Multivariate Structural Modeling of Plasticity in Fluid Intelligence of Aged Adults

This article is concerned with modeling the structure of cognitive intervention effects in adult intellectual development. Within the Cattell-Horn (Cattell, 1971 ; Horn, 1970, 1982) theory of fluid and crystallized intelligence, the ability of older adults to improve their performance on fluid intelligence tests, with or without tutor-guided instruction in test-relevant cognitive skills, is examined using multidimensional longitudinal structural equation models. The models reflect in some of their parameters substantively interesting aspects of temporal change at the ability level, and are applied to data from a two-group cognitive intervention study of plasticity in fluid intelligence of older adults (Baltes, Dittmann-Kohli, & Kliegl, 1986). Training effects are found in the experimental group that are more salient than the practice/experience effects in the control group; both types of effects are maintained over a six month testing period. Exploration of the nature of the practice effects in the control group reveals patterns of test performance improvement that are equivalent to those in the experimental group, as reflected in the interrelationships among initial ability levels and maximal improvements at the ability dimensions relative to initial status. The results suggest (a) considerable plasticity in adult intellectual functioning, and (b) that healthy aged adults may be capable - based on experience with repeatedly presented measures - to markedly improve their performance on fluid intelligence tests by capitalizing on interdependencies between fluid subabilities.