Importance of Considering User’s Social Skills in Human-agent
Interactions
Is Performing Self-adaptors Appropriate for Virtual Agents?
Tomoko Koda and Hiroshi Higashino
Graduate School of Information Science and Technology, Osaka Institute of Technology, Osaka, Japan
Keywords: Conversational Agents, Gesture, Self-adaptors, Non-verbal Behaviour, Social Skills, Evaluation.
Abstract: Self-adaptors are bodily behaviours that often involve self-touch that is regarded as taboo in public.
However, self-adaptors also occur during casual conversations between friends. We developed a virtual
agent that exhibits self-adaptors during conversation with users. Our continuous evaluation of the
interaction between the agents that exhibit self-adaptors and without indicated that there is a dichotomy on
the impression on the agents between users with high social skills and those with low skills. People with
high social skills feel more friendliness toward an agent that exhibits self-adaptors than those with low
social skills. The result suggests the need to tailor non-verbal behaviour of virtual agents according to user’s
social skills.
1 INTRODUCTION
Intelligent virtual agents (IVAs) that interact face-to-
face with humans are beginning to spread to general
users, and IVA research is being actively pursued.
IVAs require both verbal and nonverbal
communication abilities. Among those non-verbal
communications, Ekman classifies gestures into five
categories: emblems, illustrators, affect displays,
adapters, and regulators (Ekman, 1980). Self-
adaptors are non-signalling gestures that are not
intended to convey a particular meaning (Waxer,
1988). They are exhibited as hand movements where
one part of the body is applied to another part of the
body, such as picking one’s nose, scratching one’s
head and face, moistening the lips, or tapping the
foot. Many self-adaptors are considered taboo in
public, and individuals with low emotional stability
perform more self-adaptors, and the number of self-
adaptors increases with psychological discomfort or
anxiety (Ekman, 1972, Waxer, 1988, Argyle, 1988).
According to Caso et al. self-adaptor gestures were
used more often when telling the truth than when
lying (Caso, 2006).
Because of its non-relevance to conversational
content, there has not been much IVA research done
on self-adaptors, compared with nonverbal
communication with high message content, such as
facial expressions and gazes. Among few research
that has dealt with an IVA with self-adaptors, Neff
et al. reported that an agent performing self-adaptors
(repetitive quick motion with a combination of
scratching its face and head, touching its body, and
rubbing its head, etc.), was perceived as having low
emotional stability. Although showing emotional
unstableness might not be appropriate in some social
interactions, their finding suggests the importance of
self-adaptors in conveying a personality of an agent
(Neff, 2011).
However, self-adaptors are not always the sign
of emotional unstableness or stress. Blacking states
self-adaptors also occur in casual conversations,
where conversant are very relaxed (Blacking, 1977).
Chartrand and Bargh
have shown that mimicry of
particular types of self-adaptors (e.g., foot tapping and
face scratching) can cause the mimicked person to
perceive an interaction as more positive, and may lead
to form rapport between the conversants
(Chartrand,
1999).
We focus on these “relaxed” self-adaptors
performed in a casual conversation in this study. If
those relaxed self-adaptors occur with a conversant
that one feels friendliness, one can be induced to feel
friendliness toward a conversant that displays self-
adaptors. We apply this to the case of agent
conversant, and hypothesize that users can be
induced to feel friendliness toward the agent by
115
Koda T. and Higashino H..
Importance of Considering User’s Social Skills in Human-agent Interactions - Is Performing Self-adaptors Appropriate for Virtual Agents?.
DOI: 10.5220/0004751801150122
In Proceedings of the 6th International Conference on Agents and Artificial Intelligence (ICAART-2014), pages 115-122
ISBN: 978-989-758-016-1
Copyright
c
2014 SCITEPRESS (Science and Technology Publications, Lda.)
adding self-adaptors to the body motions of an
agent.
Because self-adaptors have low message content
and are low in relevancy to the contents of
conversations, they are believed to be actions that
are easily ignored during a conversation. Social
skills, on the other hand, are personal characteristics
that make interpersonal relationships smooth. They
are defined as “skills that are instrumental in
conducting smooth personal relationships” (Hayashi,
1982). People with high social skills are believed to
be able to read nonverbal behaviours in
communication with partners and use them
advantageously in communication. Furthermore,
persons with high social skills are believed to have a
tendency to use a great amount of nonverbal
communication behaviours in order to make
communication with conversation partners richer.
We focused on this characteristic of social skills and
considered that it could have the same effect when
applied to non-verbal behaviour of an agent.
Psychologists have found that people prefer
personalities similar to their own (Izard, 1960, Duck,
1973). Reeves and Nass’ research on the social
responses of people to media indicated users showed
a tendency to prefer computers with personalities
similar to theirs (Reeves, 1996). These findings
suggest that users would also prefer agents with
similar personalities. Because of the characteristics
of social skills, we conjectured that people with high
social skills would consider self-adaptor-performing
agents to have personalities similar to theirs. Thus,
in this study, we made the following hypothesis:
“Compared with people with low social skills,
people with high social skills have a greater sense of
friendliness toward an agent that exhibits self-
adaptors.” We conducted an experiment to verify
this hypothesis.
Many research studies have been done on
interactions between agents and users. However,
most of these studies evaluate transient interactions;
there have been few studies evaluating continued
interactions between agents and users. One
representative study is research on relational agents
by Bickmore. They state that building trust is critical
for continued interactions between users and agents
(Bickmore, 2001, Bickmore 2010, Vardoulakis,
2012). In our study, we took the view that
impressions of self-adaptors in informal
communication are formed through multiple
interactions. Thus, we did not evaluate impressions
after one trial, but instead evaluated multiple
interactions between agents and users by conducting
multiple trials and evaluations. We believe the
results of this study can be applied to the
development of agent applications that require long-
term interactions, i.e., counselling agents, by
evaluating the effects of displaying self-adaptors
with IVAs.
2 EXPERIMENTAL
CONVERSATIONAL AGENT
We conducted a pre-experiment in order to examine
when and what kind of self-adaptors occur during a
casual conversation between friends. We invited
four pairs who are friends for more than three years
(they are university students who study together) to
record their conversation. The recordings were more
than 20 minute long but we evaluated the last 10
minutes when the conversation was active and they
were not nervous about being videotaped. Based on
the results of video analysis of the conversations, we
found the following three types of self-adaptors
occurred most frequently in most pairs: “touching
hair,” “touching cheek,” and “touching nose.” Each
stroke occurred once as a slow movement. The
timing was either at the beginning or at the end of an
utterance. The self-adaptors implemented for the
agents in (Neff, 2011) were repetitive quick hand
scratches, rubbing, tapping, etc., as we see when the
human conversant is nervous. We did not find those
nervous repetitive movements during the casual
conversations in the pre-experiment.
The agent character and animation of the three
types of self-adaptors were created using Poser
(http://poser.smithmicro.com/poser.html). Figure 1
shows the agent carrying out the movements of
“touching hair”, “touching nose”, and “touching
cheek”. We found no literature that explicitly
described the form of the movement (e.g., how the
nose has been touched, in which way, by which part
of the hand etc.), we mimicked the form of the
movements of the participants in the pre-experiment.
Besides these self-adaptors, we created animation of
the agent making the gestures of “tilting its head”
and “placing its hand against its chest.” These
gestures were carried out by the agent at appropriate
times in accordance to the content of the
conversation (“head tilting” at the end of a question,
“hand against chest” when addressing the agent
itself) regardless of experimental conditions in order
not to let self-adaptors stand out during a
conversation with the agent.
The agent’s conversation system was developed
in C++ using Microsoft Visual Studio 2008. The
ICAART2014-InternationalConferenceonAgentsandArtificialIntelligence
116
agent’s voice was synthesized in a woman’s voice
using the Japanese voice synthesis package AITalk
(http://www.ai-j.jp/). The contents of the
conversations were casual (the route to school,
residential area, and favourite food, etc.).
Conversation scenarios, composed of questions from
the agent and response choices, were created
beforehand, and animation of the agent that reflected
the conversational scenario was created. By
connecting animated sequences in accordance of the
content of the user’s responses, the system realized a
pseudo-conversation with the user. The conversation
system had two states. The first state was the agent
speech state, in which an animated sequence of the
agent uttering speech and asking questions to the
user was shown. The other state was the standby for
user selection state, in which the user chose a
response from options displayed on the screen above
the agent. In response to the user’s response input
from a keyboard, animated agent movie that
followed the conversation scenario was played back
in the speech state.
The interaction between the agent and a
participant was restricted as a pseudo conversation.
1) The agent always asks a question to the
participant. 2) Possible answers were displayed on
the screen and the participant selects one answer
from the selection from a keyboard. 3) The agent
makes remarks based on the user’s answer and asks
the next question. The agent performs three self-
adaptors during one interaction in the “with self-
adaptor” condition. The reason we adopted the
pseudo-conversation method was to eliminate the
effect of the accuracy of speech recognition of the
users’ spoken answers, which would otherwise be
used, on the participants’ impression of the agent.
3 EXPERIMENTAL METHOD
The participants in the experiment were 24 Japanese
undergraduate and graduate students (12 male and
12 female), aged 19-24 years. Their social skills
were measured beforehand using KiSS-18 (Kikuchi's
Scale of Social Skills: 18 items) (Kikuchi, 2004).
KiSS-18 is a widely used scale for social skills in
social, clinical, industrial, and educational
psychology as well as nurse-education. Before the
start of the experiment, they were separated into a
high social skills (HSS) group and a low social skills
(LSS) group. Because the average scores on the
social skill scale for Japanese adult males and
females are 61.82 and 60.10, respectively (Kikuchi,
2004), we used these scores as reference and
Figure 1: Agents that exhibit “touching hair” (top),
“touching nose” (middle), and “touching cheek” (bottom)
self-adaptors.
established the HSS group as having a score of 63 or
above (11 participants) and the LSS group as having
a score of 58 or below (13 participants).
The participants in the HSS group and the LSS
group each carried out five rounds of conversation
with either an agent that performed self-adaptors (7
participants in the HSS group and 7 participants in
the LSS group) or an agent that did not perform self-
adaptors (4 participants in the HSS group and 6
participants in the LSS group). Each participant
conducted one conversation with the agent per day,
and the type of agent (with or without self-adaptors)
was kept the same for all trials of the experiment.
The duration of one interaction is about 2 minutes.
The difference between the two types of agents lay
only in whether or not that the agent performed self-
adaptors. The agents’ appearance, voice, timing and
number of gestures (tilting its head and placing its
hand against its chest), and conversation contents
were the same. Also, we prepared five conversation
scenarios so that the contents of conversations would
differ for each experimental trial. The order of the
conversation scenarios for the trials was the same
regardless of the type of agent. For the second trial
and after, expressions such as “I’m glad we can talk
again” were included to express the fact that this was
not the first time the participant was conversing with
the agent. The conditions of the experiment were
ImportanceofConsideringUser'sSocialSkillsinHuman-agentInteractions-IsPerformingSelf-adaptorsAppropriatefor
VirtualAgents?
117
social skills (HSS group, LSS group), type of agent
(with self-adaptors, without self-adaptors), and trial
number (1st, 2nd, 3rd, 4th, 5th).
After each conversation, the participated rated
their impressions on the agent using a semantic
differential method on a scale from 1 to 6. For the
participants’ evaluation of impressions, a total of 23
pairs of adjectives, consisting of the 20 pairs from
the Adjective Check List (ACL) for Interpersonal
Cognition for Japanese (Hayashi, 1982) and three
original pairs (concerning the agent’s “humanness,”
“annoyingness,” and “naturalness”), were used. The
list of adjectives is shown in Table 1 in the result
section. After the end of the 5th trial, a post-
experiment survey was conducted in order to
evaluate the participants’ subjective impression of
overall qualities of the agent, such as the naturalness
of its movements and synthesized voice.
4 RESULTS
4.1 Analysis of Friendliness Factor
Factor analysis (FA) was conducted on the agent’s
impression ratings obtained from the experiment.
The results of FA using the principal factor method
are shown in Table 1. Three factors were extracted,
and we named them as “friendliness” “cautiousness
and “tolerance.” We see that when the participants
perceive the agents interpersonally and rate their
impressions, these three factors have a large effect.
The first factor “friendliness” is composed of
adjectives such as humanlike, friendly, natural,
pleasant, and social. The second factor
“cautiousness” is composed of adjectives such as
cautious, mature, pertinent, and quiet. The third
factor “tolerance” is composed of adjectives such as
calm, and broad-minded.
We totalled the measured scores of adjectives
highly correlated to each FA-extracted factor (high
factor loadings), then we used the total score of each
factor for analyses. We ran three-way ANOVA with
factors “social skills” (HSS, LSS), “self-adaptors”
(with, without), and “number of trials” (1st, 5th)
(repeated measures). The dependent variable was the
total score on perceived friendliness of the agent.
For friendliness, significant second-order
interaction (p<0.01) was seen between the factors
social skills, self-adaptors, and number of trials.
Figure 2 shows the results of multiple comparisons
of friendliness by social skills for treatments of self-
adaptors and number of trials. Significant
Table 1: Results of Factor Analysis (after Promax
rotation).
differences of friendliness ratings (p<0.05) are seen
between the social skills in the condition of “with
self-adaptors in the 1st trial” (HSS group: 64.0 (SE
8.0) > LSS group: 51.9 (SE 11.6)) and of “with self-
adaptors in the 5th trial” (HSS group: 69.1 (SE 6.5)
> LSS group: 52.0 (SE 13.2)). Compared with the
LSS group, the HSS group rated significantly higher
friendliness toward the self-adaptor-performing
agent after both the 1st and the 5th trial.
Next, the results of multiple comparisons of
friendliness by self-adaptors for treatments of the
factors social skills and number of trials are shown
in Figure 3. No significant difference between with
self-adaptor and without could be seen for any of the
treatments of social skills and number of trials.
The results of multiple comparisons of
friendliness by number of trials for treatments of
social skills and self-adaptors are shown in Figure 4.
A significant difference of friendliness ratings
(p<0.05) are seen between the 1st trial and the 5th
trial in the condition of “high social skills and with
self-adaptors” (1st: 64.0 (SE 8.0) < 5th: 69.1 (SE
6.5)), and of “low social skills and without self-
adaptors” (1st: 51.9 (SE 11.6) < 5th: 55.2 (SE 16.3)).
Participants in the HSS group evaluated agents that
performed self-adaptors to be significantly friendlier
after the 5th trail than after the 1st trial. Participants
in the LSS group rated agents that did not perform
self-adaptors to be significantly friendlier after the
5th trial than after the 1st trial.
Three-way ANOVA of social skills, self-
adaptors, and number of trials was con-ducted using
cautiousness’s scale of measurement. None of the
factors showed significance in their main effects and
interactions.
ICAART2014-InternationalConferenceonAgentsandArtificialIntelligence
118
Figure 2: Results of Multiple Comparisons of Friendliness
Scores by Social Skills for Treatments of self-adaptors and
Number of Trials.
Figure 3: Results of Multiple Comparisons of Friendliness
Scores by Self-adaptors for Treatments of the Factors
Social Skills and Number of Trials.
Figure 4: Results of Multiple Comparisons of Friendliness
Scores by Number of Trials for Treatments of Social
Skills and Self-adaptors.
4.2 Analysis of Tolerance Factor
We ran three-way ANOVA with factors “social
skills” (HSS, LSS), “self-adaptors” (with, without),
and “number of trials” (1st and 5th) (repeated
measures). The dependent variables were
participants’ ratings on perceived tolerance of the
agent. Significant first-order interaction (p<0.05)
was seen between social skills and number of trials.
The results of multiple comparisons of tolerance
scores by social skills for treatments of number of
trials are shown in Figure 5. For the 1st trial, a
significant difference between social skills is seen
(p<0.05; HSS group: 8.9 (SE 1.4) > LSS group: 7.2
(SE 1.9)). In the case of the 1st trial, compared with
the LSS group, the HSS group evaluated the agent to
be significantly more tolerant.
The results of multiple comparison of number of
trials for the treatment of social skills factor are
shown in Figure 6. For participants with low social
skills, significant difference is seen between the
number of trials (p<0.05; 1st trial: 7.2 (SE 1.9) < 5th
trial: 8.1(SE 1.6)). Participants in the LSS group
evaluated the agent to be significantly more tolerant
after the 5th trial than after the 1st trial.
Figure 5: Results of Multiple Comparison of Tolerance
Scores by Social Skills for Treatments of Number of
Trials.
Figure 6: Results of Multiple Comparisons of Tolerance
Scores by Number of Trials for the Treatment of Social
Skills.
ImportanceofConsideringUser'sSocialSkillsinHuman-agentInteractions-IsPerformingSelf-adaptorsAppropriatefor
VirtualAgents?
119
4.3 Analysis of Post-Experiment
Survey
A two-way ANOVA of social skills and self-
adaptors was conducted using the post-experiment
survey’s scores (given on 8-point scale from 1: Low
to 8: High). Significant interaction (p<0.05) between
social skills and self-adaptors is seen for the
question, “Were you bothered by the agent’s
actions?” The results of multiple comparisons of
self-adaptors for each level of the social skills factor
showed a significant difference in self-adaptors
(p<0.05; with self-adaptor: 5.43 > without self-
adaptor: 3.83) for the LSS group. The LSS group felt
significantly more bothered by the agent with self-
adaptors than by the agent without self-adaptors.
Concerning the question, “Was it easy to listen to
the agent’s voice?” social skills’ main effect was
significant (p<0.05). Compared with the LSS group
(4.69), the HSS group evaluated the agents as
significantly easier to listen to (6.18).
5 DISCUSSION
5.1 Discussion of Results of Analysis of
Friendliness Factor
From Figure 2, we see that compared with the LSS
group, the HSS group felt a significantly higher
sense of friendliness toward the agent with self-
adaptors, both after the 1st trial and the 5th trial.
From this finding, we can say that regardless of the
number of trials in this experiment, the HSS group
had a significantly higher sense of friendliness
toward the agent that performed self-adaptors than
the LSS group did. Also, because there was not
much difference between the LSS group’s scores for
the condition of self-adaptors and number of trials,
we believe that it was not the case that the LSS
group did not have a sense of friendliness toward the
agent with self-adaptors; rather, the HSS group felt a
stronger sense of friendless toward the agent. This
result supports the hypothesis, “Compared with
people with low social skills, people with high social
skills feel a greater sense of friendliness toward the
agent that exhibits self-adaptors.”
From Figure 3, we see that the results of multiple
comparisons of friendliness scores by self-adaptors
for treatments of the social skills and number of
trials show that there was no significant difference
between “with self-adaptors” and “without self-
adaptors” for any of the treatments of social skills
and number of trials. This result suggests that there
was no significant difference in friendliness due to
only self-adaptors. Although not significant, there
was relatively a large difference in the friendliness
scores between self-adaptor conditions for the HSS
group, but only a small difference was seen for the
LSS group. This finding also suggests that it was not
that the LSS group lacked a sense of friendliness
toward the agent with self-adaptors, but rather, the
HSS group felt a stronger sense of friendliness
toward the agent with self-adaptors.
Figure 4 indicates that the HSS group rated the
agent with self-adaptors as significantly friendlier
after the 5th trial than after the 1st trial. The LSS
group rated the agent without self-adaptors as
significantly friendlier after the 5th trial than after
the 1st trial. From this finding, we can say that when
it comes to continued interactions with an agent, a
sense of friendliness increased for the HSS group as
a result of the agent’s performing self-adaptors. In
contrast, a sense of friendliness increased for the
LSS group as a result of the agent’s without self-
adaptors. This result also supports our hypothesis.
Also, because in continued interactions with the
agent the LSS group experienced an improved sense
of friendliness toward the agent without self-
adaptors, in contrast to the HSS group’s
experiencing an improved sense of friendliness
toward the agent with self-adaptors, we can say that
there is a dichotomy between the evaluation of
friendliness by the HSS group and the LSS group
with regards to the agent performing self-adaptors.
From these results, our hypothesis was supported.
They also suggest the need to develop agents that
meet the level of the users’ social skills when
enabling agents with self-adaptors. Also suggested
by the results is the possibility that a sense of
friendliness toward the agent by users can be
increased in a continual manner by taking into
account the level of the users’ social skills and
whether or not to have the agent perform self-
adaptors during continued interactions.
5.2 Discussion of Results of Analysis of
Tolerance Factor
Figure 5 indicates that compared with the LSS group,
the HSS group rated the agent as significantly more
tolerant after the 1st trial. Figure 6 indicates that the
LSS group rated the agent as significantly more
tolerant after the 5th trial than after the 1st trial.
These results suggest that although the LSS
group did not rate the agent as tolerant compared
with the HSS group, their evaluation of tolerance
ICAART2014-InternationalConferenceonAgentsandArtificialIntelligence
120
increased during continued interactions with the
agent. On the contrary, the HSS group’s evaluation
of the agent’s tolerance did not increase during
continued interactions. However, the HSS group
rated the agent as more tolerant than the LSS group
did from the first interaction.
5.3 Discussion of Post-Experiment
Survey
Regarding the question, “Did you feel bothered by
the agent’s actions?” the LSS group was
significantly more bothered by the agent with self-
adaptors than the agent without. This result is related
to the LSS group’s low evaluation of the friendliness
of the agent with self-adaptors. Being bothered by
the agent’s actions affected the evaluation of
friendliness negatively. For the LSS group, “being
bothered” was probably considered the same as “not
being able to stand it.” Because no difference in the
bothered-ness was seen between self-adaptor
conditions for the HSS group, this suggests that the
LSS group had an oversensitive response to the
agent’s performance of self-adaptors.
Regarding the question, “Was it easy to listen to
the agent’s voice?” the HSS group rated the agents’
voice as significantly easy to listen to compared with
the LSS group. The agent’s voice was the exactly
the same for the HSS group and LSS group. The
results suggest the possibility that in general, the
HSS group had positive view of the agent, whereas
the LSS group had a negative view of the agent.
5.4 Limitations and Future Work
This research is still at a starting phase, thus has
several limitations. Firstly, we need to conduct more
fine grained study on the self-adaptor in human-
human interactions. For example, we need to
conduct close observations on the form and
movements of self-adaptors with larger samples. In
this research we had only four pairs of conversations.
Secondly, on the implementation of self-adaptors
to the agent, our next work should include both
relaxed and stressful self-adaptors. While we used a
female figure of an agent in this experiment,
implementing a male agent and evaluation by both
genders are also needed.
Thirdly, we cannot exclude the effects of the
conversational content when evaluating perceived
impression on the agent. Although self-adaptors are
indirectly related to what is being said, and we
carefully designed the conversation scenarios so as
not to leave any particular impression on the topic, it
is hard to evaluate self-adaptors completely isolated
from the content of the conversation.
Future work should also consider cultural
differences in expressing and perceiving self-
adaptors, since there are culturally-defined
preferences in bodily expressions (Johnson, 2004,
Rehm, 2007, Rehm, 2008, Aylett, 2009) and in
facial expressions (Koda, 2009, Rehm, 2010), and
allowance level of expressing non-verbal behaviour
are culture-dependant.
6 CONCLUSIONS
Our results suggest the importance of changing the
level of displaying self-adaptors of IVAs according
to the users’ social skills. The dichotomy between
the use’s social skills suggests that it is possible to
continually improve users’ sense of friendliness
toward IVAs by combining the presence of self-
adaptors with the user’s level of social skills during
continued interactions with agents. We believe that
it is possible to efficiently elicit users’ sense of
agent’s friendliness for both people with high social
skills and with low social skills by finely adjusting
the appropriate timing of the agent’s performances
of self-adaptors and their frequency depending the
user’s level of social skills. Because users with high
social skills frequently make nonverbal movements
such as gestures and nods, and users with low social
skills have a low frequency of these nonverbal
movements, it is possible to use tools such as Kinect
sensors to detect users’ movements and frequency
during conversations and estimate their level of
social skills. If we can develop agents that use the
estimation results to automatically control the
number and frequency of self-adaptors and draw out
a sense of friendliness from users, we can sustain
high-quality agent interactions. The results of this
research could be applied to the development of
IVAs with which users require long-term interaction,
such as counselling agents.
ACKNOWLEDGEMENTS
This research is supported by a Grant-in-Aid for
Scientific Research (C) 23500266) (2011-2013)
from the Japan Society for the Promotion of Science.
ImportanceofConsideringUser'sSocialSkillsinHuman-agentInteractions-IsPerformingSelf-adaptorsAppropriatefor
VirtualAgents?
121
REFERENCES
Argyle, M., 1988. Bodily communication. Taylor &
Francis.
Aylett, R., Vannini, N., Andre, E., Paiva, A., Enz, S., Hall,
L., 2009. But that was in another country: agents and
intercultural empathy. In Proceedings of The 8th
International Conference on Autonomous Agents and
Multiagent Systems-Volume 1. pp. 329-336.
International Foundation for Autonomous Agents and
Multiagent Systems.
Bickmore, T. and Cassell, J., 2001. Relational Agents: A
Model and Implementation of Building User Trust, In
Proc. of CHI 2001, pp. 396-403. ACM Press.
Bickmore, T, Schulman, D, Yin, L., 2010. Maintaining
Engagement in Long-term Interventions with
Relational Agents. In International Journal of Applied
Artificial Intelligence special issue on Intelligent
Virtual Agents 24(6), pp. 648-666.
Blacking,j.,(ed)., 1977. The Anthropology of the Body
Academic Press.
Caso, L., Maricchiolo, F., Bonaiuto, M., Vrij, A., and
Mann. S. 2006. The Impact of Deception and
Suspicion on Different Hand Movements. Journal of
Nonverbal Behavior, 30(1), pp. 1-19.
Chartrand, T. L., and Bargh, J. A. 1999. The chameleon
effect: The perception–behavior link and social
interaction. Journal of Personality & Social
Psychology, 76, pp. 893–910.
Duck, S.W. 1973. Personality similarity and friendship
choice: Similarity of what, when? Journal of
Personality, 41(4), pp. 543–558.
Ekman, P., Friesen, W.V., 1972. Hand movements. In
Journal of Communication 22, pp. 353-374.
Ekman P., 1980. Three classes of nonverbal behavior. In
Aspects of Nonverbal Communication. Swets and
Zeitlinger.
Hayashi, T., 1982. The Measurement of Individual
Differences in Interpersonal Cognitive Structure. In
Experimental Social Psychology 22, pp. 1-9 (in
Japanese).
Izard, C. 1960. Personality similarity and friendship. The
Journal of Abnormal and Social Psychology 61(1), pp.
47-51.
Johnson, W., Marsella, S., Mote, N., Viljhalmsson, H.,
Narayanan, S., Choi, S., 2004. Tactical language
training system: Supporting the rapid acquisition of
foreign language and cultural skills. In Proc. of
InSTIL/ICALLNLP and Speech Technologies in
Advanced Language Learning Systems.
Kikuchi, A., 2004. Notes on the researches using KiSS-18
Bulletin of the Faculty of Social Welfare. Iwate
prefectural university (in Japanese).
Koda, T., Ishida. T., Rehm, M., and Andre, E., 2009.
Avatar culture: cross-cultural evaluations of avatar
facial expressions. In Journal of AI & Society, Vol.24,
No.3, pp.237-250. Springer London.
Neff, M., Toothman, N., Bowmani, R., Fox Tree, J. E.,
Walker, M., 2011. Don’t Scratch! Self-adaptors
Reflect Emotional Stability. In Vilhjalmsson, H. H. et
al. (Eds.): IVA 2011, LNAI 6895, pp. 398-411.
Springer-Verlag.
Reeves, B. and Nass, C., 1996. Media Equation : How
People Treat Computers, Television and New Media
like Real People and Place. Univ. of Chicago Press
.
Rehm, M., Andre, E., Bee, N., Endrass, B., Wissner, M.,
Nakano, Y., Nishida,T., Huang, H., 2007. The cube-g
approach -coaching culture-specific nonverbal
behavior by virtual agents. In Organizing and learning
through gaming and simulation: proc. of Isaga 2007 p.
313.
Rehm, M., Nakano, Y., Andre, E., Nishida, T., 2008.
Culture-specific first meeting encounters between
virtual agents. In Intelligent virtual agents. pp. 223-
236. Springer-Verlag.
Rehm, M., Nakano, Y., Koda, T., and Winschiers-
Theophilus, H., 2010. Culturally Aware Agent
Communications. In Marielba Zacharias, Jose Valente
de Oliveira (eds): Human-Computer Interaction: The
Agency Perspective. Studies in Computational
Intelligence, Vol. 396, pp. 411-436. Springer.
Vardoulakis, L., Ring, L., Barry, B., Sidner, C., and
Bickmore, T., 2012. Designing Relational Agents as
Long Term Social Companions for Older Adults. In
Proc. of Intelligent Virtual Agents conference.
Springer-Verlag.
Waxer, P., 1988. Nonverbal cues for anxiety: An
examination of emotional leakage. In Journal of
Abnormal Psychology 86(3), pp. 306-314.
ICAART2014-InternationalConferenceonAgentsandArtificialIntelligence
122