GENDER BIAS AGAINST FEMALE CHILDREN IN INDIA

Transcription

1 In: Gender Issues and Empowerment of Women ISBN: Editors: Manoranjan Pal, Prasanta Pathak et al. O Nova Science Publishers, Inc. Chapter 10 GENDER BIAS AGAINST FEMALE CHILDREN IN INDIA P. Arokiasamy* and J. Pradhan A major priority in the United Nations millennium development goals is to achieve gender equality and women's right. Achieving goals relating to gender equality are critical for achieving all the major millennium development goals (MDGs). South Asian countries and India are societies with strong patriarchal norms and high degree of son preference. Gender discriminations are pervasive in south Asian countries. In the Indian context, extremely high levels of gender discrimination against female children have been reported in the provision of health care, nutrition, education and resource allocation in northern and western states of India. In recent times, the natural biological laws of human reproduction of mankind for balancing its natural sex ratio, have been distorted by man-made norms, customs, traditions, religious beliefs and more recently by sophisticated medical technology resulting in lower sex ratio in India. In India, there has been a steady decline of sex ratio from 972 in 1901 to 933 females per 1000 males in From 1961 to 2001, sex ratios for children under age 6 have become more masculine (Das Gupta and Bhat, 1997; Desai, 1994; El-Badry, 1969; Miller, 1981; Paraswaman and Roy, 1991). Traditions, values and customs crusted over time have resulted in the insatiable desire for sons. Sons are preferred over daughters for a number of economic, - social and religious reasons, including financial support, old age security, property inheritance, dowry, family lineage, prestige and power, birth and death rituals and beliefs about religious duties and salvation (Dyson and Moore, 1983; Das Gupta, 1987; Das Gupta and Bhat, 1997; Basu, 1989; Miller, ; Caldwell and Caldwell, 1990). Consequently, women and girl children are accorded lower status in the Indian society. Women in India face discrimination in terms of several political and economic opportunities as a result of their inferior status. Majority of women cannot inherit parental property and political and employment participation are very limited. Gender inequalities Reader, International Institute for Population Sciences (IIPS), Deonar, Mumbai E mail-

2 188 P. Arokiasamy and J. Pradhan prevail in work, education, allocation of food, health care and fertility choices. On the other hand, at the family level women are exclusively burdened with household chores- cooking, cleaning, collecting fuel, water and caring elderly and children (Arokiasamy, 2004). In this context, this paper assesses evidences concerning gender bias against female children in India. The gender bias against female children in health, nutrition, education and the consequent gender gaps in these ind~cators tend to act as barriers in the efforts of Government to achieve the UN millennium development goals. Education and Employment The economic and social rates of returns of education are quite high, and on the whole, higher for women than for men. In patrilineal and patrilocal communities, however, daughter's education is viewed as a waste, because parents expected returns from educated daughters do not exceed the costs. As a result, female education as an investment becomes unattractive to parents. In most developing countries, women are relatively less educated than men. Although the benefits of women's education are well recognized, a number of cultural factors contribute to the gender gaps in education, with varying intensities across the states of India. Girls are either not sent to schools or do not receive the same quality and level of education, as do boys. In India, gender inequality in enrolments is worse at the secondary and tertiary level than at the primary level. Child ~ortality In the absence of a biological basis, evidence of excess female child mortality is an important indicator of gender inequality. A review of demographic health survey findings revealed that 27 out of 44 DHS countries had higher mortality for girls than boys in 1-4 years, although the average excess female mortality in 44 countries was only 2 percent (Arnold, 1997). In India, trends in child mortality by sex from the Sample Registration System indicate that the levels of excess female chiid mortality as a result of son preference have increased during the last several decades (Registrar General, 2001, 2003). Further, data from the National Family Health Survey-1 ( ) indicates that child mortality for girls in India as a whole, at 42.0 per 1000, is 43 percent higher than for boys at 29.4 per 1000 (IIPS, 1995). The corresponding rate from NFHS-2 was 42 per 1000, which is 47 percent higher than that of boys at 28. Nutrition Girls are more likely to be malnourished than boys in northern and western states (Sen and Sen Gupta 1983; Arnold et al, 1998; Pebley and Amin, 1991). Gender differentials in nutritional status are reported during infancy, with discriminatory breastfeeding and

3 Gender Bias against Female Children in India 189 supplementation practices. Infant girls are breastfed less frequently, for shorter duration, and over shorter periods than boys (Wyon and Gordon, 1971; Das Gupta, 1987). However, national family health survey indicated some varying evidences where boys and girls are about equally likely to be stunted, underweight but boys were slightly more likely than girls to be wasted. Health Care Provision In the demographic literature, there are frequent debates on the origin of differences in. morbidity and mortality of children, in particular, the relative role of biological and behavioural factors (Lopez and Ruzicka, 198 1; Preston, 1976; Langford, 1984; Waldrom, 1983; Hill and Upchurch, 1995; United Nations, 1998). Overall, biological factors are considered to be less important, while societal discrimination of girls in nutrition, preventive and curative health care seeking is found to have an impact on morbidity and mortality. Female selective infanticide is an extreme form of societal discrimination, but evidence suggests that its prevalence is too low to make a significant impact on excess female mortality. Evidence of sex selective abortion is yet another severe form of gender discrimination, which recent studies have documented (Arokiasamy, 2004; Arnold et al, 1998). Gender differences in health care are recognized as the direct consequence of discrimination against females in seeking health care. Differentials in treatment of children by sex are known to be directly linked to the differences in mortality of boys and girls (D'Souza and Chen, 1980; Das Gupta, 1987). In India, evidence of discrimination of girls in both preventive (immunization) and curative (treatment) care is reported with varying degrees amongst the states. Studies have recognized female discrimination in health care as the main pathway for excess female child mortality. Boys are more likely than girls to be taken to a health facility when Gupta, 1987; Govindasamy and Ramesh, 1996). Boys had higher immunization rates than did girls in all the states except Goa and Kamataka, although the extent of this difference varied amongst the states. Even when such discrimination in health care is not leading to fatality, it can still produce greater frailty among survivors and thus is an important child health issue in itself (Mosley, and Chen, 1984). Poor health has implication for surviving girls. Their poor health in reproductive years may be perpetuated across generations. Several studies have recognized the cultural diversity between north andsouth Indian states. North Indian kinship structure with exogamous mamage system favour strong son preference and lower female autonomy compared to south Indian kinship structure of endogenous mamage system (Dyson and Moore, 1983; Karve, 1965). Though, recent evidences suggest some blurring of north- south disparity in gender discrimination. The focus of this analysis is to examine the evidences of regional differences in gender bias against female children. A multiple indicator approach is used unlike, earlier studies,

4 190 P. Arokiasamy and J. Pradhan which have focused on selected indicators when dealing with the issue of gender bias against female children. The above pathways of gender bias cumulatively contribute to excess female mortality and gender inequalities. Accordingly, in this paper gender biases against female children are examined on five domains namely, school attendance, nutrition, immunization, medical treatment of children and child mortality. The levels of gender bias in adult literacy and work participation are lastly compared with these indices of gender bias against female children. DATA SOURCES AND METHODS OF ANALYSIS The national family health survey-2 (NFHS, ) data is used in this analysis. The NFHS-2 provides information on fertility, fertility preferences, family planning, child mortality and child health indicators. The survey also provides data on child health related measures'such as child immunization; incidence of diarrhoea, respiratory infections, and fever; treatment seeking for these diseases; and anthropometric indicators of nutritional status such as height and weight, for children under age 3. In this analysis, information on treatment of children if the child had fever, cough and diarrhoea are used. On preventive care, information on whether the child was fully immunized or not with respect to the recommended expanded programme on immunization (EPI) vaccines that includes BCG, three doses of DPT and polio and measles are used. On nutrition, data on weight, height and breastfeeding are used. On education, information on school attendance is used. The extent of gender bias against female children is examined based on sex differences on all these five dimensions. Methods of Measuring Gender Bias Multiple indicators are used to study the extent of gender bias in major states of India. Five indices of gender biases relating to school attendance (Rs), nutrition (Rn), immunization (Ri), treatment of illness (Rt), child mortality (Rcm) and overall index of gender bias (Rc) are defined. Rc represents an aggregate index of gender bias against female children, which is compared with F/M ratio on child mortality. This gender bias index is also compared with gender gaps in overall literacy and work participation rate. The variables used in the calculation of indices of gender bias are: PGAS MDBG PGBZSDWA PGG2SDHA PGBZSDWH PGI PGIC PNIG Proportion of girls attending school Medium duration of breastfeeding for girls Proportion of girls below threshold of 2SD of median weight-for-age Proportion of girls below 2SD of med~an height-for-age Proportion of girls below 2SD of median weight-for-height Proportion of girls fully immunized Proportion of girls not having immunization card Proportion of girls not immunized

5 Gender Bias against Female Children in India 191 PGHC - Proportion of girls who had cough PGHF - Proportion of girls who had fever PGHD - Proportion of g~rls who had diarrhoea PGHDB - Proportion of girls who had diarrhoea with blood PGTHPD - P.roportion of girls taken to health provider (for diarrhoea) PGTHPARI - Proportion of girls taken to health provider (for acute respiratory infection) CMG - Child mortality rate for females Similarly, PBAS, MDBB, PBB2SDWA, PGB2SDHA, PBB2SDWH, PBI, PBIC, PNIB, PBHC, PBHF, PBHD, PBHDB, PBTHPD, THPARI, CMB represents the proportions for boys. The indices are calculated are as follows: 1. Sex ratio in school attendance lndex of gender bias in schooling: PGAS(6-10) where, R1 = R2 = PBAS(6-10) PGAS(11-14) PBAS(l1-14) 2. Sex ratio in nutritional status lndex of gender bias in nutrition: MDBG PBBZSDWA PBBZSDHA PBBZSDZWH where, R3 = - MDBB' R4 = PGBZSDWA' R5 = PCBZSDHA' R6 = PcBzsDzw~ 3. Sex ratio in immunization Indices of gender bias in immunization: PC1 PGlC PNBl where, R7 = - R8 = - and R9 = - PBI' PBIC PNCI' 4. Sex ratio of disease incidence Index. of disease incidence: PBHC PBHF PBHD PBHDB where, R10 = - R11= - R12 = - and R13 = - PCHC' PCHF' PGHD PGHDB'

6 P. Arokiasamy and J. Pradhan F 5. Sex ratio in treatment seeking lndex of gender bias in treatment seeking: R14+R15 PCTHPD PCTHPARI Rt =, where R14 = - PBTHPD 6. Sex ratio in child mortality lndex of gender bias in child mortality: CMG Rcm = - CMB and R15 = ps,,.hpar,. 7. A composite index of gender bias against female children has also been computed by aggregating the four sub indices of gender bias as shown below. Composite index of gender bias: Schooling The extent of gender bias in schooling in terms of female to male ratio of school attendance rate by age of child, 6-10 years and years is presented in Table 1. Female deprivation in schooling is very high in Rajasthan, Bihar, Uttar Pradesh, Gujarat, Andhra Pradesh, Jammu and Kashmir and Orissa. The ratio of female to male school attendance - indicates 20 to 40 percent lower attendance for girls compared to boys in these states. Comparatively, the gender gap in female schooling is percent in the states of Assam, Punjab, Maharashtra, Karnataka and Tamil Nadu. The F/M ratios for school'attendance indicate lesser gender differences for children in age 6-10 years and more pronounced differences in ages 10+ years, due to higher dropout of girls compared to boys when they move to higher standards. While the gap in primary enrolment is the first stage discrimination of girls not being sent to school, discontinuation of girls at middle and secondary level is the second stage discrimination. Such gender gap in school attendances is an important pathway of gender bias against female children. Nutrition Table 2 shows that sex differential in nutritional status of children under 3 years in terms of underweight and stunting is less pronounced (5-10 percent gap). The malelfemale ratios in nutritional status of children indicate moderate levels of nutritional neglect of female children in Gujarat, Uttar Pradesh, Maharashtra, Rajasthan, Delhi, Haryana, Punjab and Andhra

7 Gender Bias against Female Children in India 193 Pradesh. The nutritional deprivation of female children in terms of weight for age and breastfeeding is sharper in these states. However, boys are more likely than girls to be wasted. The reason for lower level of female wasting may not be limited to differential food intake alone as girls are physically less active than boys or may be less prone to nutritional inadequacy: A reason being their relative lower requirements. Although, nutritional neglect indicates lesser gender gaps cumulatively, long-teim nutritional inadequacy tend to have Inany negative consequences such as higher risk of disease, deterioration in physical and mental health and intergenerational iinpact of nutritional inadequacy passed on o their children. Table 1. Gender bias in school attendance rate by age of the children, Percent of children i tending school States 6-10 years - Male Female Jammu and Kashmir Himachal Pradesh Punjab Haryana Delhi Rajastban Uttar Pradesh Madhya Pradesh Bihar Assam West Bengal Orissa Gujarat Maharashtra Andhra Pradesh Karnataka Tamil Nadu Kerala Note: * ratio of 1 indicates genderneutral; ratic years F/M (R,)* Male 1 Female 1 F/M (R2)* <I indicates female disadvantage. Rs* Immunization Of the three indicators of immunization, two relate to utilization and non-utilization of immunization. The femalelmale ratios presented in table 3 show evidence of pronounced gender differences in full immunization coverage in most states among children between 1-2 years of age. Girls have tremendous disadvantages in immunization coverage. The F/M ratio of children who had received all the required vaccination varied from for Assam to 1.16 for Andhra Pradesh.

8 194 P. Arokiasamy and J. Pradhan Table 2. Gender bias in nutritional status of the children under age 3 years by states, Kamataka Tamil Nadu Kerala Note: * ratio of 1 indicates gender neutral; ratio <1 indicates female disadvantage The ratios indicate no evidence of gender bias in Kerala, Tamil Nadu, West Bengal, Orissa and Andhra Pradesh. The second indicator that is examined is children who did not receive any vaccination at all. These were the children possibly from households with very little access to health care and poor socio-economic condition. The MIF ratios of children not given vaccination ranged from 0.53 in Himachal Pradesh to 1.33 in Kerala. The ratio value of above one indicates female advantage in Maharashtra, Andhra Pradesh, Karnataka and Kerala. Haryana and Gujarat are surprising addition to this group. Very high level of female disadvantage is found in most northern states. The value attached to the safe keeping of vaccination cards is an indication of interest shown by the 'parents, especially the mother, in the health and well being of the individual child. Therefore, it is useful to examine gender disparity (femalelmale ratio) in terms of possession of vaccination card. The ratios

9 Gender Bias against Female Children in India 195 vary from 0.57 to 1.I7 across the states. The ratio value above one in West Bengal, Andhra Pradesh, Karnataka, Tamil Nadu and Kerala indicates that female children have been in advantages position. The overall index of gender bias in immunization shows very high levels of fernale neglect. The index value is lower than one by percent in Assam, Himachal Pradesh, Punjab, Madhya Pradesh, Bihar, and Jarnmu and Kashmir. Moderate levels of female child neglect are found in Uttar Pradesh and Orissa with index value lower than one by 5-20 percent. The index value indicates no evidence of female child neglect in immunization in the south Indian states. Table 3. Gender bias in immunization of children age months by states, States Jammuand Kashmir Hirnachal Percent of children Fully immunized M F F/M (R7)* With vaccination card M F F/M (RBI* ' Not Vaccinated M F M/F (R9)* Ri* Note: * ratio of 1 indicates gender neutral; ratio <1 indicates female disadvantage.

10 I Table 4. Sex differentials in the incidence of illness of children under 3 years, Percent of children who had

11 Gender Bias against Female Children in India 197 Incidence of Illness and Treatment Seeking by Sex When examining gender bias in treatment of illness, as a priori, it is important to examine sex difference in reported illness prevalence. The linkage between severity and duration of infection, the medical treatment received, its timing and case fatality rates by sex are clearly complex. Diarrhoea1 disease, malaria and acute respiratory infections are leading causes of infant and child mortality in developing countries including India, but incidence of these diseases vary by sex. In NFHS-2 questions were asked on fever, cough accompanied by breathing and diarrhoea with blood with reference period 'is last two weeks. However these data have several limitations and recall biases. Recall biases tend to vary across the Indian states for different symptoms firstly, due to varying levels in literacy awareness of diseases and treatment seeking. These are likely to affect male-female comparisons. Secondly, if sex of the child affects the perceived severity of condition, the data obtained about boys and girls will be affected differently. Apparently, lower female morbidity might result from a bias in reporting incidence of sickness among female children although the question asked is the same (Hill and Upchurch, 1994). The problem might also arise if mothers perceived girls illness, only when they have more severe illness than boys, because such differences would affect treatment seeking. Therefore, it is necessary first to examine sex differences in the incidence of illness. Table 4 shows the sex differentials in the incidence of illness such as diarrhoea, fever and acute respiratory infection. Results indicate that boys had higher incidence of cough in all the states except Gujarat. The incidence of fever is also reported to be higher for males in most states. However, the reported incidence of diarrhoea with blood is higher for females compared to males. With respect to the incidence of all the three diseases, girls have lower reported incidence compared to boys. In patriarchal societies, women and girl children particularly face the greatest problem in acquiring adequate health care. An important way in which gender bias manifests is that girls are less likely to be taken for health care as often or as early in their illness as boys because girls may receive less attention from parents. Also first-born girls tend to receive better health care than their younger sisters. Such neglect of girls in the use of curative health services is an important link that directly account for female child mortality disadvantage. Previous analyses tended to focus on whether among those children who were sick, are there sex differentials in being treated at all or type of treatment received? This relationship is confounded by sex differentials in the timing of use of different providers and with the severity of disease. In this context, sex differentials in the source of treatment are more helpful to explain sex differential in mortality. For instance, girls might be taken to professional health provider for treatment only after their illness has deteriorated as a measure to save their lives. In Uttar Pradesh, northern India, boys were taken to city hospitals when warranted while girls were taken to less qualified doctors. Basu (1989) examined both the source and type of treatment in conjunction. Her study focused on two groups of children aged less than 12 years living in a resettlement slum in New Delhi; the parents of one group had migrated from Uttar Pradesh and those of other from Tamil Nadu. In addition to sex differences in the proportion of children not receiving treatment in both groups, the north Indian girls were more likely to receive treatment from a non-professional health provider than boys. Therefore, evidences of sex differentials in

12 198 P. Arokiasamy and J. Pradhan proportion being treated by a professional health care provider are more crucial. Since data on the timing of use of different health services was collected in this survey, we examine sex differences in the treatment of children. The F/M ratio in the proportion of children, who were taken to professional health provider, is presented in Table 5. Table 5. Gender bias in treatment seeking for diarrhoea and acute respiratory infection (ARI) among children age 0-35 months, States ( Percent of children treated by professional health provider for Diarrhoea M F Jarnrnu and Kashrnir Hirnachal Pradesh , Punjab Haryana Delhi Rajasthan Uttar Pradesh Madhya Pradesh Bihar Assam West Bengal Orissa Gujarat Maharashtra Andhra Pradesh Karnataka Tamil Nadu ' Kerala Note: * ratio of 1 indicates gender neutral; ratio <1 indicates female disadvantage. F/M (R14)* ARI M F F/M (R15)* I Rt*! For treatment of diarrhoea, the ratio ranges from 1.05 in favour of girls to 0.83 in favour of boys. In terms of being taken to a health provider for diarrhoea1 treatment, girls are at a significant disadvantage in six states namely Madhya Pradesh, Assarn, West Bengal, Orissa, Andhra Pradesh and Karnataka. The range in values of F/M ratio widens for treatment of acute respirators infections (AN) showing higher levels of gender bias in treatment seeking for ARI in these states. The evidence of sex differences in treatment seeking both for diahorrea and ARI suggests very significant neglect of female children although the incidence of diarrhoea with blood is higher for females. Child Mortality Excess female child mortality (in.age 1-4) is a more precise outcome indicator of gender bias. A large number of studies have explored the dynamics of excess female child mortality in the northern states of India, where excess female child niortality is the highest in the world.

13 Gender Bias against Female Children in lndia 199 Evidences have shown that mortality among female children is Lower than male children during neonatal period which reverses to female disadvantage in the postneonatal period (Arokiasamy, 2004). The reversal from male disadvantage in neonatal mortality to female disadvantage in post neonatal mortality is abrupt for most states of the north and north central region where excess female child mortality is more than 50 percent. Table 6. Logistic regression analysis of gender bias controlling for selected variables, States Jammu and Kashmir Himachal Pradesh Punjab Ha~yana Delhi Rajasthan Uttar Pradesh Madhya Pradesh Bihar Assam West Bengal Orissa Gujarat Maharashtra Andhra Pradesh Kamata ka Tamil Nadu I Kerala Note: ***<0.01, **<0.05,. Odds of female childc'being Fully Attending Underweight Taken to immunized I school I I health + I provider (Diarrhoea) Taken to health provider (ARI) (6-14 years) ** ** *** *** *** 0.88* 0.84** 0.65** *** 0.53** 0.46** 0.60*** ** 0.84* 0.75** 0.83** 0.57* 0.69* I ** ** * male child as reference category. Note: 1 : variables worn age, education, place of residence, casteltribe and standard of living index are used as control factors. For India, mortality among female children in age 1-4 was 43 and 47 percent higher than among male children respectively in NFHS-1 ( ) and in NFHS-2 ( ). The malefemale ratio of child mortality (Rcm) from NFHS-2 presented in Table 7 indicates the level of excess female child mortality by states. The ratios show extremely high level of excess female child mortality with the highest in Punjab and Haryana followed by Uttar Pradesh, Rajasthan and Asaarn. Excess female child mortality is comparatively lower in south Indian states with Kerala showing no evidence of excess female child mortality at all. Multivariate Analyses Logistic regression models are used to assess the net effect of variable sex of the child on health, nutrition, treatment, and school attendance, by controlling for related demographic and socio-economic status variables. Five logistic regression models are estimated with dependent variables (a) child attending school, (b) nutritional status below -2SD, (c) fully immunized,

14 200 P. Arokiasamy and J. Pradhan (d) child with diarrhoea was taken to a health provider and (e) child with ARI was taken to a health provider. As the main aim is to examine the extent of gender bias, only sex wise odds ratios have been estimated adjusting for other variables. The odds of female children being discriminated against are estimated, controlling for several socioecono~ni covariates such as, women's age, education, household standard of living index, caste and residence. Table 7, Comparison of indices of gender bias with literacy rate, work participation rate and population below poverty line State Jammu and Kashmir Himachal, hadesh Punjab Haryana Delhi Rajasthan Uttar Pradesh Madhya Pradesh Bihar Assam West Bengal Orissa Gujarat Maharashtra Andhra Pradesh Kamataka Tamil Nadu Kerala Note: '* ratio of Indices of gender bias Ri* ' Rs* Rn* Rt* I indicates gender neutral; neutral; <I= Female advantage. Rc* ratio <I Rcm** indicates Literac Y Rate7+ (Lr=F/ Population Below Poverty Line (%) The above indices of gender bias against female children in terms of these multiple indicators are based on &adjusted differences. The net effects of gender bias in various domains of neglect are therefore assessed using logistic regression models. Logistic regression model estimates presented in Table 6 with dichotomous categories of 'yes' and 'no', for four dependent variables, namely children fully immunized, underweight, attending school, treated for diarrhoea and treated for ARI. The odds of female child receiving care in yes category with male as reference category are estimated controlling for background variables such as women's age, education, household standard of living, place of residence and caste. An odds ratio of I.OO indicates that the odds of that outcome are the same for boys and girls female Work Panicipatio n Rate (Wr=F/M)* disadvantage; ' ** I=Gender

15 Gender Bias against Female Children in India 20 1 In most states, the odds of female children receiving immunization are significantly lower. Steeply lower odds ratios are indicated in the states of Punjab, Delhi, Madhya Pradesh and Assam suggesting that gender bias is very strong in these states. Besides Kerala, the surprising exceptions are, Haryana, Rajasthan, Orissa, Andlira Pradesh, where odds ratios are greater than one but not significant. The odds of females being underweight are significantly higher in a majority of states. Strong nutritional deprivations among females are noticed in West Bengal, Gujarat, Haryana, Punjab and Andhra Pradesh. Evidence of gender discrimination is strong with respect to treatment seeking for acute respiratory infection. The odds of female children being taken to a health provider is (lower about percent) 1n the states of Jammu and Kashmir, Rajasthan, Uttar Pradesh, Bihar, Orissa, and Andhra Pradesh. The pattern of sex differentials in treatment seeking for diarrhoea similarly indicates evidence of a strong gender bias against female children. Socio-Economic Condition and Gender Bias The levels in various sub-indices and composite indices of gender bias with female disadvantage in child mortality and in development indicators are further compared across the states. Levels of excess female child mortality indicate positive association with levels of gender gap in school attendance and nutritional neglect of female children. Gender bias.in immunization similarly shows positive correspondence with excess female child mortality. Neglect in treatment, however indicates no clear trend. This needs to be explored further. Also a composite index of gender bias shows no strong association with the level of excess female child mortality. The composite index of female bias indicates a strong positive relation with the gender gap in adult literacy and poverty ratio. Per-capita income of the states shows a good correspondence with the composite gender bias index. However, femalelmale ratio in work participation indicates no relation with the composite index of gender bias. j The four sub indices of gender bias and the composite (cumulative) index of gender bias provide a comprehensive evidence base about the extent of gender bias against female children across the Indian states. The indices show significant degree of gender bias against female children 'in most north and north central states of India. However, the indices of immunization and school attendance indicate consistent and sharper sex differences suggesting systematic neglect of girls. In the south Indian states and Maharashtra, the extent of gender bias is marginal with Kerala showing advantageous position for females. A comparison of gender bias in terms of composite index and female disadvantage in child mortality indicates good correspondence among majority of the states. However, in Haryana, Assam, and Orissa the two ratios of composite index of gender bias and that of child mortality indicate mixed trends. While F/M ratios of child mortality indicate greater degree of gender inequalities, the ratio of'composite index shows a narrower degree of gender inequalities in the states. The intensity of gender bias or the lack of it is highlighted in the

16 202 P. Arokiasamy and J. Pradhan summary table of gender bias pattern across the states. Gender inequalities are pronounced for indicators of immunization, school attendance, treatment and child mortality. The extent of gender bias demonstrated in this analysis has important implications for achieving millennium development goals. The path to halving and further eliminating female discrimination in school attendance, nutrition, health care use and female disadvantage in child mortality appear to be long in India. Reaching millennium development goals in states with overall grading of intense gender bias, will take longer than what is currently envisaged. In relation to each of the millennium developlnent indicators, specific policy intervention is critical for narrowing gender differentials and for achieving millennium development goals. Arokiasamy, P "Regional patterns of sex bias and excess female child mortality in India" Population-E, 59(6): Arnold, F., M. K. Choe and T.K. Roy "Son Preference, the family building process and child mortality in India" Population Studies 52: Arnold, F Gender preferences for children. DHS comparative studies no: 23. Calverton, MD: Macro International. Basu, A.M "Is discrimination in food really necessary for explaining sex differentials in childhood mortality?" Population Studies 43(2): Caldwell, P. and J. C. Caldwell Gender implication for survival in South Asia. Health transition working paper Canberra: National Centre for Epidemiology and PopuIation Health, Australian National University. D'Souza, S. and L.C Chen "Sex differentials in mortality In Bangladesh Population and Development Review 6(2): Das Gupta, M "Selective discrimination against female children in rural Punjab, India" Population and Development Review 13(1): Das Gupta, M. and P. N. Mari Bhat "Fertility decline and increased manifestation of sex bias in India" Population Studies 51(3): Desai, S Gender inequalities and demographic behaviour: India. New York: Population Council. Dyson, T. and M. Moore "On kinship structure, female autonomy, and demographic behaviour in India" Population and Development Review 9(1): El- Badry, M.A "Higher female than male mortality in some countries of south Asia: A digest" American Statistical Association Journal 64(3): Govindasamy, P. and B. M. Rarnesh "Maternal education and gender bias in child care practice in India" Paper presented at the annual meeting of the Population Association of America, New Orleans, May Hill, K. and D. M. Upchurch "Gender differences in child health: evidence from demographic and health surveys" Population and Development Review 2 l(1): International Institute for Population Sciences (IIPS) National family health survey, India, Mumbai: IIPS. International Institute for Population Sciences (IIPS) National family health survey, India, Mumbai: IIPS.

17 Gender Bias against Female Children in India 203 Karve, I Kinship organization in India. Mumbai: Asia Publishing House. Langford, C. M "Sex differentials in mortality in Sri Lanka: Changes since the 1920s. Journal of Biosocial Science 16(3): Lopez, A. D. and Ruzicka, and L.T. (eds) Sex d~flerentials in mortality. Canberra: Australian National University. Miller, B. D The endangered sex: neglect of.female children in rural north India. Ithaca, New York: Cornell University Press. Mosley, W. H. and L. C. Chen "An analytical framework for the study of child survival in developing countries" Population and Development Review (Suppl.) 10: Parasuraman, S. and T. K. Roy "Some observations of the 1991 census population of India." Journal of Family Welfare 37 (3): Pebley, A. R. and S. Amin "The impact of public health intervention on sex differentials in childhood mortality in rural Punjab, India" Health Transition Review l(2): Preston, S. H "Mortality pattern in national population, Chapter 6, Cause of death responsible for variation in sex mortality dzflerential" New York: Academic Press. Pp Registrar General, Compendium of India's fertility and mortality Indicators, , Sample Registration System, Ofice of Registrar General, New Delhi. Registrar General, Sample Registration System Bulletin, Office of Registrar General, New Delhi. Sen, A and S. Sengupta "Malnutrition of rural children and the sex bias" Economic and Political Weekly (Annual number, May) 18: United Nations "Too young to die: genes or gender?" New York: United Nations, Population Division. Waldron, I "Sex differences in human mortality: the role of genetic factors" Social Science and Medicine 1 7(6): Wyon, G., and J.E. Gordon "The Khanna study: population problem in rural Punjab. " Cambridge, MA: Harvard University Press.