New regression-based norms for letter and semantic fluency tests on Norwegian older adults.

This study presents updated regression-based norms for verbal fluency tests: "FAS," "Animals," and "Supermarket" in older Norwegian adults. By examining the potential modulating effects and putative interaction effects of age, education, and sex, we aim to investigate cognitive aging and the effects of education. The study includes 449 healthy participants, aged 70-92 years, from the NorFAST (n = 254) and HUNT (n = 195) studies. Regression-based norms for all three verbal fluency tasks were produced. Models were assessed for pertinent linear, curvilinear, or interaction effects of age, education, and sex. We compared our norms to previously published norms. The task "FAS" was predicted by education (b = 0.773, partial R2 = 0.117, p < .001) and sex (b = 1.110, partial R2 = 0.037, p < .001), and the task "Animals" was predicted by education (b = 0.313, partial R2 = 0.021, p < .01) and age (b = -0.123, partial R2 = 0.045, p < .001). Higher education and being female were associated with better performance, whereas older age was associated with poorer performance. Finally, for the task "Supermarket," we found an interaction effect between age and education (b = -0.063, partial R2 = 0.034, p < .01), but the major predictor was sex (b = 1.446, partial R2 = 0.068, p < .001). The norms were adjusted for the demographic variables that predicted the individual tasks. The illustrative norm comparison estimated similar mean performance but larger differences between the "exceptionally low" scores. The results also indicate that only processing speed declines with age, and that higher education has minimal effect on test-performance.