ABSTRACT
M ANY excellent longitudinal studies of cognitive change have had the same basic aims. The most general has been to determine the average form of the trajectory of agerelated change and, in particular, whether or not the average rates of change accelerate in old age (e.g., Hertzog & Schaie, 1988; Rabbitt, 1993a; Schaie & Strother, 1968). A corollary aim has been to determine whether rates of change differ between different mental abilities or are similar for all (e.g., Arenberg 1974; Colsher & Wallace, 1991; Heron & Chown, 1967; Hertzog & Schaie, 1988; Hultsch, Hertzog, Small, McDonald-Miszczak, & Dixon 1992; Johansson, Zarit, & Berg, 1992; Lansen, 1997; Owens, 1953, 1966; Powell, 1994; Rabbitt, 1993a; Schaie, 1996; Schaie & Labouvie-Vief, 1974; Schaie & Strother, 1968; Schaie & Willis, 1993; Terman & Oden, 1947, 1959). A third aim has been to test how rates of cognitive change are affected by demographic factors such as educational and social advantage (e.g., Bosworth, Schaie, & Willis, 1999; Evans et al., 1993; Fomer, 1972), by gender (e.g., Bosworth et al., 1999; Voitenko, & Tokar, 1983), by epidemiological factors such as general health (e.g., Bell, Rose, & Damon, 1972; Birren, Butler, Greenhouse, Sokoloff, & Yarrow, 1963; Costa & McCrae, 1980; Mclnnes & Rabbitt, 1997; Rabbitt, Bent, & Mclnnes, 1997), by specific pathologies (e.g., Hertzog, Schaie, & Gribbin, 1978), by maintenance of physical mobility and engagement in everyday physical activities (e.g., Clark, 1960; Clement, 1974; Dirken, 1972; Mclnnes & Rabbitt, 1997), or by genetic factors (e.g., Bank & Jarvik, 1978; Payton et al., 2003; Pendleton et al., 2002; Terman & Oden, 1947, 1959). This raises the general issue of the extent and etiology of individual differences in trajectories of aging. Prima facie, because individuals are affected in different ways and to different extents by their lifestyles, health histories, and genetic factors, we might expect that their trajectories of aging correspondingly diverge so that variance in performance between members of a sample will increase as the members age (Morse, 1993; Rabbitt, 1982, 1993a). It follows
that individual differences in rates of change provide more information about the functional determinants of cognitive aging than do average trajectories of decline.
