Abstract
This article is concerned with modeling the individual latent growth curves of older adults on measures of fluid intelligence in a 2‐group cognitive intervention study (Baltes, Dittmann‐Kohli, & Kliegl, 1986). Second‐order polynomial growth curves reflecting initial test performance improvement followed by relative stability/drop are fitted to the recorded scores of each participant. Experimental subjects are found to exhibit mean ability growth outperforming that shown by untrained subjects. Both the control and experimental groups, however, demonstrate the same pattern of interrelationships among the individual ability intercepts, linear slopes, and quadratic bend parameters of the growth curves. The results suggest substantial amount of plasticity in fluid intelligence of older adults, and that elders may be capable of considerably improving their performance on repeatedly presented fluid measures by capitalizing on interrelationships between fluid intelligence subabilities.