Educational level is the most consistent nonbio- logical predictor of both cognitive level and rate of change in prior longitudinal studies and meta- analyses ( Albert et al., 1995; Anstey & Christensen, 2000; Schaie, a ). Moreover, education predicts cognitive change not only in old age but also throughout adulthood ( Farmer et al., 1995; Lyketsos et al.,1999; Schaie, b ). Consistent with co-constructionist approaches, education is reported to most consistently predict change in crystallized abilities, memory, and mental status, and is less consistently predictive of change in fluid abilities and speed. The effects of education on cognitive change remain when controlling for factors such as age, gender, race, and health. In the MacArthur study of successful aging, education best predicted change in cognition ( Albert et al., 1995 ). Secular trends in education are well documented. Educational attainment, particularly in post-secondary education, has increased significantly across birth cohorts in the first half of the twentieth century. In 2000 15% of 65 elders had attended college, compared to almost 50% of Baby Boomers. Hauser and Featherman (1976) reported a total increase of about 4 years of education from birth cohorts 1897 to 1951. Intergenerational differences in schooling peaked among men born shortly after World War I, and a deceleration has occurred across more recent cohorts.
Intergenerational differences between successive generations, approximately 20 to 30 years apart, range from 2 to 4 years (cf. also Willis & Schaie, b ).