17 Feb 2026
New appointee Peter Edelsbrunner investigates how the acquisition of knowledge can be statistically validated. In his teaching, he also introduces students to research methods.
Tests seem an easy way to sample knowledge. But whether or not a test genuinely measures what it is intended to measure is a question that preoccupies Peter Edelsbrunner. The recently appointed Professor for Methods of Empirical Educational Research at LMU, who took up this post in April 2025, conducts research into how studies in educational science arrive at their findings: “I look at which statistical method is best suited to which question,” he explains.
Peter Edelsbrunner has four monitors on his desk. He uses one of them exclusively for statistical analyses, but when dealing with large amounts of data, these are run on mainframe computers. | © LMU/Johanna Weber
To evaluate whether an educational policy measure works, traditional statistical models are, for example, used to identify a possible average effect. But Edelsbrunner goes a step further: What interests him is which intervention works better or worse for which individual child, and why.
The professor is keen to ensure that possible conclusions drawn from tests about the attributes of those who are learning are really true. Only then, he argues, can educational policy measures be rationally substantiated and properly implemented.
“At the same time, these insights lay the foundation for what we call adaptive teaching, which is tailored to individual children,” he explains. Interest has been expressed from many quarters at LMU – hence the wealth of opportunities to apply Edelsbrunner’s research question. Examples include the fields of educational-psychological research, special needs education and research into applied statistics.
Peter Edelsbrunner studied psychology at the University of Graz, the Austrian city where he grew up. He majored in differential psychology, which examines the differences between individuals. To pursue his doctorate, he moved to ETH Zurich in 2012. And this became his point of entry to what, for him, was the “completely new field of teaching/learning research”. His doctoral thesis was devoted to how children learn science and the question of how they build up their knowledge step by step.
After earning his doctorate, Edelsbrunner stayed on at ETH – initially as a postdoctoral researcher and later as a Senior Lecturer – ahead of a visiting professorship at LMU in 2024. Before switching to his current position here, he worked as a postdoctoral researcher at the Chair of Physics Education, where he was able to build on experience gained over many years: “I have now been involved in teacher training for more than ten years,” he notes. “So, I spend my time showing teachers how they can teach well. I am continuing to do this at LMU, where I train statistics tutors, working with them to practice the development, preparation and delivery of tuition for a year in the context of statistics and research methods.” The professor also shares his expertise in teaching research methods at LMU’s Open Science Center.
Edelsbrunner teaches statistical models and research methods as part of teacher training study courses, too. He devotes a lot of time to thinking about how these methods can best be communicated to the students. This semester, he picked a highly practical example for one of his lectures:
“We were investigating the purpose of the morning assembly that is commonly held at nurseries. To do so, I introduced a lot of research methods, from observation and interviews to questionnaires and experiments. I was trying to show the students what statistical and research methodological thinking is there for: to reliably validate assumptions we have about educational practice.” To date, students’ feedback regarding this approach has been positive.
Edelsbrunner’s research reflects his interest in deviations from certain statistical rules. For example, he explores how accurately and to what level of detail a statistical model must reproduce all the data it works with. “The results may differ depending on the model. So, you always have to ask what the intended purpose is.”
The professor’s most recent and, to his mind, most important publication tackles the statistical modeling of knowledge tests. “It looks at whether the tasks in a test have to correlate to one another,” he explains. “In many areas of statistical modeling, it is assumed that all exercises in one and the same test are closely interrelated, because they are all measuring the same thing. If you follow this logic, you should find that a person who correctly solves task 3 also does well with task 5. In knowledge tests, however, I and other researchers are increasingly questioning this assumption.”
In this publication, Edelsbrunner touches on a statistical index that has been firmly established for decades: Cronbach’s Alpha, which assesses how intrinsically consistent a test is. “That is both important and dangerous,” Edelsbrunner asserts. “On the one hand, our work encourages new ways of thinking. On the other, it is a fine line in statistics to call established parameters into question without losing the benefit of useful standards.”
The professor’s applied research concerns itself with the way learning individuals understand scientific thinking throughout their entire life. The process begins in early childhood with the question of how children systematically learn to find things out and test their assumptions. It continues at school: Among elementary school children, Edelsbrunner has, for example, studied whether there are later benefits from exposing them to age-appropriate physics tuition early on. “For instance, we asked about how learning is transferred, i.e. how early learning in science can generate positive stimulus for learning in later years. We were able to confirm that this can indeed be the case – even when secondary education is not explicitly linked to the same contents.”
Among older children, Edelsbrunner is very interested in discovering what they think about knowledge and science: “The direction of travel is increasingly toward media skills: How can children and young people find out whether a piece of information is reliable?”
In future projects, the educational researcher plans to link the acquisition of scientific knowledge to a general understanding of science, and to also teach children to question the scientific information that is disseminated in various media. “In the next few years, I would like to bring these two strands together, encouraging schoolchildren to think coherently about scientific matters. It would be great if we could then also foster an interest in science among schoolchildren who would otherwise not really care about it.”