Focus is on the use of longitudinal designs in examinations of causal order, criminal career patterns, peer group influences, treatment, treatment and criminal justice intervention effects, and effects of ordinary life events. It is argued that the longitudinal or cohort design is not justified on methodological grounds, that it has produced illusory substantive findings, and that it has promoted policy conclusions of doubtful utility. In addition, it is noted that such designs are very expensive and have opportunity costs that have not been properly evaluated. It is concluded that many of apparent benefits attributed to longitudinal research can be realized by carefully designed and reasonably conceptualized cross-sectional studies, at substantially reduced cost. 56 references. (Author abstract modified)