Do historical values of independent properties add something to the explanation of crime – and if they do, what do they add? In order to find out, this study analyzes longitudinal relationships expressed in linear models using point-of-time (here simply “point”) and change data and expressed in product models using point data. The linear models will be of ordinary type – the quasi-linear models have shown to give explanations of crime that are not much better than the explanations given by ordinary linear models, and they will therefore not be used.