K. Warner Schale
Pennsylvania State University
The purpose of this chapter is to trace in some detail the evolution of my thinking on how to design studies that will characterize developmental progressions with minimal conceptual ambiguity. As part of this review it will become clear that my approach has obviously been affected by the impact of the contributions of numerous colleagues, whether as research collaborators or as critics. Equally important, however, has been the continuous effort to collect developmental data in such a way that these designs could be put to empirical test. Indeed here is a prime example of the dialectical interaction of puzzling data sets resulting in the examination of the appropriateness of standard research designs, and the consequent specification of alternate designs that lead to the collection of new data sets to fit the new paradigms.
The work that I wish to review began with the realization that data on the adult development of mental abilities showed wide discrepancies between cross-sectional and longitudinal data collected on the same subject population over a wide age range. In particular, it became evident that for some dependent variables substantial age differences obtained in cross-sectional studies could not be replicated in the longitudinal data while for other dependent variables, longitudinal age changes reflected more profound