1467. Engineering students' strategies and tactics observed through LMS interactions
Invited abstract in session TE-12: Learning Analytics, cluster Analytics and Data Science.
Tuesday, 16:15-17:45Room: FENP201
Authors (first author is the speaker)
1. | Sergio Celis
|
School of Engineering and Sciences, Universidad de Chile | |
2. | Esteban Villalobos
|
University of Chile | |
3. | Juan Ross
|
University of Chile |
Abstract
Supporting first-year engineering students in their learning process is daunting; courses, such as physics and mathematics, demand students to rapidly adapt to the university pace, acquiring a significant amount of new content through continuous problem-solving and examinations. This context usually results in higher failure rates and dropouts than in other fields. To support students, universities launch initiatives based on well-established theoretical frameworks, such as self-regulation and learning approaches. These frameworks rely mainly on filling self-report questionnaires, facing low response rates and not necessarily well-tailored to particular courses. In response to these challenges, learning analytics has used student data to classify students and recommend specific actions. Defining strategies and tactics is one approach to model learning data; learning strategies consist of the collection of actions students frequently use over a certain period, for instance, over a semester. Tactics are students' actions to address a specific task, such as assignments or examinations. Here, we seek to classify strategies and tactics in a traditional first-year engineering course, by analyzing their daily and weekly interaction with their learning management system (LMS).
The sample includes 52 students from a first-year course, "Introduction to Modern Physics." This course follows a traditional setting. Both courses used the same LMS. Thus, during the rest of the semester, students were separated only into two groups to send messages. The first group combined the deep-learning students and the first quantile of students with the highest reading interaction level.On the other hand, the second was composed of the students who belong to the three lower quantiles of reading interactions.
Regarding student tactics, we found four distinctive groups. The most commonly used tactic was reading mildly, which can be interpreted as following the course progress with a low level of engagement. Then, administrative-mild shows a lower level of engagement since the variable is unrelated to course content. On the other hand, reading intensely suggests a high level of engagement, usually expanded in the time previous to examinations. Reading and testing demonstrate an even higher level of engagement because those students used the LMS not only as an informative platform.
Keywords
- Behavioural Operational Research
- Data Mining
Status: accepted
Back to the list of papers