Citation

Murray GF, Erickson PG. Soc. Sci. Res. 1987; 16(2): 107-118.

Copyright

(Copyright © 1987, Elsevier Publishing)

DOI

10.1016/0049-089X(87)90011-1

PMID

unavailable

Abstract

The relative merits of cross-sectional and longitudinal designs in resolving problems of causal order have been much debated in the recent literature. The findings of recent panel studies have been in conflict with those obtained from cross-sectional studies where the dependent variable involved a measure of past behavior. The utility of a prospective measure, however, has not been assessed within a longitudinal design. In this panel study the dependent variable at t1 is a projected measure of criminal behavior. Behavior at t2 provides a comparable measure of subsequent criminality. The analyses support the notion that prospective measurement provides a good estimate of actual, subsequent behavior.