Twenty years of student surveys reveal changing attendance and experience

Updated Sep 07, 2026

HEPI published What Matters Most? 20 years of the student experience on 14 May 2026, jointly with TechnologyOne. Gosia Turner and Rose Stephenson analysed existing Student Academic Experience Survey data; this was not a new survey or a change to NSS. Publication details.

Reading the long-term evidence

The dataset contains 206,512 responses across survey waves. It is not a panel following the same students for 20 years. Coverage began with first- and second-year undergraduates in England, expanding in 2012 to later-year undergraduates across the UK. Some years had no survey, and the analysis excludes 2009.

The report uses annual weights and harmonised variables. Some models use only recent years because their questions were introduced later. It describes associations between students’ views of teaching, feedback, belonging and their wider experience; these should not be presented as experimentally established causes.

Respondents reporting attendance at all scheduled classes fell from 63% in 2006 to 48% in 2025. The mean gap between scheduled and attended teaching was just over one hour initially and 2.4 hours in 2025. These are self-reported hours. Changing delivery, employment and commuting are possible context, not a complete causal explanation. Report: methodology, pages 23–26; attendance, pages 61–63.

Questions for local investigation

Our practical suggestion is to use the findings to frame questions rather than diagnose an institution. A national association cannot identify why a particular student missed teaching or why a course’s value-for-money ratings changed.

Ask how timetables, access, work commitments and teaching arrangements interact in the local context. Include students whose circumstances may make participation difficult, and distinguish reported experience from administrative records.

Keep questions and cohorts visible when comparing surveys. Do not combine NSS, postgraduate surveys and local module evaluations into a single trend simply because comments use similar words.

The NSS open-text analysis methodology can help plan review of a suitable comment set. Comments can suggest explanations to investigate, while a claim that an intervention improved attendance or belonging needs an appropriate evaluation.

Review note, 7 September 2026: clarified repeated survey waves, changing geographical and year-group coverage, weighting and self-reported attendance. Replaced causal “drivers” language with bounded interpretation of associations.