Clustering the Relationship Between Scaffolding and Students' Characteristics Through Data Mining

• In Event: Poster Session 4
  In Poster Session: Research on Teaching With Technology

Abstract

Data mining is an effective methodology to find out the patterns of data and to establish the relationship between the included moderators. This study utilized an EM algorithm that is one of data mining techniques for clustering to identify the effectiveness of computer-based scaffolding with a combination of several sub-categories from scaffolding and students characteristics. The results of the clustering analysis showed how the characteristics of scaffolding can be grouped with each other, and which groups are the most effective in Problem-centered Instructional Models for STEM education. The best combination of scaffolding characteristics that have the highest effect size was conceptual scaffolding with fading and adding functions in the context of problem-solving for graduate students in Science Education.

Authors

• Nam Ju Kim, University of Miami

• Brian R. Belland, Utah State University

• Yong Seog Kim, Utah State University