Education, cohort replacement, and (mis)assumptions in social science research.

摘要

Wilson and Gove (1999a; 1999b) argue that due to a (mis)assumption about what a control for education means, prior work by others has misinterpreted findings regarding an education-adjusted intercohort decline in vocabulary ability. The current research extends these findings by investigating the underlying processes of educational attainment over the course of the last half century.; Research literature on educational attainment attributes the pattern of increase of educational attainment over time to a process of cohort replacement in which intercohort change is the primary mechanism underlying the replacement process. Under the assumption that cohorts do not significantly increase their educational levels after age 24, education can therefore be “fixed” for all persons after that age. In the current research, enrollment data are presented that demonstrate substantive changes in the demographics of persons enrolled in higher education; specifically that persons who are older than age 24 have returned to school in record numbers. Using General Social Survey (GSS), Census, and Current Population Survey (CPS) data spanning the latter half of this century, I demonstrate that the strength of the replacement effect in terms of educational attainment is decreasing, and that intercohort differences have diminished in comparison to intracohort change as a contributor to increases in educational attainment. I further explore these findings with respect to race and gender. These findings have broad implications for social science researchers in (1) the way education is treated in statistical analyses, and (2) for the interpretation of results that include an educational control variable.