Optimising online course design and student engagement using learning analytics

Gediminas Lipnickas, Gosia Ludwichowska-Alluigi, Joanne Harris, Bora Qesja

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

In recent years, higher education has experienced a rapid transition to online learning due to the Covid-19 pandemic and increased access to technology globally. As universities continue to shift to online delivery, an improved understanding of how students’ engagement, performance, and retention are affected by the online learning design is needed.

Although past studies attempt to investigate the relationship between students’ engagement and their overall course performance (Barthakur et al., 2021; Matcha et al., 2020) to date, few studies have examined specifically what course design elements within a 100% self-regulated online learning environment promote students’ engagement and improve their performance. Past studies use learning analytics (LA) to measure students’ interactions with the online environment but with limited application of learning theory (Lockyer et al., 2013), and only in traditional university courses with an online component (i.e., blended), or courses that transitioned to online delivery (Jaggars & Xu, 2016). While general guidelines for online course design exist, there is limited understanding of what design elements affect students’ engagement to achieve better learning outcomes in a 100% online learning environment at a program level (Barthakur et al., 2021).

This preliminary study precedes a large-scale study across multiple programs to address this gap. Drawing on learning analytics data we examine UniSA Online courses which are delivered 100% online and share the same learning design template, making it easy to identify differences in students’ engagement patterns and their link to performance. All course design elements and students’ activities for two business courses from 2018 to 2022 were coded using the Open University Learning Design Initiative (OULDI) design framework (Canole, 2010). Regression and cluster analyses revealed the types of activities and their impact on student engagement and performance in the online course environment. While engagement is measured by students’ attempts at learning activities and accessing learning resources on the course website, performance is captured by students’ grades.

We found that online courses tend to be characterized by primarily assimilative content, but different types of assimilative content and tasks show varying degrees of student engagement. Students primarily engage with course forums, mainly reading posts of other students and the facilitators’ feedback. Assessment-related resources were the next most viewed elements of the course as students accessed those resources multiple times. The third most accessed content was productive content in the form of non-graded quizzes allowing students to practice and review feedback. Viewing course content videos had the highest impact on student grades, followed by assessment instructional videos. Productive content had a significant impact on student performance despite some of this content not being among the most engaged with.

Extending the study across multiple programs and cohorts will contribute to the identification of more effective online design practices to improve the quality of online higher education, meet the changing needs of students, and achieve better learning outcomes.
Original languageEnglish
Title of host publicationRethinking and reshaping higher education
Publication statusPublished - 27 Sep 2022
EventHERGA Conference 2022: Rethinking and reshaping high education - University of Adelaide, Adelaide, Australia
Duration: 27 Sep 202227 Sep 2022
http://www.herga.com.au/

Conference

ConferenceHERGA Conference 2022
Abbreviated titleHERGA 2022
Country/TerritoryAustralia
CityAdelaide
Period27/09/2227/09/22
Internet address

Fingerprint

Dive into the research topics of 'Optimising online course design and student engagement using learning analytics'. Together they form a unique fingerprint.

Cite this