Student behavioural profiles in blended learning courses

By Daniel Johnston

Updated Apr 23, 2026

Blended learning gives students more ways to engage, but it also makes disengagement harder to spot. If instructors want to improve participation, they need to understand how students actually use online materials, not just how they say they use them.

For insight into students' self-regulation in this context, the work of Lust et al. [1] is a useful starting point. There is also a growing body of research on software that helps educators examine engagement directly. One example is the CLAS tool, which was the subject of a previous Student Voice Blog article on collaborative learning and student engagement [2].

That distinction matters because survey-based studies can be distorted by "social desirability bias", where participants give answers that make them look conscientious rather than answers that reflect what they really do [3]. In practice, some students may describe themselves as engaged "good students" while behaving quite differently in the learning environment. The paper considered in this article avoids that problem by examining the behavioural profiles that emerge when students use a video annotation tool in a blended course.

Studying students' interaction with a video annotation tool, Mirriahi et al. [4] set out to answer three practical questions:

Are there specific archetypes that emerge from analysis of students’ engagement?
Does the delivery method of a course help to develop students’ behaviours?
If the archetypes are found for the first question, do these translate into notable differences in attainment?

The researchers used the CLAS tool, which featured in an earlier Student Voice Blog article. All participants came from the same higher education institution in North America, studied the same discipline, and had no prior experience of using CLAS or a similar tool. Their course, like many blended courses that depend on clear timing, accountability, and workload control, had two phases: phase 1 focused on developing knowledge and understanding, while phase 2 focused on applying that phase 1 content. When students watched a video lecture, the researchers recorded the following data for each student.

How often they fast-forwarded each video
How often they rewound each video
How many times they watched a video continuously, without pausing, fast-forwarding, or rewinding
How many times they paused each video
How long each video was played for

For the final measure, the authors note that play time does not guarantee attention, because the software can only track mouse clicks. After collection, the data was analysed using hierarchical cluster analysis.

Comparing each student's graph with the wider group revealed four distinct behavioural archetypes. These clusters, A to D, were labelled "minimalists", "task-focused", "disenchanted", and "intensive" as a shorthand way to describe different patterns of learning engagement. The authors note that it is possible to divide the participants into more groups, but overlap increases when too many categories are used.

Cluster A: The Minimalists

These students made up the second largest group. Based on their behavioural metrics, they engaged much less actively with the video content, especially when it came to making annotations. The researchers suggest that this is not necessarily a sign of laziness. It may instead point to a more social or less tool-dependent approach to learning.

Cluster B: The Task-Focused Students

This was the largest cluster, accounting for 37% of participants. In contrast to the minimalists, these students showed the highest overall level of engagement with the video lectures.

Cluster C: Disenchanted Students

Encouragingly, this was the smallest cluster, representing 13% of participants. As the name suggests, this group showed relatively low interaction with the video content compared with the task-focused students. They did engage at first, but that effort faded over time, leading the authors to comment on their "limited sustained effort compared to students in clusters B and D".

Cluster D: Intensive Students

This group was also relatively small, only slightly larger than cluster C. Its defining characteristic was sustained effort and apparent internal drive. Most notably, these students engaged more actively with the video content than any other cluster, except on the fast-forwarding measure.

Across the different courses studied, the researchers were also able to address their second question. Students who were not incentivised, through grading, to engage with the content were less likely to align with clusters B or D. Instead, they were more likely to adopt a minimalist or disenchanted pattern. The authors also observed that if a student was incentivised during phase 1 but not phase 2, they did not usually maintain their earlier behaviour. This supports the wider point, also cited in [4], that engagement depends on external factors such as course design and delivery mode as well as internal factors.

What should educators take from this research? First, consistency matters. Students appear to benefit academically when expectations, incentives, and delivery methods remain stable across classes and across the year. Second, if video lectures are going to remain part of blended or fully online delivery, instructors benefit from understanding how student engagement can be sustained in online modules and which engagement patterns are emerging in their own classes. That makes it easier to design better interventions, tailor support, and avoid treating all disengagement as the same problem. Implementing video annotation tools and standardising teaching practices is not simple, but this study suggests the payoff can be stronger engagement and better outcomes.

FAQ

Q: How do students perceive their own engagement and student needs in relation to the identified behavioural archetypes (minimalists, task-focused, disenchanted, intensive)?

A: Students' perceptions are likely to vary. Some will recognise their own engagement patterns, while others may misread them because of personal bias or different ideas about what "good engagement" looks like. That is why student voice matters. Direct feedback from students can add context to the behavioural data collected through tools like CLAS, helping educators understand why a pattern appears and what kind of support or teaching change might help.

Q: What role does text analysis play in identifying and understanding the nuances of student engagement in online learning environments?

A: Text analysis tools for education help educators see what behavioural data alone cannot show. By analysing annotations, forum posts, and feedback comments, institutions can identify confusion, frustration, motivation, and reactions to teaching approaches that are not visible in click data alone. Used alongside behavioural measures, student voice gives a fuller picture of how, and why, students engage in online learning environments.

Q: How can educators effectively use the insights from behavioural archetypes and text analysis to tailor their teaching strategies for different groups of students within the same course?

A: Educators can use behavioural archetypes to segment support instead of applying the same intervention to every student. Minimalists may benefit from more collaborative or interactive activities, task-focused students may respond well to stretch work, disenchanted students may need earlier re-engagement and clearer incentives, and intensive students may benefit from greater autonomy or more advanced tasks. Pairing those patterns with text analysis helps ensure that any teaching change reflects students' actual experiences, not just their observed clicks.

References:

[1] Lust, G., J. Elen, and G. Clarebout, Regulation of tool-use within a blended course: Student differences and performance effects. Computers & Education, 2013. 60(1): p. 385-395.
DOI: 10.1016/j.compedu.2012.09.001

[2] Risko, E.F., et al., The Collaborative Lecture Annotation System (CLAS): A New TOOL for Distributed Learning. IEEE Transactions on Learning Technologies, 2013. 6(1): p. 4-13.
DOI: 10.1109/TLT.2012.15

[3] Beretvas, S.N., J.L. Meyers, and W.L. Leite, A Reliability Generalization Study of the Marlowe-Crowne Social Desirability Scale. Educational and Psychological Measurement, 2002. 62(4): p. 570-589.
DOI: doi/10.1177/0013164402062004003

[4] Mirriahi, N., et al., Uncovering student learning profiles with a video annotation tool: reflective learning with and without instructional norms. Educational Technology Research and Development, 2016. 64(6): p. 1083-1106.
DOI: 10.1007/s11423-016-9449-2

Request a walkthrough

Book a free Student Voice Analytics demo

See all-comment coverage, sector benchmarks, and reporting designed for OfS quality and NSS requirements.

All-comment coverage with HE-tuned taxonomy and sentiment.
Versioned outputs with TEF-ready reporting.
Benchmarks and BI-ready exports for boards and Senate.

Prefer email? info@studentvoice.ai

UK-hosted · No public LLM APIs · Same-day turnaround