Aleven - Causal Argumentation Game

Policy World: Combining an intelligent tutor with an educational game

Summary Table

Study 1

PIs	Matt Easterday, Vincent Aleven
Other Contributers
Study Start Date
Study End Date
LearnLab Site
LearnLab Course
Number of Students
Total Participant Hours
DataShop

Abstract

In the educational game Policy World, students search for and analyze scientific evidence that they then use to debate a computer opponent on topics like school-choice. Unfortunately, commonly-used game mechanics that impose stiff penalties to build anticipation and motivation (e.g., restarting a level after “dying”) conflict with traditional cognitive tutoring mechanics such as immediate, direct assistance. In this study I compare the effects of immediate vs. delayed cognitive assistance on learning and motivation in an educational game. The laboratory study with compares 80 college undergraduates randomly assigned to two groups. Students in the traditional-tutoring group receive immediate, step-level cognitive assistance after making an error. Students in the cognitive-game group receive only situational game feedback. However, after failing a level, these students are allowed to replay the level and are then given immediate, step-level cognitive assistance. Outcome measures include both learning (e.g., whether the student picks a policy position based on evidence and can cite evidence to support that positions) and motivation (e.g., the student’s time on task and self-reported attitude toward the game). Log data will be analyzed for major sources of errors in: searching for evidence, comprehending and evaluating causal claims, creating diagrammatic representations of causal claims, synthesizing multiple claims, and choosing recommendations. The study will determine whether we can increase motivation in traditional tutors by adding game mechanics like fantasy and opposition or if we must fundamentally alter how we provide assistance in the context of an educational game.