digital chameleons
DESCRIPTION
TRANSCRIPT
Digital ChameleonsAutomatic Assimilation of Nonverbal Gestures in Immersive Virtual Environments
Bailenson, J. N., & Yee, N. (2005)
Eun Joung Cho
Look at these pictures.
And these . . .
Were you smiling?
Did you feel sad?
Motor mimicry
People change their own face similar to their interactant.
e.g. Our face is distorted when the other is in pain.
Chameleon effect
Mimicked behaviours lead to interactants’ favor.
Synchronization
Chameleon effect
• Accents
• Speech patterns
• Syntax
• General mood
• likable
• prosocial behaviour
• rapport persists
A salad, B main dish,
andC dessert,
please
A salad,B main dish,
andC dessert
…thank you, ma’am.
… thank you, ma’am.
(van Baaren, Holland, Steenaert, & van Knippenberg, 2003)
Using mimicry in Virtual world
1. Strategic conversation with others using avatar.
2. Sociable computer system.
Because the information is all digital
1. Frequency (how many times)
2. Thoroughness (how many types of gestures)
3. Intensity (exact mirror or only an approximation)
2005Digital Chameleons
2008Detecting digital chameleons
2010Effects of Facial Similarity on User Responses to Embodied Agents
Experiment video clip
2005
• Participant’s gender
• Agent’s gender
• Agent’s behaviour (mimic or recorded)
IV
• agent’s social presence (realistic)
• agreement (agent’s persuasion)
• impression of the agent (positive)
DV
Digital Chameleons
1.
2. Over 28˚, participants cannot see the avatar.
• Female : mimic -> do not
• Male : recorded -> do
2008
Agent’s behaviourIV
• trustworthiness
• warmth
• information
• agreement
DV
• Mirror-mimic (exactly mirrored)
• Congruent-mimic (reverse-mirrored)
• Axis-switch (mirrored along a different axis)
Detecting digital chameleons
• Participants rated the presenter as
less trustworthy when they detected the mimic.
less friendly (warmth)
• Participants were more likely
to detect the mirror-mimic condition than others.
2010
IV
• Involvement (How much do you feel connected to X?)
• Distance (How much does X leave you with cold feelings?)
• UseIntention (How much do you want to use X again?)
DV
Effects of Facial Similarity on User Responses to Embodied Agents
• Designed similarity
(facial similar vs. dissimilar)
• Designed affordance
(aid vs. obstacle)
• Similar – dissimilar : no difference
• Aid – obstacle : difference
On average participants feel
more using intention with an aiding agent that was facially similar.
higher involvement with an aiding agent.
less distant with an aiding agent.
To sum up,
Mimicked behaviour more social presence / better impression
But,
If the users notice the mimicry, it does not work
And,
The users feel positive when just the agents are helpful.
Thank you
Eun Joung Cho