Shaping the Play Experience: Examining Feedback and Choice in Games for Learning

• In Event: Exploring Influences of Game Types and Design Decisions: Issues of Learning, Interaction, and On/Offline Boundaries

Abstract

Purpose

This study examines how two factors of in-game play (type of feedback and choice of non-player character) impact learning outcomes and motivation in a mathematics-content educational video game.

Theoretical Framework

Feedback and choice are two factors that multimedia instruction and video games have in common. Feedback has a rich history in instructional theory (Azevedo & Bernard, 1995; Bangert-Drowns, Kulik, Kulik, & Morgan, 1991; Shute, 2008) and can promote intentional and purposeful processing of information (Moreno & Valdez, 2005). In video games, feedback is half of the “circular model of gameplay,” where the “gamer’s input and the game’s output reciprocally influence each other” (Heaton, 2006).

Choice is another important design factor. In multimedia learning choice can influence both learner-interface and learner-content interactions (Chou, 2003). Yee (2006) found that choices concerning appearance, accessories, and color scheme in games were important sub-components of player motivation.

Given game designers' ability to “design the structures and contexts in which play takes place” (Salen & Zimmerman, 2004, p. 67), what is the most effective combination of feedback and choice for learning and motivation?

Method and Data Sources

To explore this question, an educational game designed to teach the properties of angles was used across two conditions (feedback and choice) with four, randomized experimental groups (informative vs. elaborative feedback; choice vs. no choice of avatar). In the feedback condition, when a player makes a mistake an avatar appears and provides either Informative (I) or Elaborative (E) feedback. Informative feedback tells the player what they did wrong. Elaborative feedback (E) tells the player what they should do to solve the problem. In all conditions, feedback is provided by an avatar, however, in the Choice (C) condition players chose one avatar from a list of six. In the Non-choice (N) condition, avatars are assigned in the same proportions as those chosen in the C condition.

N = 103 6th and 7th graders were randomly assigned across four conditions: EC, EN, IC, and IN, and pretested for their knowledge of angles and ability to solve angles problems. A one-way ANOVA comparing pre-test scores found no significant difference (F=.359, p=.783) between groups. Participants were a given one, 30-minute play session followed by a paper-based post-test. In-game play data and motivation-related self-reports were also collected.

Results

A paired samples T-Test across all groups revealed a significant difference (t=-4.605, p<.001) in positive learning outcome from pre- to post-test. Examining learning outcomes by group, a paired samples T-Test shows a significant pre- to post-test change for the Choice condition regardless of feedback type (CE: t=-4.630, p<.001; CI: t=-3.581, p=.002). The non-choice group showed no significant pre-to-post change. A two-way ANVOA confirmed that choice was the significant factor for the learning outcomes (F=9.283, p=.003). No differences in motivation were found.

Scholarly Significance

This research has implications for educators and game designers. It informs our understanding of the potential influence of choice and feedback types in educational games, and how specific designs that shape game play may influence learning.

Authors

• Charles K. Kinzer, Teachers College, Columbia University

• Daniel L. Hoffman, Teachers College, Columbia University

• Selen Turkay, Harvard University

• Nilgun Gunbas, Kafkas University

• Pantiphar (Dao) Chantes, Teachers College, Columbia University