How can universities of applied sciences get more out of student evaluations?

Students fill in questionnaires, programme committees meet and quality staff write reports. Still the question lingers whether all that measuring leads to better education. I wanted to know how it works behind the scenes, and between March and June 2026 I spoke with thirteen people who do this work, from a lone quality officer at a small teacher training college to a policy adviser at a large university of applied sciences.

What kept coming back

Five themes came back in almost every conversation. These are the main ones.

1. Open answers go unread. All thirteen named processing open answers a bottleneck. One of them analyses hundreds of comments from the National Student Survey by hand every year. Others anonymise the text themselves and run it through an AI tool, or leave it altogether.
2. The feedback loop rarely closes. Eleven of thirteen found the reporting back to students weak. The chain breaks in several places: between collecting and analysing, between analysing and deciding, and between deciding and a visible change.
3. Satisfaction is not the same as quality. Seven of thirteen raised this themselves. At the same time, one of them stressed that the student's view remains fully valid.
4. Scale decides everything. With fifteen students you simply read the open answers yourself. With hundreds it is a different story.
5. AI is wanted, but the organisation holds back. Eight of thirteen wanted AI that groups open answers into themes. Privacy, procurement and policy often mean it hasn't happened yet.

“The collecting all works out fine, but what is done with it exactly, well.”

Two gaps, not one

The analysis gap is the largest: open answers are a gold mine that can't be mined by hand. But there is a gap on the collecting side too. A questionnaire with fixed questions often yields shallow answers, like "fine" or "good", and then there is little to analyse. Whoever can follow up better while collecting has less to make up for on the analysis side.

What you can do on Monday

You don't have to replace your whole evaluation system. A few things you can do right away:

• Pick one course or term to try, ideally an evaluation that was planned anyway. A small, scoped trial says more than a plan.
• Add one open question that asks why, and read the answers within a week. Then you can still adjust before the term ends.
• Decide up front what you report back and to whom, so the loop doesn't end at the report.
• Keep your standard questionnaire for comparison across years, and put the conversation next to it.
• Involve the privacy officer early, especially if you want to use AI. See also the privacy checklist.

Febaso grew out of this research: an analysis and conversation partner that groups open answers into themes with counts and quotes, and asks a follow-up where the why matters. It works alongside your existing evaluations, not instead of them.