Food Insecurity and Violence Against Children at Homes in Southern Uganda: A Propensity Score Analysis

Background: Food insecurity and violence against children at home is widely spread and can have serious adverse implications for the mental and physical health of children and adolescents, resulting into poor quality of life that may persist even into adulthood (UNICEF, 2018; WHO, 2019). According to the most recent estimates, globally 54-76% of 2-17 children and 50-82% of 2-17-year-old children on the African continent have experienced some form of past-year violence at homes (Hillis, Mercy, Amobi, & Kress, 2016). In Uganda alone, half of adolescent girls and two thirds of boys (13-17 years old) reported an experience of violence in general and an estimated 44% of young girls 13-17 years and aged 58% of young boys report having experienced physical violence (UNICEF, 2018).

In addition, it is estimated that more than 1 in 3 people are classified as food-insecure in Uganda, (USDA, 2018). Food security is directly connected to a multitude of economic factors (USAID, 2019) including poverty, and has been grouped among the characteristics of households that are predisposed to family violence (Jackson, Lynch, Helton & Vaughn, 2018). However, there is limited research on links between food insecurity and violence against children at homes in low resource settings. Therefore, this paper aims to explore how food insecurity and thus poverty in general is linked to violence against children at homes in Uganda.

Methods: This paper uses the data from 14-17 old 1,260 adolescent girls from 47 secondary schools in southern Uganda. The data was collected at baseline stage from Suubi4Her study that aims to explore the impact of an economic empowerment strengthening intervention on health outcomes of adolescent girls.

To establish effects of food insecurity on risk of physical violence against children within the home, the propensity scores are created by modeling a logistic regression (Rosenbaum & Rubin, 1983). Then one-to-one match is made to make sure that distribution of adolescent girls with food insecurity is closer to that of adolescent girls with secure food supplies, indicating correction of confoundedness. The final sample includes 644 cases—evenly matched between food insecure and secure adolescents. After matching, Poisson regression analysis with robust standard errors is used to explore the effect of poverty on incidents of physical violence at home. The PSM is conducted using STATA/PSMATCH2.

Results: Our analysis indicate that 28.97% of adolescent girls report inadequate food/food insecurity and 57.14% of all girls also indicate exposure to physical violence at home. The further analysis shows that girls who have experienced food insecurity are more likely to report higher rate ratio for physical violence in the home by almost 18% (95% CI: 1.006 - 1.375, p=0.042) after controlling family demographics as constant.

Implications: Our findings indicate that food poverty/insecurity is associated with higher rates of physical violence in the home. This finding could have implication for social work practice and child protection policies targeting decreasing violence against children.