ABSTRACT
This chapter analyzes longitudinal relationships expressed in linear models using point-of-time and change data and expressed in product models using point data. It describes that there is not much to be gained in the linear explanation of the variance in a crime factor by adding historical data of these properties. The important properties SOCAID and OCCDICH can only be described with crime-synchronous values, and this makes for problems when interpreting coefficients. The longitudinal analyses have offered opportunities for testing the social control explanation of crime, opportunities that are much better than the opportunities provided by cross-sectional analyses. The possibility that there may be micro-level interaction that cannot be expressed in the type of macro-level terms often used in criminological research is of great interest when evaluating the attempts made in this study to model crime and the change in crime.
