ABSTRACT
Crime of the residents is adequately described by one single factor, and this factor has very strong bivariate relationships with certain independent properties. The crime factor has been found to be cross-sectionally and bivariately related to several independent properties in similar ways in the reference years of 1980, 1985 and 1990. The longitudinal relationships have been analyzed in various ways, by considering linear and product models, point and change data, different time periods, micro- and macro-level interaction, and data for individuals with different patterns of moves. Linear analyses have been performed of the change in crime using changes in various independent factors as regressors, including changes in aggregated micro-level bivariate interaction within and between individuals, for the undivided area population as well as for the long-term residents and the newcomers. The cross-sectional relationship can be assumed to describe the total of the relationships caused by their influences on crime and their non-causal relationship with crime.
