Geodemographics and spatial interaction: an integrated model for higher education
A. D. Singleton; A. G. Wilson; O. O’Brien (2012). Journal of Geographical Systems, 14(2), 223-241. DOI: 10.1007/s10109-010-0141-5
Abstract
Spatial interaction modelling and geodemographic analysis have each developed as quite separate research traditions. In this paper, we present an integrated 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 providing a quantitative tool that can be used by HE stakeholders to explore policies relating to such topics as widening access to under-represented groups.
Extended Summary
This research develops an integrated model combining spatial interaction modelling with geodemographic classification to predict student flows from English schools and colleges into higher education institutions. The study addresses persistent inequalities in university participation rates between different social groups, despite overall growth in student numbers. The research integrates comprehensive national datasets from the Department for Education’s National Pupil Database, Learning and Skills Council records, and Higher Education Statistics Agency data, covering students who qualified with A-levels or equivalent qualifications during 2005-2007. These data were coded using the Office for National Statistics Output Area Classification, which categorises neighbourhoods into seven geodemographic ‘Super Groups’ based on social and economic characteristics. The model framework divides England into 150 local authority zones as origins and 88 universities as destinations. A key innovation involves calibrating distance decay parameters for each geodemographic group, reflecting different travel behaviours and higher education participation patterns. The research found that students from ‘Prospering Suburbs’ areas travel greater distances to attend university, whilst those from ‘Multicultural’ neighbourhoods show sharp distance decay in participation. The model incorporates university attractiveness factors based on rankings and capacity constraints, using both singly and doubly constrained spatial interaction equations depending on whether institutions have surplus places. Testing reveals generally encouraging results, with the model successfully predicting aggregate flows to institutions like the University of Manchester and from areas like Norfolk. However, some discrepancies emerge, particularly under-predicting flows from affluent areas that may reflect regional income variations not captured in census-based geodemographic classifications. The research demonstrates how behavioural insights from neighbourhood classifications can enhance traditional spatial models, creating a more nuanced understanding of educational access patterns. This integrated approach offers practical value for higher education stakeholders by providing a quantitative tool for exploring policy scenarios around widening participation and understanding geographic inequalities in university access. The model represents an important step towards evidence-based policy making in educational planning, though further refinement is needed to incorporate subject-specific factors and alternative impedance measures beyond simple distance.
Key Findings
- Geodemographic groups show distinct distance decay patterns, with ‘Prospering Suburbs’ students travelling further than other neighbourhood types.
- The integrated model successfully predicts aggregate student flows between local authorities and universities across England.
- Students from affluent areas travel greater distances to university than traditional spatial models would predict.
- Combining spatial interaction modelling with geodemographic classification provides enhanced behavioural insights for higher education policy.
- The model framework offers a practical tool for testing widening participation scenarios and understanding geographic educational inequalities.
Citation
@article{singleton2012geodemographics,
author = {A. D. Singleton; A. G. Wilson; O. O’Brien},
title = {Geodemographics and spatial interaction: an integrated model for higher education},
journal = {Journal of Geographical Systems},
year = {2012},
volume = {14(2)},
pages = {223-241},
doi = {10.1007/s10109-010-0141-5}
}