Integrating Generative AI and Progressive Guided Scaffolding Mechanisms in Educational Games to Facilitate Research Design and Statistical Learning

Abstract

The teaching of quantitative research design in the social sciences is crucial, but learners’ learning motivation are limited and there is a lack of more case studies and timely diagnostic guidance. Utilizing case scenarios and giving scaffolding guidance helps to address these limitations. In this study, we designed an educational game that combines case studies and progressive guidance on the Generative AI (GAI) scaffolding. This study develops an innovative scaffolding guidance module for GAI scripts. When a player asks a question, the player will not be told the answer directly. Instead, it gradually guides the player to find the research design problem and think in the direction of appropriate analytical methods. Learners play the role of an anxious graduate student facing a research bottleneck. For a limited time, he can have a discussion with the scaffolding guide to the NPC played by GAI. A total of 18 people participated in the empirical evaluation of this study. The study found that learners had high flow, low anxiety, found the game fun during the game, and had a desire to play again. (All scores are significantly higher than 3, i.e., the median of the scale.) Learners felt that this game enhanced research design thinking more than the conventional curriculum. 72% of the participants felt that the game helped in understanding the concepts of the research design. 78% of the participants felt that the NPC characters would give guiding hints to help learners find the information they need to solve problems online.

Save as a Favourite

Author Information
Yu-Kai Chu, National Taiwan University of Science and Technology, Taiwan
Huei-Tse Hou, National Taiwan University of Science and Technology, Taiwan