| Article ID | Journal | Published Year | Pages | File Type |
|---|---|---|---|---|
| 350296 | Computers in Human Behavior | 2015 | 7 Pages |
•Jung and attractive female agent positively impacts interest in learning.•Old and unattractive male agent has no impact on the motivation.•Agent appearance did not impact performance on a final exam.•Initial evidence for importance of utilizing personalized agents in learning tutoring.
Background: Anthropomorphic (human-like) agents are commonly applied in e-learning as tutors for students, but we still do not know which particular features of an agent contribute to an enhancement of learning and motivation. Method: In this study a sample of university students was offered online learning cases as exam preparation aid and students were randomly assigned to one of three groups. The first group was tutored by a virtual male agent, the second by a virtual female agent. The cases of the third group were not assigned with an agent. We investigated if and under which circumstances a virtual tutor influences the students’ exam performance, interest in the course material and their enjoyment during learning. Results: While the female agent (rated as young and attractive) had a positive impact on the interest in the course material, the male agent (rated as old and less attractive) had no positive influence at all on motivation or performance of the students on the exam. Conclusions: Appearance features of tutoring agents in e-learning independently influence the student’s motivation and performance. Thus, interest can be manipulated by different aspects of the agent dependent on the primary goal while the agents have no effect on performance. We therefore suggest implementing personalized tutors in e-learning tutorials; however, we recommend more research examining the underlying mechanisms behind the positive effects on interest and look for ways to transfer them to performance.
