Research That Matters (January 17 - 20, 2008)

Background: People with disabilities are more likely than those without to experience material hardships (She & Livermore, 2006). A major form of material hardships is food insecurity, which is almost twice in prevalence among the disabled population as that among the non-disabled population. A possible explanation is that people with disabilities do not have sufficient financial resources (income and assets) to adequately meet their basic needs for food. In other words, financial resources may be an important mediating factor between disability and food insecurity status. This study will use the 1999 Panel Study of Income Dynamics (PSID) data to test this proposition and identify the direct and indirect effect of disability on food insecurity. The findings of this study will have important implications for food assistance programs and disability policy.

Method: The data used in this study are from the 1999 PSID. The sample included 5,430 households, of which 17.5% (n=947) are headed by individuals with a disability. Food insecurity status is measured by the Food Insecurity Core Module developed by the USDA. Based on this module, two measures are created: one is the food insecurity score (0-10) with larger scores indicating higher levels of food insecurity, and the other is a categorical measure with three levels ("food secure," "food insecure," and "food insecure with hunger"). Disability is a dichotomous measure of "work limitation" in the PSID. Control variables include the household head's gender, race, education, employment status, household size, and food assistance participation. Multiple regression models are tested hierarchically, first with disability status entered along with the other control variables, and then with income and assets being added.

Results: OLS regression model with the food insecurity score regressing on disability status and the control variables is statistically significant (F=105.77, p<.0001, R-square=0.197). Disability status is significantly associated with food insecurity (b=0.137, t=6.58, p<.0001). The addition of income and assets, however, does not significantly change the parameter estimate of disability status, indicating that disability has a unique effect on food insecurity, not necessarily through the mediation of financial resources. Logistic regression presents similar results. In addition, what is interesting is that neither employment status nor food assistance participation is significantly associated with food insecurity status.

Implications: Although the findings do not support the prevailing view that high incidences of food insecurity among people with disabilities are attributed to inadequacy of financial resources, the conclusion needs to be examined by further research. In addition, future studies should particularly focus on disability related issues that may lead to food insecurity. For instance, given the same level of income, people with disability might have higher levels of health expenses, which has impeded their financial capacity for food consumption. Lack of transportation services and assistance in food preparation might also contribute to food insecurity. Food assistance programs targeting people with disabilities should also consider these factors in order to effectively address the food insecurity problem among this population.