Effects of a Pedagogical Agent’s Emotional Expressiveness on Learners Perceptions Enilda Romero Ginger Watson, Ph.D. Darden College of Education Old Dominion University
Jul 22, 2015
Effects of a Pedagogical Agent’s Emotional Expressiveness on Learners Perceptions
Enilda Romero
Ginger Watson, Ph.D.
Darden College of Education
Old Dominion University
INTRODUCTION
Computer–Mediated
Instruction
Content Representation
Animated Pedagogical
Agents
Emotion
•Web–based Instruction•Instructional Simulations•Serious Games
•Text•Audio•Images•Animations
•Interface Agent Characteristics•Character Building Qualities
•Low Intensity•High Intensity
LITERATURE REVIEW
Quantitative Inquiries
Positive Results for both Learning and Learner’s Perception (Lester et al., 1997)
Failed to Obtain Significant Results for Either Variable (Craig et al., 2002)
Qualitative Inquiries
Phenomenological Exploration of Conversational Agent (Veletsianos & Miller, 2008)
PURPOSE STATEMENT
To describe the lived experiences of the learners as they interact with an emotionally expressive animated pedagogical agent in a computer–mediated instructional environment.
RESEARCH QUESTIONS
a. How do learners perceive low intensity emotionally expressive agents?
b. How do learners perceive high intensity emotionally expressive agents?
OPERATIONALIZED VARIABLES
Low Intensity Emotion- Soft emotional tone in the voice
- Eyebrow, mouth, and cheek movements were subtle and gentle
- Head movement was limited and delicate.
High Intensity Emotion- Strong emotional tone in the voice, with
very high pitch and intonation
- Eyebrow, mouth, and cheek movements were exaggerated
- Significant amount of head movement.
PARTICIPANTS (8)
50%
25%
25%
Experience with Computers
Competent
Proficient
Expert
37%
63%
Gender
Male
Female
75%
25%
Age
18 – 25
25 – 35
RESULTS
• Perceived Importance
– “Emotions are a huge part of people and that
if you take that out of a teacher it would be
harder for the students to relate and they will
get bored.”
– One participant mentioned that they could not see the difference between using an emotionally expressive agent and just videotaping a real person.
RESULTS
• Perceived Enjoyment
– Regardless of agent expressiveness, low or high intensity, participants were both critical and complimentary regarding their perceived enjoyment of the animated agent.
RESULTS
• Perceived Humanness of Agent
– All participants perceived the
agent as “professional looking”, “sharp
dresser,” “very modern” and someone that “looked like a professor.”
– “I noticed that sometimes the voice and the
movement of the mouth did not quite line up.
For a second it discredited the avatar.”
RESULTS
• Perceived Barriers of Implementation
– “I had to focus on what it was saying because
his face distracted me from his voice.”
– “The funny faces he made with the smile and
moving eyebrows distracted me into watching
them instead of listening to what he had to
say.”
RESULTS
• Suggestions for Improvement
– “I could easily spot the same pattern of
movements and expression from the avatar. A
richer library of expressions would make a
great difference.”
– “Sometimes I imagined a person reading a
script.”
• Low Intensity Emotion– Perceived the agent as an important element
because it used emotion to provide subtle cues that highlighted important content in the lecture.
– Perceived this agent as friendly and professional looking.
– Did not perceived the agent as human because it lacked proper synchronization between the topic of the dialogue and the expression of emotion.
– Felt it distracted.
• High Intensity Emotion– Perceived the agent as an important because
the emotion contributed to their character as instructors of a lecture.
– Enjoyed the emotion and movement of the agent.
– Felt the agent was not visually pleasant.
– Perceived the agent as distracting.
Design Recommendations
• To utilize animated agents in scenarios that involve performance or decision making
• To design the agent in a manner that it is visually pleasant to the learner
• To develop the agent in a manner that it presents a consistent synchronization
• Number of participants
• The content limited the facial and verbal expressions of the agent to mostly negative emotions
Limitations
ACKNOLEDGEMENT
Dr. Danica HaysAssociate Professor, Old Dominion University
Nicole CoxBiology Student, Old Dominion University
Questions or Comments?
Contact Information:
Enilda Romero
Dr. Ginger Watson