Content uploaded by W. Lewis Johnson
Author content
All content in this area was uploaded by W. Lewis Johnson
Content may be subject to copyright.
Interactive Pedagogical Drama for
Health Interventions
Stacy C. Marsella, W. Lewis Johnson & Catherine M. LaBore
Center for Advanced Research in Technology for Education (CARTE)
USC / Information Sciences Institute
4676 Admiralty Way, Marina del Rey, CA 90292 USA
marsella@isi.edu
Abstract. The goal of Interactive Pedagogical Drama (IPD) is to exploit the edifying
power of story while promoting active learning. An IPD immerses the learner in an
engaging, evocative story where she interacts with realistic characters. The learner
makes decisions or takes actions on behalf of a character in the story, and sees the
consequences of her decisions. The story’s characters are realized by autonomous
agents. We discuss IPD in the context of Carmen’s Bright IDEAS (CBI), a
multimedia title designed to teach problem solving skills to mother’s of pediatric
cancer patients. CBI was an exploratory arm of a clinical trial and here we discuss
key creative and technical aspects of the design and results from that arm.
Topics: Pedagogical Drama; Virtual Learning Environments
SubTopics: Agent-based virtual characters; Health Interventions
The use of drama as a pedagogical tool is a constant across cultures and throughout history.
In Poetics, Aristotle argued over two millennia ago that learning and drama are interwoven:
that drama is an imitation of life and humans learn through enjoyment of that imitation.
More recently, research in psychology has argued that narrative is central to how we
understand the world and communicate that understanding [1]. And of course, the
engrossing, motivational nature of story is unmistakable; the world now consumes stories in
various media with a “ravenous hunger” [10].
Yet stories can also have a drawback from a learning perspective: they typically place
the learner in the role of passive audience instead of active learner. The goal of Interactive
Pedagogical Drama (IPD) is to exploit the edifying power of story while promoting active
learning. An IPD immerses the learner in an engaging, evocative story where she interacts
openly with realistic characters. The learner makes decisions or takes actions on behalf of a
character in the story, and sees the consequences of her decisions. The learner identifies
with and assumes responsibility for the characters in the story, while the control afforded to
the learner enhances intrinsic motivation [7]. Since the IPD framework allows for stories
with multiple interacting characters, learning can be embedded in a social context [17]. We
take a very wide view of the potential applications of interactive story and IPD in particular.
We envision interactive story as a means to teach social skills, to teach math and science, to
further individual development, to provide health interventions, etc.
In creating an IPD, the demands of creating a good story, achieving pedagogical goals
and allowing user control, while maintaining high artistic standards, must all be balanced.
To ensure a good story, dramatic tension, pacing and the integrity of story and character
must be maintained. Pedagogical goals require the design of a pedagogically-appropriate
“gaming” space with appropriate consequences for learner choices, scaffolding to help the
learner when necessary and a style of play appropriate to the learner’s skill and age. To
provide for learner control, an interaction framework must be developed to allow the
learner’s interactions to impact story and the pedagogical goals. These various demands can
be in conflict, for example, pedagogically appropriate consequences can conflict with
dramatic tension and learner control can impact pacing and story integrity.
Appears in AIED 2003, 11th International Conference on Artificial Intelligence in Education, Australia, 2003.
We have developed an agent-based approach to interactive pedagogical drama. Our first
IPD was Carmen’s Bright IDEAS (CBI), an interactive, animated health intervention
designed to improve the social problem-solving skills of mothers of pediatric cancer
patients. Parents of children with chronic diseases are often poorly equipped to handle the
multiple demands required by their ill child as well as the needs of their healthy children,
spouse and work. Critical decisions must be made that affect family and work. To help train
parents in the problem-solving skills required to address such challenges, CBI teaches a
method for social problem-solving called Bright IDEAS [16]. Each letter of IDEAS refers
to a separate step in the problem solving method: Identify a solvable problem, Develop
possible solutions, Evaluate options, Act on plan and See if it worked. Prior to CBI, the
Bright IDEAS method was taught in a series of one-on-one sessions with trained
counselors, using worksheets that helped a mother detail her problems in terms of IDEAS
steps. The purpose of Carmen's Bright IDEAS is to teach mothers how to apply the Bright
IDEAS method in concrete situations. Mothers learn more on their own and at times of their
own choosing, and rely less on face-to-face counseling sessions. An early version of the
CBI system was first described in [12], and was subsequently further developed and tested.
Although the design of agent-based interactive drama systems is an increasingly active
area of agent research [e.g., 4,9,15,18], it is also an open area of research with few fielded
applications. Further, their use in pedagogical applications is far less common and raises
additional challenges, especially in such a challenging application area as psychosocial
intervention. CBI has been tested as an exploratory arm of a clinical trial of the Bright
IDEAS method at seven cancer centres across the U.S. This paper describes design changes
made in the course of preparing CBI for clinical trial, as well as results from its evaluation.
1. The Design of Carmen’s Bright IDEAS
In our basic design for interactive pedagogical drama, there are five main components: a
cast of autonomous character agents, the 2D or 3D puppets which are the physical
manifestations of those agents, a director agent, a cinematographer agent, and finally the
learner/user who impacts the behavior of the characters. Animated agents in the drama
choose their actions autonomously but also follow directions from the learner and/or a
director agent. Director and cinematographer agents manage the interactive drama's
onscreen action and its presentation, respectively, so as to maintain story structure, achieve
pedagogical goals, and present the dynamic story so as to achieve best dramatic effect. The
design of all these agents requires both general capabilities as well as knowledge specific to
the interactive drama that is being created.
The story for Carmen’ Bright IDEAS was developed
via collaboration between the authors, clinical colleagues
at two of the seven sites where the Carmen was to be
tested and a professional scriptwriter (Jonathan Kaplan).
The story and design features of the interactive
experience went through a sequence of formative
evaluations with mothers at one of these sites. The story
is organized into three acts. The first act reveals the back
story; various problems Carmen is facing, including her
son’s cancer, her daughter Diana’s temper tantrums, work problems, etc. The second, main,
act takes place in an office, where Carmen discusses her problems with a clinical counselor,
Gina, who suggests she pick a solvable problem and use Bright IDEAS to help her find
solutions. See Figure 1. With Gina's help, Carmen goes through the initial steps of Bright
IDEAS, applying the steps to one of her problems and then completes the remaining steps
on her own. The final act reveals the outcomes of Carmen's application of Bright IDEAS.
Figure
1
: Gina & Carmen (right)
The design goal for the overall CBI experience was to mirror, but be distinct from, the
kinds of problems the learner faces. The learner was supposed to be transported into the
drama away from the stress of their own problems. Whereas the emotions evoked would be
strong, the transportation would make them more manageable. To realize this goal, we
chose a presentational, third person experience where learners influenced Carmen but were
not immersed in the drama. Characters would not directly talk to the learner about their own
problems, which might be insulting or too distressing. Presentational drama would allow for
psychological distancing and the freedom to explore coping strategies through Carmen.
Further, it reduced interactivity demands – stressed mothers did not need to control their
character moment-to-moment since it had its own motivations. The interactivity would
nevertheless allow active learning, promote identification with Carmen and ideally lead to
belief in the efficacy of the skills being taught. To the extent the learner explored
unproductive coping strategies with Carmen, we wanted those strategies projected back into
the drama where they would be dealt with.
1.1. Interaction Model
The interaction model we designed for
CBI is what we call a rubber-band model. See
Figure 2. Both Gina and the learner exert
influence over Carmen. It is Gina's job to
keep the social problem solving on track so
that the story proceeds to a successful
outcome by effectively responding to
Carmen's cognitive and emotional state, at
times motivating her through dialog to work through the steps of IDEAS on some problem
or alternatively calming or reassuring her. The human mother interacts with the drama by
making choices for Carmen such as what problem to work on and how she should cope with
the stresses she is facing. The learner can choose alternative internal thoughts for Carmen,
such as “I hope this helps with Diana.” These were presented as thought balloons (see
Figure 3). Both Gina’s dialog moves and the learner’s choices influence the cognitive and
emotional state of the agent playing Carmen, which in turn impacts her behavior and dialog,
perhaps in conflicting ways. In this interaction model, the Gina agent is both on-screen
character and the drama’s director. The Carmen agent has her own cognitive and emotional
dynamics (hence rubber-band metaphor) that ensure her behavior is believable, regardless
of how Gina and the learner influence her.
The combination of Gina's motivation of
Carmen through dialog and the learner’s
emotional impact on Carmen creates tension, a
tug-of-war between Gina’s attempts to motivate
Carmen and the initial, possibly less positive,
attitudes of the Carmen/learner pair. As the
learner plays a role in determining Carmen's
attitudes, she assumes a relationship in this tug-of-
war, including, ideally, an empathy for Carmen
and her difficulties, a responsibility for the
onscreen action and perhaps empathy for Gina. If
Gina gets Carmen to actively engage in applying the IDEAS technique with a positive
attitude, then she potentially wins over the learner, giving her a positive attitude.
Regardless, the learner gets a vivid demonstration of how to apply the technique.
Figure 2: Rubber Band Model
Figure 3: Thought Balloons
2. Agent Models
The interactivity in CBI largely centers on the main act in Gina’s office, where there is a
discussion of serious and disturbing problems confronting Carmen. In building agents for
these characters, a central challenge was to faithfully model the emotions and coping
behaviors associated with these issues and reflect them in the agent’s behaviors. In addition,
it was necessary to model the inherent structure of this act, the discourse strategies Gina and
Carmen might use in their discussion of how to apply Bright IDEAS to those problems.
Each onscreen character is realized by an agent architecture that has goals, (discourse)
strategies to realize those goals and dialog tactics that flexibly interpret those strategies
allowing the agent to react in goal-directed fashion to the unexpected. The architecture is
realized by a collection of modules: problem solving and dialog model how to use dialog to
achieve goals, emotional appraisal models how the agent emotionally evaluates events and
behavior generation arbitrates the agent’s nonverbal behavior. The full details of this
architecture are beyond the scope of this paper (see [11,12] and authors for more details).
The agent models and learner interaction are grounded in the research that influenced
the Bright IDEAS pedagogy: cognitive appraisal models of human emotion. In appraisal
models, emotions and coping responses stem from appraisal of events in terms of their
significance to the individual. Appraisal leads to a characterization of a stressful event along
several dimensions, including ego-involvement, expectancy, accountability and coping
potential. Ego involvement concerns how an event impacts the collection of commitments,
goals, concerns or values that comprise a person’s ego-identity, such as self- and social-
esteem. Expectancy concerns whether the situation will likely get better. Accountability is
an evaluation of who is at fault. Coping potential is an evaluation of how a situation can be
coped with, leading to a wide range of possible coping behaviors such as taking action to
change the world, avoidance of stressful situations, trying to focus on the positive, or denial.
The key to the pedagogy is to teach how to choose and carry out the right strategy for a
given situation and to maintain a realistic belief in one’s own efficacy.
For example, at one point Gina asks Carmen why her daughter is having tantrums.
Carmen feels anxious about being judged a bad mother (ego-involvement) and the learner
may choose a thought that reinforces the anxiety (e.g., by reducing expectancy). Carmen
copes (problem solving) by dismissing the significance of the tantrums (dialog model):
“She is just being babyish, she wants attention.'” Based on Carmen's dialog and emotional
state, behavior generation selects relevant behaviors (e.g., fidgeting with her hands). Her
dialog also feeds back to emotional appraisal. She may now feel guilty for “de-humanizing'”
her child and display that feeling (physical focus) and then go on to openly blame herself.
Carmen can go through this sequence of interactions solely based on the flux in her
emotional reaction to her own behavior. Gina, meanwhile, will emotionally appraise
Carmen's seeming callousness and briefly reveal shock (e.g., by raised eyebrows), but that
behavior is quickly overridden if her dialog model decides to project sympathy.
2.1. Dialog Module
The dialog module uses high-level strategies to drive the discourse through a scene. For
example, the main act in CBI is Gina’s goal of getting Carmen to apply IDEAS to one of
her problems. Gina has an abstract strategy to do this (reassure Carmen, suggest they use
IDEAS, ask her to choose a problem and guide her through the steps of IDEAS). Gina also
has alternative sub-strategies that can be used to solve these steps, e.g., she can help Carmen
develop solutions by prompting her to enumerate them, or she can help Carmen identify
features of a problem by asking a sequence of questions.
These strategies both recapitulate the original script as well as variablize it at multiple
levels, both in terms of the alternative strategies and sub-strategies that can be selected as
well as by how the strategies are interpreted. The agent realizes dialog turn-by-turn by
flexibly executing the high-level strategies using a state machine built out of 12 types of
dialog acts. The machine allows the agent to adapt to twists and turns in the dialog caused
by the autonomy of the agents and the learner’s interactions. For example, if Carmen
doesn’t answer a question properly, Gina may try an alternative path in the state machine to
get the answer, such as reassuring Carmen and re-asking the question. In addition to acts
like Ask and Answer, there are ones that are tightly coupled to emotional state and
pedagogy: Reassure, Praise, Agree/Sympathize, Resign (Give-up) and Summarize.
In order to maximize expressive effect, recorded dialog of voice actors was used instead
of speech synthesis. A significant amount of variability in the generated dialog is supported
by breaking the recordings into meaningful individual phrases and fragments and by
recording multiple variations (in content and emotional expression). Additionally, different
expressive behaviors could be associated with a phrase (discussed below). There are 480
dialog fragments in the clinical trial version of CBI. The agents compose their dialog on the
fly, using annotations attached to the fragments to understand each other and decide how to
respond. The agents experience each fragment’s annotation in order, so their internal state
and appearance can be in flux over the dialog segment. Annotations include:
• Dialog content: Dialog Act, Speaker and Addressee
• Emotional content: Coping Act (e.g. denial)
• Propositional content: Main referent (e.g. Diana) & Topic (e.g., temper tantrums)
2.2. Emotional Appraisal and Behavior Generation
The emotion model in CBI has several unique features required to realize psychosocial
drama that set it apart from most models of emotions used in agent systems. Emotions can
stem from appraisals of the other agent’s dialog; but an agent can also react emotionally to
its own dialog. The expression of emotion also stems from two sources, appraisals [3] as
well as the intention to communicate emotions [8,13] derived from the dialog act. Thus an
agent can communicate emotions that they do not feel. The expression of emotion and
nonverbal behavior is sequenced by the behavior generation module over the course of an
agent’s dialog fragments, beginning with an emotional expression based on the appraisal of
previous dialog, followed by other nonverbal behavior (e.g., gestures) selected based on
current dialog fragment and emotional state and ending with the appraisal of that fragment.
As the above tantrum example revealed, a key aspect of the emotion modeling is the
agent's ego identity. Carmen’s ego ideal of being a good mother, and desire to be perceived
as one (social esteem) leads to her anxiety about discussing Diana's tantrums with Gina.
Ego identity is modeled as a collection of role ideals (Carmen wants to be a good mother),
concerns (good mothers want their children to be happy and healthy) and responsibilities
(good mothers are responsible for their child’s behavior). The system also models relations
(Gina is a parental-surrogate for Carmen). Appraisal rules derive emotions from these
various representations (e.g., an event is negative if it violates a concern; if talking about
negative event which agent feels responsibility for then increase guilt).
3. Clinical Trials
The overall clinical study of Bright IDEAS, of which CBI was a smaller exploratory arm,
was designed to validate the Bright IDEAS technique for teaching social problem solving
skills. This larger study involved each learner interacting with a clinical research assistant
(RA) once a week over the course of an 8-week intervention. This intervention involved
face-to-face discussion of the mother’s problems and guided instruction on how to use the
Bright IDEAS technique in general and its worksheets in particular. Evaluations were done
using a standard social problem solving skills instrument. In the CBI arm, the system was to
replace the standard interaction with the RA on 3 sessions, with the other 5 being the
standard RA intervention. Each of these 3 sessions was supposed to involve a separate
problem-solving vignette, for example, Carmen’s difficulties with Diana’s temper tantrums,
understanding the medical treatment, and work-related issues. RAs at the various sites
instructed subjects on how to use CBI, and were available to answer the subjects’ questions.
System development had to accommodate the clinical trial’s coordinated recruitment of
mothers across the USA. Due to the tight constraints, there was insufficient content created
in time to detail Carmen working through solutions to three separate problems. Only one
full vignette was completed, the Diana tantrum problem. To compensate, two other
modified vignettes were done, a task that was facilitated by the flexibility of the agent
infrastructure. A reduced content vignette was created concerning Carmen’s problems at
work in which the system would at key junctures ask the learner how she would address
Carmen’s problems (“Can you help Carmen DEFINE solutions?”) and write down these
answers on Bright IDEAS worksheets. This reduced the need for more voice recording and
animation, while providing more concrete experience in using Bright IDEAS. In addition, a
third vignette was created, a variant of the Diana problem but with text windows popping
up at key pedagogical junctures with commentary on the on-screen events in terms of Bright
IDEAS concepts. For example, at a moment in which Carmen was despondent, a window
would pop up, stating “Carmen CAN solve some of her problems. She needs to be more
optimistic.” Both of these techniques tended to break the “fourth wall” of the drama and
thus opened up additional concerns and evaluation issues.
Being part of a larger study also led to other changes. Most significantly, recruitment of
mothers for other arms of the larger study was behind schedule. To compensate, sites had to
shift recruitment away from the Carmen arm and some sites did not run any subjects at all.
For example the Children’s Hospital LA (CHLA) site played a key role in the design of the
system, but was then redirected to another arm of the study. This led to distributional
differences between the other arms of the study and the CBI study, which made direct
comparison difficult. The populations tested at each site differ along several dimensions,
such as what type of cancer, the family’s social-economic status, computer expertise,
marital status, ethnicity and whether the mother is a native English speaker.
Twenty-six subjects started the Carmen arm, sixteen of which used CBI and completed
the subsequent training sessions and post-test evaluations. The main instrument for the
evaluation of CBI was a questionnaire designed by the authors that used 5-point scale items
(1-5; e.g., from strong disagreement to strong agreement with “somewhat” in the middle).
Key issues addressed included :
• Helpfulness in remembering the steps of Bright IDEAS (BI)
• Helpfulness in understanding how to apply BI to the learner’s own problems
• Relative difficulty of Carmen’s problems and her ability to cope
• Clarity, interestingness, believability of the story
• Assessment of the interactivity via thought balloons
• Assessment of interface features that were outside the drama:
o First vignette’s showing Carmen’s worksheets being filled out by Carmen
o Second vignette’s stopping the story so learner could fill out worksheets
o Third vignette ’s commentary on the unfolding drama
Table 1 presents results for key questions. The mothers felt that Carmen’s Bright IDEAS
helped them remember the steps of Bright IDEAS and understand how to apply them to
their own problems. Anecdotally, one mother commented that CBI made Bright IDEAS
more concrete. They also assessed the story as believable, interesting and convincing. We
believe this to be a remarkable achievement; it should not be forgotten that the learner is
interacting with 2-D cartoon-like animated characters about deeply distressing problems and
the possibility of a negative reaction was a priori a deep concern of the developers. The
mothers also tended to find the amount of interaction about right. To our surprise, they also
found the elements that break the fourth wall helpful. Overall, the results were positive.
Although there are a small number of subjects and the assessment is subjective, many of the
results are significantly above mean. We believe this argues for the potential effectiveness
of interactive drama.
Category/Question Range Mean 95% Confidence Level
Interactive Story
Story Believable 2-5 4.5 .47
Conflict w/Diana Convincing 3-5 4.4 .43
3rd Vignette’s commentary useful 2-5 4.0 .51
Interaction (1-too little, 3-about right, 5-too much) 2-4 3.0 .19
Pedagogical effectiveness
CBI helped understand how to apply BI 3-5 4.3 .45
CBI helped recall of BI steps 2-5 4.1 .49
Self-Motivated (1-not, 3-somewhat, 5-very) 2-5 3.4 .58
Table 1: Results
Additionally, nine mothers said they would prefer to interact with Carmen as opposed to
seeing a video (5) or reading a story (1). In an open-ended question on what they liked,
mothers often stated or suggested they identified with Carmen, liked her “real” problems
and said CBI helped them with their problems. We also asked for criticisms, but these
tended to be muted (system “a bit slow”, “lack of options in dealing with problems”). And
under suggestions, mothers stated that there should be more problems for Carmen to
address, often giving specific examples, and wanted an increased ability to interact on the
problem solving. The RAs also filled out a questionnaire on each mother’s experience. In
open comments, they noted that several mothers reported identifying with Carmen or her
problems, although one mother reported that Carmen should stop feeling sorry for her self.
One RA commented that identification led to a more positive interaction with CBI. Thus,
both instruments indicate a definite need for more tailoring. RAs also identified technical
glitches. Finally, subjects were given the same instrument of the larger study designed to
evaluate problem-solving skills. Analysis of this final outcome measure is still underway.
We conjecture that effectiveness will depend on site and perhaps RA supervising the
intervention.
4. Related Work
This work complements, but is distinct from, the work of Schank, Klein, and others on the
edifying narratives of personal experience (“war stories”) [14,5]. Such narratives are
fundamentally non-interactive, unlike IPD. Some story-oriented learning methods (e.g.,
Jasper series [2]) promote active discussing and problem solving in the context of stories,
but do not provide learners a way of seeing the consequences of their decisions. The goal-
based scenarios of Schank [14] are closer in spirit, however IPD takes into account both the
dramatic and pedagogical decisions that must be made in managing an interactive story.
There is also related work in the areas of managing interactive drama [4,9,15,18]. For
example, Weyhrauch[18] structures the space of all possible stories in terms of possible
permutations of plot points or important moments and a preference function on which
orderings of those points it prefers. When an important plot transition is recognized then the
manager does a forward projection of all possible extensions of the current story, chooses
one that it prefers and proceeds to try to manipulate the world so that story occurs. In
contrast, CBI’s emphasis is on flexibly managing the dialog interactions to motivate the
characters to achieve the story’s goals.
5. Conclusion
The results of the exploratory evaluation are promising for the use of IPD in health
interventions. Although the expectation that a system like CBI could substitute for time
spent with a trained clinical researcher teaching Bright IDEAS is bold, the reality is that the
current alternative of repeated one-on-one sessions with RAs is not economically feasible
for reaching a larger audience. Similarly, discussion groups or role-playing games also
require specially trained facilitators and would be difficult for these busy mothers to
participate in. On the other hand, passive presentation methods are not likely to promote
active learning. Interactive Pedagogical Drama could fill a void in making effective health-
intervention training available to the larger public at their convenience. The training task for
Carmen’s Bright IDEAS was a difficult one, fraught with many potential pitfalls. The fact
that it was so well received by the mothers was remarkable, and bodes well for applying
IPD to other training and learning tasks.
Acknowledgements
We would like to thank our clinical collaborators, particularly O.J. Sahler, MD, Ernest Katz, Ph.D.,
James Varni, Ph.D., and Karin Hart, Psy.D. Supported in part by the National Cancer Institute under
grant # R25CA65520.
References
[1] Bruner, J. (1990). Acts of Meaning. Harvard Univ., Cambridge, MA
[2] Cognition and Technology Group at Vanderbilt. (1992). The Jasper series as an example of anchored
instruction: Theory, program description and assessment data. Educational Psychologist, 27, 291-315.
[3] Gratch, J. (2000) Émile: marshalling passions in training and education. Proceedings of the Fourth
International Conference on Intelligent Agents, Barcelona, SPAIN. 2000.
[4] Kelso, M.T., Weyhrauch, P. & Bates, J. (1993) Dramatic Presence. In Presence: Journal of Teleoperators
and Virtual Environments, 2, 1, MIT Press.
[5] Klein, G. (1998). Sources of Power: How People Make Decisions. Cambridge, MA: MIT Press
[6] Lazarus, R.S. (1991) Emotion and Adaptation. Oxford Press.
[7] Lepper, M.R. & Henderlong, J. (2000). Turning _play_ into _work_ and _work_ into _play_: 25 years of
research in intrinsic versus extrinsic motivation. In Sansone & Harackiewicz (Eds.), Intrinsic and Extrinsic
Motivation: The Search for Optimal Motivation and Performance, 257-307. San Diego: Academic Press.
[8] Lester, J. C., Towns, S. G., Callaway, C. B., Voerman, J. L., & FitzGerald, P. J. (2000). Deictic and
Emotive Communication in Animated Pedagogical Agents. In J. Cassell & S. Prevost & J. Sullivan & E.
Churchill (Eds.), Embodied Conversational Agents (pp. 123-154). Cambridge: MIT Press.
[9] Machado, I. & Paiva, A. (1999) Heroes, villains, magicians,…: Believable characters in a story creation
environment. In AIED-99 Workshop on Animated and Personified Pedagogical Agents, Le Mans, 39-46.
[10] McKee, R. (1997). Story. Harper Collins, NY, NY.
[11] Marsella, S. (2002) Pedagogical Soap. In Dautenhahn, Bond, Canamero & Edmonds (Eds) Socially
Intelligent Agents: Creating Relationships with Computers and Robots. Kluwer.
[12] Marsella, S. Johnson, W.L. & LaBore, C. 2000. Interactive Pedagogical Drama. In Proceedings of the
Fourth International Conference on Autonomous Agents, 301-308.
[13] Poggi, I. & Pelachaud, C. (1999) Emotional meaning and expression in performative faces. In
International Workshop on Affect in Interactions: Towards a New Generation of Interfaces, Siena, Italy, 1999.
[14] Schank, Roger C. (1996) Goal-Based Scenarios: Case-Based Reasoning Meets Learning by Doing. In
Leake, David, Eds. Case-Based Reasoning: Experiences, Lessons & Future Directions, AAAI Press 295-347.
[15] Swartout, Hill, Gratch, Johnson, Kyriakakis, LaBore, Lindheim, Marsella, Miraglia, Moore, Morie,
Rickel, Thiebaux, Tuch, Whitney & Douglas (2001) Toward the Holodeck: Integrating graphics, sound,
character and story. In Proceed. of Fifth International Conference on Autonomous Agents, Montreal, 409-416.
[16] Varni, J.W., Sahler, O.J., Katz, E.R., Mulhern, R.K., Copeland, D.R., Noll, R.B., Phipps, S., Dolgin, M.J.,
& Roghmann, K. (1999) Maternal problem-solving therapy in pediatric cancer. Journal of Psychosocial
Oncology, 16, 41-71.
[17] Vygotsky, L. (1978). Mind in society: The development of higher psychological processes. (M. Cole, V.
John-Steiner, S. Scribner, & E. Souberman, Eds. and Trans.). Cambridge, England: Cambridge Univer. Press.
[18] Weyhrauch, P. (1997) Guiding Interactive Drama. Ph.D. Thesis. Tech Report CMU-CS-97-109. CMU.
