Abstract:

There is still controversy over whether the pedagogical agent (PA) is effective. The PA is a virtual character presented on the interface to promote student learning. This study carefully examined the effect of PA and tested the image principle (when the image of the PA is displayed on the screen, it does not necessarily promote deeper learning) in an interactive e-learning environment, designed to teach animal classification. A total of 91 sixth grade students were recruited and divided into groups with and without PA. The amount of information provided by experimental group and control group was equal. For example, the system required students to observe turtles. The group without PA guided students through text boxes, whereas the group with PA guided students through dialogues. However, what they said was the same, and the hints provided by the system after students observed turtles were also the same. In addition, the appearance of PA was well-designed and excluded the influence of gender (a cartoon character in the shape of a lily). The results indicated experiment group and control group both reached significant difference between pre-test and post-test on comprehension. However, the analysis of covariance showed no significant effect between the two groups (p = .399) on comprehension. This study confirmed the image principle and questioned that in some experiments, PA was effective because they provided more information. If a PA does not provide additional support, it is no different from the group without PA. Therefore, when drawing conclusions about the experiment of PA, researchers should inform readers whether the amount of information in the two groups is equal.

Keywords:

pedagogical agent, image principle, e-learning.