MEASURING SELF-REPORTED DEVIANCE - CROSS-SECTIONAL OR PANEL DATA

The use of longitudinal designs in research on deviant behavior has recently been questioned on theoretical, methodological, and even economic grounds. One dimension of this questioning has been based on the ability of cross-sectional research to adequately address the issue of causality as demonstrated by the comparability of findings for both cross-sectional and longitudinal studies. This research also addresses the comparability of cross-sectional data with longitudinal data by testing modern perceptual deterrence theory hypotheses utilizing a three-wave panel study of self-reported driving while under the influence of alcohol among adults. Findings suggest that cross-sectional measures of drunken driving behavior are strongly correlated with longitudinal measures. However, comparisons of models based on modern deterrence theory produce inconclusive results concerning the ability of cross-sectional measures to approximate longitudinal measures of self-reported deviance. © 1990.