using learning analytics to predict students’ performance in moodle lms

USING LEARNING ANALYTICS TO PREDICT STUDENTS’ PERFORMANCE IN MOODLE LMS

A case of Mbeya University of Science and Technology

(MUST)

 Dr. Joel S. MtebeUniversity of Dar es Salaam,Tanzaniahttp://works.bepress.com/mtebe/

Motivation, problem area

• Increased adoption of LMS in Sub-Saharan Africa

• Moodle sites (December 2015)– South Africa (390)– Kenya (109)– Tanzania (49)– Uganda (36)

• About 80% of institutions in Tanzania are using various LMS.

Motivation, problem area

• No reliable evidence to suggest that institutions have been utilizing LMS in a bid to improve students’ performance.

.

Research Objectives

• This study developed Learning Analytics tool to examine students’ activities in Moodle LMS through utilizing data from the system log.

• Developed Learning Analytics tool was used to analyze data from the Moodle at Mbeya University of Science and Technology (MUST)

• To determine the causal relationship between students’ activities in Moodle and final students’ results

Research approach, Methodology

• Quantitative research design with data gathered from LMS log using the developed Learning Analytics tool.

• The variables – the number of downloads– number of forum posts – number of peer interactions – the time spent on the system – the number of logins performed and – number of exercises performed by student.

Research approach, Methodology

• Data obtained from the Learning Analytics tool to compare with students scores in the final exam through linear regression analysis.

• Two (2) courses were used

– Applied Biology I (117 students) – Services and Installation II (60 students)

Major Outcomes/Results

• Dashboard interface

Students’ Logins Frequency Searched For 111 Students

8

Applied Biology I Regression results

9

Course 1

• The results show that

– Peer Interaction (beta value = 19.6%), and

– Forum Posts (beta value = 77.1%)

– have shown to have positive significant effect on students’ performance

• Other variables were not significant

10

Services and Installation II regression results

11

Course 2

• The results show that – Forum Posts (beta value=48.5%), and

– Exercise (beta value=51.5%)

– have shown to have positive significant effect on students’ performance in the exam for courses offered via the LMS.

• Other variables were not significant

12

Discussion and Conclusion

• Forum Posts has shown to have significant effect on students’ learning performance in both courses.

– This means, students who were active in the discussion forums during course delivery performed better that those who were less active.

Discussion and Conclusion

• Peer Interaction had impact on students’ performance in Analysis for Applied Biology course. – in order to increase students performance for student taking courses

offered via LMS instructors at MUST must promote for peer interactions amongst students.

• Exercises was found to be the contributing factor in students’ final score in the Service and Installation II course. – Instructors at MUST should consider giving their students more

exercises in order to improve their final scores

SUGGESTION FOR FUTURE

• Logs simply record learners’ behavior in LMS, but they do not explain why some of the factors were significant and some were not significant.

• Courses were not purely online – there are many offline activities such as reading course-

related books or discussions with peers in offline environment were not tracked in the LMS.

Conclusion

• The study designed and developed Learning Analytics tool and use the tool to determine the factors that have impact on student performance.

• The study has shown that Learning Analytic tools are the powerful tools that can be used by our institutions to predict the success and failure of the installed LMS in our premises.

16

Contacts

• Dr. Joel S. Mtebe

Center for Virtual LearningUniversity of Dar es Salaam

jmtebe@gmail.com+255715383366

Web: http://works.bepress.com/mtebe

Slide Title

End