Developmental trajectory groups: Fact or a useful statistical fiction?

A rapidly growing literature in criminology and psychology uses a group-based methodology to identify and analyze developmental trajectories. Some confusion has arisen about the interpretation of this novel statistical model and with it some degree of cautionary commentary. We begin with a discussion of the role of trajectory groups as a statistical device for approximating population differences in developmental trajectories. We then discuss three misconceptions about group-based trajectory modeling that stem from misunderstandings about the approximating role of trajectory groups: (1) individuals actually belong to a trajectory group, (2) the number of trajectory groups is immutable, and (3) the trajectories of group members follow the group-level trajectory in lock step. We also point out that group-based statistical modeling is not bound at the hip to the testing of taxonomic theories, and can just as well be used to test nontaxonomic theories.