A Multilevel Analysis of Partner Violence

 描述（由申请人提供）：在全球范围内，关于亲密伴侣暴力的研究是基于妇女报告的接触情况，对男子实施暴力的决定因素知之甚少。在较贫困环境中的初步证据表明，童年经历和当地性别规范可能会协同影响男子的犯罪风险，但这些风险因素尚未一起研究。在孟加拉国这样的重男轻女的历史环境中，年长男子关于两性关系中的平等-不平等和暴力侵害妇女行为的规范可能会影响到社会地位较低的成员，包括年轻男子的规范。与年长的男性相比，年轻男性更有可能在婚姻中建立行为模式，他们的行为，包括IPV的实施，可能更容易改变。目标。我们将利用孟加拉国一项基于人口的男性调查的数据，开发和测试一种新的测量模型，用于测量社区一级的老年男性性别规范。我们将在多层次分析中使用这一措施，以探讨是否生活在赞成性别平等和反对暴力侵害妇女的高级男子中的地方规范中，直接和/或通过缓冲年轻男子在童年时期遭受暴力的影响，降低了年轻男子实施身体IPV的风险。数据样本包括961名35-49岁的老年男子和564名18-34岁的已婚男子，他们居住在50个城市和65个农村社区，参加了孟加拉国男子对性别和暴力侵害妇女行为的态度和做法调查。方法.我们将使用三层项目反应理论（IRT）方法来探索和改进老年男性社区水平性别规范的测量，包括评估测量的维度，可靠性，测量在城市和农村社区之间的不变性，以及社区内的异质性。我们将使用多层次逻辑回归来评估（1）社区性别规范与年轻已婚男性犯下任何前一年身体IPV的概率之间的关系，以及（2）年轻男性在童年时期遭受暴力与他犯下最近身体IPV的概率之间的关联是否取决于社区层面的性别规范。敏感性分析将测试结果的稳健性。创新在全球范围内，缺乏关于（1）男子实施IPV和（2）使用先进方法衡量社区一级男子性别规范的研究。因此，人们对社区一级的性别规范（通过这种测量模式制定的）与年轻男子实施IPV的直接联系以及通过缓冲童年遭受暴力的潜在不利影响而建立的联系知之甚少。这些数据提供了一个难得的机会来探索这些研究不足的关系。我们的新的分析策略将推进社区性别规范的男性和男性的IPV犯下的多层次的决定因素的研究领域。我们的研究结果将为国家政策和社区计划提供信息，以减少孟加拉国男性的IPV犯罪，并对类似环境中的研究和政策产生影响。

 DESCRIPTION (provided by applicant): Globally, research on intimate partner violence (IPV) is based on women's reported exposure, with little known about the determinants of men's perpetration. Nascent evidence in poorer settings suggests that childhood experiences and local gender norms may synergistically influence men's risks of perpetration, but these risk factors have not been studied together. In historically patriarchal settings like Bangladesh, norms among senior men about equity-inequity in gender relations and violence against women may influence those of lower-status community members, including younger men. Compared to more senior men, younger men are more likely to be establishing patterns of behavior in marriage, and their behaviors, including the perpetration of IPV, may be more amenable to change. AIMS. Using data from a population- based survey of men in Bangladesh, we will develop and test a novel measurement model for senior men's community-level gender norms. We will use this measure in a multilevel analysis to explore whether living amidst local norms among senior men favoring gender equity and disfavoring violence against women lowers the risk of perpetrating physical IPV among younger men, either directly, and/or by buffering the influence of younger men's exposure to violence in childhood. DATA. The sample includes 961 senior men 35-49 years (y) and 564 ever-married men 18-34 y nested in 50 urban and 65 rural communities who took part in the survey of Men's Attitudes and Practices Regarding Gender and Violence Against Women in Bangladesh. METHODS. We will use 3-level item response theory (IRT) methods to explore and refine the measurement of community- level gender norms among senior men, including an assessment of the measure's dimensionality, reliability, measurement invariance across urban and rural communities, and within-community heterogeneity. We will use multilevel logistic regression to assess (1) the relationship between community gender norms and a younger ever-married man's probability of perpetrating any prior-year physical IPV and (2) whether the association of a younger man's exposure to violence in childhood with his probability of perpetrating recent physical IPV depends on community-level gender norms. Sensitivity analyses will test the robustness of the findings. INNOVATION. Globally, research is lacking on (1) men's perpetration of IPV and (2) the measurement of community-level gender norms among men using advanced methods. Thus, little is known about the associations of community-level gender norms (developed through such measurement models) with younger men's perpetration of IPV, directly and by buffering the potentially adverse influence of childhood exposure to violence. These data offer a rare occasion to explore these understudied relationships. Our novel analytic strategy will advance the fields of study on community gender norms among men and the multilevel determinants of men's IPV perpetration. Our results will inform national policies and community programs to reduce men's perpetration of IPV in Bangladesh, with implications for research and policy in similar settings.