Modeling determinants of women's decision-making autonomy on agriculture and household matters in rural Ethiopia: a mixed effects beta regression tree analysis.

Despite existing research on women's decision-making autonomy (WDMA), there remains a significant gap in understanding the complex factors influencing women's autonomy. This study investigated the determinants of WDMA in agriculture and household matters in rural Ethiopia. The study used data from nine rural districts in the Oromia and the former Southern Nations Nationalities and Peoples region in Ethiopia. A mixed effects beta regression tree model was used to identify variables impacting WDMA across diverse participant segments. The analysis incorporated various fixed effects, including family size, women's and their partners' years of education, women's age, women's perception towards cultural norms on gender differences, religion, household wealth quintiles, and participation in social groups. Random effects were taken into account by including the districts of the households and the survey rounds, while the gender of the household leader and marital status served as partitioning variables. Factors that were found to be significantly associated with WDMA among the married women from male-led households were: the age of the women, the perception of women towards cultural norms on gender difference, if women were Muslim compared to Orthodox Christian, richness of households and participation in social groups. In the case of women not in marriage and who were from male-led households, we identified the same factors to be associated with increased WDMA. Among women in female-led segments, only the age of the women and Orthodox Christianity compared to Muslim followers were contributing significantly to a higher WDMA. Interventions aimed at increasing women's decision-making autonomy should consider acting on some of these determinants so that women can play a more prominent role in agricultural and household matters.