A sample of 704 cognitively intact individuals (M age = 63.7 years) performed a battery of cognitive tests on as many as three occasions, at approximately 3-year intervals. The authors used random effects models to analyze cross-sectional relationships between cognitive performance and state anxiety and longitudinal relationships between cognitive change and neuroticism, after controlling for gender, age, and education. Cross-sectionally, higher state anxiety was associated with poorer performance on Wechsler Adult Intelligence Scale Synonyms, WIT III Analogies, Koh's Block Design, two measures of visual learning (Names and Faces and Thurstone's Picture Memory), and, for men, CVB-Scales Digit Span Test and Card Rotations. In longitudinal models, the main effects for neuroticism were significant for Block Design, Symbol Digit, and Names and Faces, but there were no significant interactions among neuroticism, gender, and time. These results provide some support for Eysenck's processing efficiency theory but none for neuroticism as a risk factor for cognitive decline in normal aging.