Presenter Information

Abigail GentryFollow

Location

Purdue University, West Lafayette

Keywords

"STEM MOOCs", " Topic Modeling", "Text Mining", "Post-course Survey Analysis"

Abstract

Massive Open Online Courses (MOOCs) are offered for many different subject areas such as STEM, Arts, Medicine, and Business, with differences in each subject area. In this study, we focus on STEM MOOCs and their unique challenges. In order to design better MOOCs to encourage engagement, it is important to understand learners expectations and preferences based on the subject area by analyzing end-of-course evaluation surveys of STEM MOOCs. We analyzed open-ended learner feedback for two post-survey questions for 110 unique STEM MOOCs (some with multiple re-runs) with a total of 8100 responses, offered on the FutureLearn platform. The end-of-course survey questions analyzed in this study were: a) What was your favorite part of the course, and why? and b) How could the course be improved? We used the Latent Dirichlet Allocation (LDA) topic model to determine prominent topics in the responses, and then determined the theme of each topic by qualitatively examining a) top words that defined the topic, and b) learners’ responses most representative of the topic.

Our results indicate that most STEM MOOC learners enjoyed certain aspects such as course assessment, content and teaching methods. Learners wanted improvements in mentor guidance and peer interaction, since discussion groups were often unorganized and did not provide accurate information. STEM MOOC learners also want to see more accurate course description and expected level of difficulty. The results of this study could help instructors, instructional design staff and MOOC platforms improve STEM course design to provide a better learner experience.

Comments

Poster only!

ASEEREG_rev2.pdf (769 kB)
Poster

Share

COinS
 

(POSTER) A Bird's-Eye View of What Learners Like and Dislike about STEM MOOCs Using Topic Modeling

Purdue University, West Lafayette

Massive Open Online Courses (MOOCs) are offered for many different subject areas such as STEM, Arts, Medicine, and Business, with differences in each subject area. In this study, we focus on STEM MOOCs and their unique challenges. In order to design better MOOCs to encourage engagement, it is important to understand learners expectations and preferences based on the subject area by analyzing end-of-course evaluation surveys of STEM MOOCs. We analyzed open-ended learner feedback for two post-survey questions for 110 unique STEM MOOCs (some with multiple re-runs) with a total of 8100 responses, offered on the FutureLearn platform. The end-of-course survey questions analyzed in this study were: a) What was your favorite part of the course, and why? and b) How could the course be improved? We used the Latent Dirichlet Allocation (LDA) topic model to determine prominent topics in the responses, and then determined the theme of each topic by qualitatively examining a) top words that defined the topic, and b) learners’ responses most representative of the topic.

Our results indicate that most STEM MOOC learners enjoyed certain aspects such as course assessment, content and teaching methods. Learners wanted improvements in mentor guidance and peer interaction, since discussion groups were often unorganized and did not provide accurate information. STEM MOOC learners also want to see more accurate course description and expected level of difficulty. The results of this study could help instructors, instructional design staff and MOOC platforms improve STEM course design to provide a better learner experience.