Data-informed teaching in practice at Ghent University and the University of Strasbourg

Asking learners to answer a question on their phone or laptop has become everyday practice, in lecture halls and training rooms alike. But a set of answers is only a starting point. What matters is what those answers reveal about learning, and what you choose to do with them.

In our webinar Data-informed teaching and training, Krijn Boom (Educational Support Officer and Digital Learning Policy Officer, Ghent University) and Vincent Delair (lecturer at the University of Strasbourg and trainer at the CFA de l'IUMM) shared how they read that data at two very different scales: across a university, and inside a single session.

1. It starts by making the signal exist

Before any dashboard, there is a more basic question: does the data exist at all? In a traditional session, teachers and trainers read the room through a few nods, a silence, and the same two hands going up. Whatever the rest of the group thinks stays in the room.

Active learning changes that. When you ask a question and everyone answers, what used to be invisible becomes something you can read. That includes the learners who would never have spoken up, because they were nervous, unsure, or simply never got the chance. Anonymity plays a key role here: it makes it safe to give a wrong answer, which is precisely the answer you need to see.

This is why Wooclap doesn't just collect data: it creates learning data. And the value of that data depends on a pedagogical choice made long before any dashboard is opened: what you decide to ask, and how. A question that checks understanding produces a very different signal from one that only checks attention. With more than 20 question types available, question design is where data-informed teaching really begins.

2. Data informs, people decide

At Wooclap, we say data-informed, not data-driven. Data is one input among others: it can't capture a group's history, an organization's culture or the mood in the room. A dashboard shows you where to look. It doesn't tell you what it means, and it doesn't make the call.

That interpretation stays with the people who know their learners: teachers and trainers in the session, learning designers and digital learning teams across the institution or the company.

It also helps to separate two kinds of data:

• Engagement data (how many joined, how many answered) shows that a session was active, but says little about what learners understood.
• Learning data (which questions caused confusion, who has mastered a topic and who hasn't yet) shows what was really understood, and what to revisit next.

Attendance, completion and participation rates tell you that an activity happened. Learning data tells you what learners actually took away from it. Used this way, data serves learners first: it gets them support when they need it, rather than simply keeping an eye on them.

3. Uncovering trends across a large campus

Ghent University has 11 faculties and around 50,000 students. It has been at the forefront of active teaching for more than a decade, and has used Wooclap for six years. For Krijn Boom, data is knowledge: it is the language he uses with the board and policy advisors to get decisions made.

While analyzing annual usage patterns, the digital learning team spotted around 90,000 responses during the month of January. Yet in January, the university is more or less closed: there are no classes, and students are preparing for their exams.

Most of these answers came from self-paced questionnaires that students complete on their own time. Krijn has two hypotheses: teachers may be using them to assess their classes, or offering them as a way to self-assess before exams. His next step is to identify the teachers behind these questionnaires and ask them why, so that a strong use case can be shared across the university.

A good example of data showing where to look, without telling you what it means.

4. Data as a compass: a practical 3-step method

Vincent Delair learned the value of data in industry, where he built corporate training programs before teaching at university and training apprentices. Back then, he lacked a reliable way to assess his trainees' level. Today, he uses learners' answers as a compass: every answer, and every missing answer, shows where each learner stands on the way to the training goal.

His method follows three steps, which together combine diagnostic, formative and pre/post assessment:

1. At the start: a diagnostic. After an icebreaker, a short questionnaire built on the training objectives reveals the group's starting level. Vincent uses it to balance groups for collaborative activities and to identify the learners who will need more frequent attention.
2. During the lesson: a learning questionnaire. He checks the same concept in different ways, using several question types, so that a lucky guess can't pass for understanding. The results show what landed and what didn't, so he can target remediation and adjust how he teaches. Sometimes he follows up with a short group activity where teams present their answer in 180 seconds, then debrief together.
3. Before exams: the same diagnostic, again. Comparing results with the starting point, and with the training target (85% in his example), shows how effective the training has been and which learners need a final push. Learners also see their own progress.

5. Simple tools to analyze and improve

Krijn and Vincent rely on two Wooclap features, one for each scale.

• For administrators: Admin Analytics. Found in the Admin Dashboard, this self-serve feature shows how Wooclap is used across the institution or organization: active presenters, events, answers received and trends over time. Its pedagogical layer goes beyond volume: the Content view shows whether teams use Wooclap to check understanding, gather feedback or present information, and which question types they rely on most. Krijn’s team uses it in its regular reviews of digital learning tools, and this is how the January surge became visible.
• For teachers and trainers: Session Overview. As soon as a session ends, this AI feature gives a clear read on participation, question accuracy and score distribution, to separate those who have mastered a topic from those who haven’t yet. AI recommendations then suggest what to try next time, based on your questions and your learners’ answers: a new question type, a topic worth revisiting, or a setting such as the “I’m confused” button, which lets learners signal in real time where they need support. Vincent uses these recommendations to improve his questionnaires. Session Overview doesn’t decide for you: it gives you a starting point.

6. The same signal, two scales of decisions

Krijn and Vincent work at very different levels, but they read the same signal. In a session, it tells a teacher or trainer what to re-explain now and what to redesign for next time. Across an institution or a company, it shows where teaching practice actually needs support, rather than where you assumed it did.

As Krijn put it, the age of teaching from the ivory tower, repeating the same content every year because it has always been done that way, is over. Teachers and trainers want to listen to their learners and keep improving. Data is one of the ways they do it, as long as it informs their judgment rather than replacing it.