantaged.

Appendix C.2: Electricity and Fertility Decline:
The social and economic impact of rural electrification program in Bangladesh is examined in Barkat et al. (2002). Their study is based on two villages with 14,000 respondents and 2,491 households, of which 1,380 are electrified and 1,111 non-electrified. They show its impact on economic aspects including income, employment, and poverty reduction; social and cultural aspects including education, health and family planning; demographic aspects including birth and death rates. Their main findings are:

1. The overall literacy rates and the quality of education (measured by expenditure of education, dropout rate and time spent for study at night by student) are much improved in the electrified households than in non-electrified households. This quality improvement in the electrified households works through many channels: more time available for study after sunset, the quality of that time due to sufficient light and fan for comfort, strengthening the knowledge-base due to access to TV (which in turn increases the appetite for learning), parents (especially mothers/other elder female members) devote more time in assisting children's education compared to before electricity etc. In terms of knowledge about the crucial public health issues, respondents in the electrified households are reported to become more aware than those in the non-electrified households. 56% of them consider TV as the main source of knowledge (the corresponding figure for TV is 28% in the non-electrified households in electrified villages, and 17% in the non-electrified villages). TV played an immense role as the major source of such enhanced knowledge on health issues. The authors also note that income and employment opportunities are higher than non-electrified households due to the fact that in the electrified households women can spend more time for their household work such as sowing, harvesting, drying, etc. after returning from outside activities.

2. All those factors are related to fertility transition. The reported mean number of children ever born to women is 4.3 in both electrified households and households in the non-electrified villages. However, the mean number of deaths is relatively lower in the electrified households (50 ever-reported deaths per 100 households) than in the non-electrified villages (62 deaths) and in the non-electrified households in the electrified villages (59 deaths). In the electrified households, not only the mean number of the ever-died is relatively low, but also both the incidences of death and severity of death (measured in terms of death of 3 or more members) are less pronounced. As a result, the demographic consequences are better: the proportion of those still surviving to the ever born is higher in the electrified households (88.4%) than in the households of non-electrified villages (85.8%). The average household size of the electrified households is slightly higher (6 persons per household) than in the non-electrified households (5.4 and 5.7 respectively for those in the electrified and non-electrified villages). It is likely that this slightly high average household size of the electrified household is due to less poverty-induced out-migration of family members, higher incidence of joint-family structure, and more job opportunities in the electrified areas. The dependency ratio is lowest (0.64) in the electrified households, highest (0.73) in the non-electrified households of electrified villages and in-between (0.68) in the households of non-electrified villages. Thus, compared to the non-electrified, the same number of active population supports a smaller number of dependent populations in the electrified households. Estimates show that availability of electricity in the household contributes to 15.7% of the reduction in TFR (based on comparison of TFRs in two extreme samples), but the contribution of the availability of electricity in the village but not in the household is only 2%. The TFR of the poor in the electrified (2.7) was 26% less than that of the poor in the non-electrified villages, and it was even 7.5% less than that of the rich in the non-electrified villages (2.9). Thus, electricity not only contributes to overall TFR decline, but also to the significant reduction in TFR among the poor. Undoubtedly the refinement of these estimates are needed with respect to the entire demographics of population and reproductive health programs. It should also be noted that in addition to electricity, there exist many other determinants of fertility (other than the family planning program per se) such as income, employment, education and age at marriage.

Appendix Table C.3: Pearson Correlation Coefficient between education level and NGOs members and mass media

"**" Correlation is significant at the 0.01 level, "*" Correlation is significant at the 0.05 level, (2-tailed)