Literature Review — Student Retention in UK Higher Education

Search strategy

Searches were conducted in Scopus, ERIC, the British Education Index and Web of Science between January and February 2026. The primary string was ("higher education" OR "university") AND ("student retention" OR "student attrition" OR "withdrawal") AND ("United Kingdom" OR UK OR England OR Scotland OR Wales), adapted to each database’s syntax.

Inclusion was restricted to peer-reviewed empirical studies published from 2010 onwards, reporting on UK institutions, and available in English. The 2010 boundary was chosen because the 2012 fee reform materially altered the financial context of withdrawal decisions, rendering earlier findings of limited comparability. Grey literature from the Office for Students, HESA and Advance HE was included separately, as UK sector-level retention data is largely published outside the peer-reviewed literature.

The searches returned 1,284 records, reducing to 61 studies after duplicate removal, title and abstract screening, and full-text eligibility assessment.

Theme two: the limits of integration theory in a commuter context

Tinto’s (1993) model of academic and social integration remains the dominant theoretical frame in retention research, and the UK evidence broadly supports its central claim that integration predicts persistence. The relationship, however, is considerably weaker in studies of commuter and mature students — a population that has grown substantially in the UK since 2012 and which the model was not developed to describe.

Thomas (2012) found belonging to be the strongest predictor of continuation across a large multi-institution sample, apparently confirming Tinto. Yet studies focusing specifically on commuter students report that social integration measures explain markedly less variance, with practical and financial factors dominating. This is not a contradiction so much as a boundary condition: integration theory assumes a residential campus experience in which social and academic life are co-located. Where they are not, the mechanism the theory describes has limited purchase — a limitation acknowledged only rarely in the UK literature and never, to this reviewer’s knowledge, tested directly.