Neighbourhood Composition and Educational Aspirations: Heterogeneity, Scaling, and Heterogeneous Scaling of Spatial Effects.
Contextual effects can be heterogeneous in terms of a “classical” interaction; different groups can be differently exposed to or affected by the mechanisms that mediate particular contextual impacts.
Spatial dimension adds an additional challenge in terms of the scaling of contextual effects; In particular, research on the modifiable areal unit problem (MAUP) has shown that the definition of the area where the explanatory factors are measured typically affects the results, even when the definitions are relatively similar.
In addition there might be the combination of both challenges: heterogeneity of scaled effects. It implicates that group specific contextual effects are not varying constantly over the whole geographical scale of measurement but groups show different scale shapes.
The aim of this paper is to address these challenges using the example of the impact of social composition of the close living environment on young people’s educational aspirations. In previous research, neighbourhood conditions have frequently been considered as among the main explanatory factors for educational outcomes.
We make use of ego-centred areas of different radii to measure contextual indicators on different scales and vary effects of the living environment with respect to individual characteristics such as social and migration background. For this purpose we utilise individual data of German Socio-economic Panel combined with georeferenced small-area context information on population’s academic composition. We will present our aggregation approach and provide technical tools developed by our research group (aggind.ado).
We can observe clear associations between the quality of the neighbourhood and academic ambitions; however, only on the close range for non-immigrants from non-academic family background.