Community Influences on Intimate Partner Violence in India - Women's Education, Attitudes Towards Mistreatment and Standards of Living

Publish in



Please download to get full document.

View again

of 7
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
India - education to work towards reducing intimate partner violence
  Community influences on intimate partner violence in India: Women’s education,attitudes towards mistreatment and standards of living q Michael H. Boyle * , Katholiki Georgiades, John Cullen, Yvonne Racine McMcaster University Hamilton, Ont. Canada a r t i c l e i n f o  Article history: Available online 18 July 2009 Keywords: IndiaIntimate partner violenceEducationClusteringSurveys a b s t r a c t Intimate partner violence (IPV) directed towards women is a serious public health problem. Women’seducation mayoffer protection against IPV, but uncertainty exists over how it might reduce risk for IPV atthe communityand individual levels. The objectives of this study are to: (1) disentangle community fromindividual-level influences of women’s education on risk for IPV; (2) quantify the moderating influenceof communities on individual-level associations between women’s education and IPV; (3) determine if women’s attitudes towards mistreatment and living standards at the community and individual levelsaccount for the protective influence of women’s education; and (4) determine if the protective influenceof education against IPV is muted among women living in communities exhibiting attitudes moreaccepting of mistreatment.Study information came from 68,466 married female participants in the National Family Health Surveyconducted throughout India in 1998–1999. Multilevel logistic regression was used to address the studyobjectives. IPV showed substantial clustering at both the state (10.2%) and community levels (11.5%). Atthe individual level, there was a strong non-linear association between women’s education and IPV,partially accounted for by household living standards. The strength of association between women’seducation and IPV varied from one community to the next with evidence that the acceptance of mistreatment at the community level mutes the protective influence of higher education. Furthermore,women’s attitudes towards mistreatment and their standards of living accounted for community-levelassociations between women’s education and IPV.Place of residence accounted for substantial variation in risk of IPV and also modified individual-levelassociations between IPV and women’s education. At the community level, women’s education appearedto exert much of its protective influence by altering population attitudes towards the acceptability of mistreatment. However, there was no residual association between women’s education and IPV at thecommunity level once living standards are taken into account. While women’s education providesstrong, independent leverage for reducing the risk of IPV, planners must keep in mind importantcommunity factors that modify its protective influence.   2009 Elsevier Ltd. All rights reserved. Intimate partner violence (IPV) is the intentional use of physicalforce (beatings, rape) to inflict harm on a spouse or partner. It isa widespread public health problem with serious consequences(Ellsberg, 2006). A recent investigation of Women’s Health andDomestic Violence against Women estimated the lifetime preva-lence of IPV directed towards women to vary from 15 to 71% in 15sites from 10 countries (Garcia-Moreno, Jansen, Ellsberg, Heise, &Watts, 2005). The physical and mental health sequelae of IPV directed towards women include increased mortality, injury anddisability, worse general health, chronic pain, reproductive disor-ders, depression, PTSD, alcohol abuse and drug abuse (Golding,1999; Plichta, 2004). Given that IPV is a product of social context( Jewkes, 2002), it is not at all clear that traditional medicalapproaches such as individual screening and intervention (Mac-Millan et al., 2006) will representeffective orefficient strategies forreducing IPV. An alternative is to focus on the social determinantsof IPV with a view to identifying modifiable characteristics forprevention. q Michael Boyle is supported by a Canada Research Chair in the Social Determi-nants of Child Health. Katholiki Georgiades is supported by an Ontario MentalHealth Foundation New Investigator Fellowship. The authors thank Jon Rasbash forhis helpful comments on the manuscript. *  Corresponding author. Department of Psychiatry and Behavioural Neurosci-ences & Offord Centre for Child Studies, Hamilton Health Sciences, McMcasterUniversity, Chedoke Site, Central Building 303, Hamilton, Ont., Canada L8N 3Z5.Tel.:  þ 1 905 521 2100; fax:  þ 1 905 521 4970. E-mail address: (M.H. Boyle). Contents lists available at ScienceDirect Social Science & Medicine journal homepage: 0277-9536/$ – see front matter    2009 Elsevier Ltd. All rights reserved.doi:10.1016/j.socscimed.2009.06.039 Social Science & Medicine 69 (2009) 691–697  Violence directed towards women in India has attracted specialattention (Sharma, 2005). Historically, patriarchal prerogativeshave dominated family relationships in that country and culturallysanctioned use of physical violence by men against women hasbeen condoned as a mechanism of control and conflict resolution(Segal, 1999; WHO, 2002). In addition, socio-economic influencesthat elevate risk for IPV directed towards women – poverty(Bangdiwala et al., 2004; Hindin & Adair, 2002; Jewkes, 2002;Koenig, Stephenson, Ahmed, Jejeebhoy, & Campbell, 2006), youngage at marriage, low education (Bangdiwala et al., 2004; Gage,2005; Jewkes, 2002; McQuestion, 2003) and multiparity(McQuestion, 2003) – have been endemic to India in the past.India, however, contains substantial diversity in language,culture, socio-economic well being and demography (Dyson &Moore,1983; Haub & Sharma, 2006). These factors can be expectedto influence normative expectations for gender roles, relationsbetween men and women and the acceptability of using physicalviolence, providing a basis for area-patterning of IPV. Althoughmostly rural, Indiaisalsoan emergingeconomic powerandsubjectto forces of modernity, including increased levels of literacy andeducation for both men and women. These forces are most likelytobe concentrated in selected geographic areas and also contribute tospatial differentiation in the risk for IPV. Women’s education in India and risk for IPV  The second National Family Health Survey (NFHS-2) conductedthroughout India in 1998–1999 (International Institute for Pop-ulation Sciences (IIPS) and ORC Macro, 2000) reported a strongnegative gradient between educational attainment for marriedwomen and their reports of being beaten or physically mistreatedinthepast12months(recentIPV):illiterate,14.1%; < middleschool,8.8%; middle school complete, 7.0%; and high school complete andabove 3.6%. In a recent study of the same data set, these gradedindividual-level effects persisted after adjustment for a number of covariates (Ackerson, Kawachi, Barbeau, & Subramanian, 2008).Demonstrating contextual-level effects for women’s educationaggregatedtothearealevelmustexplainbetween-areavariationinIPV after controlling for potential confounding variables, includingindividual education. The same report by Ackerson et al. (2008)estimated that 6% of the total variation in recent IPV was attribut-able to male and female literacy aggregated to the neighbourhoodlevel. In contrast, two studies, one in Bangladesh (Koenig, Ahmed,Hossain, & Mozumber, 2003) and the other in north India (Koeniget al., 2006) found no empirical evidence to support contextual-level influence of women’s education on IPV. These negative find-ings could be due to the lack of IPV clustering – no statisticallysignificant place-to-place variations in IPV were reported – and/orthe dominance of other individual and community-level variablesincluded in the regression models. Mechanisms of effect at individual and community levels Women’s education is at the intersection of many influences.Educational opportunities for girls will be subject to: 1) parentalresources, beliefs and attitudes; 2) accessibility of schools andeducational resources; and 3) community norms and expectations.Education can be expected to have an indirect effect on the risk of IPV by influencing the beliefs, self-image and capability of youngwomen. This should lead to changes inwomen’s attitudes towards,andacceptanceof,physicalmistreatmentasameansofsubjugationand conflict resolution: reducing individual tolerance for suchbehavior should lower risk of exposure. Of course, the by-productsof education for women go beyond changes in attitudes; theyinclude higher living standards that come from increases inpersonal capability and opportunities for employment, as well asimproved socio-economic circumstances and lower risk of IPV through the process of assortive mating.At a theoretical level, there are reasons to believe that women’seducational attainment aggregated to the community level mightexert protective influences on women’s risk for IPV. Highereducation is aligned with moreliberal norms and values pertainingto women’s rights and less acceptance of violence as a means of resolving conflicts. As more and more women in a particular areaare exposed to higher education, the percent of the populationembracing liberal views about gender equality and opportunityshould increase. The increasing presence of these norms, valuesand attitudes may take on a collective influence, loweringcommunity tolerance for physical mistreatment and perhapsleading tosanctions against suchbehavior. This collective influencewould be based on processes associated with social learning andimitation and emerge from repeated social interactions andexchanges that come to define the parameters of acceptablebehavior (Kravdal, 2004).In addition to the increased presence of liberal views aboutgender equality and opportunity, we can expect increases inwomen’s education at the community level to be associated withimproved living standards, population health, public infrastructureand institutional capacity for serving the public good. It is possiblethat the effects of women’s education on risk for IPV are indistin-guishable from those associated with community socio-economiccircumstances. Indeed, areas with higher living standards mayprovide greater opportunities for women’s education. Demon-strating an independent effect for women’s education aggregatedto the community level would provide compelling evidence for itsprotective influence. Community-level modification of individual-level effects  Just as communities mayexert a direct influence on risk for IPV,they may also mute or augment risk or protection operating at theindividual level. A person-environment fit framework lies behindthese community-level influences that become manifest in cross-level interactions. Serving to illustrate this is the reported interac-tion between women’s education and neighbourhood literacy inthe study by Ackerson et al. (2008). The protective effects of education were strongest for educated women ( > 6 years) living inhigh literacy neighbourhoods but weaker for womenwith levels of education discordant with their neighbourhood. In the presentstudy, we associate higher levels of educational attainment withincreased knowledge, an enhanced capacity to access and to useinformation, more autonomy and more liberal ideas about thestatus of women ( Jewkes, 2002). The influence of these individual-level characteristics on IPV could be muted or even reversed incircumstances where regressive ideas about the status and role of women are dominant (Sugarman & Frankel, 1996). Objectives This study examines area-patterning of IPV in India witha special focus on women’s education, attitudes towards mistreat-ment and living standards assessed at the community level andindividual levels. The objectives of the study are to: (1) disentanglecommunity from individual-level influences of women’s educationon risk for IPV; (2) quantify the moderating influence of commu-nities on individual-level associations betweenwomen’s educationand IPV; (3) determine if women’s attitudes towards mistreatmentand living standards at the community and individual levelsaccount for community-level associations between IPV and wom-en’s education; and (4) determine if the protective influence of  M.H. Boyle et al. / Social Science & Medicine 69 (2009) 691–697  692  education against IPV is muted among women living in areasexhibiting attitudes more accepting of mistreatment. The studybuilds on the work of  Ackerson et al. (2008) by extending ourunderstanding of the factors at the community and individuallevels associated with educationwhich might explain its protectiveinfluence on IPV. Methods In the NFHS-2, female interviewers used standard surveyquestionnaires administered face-to-face to collect health-relatedinformation from a nationally representative probability sample of womenaged15–49yearslivinginhouseholddwellings.Thesurveyused a stratified, multi-stage, cluster design based on the 1991Census. Each state was divided into urban and rural areas. In urbanareas,wardslistedinthe1991Censuswerestratifiedbydistrictandfemale literacy, and a sample of wards was selected with proba-bilityproportionaltosize(PPS)(stage1).Next,selectedwardsweredivided into census enumeration blocks (CEBs) of about 150–200households each and one CEB was taken from each ward with PPS(stage 2). After listing households within selected CEBs, about 30households were chosen with equal probability for enlistment inthe study (stage 3).In rural areas, villages listed in the 1991 Census were stratifiedby district and several demographic features. Contiguous villageswith less than 50 households were linked to form primarysampling units (PSUs) with more than 50 households, while largervillages, with more than 500 households, were segmentedinto smaller PSUs. Next, villages and PSUs were selected with PPS(stage1).AfterlistinghouseholdswithinselectedvillagesandPSUs,about 30 households were selected with equal probability forenlistment in the study (stage 2). To our knowledge, none of thespatial areas selected for the NFHS-2 were contiguous giving uslittle reason to be concerned about spatial autocorrelation. Sample for analysis A total of 91,196 households were enlisted (97.5% response) and89,199 women completed an interview (95.5% response). To beeligible for the analysis, women had to be usual residents in thehousehold as well as married and living with their spouse at thetime of interview ( N  ¼ 79,160). In households with more than onewoman, one was selected at random to prevent household clus-tering from inflating area estimates ( N  ¼ 69,750). (There were toofew women in households to model household clustering asa separate level.) Also, women had to have complete informationon the study variables to be included ( N  ¼ 69,191) and areas withfewer than 10 eligible women in the study were dropped from theanalysis to reduce statistical overlap between variables measuredonindividualwomenandthesamevariablesaggregatedtothearealevel ( N  ¼ 68,466). Concepts and measures Intimate Partner Violence (IPV) Three questions taken from the Status of Women module wereused to classify IPV in the last 12 months. The stem question read,‘Since you completed 15 years of age, have you been beaten ormistreatedphysically byany person?’Womenresponding yeswereasked to identify their relationship to all persons responsible forsuch acts and then to report the aggregate frequency of occurrencein the last 12 months in three categories: once, a few times andmany times. Women answering yes to the stem question,identifying their husband as the perpetrator and reporting anoccurrenceinthepast12monthswereclassifiedaspositiveforIPV. Independent variables These included women’s educational level, household standardof living and women’s attitudes towards acceptance of partnermistreatment. Woman’s educational level was measured by thetotal number of years in school as reported by the woman.Household standard of living was measured by a cumulativeweighted index of 30 durable goods developed to assess the stan-dard of living for households participating in the NFHS-2. The scalescores were converted to a standard score with mean 0.0 andstandard deviation of 1.0. Coefficient alpha for the scale is 0.85. Theapproach is similar to that used by Filmer and Pritchett (1999,2001) to create the asset index which performs as well asconsumption expenditures (Bollen, Glanville, & Stecklov, 2001;Houweling, Kunst, & Mackenbach, 2003).Acceptance of partner mistreatment was measured by a six-itemindex comprised of binary indicators coded  0  (absent) versus 1  (present) for a woman who agreed that a husband is justified inbeatinghiswifeunderoneormoreofthefollowingcircumstances:1,if he suspects her of being unfaithful; 2, if her natal family does notgive expected money, jewellery, or other items; 3, if she showsdisrespectfor in-laws;4,ifshe goesoutwithouttellinghim; 5,ifsheneglects the house or children; and 6, if she doesn’t cook food prop-erly. This index is a linear combination (the largest principalcomponent) of the 6 indicator variables weighted by component-scoreweightsthatmaximallyaccountfortheirco-variation.Thefirstprincipal component accounted for 52% of the variance and coeffi-cient alpha for the scale was 0.82. Control variables The individual controlvariables included: woman’s ageinyears,family structure, total number of children, working outside thehome,exposuretophysicalmistreatment(otherthanIPV)sinceage15 years and urban residency. The stem question used to code IPV provided the basis to classify exposure to violence other than IPV.Women were classified positive for exposure since age 15 if theyidentified one or more perpetrators from the following groups:immediate family (mother, father, step mother, step father, son,daughter, brother or sister); extended family (mother-in-law,father-in-law, son-in-law, daughter-in-law, brother-in-law, sister-in-law, other relative); and other (friend, acquaintance, teacher,employer, stranger). Family structure was coded  0  (extended)versus  1  (nuclear). Extended families included the presence of parents or in-laws, whereas nuclear families consisted of a couplewith or without children in the home. Urban residency coded  1 (yes) versus  0  (no) was assessed at the community level. Community measures Because India has relatively few states ( n ¼ 26) in contrast tolarge numbers of small areas or communities (PSUs in this study),we focus on the latter for estimating specific area-level influences.The community-area measures included: women’s education,household standard of living and attitudes acceptant of partnermistreatment. These community-level measures were derived bysummingvaluesobtainedonindividualwomenineachcommunityand dividing by the total number of women respondents living ineach one. The aggregation of individual-level measures to repre-sent group characteristics is recommended for testing ecologicaland individual-level hypotheses in multilevel studies (Blakely &Woodward,2000).Onaverage,theseecologicalmeasuresarebased M.H. Boyle et al. / Social Science & Medicine 69 (2009) 691–697   693  on 22 womenper community (from 10 to59) – a sufficient numberto generate reliable estimates.  Analysis This study uses multilevel logistic regression analysis to modelIPV reported by women. In logistic regression analysis, the proba-bility of response (  p ) is converted to an odds:  p /(1-  p ); and thentransformed to a logit (  p ): ln [  p /(1-  p )] so that expected responsevalues can be expressed as a linear function of the explanatoryvariables in the logit scale. Multilevel logistic regression analysis isused to model binary response variables that are nested hierar-chically. In this study, reports of IPV are nested in  i  women, from  j communities, located in  k  states. Multilevel analysis provides thebasis to model correlated responses by partitioning residual errorsassociated with each level in the hierarchy and expressing theseerrors as patterns of variation (random effects variances). Theserandom effects variances are used to estimate intra-class correla-tion coefficients (ICC) which quantify the extent of clustering ata particular level.The  b  coefficients in multilevel logistic regression analysis arecalled ‘fixed effects’ and quantify the strength of associationbetween dependent and independent variables. Fixed effectsregressions modelled at one level (e.g., individual) can be allowedto vary randomly at higher levels, in a way that is analogous togeneratingseparateregressionsforeachgroupandthenestimatingtheir variability. For example, specifying the regression of IPV onwoman’s education as a random effect at the state level wouldproduce an estimate of how much the  b  coefficients for women’seducation vary from one state to another. The multi-level modelused in the analyses is depicted below: Logit   p ijk   ¼  X  0 ijk b þ Y  0 ijk w  jk þ n k þ w  jk ; þ 3 ijk : p ijk  is the probability of woman  i , from community  j , in state  k ,reporting IPV.  X  0 ijk  is the vector of covariates (fixed effects) corre-sponding to measures taken at the woman and community levels. Y  0 ijk  is a subset of covariates selected from  X  0 ijk  at the woman leveland allowed to vary randomly at the community level.  n k ,  w  jk , and 3 ; ijk  are random effects intercepts – unexplained residual variationat the state, community and woman-level, respectively.All estimates were derived using second order penalized quasi-likelihood (PQL) and iterative generalized least-squares estimationin  MLwiN   (Rasbash, Steele, Browne, & Prosser, 2004). Residualvariation at level 1 ( 3 ijk ) is assumed to have a standard logisticdistribution with mean zero and variance  p  2  /  3 ¼ 3.29. At subse-quent levels (communities,  w  jk ; states,  n k ), the ICC is given by theestimated residual variation at each level divided by total residualvariation. In these models, the interpretation of fixed effects esti-mates or the  b s  (e.g., the regression of IPV on women’s educationconverted from logarithms of the odds ratios to odds ratios) is thesame as it would be in ordinary logistic regression.Objective one is addressed by estimating IPV clustering at thestate and community levels from the multilevel null model andthen determining the unadjusted fixed effects regression coeffi-cients of IPV on women’s levels of education assessed individuallyand aggregated to the community level. Objective two is addressedby allowing the regression of IPV on women’s education to varyrandomly at both the state and community levels (randomregression coefficients: model 1). Objective three is addressed bycontrolling for women’s attitudes towards mistreatment and livingstandards assessed at the community and individual levels. This isdoneintwostepstoseparateouttheexplanatoryeffectofwomen’sattitudes towards mistreatment from their living standards.Objective four is addressed by including a cross-level interactiontermbetweenwomen’sattitudestowardsmistreatmentassessedatthecommunitylevelandwomen’s education assessed individually.In preparation for the analysis, we centered the following vari-ables (subtracted the overall means from their observed values):women’s education at the community and individual levels;women’s age in years and total number of children. Rescalingprovides morerealistic modelled estimates for theaverage levels of IPV and reduces problems of multi-collinearity when testing forstatistical interactions. Results Table 1 presents summary information on the total sample.There are 26 states, 3118 areas (clusters) and 68,466 women. Thereis substantial variation in the prevalence of IPV between states (1.5to 19.0%) and between clusters (0.0 to 61.5%). About 31.3% of women live in urban areas and 51.4% live in nuclear families. Onaverage, women have 3.90 years of education and 9.7% of thesample reported exposure to IPV in the last 12 months.Correlations among the study variables are shown inTable 2. Atthe area level, women’s education, living standards and urbanresidence are associated negatively with acceptance of mistreat-ment. Cross-level associations between analogous variables (e.g.,women’s education assessed individually and aggregated to thearea level,  r  ¼ 0.664) are strong. Women’s education and standardsof living assessed individually exhibit the strongest associationswith IPV ( r  ¼ 0.139 and  r  ¼ 0.173, respectively).Table 3 presents the multilevel null model of IPV in column one.IPV shows evidence of variability (clustering) at both the state(0.426) and community (0.481) levels – equivalent to about 10.2and 11.5 percent of total variance, respectively. In model 1, unad- justed associations between IPV and women’s education assessedat the community and individual levels are statistically significant,strong and negative in direction. The significant quadratic term(OR  ¼ 0.989) indicates that the reduction in risk of IPV is propor-tionally stronger at higher levels of women’s education (non-lineareffect). The random regression coefficients for the linear compo-nent of women’s education are statistically significant at the state(0.0013) and community levels (0.0046): these coefficients  Table 1 Sample characteristics.Level of analysis  M   (SD)Variables %Level 3, states ( n ¼ 26)Number of communities  M   (SD) 119.9 (73.2)% prevalence IPV (min/max) 8.5 (1.5/19.0)Level 2, communities ( n ¼ 3118)Number of women  M   (SD) 22.0 (6.9)% prevalence IPV (min/max) 9.5 (0.0/61.5)Women’s education  M   (SD) 3.9 (3.2)Standard of living  M   (SD) 0.01 (0.68)Acceptance of mistreatment  M   (SD) 0.01 (0.60)% urban residence 31.3Level 1, women ( n ¼ 68,466)Education in years  M   (SD) 3.9 (4.7)Standard of living  M   (SD) 0.00 (1.0)Acceptance of mistreatment  M   (SD) 0.00 (1.0)% exposed to mistreatment not IPV 4.9Age in years  M   (SD) 31.8 (8.3)% nuclear family 51.4Total number of children  M   (SD) 3.1 (2.1)% working 35.0% exposed to IPV 9.7 n ¼ sample size;  M  ¼ mean; SD ¼ standard deviation. M.H. Boyle et al. / Social Science & Medicine 69 (2009) 691–697  694
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks