The present study was conducted among the ever-married women in their reproductive age group (15-49) years in Bangladesh. The required data and necessary information were taken from Bangladesh Demographic and Health Survey (BDHS), 2004. It was the fourth survey in a series of national-level population and health survey utilized a multistage cluster sample, based on the Bangladesh Census, 2001.

Chi-square test was used to find the association between ages at marriage with background factors of Bangladeshi females. Logistic regression analysis was applied to identify the risk factors and to predict the probability of success. The general logistic model expresses a qualitative dependent variable as a function of several independent variables, both qualitative and quantitative (Fox, 1984). If P is the probability of getting married at the age of 18+ years, then

Figure 1

where, β is a vector of the unknown coefficient and X is a vector of covariates that affect the legal age at marriage. The general logistic regression model can thus be expressed as:

Figure 2

which express the log odds of current users as a linear function of the dependent variable.

In Bangladesh, the legal minimum age at marriage for women is 18 years. In logistic model, the dichotomous response variable was assigned the value 1 for getting married at the age of 18+ years and 0 otherwise. The explanatory variables were: geographic region, respondents and their husbands’ education, working status, religion, watching TV and husband’s occupation. Statistical analyses were carried out using SPSS software.

Marriage in Bangladesh, as in neighboring North India, was traditionally very early (Maloney et al., 1981; Lindenbaum 1981). Among Hindus, in particular, child marriage – marriage before menarche - was common. It was also widely believed that early marriage for girls made them more malleable and accepting of their new family circumstances, and in particular of the authority of their husbands, and in-laws. Early marriage is perceived particularly by Hindus to be in keeping with the notion of the Joint Family, where ideally property is held jointly with the husband’s brothers and father – while he is still alive. Maloney et al., (1981) reported that many of their Hindu respondents believed that it was meritorious if a girl was married by the age of eight years, but a disgrace if she had her first menstruation while still living in her natal household. Under Muslim law guardians did not have the right to contract marriage for a pre-pubescent girl (Jhabvala 1975; Maloney et al., 1981), but clearly the Hindu example had influenced Bangladesh’s Muslims. Apart from protecting family norms, the major pressure for early female marriage was the fear of the consequences of delayed marriage for the girl and her family. For Hindus, it was critical to prevent miscegenation between different castes: this would result in ritual pollution with disastrous consequences for all families involved. For Muslims premarital sexual activity or even ‘unchaste’ behavior challenges not only the young woman’s virtue but also the ijat (or izzat - honour) of her family members (Maloney et al., 1981; Kotalova 1996; White 1992). This was closely linked to the Islamic concept of ‘Purdah’ (known in Bangladesh as Parda).

Female age at marriage has been rising intermittently since the 1960s. The simulate mean age at marriage (SMAM) was 13.9 years at the 1961 census – an apparent decline from 14.4 in 1951, but it rose to 15.9 by the 1974 census, 16.4 by the 1981 census and 18.1 by the 1991 census (Xenos and Gultiano 1992; United Nations 2000). The Bangladesh Fertility Survey of 1975-76 (Maloney et al., 1981) found that mean age at marriage of women who first married in 1927 was 10.9 years, while for women who married in 1957 it was 13.0 and for the youngest marriage cohort it was about 15. The 1999/2000 Bangladesh Demographic and Health Survey (BDHS) found a median marriage age among all women 15-49 (including both married and unmarried women) of 14.7 years. Older women aged 45-49 years recorded 13.8 years, and those aged 20-24, 16.1 years (Mitra et al., 2001). Age at marriage was varied by region, education and urban/rural residence. It was earliest in the two divisions believed to be most influenced by Hindu ideals, Rajshahi and Khulna, even though these areas are said to be those where Purdah is weakest. This was also despite the fact that Hindus now, in general, marry later than Muslims, in part because of higher education rates. For the ages 20-49 years, women with no education married at 14.0 years, primary incomplete 14.4, primary complete 15.0, and secondary plus 17.2 years. Secondary school often means delaying marriage. Women aged 45-49 years with secondary plus schooling married 2.1 years later than women with no education (15.6 years versus 13.5); for women aged 20-24 years the difference was 4.5 years (19.0 versus 14.5 years). Urban women aged 20-49 married at 16.2 years as compared to 13.7 years among rural women, a difference of 2.5 years. The difference increased from 0.9 years (14.6 years versus 13.7) for women aged 45-49 to 2.6 years (18.3 years versus 15.7) for women 20-24 years. This implies the rural-urban difference was larger than suggested by an uncontrolled comparison, but also that the change over time within the urban population was exaggerated.

Table 1 reveals the background information of the respondents living both in rural and urban areas of Bangladesh. Among the six administrative regions, the maximum number of rural women (24.0%) were from Rajshahi region, followed by Dhaka (19.4%), Chittagong (17.3%), Khulna (14.9%), Barisal (13.4%) and Sylhet (11.0%), respectively. In case of urban areas, the maximum number of women were from Dhaka region (28.8%), followed by Chittagong (19.7%), Rajshahi (19.3%), Khulna (15.0%), Barisal (9.0%) and Sylhet (8.1%), respectively. The illiteracy rate was higher among rural women (42.4%) than that of urban (31.4%). Women empowerment should be enhanced through the participation of women in income related activates as the results indicated that about 80% of the rural and 75% of urban women did not have any job (are not working). The results also showed that the majority of the respondents (about 80% of rural and 90% of urban) were Muslims by religion. Occupations of the husband were categorized into two groups such as manual (includes those who were involved directly with physical labor to earn) and non-manual (includes those who were involved with different professional, clerical and business related activities to earn). The results indicated that more than 75% of the rural husbands were engaged with manual occupation and about 62% for urban.

Figure 3

Table 1: Distribution of the respondents by background characteristics

To see the association between age at marriage and various selected background characteristics of rural-urban differentials in Bangladesh, a well-known statistical tool namely- Pearson Chi-square test procedure was used and the results are presented in Table 2. The results revealed that there were significant variations in legal age at marriage (among rural and urban women, it was 7.4% and 13.9%, respectively) among women with different socio-economic and demographic characteristics. Among the selected background characteristics - geographic region, respondents and their husbands’ education, religion, watching TV and husband’s occupation were significantly associated with the age at marriage of the ever-married women of Bangladesh.

Figure 4

Table 2: Results of Chi-square test between age at marriage (

The results of logistic regression analysis are presented in Table 3 in the form of logistic regression coefficients, p-value, and relative odds ratio corresponding to the selected explanatory variables both for rural and urban areas.

Dhaka division was considered as the combination and interactions of all other divisions, therefore, considering Dhaka as reference category, the regression coefficients both for rural and urban ever-married women of Bangladesh under different geographic regions were being calculated. Except for the rural areas of Barisal division, all results showed the statistically significant effect on age at marriage. For rural areas, the odds ratio corresponding to Chittagong and Sylhet divisions were found 1.737 and 3.131, respectively. This clearly indicated that the women of Chittagong and Sylhet divisions had 1.737 and 3.131 times respectively more probability in getting their married after the legal age at marriage (18+ years) than the women of Dhaka division (reference category). On the other hand, for urban areas, the odds ratio corresponding to Chittagong and Sylhet divisions were found 1.504 and 1.986 respectively. This also indicated that the women of Chittagong and Sylhet divisions had 1.504 and 1.986 times respectively more probability in getting their married after the legal age at marriage (18+ years) than that of the women of Dhaka division (Table 3).

It is apparent from the results that there are significant regional variations in age at marriage among the women of Bangladesh. Therefore, to order to achieve the replacement level fertility within 2015, various programs should be taken to remove the regional variations in age at marriage among the women of Bangladesh.

Education is the key determinant of the life style and status enjoys in a society. For this reason, the regression coefficients corresponding to different levels of education both for respondents and their husbands were calculated. It was seen that the results had statistically significant effect (both for rural and urban areas) on age at marriage except corresponding to primary level of education.

From the results, it is also evident that the women and their husbands of urban areas with secondary and higher secondary+ level education have higher age at marriage (married after the age of 18 years) than their rural counterparts.

In order to identify the impact of working status on age at marriage of the women, the respondents were categorized into working and not working. The regression coefficients demonstrated that working status of urban areas had significant effect on age at marriage. The odds ratio corresponding to working women of urban areas was 1.265. It means that the women of urban areas of Bangladesh have 1.265 times higher probability in getting married after the age of 18 years than the non working women of urban areas (reference category). In the contrary, the working women of rural areas had an effect on age at marriage, with working women having 1.161 times more likely to go for married after the age of 18 years, but it was not statistically significant (Table 3).

Religion is another important and highly significant factor influencing age at marriage of the respondents. The logistic co-efficient indicated that the highest occurrence of married after the age of 18 was among non-Muslim women of urban areas, followed by non-Muslims women of rural areas. It appears that the rural and urban non-Muslim women are 1.981 and 2.122 times respectively more likely to go for getting married after the age of 18 than the Muslim women (Table 3).

Mass media can play a strong role by creating awareness about the early marriage related complications and the bad effect of these complications in the future health of mothers and their new born babies. The result showed that the rural and urban women who watched TV were 1.101 and 1.579 times respectively more likely to go for getting married after the age of 18 than that of the women who did not watch. The result also showed the highly significant (statistically) effect for urban areas but for rural areas, this effect was not statistically significant.

Occupation of the husband had an effect on the female age at marriage, with husbands having non-manual occupation being 0.966 times less likely for rural and 1.1.3 times more likely for urban for getting married after the age of 18, but these effect were not statistically significant.

Table 3: Logistic regression estimates of the effects of different background characteristics on age at marriage of rural–urban differentials in Bangladesh

Figure 5