Exploring Residential Mobility As a Spatial Practice Via Qualitative Interviews, Quantitative Survey and Agent Based Modeling

The paper reports on a collaboration of sociologists and landscape ecologists in understandingfuture residential patterns in a medium sized shrinking city in Germany. Shrinking cities are an especially interesting case here, because numbers of inner city relocations are in general higher than in other cities (Sturm et al. 2013 for Germany). This is due to housing over-supply with more and cheaper offers being available. Like this, patterns of socio-spatial segregation can alter faster.

In a step-wise design, we included inductive and deductive steps, qualitative and quantitative approaches and finally the construction of an agent-based model to simulate futurehouseholds residential mobility patterns. Firstly, we explored mobility practices by means of in-depth interviews from which we derived context-appropriate information on the orientations and decision criteria of residents. Here, space is included as the life worlds of interviewees.

Secondly, from these results, we constructed a classic questionnaire survey to achieverepresentative data on the decision criteria among different social and demographic groups. In this step, administrative spatial boundaries were introduced in order to standardizequestionnaire responses. From the survey results, thirdly, we develop rules of behavior of households and simulate their decisions in an agent-based model (ABM) as a method to understand patterns emerging from individual behavior. Thus, ABMs use abstract spatial categories and typologies of spatial practices of agents (households in our case). In ABMs, space can be both an independent variable (thus influencing behaviour) and a dependent variable (if the space is influenced by human actions) In our case, space is an independentvariable in that location choice depends on the socio-spatial structures and the images attached to certain places and neighbourhoods, on distances to work-places, social networks, infrastructure access etc. Space is a dependent variable when analyzing collective householdsmigration patterns and the changed patterns of socio-spatial segregation.