Longitudinal neighborhood profiles in delinquency: The decomposition of change

The methods that traditionally have been used to examine changing spatial distributions of crime and delinquency rates are not able to analyze parsimoniously the full range of within-community variation in these rates over time. In this paper, we present a variant of Bryk and Raudenbush's(Psychol. Bull. 101: 147–158, 1987) hierarchical linear model that can simultaneously examine the full range of within- and between-group variation contained in longitudinal data sets. The method is illustrated through an analysis of the delinquency rates characterizing Chicago's neighborhoods between 1930 and 1970.