WM_Data: Derived Dataset on Attitudes Toward Domestic Violence Among Ever-Married Women in Bangladesh (2019)
Description
This dataset is a researcher-created, anonymized, derived dataset based on the 2019 Multiple Indicator Cluster Survey (MICS) conducted in Bangladesh. It focuses on attitudes toward the justification of domestic violence among ever-married women aged 15–49 and includes selected socio-demographic and household-level variables relevant to this outcome. The dataset was constructed by filtering the original MICS women’s questionnaire data and recoding variables related to education, wealth status, age, marital history, geographic division, access to technology, and other contextual factors. A binary outcome variable representing attitudes toward the justification of domestic violence was generated following standard MICS questionnaire items. Survey weights were retained to support population-representative analysis. All direct and indirect personal identifiers were removed, and the dataset does not allow identification of individuals or households. The original MICS survey followed internationally accepted ethical standards, including informed consent. This derived dataset is intended for academic and policy-oriented research, including statistical and machine learning–based analyses of gender-related attitudes in Bangladesh.
Files
Steps to reproduce
1. Obtain access to the Bangladesh Multiple Indicator Cluster Survey (MICS) 2019 women’s questionnaire data from UNICEF, subject to MICS data access conditions. 2. Load the original MICS women’s dataset into SPSS (or an equivalent statistical software capable of reading SPSS .sav files). 3. Filter the dataset to include ever-married women aged 15–49. 4. Select variables related to attitudes toward the justification of domestic violence and relevant socio-demographic characteristics, including education, wealth index, age, marital history, geographic division, and access to technology. 5. Recode domestic violence attitude variables to construct a binary outcome variable representing justification versus non-justification, following standard MICS questionnaire items. 6. Clean the data by handling missing values, applying consistent labeling, and removing any remaining indirect identifiers. 7. Save the processed dataset as an SPSS file (wm_data.sav), which constitutes the derived dataset shared here. Model training, class balancing, and explainability analyses were conducted separately and are not required to reproduce the dataset itself.
Institutions
- Khulna University School of Science Engineering and TechnologyKhulna