Seductions of Method: Rejoinder to Nagin and Tremblay’s “developmental Trajectory Groups: Fact or Fiction?”

Remember LISREL? Many of us can recall a time when papers were accepted (or rejected) in peer-reviewed journals based on the choice of this then-preferred method. Alas, criminology has witnessed a recurring sequence of hegemonic statistical methods. Before LISREL came path analysis and before that factor analysis was all the rage; in the 1980s event history analysis ruled for a brief stint and since the 1990s HLM has come on strong. It is even common to read in abstracts nowadays that a paper applies (fill in the acronym for method) as if that were a contribution in itself. We might include as well the criminal career paradigm that reigned for years largely on grounds of method (for example, how best to measure lambda). Enter group-based trajectory modeling, aka TRAJ, self-described by Nagin and Tremblay (2005: 874) as the current “statistical method of choice” for developmental research. In a remarkably short time by academic standards, more than fifty papers have been published using group-based modeling, and many more are surely in the pipeline. Is TRAJ just another statistical fad that will pass or has it transformed our knowledge and the way we conduct research?