DIGITAL LIBRARY
INSTRUCTIONAL REDESIGN BY LEARNING ANALYTICS
1 Humboldt Universität Berlin (GERMANY)
2 Technische Universität Chemnitz (GERMANY)
About this paper:
Appears in: ICERI2023 Proceedings
Publication year: 2023
Pages: 9137-9144
ISBN: 978-84-09-55942-8
ISSN: 2340-1095
doi: 10.21125/iceri.2023.2340
Conference name: 16th annual International Conference of Education, Research and Innovation
Dates: 13-15 November, 2023
Location: Seville, Spain
Abstract:
The use of learning analytics (LA) in schools can counteract the digital divide. The educational disadvantages and social inequalities can be tracked, and steps can be taken to solve them by integrating AI-supported learning analytics in the teaching-learning process in school. Therefore, the uncertainties and mistrust in dealing with new technical procedures should be overcome through knowledge and competence transfer. In this study, an effort has been made to introduce the concept of learning analytics and its possible embedment in the pedagogical practice of current and prospective teachers to reduce reservations and increase interest in LA.

We created LA learning units to provide educators, including those from the domain unfamiliar with technology, a fundamental introduction to LA and arouse their interest in integrating LA into their didactical practice. Three modules are developed with a view from everyday school life and pedagogical practice:
In the first module, the individual analysis levels of LA are explained with typical analysis procedures and the respective benefits for teachers in everyday school life.

The second module is the main module. It addresses the increasing challenges of teachers dealing with heterogeneous school classes and supporting students individually. In this module, LA is presented as a solution and the current greatest potential for generating adaptive individual learning paths. To enable teachers to use this potential, it is explained how their teaching can be formalized and modeled with the help of a domain model, learner model, and didactic model to control and design teaching/learning processes with LA quasi-co-creatively.

Starting from analog teaching, it is explained how learners are digitally modeled based on their characteristics or data, the subject matter is modeled as an ontology similar to a concept map, and a didactic model is created from the lesson planning or instructional design. This has enabled us to recognize that instruction can be designed by humans and machines according to the same principles. We have thus created a formal framework for combining analog and digital instruction and making it machine-readable to apply LA procedures.

So that teachers can design lessons within this framework and facilitate adaptive learning pathways, we explained didactic concepts and principles, encompassing Constructive Alignment as a prerequisite for the creation of adaptive learning paths with (Micro-)Learning Units, formulating appropriate learning outcomes, designing learning activities and reviewing learning outcomes.

In the third module, the design and use of LA dashboards are presented to observe and accompany individual learning states and learning achievements.
The modules are taught with Micro Learning Units and learning outcomes are formatively tested. In this study, we examine learner data sources, formative test scores, interaction, and forum data, and a survey at the beginning and the end of each to assess whether the learning objective has been achieved: whether uncertainties and mistrust have been overcome, reservations reduced and interests in LA increased, or whether we need to redesign the learning units on LA.
Keywords:
Learning Analytics, Teachers further education.