9375Georg-August-Universität Göttingen, Gottingen, Germany.
J Interpers Violence. 2022 Dec;37(23-24):NP22704-NP22731. doi: 10.1177/08862605211072179. Epub 2022 Feb 8.
The purpose of this paper is to identify the risk factors for emotional intimate partner violence (IPV) against women and girls with children in Mexico from an ecological perspective. To that end, we generate a dataset with 35,004 observations and 42 covariates, to which we apply an additive probit model estimated with a boosting algorithm to overcome high-dimensionality and simultaneously perform variable selection and model choice. The dataset integrates 10 information sources, allowing us to properly characterize the four levels of the ecological approach, which is the first contribution of this paper. In addition, there are three key contributions. First, we identify a number of factors significantly linked to emotional IPV against women with children: age, age at sexual initiation, age at marriage (or cohabitation), autonomy regarding professional issues, social support networks, division of housework, the community's Gini index, women's economic participation in the municipality, and the prevalence of crime against males in the region. Second, we discuss some risk factors whose effects have not been examined or have been underexplored for Mexico; these include women's decision-making autonomy, social support networks, distribution of housework, the community's economic inequality, and criminality. Third, we identify specific risk subgroups that are generally overlooked: women who had their first sexual intercourse during childhood and women who got married (or moved in together with a partner) late in life. The main results suggest that strategies aiming to promote women's social and economic empowerment and reduce criminality should also incorporate a gender component regarding emotional violence against women with children in the context of intimate relationships.
本文旨在从生态视角出发，确定墨西哥有子女的妇女和女童遭受情绪亲密伴侣暴力（IPV）的风险因素。为此，我们生成了一个包含 35004 个观测值和 42 个协变量的数据集，并应用了一个加性概率模型，该模型使用提升算法进行估计，以克服高维性并同时进行变量选择和模型选择。该数据集集成了 10 个信息源，使我们能够正确描述生态方法的四个层次，这是本文的第一个贡献。此外，还有三个关键贡献。首先，我们确定了一些与有子女的妇女遭受情绪 IPV 显著相关的因素：年龄、初次性行为年龄、婚姻（或同居）年龄、专业问题的自主权、社会支持网络、家务分配、社区基尼指数、妇女在市一级的经济参与度以及该地区针对男性的犯罪率。其次，我们讨论了一些风险因素，这些因素的影响尚未在墨西哥得到检验或研究不足，包括妇女的决策自主权、社会支持网络、家务分配、社区经济不平等和犯罪。第三，我们确定了一些通常被忽视的特定风险亚组：童年时期就有初次性行为的妇女和晚年结婚（或与伴侣同居）的妇女。主要结果表明，旨在促进妇女的社会和经济赋权以及减少犯罪的战略还应包含一个性别组成部分，即针对有子女的妇女在亲密关系中遭受情绪暴力的问题。
