Reading: Data-based decision-making

2. Support for teaching and learning decisions

Data analysis of information generated through students’ interaction with educational technologies has been recognised as a valuable approach that enables better insights and a deeper understanding of learning processes. This has encouraged the development of strategic, data-informed decision-making in education, which was crucial for the emergence of a new field of research and practice — learning analytics (Gašević et al., 2019).

Definition. For decades, teachers have monitored student progress, designed assessments, and used evidence to improve learning. However, learning analytics brings novelty and enhances these practices through the use of new digital data and analytical methods from data science and artificial intelligence.

Benefits and challenges. Learning analytics can be based on different types of data, which will be discussed in the following sections of the e-course. The emphasis is on data generated through interactions between students and information technology, primarily learning management systems (Divjak & Maretić, 2017). Data-driven learning analytics enables the creation of insights, decisions, and actions aimed at improving learning and teaching (Chatti et al., 2020). Insights into students’ learning processes and learning difficulties help teachers identify weaknesses in learning and teaching activities and provide constructive feedback that can guide and improve further learning (Gašević et al., 2015). Furthermore, learning analytics supports personalised learning (Chatti & Muslim, 2019) and the development of flexible learning pathways (Tsai et al., 2018).

In addition to improving the quality of learning experiences, learning analytics helps identify factors influencing student success and completion rates (Tsai et al., 2018), thus playing an important role in identifying students at risk of dropping out and in planning interventions aimed at maximising student retention (Susnajk et al., 2022). Learning analytics can also provide useful insights directly to students themselves by supporting self-regulated learning (Schumacher & Ifenthaler, 2018). Learning analytics, therefore, supports data-informed decision-making at multiple levels: from the institutional level to the level of teachers and, ultimately, students.

Despite the significant increase in interest among higher education institutions in using learning analytics to improve learning and teaching, as well as growing funding opportunities, research has shown that the implementation of learning analytics in higher education is still not systematic or widespread, and that challenges remain for effective data use (Tsai et al., 2018). There are various reasons for this, such as the lack of policies and strategic support, insufficient capacities for data-driven decision-making, and differing levels of stakeholder engagement (Tsai & Gašević, 2017), which affect the usefulness of learning analytics tools and the relevance of policies (Tsai et al., 2022), as well as institutions’ experience with learning analytics (Tsai et al., 2021). Furthermore, the challenge of trust in learning analytics is also present (Tsai et al., 2021).

An additional key challenge in creating the value of learning analytics is data quality, as the quality of outcomes largely depends on the quality of input data (Sheikh et al., 2022). Bias and unfairness that may arise from data affect the trustworthiness of learning analytics, especially when it comes to predictive models (Li et al., 2022).

Finally, even after adoption and significant investments in learning analytics, some universities have failed to achieve their strategic goals. It is therefore emphasised that using learning analytics to achieve strategic value requires investment not only in infrastructure, but also in staff professional development and strategic planning (Sheikh et al., 2022).

Relevant knowledge and skills related to strategic planning can be acquired in the first e-course of the programme “Management of the Digital Transformation of Educational Institutions”, entitled Strategic Planning for Digital Technology Implementation.

Connection with related fields. There are several “sister” research fields with which learning analytics shares similar goals. One of them is Educational Data Mining (EDM). Similar to learning analytics, EDM applies machine learning, data mining algorithms, and statistics to educational data (Dutt et al., 2017). However, EDM is more focused on technology and models, whereas learning analytics is more focused on learning itself (Sghir et al., 2023), although the two fields overlap more often than they differ (Rienties et al., 2020). Another field closely related to learning analytics is discussed in the following section.