Modelling residential mobility behaviour using a commercial data set: An analysis of mover/stayer characteristics across the life-course

Residential mobility is a key mechanism in the evolution of local population size and structure and is of importance to policy makers tasked to provide resources and services. However, while the broad spatial and compositional characteristics of (aggregate) migration flows are fairly well understood, a greater understanding of the more personal (individual-level) characteristics of movers and non-movers, for instance their neighbourhood satisfaction, household income and/or plans for a future moves, is essential if we are to fully understand the processes and patterns behind residential mobility and immobility. This paper exploits a bespoke commercial data set, Acxiom’s Research Opinion Poll (ROP), for the analysis of individual residential mobility behaviour across the life-course. In doing so, it uncovers some interesting associational patterns specifically related to some of the characteristics of movers vis-à-vis stayers that have, until very recently, been seriously understudied due to the lack of suitable data. However, since the analysis draws on a commercial data set hitherto unused for population analysis, the first part of the paper is concerned with investigating whether there is a practical need for sampling weights, designed to account for the unequal probabilities of selection in a sample for which the user has no prior information on the sampling design/strategy employed. The comparison of like-for-like weighted and unweighted binary logistic regression models suggests a good deal of stability and reliability across the data, but particularly for the model estimates derived from the pooled (combining 2005, 2006, 2007) ROP data, where the effect size and directional relationships are in close agreement. The substantive analytical focus in the second part of the paper capitalises on the confidence demonstrated in utilising pooled data, and the associated practical advantages gained with increased sample size and an inherently flexible data source, to explore how the complex and interlinked micro-level characteristics of movers and non-movers vary according to an individual’s life-course stage. One important conclusion from this analysis relates to the relative unimportance of what are traditionally thought of as labour market characteristics. In contrast, however, characteristics associated with the housing market are found to be of great substantive relevance. The paper suggests such findings are likely to occur as a result of measuring movers as a single homogenous group, irrespective of the distance travelled between origin and destination residence. Moreover, a focus on the more some of the less commonly observed behaviours/characteristics of (non)movers uncovers results worthy of attention. Future plans to move are found to be negatively associated with mobility, especially for those in their early adulthood, something which, at first sight, appears to contradict the cumulative inertia hypothesis. Furthermore, across the life-course, greater neighbourhood satisfaction is found to be consistently and rather strongly associated with those who have recently moved as opposed to those who remained in situ. Yet interestingly, all things being equal, a positive additional effect is associated with homeowners with a negative additional effect for renters regardless of type. The paper concludes by suggesting that reliable approximations for directional associations can be drawn from the ROP without the need for sampling weights; and calls for the analysis presented here to be extended, both technically and analytically, through the use of a multilevel statistical framework