Spatial interaction modelling and geodemographic analysis have each developed as quite separate research traditions. In this paper, we present an inte- grated model that harnesses the power of spatial interaction modelling to behavioural insights derived from a geodemographic classification. This approach is applied to the modelling of participation in higher education (HE). A novel feature of the paper is the integration of national schools, colleges and HE data; a national model is then calibrated and tested against actual recorded flows of students into HE. The model is implemented within a Java framework and is presented as a first step towards pro- viding a quantitative tool that can be used by HE stakeholders to explore policies relating to such topics as widening access to under-represented groups.