Using the Cooperative Board Game Pandemic to Study Teamwork
Emily C. Anania, Joseph R. Keebler, Katlin M. Anglin, & Jason P. Kring
Embry-Riddle Aeronautical University
Teamwork is paramount in most modern-day career fields. It is important for Human Factors students and
professionals to understand the knowledge, skills, and attitudes (KSAs) requisite for excellent teamwork and to
grasp the many hurdles that exist in appropriately measuring its major constructs. Although theories can be
imparted didactically, and ideas for measurement discussed, conducting team-based research continues to be
extremely difficult, in particular for measuring team behaviors. This paper discusses a potential remedy to this
issue through the team-based game “Pandemic©.” For this game, players work together on teams of 2-4
individuals with the goal of curing a set of 4 diseases that are eradicating humanity. Pandemic forces
individuals to be resourceful and work together - and exemplifies many of the behavioral, attitudinal and
cognitive components of teamwork. This paper discusses: a) the KSAs, team processes, and measurable
outcomes that can be studied through Pandemic, b) aspects of the environment in Pandemic that can be
manipulated, and c) an example of a study currently underway using Pandemic.
Conducting teamwork research in experimental settings
can prove difficult. Oftentimes, research is conducted in
complex simulation environments which requires considerable
costs and extensive training of participants to achieve
reasonable levels of performance. These issues can lead to
longer experimental times, as well as fatigued participants and
researchers. Team research is varied in methodology, but
games can offer unique testbeds for studying teams in action.
Bowers, Salas, Prince, and Brannick (1992) introduced
the potentiality of using low-cost, low-fidelity simulators in
order to study teams. Bowers and Jentsch (2001) suggest that
commercial, off-the shelf (COTS) programs are a better
alternative to costly simulations created solely for the purpose
of research, and that some COTS games are relatively easy to
implement, while still being useful for measurement of team
constructs. Indeed, computer games (Edwards, Day, Arthur, &
Bell, 2006) and computer simulations (Mathieu, Heffner,
Goodwin, Salas, & Cannon-Bowers, 2000) have both been
successfully utilized to study team mental models. However,
these technology-based approaches have the propensity to be
time-consuming, and data collection is at risk if the
technology malfunctions. Additionally, it is often difficult to
control a large set of variables in an automated game setting.
Therefore, COTS video games can be difficult to select to
ensure alignment with the specific nature of a given research
The present paper reviews the use of a board game as a
potential remedy to these problems. Specifically, we explain
the use of a cooperative board game, Pandemic, as an
experimental environment. This game is easy to learn and
implement, and forces players to act as a coordinated team in
order to win. This can serve as an excellent test-bed for
understanding teamwork, team processes, and outcomes. In
this paper, we first discuss the gaming environment of
Pandemic, and then describe some of the major theoretical
components of teamwork that we believe can be examined and
manipulated within the Pandemic environment. We conclude
by outlining a set of potential studies highlighting the utility of
the game in a real experimental setting.
PANDEMIC GAMING ENVIRONMENT
Pandemic is a multi-player cooperative game where
players attempt to eradicate the world of four diseases. The
game is available as both a board game, as well as an iTunes
application. The board game costs roughly 30 dollars, and the
iTunes app version costs $6.99, making either an affordable
platform. Both versions of the game have identical rules. The
game board consists of a map of the world broken into four
regions with major regional cities spread across the continents.
The four diseases, represented by small cubes corresponding
to the color of the region where they originate, are initially
relegated to four geographical locations – red for Asia, black
for Eastern Europe and southern Indo-Eurasia, blue for
Western Europe and North America, and yellow for Africa
and South America (Figure 1).
Figure 1. Board for the game Pandemic. Notice the colored
regions indicating the different diseases.
As the game progresses, the disease cubes “spread” as the
players draw cards from the infection deck that represent the
Copyright 2016 by Human Factors and Ergonomics Society. DOI 10.1177/1541931213601405
Proceedings of the Human Factors and Ergonomics Society 2016 Annual Meeting 1770
prevalence of the disease in a particular city (see Figure 2). If
at any time a city contains more than three cubes it
“outbreaks” and disperses cubes to all adjacent cities. To
combat the diseases, players collect player cards, at a rate of
two per turn, containing the names of the various cities
throughout the world. Each card is further colored given the
region of the city it is from (see Figure 3). To “cure” a disease,
players must coordinate their actions such that one of them
can discard 5 cards of the same color. The game is designed to
make it extremely difficult to achieve this goal – making the
game challenging to win. During a turn, the player receives
four action points that can be spent on various activities such
as moving around the board, curing disease cubes, or curing a
disease by discarding five of the same colored cards. Between
each player’s turn, an “infection” card is turned over,
indicating which cities to place disease cubes on before then
next player’s turn. Although these rules are relatively basic,
the game state can become very complex in relation to the
various roles players can take on and where cubes are placed
at the beginning of the game. Throughout the next section we
will focus on aspects of the game environment that can be
manipulated to answer team-based research questions.
Figure 2. Example of infection cars with one from each
Figure 3. Example of player cards with one from each
The players have access to seven different specialist roles
that alter the rules of the game in different ways depending on
their profession. For instance, the Scientist only needs to
discard four cards (instead of five) to cure a disease, whereas
the Medic can cure all disease cubes on a city for 1 action
point instead of the usual rule of 1 point per cube. Although
full knowledge of the rules is necessary to understand the
value of individual characters, we will briefly describe how
each character affects the team’s ability to win the game.
Understanding the different player roles is integral to
manipulating the game environment for research purposes (see
Scientist: Cure diseases easier by using less cards than
normally needed (four instead of five).
Researcher: Allows easier exchange of cards. Normally,
players must have both of their pawns in the same city as the
player card they wish to exchange. For instance, to exchange
Moscow, both players must be in Moscow and it must be the
turn of the player giving the card. The Researcher instead can
take any card from any player, as long as they are both co-
located in the same city.
Medic: The Medic can remove disease cubes from cities
much easier than other players. Usually, each player can only
remove one disease cube per action point, but the Medic can
remove all disease cubes for 1 action point.
Dispatcher: The Dispatcher gives the team an advantage
in their movements – instead of spending movement points
each turn on moving their own pawn, the Dispatcher can
instead move someone else’s pawn. Further, they can spend an
action point to move one pawn to any other pawn on the
Quarantine Specialist: The Quarantine Specialist creates
a “safe zone” in the city they are located in as well as all
adjacent cities. No disease cubes can be placed on any of these
cities while the specialist is nearby.
Operations Expert: The Operations Expert can build
research stations in any city as an action, instead of the usual
need to discard the city card where the station will be built.
They can also move from research stations to any other city
simply by discarding any card.
Contingency Planner: Special event cards in the player
deck can be re-used by the Contingency Planner, where they
are usually removed from the game instead.
Proceedings of the Human Factors and Ergonomics Society 2016 Annual Meeting 1771
Figure 4. Player role cards.
Major Constructs of Teamwork
Below we briefly discuss each of the major constructs of
teamwork, conceptualized by Salas et al. (2014), that are
elicited by Pandemic and how they can be studied. To
summarize, these constructs are:
Composition: The respective roles and expertise of the
Cognition: Shared mental models, transactive memory,
and team understanding of each other’s knowledge.
Communication: The pushing and receiving of
information between members of the team.
Cooperation: Intrinsic and extrinsic motivation that
pushes the team to perform well together.
Coordination: Actions in concert needed to complete a set
of shared goals.
Coaching: The manifestation of leadership in relation to
resource management and goal setting.
Conflict: The lack of overlap between different team
The following sections explore how Pandemic can be
used to study each of these constructs.
Composition. Team composition has a strong influence
over team performance. Because Pandemic includes various
team roles which players occupy during the game, it offers the
opportunity for a very clear team composition manipulation.
Team composition can refer to either composition of the team
from the seven available roles, or composition of the team
members who are playing those roles. Variables manipulated
could include surface-level characteristics (i.e., sex, race, or
age) or deep-level characteristics (i.e., personality, values,
team roles, or gender roles).
Previous team composition literature has focused on
gender composition, personality, and team roles. Composition
is most often studied in conjunction with team performance
and cohesion. Research by Mathieu, Tannenbaum,
Kukenberger, Donsbach, and Alliger (2015) has focused on
validating a measure of what they call Team Role Experience
and Orientation (TREO), but has not explored which of these
roles must be present in order to have a successful team.
Teams for Pandemic could be selected and measured on the
TREO, or on other characteristics in order to explore the
necessity of different personalities and team roles on a given
Manipulating composition. Manipulating composition can
be easily done in Pandemic through pre-choosing the various
player roles. Normally the rules state that the player roles are
to be randomly distributed amongst the player prior to starting
the game. Instead, researchers could choose a pre-determined
set of player roles, and then administer them to players either
randomly or via matching with individuals based on their
personalities or some other factor relevant to the current
research question. Finally, there is arguably a “best” team
when it comes to composition in Pandemic. Specifically,
having the Dispatcher, Researcher, Medic, and Quarantine
Specialist provides an ideal set of abilities that is arguably
better than any other mixture of roles. Researchers could
manipulate which of these roles is present or absent as a
function of how well the human team is playing the game.
This could answer interesting questions – Do individuals who
play the game well need these specific roles? What roles are
the most flexible in terms of being needed on the team for
success? What individual differences and characteristics of
participants aligns them with a particular in-game role?
Cognition. Training or familiarity with game rules is
needed to play Pandemic with any degree of success.
However, in a laboratory setting, participants will not always
be familiar with typical gameplay. The game changes as
people become more familiar with the rules, or more familiar
with their teammates. Team cognition is a positive predictor of
not only performance, but also motivational states (Dechurch
& Mesmer-Magnus, 2010). Team cognition can be studied
through Pandemic with assessments, before or after the game,
about knowledge and beliefs of the game, as well as players’
strategies and plans.
Manipulating Cognition. One of the major ways to
manipulate cognition is through the player card system.
Normally the rules state that players cannot show each other
their cards, and instead can only tell each other what cards are
in their hands. This rule could be changed so that players play
with their cards face up, which would imitate a high level of
team cognition and shared understanding. This rule could also
be changed to limit the players even more, demonstrating a
low level of team cognition. For instance, players could
instead only be allowed to show or talk about one card per
turn, or only cards of a particular color.
Communication. Much like team performance in all
contexts, communication between teammates is necessary to
do well in Pandemic. Team members need to share
information with each other – namely what cards they are
holding. Without knowledge of other people’s abilities and
possible actions, team members are unable to effectively play
the game and eradicate diseases.
Communication is difficult to measure, but not
impossible. For example, based on a class exercise that used
Pandemic to study teamwork, one student developed a
software platform to observe and record communication
Proceedings of the Human Factors and Ergonomics Society 2016 Annual Meeting 1772
between team members (Keebler, DiazGranados, & Smith,
2014). Additionally, video recording can be used to study
teams after they have already played the game.
Manipulating Communication. Communication can be
experimentally manipulated through prior training of the team.
Specifically, teaching the team members to use closed-loop
communication could enhance team performance and change
the way the teams structure their decisions. If the researchers
are interested in studying virtual teams, players could instead
be in separate locations and play via some type of interface.
The amount of communication in the game could also be
manipulated by dictating which players are allowed to speak,
and when. In some video games, it is common for users to
have preset expressions or phrases with which they can
communicate – instituting this in Pandemic would also be an
interesting way to manipulate communication.
Cooperation. Pandemic is a uniquely cooperative game,
and therefore forces players to be reliant on each other for
success. Cooperative action arises out of “attitudes, beliefs,
and feelings of the team” (Salas et al., 2014, p. 6). That being
said, cooperation is not necessarily the main aim of all players
involved. Some participants may not have prosocial
tendencies, and will favor a competitive strategy, even in a
supposedly cooperative environment. Cooperative goals are
hard to measure in the moment, but attitudinal states, beliefs,
and feelings are relatively easy to capture.
Manipulating Cooperation. In order to manipulate
cooperation, researchers would likely need to give players
individual goals in the game (e.g. your pawn must make it to
Paris by Turn 3), as well as team goals (e.g. the team must
cure at least 2 diseases). This change in gameplay would
necessitate that players have motivation to not only succeed as
a team, but also as an individual. By manipulating players’
perceptions of the game, cooperation can be studied.
Coordination. The interdependency of Pandemic requires
that players very closely coordinate their actions, sometimes
several moves ahead. Coordination can be manipulated or
observed through the amount of information shared between
players, and to what degree that information is utilized. For
example, Pandemic’s diseases are “spread” through outbreaks
of disease cubes, and these could possibly serve as a metric for
measuring coordination – how many cubes did the team clear
from the board versus how many did they leave un-cleared?
Winning Pandemic is a difficult endeavor, and simply winning
the game is not necessarily a good measure of team
Manipulating Coordination. One specific role in
Pandemic is integral to coordination: the Dispatcher. The
Dispatcher is uniquely able to move other players’ pawns in
an effort to aide gameplay by bringing the pawns together so
that players can “share knowledge.” If the Dispatcher were to
be taken out of play, either before the game, or during the
game, coordination would likely decrease, and the team would
demonstrate a low level of coordination.
Coaching. Coaching (or leadership behaviors) during
Pandemic can significantly affect team performance. As
leadership does not always come from a sole leader, different
team members can also offer help guide behavior. This means
that leadership can be studied in distinctly different ways
using Pandemic. Leadership behaviors can simply be observed
through gameplay and correlated with any other variables of
interest. However, a confederate player (or multiple) may be
utilized – either to serve as a team “leader” or to serve as a
team follower, while evaluating the leadership of the
Manipulating Coaching. The use of confederate players
would be one manipulation for researching coaching and
leadership behaviors. One confederate player may be utilized
in order to test out different leadership strategies and their
effect on team outcomes. A team of confederates with one
participant may also be useful for evaluating leadership
potential and effectiveness.
Conflict. Pandemic presents many opportunities for
conflict between team members as well as conflicts within the
game that the team needs to solve, considered as relationship-
and task-based conflicts (Salas et al., 2014). Conflict will arise
naturally in Pandemic, situations can also be constructed in
order to cause conflict between team members. In most teams,
some type of conflict will challenge the people involved.
Pandemic offers a natural avenue through which to study the
effects of conflict on group performance as well as on other
Manipulating Conflict. As stated, stacking the decks can
cause conflict at a specific point in the game, controlled by the
experimenter, in order to study task-based conflict.
Relationship-based conflicts can be achieved through the use
of a confederate player to instigate different conflicts.
EXAMPLE STUDY ON TEAM ADAPTATION
As Pandemic requires players to use complex reasoning
and inference as well as coordinate actions, the team is also
required to quickly adapt strategies to avoid or limit outbreaks.
For example, when “Epidemic” cards are turned over, a
disease appears full strength in a previously uninfected city,
jeopardizing the team’s efforts toward winning the game.
Thus, the team needs to adapt to disruptive, unexpected
events, which makes Pandemic an ideal tool for research on
team adaptation. Maynard, Kennedy, and Sommer (2015)
define the team adaptation process as “…adjustments to
relevant team processes (i.e., action, interpersonal, transition)
in response to the disruption or trigger giving rise to the need
for adaptation” (p. 656). According to this model, behavioral
changes related to action processes (e.g., monitor progress,
coordinate with team members), interpersonal processes (e.g.,
conflict management, motivation and confidence building),
and transition processes (mission analysis, planning, strategy
formulation) determine how well a team adapts. However, a
valid and reliable measure of these behavioral changes does
not currently exist.
To address this need, we are using Pandemic to test and
validate a behaviorally-anchored rating scale called the Small
Team Adaptation Rating Scale (Anglin & Kring, 2016), or
STARS. The preliminary version captures the frequency of
behaviors during a task or mission organized into action,
interpersonal, and transition processes. As four-person teams
Proceedings of the Human Factors and Ergonomics Society 2016 Annual Meeting 1773
play the game, we are using STARS to measure how often
teams engage in behaviors related to the three types of
processes, particularly after the appearance of “Epidemic”
cards, and determine if other behaviors should be included in
the measure. After refining STARS, we plan to conduct a
second study, again with four-person teams. Based on ratings
of each team’s performance and their STARS ratings, derived
from video recordings, we will determine how well the metric
differentiates between low- and high-adapting teams. As
Pandemic is an efficient, low-cost way to capture the team
adaption processes, this game will help validate the STARS
measurement tool, a key benefit that can be used in future
investigations to compare team adaptation in different settings
and mission scenarios.
This paper consists of a brief, high-level review of the
game Pandemic with a specific focus on the game
environments potential as a tool for team research. It is our
hope that this will aid others who study teamwork in using a
low-cost, easily implemented, and easy to learn environment
that can be used to study multiple facets of teamwork with a
low barrier of entry for both the researchers and the
participants. Although we have highlighted some major
constructs of teamwork and provided examples, this is just a
small portion of the potential research questions that could be
answered and studied in this environment. It is our hope that
others will utilize this work to further their own research
Anglin, K. M, & Kring, J. P. (2016). Lessons from a space
analog on adaptation for long-duration exploration
missions. Aerospace Medicine and Human Performance,
Bowers, C. A., & Jentsch, F. (2001). Use of commercial, off-
the-shelf, simulations for team research. Advances in
Human Performance and Cognitive Engineering
Research, 1, 293-318.
Bowers, C., Salas, E., Prince, C., & Brannick, M. (1992).
Games teams play: A method for investigating team
coordination and performance. Behavior Research
Methods, Instruments, & Computers, 24(4), 503-506
DeChurch, L. A., & Mesmer-Magnus, J. R. (2010). The
cognitive underpinnings of effective teamwork: A meta-
analysis. Journal of Applied Psychology, 95(1), 32-53.
Keebler, J. R., DiazGranados, D., & Smith, D. C. (2014,
September). Learning team theories and measurement
through the game Pandemic. In Proceedings of the
Human Factors and Ergonomics Society Annual Meeting
(Vol. 58, No. 1, pp. 442-446). SAGE Publications.
Leacock, M. (2012). Pandemic. Quebec, Canada: Z-Man
Mathieu, J. E., Tannenbaum, S. I., Kukenberger, M. R.,
Donsbach, J. S., & Alliger, G. M. (2015). Team role
experience and orientation: A measure and tests of
construct validity. Group & Organization
Management, 40(1), 6-34.
Maynard, T. M., Kennedy, D. M., & Sommer, A. S. (2015).
Team adaptation: A fifteen-year synthesis (1998–2013)
and framework for how this literature needs to “adapt”
going forward, European Journal of Work and
Organizational Psychology, 24(5), 652-677, DOI:
Salas, E., Shuffler, M. L., Thayer, A. L., Bedwell, W. L., &
Lazzara, E. H. (2014). Understanding and improving
teamwork in organizations: A scientifically based
practical guide. Human Resource Management, 54(4),
Proceedings of the Human Factors and Ergonomics Society 2016 Annual Meeting 1774