Objectives: To examine contributions of self-reported vision impairment to disability by persons 65 years and older. Method: We used data from the National Health and Nutrition Examination Survey (NHANES IV; 1999-2008; weighted N = 6,550). The four-staged Nagi model of the disability process, incorporating biomarkers, disease, and physical functioning, was used as a framework to guide variable selection. An unadjusted ordinal logistic regression model testing effects of vision on daily-life activities was compared with models that controlled covariates associated with each stage of the disability model. Results: For older adults with vision impairments, effects of poor vision in predicting disability were substantially reduced or eliminated when other health covariates were controlled. Where significant effects of poor vision remained, the strongest effect was found for doing leisure activities at home (OR = 5.29; 95% CI [3.14, 8.90]). Discussion: Results suggest the importance of considering multiple health factors when formulating comprehensive vision rehabilitation interventions.