This paper is concerned with the idea of invariant preference structures. At the outset, it is argued that empirical tests on the transferability of preference functions to other regions are necessary to substantiate the claim that behavioural models represent valuable ways of assessing the impact of environmental planning programs. Subsequently, the findings of an empirical study of spatial shopping choice behaviour in two Dutch regions are considered. It is demonstrated that, using Rushton's preference scaling methodology, the preference functions pertaining to one group of respondents in one particular region can be used successfully to predict aggregate spatial shopping patterns of an entirely different group of respondents living in a ...