Date of Award


Degree Type


Degree Name

Doctor of Philosophy (PhD)


Social Sciences


Colleen Heflin

Subject Categories

Public Affairs, Public Policy and Public Administration | Public Policy | Social and Behavioral Sciences


This dissertation consists of three chapters pertaining to Supplemental Nutrition Assistance Program (SNAP) participation, food insecurity, cognition impairment, and sibling correlation in SNAP participation over the life course. Participation in SNAP among eligible adults 60 and older is much lower than among the younger population, and rates continue to decline throughout the life course while, at the same time, the risk of cognitive impairment increases. The relationship between food insecurity and health outcomes among adults has garnered increasing attention. Some previous studies found an association between food insecurity and cognition outcomes among older adults. However, they were hampered by not considering midlife as a specific food insecurity exposure window and the effect of this hardship over a long-run time period. Even though siblings are essential players in family dynamics, previous research that carefully considered siblings’ role in welfare participation is limited. Due to the high administrative burden associated with SNAP eligibility processes, the first paper examines if cognitive impairment is associated with low uptake of SNAP among the low-income older adult population. The second investigates the association between individuals’ food insecurity in their fifties, measured as exposure experience at some point and total years of the exposure, and the cognitive ability and incident dementia in later years. The third paper adopts a nuanced familial paradigm in conceptualizing an individual’s social network, thereby foregrounding an exploration into the influence of parental and sibling interrelationships on the likelihood of an individual’s SNAP participation. In Chapter 1, co-authored with Colleen Heflin, I estimate linear probability fixed-effects models to assess the effect of cognitive decline on the likelihood of SNAP participation among eligible adults aged 60 and above, controlling for observed characteristics that change over time as well as individual, time, and state fixed effects using panel data from the Health and Retirement Study. The results show that the reduced levels of cognitive functioning that rise to the classification of dementia were strongly associated with reductions in the probability of SNAP take-up among eligible older adults. Results were particularly salient for females and those living alone. One barrier to SNAP take-up among older adults may be cognitive impairment, with the size of the effect differing by gender and living arrangement. Policymakers may want to consider initiatives to increase SNAP participation among older adults, including a focus on further simplification of eligibility and recertification processes that reduce administrative burden. Chapter 2 examines how exposure to food insecurity (FI) and the cumulative exposure duration during the age of fifties is associated with subsequent cognition ability and dementia onset risk using HRS data from 1995 to 2020. The findings suggest that food insecurity exposure experienced in age 50-59 is associated with higher dementia onset risk and a lower total cognition function score after age 60. In addition, each additional year of food insecurity exposure in the midlife stage is associated with a dose-response increase in dementia onset risk and a reduction in total cognition functioning score in later life. This study strengthened the literature that both the timing and extent of food insecurity exposure matters for later-life cognitive health and late-onset dementia. This study indicates that life course disadvantage accumulates during midlife to predict worse later-life cognitive function, which provides strong evidence for the cumulative inequality model. Chapter 3 exploits data from the Panel Study of Income Dynamics (PSID) from 1975 to 2019 and estimates three-level mixed effects logistic models to assess the relationship between individuals’ family network and their SNAP participation. The analysis has leveraged sibling dyads to provide nuanced insights into intergenerational and sibling correlations in SNAP participation, simultaneously accounting for the effects of time-invariant early-life shared environments and time-variant parental, inter-sibling, and external environmental factors. I found that early-life SNAP exposures and current parental on SNAP are associated with an individual’s SNAP participation in adult life. However, the influence of early-life exposure and sibling’s SNAP participation are nuanced and vary across gender and race. The study contributes to literature by combining both the intergenerational and intragenerational perspectives in understanding of welfare participation across the lifespan of two generations.


Open Access

Included in

Public Policy Commons