Distinguishing the various aspects of a 'criminal career,' including participation, frequency, and termination, is likely to produce more precise knowledge about offending and its causes than a single aggregated construct such as 'criminal propensity.' By disaggregating various aspects of offending, the effects of various factors on aspects of criminal activity can be isolated. Research to identify patterns of variation or stability during criminal careers inherently requires longitudinal data because a 'career' is longitudinal by definition, and longitudinal information is necessary to address a longitudinal phenomenon. Gottfredson and Hirschi are correct in supporting randomized experiments and quasi-experiments as the best methods for testing causal hypotheses, but there is nothing in their methodological assessments that undermines longitudinal surveys as an important method in research relating to criminal careers. Longitudinal surveys are markedly superior to cross-sectional surveys in describing the natural history and developmental course of a phenomenon, in studying developmental sequences, in studying prediction, and in drawing causal inferences. 19 footnotes and 29 references.