Zurich Workshops on Learning in a Digitalized World

Adaptive teaching has become an important response to the increasing heterogeneity among students and the wish to address diverse learner prerequisites. It involves systematically adapting instruction and materials to students’ needs in order to support learning processes more effectively. In practice, however, adaptivity poses significant challenges. Teachers must continuously assess learners’ needs, monitor their progress, and adapt instructional strategies, which can be cognitively and organizationally demanding. Digital technologies offer promising support for adaptive learning and teaching. Technology-based tools can assist in diagnosing learner characteristics, tracking learning processes, and providing automated feedback, thereby enabling more flexible and scalable forms of adaptivity.

In this workshop, you will explore the conceptual foundations of adaptive learning and teaching. You will examine what adaptivity means from both theoretical and instructional perspectives, how it differs from related approaches such as individualized or personalized learning, and how digital technologies can be used to design adaptive learning environments. We will also critically reflect on the potential implications of adaptive learning for educational equity. While adaptive systems offer promising opportunities to support diverse learners, they may also involve unintended consequences, including the possible reinforcement of existing inequalities.

By participating, you will gain both conceptual and practical insights into technology-based adaptivity. The workshop aims to support you in reflecting on applications in educational practice as well as within your own research design.

Intelligent Tutoring Systems (ITS) are an educational technology that has proven to be very effective in helping students learn. They guide students during problem-solving practice, enabling them to follow their own path while solving a given problem and providing individualized stepwise feedback. Many tutoring systems also support individualized mastery learning, in which each student gets assigned their own, individualized set of problems sufficient to master targeted skills.

ITS are associated with the field of  ‚AI in Education‘. A popular kind of tutors are Cognitive Tutors. These tutors are based on a cognitive model that reproduces the mental processes involved in solving a task. A cognitive model does justice to the fact that students can successfully solve one and the same task in different ways. They are thus very useful for many students, including low-performing students, who are known to have difficulties in adapting to different ways of solving problems.

In this Summer School, you will learn concepts of ITS and you will learn to build them. You will develop a prototype of a computer-based intelligent tutor. To do this, you will use the CTAT (Cognitive Tutor Authoring Tool) developed by researchers at Carnegie Mellon University (CMU).

CTAT was developed for both non-programmers and programmers, which is why we offer two tracks during the Summer School. In track one, using behavior graphs requires no programming experience. If you can already program, you can dive into using Nools.js to build so-called rule-based tutors in track two.

During the Summer School, you will work on your own project in your area of interest. CTAT tutors usually focus on solving multi-step problems, such as those often found in mathematics, physics, or chemistry, but they are also used in language learning. Over the course of the week, you will first conduct a cognitive task analysis to understand the types of problems for which your tutor may be used and to understand students’ processes for solving them. Subsequently, you will use CTAT to implement an intelligent tutor. By the end of the week, you will have a running prototype of your tutor. During the Summer School, you will be mentored by experts from CMU (who developed CTAT) and PHZH.

Interventions in research projects often involve pre- and post-tests to determine whether changes have occurred between two measurement points. Digital learning environments, in particular, offer the opportunity to examine what happens in between these points. For example, tutors developed with CTAT (see track 2) store all interaction steps in a database, and learning management systems (LMS) such as Moodle generate large amounts of log data. Special procedures are needed to analyze this vast quantity of process data, which we will examine more closely in this track. In addition to data preparation and analysis, we will also address the interpretation of the data.

The methods we examine can also be applied to data outside digital learning environments, such as grade progression in math tests or the number of training sessions per week.

We will work with various packages in RStudio. Basic knowledge of R is helpful but not required. To facilitate exchange among participants, we will primarily work with shared datasets. However, there will also be sufficient time to discuss your own project.