Learning Management Systems: What More can we Know?

Abstract

Learning management systems (LMS) are ubiquitous among colleges and universities worldwide, however they are thought of as a transactional warehouse rather than an opportunity to understand student learning outside of the classroom. Each LMS is used differently across campuses, and even across sections of the same courses taught by different instructors. For instance, one instructor might utilize a gradebook feature which allows students to view their assignment grades, while another instructor in the same course might use a different method to distribute grades for assignments. But is there a differential relationship between tool usage and student engagement or performance in the course across these sections? Our research addresses this issue and seeks to understand the nature of how LMS tools are used by students and how the use of those tools may shed insight on student learning or engagement.We ground our work theoretically using the Academic Plan Model to understand how freshman engineering students' use of LMS tools relate to their performance in the class. The Academic Plan Model details potential influences on curriculum design at the course, program, and institutional levels. As the Model suggests, faculty members may (or should) consider learners, instructional resources, and instructional processes when developing their curricular plans. Prior research within and outside engineering, however, has shown that faculty tend not to draw on available data when considering these components, if they even consider them at all. Our study presents an idea for bringing data into those considerations by focusing on the course-level activities of students within an LMS. We empirically describe an opportunity educators have to understand what can be learned from investigating LMS student data. Specifically, our data set consists of student LMS log files (approximately 15 million rows) for one engineering course (36 sections and 876 students) from Fall 2013.Results show clear patterns of student engagement with different LMS tools across final grades within first year engineering courses. Additionally, certain tool usage relates more strongly with course performance. By understanding how and when students use those tools in particular, faculty members may be able to create more data-informed course plans and provide empirically driven feedback to students on their levels of engagement in the class.

Constructing Insights on Learning Analytic Student Activity Data from an Online Undergraduate Construction Management Course

“…To explore how overall student activity correlates to performance, we used Pearson's correlation test on the students' individual total activity and their final grades. Contrary to some literature [5], [10], [13], [16], we did not find a statistically significant correlation at the specified level of p < 0.05 between individual students' total activity and their final grade, as noted by the weak positive trend in Figure 6. One hypothesis for the weak correlation is that multiple factors, such as personal situations or exam time limits, affect student grades.…”

“…For example, 4 PM is 4:00:00 The second phase of the analysis was determining if student activity correlated with performance in the course. Studies have found that LMS tracking variables such as specific tool use [5], [16], content interactions [13], and quantity of online sessions [10] correlate with the students' final grades. Our study used Pearson's correlation test, similar to Brozina and Knight (2015), to assess the correlation of student activity with course performance.…”

“…Studies have found that LMS tracking variables such as specific tool use [5], [16], content interactions [13], and quantity of online sessions [10] correlate with the students' final grades. Our study used Pearson's correlation test, similar to Brozina and Knight (2015), to assess the correlation of student activity with course performance. We also used Pearson's correlation to determine if discussion posts and replies correlated with course grades, as proven by Macfadyen and Dawson (2010) [10] and Kim et al (2016) [11].…”

“…Learning analytics have provided instructors with critical information about their students' behaviors in courses. For example, learning analytics supply insight into the number and time of student interactions [12]- [14] and the frequency of viewing content pages [15] and tools [10], [16]. Student behavior analytics is often compared to student performance and proven to correlate significantly.…”

Using Learning Analytics and Student Perceptions to Explore Student Interactions in an Online Construction Management Course

“…The evolution of online learning has produced unprecedented amounts of student data for understanding the individual learning dynamics of online learners (Brozina and Knight 2015). One type of big data (massive sets of data computationally analyzed for patterns and trends) central to online learning is learning analytics (LA).…”