172 Irving Lorge 



sampling and evaluating living samples introduces a bias. 

 The survivors of a cohort at any age must differ somewhat 

 from those who succumbed to the vicissitudes of life (cf. 

 Thorndike et al., 1934). Equally significant, moreover, is the 

 fact that the samples of differing age groups may have had 

 different environment and experiences, e.g. socially, in the 

 values and attitudes acquired; or educationally, in the amount 

 and quality of schooling obtained; or, vocationally, in the 

 duration and intensity of the work load carried. 



The cross-sectional method, therefore, allows comparisons 

 between measures of differing age groups, i.e., it provides 

 estimates about age differences rather than evidence about age 

 changes. The method used to provide evidence about changes 

 brought about in the ageing process is the longitudinal or 

 follow-up. Essentially, the longitudinal method, at its best, 

 studies some trait or traits in the same sample of individuals at 

 successive periods in their lives. The follow-up studies of the 

 intellectually gifted (Terman et al., 1947), of the vocational 

 careers of adolescents (Thorndike et ah, 1934) and of the intelli- 

 gence test performances of college students thirty years later 

 (Owens, 1953) are typical longitudinal studies. The assump- 

 tion basic to the method is that nature (in the sense of genetic 

 constitution) is more adequately controlled since the same 

 individual is being evaluated at different times during his life- 

 span. It must be recognized, however, that any cohort, to 

 some degree, may be a function of its own special environ- 

 ment. A cohort born in 1886 would have been brought up in 

 circumstances differing significantly from that to which the 

 cohort now being born will be exposed. To this degree, then, 

 the longitudinal method, by the very homogeneity of environ- 

 ment for any cohort, has some limitations. The difficulties in 

 maintaining effective liaison with all members of the cohort 

 or its sample, of keeping and training a loyal and efficient 

 research staff to do the testing, are not nearly so crucial as are 

 the difficulties introduced by the impact of social, political, 

 economic, or military events on a specific cohort over its life- 

 span. Except for Owens' study of intelligence and Kallmann's 



