Serial neuropsychological assessment is often conducted to monitor changes in the cognitive abilities of individuals over time. Because practice effects occur and the reliability of test scores is less than perfect, it is difficult to judge whether varying test results should be attributed to chance trends or to real changes in underlying cognitive abilities. In a large sample of adults (age range, 49–81 years), we evaluated the influence of age, gender, and education on test–retest changes in performance after 3 years on Rey's Verbal Learning Test (VLT), the Stroop Color-Word Test (SCWT), and the Letter Digit Substitution Test (LDST). A new statistical method was applied to assess the significance of changes in test performance (i.e., the regression-based change method). The results showed that test–retest changes differed as a function of age for the VLT Total recall 1–3, VLT Total recall 1–5, VLT Delayed recall, and LDST measures. An age × gender interaction was found for the SCWT Interference change score, suggesting that the age-related decline in executive functioning after 3 years was more pronounced for males than for females. A normative change table with appropriate corrections for the relevant independent variables was established. (JINS, 2008, 14, 71–80.)