Violence in the family affects a woman's future

Violence in the family affects a woman's future


by Raffaella Piccarreta, Department of Decision Sciences

Living in an abusive family has a strong impact on the life choices and behaviors of young women, who are forced to start their transition to adulthood earlier. An impact so negative as to cancel out any positive effects due to the well-being and development of the area in which they live. This is the most important evidence that emerges from the study Colombian women's life patterns: a multivariate density regression approach, conducted with Sara Wade, Andrea Cremaschi and Isadora Antoniano-Villalobos.
In this study, in fact, in addition to some factors recognized as important in other contributions – the cohort of birth, the region and the area (rural or urban) of residence – we also consider if and how having grown up in a violent family context contributes to influencing the transition to adult status of a woman and possibly her capacity for self-determination (agency).

To do this, we use data from the 2010 Colombian survey as part of the Demographic and Health Survey Program (DHS, https:// to study some focal events in the transition to adult status – such as age at sexual debut, first child, first cohabitation or marriage, and participation in the world of work (at the time of the interview) – on their interaction,  and on their possible relationship with some characteristics of the family context.

We chose to focus on Colombia because the context is particularly difficult for women. In addition to the patriarchal traits of society, shared with other Latin American countries, Catholic society is characterized by considerable inequalities with regard to most indicators of well-being, which are critical for the poorest and least educated strata of the population living in rural areas. The situation of women is further complicated by gender inequalities in job opportunities, and by social norms that influence young women's choices and actions regarding fertility. In particular, despite the undoubted progress recorded since the 2000s, the pregnancy rate among Colombian adolescents remains very high, especially – again – among the most disadvantaged strata of the population.

The analysis we are interested in presents difficulties at the methodological level: in fact, we jointly study dependent variables of diverse nature, and with different measurement scales, on the basis  of data that can be censored because some women – especially the younger ones – may not have yet experienced the events of interest at the time of the interview; moreover, the explanatory variables available are relatively few. We have therefore developed a very flexible non-parametric model, which makes the most of the available information, and which allows a very detailed analysis of the events of interest, conditionally to the background variables taken into consideration.

Our results highlight marked differences between urban and rural areas, the latter characterized by an earlier transition to the adult state. These areas are indeed characterized by lower levels of education and well-being, which confirms studies on the risks associated with the most distressed situations, as supported by similar results observed for poorer regions.
In particular, with reference to participation in the world of work, our model suggests that the probability of working increases with higher age of having the first child. In particular, a lower probability of working for young mothers is observed, which persists even with increasing age. This highlights the negative and permanent effects of early motherhood.

Latest Articles Opinion

Go to archive
  • The key that opens the black box

    Mathematical optimization can enable a better understanding of the deep structure of neural networks, so as to arrive at an artificial intelligence that is not only generative but also explainable, i.e. capable of justifying its own choices. Only in this way, AI can prove useful in assisting political, ethical, financial or commercial decisions

  • Side effects

    A study on a large sample of covered bonds issued by European listed banks in the 20052016 period invites regulators and monetary authorities to pay attention to indirect risk

  • Does the media propagate stereotypes?

    The answer comes from computer vision and text analysis techniques used in a Bocconi research study which highlights how modern machine learning tools can broaden the scope of social science

Browse the magazine in digital format.

View previous issues of Via Sarfatti 25



Mon Tue Wed Thu Fri Sat Sun
    1 2 3 4 5
6 7 8 9 10 11 12
13 14 15 16 17 18 19
20 21 22 23 24 25 26
27 28 29 30 31