At least three major outcomes of the intra-urban migration process can be identified and modelled. These include the decision to move or stay, the nature of housing chosen by mover households, and locational choice. Empirical models are formulated and tested using data from a household mobility survey undertaken in Christchurch in 1974. Particular emphasis is given to formulating structural models of locational choice, guided by the following general premises: 1. that relocation is undertaken within and is constrained by the existing urban structure, and 2. locational choice behaviour is regarded as the outcome of the interaction of two basic factors: the behavioural propensities of individuals (e.g. preferences) and the constraints (both area and household level) confronting those individuals. Discriminant models of the type: L = f (P, C) are generated in which measures of household preferences (P) and constraints (C) combine to allocate mover households among locations (L) within the city. The suggestion is that urban spatial structure and the behaviour of individuals and households within that structure be considered together, and that models of locational choice should be based on behaviour - structure linkages. Locations are classified via a principal components-hierarchical grouping analysis route, using an array of population and housing variables, In addition to the identification of variables necessary for an explanation and prediction of household locational choice, important features of the modelling exercise include: 1. an examination of the influence of vacancy structure and pattern on model performance 2. an assessment of the influence of area homogeneity on the explanatory and predictive power of a location allocation model, 3. an investigation of the role of an area's location within the city system in locational choice modelling, and 4. model validation.