Key points are not available for this paper at this time.
Selective attrition effects in subjects born between 1885 and 1935 (age from 21 to 75 years of age at first test) are examined in a sequential design involving the components of cohort, occasion, retest-participation, and sex, 490 male and female subjects tested in 1956 on ten cognitive variables were classified into retest-participants and nonparticipants on the basis of a follow-up in 1963. An additional 867 subjects first tested in 1963 are also grouped into those who were available for retesting in 1970 and those who dropped out. Strong effects of participation were obtained on all measurement variables, with retest-participants scoring higher than dropouts on all variables. In dications of cohort-related and age-related differences in participating behavior were also found. These results further substantiate the contention that cohort differences, and age and cohort-related differences in volunteering behavior, seriously jeopardize the validity of conventional longitudinal and cross-sectional studies.
Schaie et al. (Sun,) studied this question.