Objectives. We examined the role of women’s education and proximate educational context on intimate partner violence (IPV).

Methods. We examined a sample of 83627 married women aged 15 to 49 years from the 1998 to 1999 Indian National Family Health Survey. We used multilevel multiple logistic regression modeling to estimate the relative effect of women’s and their husband’s levels of education, spousal education differential, and community-level literacy on women’s risk of recent and lifetime IPV.

Results. In adjusted models, odds of recent IPV among women without any education were 5.61 times (95% confidence interval [CI] = 3.53, 8.92) those of college-educated women, and odds among wives of uneducated men were 1.84 times (95% CI=1.44, 2.35) those of wives of college-educated men. Women with more education than their husbands were more likely than those with educational parity to report recent IPV (odds ratio [OR]=1.18; 95% CI=1.05, 1.33). The results were similar for lifetime IPV. After we controlled for individual factors, as community male and female literacy levels increased, likelihood of IPV declined.

Conclusions. Although increasing women’s levels of education is crucial to reducing IPV for women, proximate educational context is also an important factor in reducing this public health burden.

Intimate partner violence (IPV) is an important global public health concern13 related to adverse outcomes such as physical trauma, mental illness, psychosomatic illness, poor health-related behaviors, poor birth outcomes, suicide, and murder.2,4 Recent international estimates indicate that the percentage of women with a lifetime experience of IPV is as high as 71% (rural Ethiopia)2 and falls between 21% and 47% in the majority of countries.1 IPV is acute in South Asia, a region where issues related to basic gender equity have been much discussed.58 In India, which accounts for 16% of the global female population and 76% of the South Asian female population, some 40% of women have reported being physically abused by their husbands during their adult lives.9

We investigated the role of proximate educational context in women’s likelihood of experiencing IPV. We hypothesized not only that a woman’s own educational attainment is protective against IPV, but that both the household and the communitywide educational contexts of a woman are a factor in IPV. The need to consider proximate educational context—which can be conceptualized as the level of education among all people near to each woman (e.g., partner’s level of education, educational differential between woman and partner, level of education in the community of residence)—is crucial, given the costs and benefits associated with education of women and of people generally.

In the United States, there is a growing body of evidence of the relation between neighborhood context and domestic violence. This research, theoretically grounded in the concepts of social disorganization10 and collective efficacy,11 has found that women who live in areas marked by high levels of poverty,12,13 unemployment,14 and neighborhood disadvantage15 are more likely to experience IPV. This work has been extended to India, with one study finding that IPV is associated with the district murder rate in Uttar Pradesh.16 These findings of area effects on IPV suggest that area educational levels could be both an independent determinant of a woman’s likelihood of experiencing IPV and a potential modifier of the relation between her own education and her likelihood of experiencing IPV. Indeed, education is related to norms of the acceptability of IPV,17 which in turn influences a woman’s likelihood of being abused.1820

Findings from this research can inform advocates and policymakers of the need to focus attention and financial resources on education in the drive to reduce IPV in India. Using the most recent nationally representative population-based data from India, we examined the relative importance of a woman’s education and proximate educational context in IPV.


We used the 1998 to 1999 Indian National Family Health Survey (INFHS), a nationally representative cross-sectional study of 92 447 households.21 Trained data collectors interviewed an adult member in each selected household to obtain socioeconomic and demographic information about the household and its family members, achieving a household response rate of 98%. Data collectors performed face-to-face interviews with 90 303 ever-married women aged 15 to 49 years from responding households. The women’s survey had a response rate of 96% and provided information on IPV. Given the sensitive nature of the women’s survey, extensive efforts were made to protect the participants and to ensure the quality of the data collected.22

We restricted the analysis to married women and excluded divorced, widowed, or separated women (n = 5441); we also excluded 1235 women because of missing information on IPV and other predictor variables, yielding a final analytic sample of 83 627 women. These women were located in 3215 primary sampling units in 440 districts in the 26 Indian states. Primary sampling units, hereafter termed neighborhoods, were villages or village clusters in rural areas and census enumeration districts in urban areas.

Outcome Measures

We analyzed 2 outcome measures: lifetime IPV and recent IPV. A screening question asked, “Since you completed 15 years of age, have you been beaten or mistreated physically by any person?” Two follow-up questions asked “Who has beaten you or mistreated you physically?” and “How often have you been beaten or mistreated physically in the last 12 months: once, a few times, many times, or not at all?” A woman was considered to have experienced lifetime IPV if she reported ever being physically abused by her husband since age 15. A woman was considered to have experienced recent IPV if she reported being physically abused by her husband in the previous 12 months.


Individual and proximate educational context for each woman was defined in the following ways. First, we recorded the educational attainment of the woman in terms of the following categories, each representing significant milestones in the formal Indian education system: 0 years, 1–5 years, 6–8 years, 9–10 years, 11–12 years, or 13 or more years. Second, we considered the educational attainment of the spouse in the same manner. Third, we analyzed the educational discrepancy between spouses with 3 categories in which (1) the husband had achieved a higher educational milestone, (2) the wife had achieved a higher educational milestone, or (3) the couple had educational parity. Finally, a woman’s community educational context was conceptualized as neighborhood (i.e., community-level) male and female literacy levels and defined as the proportion in each woman’s residential neighborhood of respondents of each gender (15 years or older) that had completed at least 1 year of formal education. Area-level literacy measures were calculated from the INFHS household survey data set, which contained 333835 men and women 15 or older residing in 3215 neighborhoods, and were specified as tertiles for analysis. Tertiles for female literacy were specified with the following percentage ranges: 0%–35.44%, 35.45%–63.99%, and 64%–100%. Tertiles for male literacy were specified with these percentage ranges: 3%–69.39%, 69.40%–86.29%, and 86.30%–100%.


We included several socioeconomic and demographic variables: age in years, age at marriage, religion, social caste, standard of living, employment status, and location of neighborhood (Table 1). Ages were grouped in 5-year ranges, and age at marriage was categorized as under 15, 15 to 17, 18 to 20, and over 20 years. Religion was recorded as Hindu, Muslim, Christian, Sikh, or Other. Caste was determined according to the identification of the head of the household as belonging to a scheduled caste, scheduled tribe, other backward class, or the general class.

Briefly, scheduled castes are those with members in the lowest level of the caste system and who have suffered the greatest burden of deprivation.23 Scheduled tribes are approximately 700 officially recognized social groups characterized by their geographic isolation and limited social and economic interaction with the rest of India.24 “Other backward class” is a legislatively defined group representing those who have historically suffered significant social deprivation not as severe as that suffered by scheduled castes and tribes. Thus, the general class is a residual caste grouping that traditionally has been viewed as better off than the above groupings.

Standard of living, defined in terms of living environment and material possessions, is a reliable and valid measure of household material well-being or wealth.25 Each person was assigned a standard of living score based on a linear combination of scores for 19 different household characteristics, such as quality of the home, type of fuel used for cooking, and ownership of a bicycle or television, weighted according to a factor-analysis procedure.26 The analytic models used quintiles of these weighted scores. Employment was classified according to whether the woman was not working, performing unpaid work, or working for pay in a manual, nonmanual, or agricultural profession. Finally, information from the 1991 Indian National Census was used to create categories defining whether each neighborhood was in an urban area of more than 1 million people (large city), an urban area of between 100 000 and 1 million people (small city), an urban area of less than 100 000 people (town), or a rural area (village). The Indian National Census defines a rural area as having at least 1 of 3 characteristics: (1) fewer than 5000 residents, (2) population density less than 1000/sq mi, or (3) at least 25% of the adult male population being employed in agriculture.

Statistical Analysis

Data were analyzed with logistic multilevel modeling procedures,27 the strengths and relevance of which have previously been described.2830 We specified a 4-level model for each binary response y, reporting lifetime or recent IPV or not, for individuals i living in neighborhood j in district k in state l of the form


Probability was related to a set of categorical predictors X and a random effect for each level by a logit-link function as


The linear predictor on the right-hand side of the equation consisted of a fixed part (β0 + βX) estimating the conditional coefficients for the exposure variable (and covariates) and 3 random intercepts attributable to neighborhoods (u0jkl), districts (v0kl ), and states (f0l), with each assumed to have an independent and identical distribution and variance estimated at each level. All models were created by penalized quasi-likelihood approximation with second-order Taylor linearization as implemented in MLwiN version 2.02.31,32

Sixteen percent and 9.3% of women reported lifetime IPV and recent IPV, respectively (Table 1). Nearly half of all women had no formal education, and just over 5% had 13 or more years of education. By contrast, 25% of the respondents’ husbands had no formal education and 11% had 13 or more years of education. Approximately 40% of the women had the same level of education as did their husband, 9% had a higher level of education, and half had a lesser level.

Woman’s Education

The analyses, adjusted for age, revealed an inverse relation between lifetime or recent IPV and a woman’s educational attainment (Table 2). This relation was attenuated substantially after adjustment for covariates but remained statistically significant (P < .001). Women with no formal education were 4.5 times more likely (95% confidence interval [CI] = 3.37, 6.25) to report lifetime IPV and 5.6 times more likely (95% CI = 3.53, 8.92) to report recent IPV compared with those schooled for more than 12 years.

Husband’s Education

Educational gradients in lifetime and recent IPV were also observed when we analyzed husbands’ education. Higher educational levels for husbands were associated with lower odds of lifetime and recent IPV (Table 2). Even after adjustment for covariates and a woman’s own education, we observed that women married to husbands with no formal education were much more likely to report lifetime IPV (odds ratio [OR] = 1.82; 95% CI = 1.50, 2.20) or recent IPV (OR = 1.84; 95% CI = 1.44, 2.35) than were those married to husbands who were college educated (P < .001).

Spousal Education Differential

Spousal education differential was a strong predictor of women’s likelihood of reporting lifetime IPV (P = .002) and recent IPV (P = .01) in models adjusted for women’s and husbands’ education as well as other covariates (Table 2). Women who were more educated than their husbands were more likely to report lifetime (OR = 1.18; 95% CI = 1.07, 1.29) and recent IPV (OR = 1.18; 95% CI = 1.05, 1.29). Women with educational levels lower than their husband’s were not substantially different from the parity group (the reference group).

The strength of the association between women’s education and IPV declined by one fourth between models that adjusted for individual characteristics excluding husband’s education and education differential, and models that adjusted for all individual sociodemographic and education variables as well as contextual education variables.

Neighborhood Literacy

Even after adjustment for woman’s and husband’s education, spousal educational differential, and the covariates, there remained substantial variation in both lifetime and recent IPV at the neighborhood level, suggesting the potential importance of neighborhood-level factors for IPV. Using the latent variable approach to estimate the intraclass correlation,33 we found that 5% (P < .001) of the total variation in lifetime IPV and 6% (P < .001) of the total variation in recent IPV was attributable to neighborhood-level factors. Neighborhood male and female literacy were inversely related to lifetime and recent IPV (Table 3). We observed a graded relation between lifetime IPV and neighborhood male literacy. The odds for lifetime IPV of women living in neighborhoods in the middle and lowest tertiles of male literacy were 1.08 (95% CI=1.00, 1.18) and 1.12 (95% CI=1.01, 1.23) times greater, respectively, than those of women living in the highest-tertile neighborhoods. The odds for lifetime IPV of women living in neighborhoods in the middle and lowest tertiles of female literacy were 1.18 (95% CI = 1.07, 1.29) and 1.10 (95% CI = 0.99, 1.24) times greater, respectively, than those of women living in the highest-tertile neighborhoods. Although the 2 lower tertiles of neighborhood female literacy were associated with higher odds of lifetime IPV, the observed odds were not consistent with a dose–response group relationship. The observed relations between neighborhood literacy and recent IPV were similar to those between neighborhood literacy and lifetime IPV.

We also tested for interactions between neighborhood literacy levels and woman’s education (Figure 1). The tests revealed that the overall protective effect of living in neighborhoods with high literacy levels was stronger for women with higher levels of education (> 6 years). The gradient between woman’s education and risk of lifetime IPV was substantially weaker in low-literacy neighborhoods. In other words, women with lesser levels of education gain very little if anything in terms of lifetime IPV risk by living in high-literacy neighborhoods, but highly educated women appeared to lose much of the protection their education provides if they live in low-literacy neighborhoods.

Our study yielded 4 key findings. First, we found independent effects of a woman’s and husband’s educational attainment on the woman’s likelihood of reporting IPV, with the effects being stronger for the woman’s education but their husbands’ education still capturing about 40% of the woman’s educational effects. Second, women whose educational attainment was higher than their husband’s were more likely to report IPV. Third, neighborhood male and female literacy were inversely associated with IPV, independent of individual-level factors. Finally, neighborhood literacy modified the relation between a woman’s education and IPV so that the protective effects of neighborhood literacy were stronger for women with a middle to high level of education.

The findings concerning the protective effects of a woman’s education against IPV, based as they are on a large nationally representative data set after control for a range of socioeconomic and demographic risk factors, is an important corroboration of previous findings,16,34 because several other studies reported a null association.3538 The link between a woman’s education and risk of IPV can be made through a number of pathways. Because IPV in India is associated with economic disadvantage,16,39 individual education could provide a woman with skills and knowledge that improve her ability to manage a household despite reduced economic resources and thereby decrease her risk of IPV. Furthermore, education could provide a woman with more opportunities for financial independence, allowing her to leave an abusive husband20 and providing her husband with an incentive to refrain from abusing her. Apart from the knowledge that abusing an educated wife could spur her to leave him, a husband may also value and respect an educated wife more than he would an uneducated wife, thus, providing her additional protection from abuse.36

The attenuation in the relation between woman’s education and IPV after adjustment for husband’s education, spousal education differential, and neighborhood literacy indicates the importance of women’s educational context in relation to risk of abuse. Although our finding of an inverse relation between husband’s education and risk of IPV has some precedent in the literature from India,16,17,35,37,39 other researchers have also found a null association34 or even a direct association.38 Qualitative research suggests that men abuse their wives when they do not have socially acceptable ways to vent their frustrations in public.20 Educated men may be more adept at preventing or coping with these frustrations if their education provides them with economic security and social status among their peers. Education may also allow men to avoid other behaviors and conditions, such as substance abuse,35,37 gambling,34 and sexual dysfunction,34 that have been linked with domestic violence. Additionally, because some men use IPV as a means of asserting their authority over women,20 those with less education are more likely to believe that they are justified in controlling the actions of their wives and in using physical force to accomplish this domination.17

Although our finding that women who are more educated than their husbands are at higher risk of abuse corroborates similar findings from other developing countries,40,41 it is unique in India, because previous studies found no relation between spousal education differential and IPV.36,42 Although we recognize the limited effect of educational differential in relation to the additive effects of women’s and husband’s education, we believe that this association highlights the relationships between gender and power that are key to understanding the problem of IPV. One reason that educational differential may be related to IPV risk involves a man’s sense of his own masculinity, which may be derived from perceptions of personal power43 and domination over his wife.20 Thus, a man with less education than his wife may be more likely than other men to commit abuse if he is threatened by any status that his wife’s education may confer on her.44

The large differences in domestic violence victimization trends between states, districts, and neighborhoods, even after accounting for individual sociodemographic characteristics, emphasize the role that community context plays in IPV across India. Our study found neighborhood literacy to be associated with IPV and to modify the association between individual education and IPV. Previous research suggests that violence against women is associated with community attitudes that either are accepting of or indifferent to IPV,16 and that the social norms underlying these attitudes are unevenly distributed across the country’s geography.45 Because education is an important factor in determining the acceptability of abuse,17 a community’s level of education may contribute to a woman’s risk of abuse by influencing a husband to view IPV as either normative or nonnormative behavior. Community education could also act through social norms to influence the institutional resources and interpersonal support that are available to assist abused women.46 In this way, neighborhood literacy could also interact with a woman’s individual education, because even if she is highly educated and able to support herself financially, strong community social norms prohibiting divorce in a low-literacy area may prevent an abused woman from leaving her husband.


The global measure of domestic violence used in the INFHS has been found to be less likely to elicit a report of violence victimization than measures that include multiple behaviorally specific questions about what types of abuse the respondent has experienced.47 This partially accounts for why the prevalence of abuse reported in our study is smaller than that reported in previous research with Indian women.9 In fact, previous research in Nicaragua that used a global measure of IPV similar to the one we used found an inverse relation between education and abuse, and the tool using multiple behaviorally specific questions found no relation, in a similar population.47 Because it is impossible to know whether women in our study were underreporting their experiences of abuse—and if so, which women they were—it is impossible to rule out the possibility that the inverse association between education and IPV has been inflated in this study. However, the cultural norms related to IPV in India, as evidenced by the high rate of IPV9 and the professed level of tolerance of many Indians toward IPV,17 suggest that the stigma against discussing personal experiences of IPV is likely not yet very strong in India. Given this and the very strong association observed between education and IPV in the current study, a differential attitude among sociocultural groups toward reporting IPV is unlikely to account for the observed differences in IPV risk.

Another limitation of this study is the lack of a dose–response group relationship found in the relation between neighborhood female literacy and IPV. Our results indicate that women living in neighborhoods of moderate literacy are at higher risk of IPV than those living in high-literacy or low-literacy neighborhoods. Findings in neighboring Bangladesh suggest that IPV often temporarily increases when women-focused development programs are introduced into a community, because of the threat husbands perceive to their hegemony, before dropping below baseline levels.48,49 Similarly, men who live in neighborhoods in which the female literacy rate has just started to rise may feel threatened and respond with violence.


To our knowledge, our study was the first study of a nationally representative sample to investigate the relation between multiple aspects of the educational context and IPV in India. IPV remains a problem of critical importance in India associated with poor reproductive outcomes,50,51 asthma,52 injury,46 psychological dysfunction,9 suicide,53 and murder.54 Our findings affirm the important independent effects of a woman’s own education and her proximate educational context on IPV, and provide a basis for considering the importance to women’s health of costs and benefits associated with others’ education. Although our study cannot establish a causal relationship, it does suggest that educating girls could provide them with significant protection against IPV as adults, above and beyond the benefits that accrue in terms of earning potential and quality of life. The need to also explicitly consider the role of proximate educational context emphasizes that interventions aimed only at educating women may not be sufficient to make a significant impact on IPV. Interventions aimed at eliminating IPV should pay significant attention to educating men and raising levels of education in communities in general.

TABLE 1— Descriptive Characteristics of Married Women (N = 83 627): Indian National Family Health Survey, 1998–1999
TABLE 1— Descriptive Characteristics of Married Women (N = 83 627): Indian National Family Health Survey, 1998–1999
 Sample, No. (%)aLifetime IPV, No. (%)bRecent IPV, No. (%)c
Total83 627 (100.0)13 460 (16.1)7 749 (9.3)
Location of neighborhoodd
    Large City9 213 (11.0)993 (10.8)556 (6.0)
    Small City5 319 (6.4)689 (13.0)361 (6.8)
    Town11 465 (13.7)1 464 (12.8)776 (6.8)
    Village57 630 (68.9)10 314 (17.9)6 056 (10.5)
Age, y
    45–497 205 (8.6)1 047 (14.5)394 (5.5)
    40–449 742 (11.7)1 584 (16.3)706 (7.3)
    35–3912 640 (15.1)2 193 (17.4)1 139 (9.0)
    30–3414 865 (17.8)2 675 (18.0)1 525 (10.3)
    25–2917 272 (20.7)2 964 (17.2)1 844 (10.7)
    20–2415 316 (18.3)2 249 (14.7)1 539 (10.1)
    15–196 587 (7.9)748 (11.4)602 (9.1)
Age at wedding, y
    ≥ 2113 335 (16.0)1 029 (7.7)562 (4.2)
    18–2023 688 (28.3)2 992 (12.6)1 721 (7.3)
    15–1731 554 (37.7)5 721 (18.1)3 369 (10.7)
    < 1515 050 (18.0)3 718 (24.7)2 097 (13.9)
    Hindu65 167 (77.9)10 721 (16.5)6 153 (9.4)
    Muslim9 921 (11.9)1 769 (17.8)1 018 (10.3)
    Christian4 506 (5.4)497 (11.0)283 (6.3)
    Sikh1 988 (2.4)217 (10.9)127 (6.4)
    Other/missing religion2 045 (2.5)256 (12.5)168 (8.2)
    General35 159 (42.0)3 992 (11.4)2 199 (6.3)
    Scheduled caste14 174 (17.0)3 213 (22.7)1 972 (13.9)
    Scheduled tribe9 967 (11.9)1 669 (16.8)1 009 (10.1)
    Other backward class24 327 (29.1)4 586 (18.9)2 569 (10.6)
    Not working54 159 (64.8)6 919 (12.8)2 199 (6.3)
    Unpaid8 704 (10.4)1 667 (19.2)1 972 (13.9)
    Paid nonmanual3 693 (4.4)375 (10.2)1 009 (10.1)
    Paid agricultural10 603 (12.7)2 870 (27.1)2 569 (10.6)
    Paid manual6 468 (7.7)1 629 (25.2)2 199 (6.3)
Living standard, quintile
    Fifth (highest)17 124 (20.5)987 (5.8)455 (2.7)
    Fourth17 264 (20.6)2 017 (11.7)1 065 (6.2)
    Third16 896 (20.2)2 843 (16.8)1 530 (9.1)
    Second16 676 (20.0)3 517 (21.1)2 101 (12.6)
    First (lowest)15 667 (18.7)4 096 (26.1)2 598 (16.6)
Woman’s education, y
    ≥ 134 374 (5.2)74 (1.7)29 (0.7)
    11–123 678 (4.4)203 (5.5)96 (2.6)
    9–109 838 (11.8)711 (7.2)402 (4.1)
    6–810 520 (12.6)1 273 (12.1)1 239 (6.8)
    1–513 666 (16.3)2 330 (17.1)1 239 (9.1)
    No formal schooling41 551 (49.7)8 869 (21.3)5 265 (12.7)
Husband’s education, y
    ≥ 139 313 (11.1)425 (4.6)226 (2.4)
    11–127 002 (8.4)627 (9.0)334 (4.8)
    9–1017 263 (20.6)2 030 (11.8)1 154 (6.7)
    6–813 745 (16.4)2 196 (16.0)1 258 (9.2)
    1–514 936 (17.9)3 095 (20.7)1 696 (11.4)
    No formal schooling21 368 (25.6)5 087 (23.8)3 081 (14.4)
Spousal education differential
    Same level33 897 (40.5)6 023 (17.8)3 538 (10.4)
    Wife is more educated7 722 (9.2)1 084 (14.0)609 (7.9)
    Husband is more educated42 008 (50.2)6 353 (15.1)3 602 (8.6)
Neighborhood female literacy, tertile
    Highest26 695 (31.9)2 565 (9.6)1 297 (4.9)
    Middle27 792 (33.2)4 852 (17.5)2 721 (9.8)
    Lowest29 140 (34.8)6 043 (20.7)3 731 (12.8)
    Neighborhood male literacy, tertile
    Highest26 970 (32.0)2 804 (10.4)1 441 (5.3)
    Middle28 242 (33.8)4 727 (16.7)2 638 (9.3)
    Lowest28 415 (34.0)5 929 (20.9)3 670 (12.9)

Note. IPV = intimate partner violence.

aNumber and percentage of women with that descriptive characteristic out of the total sample.

bNumber and percentage of women with the characteristic who have experienced IPV since age 15 years.

cNumber and percentage of women with the characteristic who have experienced IPV in the past year.

dA large city was defined as having of more than 1 million people, a small city as having between 100 000 and 1 million people, and a town as an urban area with less than 100 000 people. A village was a rural area that had at least 1 of 3 characteristics: fewer than 5000 residents, population density less than 1000/sq mi, or at least 25% of the adult male population being employed in agriculture.

eScheduled castes are those with members in the lowest level of the caste system. Scheduled tribes are approximately 700 officially recognized social groups characterized by their geographic isolation and limited social and economic interaction with the rest of India. “Other backward class” is a legislatively defined group representing those who have historically suffered significant social deprivation not as severe as that suffered by scheduled castes and tribes.

TABLE 2— Age-Adjusted and Fully Adjusted Models of Lifetime and Recent Intimate Partner Violence, by Sample Education Characteristics: Indian National Family Health Survey, 1998–1999
TABLE 2— Age-Adjusted and Fully Adjusted Models of Lifetime and Recent Intimate Partner Violence, by Sample Education Characteristics: Indian National Family Health Survey, 1998–1999
 Lifetime IPVRecent IPV
 Age Adjusted,a OR (95% CI)Fully Adjusted,b OR (95% CI)Age Adjusted,a OR (95% CI)Fully Adjusted,b OR (95% CI)
Woman’s education, y
    ≥ 13 (Ref)
    11–123.23 (2.43, 4.29)2.59 (1.93, 3.48)3.64 (2.33, 5.69)2.84 (1.80, 4.49)
    9–104.55 (3.52, 5.89)2.94 (2.23, 3.88)6.04 (4.02, 9.06)3.75 (2.45, 5.76)
    6–87.89 (6.13, 10.16)3.91 (2.94, 5.20)10.07 (6.75, 15.03)4.77 (3.08, 7.39)
    1–511.72 (9.13, 15.04)4.60 (3.43, 6.17)14.18 (9.53, 21.10)5.33 (3.41, 8.33)
    No formal schooling15.24 (11.90, 19.51)4.59 (3.37, 6.25)19.59 (13.21, 29.06)5.61 (3.53, 8.92)
Husband’s education, y
    ≥ 13 (Ref)
    11–122.02 (1.76, 2.31)1.27 (1.10, 1.47)1.91 (1.59, 2.29)1.18 (0.97, 1.42)
    9–102.96 (2.64, 3.33)1.47 (1.29, 1.68)2.93 (2.51, 3.43)1.43 (1.20, 1.69)
    6–84.15 (3.70, 4.66)1.64 (1.43, 1.88)4.08 (3.49, 4.76)1.58 (1.32, 1.90)
    1–55.46 (4.87, 6.12)1.83 (1.57, 2.12)5.19 (4.45, 6.05)1.72 (1.42, 2.09)
    No formal schooling6.12 (5.47, 6.85)1.82 (1.50, 2.20)6.15 (5.29, 7.15)1.84 (1.44, 2.35)
Spouse education differential
    Same level (Ref)
    Wife is more educated0.91 (0.84, 0.98)1.18 (1.07, 1.29)0.89 (0.80, 0.98)1.18 (1.05, 1.33)
    Husband is more educated0.84 (0.81, 0.88)1.01 (0.93, 1.10)0.82 (0.78, 0.86)1.04 (0.93, 1.17)

Note. IPV = intimate partner violence; OR = odds ratio; CI = confidence interval.

aEach age-adjusted model included age in years and 1 education variable: woman’s education, husband’s education, or spousal education differential.

bModels included location, current age in years, age at wedding, religion, caste, employment, living standard, woman’s education, husband’s education, and spousal education differential.

TABLE 3— Odds Ratios (ORs; with 95% Confidence Intervals [CIs]) for Reported Intimate Partner Violence and Neighborhood Literacy: Indian National Family Health Survey, 1998–1999
TABLE 3— Odds Ratios (ORs; with 95% Confidence Intervals [CIs]) for Reported Intimate Partner Violence and Neighborhood Literacy: Indian National Family Health Survey, 1998–1999
 Lifetime IPV, OR (95% CI)Recent IPV, OR (95% CI)
Neighborhood female literacy, tertile, %
    Highest (Ref)1.001.00
    Middle1.18 (1.07, 1.29)1.20 (1.07, 1.35)
    Lowest1.10 (0.99, 1.24)1.14 (0.99, 1.31)
Neighborhood male literacy, tertile, %
    Highest (Ref)1.001.00
    Middle1.08 (1.00, 1.18)1.07 (0.96, 1.19)
    Lowest1.12 (1.01, 1.23)1.14 (1.01, 1.28)

Note. IPV = intimate partner violence. One contextual education variable was modeled at a time. Models include location, current age in years, age at wedding, religion, caste, employment, living standard, woman’s education, husband’s education, and spousal education differential.

Subramanian is supported by the National Institutes of Health (award NHLBI-1-K25-HL081275).

We acknowledge the support of Macro International Inc, Washington, DC, for providing access to the 1998 to 1999 Indian National Family Health Survey data.


1. Garcia-Moreno C, Jansen HA, Ellsberg M, Heise L, Watts CH. Prevalence of intimate partner violence: findings from the WHO multicountry study on women’s health and domestic violence. Lancet. 2006;368:1260–1269. Crossref, MedlineGoogle Scholar
2. Heise L, Garcia-Moreno C. Violence by intimate partners. In: Krug EG, Dahlberg LL, Mercy JA, Zwi AB, Lozano R, eds. World Report on Violence and Health. Geneva, Switzerland: World Health Organization; 2002. Google Scholar
3. Watts C, Zimmerman C. Violence against women: global scope and magnitude. Lancet. 2002;359:1232–1237. Crossref, MedlineGoogle Scholar
4. Campbell JC. Health consequences of intimate partner violence. Lancet. 2002;359:1331–1336. Crossref, MedlineGoogle Scholar
5. Sen A. Missing women. BMJ. 1992;304(6827):587–588. Crossref, MedlineGoogle Scholar
6. Basu AM. Culture, the Status of Women, and Demographic Behaviour: Illustrated With the Case of India. Oxford, England: Oxford University Press; 1992. Google Scholar
7. Tinker I. Persistent Inequalities: Women and World Development. New York, NY: Oxford University Press; 1990. Google Scholar
8. Das Gupta M, Chen LC, Krishnan TN (Eds). Women’s Health in India: Risk and Vulnerability. Bombay, India: Oxford University Press; 1995. Google Scholar
9. Kumar S, Jeyaseelan L, Suresh S, Ahuja RC. Domestic violence and its mental health correlates in Indian women. Br J Psychiatry. 2005;187:62–67. Crossref, MedlineGoogle Scholar
10. Sampson RJ, Groves WB. Community structure and crime: testing social disorganization theory. Am J Sociol. 1989;94:774–802. CrossrefGoogle Scholar
11. Sampson RJ, Raudenbush SW, Earls F. Neighborhoods and violent crime: a multilevel study of collective efficacy. Sci. 1997;277:918–924. Crossref, MedlineGoogle Scholar
12. Cunradi CB, Caetano R, Clark C, Schafer J. Neighborhood poverty as a predictor of intimate partner violence among White, Black, and Hispanic couples in the United States: a multilevel analysis. Ann Epidemiol. 2000;10:297–308. Crossref, MedlineGoogle Scholar
13. Miles-Doan R, Kelly S. Geographic concentration of violence between intimate partners. Public Health Rep. 1997;112:135–141. MedlineGoogle Scholar
14. O’Campo P, Gielen AC, Faden RR, Xue X, Kass N, Wang MC. Violence by male partners against women during the childbearing year: a contextual analysis. Am J Public Health. 1995;85:1092–1097. LinkGoogle Scholar
15. Benson ML, Fox GL, DeMaris A, Van Wyk J. Neighborhood disadvantage, individual economic distress and violence against women in intimate relationships. J Quant Criminol. 2003;19:207–235. CrossrefGoogle Scholar
16. Koenig MA, Stephenson R, Ahmed S, Jejeebhoy SJ, Campbell J. Individual and contextual determinants of domestic violence in North India. Am J Public Health. 2006;96:132–138. LinkGoogle Scholar
17. Martin SL, Moracco KE, Garro J, et al. Domestic violence across generations: findings from northern India. Int J Epidemiol. 2002;31:560–572. Crossref, MedlineGoogle Scholar
18. Go VF, Johnson SC, Bentley ME, et al. Crossing the threshold: engendered definitions of socially acceptable domestic violence in Chennai, India. Cult Health Sex. 2003;5:393–408. CrossrefGoogle Scholar
19. Nayak MB, Byrne CA, Martin MK, Abraham AG. Attitudes toward violence against women: a cross-nation study. Sex Roles. 2003;49(7–8):333–342. CrossrefGoogle Scholar
20. Krishnan S. Do structural inequalities contribute to marital violence? Ethnographic evidence from rural South India. Violence Against Women. 2005;11:759–775. Crossref, MedlineGoogle Scholar
21. International Institute for Population Sciences, ORC Macro. National Family Health Survey, 1998–1999. Mumbai, India: International Institute for Population Sciences; 2000. Google Scholar
22. International Institute for Population Sciences. 1998–1999 Indian National Family Health Survey Interviewer’s Manual. Mumbai, India: International Institute for Population Sciences; 1998. Google Scholar
23. Subramanian SV, Nandy S, Irving M, Gordon D, Lambert H, Davey Smith G. The mortality divide in India: the differential contributions of gender, caste, and standard of living across the life course. Am J Public Health. 2006;96:818–825. LinkGoogle Scholar
24. Subramanian SV, Davey Smith G, Subramanyam M. Indigenous health and socioeconomic status in India. PLoS Med. 2006;3(10):e421. Crossref, MedlineGoogle Scholar
25. Filmer D, Pritchett LH. Estimating wealth effects without expenditure data–or tears: an application to educational enrollments in states of India. Demography 2001;38:115–132. MedlineGoogle Scholar
26. Rutstein SO, Johnson K The DHS Wealth Index: DHS Comparative Reports 6. Calverton, Md: ORC Macro; 2004. Google Scholar
27. Goldstein H. Multilevel Statistical Models. London, England: Arnold; 2003. Google Scholar
28. Blakely T, Subramanian SV. Multilevel studies. In: Oakes JM, Kaufman JS, eds. Methods in Social Epidemiology. San Fransisco, Calif: Jossey-Bass; 2006:316–340. Google Scholar
29. Subramanian SV. The relevance of multilevel statistical methods for identifying causal neighborhood effects. Soc Sci Med. 2004;58:1961–1967. Crossref, MedlineGoogle Scholar
30. Subramanian SV, Jones K, Duncan C. Multilevel methods for public health research. In: Kawachi I, Berkman LF, eds. Neighborhoods and Health. New York, NY: Oxford University Press; 2003:65–111. Google Scholar
31. Goldstein H, Rasbash J. Improved approximations for multilevel models with binary responses. J Royal Stat Soc. Series A. 1996;159:505–513. CrossrefGoogle Scholar
32. Rasbash J, Steele F, Browne W, Prosser B. A User’s Guide to MLwiN, Version 2.0. Bristol, England: Centre for Multilevel Modeling; 2005. Google Scholar
33. Goldstein H, Browne W, Rasbash J. Partitioning variation in generalised linear multilevel models. Underst Stat. 2002;1:223–232. CrossrefGoogle Scholar
34. Verma RK, Collumbien M. Wife beating and the link with poor sexual health and risk behavior among men in urban slums in India. J Comp Fam Stud. 2003;34:61–74. Google Scholar
35. Khosla AH, Dua D, Devi L, Sud SS. Domestic violence in pregnancy in North Indian women. Indian J Med Sci. 2005;59:195–199. Crossref, MedlineGoogle Scholar
36. Panda P, Agarwal B. Marital violence, human development and women’s property status in India. World Dev. 2005;33:823–850. CrossrefGoogle Scholar
37. Peedicayil A, Sadowski LS, Jeyaseelan L, et al. Spousal physical violence against women during pregnancy. BJOG. 2004;111:682–687. Crossref, MedlineGoogle Scholar
38. Rao V. Wife-beating in rural south India: a qualitative and econometric analysis. Soc Sci Med. 1997;44:1169–1180. Crossref, MedlineGoogle Scholar
39. Martin SL, Tsui AO, Maitra K, Marinshaw R. Domestic violence in northern India. Am J Epidemiol. 1999;150:417–426. Crossref, MedlineGoogle Scholar
40. Burazeri G, Roshi E, Jewkes R, Jordan S, Bjegovic V, Laaser U. Factors associated with spousal physical violence in Albania: cross sectional study. BMJ. 2005;331:197–201. Crossref, MedlineGoogle Scholar
41. Flake DF. Individual, family, and community risk markers for domestic violence in Peru. Violence Against Women. 2005;11:353–373. Crossref, MedlineGoogle Scholar
42. International Center for Research on Women, Centre for Development and Population Activities. Domestic Violence in India: A Summary Report of a Multi-Site Household Survey. Washington, DC: International Center for Research on Women; 2000. Google Scholar
43. Jewkes R. Intimate partner violence: causes and prevention. Lancet. 2002;359:1423–1429. Crossref, MedlineGoogle Scholar
44. Krishnan S. Gender, caste, and economic inequalities and marital violence in rural South India. Health Care Women Int. 2005;26:87–99. Crossref, MedlineGoogle Scholar
45. Dyson T, Moore M. On kinship stucture, female autonmy, and demographic behavior in India. Popul Dev Rev. 1983;9:35–60. CrossrefGoogle Scholar
46. Panchanadeswaran S, Koverola C. The voices of battered women in India. Violence Against Women. 2005;11:736–758. Crossref, MedlineGoogle Scholar
47. Ellsberg M, Heise L, Pena R, Agurto S, Winkvist A. Researching domestic violence against women: methodological and ethical considerations. Stud Fam Plann. 2001;32:1–16. Crossref, MedlineGoogle Scholar
48. Ahmed SM. Intimate partner violence against women: experiences from a woman-focused development programme in Matlab, Bangladesh. J Health Popul Nutr. 2005;23:95–101. MedlineGoogle Scholar
49. Schuler SR, Hashemi SM, Badal SH. Men’s violence against women in rural Bangladesh: undermined or exacerbated by microcredit programmes? Dev Pract. 1998;8:148–157. Crossref, MedlineGoogle Scholar
50. Jejeebhoy SJ. Associations between wife-beating and fetal and infant death: impressions from a survey in rural India. Stud Fam Plann. 1998;29:300–308. Crossref, MedlineGoogle Scholar
51. Martin SL, Kilgallen B, Tsui AO, Maitra K, Singh KK, Kupper LL. Sexual behaviors and reproductive health outcomes: associations with wife abuse in India. JAMA. 1999;282:1967–1972. Crossref, MedlineGoogle Scholar
52. Subramanian SV, Ackerson LK, Subramanyam MA, Wright RJ. Domestic violence is associated with adult and childhood asthma prevalence in India. Int J Epidemiol. 2007;36:569–579. Crossref, MedlineGoogle Scholar
53. Vizcarra B, Hassan F, Hunter WM, Munoz SR, Ramiro L, De Paula CS. Partner violence as a risk factor for mental health among women from communities in the Philippines, Egypt, Chile, and India. Inj Control Saf Promot. 2004;11:125–129. Crossref, MedlineGoogle Scholar
54. Shaha KK, Mohanthy S. Alleged dowry death: a study of homicidal burns. Med Sci Law. 2006;46:105–110. Crossref, MedlineGoogle Scholar


No related items




Leland K. Ackerson, ScD, MPH, Ichiro Kawachi, MD, PhD, Elizabeth M. Barbeau, ScD, MPH, and S.V. Subramanian, PhDLeland K. Ackerson, Ichiro Kawachi, and S.V. Subramanian are with the Department of Society, Human Development, and Health, Harvard School of Public Health, Boston, Mass. Elizabeth Barbeau is with the Department of Society, Human Development, and Health, Harvard School of Public Health, Boston, and the Dana-Farber Cancer Institute, Boston. “Effects of Individual and Proximate Educational Context on Intimate Partner Violence: A Population-Based Study of Women in India”, American Journal of Public Health 98, no. 3 (March 1, 2008): pp. 507-514.


PMID: 18235066