Abstract:

Housing affects who can participate and thrive in higher education, yet research and policy often treat affordability and accessibility as separate issues. This study links the pre-market search stage to lived housing experience and spatially uneven equity among university students. Using a cross-sectional survey from a German university city (socio-demographics, housing search, costs, dwelling attributes, travel, Likert satisfaction, and open-text feedback), we geocode residences and derive network-based travel times and amenity proximity from open data.

We operationalize:
(1) choice scarcity during the search—application intensity and whether students had more than one viable option at acceptance;
(2) a Composite Satisfaction Index combining Likert scales with multilingual sentiment from open responses; and
(3) a Composite Equity Index integrating affordability, accessibility, and satisfaction.

Models include logistic regressions for choice-scarcity outcomes, effect-size comparisons on satisfaction with multiplicity control, global OLS for equity, and Geographically Weighted Regression (GWR) to reveal local penalties.

Results show that international students submit substantially more applications and are less likely to have multiple viable offers at the moment of acceptance. They report lower satisfaction on both scales and sentiment, even when mean objective access looks similar—suggesting hidden burdens (compromise on quality, crowding, or management). The Equity Index is lower for international students; while rent and room size have stable associations, the penalties for international status and travel time vary by neighbourhood. GWR maps identify corridors and districts where disadvantages concentrate—often where bridge crossings, headways, or market tightness amplify small geographic frictions.

Implications are actionable and place-sensitive:
(i) reduce choice scarcity upstream;
(ii) target support where local penalties are strongest; and
(iii) prioritise quality/management improvements that drive dissatisfaction.

Methodologically, the paper offers a reproducible workflow that integrates sentiment, composite indices, and spatially varying coefficients to move beyond single-indicator diagnostics.

Keywords:

Student housing, higher education, equity, residential satisfaction, affordability, accessibility, sentiment analysis, geographically weighted regression.