Modeling the incubation effect among students playing an educational game for physics

Date of Award

2018

Document Type

Dissertation

Degree Name

Doctor of Philosophy in Computer Science

Department

Information Systems & Computer Science

First Advisor

Rodrigo, Ma. Mercedes T., Ph.D.

Abstract

This research investigated the phenomenon called Incubation Effect (IE) in the context of Physics Playground, a computer-based learning environment, and extracted features that would predict the incidence of revisiting an unsolved problem and its positive outcome. A logistic regression model was developed and found coarse-grained level features that predict IE such as time of revisit, students productivity, and problem difficulty. Fine-grained analysis used LSTM, a deep learning technique, which improved the performance of the IE model. A combination of a dimension reduction and visualization technique called T-SNE and X-means clustering were used to interpret the learned features and found that the coarse-grained features are consistent with the fine-grained features but action level features were also discovered such as higher incidence of erase and hover tutorial, lower incidence of pause, and improvement in drawing ramp and springboard during the revisit after incubation. These features were discussed and how they could be translated into game mechanics that could improve students performance in computer-based learning environments.

Comments

The C7.T344 2018

Link to Full Text

Share

COinS