Further Reflections on Modeling and Analyzing Developmental Trajectories: A Response to Maughan and Raudenbush

Abstract

In this article, the authors respond to the Raudenbush and Maughan commentaries elsewhere in this volume. Stephen Raudenbush's principal criticism of the groupbased trajectory model is that it reifies the idea that people follow a small number of immutable trajectories of behavior. This criticism reflects a fundamental misunderstanding of the statistical role of trajectory groups. Trajectory groups describe the trajectory of behavior that has actually occurred; that behavior is not reified but real. There is nothing in a trajectory group model that asserts the behavior pattern is permanent, that no intervention can change it, or that it will continue beyond the time period of the observed data. The question of whether a group-based trajectory model or a hierarchal linear model can provide a better statistical representation of change is an empirical rather than philosophic question. The answer will undoubtedly be context-specific. Barbara Maughan's comments were cast as reflections froma developmental psychopathology perspective. The authors are in complete agreement with Maughan's observations about the importance of using diverse methods in studying developmental psychopathology and of unpacking the developmental trajectories of the constituent components of antisocial behavior. The authors would only add that the developmental origins of these behaviors should be studies from as early in life as possible.