Cognitive decline in old age: separating retest effects from the effects of growing older.

Repeated administration of cognitive tests improves test performance, making it difficult for researchers to gauge the true extent of age-related cognitive decline. The authors examined this issue using data from the Rush Religious Orders Study and linear mixed-effects models. At annual intervals for up to 8 years, more than 800 older Catholic clergy members completed the same set of 19 cognitive tests from which previously established composite measures of cognitive domains were derived. Retest effects on some measures were substantial and continued to accumulate even after 8 annual test readministrations, but effects on other measures were minimal. Across cognitive measures, retest effects were not related to age, sex, or education. Individual differences in retest effects were substantial but not consistent across cognitive measures. The results suggest that retest-based improvement in cognitive test performance can be substantial and persistent.