In Asia, Child Mortality Is Not Linked to Women's Autonomy or Religion

Lisa Remez, Guttmacher Institute

First published online:

An analysis of 15 Muslim and non-Muslim community pairs in four countries in Asia shows no consistent relationship between differences in levels of female autonomy and child mortality.1 In many areas, Muslim women do not have lower levels of autonomy than women of other religions. Moreover, indicators of women's autonomy do not have a strong or uniform association with child mortality.

To test common assumptions about relative levels of female autonomy and child mortality in Muslim and non-Muslim communities, and to determine whether female autonomy is associated with child mortality, the researcher used data from four of the five countries included in the Survey on the Status of Women and Fertility of 1993-1994. These countries--India, Malaysia, the Philippines and Thailand--have Muslim and non-Muslim populations.

The analysis compared six indicators of female autonomy and probabilities of under-five mortality within 15 Muslim and non-Muslim community pairs. The autonomy indicators measured women's freedom of movement (in 13 pairs a summative three-item scale of 0-3 and in two the proportion of women who did not need permission to go outside the home), whether women had ever been beaten by their husband, whether they were afraid to disagree with him, whether employed women had control over their income, whether women made independent economic decisions (summative four-item scale of 0-4) and whether women decided (either on their own or with their husband) what to do when a child was ill. Cox regressions were used to examine the effect of being Muslim on the hazard of before-five mortality in three of the four countries studied, once women's level of autonomy and socioeconomic status were controlled for.

Muslims were compared with Hindus in eight community pairs (five in India and three in Malaysia), with Christians in two pairs (both in the Philippines) and with Buddhists in five pairs (three in Malaysia and two in Thailand). Although the community pairs were also examined for differences in women's socioeconomic characteristics, such as the woman's (and her husband's) educational attainment, index of household possessions and income, no relationship between religion and socioeconomic indicators was evident.

On one of the six measures of women's autonomy, Muslim women were clearly at a disadvantage. In 12 of the 15 community pairs, Muslim women had less freedom of movement than non-Muslims. For example, in the 10 pairs using a summative score, the totals in Muslim communities (based on three items measuring whether the woman could go alone to the market, the health center and the community center) ranged from 0.2 to 1.9, whereas those in non-Muslim communities ranged from 0.3 to 2.9. In the two pairs in Thailand that used a different freedom of movement measure, 39-53% of non-Muslim women could go outside the home without seeking permission, compared with 9-25% of Muslim women. However, the proportions of women who had been beaten by their husband were lower or no different among Muslims than among non-Muslims in seven of the nine community pairs for which data were available.

The pattern for fearing to disagree with a husband was also mixed: Higher proportions of Muslim women feared to disagree with their spouse when compared with Christian women in the Philippines (84% vs. 34-41%) and Buddhist women in Malaysia (37-43% vs. 19-28%). However, Muslims were either less likely or no more likely to fear expressing disagreement than Hindus in India and Malaysia, or Buddhists in Thailand. Muslim women had no less control over their earnings than women of other religions (with the main exception of Christians in the Philippines), and differences by religion in economic decision-making were similarly small (except for the comparisons between Muslims and Buddhists in three settings in Malaysia and one in Thailand). However, Muslim women had less influence when a child was sick in 10 of the 15 community pairs studied, and this difference was large in six of the 10 differing pairs.

Overall, under-five mortality levels in the decade before the survey were not uniformly higher in Muslim than in non-Muslim settings. Eight of the 15 pairs differed significantly by child mortality, and in half of these community pairs, mortality was significantly lower among Muslims than among non-Muslims.

The correlations between differences in autonomy and differences in mortality were generally small, with differences in freedom of movement most often (in nine of the 15 pairs) coinciding with differences in mortality. In six areas, Muslims had lower freedom of movement scores and higher or similar child mortality rates, and in three areas, Muslims had higher freedom of movement scores and lower mortality rates. However, the author notes that in many settings where Muslims were at a disadvantage in both autonomy and mortality, they also fared worse on socioeconomic indicators, so "it is difficult to take any association between mortality and freedom of movement at face value."

The results of the Cox regressions, performed with data for 10 community pairs in the Philippines, Malaysia and Thailand, indicate that for the effect of religion alone, the risk of dying by age five is significantly greater for Muslims than non-Muslims in four of these 10 pairs. Introducing controls for both the freedom of movement autonomy measure and socioeconomic status did not change the significant and positive results for being Muslim in three of these four community pairs.

According to the researcher, her findings do not support the idea that rates of child mortality among Muslims are associated with levels of autonomy among Muslim women. In fact, she argues that this assumption "had far from strong or consistent support in the array of settings examined here." The author suggests that the Muslim disadvantages in mortality found in some countries may be related instead to "infrastructure, disease conditions, child care practices or other context-specific factors that were not measured (or are unmeasurable)."--L. Remez


1. Ghuman SJ, Women's autonomy and child survival: a comparison of Muslims and non-Muslims in four Asian countries, Demography, 2003, 40(3):419-436.