Predicting Attitudes Toward Domestic Violence Using Machine Learning

January 2025

"Agree or disagree: A husband is justified in hitting or beating his wife if she burns the food."

Posed by the Demographic and Health Surveys (DHS) Program, this survey question revealed the global disparities in attitudes toward domestic violence. In Afghanistan, 73.2% of respondents justified violence for at least one reason, compared to just 3.4% in the Dominican Republic. Intrigued by these findings, I developed a Random Forest model that predicts the likelihood of a person justifying domestic violence based on eight key features:

• Country

• Gender

• Survey Question

• Age Group

• Education

• Marital Status

• Employment Status

• Residence Type

By analyzing these factors, the model helps identify underlying patterns that contribute to the normalization of violence against women and girls. This understanding is critical for designing more effective and targeted interventions to change harmful societal attitudes.