NSS results need more than headlines

This week, Dr Stuart Grey focuses on NSS 2026 results day: the national headlines, the disabled student gaps that still need action, and what it takes to turn large volumes of student comments into usable evidence quickly.

The episode also looks at what Student Voice AI produced for NSS customers on results day, why sentence-level labels matter, and how recent research should shape the way universities use machine-learning analysis and high-stakes teaching evaluation evidence.

In This Episode

Why stronger NSS national headlines do not remove the need for subgroup analysis.
What results day looked like across Student Voice AI's NSS customer work.
Why sentence-level comment analysis helps teams avoid flattening mixed student feedback.
How Valeriya Minakova, Ryan Patterson and colleagues frame machine learning as a complement to academic judgement.
Why Shaylen Stone, Gerard Jefferies, Megan Lee and Cindy Davis's wellbeing research is a warning about high-stakes evaluation use.
Why disabled student gaps in organisation, management, and student voice need comment-level evidence alongside the scores.

Student Voice Practice

NSS results become more useful when teams read national context, local score movement, and student comments together. Comment analysis should preserve enough detail for staff to see what students actually meant, while keeping outputs reviewable and proportionate.

Sentence-level labelling helps separate praise, concerns, and mixed experiences within the same comment. That makes the evidence easier to route to the right people and reduces the risk that one average score carries too much of the story.

Research Spotlight

Across the Sector

Practical Takeaway

Treat NSS results day as the start of evidence work, not the end of reporting. Use the national headlines, but read them alongside subgroup gaps and the comments students wrote in their own words.

Subscribe to The Student Voice Weekly: https://www.studentvoice.ai/blog/newsletter/

Transcript

Hi, and welcome to Student Voice Weekly. I'm Dr Stuart Grey, founder of Student Voice, and Wednesday was NSS results day!

Today I'd like to talk about the National Student Survey, because this week the results landed and, for many people in universities, that creates a very particular kind of day. There are sector headlines, local comparisons, subject-level scores, action planning templates, and a lot of pressure to say quickly what the results mean.

The national picture this year is broadly positive. The Office for Students published the NSS 2026 results, with a response rate of 71.8 per cent and around three hundred and sixty-two thousand responses. Teaching was 88.1 per cent positive in England, and student voice was 80.2 per cent positive.

So there is a genuine sector headline there. Students are, overall, more positive across the main themes.

But the more important signal is underneath that headline. Disabled students remained less positive across every theme, with the biggest gaps in organisation and management, and in student voice. That changes how universities should read the results. A stronger average does not mean the experience is working equally well for everyone.

The key question is: what are students actually describing underneath the movement in the scores?

For us at Student Voice AI, results day was a very compressed practical version of that problem. Across fifteen institutions, we processed evidence from twenty-five thousand NSS respondents. That meant forty-four thousand comments, split between twenty-three thousand positive comments and twenty-one thousand negative comments.

Under the surface, though, the more useful view was at sentence level. We were working with about ninety-six thousand comment sentences, roughly one hundred and twenty-five thousand category labels, and nearly two hundred and ninety thousand sentiment labels. We also generated over six hundred spreadsheet files and around two hundred summary reports.

The annual release arrives all at once, but the work it creates is not just about producing a headline slide. It is about turning a very large pile of student language into evidence that teams can actually use.

If that takes weeks, the score starts to do too much of the work. People build a story around the number before the comments have had a chance to explain what students meant. That is where student feedback can become a little bit hollow. You have data, but not enough conversation around it.

This is why sentence-level analysis matters. A single NSS comment can contain praise for teaching, frustration about feedback timing, and a complaint about timetable changes. If the whole comment is treated as one positive or negative thing, the analysis gets flattened. If the comment is broken into sentences, the different parts can go to the right themes.

So a sentence about enthusiastic lecturers can sit under teaching. A sentence about slow feedback can sit under assessment and feedback. A sentence about unexpected timetable changes can sit under organisation and management. The categories show what the comment is about. The sentiment labels show whether that part of the experience was positive, negative, or more mixed.

That does not remove the need for human judgement. It makes the human judgement better. The reason we do this is to give quality teams, planning teams, student experience teams, and academic colleagues a map of what students are saying, rather than a wall of text or a single average.

And that connects directly to one of the research pieces in the newsletter. Valeriya Minakova, Ryan Patterson and colleagues looked at machine-learning analysis of mid-semester teaching evaluation comments in large Canadian courses. What I like about the paper is that it does not treat machine learning as magic. It asks whether instructors actually found the output useful.

The answer was broadly yes, but with an important condition. Instructors valued the analysis when it surfaced themes they recognised, ranked issues in a way that helped them prioritise, and presented the results clearly enough for them to use. The labels, visualisations, and report design were not decoration. They were part of whether the analysis felt credible.

That is exactly the point for NSS work. It is not enough to say that comments have been categorised. The categories need to make sense to the people who know the course. The outputs need to be reviewable. Staff need to be able to look at the evidence and say, yes, that sounds right, or actually, this label needs refining, or this theme means something different in our context.

The second paper in the newsletter adds a warning that I think is especially important straight after results day. Shaylen Stone, Gerard Jefferies, Megan Lee and Cindy Davis looked at student evaluations of teaching and academic wellbeing in Australia. Their study is a reminder that scored and written evaluations can affect staff quite deeply, especially when they drift into promotion, appraisal, or performance narratives.

That does not mean universities should ignore student criticism. Students need to be able to say when something is not working. But institutions have to be careful about how comments are released, interpreted, and used. A single hostile remark should not become the story of someone's teaching. A small movement in a score should not become a crude judgement about a colleague's value.

So the responsible approach is to report patterns rather than isolated remarks. Look for repeated concerns. Keep the raw student language available for review where it is appropriate, but do not let one sharp comment dominate the interpretation. Separate developmental feedback from high-stakes judgement. And make sure that student voice work does not create avoidable harm for staff while trying to improve the student experience.

This is where the national NSS story and the comment-analysis story meet. If disabled students are less positive about organisation and management, or about whether their voice is heard, the next question cannot just be, how do we raise the score. The better question is, where does that weaker experience show up in students' own words?

It might show up in comments about adjustment processes. It might be about inconsistent communication. It might be about timetables, placements, access to learning resources, or whether students can see any response after they give feedback. Those are different problems. They need different owners and different kinds of action.

And this is also why averages are dangerous when they are left on their own. A university can improve overall and still have groups of students whose experience is not improving enough. A subject can look fine at the headline level and still have a very specific problem in assessment communication or course organisation. Comments are where the real value is, because they help explain what the number cannot.

So if I were looking at NSS results this week, I would want three layers in front of me. First, the national context, so we know what is happening across the sector. Second, the local score movement, so we know where our own subjects and services have shifted. And third, the comment evidence, broken down in a way that shows what students were actually talking about.

That third layer is the one that often gets squeezed. It is also the one most likely to tell you what to do next.

The practical implication is quite simple. Treat NSS results day as the start of evidence work, not the end of reporting. Use the national headlines, but do not stop there. Read the comments through the lens of subgroup gaps. Separate broad themes into the actual issues students raised. Keep the analysis reviewable. And try and get the evidence to the people who can act while there is still time to shape plans for the next year.

So the takeaway is this: stronger NSS headlines are welcome, but the improvement work sits in the comments. The job is to turn those comments into careful, traceable evidence without losing the human judgement that makes the evidence useful.

That is it for this week. The full set of links and written summaries is in the Student Voice Weekly newsletter. If you work with student feedback and want the research, regulation, and sector signals in one place each week, you can subscribe at studentvoice.ai. And if this was useful, please share it with someone who works on student experience, teaching, quality, or planning, or leave a review in your podcast app.