Serious Games: A New Domain for Pedagogical Agent Research?

• In Event: Symposium: Future Directions for Pedagogical Agent Research

Abstract

Objectives and theoretical framework

Serious Games are a promising instructional format combining playing and learning (Ritterfeld, Cody & Vorderer, 2009). Game characters such as avatars and non-player-characters play an important role in game design. However, recommendations on how to design these characters are scarce. Research on the design of pedagogical agents as characters in multimedia learning environments has a longer tradition summarized in a review and a meta-analysis (Heidig & Clarebout, 2011; Schröder et al., 2013). Two significant design features of pedagogical agents are their gender and their perceived competence (Baylor & Kim, 2005). In a technical domain, male pedagogical agents that take an expert role should result in more interest and enjoyment than female pedagogical agents or agents designed as peers (Baylor & Plant, 2005). As non-player-characters in serious games provide feedback and support for learning, they can be viewed as pedagogical agents – transferring the design guidelines for pedagogical agents to serious games.

Methods

Design and materials

A 2x2-design was applied varying the pedagogical agent’s gender and instructional role (peer vs. expert, figure 1) in a serious game called “SERENA SUPERGREEN” (figure 2). This serious game aims at facilitating the interest of girls in the age of 13-16 years in technical domains.

Data sources

The participants were 69 middle school students (9th-grade, 35 girls, 34 boys). The mean age was 14.51 years (SD=.61). The participants were randomly assigned to the treatment groups (figure 1). They played SERENA for 20 minutes and answered questionnaires on enjoyment and technical interest before and after playing. Then, they answered the Agent Persona Instrument (API, Ryu & Baylor, 2005) and a participant questionnaire.

Results

ANOVAs for repeated measures on the enjoyment scores (before and after playing) and the technical interest scores showed no changes over time (enjoyment: F(1, 63)=1.10, p=.17; interest: F(1, 62)=2.03, p=.16). Neither the pedagogical agents gender (enjoyment: F(1, 63)=1.14, p=.29; interest: F(1, 62)=.01, p=.95) nor its instructional role (enjoyment: F(1, 63)=1.05, p=.31; interest: F(1, 62)=.96, p=.33) or the interaction of both (enjoyment: F(1, 63)=.78, p=.38; interest: F(1, 62)=.91, p=.35) affected the enjoyment or the technical interest of the participants.

The scores on the Agent Persona Instrument did not differ among the four pedagogical agent groups. Nevertheless, explorative analyses revealed that girls perceived pedagogical agents designed as peers as more helpful, credible, motivating and supportive compared to pedagogical agents designed as experts. Boys however, perceived pedagogical agents designed as experts as more credible and more effectively in presenting information than peer-pedagogical agents (see table 2-3).

Discussion

In sum, the design of the pedagogical agents did not affect enjoyment or technical interest of the players. This might be due to the relatively short playing time. However, the pedagogical agents were perceived differently according to their gender and role depending on the gender of the player. Thus, the design of non-player characters in serious games should consider the characteristics of the character, the player, the context and the instructional role. Just as proposed in the „Pedagogical Agents-Levels of Design (PALD)“-model (Heidig & Clarebout, 2011).

Authors

• Steffi Heidig, University of Applied Sciences Zittau/Goerlitz

• Felix Kapp, TU Dresden

• Susanne Narciss, Technische Universitaet Dresden