Overview

This study examines how having a physical disability influences one’s risk of being food insecure, given that they are above the age of 60. Southern Nevada is a unique landscape due to its incredibly scarce resources, low quality education, and lack of viable infrastructure to support an ever growing population. On the other hand, the COVID-19 pandemic highlighted glaring disparities in resource access among the elderly and how social exclusion has exacerbated the issue of food insecurity greatly. Combined with Nevada’s lack of viable resources, disabled senior adults are barred by social barriers that prevent them from accessing necessities.

Outside of the context of social barriers, disability, as a subject, is shrouded in taboo. It’s a subject many don’t feel comfortable being truthful about. While some may not know they have a disability, others will intentionally misreport theirs, due to a combination of social stigma, financial incentive, or cultural reasons. This human element to disability is often neglected in past studies and results in data that is often skewed.

This study utilizes a quantitative analysis using both multivariable regression and nonparametric bounds to examine the relationship between self-reported disability status and food insecurity among senior adults in Southern Nevada. This dataset consists of survey response data from 1,000 senior-age respondents in Southern Nevada collected last year, including questions about socioeconomic factors, disability, and food insecurity status. Given the misreported nature of disability, a nonparametric bounds approach is integrated to estimate the true effect of disability on food insecurity, controlling for socioeconomic factors like education, race, SNAP participation, healthcare access and income. The analysis includes six linear probability models (LPM): three focusing on disability and socioeconomic factors in relation to low food insecurity and three in relation to very low food insecurity. Unlike previous studies on this topic, this study utilizes a nonparametric bounds analysis to account for perfect data and misreported data (up to 5%), comparing the two in order to gauge the true size of the relationship between food insecurity and disability. This comparison enables a more honest understanding of how misreported data impacts policy responses to the issue, while integrating necessary solutions. 

Research Question

The nature of this study revolves around two, core questions: How does disability status impact the likelihood of food insecurity among senior adults (60+) in Southern Nevada? How does the misreported nature of disability influence estimations of food insecurity, and what policy measures can be applied to mitigate the effects of food insecurity? 

This study highlights systemic barriers that further exacerbate food insecurity and offers data-driven policy solutions that address these social inequities. While food insecurity has remained a relevant topic in the post-pandemic world, the elderly are an often neglected and understudied population. Senior adults with disabilities are often exempted from major food insecurity studies, even though the likelihood of experiencing health-related conditions are the highest among this group. This research question was crafted as a two-pronged approach to provide clarity on both groups. Accounting for senior adults, followed by those with disabilities, and finally the intersection of the two. This question includes a more interpersonal element by asking how the endogenous and “taboo” nature of disability may compel respondents to misreport their status, resulting in skewed data. The foundation of these questions rely on asking if the basis of food insecurity policy is wrong. If so, how does one measure misreporting and in what ways can that be accurately measured? 

Findings & Implications  

In a fully controlled regression model, weighing food insecurity and disability, while controlling for factors such as age, income, race, education, SNAP, etc., having a disability increases the likelihood of being food insecure by 22.05 percentage points (p < 0.001) and 10.53 percentage points (p < 0.01) for very low food insecurity. These findings indicate a statistically significant and strong relationship between disability and food insecurity. Education was a protective factor, with the minimum of a college degree equating to lower risk of food insecurity, an approximately 18.43 to 23.18 percentage point decrease in food insecurity risk. Income was statically significant (p < 0.001), though the margins effect per dollar was sizable small (-0.0000037), aligning with previous research that income alone was not the sole determinant of one’s food insecurity risk. SNAP participants were 14.43 percentage points more likely to be food insecure and 11.99 percentage points more likely to be food insecure (p < 0.01), reflecting a reverse causality. Female respondents had a 5.01 percentage point increase (p < 0.10) in food insecurity risk, compared to their male counterparts, suggesting a gender-based disparity.  

Utilizing the bounds analysis, the model then takes into account misreporting using a combination of the social factors outlined in the regression, along with assuming a 5% misreporting in the sample population. Under the baseline assumption, which assumes that nobody misreported their disability, being disabled is associated with a 30.8 percentage point increase risk of being food insecurity. However, when the model accounts for a 5% misreporting rate, while controlling for social factors (e.g. race, income, SNAP, education, etc.) dramatically shrinks the effect. The upper bound is only 18.8 percentage points with a 5% error and as low as 3.0 percentage points with no error. The results suggest that standard regression models may overstate the true effect of disability on food insecurity, likely due to misclassification.

Implications

These findings suggest that disability is a key predictor of food insecurity among senior adults, even accounting for socioeconomic factors. However, the nonparametric bounds reveals that the relationship observed in traditional regression models may overstate the relationship between the two factors, due to misreporting. Policy efforts should focus on improving food access for disabled adults, while addressing the barriers to food access like education, location, income, etc.. Funding programs like Meals on Wheels, which deliver cooked meals to homes, may help offset this imbalance and make food access more equitable. Researchers and policymakers should be cautious about interpreting causal effects when utilizing self-reported data.