Examining the behavioral and structural characteristics of team leadership in extreme environments

Abstract

Despite the growing interest in extreme teams, there is currently a lack of understanding concerning leadership within such teams, as the literature has predominantly focused on team leadership within the context of traditional organizations. The current study investigates team leadership within the context of teams operating in extreme environments, with a specific focus on teams operating in isolated, confined environments. We seek to identify team leadership functions as well as a subset of structural characteristics associated with team leadership in extreme environments (i.e., formality of leadership, locus of leadership, and leadership distribution). We leverage a historiometric approach to capitalize on real historical examples of extreme teams that are rich with critical information regarding actual team leadership functions occurring in extreme settings. Results suggest that team leadership functions such as team problem solving, supporting social climate, structure and planning, and sensemaking are among the most prevalent. Results also indicated that the degree to which leadership is distributed throughout the team as well as the formality of leadership varies across action and transition phases of the team's task cycle.

Full text

SPECIAL ISSUE ARTICLE
Examining the behavioral and structural characteristics of team
leadership in extreme environments
C. Shawn Burke
1
|Marissa L. Shuffler
2
|Christopher W. Wiese
3
1
University of Central Florida, Orlando,
Florida, U.S.A.
2
Clemson University, Clemson, South
Carolina, U.S.A.
3
Georgia Institute of Technology, Atlanta,
Georgia, U.S.A.
Correspondence
C. Shawn Burke, University of Central Florida,
Orlando, Florida, U.S.A.
Email: sburke@ist.ucf.edu
Funding information
Johnson Space Center, Grant/Award Number:
NNX14AK54G; NASA, Grant/Award Number:
NNX14AK54G
Summary
Despite the growing interest in extreme teams, there is currently a lack of under-
standing concerning leadership within such teams, as the literature has predominantly
focused on team leadership within the context of traditional organizations. The
current study investigates team leadership within the context of teams operating in
extreme environments, with a specific focus on teams operating in isolated, confined
environments. We seek to identify team leadership functions as well as a subset of
structural characteristics associated with team leadership in extreme environments
(i.e., formality of leadership, locus of leadership, and leadership distribution). We
leverage a historiometric approach to capitalize on real historical examples of extreme
teams that are rich with critical information regarding actual team leadership functions
occurring in extreme settings. Results suggest that team leadership functions such as
team problem solving, supporting social climate, structure and planning, and
sensemaking are among the most prevalent. Results also indicated that the degree
to which leadership is distributed throughout the team as well as the formality of
leadership varies across action and transition phases of the team's task cycle.
KEYWORDS
extreme environments, team leadership, teams
1|INTRODUCTION
The greatest leader is not necessarily the one who does
the greatest things. He is the one who gets the people
to do the greatest things (President Ronald Reagan)
Space exploration teams, polar exploration teams, long‐duration
sailboat racing teams, mountaineering teams, and provincial recon-
struction teams; what is the uniting factor among all these teams?
They are extreme teams. In essence, extreme teams are teams that
“(a) complete their tasks in performance environments with one or
more contextual features that are atypical in level (e.g., extreme time
pressure) or kind (e.g., confinement, danger) and (b) for which ineffec-
tive performance has serious consequences …” (Bell, Fisher, Brown, &
Mann, 2016, p. 2). Extreme team members are exposed to atypically
high levels of stressors that (a) appear in combination simultaneously
and (b) may be a mixture of chronic and acute in nature. The broader
literature on teams operating in what can be considered extreme con-
texts indicates that teams will often fail under such conditions—such
intense, continuously stressful situations can emphasize a focus on
self, decrease team perspective, decrease prosocial behaviors, and
negatively impact team decision‐making processes (Driskell & Salas,
1991; Driskell, Salas, & Johnston, 1999). Due to the critical nature of
extreme team performance, there have been repeated calls for more
research to facilitate an understanding of the factors that contribute
to team effectiveness within such environments (Bell et al., 2016;
Friedrich, Vessey, Schuelke, Ruark, & Mumford, 2009; Hannah,
Campbell, & Matthews, 2010; Keeton, Schmidt, Slack, & Malka, 2012).
One element that has been argued to play a critical role in
extreme team effectiveness is team leadership (e.g., Mulhearn et al.,
2016; Crosby, 2008; Hannah, Uhl‐Bien, Avolio, & Cavarretta, 2009).
The purpose of leadership in any given team is to establish goals and
set direction that will lead to accomplishing these goals (Zaccaro,
Rittman, & Marks, 2001). Previous research suggests that team
Received: 16 February 2017 Revised: 21 February 2018 Accepted: 29 March 2018
DOI: 10.1002/job.2290
J Organ Behav. 2018;1–15. Copyright © 2018 John Wiley & Sons, Ltd.wileyonlinelibrary.com/journal/job 1

leadership is a critical component of ensuring effective team processes
and team outcomes (Burke et al., 2006; Nicolaides et al., 2014; Salas,
Sims, & Burke, 2005). Effective team leadership may be a critical
lynchpin in helping teams face the challenges of extreme contexts,
by performing necessary functions that can aid in reducing stressors
specific to the task at hand as well as supporting a positive social
climate that encourages effective teamwork under difficult circum-
stances (Zaccaro, Heinen, & Shuffler, 2009). However, to date, there
has been little investigation regarding how team leaders can aid in
facilitating effective team processes and performance within extreme
contexts. Accordingly, it becomes imperative to understand what
leaders can do to mitigate challenges and facilitate optimal team
interactions within such environments so that team performance is
maximized.
In particular, although the team leadership literature offers
evidence for what types of functions may be most critical for leading
teams in traditional organizational settings, it is unclear if these
functions are similar or different in leading for extreme contexts.
Additionally, given the complexity of extreme teams, the structure of
leadership may also look very different; there may be multiple leaders
on a team, with different members sharing leadership responsibilities
or rotating leadership to ensure effectiveness during different phases
of teamwork (Zaccaro & DeChurch, 2012). Overall, the need to
understand team leadership in such contexts has become increasingly
recognized (e.g., Suedfeld, 2012). Yet Hannah et al. (2009) noted that
“leadership in extreme contexts may be one of the least researched
areas in the leadership field.”(p. 897).
Therefore, the goal of the current study is to move the literature
forward in understanding the leadership functions and structural com-
ponents of leadership occurring in extreme contexts. In doing so, we
offer two major advancements in this area. First, we identify prevalent
team leadership functions that occur in extreme team contexts, which
offers insight as to whether existing taxonomies are relevant to such
teams and, if so, what functions may be most often implemented.
Second, we identify key structural characteristics associated with lead-
ership in extreme teams, namely, the degree to which leadership is
carried out via formal or informal leaders, the degree to which leader-
ship is shared by multiple members or contained within a single indi-
vidual, and how structure of leadership may shift in response to
changing team needs over time. These advancements meet a critical
need for expanding team leadership theory and research into extreme
contexts, as well as offering practical guidance regarding the leader-
ship functions and structural characteristics that need to be addressed.
To achieve our aims, we employ historiometry (Simonton, 2003)
to analyze archival documentation of crew interaction within the con-
text of historical events wherein teams were operating in extreme
environments. In the following, we first present background on team
leadership and a set of hypotheses driving our inductive approach.
Next, we summarize our methodology and then present our results
along with a discussion of their implications. In summarizing thematic
findings across the different leadership functions and structural
components of leadership needed in extreme teams, we offer both
answers to our research questions as well as provide future avenues
for exploring and conceptualizing critical leadership needs for teams
in these extreme settings.
2|TEAM LEADERSHIP
Leadership is not a new concept, but it is a complex one that has been
extensively researched throughout history and has been connected to
a range of team and organizational performance factors (Yukl, 2006).
However, most leadership research has focused on leading individuals
or individual leader–follower relationships (DeChurch, Hiller, Murase,
Doty, & Salas, 2010), neither of which directly speaks to the processes
involved in leading teams. In focusing on team leadership, attention
must be paid to what leader(s) do to build and maintain the dynamic
processes and states that facilitate coordinated team action; a focus
is lacking within the literature on individual leadership (Kozlowski,
Watola, Jensen, Kim, & Botero, 2009).
Most recognize leadership as a process of social influence where
the role of team leadership is “to do or get done, whatever is not being
adequately handled for group needs”(McGrath, 1962, p. 5). In this
vein, leaders must engage in social problem solving, which, in turn,
moves the team towards their objective or goal(s). In exploring team
leadership in this manner, researchers have begun to identify how
leaders can facilitate the team's ability to develop the shared behavior,
cognition, and affect that allow teams to progress towards and accom-
plish team goals. In setting and maintaining the above conditions, team
leaders engage in social problem solving whereby they are responsible
for ensuring that problems impacting goal attainment are diagnosed,
solutions are generated, and plans are developed and implemented
(Zaccaro et al., 2009). In doing the above, leaders are building the
shared affective, behavioral, and cognitive capacity within the team
that facilitates coordinated interaction in response to social problems.
Leveraging work of Fleishman et al. (1991), it is argued that leaders
influence the development of team cognition through processes such
as sensemaking, sensegiving, identifying problem needs, planning, and
developing and motivating team members (Zaccaro et al., 2001).
Leaders also influence team affect in that feedback, selecting and
developing personnel, and utilizing and monitoring personnel
resources can impact team affective states such as conflict and emo-
tion. Finally, leadership processes such as planning, coordinating per-
formance strategies, developing members, motivating, and providing
feedback to members may impact motivational states such as cohe-
sion and collective efficacy (Zaccaro et al., 2001). The development
of these emergent states as well as leadership processes, which
facilitate team coordination (i.e., matching member capabilities to role
requirements, offer clear strategies, monitoring environmental
changes, providing feedback, and recalibrating action), results in teams
being able to capitalize on the potential synergy present and maximize
team performance.
The work of Kozlowski and colleagues (2009; 1996) also speaks
to this notion of functional leadership by highlighting dynamic
processes that leaders engage in to move teams through a
developmental progression. Specifically, Kozlowski et al. (2009) argue
that as team's progress through developmental transitions (i.e., new,
novice, expert, and adaptive), the role of the team leader also
changes (i.e., mentor, instructor, coach, and facilitator) as does the
team leader's focus (i.e., identification and commitment, taskwork
capability, teamwork capability, and adaptive capability). Most
recently, Burke, Monsky, and Salas (2017) have further expanded on
2BURKE ET AL.

this notion by beginning to delineate the functions that may be
most relevant within each focus point.
In essence, the prior work on team leadership argues that team
leadership creates the enabling conditions for effective team perfor-
mance (Hackman, 2002) by developing and maintaining the shared
behavior, affect, and cognition, which facilitates explicit/implicit coor-
dination, adaptation, and team self‐regulation (Kozlowski et al., 2009).
Building upon prior taxonomies, incorporating recent advancements in
team science, and further extracting the themes underlying functional
leadership theory are works by Morgeson, DeRue, and Karam (2010).
Morgeson et al. (2010) critically reviewed the literature on team lead-
ership and recent advancements in team science to identify a set of 15
leadership functions that coalesce within the dynamism present in
team process, thereby representing the state of the art with respect
to team leadership theory. This work acknowledges the dynamic
nature of team leadership by notating leadership functions that are
most relevant during two temporal phases that have been argued for
within teams: transition and action.
Transition phases have been described as those periods of time
where teams are primarily focused on mission analysis, planning, and
strategy formulation (Marks, Mathieu, & Zaccaro, 2001). Morgeson
et al. (2010) identify the following leadership functions as being impor-
tant during transition phases: composing the team, defining the
mission, establishing expectations and goals, structure and planning,
training and developing the team, sensemaking, and the provision of
feedback (see Table 1 for definitions). Further, action phases have
been described as those periods of time when the team is focused
on actual goal accomplishment (Marks et al., 2001). In this vein,
Morgeson et al. (2010) argue for the importance of the following lead-
ership functions: monitoring team, managing team boundaries, chal-
lenging the team, performing team task, solving problems, provision
of resources, encouraging team self‐management, and supporting the
social climate (see Table 1 for definitions). Using this aforementioned
theoretical and empirical basis, Morgeson et al. (2010) offer a compre-
hensive integration of the literature in the form of a broad set of lead-
ership functions. Based on their review, these leadership functions are
that which, when enacted, serve to develop and later maintain the
shared affect, behavior, and cognition shown to facilitate effective
team performance. The focus on a broad set of leadership functions
has caused some to highlight the need for researchers to pay
increased attention to context and, in doing so, engage in a “… valida-
tion of the contextual influences that enhance the efficacy of some
leadership actions and diminish others”(Zaccaro et al., 2001, p. 455).
In this vein, we next describe the context in which team leadership
is embedded such that hypotheses can be developed.
3|EXTREME TEAMS AS A CRITICAL TEAM
LEADERSHIP CONTEXT
Given the importance of context, we next describe the specific con-
text within which we examine team leadership: the context of extreme
environments. Hannah et al. (2009) define extreme environments as
an “environment where one or more extreme events are occurring
or are likely to occur that may exceed the organization's capacity to
prevent and result in an extensive and intolerable magnitude of
TABLE 1 Taxonomy of team leadership functions (adapted from Morgeson et al., 2010)
Leadership function Definitions
Transition phase Compose team “Selecting individuals who will be successful in accomplishing the team task
outlined by the organization and then ensuring that the mix of individuals is
appropriate over time as the team develops and the environment evolves”
(Morgeson et al., 2010, p. 12).
Define mission “Determining and communicating the organization's performance expectations
for the team in such a way that they are broken down into tangible,
comprehensive pieces”(Morgeson et al., 2010, p. 13).
Establish expectations/goals Creating expectations regarding what is acceptable performance and setting team goals.
Structure and plan “Determining or assisting in determining how work will be accomplished …,
who will do which aspects …, and when the work will done …” (p. 15).
Train and develop team The provision of task and/or team training through instruction, demonstration,
and ongoing coaching
Sensemaking Identifying and assigning meaning to environmental events which may impact
the team in their progress towards goals attainment. Includes the clear
transmission of interpreted events to the team.
Action phase Provide feedback Giving, seeking, and receiving task‐feedback; providing constructive feedback
regarding errors and offering advice for improvement (Cannon‐Bowers, Tannenbaum,
Salas, & Volpe, 1995).
Monitor team “Monitoring and evaluating the team's progress towards task completion, the
resources available to the team, the team's external environment, and team member
performance”(Morgeson et al., 2010, p. 20).
Manage team boundaries Communicating and coordinating with important entities outside the team in an effort to
bring in and push out key information. Also includes buffering the team from external
forces which may disrupt effective work processes.
Challenge team Playing devil's advocate and challenging the team's assumptions, methods, and processes
with the goal of facilitating the identification of the best methods for task accomplishment.
Perform team task Engaging or intervening in some aspect of the team's task.
Solve problems Participating and supporting the team in all phases of problem solving, including assessment,
development of and implementation of solutions.
Provide resources Gathering and providing informational, personnel, and material resources needed by the team.
Encourage team self‐management Behaviors that foster team autonomy.
Support social climate Behaviors which facilitate the socioemotional health of the team.
BURKE ET AL.3

physical, psychological, or material consequences to—or in close phys-
ical or psycho‐social proximity to‐organization members”(p. 898).
Within this context, the extreme environments that we focus on
include space exploration, polar exploration, and long‐duration
sailboat racing. Although all three may be considered extreme action
teams, they differ in the degree to which they are isolated and con-
fined and the nature of the extreme events that are typical within each
context. Examining team leadership within this cross section of envi-
ronments allows us to identify the aspects of team leadership that
generalize across contexts, as well as those areas of contextual differ-
ences. To assist in fostering an understanding of each context, a brief
description of each is provided. We also delineate how each context
aligns with several defining characteristics of extreme environments
(i.e., location in time, magnitude of consequence, probability of conse-
quences, physical or psychosocial proximity, and form of threat;
Hannah et al., 2009; see Table 2).
3.1 |Space exploration
Space exploration is a mission critical environment whereby teams of
two to six individuals work both independently and interdependently
to accomplish mission essential tasks. Similar to the other two con-
texts described below, errors in this environment can have life‐threat-
ening consequences. Team members tend to be highly educated and
skilled, and crews are often multicultural. Training for spaceflight is
intense, with the average astronaut being in training anywhere
between 5 and 10 years from the time they are selected to fly (Vessey,
2014). Current missions have reachback capabilities to ground sup-
port/mission control; however, depending on the exact location of
the spacecraft in orbit, communication delays may exist.
The sources that form the primary data for our study in terms of
spaceflight originate from crews operating onboard the International
Space Station, Mir, Skylab, and Salyut, as well as crews participating
as part of the Mars 500 study. Within the above contexts, mission
lengths vary based on payload and the platform but range from around
30 to 500 days—with most averaging around 6 months. For the
inflight crews, the environment represents an isolated, confined envi-
ronment where the degree of confinement and isolation is dictated by
the size of the spacecraft and location in orbit. Although spaceflight
crews represent intact crews, portions of the crew do, at times, rotate
in/out (e.g., crew replacement is possible). This is slightly different
than the other two contexts where crew replacement is nearly impos-
sible (polar exploration) or most often precipitated by an extreme
event (long‐duration sailboat racing). Moreover, mission durations
tend to be shorter than within the other contexts analyzed.
3.2 |Polar exploration
Polar exploration refers to the process of the journey to and explora-
tion of the Artic and/or Antarctica. Although polar exploration con-
tinues into present time, the primary sources that form our data set
represent polar explorations as documented by Robert Falcon Scott,
Ernest Shackleton, and Ronald Amundsen. Polar explorers typically
made their transcontinental journey by navigating through often icy
waters and marching across land. For example, in January of 1915,
Shackleton sailed on the Endurance through a thousand miles of
packed ice, only for his ship to become locked in ice (where it drifted
for 10 months) a day short of his destination. Ultimately, Shackleton
and a small skeleton crew were forced to use a small lifeboat to
journey over 850 miles of heavy seas to the closest civilian outpost.
Physical conditions were harsh within this context, crews were often
placed in life‐threatening situations, and little reachback to friends
and family back home existed. Missions often lasted several years.
3.3 |Long‐duration sailboat racing
Long‐duration sailboat racing is often considered an extreme environ-
ment due to the unpredictability that is associated with ocean racing.
Racing is characterized by variations in workload and close confined
quarters. Wave height often varies during portions of the race, with
larger waves being of enough height and force to cause a crew to be
up the whole night trying to make sure that the ship stays afloat while
simultaneously being pelted by water from above and on the deck. It is
not uncommon for the crew to be exposed to three‐story swells and
iceberg‐covered seas that must be navigated. Crews typically average
around nine crewmembers that race 24 hr a day. Races can take over
9 months to complete and may cover as many as 39,000 nautical
miles. Although crewmembers may race for more than 20 days at a
TABLE 2 Contexts mapped to characteristics of extreme environments
Spaceflight Polar exploration Long‐duration sailboat racing
Team characteristics
Isolation Moderate High Low
Confinement Moderate–high High High
Duration Moderate Long Moderate–long
Predictability Moderate Low Low
Crew size Small Large Moderate
Event characteristics
Location in time Exclusively in situation Primarily in situation Primarily in situation
Magnitude of consequences High High High
Probability of consequences High High High
Leader–follower proximity Mixed physical Close physical Close physical
Form of threat Physical, psychological Physical, psychological Physical
4BURKE ET AL.

time (and are relatively confined during this time), races have several
legs whereby the crew may stop at different ports. Crew size is smaller
than the typical crew used in polar exploration.
4|TEAM LEADERSHIP FUNCTIONS IN
EXTREME TEAM CONTEXTS
Although the three extreme contexts chosen differ in many regards,
uniting the three is a focus on environmental complexity and/or
unpredictability, confinement, and the significant durations for which
the teams are isolated and/or confined away from family, friends,
and significant others. Moreover, when an extreme event occurs,
although the consequences are most immediately directed at a small
number of individuals, the magnitude of the consequences is high in
that they are often life‐threatening. The form of threat within all three
contexts not only is primarily physical but also has psychological com-
ponents. Additionally, there is high visibility of mission success/failure
outside the immediate team. Finally, the majority of incidents we
identified regarding extreme teams primarily speak to the interactions
that occur when a team is already in the extreme environment situa-
tion (i.e., not during the initial planning phases of the team's constitu-
tion). Together, these themes result in important implications for the
types of leadership functions expected. We next break down these
functions based on the transition and action phases to better identify
when they might occur.
4.1 |Leadership functions in transition phases of
extreme teamwork
We first expect that due to the event‐based nature of these extreme
team incidents, several of the team leadership functions seen early
on in team or mission formation most likely will not appear (i.e., com-
pose/restructure team, define mission, and establish expectations/
goals). Accordingly, we expect that transition phase leadership func-
tions in the extreme team incidents identified will primarily deal with
adaptive processes (e.g., structure/planning, sensemaking, and train/
develop) as well as managing the complexity and dynamism of the
environment. Limited reachback capabilities, tight coupling and cas-
cading of events, and the high magnitude of consequences in extreme
teams essentially create a need for the leader(s) to ensure that the
team is appropriately making sense of environmental events so that
team actions can be quickly adapted in light of changing conditions.
Therefore, we propose the following:
Hypothesis 1. Sensemaking and structure/(re)planning
are critical transition phase leadership functions for
teams operating in extreme environments.
Due to the magnitude of the consequences within these extreme
contexts, we also pose that another key transition phase leadership
function is training and developing the team. Teams operating in
extreme contexts drive a need for leadership that instills a learning cli-
mate. This is similar to many of the tenets within high‐reliability orga-
nizations yet at a team level. Team training and development includes
not only ensuring that task skills do not decay but also focusing on
ensuring that teams are developing and maintaining the shared affect,
behavior, and cognition that facilitates coordination and builds capac-
ity within the team. This, in turn, allows them to more fully engage in
team behaviors such as mutual performance monitoring and back up
behavior.
Hypothesis 2. Training and developing the team is a
critical transition phase leadership function for teams
operating in extreme environments.
4.2 |Leadership functions in action phases of
extreme teamwork
When thinking about leadership during the action phases of the team's
task cycle within extreme teams, solving problems becomes a critical
leadership function. By the very definition, extreme teams operate in
environments that are complex, dynamic, and often unpredictable (see
Table 2), and therefore, engaging in problem solving in situ such that
adaptive action can be taken to an unexpected event is critical. The
magnitude of consequences as well as the probability of conse-
quences when teams fail to appropriately respond to environmental
changes are high, often life threatening within all three contexts under
study.
Two other critical leadership functions within extreme teams are
driven by the isolated and confined nature of the teams under study,
as well as their duration—monitoring the team and challenging the
team. Within teams that have moderate to high degrees of isolation
from outside entities, monitoring and challenging would seem to go
hand‐in‐hand in an effort to avoid groupthink (Janis, 1982). Monitor-
ing is also important within this context to assist in determining when
team members might become overloaded to the stressors, as other
members can provide psychosocial support or possibly serve as
backup to the overloaded team member. Additionally, monitoring of
not only task progress but also crew interaction is critical in that it
may be a mechanism by which mistakes can be caught early before
they cascade and have life‐threatening consequences.
Teams working under stressful conditions, such as those present
in extreme environments, will often lose sight of the team's overarch-
ing goal and create conditions whereby lower‐status team members
become compliant to the input of high‐status members (Driskell
et al., 1999). In such situations, monitoring and challenging the team
may mitigate the tendency of lower‐status members to not offer input
or speak up when surrounded by members of higher status. Challeng-
ing the team is also a mechanism through which the devil's advocate
role can be enacted as a way to minimize the potential of groupthink.
This leads to the following hypothesis:
Hypothesis 3. Monitoring the team, challenging the
team, and solving problems will be critical action phase
leadership functions for teams operating in extreme
environments.
Up to this point, the leadership functions that have been argued
to be critical are primarily those which are task‐focused; however, it
is also important that leaders work to facilitate and maintain the psy-
chosocial well‐being of the team. Leadership functions, which serve
BURKE ET AL.5

to support the social climate within extreme teams, are important due
to the high degrees of stress and negative affect that can arise due to
the isolation from traditional support systems. Leadership functions
that support the social climate facilitate the team's motivation to
persist under difficult situations, as well as build cohesion within the
crew. This, in turn, should work to foster the team perspective that
often falters under stress (Driskell et al., 1999). Thereby, the following
hypothesis is put forth:
Hypothesis 4. Supporting the social climate will be a
critical action phase leadership function for teams oper-
ating in extreme environments.
5|THE MANIFESTATION OF LEADERSHIP
STRUCTURE IN EXTREME TEAM CONTEXTS
Traditionally, research has primarily focused on a single source of lead-
ership most often manifested in terms of formally designated leader-
ship roles. However, the functional approach to leadership implicitly
argues that leadership can be distributed throughout the team.
Although still a relatively new area of study, there have been promis-
ing findings to suggest that this distributed structure of leadership
facilitates effective teamwork and enhances team performance (e.g.,
Carson, Tesluk, & Marrone, 2007). In light of this, team leadership
researchers have begun to move beyond an approach, which is solely
leader‐centric to one that reflects a mixture of leader and team‐centric
approaches (Zaccaro et al., 2009). Leader‐centric approaches focus
primarily on the role of the individual leader and the manner in which
they develop and maintain effective team processes and states, which,
in turn, facilitate team performance. In contrast, team‐centric
approaches emphasize principles of collective or shared leadership,
where the responsibility for directing and managing team dynamics
is distributed throughout the team (see Manz & Sims, 1980; Pearce
& Conger, 2003). Moreover, recent work has suggested that team‐
centric approaches to leadership may explain more variance in team
performance than traditional leader‐centric approaches (Ensley,
Hmieleski, & Pearce, 2006). The predominant amount of work that
has been conducted on team‐centric approaches, however, has been
done outside the context of extreme teams. Therefore, the ways in
which leadership is manifested are an important aspect to examine
within the context of extreme teams.
5.1 |Leadership formality and locus
Extending existing work on team leadership, Morgeson et al. (2010)
take a mixed approach (leader‐centric, team‐centric) and argue for
the importance of research, which explicitly examines the formality
and locus of leadership. In this vein, a framework is developed, which
begins to put forth arguments regarding the leadership source that
may be in the most efficacious position to enact specific leadership
functions within the action and transition phases. The crossing of lead-
ership formality with the locus of leadership produces four quadrants
of possible leadership sources. In this context, formality of leadership
is defined as reflecting “whether the responsibility for team perfor-
mance is formalized in the organization (formal) or whether there is
no direct responsibility for a team's leadership and performance (infor-
mal)”(Morgeson et al., 2010, p. 8). Conversely, locus of leadership
refers to “whether the leader is a member of the team and thus
engaged in part of the team's task cycle (internal) or whether the
leader is not a member of the team and thus outside the team's day‐
to‐day activities (external)”(Morgeson et al., 2010, p. 8). Moreover,
the authors provide some exemplars of the types of leaders that might
operate in each of the four quadrants. For example, when examining
internal leaders, formal exemplars include team leaders and project
managers, whereas informal examples include shared and emergent
leaders. Looking at the external sources of leadership, examples of for-
mal leaders include sponsors, coaches, and team advisors, whereas
informal examples include mentors, champions, and executive
coordinators.
The argument for a team‐centric approach to leadership whereby
leadership capacity is built within the team also sets up the possibility
of leadership being enacted by formal or informal leaders. Within
teams that routinely face the possibility of extreme events, the impor-
tance of leadership being enacted by a mixture of formal and informal
leaders increases as leadership functions are distributed throughout
the team and leader–follower proximity is high due to the isolated
nature of the teams. Additionally, when leadership functions are
enacted by a mixture of formal and informal leadership, it capitalizes
on the synergy present within the team. Further, support for the
importance of formal and informal leadership within such environ-
ments can be seen in the literature on high‐reliability teams. High‐reli-
ability teams are those which operate in mission critical environments
where the cost of errors can be life‐threatening (Wilson, Burke, Priest,
& Salas, 2005). Within such teams, a key operating tenet is deference
to expertise whereby communication of expertise is communicated
from all levels, attention is paid to those on the front line regardless
of rank, and diversity is cultivated (Weick & Sutcliffe, 2015). The focus
on pulling expertise from all levels and a focus on those on the front
line suggest a mixture of formal and informal leadership where less
attention is placed on status and more emphasis is placed on
knowledge and ability. Although these high‐reliability teams are often
not isolated and confined, the mission critical nature of many of these
teams affords them some similarities to those teams operating in
extreme contexts.
Finally, leadership in extreme contexts will be enacted almost
exclusively by internal leaders given the team's isolated and confined
nature. This is in contrast to teams operating outside extreme environ-
ments where the operating context does not place such boundaries.
Given the above, we propose the following:
Hypothesis 5. Leadership in extreme teams that are iso-
lated and confined will be primarily internal.
Hypothesis 6. Leadership in extreme teams that are iso-
lated and confined will be a mixture of formal and
informal.
5.2 |Leadership distribution
In pulling the thread that leadership sources within team may take
many forms, Morgeson et al. (2010) highlight a gap in the current
6BURKE ET AL.

literature, whereby the primary focus has been on formally appointed
leaders to the exclusion of other forms. This, in turn, leads to calls for
future research to investigate nontraditional forms of leadership struc-
tures (e.g., moving beyond formal leadership) and the need to explore
multiple sources simultaneously (e.g., formal and informal) so as to
better understand the full leadership capacity of the team (Morgeson
et al., 2010). Examples of formal sources of leadership include team
leaders (e.g., commanders and captains), project managers, sponsors,
coaches, and team advisors (Morgeson et al., 2010). Informal sources
include shared and emergent leadership as well as mentors,
champions, and executive coordinators (Morgeson et al., 2010).
Recent work is beginning to address the notion of investigating more
nontraditional forms of leadership and has found that often the
nontraditional forms have been shown to predict not only team
performance but also often over and above traditional forms
(D'Innocenzo, Mathieu, & Kukenberger, 2016; Nicolaides et al.,
2014; Pearce & Conger, 2003; Wang, Waldman, & Zhang, 2014).
The emerging literature on shared or collective leadership is of noted
importance here. However, as with the predominant amount of work
on team leadership, the majority of this research has been done within
the context of traditional teams. Given repeated calls for the impor-
tance of context in understanding teams (Gladstein, 1984; Hackman,
2003; Mathieu, Maynard, Rapp, & Gilson, 2008), this begs the ques-
tion as to the leadership sources most instrumental in extreme teams.
The distribution of leadership throughout the team serves to build
leadership capacity within the team and allows members to serve as
redundant systems for one another extending the ways in which team
members can back one another up. In the cases of teams operating in
extreme environments, this becomes critical, as the atypical stressors
present in such environments can easily overload individual team
members. Having fellow team members who have the ability, attitudi-
nal, and cognitive capacity to engage in leadership functions can
facilitate team synergy, in essence allowing members who have the
cognitive capacity to step in and back up other members in terms of
leadership activities, as needed. Additionally, it also leverages what is
known regarding high‐reliability organizations—paying attention to
expertise wherever it may lie and realizing that members have differ-
ent skill sets and good teams learn to make the most efficacious use
of member capabilities. This is especially true when working in
extreme environments.
This argument is supported through initial evidence based on
interviews conducted with NASA subject matter experts (SMEs). In
essence, results from these interviews suggest that not only is leader-
ship a key role in long‐duration, distance‐exploration missions, but also
that leadership functions are often distributed throughout the team
and implemented through a mixture of formal and informal leadership
(Burke, Monsky, & Salas, unpublished manuscript). Additional anec-
dotal evidence for the importance of leadership being distributed
throughout the team can also be seen in other extreme contexts
(i.e., adventure racing and high‐altitude mountaineering). For instance,
in talking about the sport of adventure racing, Nagle reports that “part
of our success lies in having tremendous redundancy within our team.
So we just allow leadership to flow, hour by hour, to whoever is stron-
gest at the time”(as cited in Dahle, 1999, p. 1). Similarly, in talking
about extreme mountaineering environments, Levine (2014) stated
that “One of the most important keys to successful performance …
is empowering everyone on the team to think and act like a leader”
(p. 16). Thereby, we put forth the following hypothesis:
Hypothesis 7. In terms of structure, team leadership in
extreme teams will be enacted through a mixture of
hierarchical and shared leadership, but the latter form
of leadership will be the most prominent.
6|TEMPORAL CONSIDERATIONS AND
LEADERSHIP STRUCTURE
A final consideration in examining leadership in extreme teams
relates to the impact of temporal factors on the manner in which
leadership is enacted. Marks et al. (2001) put forth a seminal paper
on task cycles within teams, which serves to delineate the dynamic
nature of team interaction during a performance episode. Specifi-
cally, Marks et al. (2001) argue that different team processes are
important at different phases of task execution. In essence, teams
perform in “temporal cycles of goal directed activity, called
episodes”(p. 359). These episodes in turn are marked by periods
of action and transition processes. Although this notion of action
and transition phases was originally proposed in terms of teamwork
behaviors, it has since been extended to team leadership (Morgeson
et al., 2010).
Despite these advancements and repeated calls to better under-
stand temporal dynamics in teams (e.g., Harrison, Price, Gavin, &
Florey, 2002; Mohammed, Hamilton, & Lim, 2009) and the role of
team leadership, there has been limited work in this vein (see
Kozlowski et al., 2009, for an exception). Thereby, in thinking about
the sources of leadership most typically seen as being functional in
extreme teams, another related question that arises is whether partic-
ular types of leadership functions are more likely to be enacted by a
particular leadership source. For example, in thinking about transition
phase leadership functions (e.g., composing the team, defining the mis-
sion, and establishing goals/expectations), one might expect that the
formal leader is in the best position to enact these activities due to
their designated authority. Moreover, within extreme teams, where
errors are of such high consequence, an external leader may be in
the best position to facilitate transition functions due to the ability
to maintain a broader situation awareness that is sometimes
constrained within teams under pressure (Driskell et al., 1999). How-
ever, there is the additional constraint within isolated confined teams
that there may be limited reachback to an external leader; therefore,
transition functions may fall more predominantly within the realm of
an internal leader.
In contrast, when thinking about the leadership functions argued
for within the action phase of teamwork (e.g., monitor team, manage
team boundaries, challenge team, and solve problems), it may be
expected that these activities may be more frequently shared through-
out the team as dictated by the ebb and flow of task demands and
requisite leadership capabilities. Given the action‐orientated nature
of these activities and the isolated nature of many extreme teams,
we would expect that these would primarily be conducted by an inter-
nal leader. Although we have set forth informal hypotheses regarding
BURKE ET AL.7

the interaction between team leadership functions and leadership
source given the relative infancy of this piece of the literature, we still
view that these are exploratory in nature. This leads us to our final set
of hypotheses:
Hypothesis 8. Leadership occurring during the transi-
tion phase of team performance will be primarily enacted
by a single leader.
Hypothesis 9. Leadership occurring during the action
phase of team performance will be primarily distributed
throughout the team.
7|METHOD
The study of team leadership within extreme teams is challenging at
best due to the nature of such teams and the relative frequency with
which such teams exist. Therefore, as a first step to understanding
team leadership in these contexts, we employed historiometry
(Simonton, 2003). Historiometry is “that collection of methods in
which archival data concerning historic individuals and events are sub-
jected to quantitative analyses in order to test nomothetic hypotheses
about human thought, feeling, and action”(Simonton, 1998, p. 269).
Benefits of this approach include the contextual richness of the data
and the corresponding external validity. Leadership is an area that
has been particularly advantageous to explore with this approach, as
it is often well documented as a source of success or failure in histor-
ical events (e.g., DeChurch et al., 2011; Vessey, Barrett, Mumford,
Johnson, & Litwiller, 2014).
7.1 |Historiometric analysis
The first step was to identify a set of extreme environments within
which teams would be embedded and working under conditions of
isolation and confinement. This initial step returned a number of
potential contexts, but contexts were down‐selected based on the
degree to which the types of stressors present in the environments
overlapped with those typically seen in long‐duration spaceflight
(Dietz et al., 2010) as this was our original context of study. In
down‐selecting extreme environments, we worked with our partners
at NASA to decide on a final set (i.e., space exploration, polar explora-
tion, and long‐duration sailboat racing). Once an initial set of extreme
environments was decided upon, searches began for archived data
sources.
Next, searches to identify the specific archival documents that
would be used to provide data as to the nature of team leadership
within extreme environments were begun. To identify potential
sources, searches were conducted in the following data repositories:
EBSCOhost database, Google, Google Scholar, NASA data reposito-
ries, amazon.com, blogs, and communities of practice. In searching
the aforementioned databases, the following keywords were used:
Antarctic exploration, long duration sailboat racing, ocean sailboat
racing, and space exploration. These primary terms were then paired
with more narrowly defined terms, specifically: (a) Antarctic explora-
tion paired with Shackleton, Scott, Amundsen; (b) long duration/ocean
sailboat racing paired with Volvo Ocean Race, America's Cup, and
Whitbread Race; and (c) space exploration paired with Mir, Skylab,
Shuttle, ISS, Salyut, and Mars500. Later, similar keywords were used
on Google to identify potential blogs or social media sources with
respect to the more recent events (e.g., long duration sailboat races
and space exploration). In total, the complete set of searches resulted
in approximately 124 potential sources to examine (see Appendix A
for a sample).
The initial set of sources was then further narrowed according to
the following four criteria: (a) Sources must describe interdependent
interaction among the crew/team; (b) teams being described must be
operating in mission critical environments where errors are often
life‐threatening; (c) teams must be representative of intact teams
operating over a significant duration (i.e., more than a few weeks);
(d) there must be a reasonable expectation of team leadership func-
tions being present and described; and (e) source must be accessible.
Thirty‐nine sources passed this criteria and moved to the next stage
—incident extraction.
7.2 |Critical incident technique
We paired historiometry with the critical incident technique as a way
to systematically extract data from the archival sources. The critical
incident technique has been described as “a method for obtaining spe-
cific, behaviorally focused descriptions of work or other activities”
(Bownas & Bernardin, 1988, p. 1120). For the purposes of this study,
each critical incident represented an observable leadership behavior
occurring within the context of teams, described the context in which
the behavior occurred, and was linked to a specific outcome.
The third author trained 15 psychology students on the critical
incident techniques. Following training, each of the 15 coders went
through three rounds of independent coding and feedback before
they were allowed to code the source material to ensure adequate skill
and similarity in their coding technique. After these three rounds of
training, the 15 coders reviewed all source material and began to
extract critical incidents related to team leadership. This initial round
of incident extraction resulted in 311 incidents (i.e., 100 incidents
for polar exploration, 101 for long‐duration ocean races, and 110 for
space exploration). These incidents were then reviewed by the first
and second authors for quality and redundancy. In judging incidents
on quality, a focus was put on incidents that (a) could stand alone in
their interpretation, (b) were clearly described, (c) where it was appar-
ent who was engaging in the leadership action described, and (d) a
clear outcome or effect of the leadership action was described. At this
time, incidents were also checked for clarity and redundancy. Redun-
dancy involved cases where a single incident and corresponding action
was pulled from different sources. In these cases, a single description
of the incident was kept. There were also a few instances where
redundant accounts of the same interaction occurred but were
described in a conflicting manner. These incidents were dropped as
the validity of the specific descriptions was not as apparent as in cases
where multiple sources describe the same incident in the same or
similar manner. At the conclusion of this process, the final data set
included 152 incidents (i.e., 40 polar exploration incidents, 66 long
duration sailboat race incidents, and 46 spaceflight incidents). The final
8BURKE ET AL.

data set included incidents extracted from the following events: polar
expeditions (Robert F. Scott, Ernest Shackleton, Ronald Amundsen,
Thomas Musgrave, and Charles F. Hall), long‐duration sailboat racing
(Clipper Round the World Yacht Race, Sydney‐Hobart Ocean Race,
and Volvo Ocean Race), and spaceflight (Space Shuttle, Mir, Mars
500, International Space Station, and Skylab).
Once the incidents had been reviewed, they were subjected to a
series of analyses using standard card sorting procedures conducted
by two teams of three SMEs. The SME teams were composed of
Industrial/Organizational psychologists and doctoral students with
extensive experience with respect to both leadership and teams. The
first SME team independently coded each critical incident with respect
to (a) the leadership behavior represented, (b) whether the leadership
behavior was enacted by a single individual or multiple individuals (i.e.,
individual versus shared), (c) formality of leadership (i.e., informal and
formal—Morgeson et al., 2010), and (d) locus of leadership (i.e., internal
and external—Morgeson et al., 2010). Intraclass correlations (ICC2)
demonstrated high levels of agreement for ratings of leader functions
(ICC2 = .98), single versus shared leadership (ICC2 = .91), formality of
leadership (ICC2 = .88), and locus of leadership (ICC2 = .91) across all
three raters.
In conducting the initial sort, the rating team used the team lead-
ership functions argued for by Morgeson et al. (2010) as a frame of
reference. However, raters were instructed not to limit themselves
to those particular categories but to use their expertise in team lead-
ership to decide the most appropriate grouping and corresponding
label for each evidenced leadership function. The second set of SMEs
conducted a back‐translation of the incidents (sorted the incidents
into the labels), assisting to ensure data integrity. This process was
completed independently for each of the three extreme contexts—
spaceflight, polar exploration, and long‐duration sailboat races. At
the conclusion of rating process, a series of consensus meetings were
held among the three primary SMEs to resolve any area of
disagreement.
8|RESULTS
8.1 |Leadership functions
One of the primary questions of interest was the degree to which
team leadership functions as argued for within the broader literature
on team leadership would generalize to teams operating in extreme
environments. In this vein, we put forth four hypotheses regarding
those leadership functions that would be expected to be most preva-
lent in these contexts based on some of the defining features of
extreme contexts.
Results of the thematic analysis indicated evidence for nearly all
the team leadership functions that appear in one of the latest integra-
tive reviews on team leadership (i.e., Morgeson et al., 2010; see
Table 1 for descriptions). The one behavior, which did not appear in
our corpus of materials, was defining the mission. Defining the mission
has been shown to be critical outside of the context of extreme teams
as it serves to create direction for the team and begins the develop-
ment of shared mental models among members. We believe that the
absence of this function within the current corpus of materials is an
artifact of the methodology and source material as very little of the
source material gathered described events that happened with the
crew prior to being in the situation. The archived documents primarily
described extreme events and team interaction once the teams were
already in the situation. In addition, to those leadership functions
argued for within Morgeson et al. (2010), four other leadership func-
tions emerged within these extreme contexts: dominance, recognition
and utilization of expertise, leading by example, and self‐management.
However, these functions were not prominently seen and therefore
are not included in our primary analyses.
Hypotheses 1–4 described those team leadership functions within
the transition and action phases of the team's task cycle that were
believed to be among the most critical to teams operating in extreme
environments (and therefore appearing the most frequently). To exam-
ine the data in relation to the hypotheses put forth, the team leader-
ship functions that emerged from the card sort were rank ordered in
terms of their frequency of occurrence. As can be seen in Table 3,
we obtained mixed results for the hypotheses put forth regarding
the primary leadership functions. Specifically, the top five team leader-
ship functions included supporting the social climate, solving prob-
lems, structure and planning, sensemaking, and monitoring the team,
respectively. Relatively speaking, structure and planning and
sensemaking collectively accounted for 74% of the transition phase
functions seen, whereas team problem solving, supporting the social
climate, and monitoring the team collectively accounted for 70% of
the action phase functions seen. These results fully support Hypothe-
ses 1 and 4 but provide mixed support for Hypothesis 3 as challenging
the team did not appear in the top five most frequently occurring
functions. Results did not support Hypothesis 2, which argued for
training and developing the team as a critical transition behavior.
Although training and developing the team (including provision of
feedback) did appear frequently, accounting for approximately 19%
of the transition functions seen, it did not fall within the top five when
collapsing across all action and transition functions.
8.2 |Leadership distribution
A second area of interest pertains to the manner in which leadership is
manifested. Specifically, Hypotheses 7, 8, and 9 relate to the degree to
which each leadership behavior was enacted by a single individual or
distributed throughout the team such that the incident described
multiple individuals exhibiting leadership functions (i.e., shared
TABLE 3 Top five leadership functions witnessed in extreme teams
Leadership
functions
Phase of
task cycle
Rank
order
% of comments
supporting rank (%)
Support social climate Action 1 21
Solve problems
•Technical, Team
Action 2 20
Structure and plan Transition 3 15
Sensemaking Transition 4 09
Monitor team Action 5 06
BURKE ET AL.9

leadership). Results, which speak to the above hypotheses, can be
found in Table 4. Hypothesis 7 was supported in that the team
leadership functions seen within the extreme teams operating in
extreme contexts were enacted through a combination of leadership
residing within a single individual as well as leadership distributed
throughout the team.
Partial support was suggested for Hypotheses 8 and 9, which
dealt with leadership form in relation to transition and action phases
within teams' task cycles. As suggested, when looking at transition
phase leadership functions, overall there was a trend for them to be
enacted primarily through a single individual as compared with being
distributed throughout the team (56% and 44%, respectively).
However, in looking at the manner in which action phase leadership
functions were enacted overall, we see a trend for a predominant
focus on enactment by a single individual as opposed to being distrib-
uted throughout the team (54% and 46%, respectively). This is in con-
trast to Hypothesis 9, which predicted a greater emphasis on shared
leadership within the action phase as compared with a leader‐centric
approach.
To further investigate these results, the manner in which each
specific leadership function was most predominantly enacted was
examined. These results can be seen in Table 4. Within this table,
the individual leadership functions are represented in the rows,
whereas the columns represent the leadership form by which they
were enacted. For example, in looking at polar exploration, the leader-
ship function of train and develop, the team was enacted through a
leader‐centric approach (individual enactment), whereas structure
and planning was enacted through both a leader and team‐centric
approach (individual, shared). Therefore, one can easily see the break-
down for each individual leadership function.
In addition, within each extreme context (polar exploration, space
exploration, and long‐duration sailboat racing), the areas in which the
cell contains a bolded “X”indicates the predominant form of leader-
ship for that function within that context. In the case, where there
are two bolded “Xs”for a particular function within a given context,
it represents that both forms of leadership appeared equally (e.g.,
see train and develop within space context). In looking at the individ-
ual leadership functions, which comprise the higher‐order categories
of transition and action phase, it becomes apparent in collapsing
across contexts there is support for Hypotheses 8 and 9. As seen in
the column labeled “across contexts,”there are more instances of
leadership being enacted in a leader‐centric manner (by a single indi-
vidual; 80%) as compared with the action phase functions, which have
a greater number of enactments through a team‐centric approach
whereby leadership is distributed throughout the team (62%—team
centric).
8.3 |Degree of formality
Further investigating the structural aspects of team leadership was
Hypothesis 6 that predicted team leadership within extreme environ-
ments would be enacted by both formal and informal leaders, thereby
taking advantage of leadership expertise wherever it may lie. Results
begin to offer support for this hypothesis in that collapsing across con-
texts, team leadership functions were seen to be enacted predomi-
nantly through formal and informal mechanisms and to a much lesser
extent a mixture of formal and informal working together simulta-
neously (see formality across context column). Moreover, while more
exploratory in nature as no hypotheses were put forth, results begin
to suggest that, in general, those transition functions seen tend to
TABLE 4 Degree to which leadership functions are shared
Form of leadership
Antarctic Space Long‐duration sailboat Across contexts
Individual Shared Individual Shared Individual Shared Individual Shared
Transition phase Functions X
Compose/restructure team XX
Establish expect and goals XxX
Structure and plan XxXxXX
Train and develop/provide feedback XXXXxX
Sensemaking Xxx XX xX
Action phase functions X
Monitor team XX X X
Manage team boundaries XXXX
Challenge team XXXX
Perform team task XX X
Solve problems Xxx XxXX
Provide resources XXX XXX
Encourage team self‐management X‐‐‐‐ ‐‐‐‐
Support social climate Xxx XX X X
The bolded entries within each cell indicate the predominant form of each leadership functions within each of the three contexts.
Nonbolded entries within each cell indicate that the leadership function was enacted through a mixture of individual and shared leadership; the nonbolded
cell entry represents the form exhibited with lower frequency.
The across context column represents the predominant manner in which each leadership function was exhibited when collapsing across contexts.
10 BURKE ET AL.

be enacted predominantly through formal leadership (52% of the
transaction functions seen were enacted this way; see the column
entitled “formality across contexts). This is in contrast to action phase
leadership functions, which were fairly evenly split between enact-
ment by a formal leader (45.7% of those action phase functions were
enacted this way) and an informal leader (42.3%).
Finally, while exploratory in nature, Table 5 also begins to illus-
trate trends for how specific functions within the action and transition
phases were manifested in terms of formality. This can be seen by
examining the cells within the table. Specifically, each cell containing
an “x”indicates that particular leadership function being enacted
through the source represented at the top of the column. In examining
the table in this manner, it becomes apparent that several leadership
functions are consistently manifested by a variety of leadership
sources (i.e., structure and planning, problem solving, sensemaking,
and supporting the social climate) and consistently through a combina-
tion of formal, informal, and a mixture of the two leadership sources.
Other functions tend to be enacted predominantly by formal leaders
(e.g., provide resources) or vary between enactment by formal and/
or informal leaders (e.g., establish expectations/goals, train and
develop, and manage team boundaries). Very few of the leadership
functions witnessed were enacted solely through informal sources of
leadership. This represents the importance of teams being flexible in
the enactment of leadership and creating a climate where attention
is paid to expertise no matter where it may reside.
8.4 |Locus of leadership
The final set of analyses are in relation to the degree to which leader-
ship functions were enacted by someone internal to the team, external
to the team, or a mixture such that the critical incident described
leadership being shared among an internal and external member.
Supporting Hypothesis 5, 90% of the team leadership functions seen
were exhibited by a leader internal to the team. Of the functions not
enacted by someone within the team, 9% were enacted by someone
outside the immediate team, with the remaining 1% being enacted
by a combination of individuals inside and outside the immediate
team. Moreover, the predominant amount of team leadership enacted
by someone outside the immediate team occurred within the context
of spaceflight and was due to the interaction with mission control.
9|DISCUSSION
Despite the high consequences of failure within teams operating in
extreme environments and the myriad of stressors present within such
environments that may serve to make it challenging to maintain the
effective processes and states, which facilitate team performance,
these teams are difficult to study given the extreme settings within
which they operate as well as their relatively low numbers compared
with more traditional teams. The current study examined team leader-
ship within a subset of extreme teams, which operate in isolated and
confined environments through the use of archival sources that
describe team interaction, and correspondingly leadership, in the con-
text of real teams operating in such environments.
The results of this examination serve to contribute to the litera-
ture on extreme teams as well as the broader literature on teams. Con-
tributions to the literature on extreme teams include the finding that
the predominant number of team leadership functions argued for
within the functional approach to leadership are present within teams
operating in extreme environments. Moreover, there is some indirect
evidence that these functions are important in these environments
TABLE 5 Formality of leadership functions
Formality of leadership
Antarctic Space Long‐duration sailboat Formality across contexts
F I Mix F I Mix F I Mix F I Mix
Transition functions 52% 24% 24%
Compose/restructure team x x
Establish expect and goals x x
Structure and plan x x x x x x x x x
Train and develop x x x x x
Sensemaking x x x x x x x x
Action phase functions 45.7% 42.3% 12%
Monitor team x x x x
Manage team boundaries x x x
Challenge team x x
Perform team task x x
Solve problems x x x x x x x x
Provide resources x x x
Encourage team self‐management x
Support social climate x x x x x x x x
Entries within each cell illustrate the manner in which each leadership function was manifested (i.e., solely through formal leadership [F], solely informal
leadership [I], or a mixture of formal and informal simultaneously [Mix]).
Formality across contexts collapses across the three contexts to examine the degree to which leadership functions within a given phase (i.e., transition or
action) were enacted by formal leaders, informal leaders, or a mixture of formal/informal simultaneously.
BURKE ET AL.11

as the predominant number of the teams examined in this sample
would be considered effective at least in the sense that they were able
to adapt and survive the unexpected events present in each environ-
ment and to a larger degree the predominant number of these teams
would be considered successful in that they completed the primary
objectives of the team (i.e., exploring space, finishing in the long dura-
tion sailboat context, or exploring polar regions).
A second contribution to the literature on extreme teams is that
our findings suggest five leadership functions that are important and
consistently seen across the contexts examined. The functions consist
of two transition phase functions (i.e., sensemaking and structuring
and planning) and three action phase functions (i.e., supporting the
social climate, monitoring the team, and problem solving). Leaders
who engage in sensemaking and structuring and planning serve to
begin to set the building blocks for the formation, maintenance, and
adaptation of the shared mental models among team members that
will later facilitate coordinated team action during the action phases.
Furthermore, monitoring the team and problem solving are critical
functions in complex and dynamic environments as the monitoring
alerts members to potential task or psychosocial issues that may arise
as teams are working in extreme environments, isolated and confined
away from family and friends. The monitoring of team behaviors
serves as initial input in the problem solving behaviors that are needed
in dynamic environments where extreme events are often tightly
coupled a have cascading effects. Finally, supporting the social climate
becomes critical in such extreme environments due to the negative
affect and feelings of isolation that can arise in such environments,
which in turn leads to a lack of motivation. From a practical stand-
point, identification of the prevalence of these five functions may
point to a key set of leadership functions that should be trained for
teams preparing to operate in such environments. Although the five
functions could be trained, they could also be used during selection
to comprise a team where members already have this capacity as a
way to build redundancy within the team.
A third contribution to the literature on extreme teams is also a
contribution to the broader literature on team leadership. Specifically,
the investigation of the structural aspects of leadership in extreme
teams (i.e., leadership distribution, formality, and locus). Taking a func-
tional perspective to leadership, we moved past solely focusing on a
leader‐centric view of leadership to one, which represents a combina-
tion of team and leader‐centric approaches. This, in turn, opens up
several possibilities for investigating the manner in which leadership
is enacted as well as the source from which leadership arises. In this
vein, we found evidence for leadership functions being enacted
through a combination of leadership distributed throughout the team
as well as leadership being enacted by a single individual. Furthermore,
results suggested that leadership in extreme teams was enacted by
both formal and informal leaders as well as formal and informal leaders
working together to enact leadership. This later result is reminiscent of
work on high reliability organizations but brought down to the team
level. Both of the above results suggest that in comprising teams to
work within such extreme environments attention needs to be paid
to selecting individuals who can alternate between leadership and
followership. In addition, formal leaders should be taught to promote
a climate where the source of leadership is not limited to high status
individuals but allowed to flow from all members on the team. In this
way serving to build the leadership capacity within the team.
A final contribution of our study deals with temporal consider-
ations. Although the team literature has long called for more attention
being paid to time, the predominant amount of work in teams exam-
ines ad‐hoc teams of short duration. The teams that were the primary
focus of our investigation were not ad‐hoc teams but primarily intact
teams operating together over long periods of time. Moreover, we
were able to begin to empirically examine how the different temporal
phases of the team's task cycle may impact structural characteristics
associated with team leadership. These results suggest that the nature
of team leadership during action and transition phases may be mani-
fested differently. Finally, although no predictions were made to this
point in the current investigation, findings begin to point to the impor-
tance of examining the specific functions that comprise transition and
action phases. For example, results suggested that certain leadership
functions were most often enacted through a mixture of formal, infor-
mal, and joint mechanisms, whereas other functions were most typi-
cally enacted through formal mechanisms. Differences were also
seen in the degree to which specific functions tended to be enacted
in a leader or team‐centric approach. Future work should more specif-
ically investigate this aspect as it begins to suggest that it is neither a
team nor leader‐centric approach that is more appropriate, but that
the efficacy of one or the other approach will vary based on the par-
ticular leadership function.
9.1 |Limitations
Although the approach taken in examining the hypotheses posed in
this study offers many benefits for understanding team leadership in
extreme teams, as with all studies, it also has its limitations. For exam-
ple, although the examination of archival accounts of historical events
provides a wealth of contextually rich information about teams oper-
ating in real contexts, it does not facilitate an understanding of the
relationship of identified leadership functions or structural characteris-
tics to their corresponding impact on team processes and emergent
states. Future research should empirically explore some of these key
relationships. Second, the fact that the source documents from which
the critical incidents were extracted were not written with our
research questions in mind is both a strength and a weakness, whereas
the fact that the sources were not written with our specific research
question in mind limits the degree of bias with respect to our specific
research question. However, it does not negate the possibility that the
individual's accounts of the events that are being described are biased.
Although attempts were made to minimize this type of bias by
collecting information from multiple sources, within the extreme envi-
ronments examined the archived perspectives were often limited.
Third, although we would argue that the extreme contexts that we
examined were more similar than different, they were purposely cho-
sen to vary along some dimensions in an attempt to cover a range of
extreme teams in isolated, confined environments. Although their var-
iation provides a greater degree of confidence in those results, which
generalize across contexts, it is possible that some of the results from
the polar exploration sample may in part be an artifact of the time in
history that the events occurred as these are much earlier in time than
12 BURKE ET AL.

either of the other two contexts. This is probably most relevant for the
analyses that deal with the distribution of leadership. A final limitation
is that the predominant amount of our source documents describe
team interaction in situ thereby potentially limiting the presence of
some transition functions that may only occur at initial team forma-
tion. Our hope is that this research generates additional interest into
the functional leadership functions within extreme teams and how
the structural characteristics associated with team leadership may vary
based on temporal needs of the team (i.e., action and transition).
9.2 |Future research
The results herein begin to provide an initial glimpse into the functions
and structural characteristics of team leadership as seen in extreme
teams. There are several threads that could be investigated by future
research. First, and perhaps foremost, future research is needed to
unpack the relationship of specific leadership functions to the team
processes and emergent states that have been argued to facilitate
performance in mission critical environments, especially those that
serve to facilitate the socioemotional health of the team. Although
the current research began to highlight how the nature of team lead-
ership may vary based on temporal factors by focusing on leadership
functions as evidenced within the transition and action phases of
teamwork, future research can further pull this thread. Specifically,
how do temporal factors such as the stage of team development or
stage of mission impact the types of leadership functions and struc-
tures, which are important? Of note is that the primary number of
the sources that were identified for analysis within the current study
dealt with team leadership once the team's had already begun their
mission. Therefore, it was not possible to examine differential impor-
tance of the various leadership functions or structural characteristics
at different points in time—with respect to either the team's life cycle
or stage of mission. It is our hope that the findings presented here and
the many new questions that emerge will serve to spur future research
in this area.
ACKNOWLEDGEMENT
This work was supported by a NASA grant (NNX14AK54G) to Dr.
Shawn Burke, Principal Investigator, Dr. Eduardo Salas and Dr. Marissa
Shuffler, Co‐Principal Investigators. The views expressed in this work
are those of the authors and do not necessarily reflect the organiza-
tions with which they are affiliated or their sponsoring institutions.
ORCID
C. Shawn Burke http://orcid.org/0000-0003-3403-8405
Marissa L. Shuffler http://orcid.org/0000-0002-6292-8950
Christopher W. Wiese http://orcid.org/0000-0002-3932-4008
REFERENCES
Bell, S. T., Fisher, D. M., Brown, S. G., & Mann, K. E. (2016). An approach
for conducting actionable research with extreme teams. Journal of
Management,20,1–26.
Bownas, D. A., & Bernardin, H. J. (1988). Critical incident technique. In S.
Gael (Ed.), The job analysis handbook for business, industry and govern-
ment (Vol. II) (pp. 1120–1137). New York: John Wiley and Sons.
Burke, C. S., Monsky, D., & Salas, E. (2017). Moving towards an under-
standing of team roles in long duration, distance exploration missions.
Unpublished manuscript.
Burke, C. S., Stagl, K. C., Klein, C., Goodwin, G., Salas, E., & Halpin, S. M.
(2006). What types of leadership behaviors are functional in teams?
A meta‐analysis. Leadership Quarterly,17, 288–307.
Cannon‐Bowers, J. A., Tannenbaum, S. I., Salas, E., & Volpe, C. E. (1995).
Defining team competencies and establishing team training require-
ments. In R. Guzzo, E. Salas, & Associates (Eds.), Team effectiveness
and decision making in organizations (pp. 333–380). San Francisco, CA:
Jossey‐Bass.
Carson, J. B., Tesluk, P. E., & Marrone, J. A. (2007). Shared leadership in
teams: An investigation of antecedent conditions and performance.
Academy of Management Journal,50, 1217–1234.
Crosby, R. (2008). An analysis of leadership behavior in extreme military
contexts. Unpublished master's thesis. Naval PostGraduate School;
Monterey, California
Dahle, C. (1999). Xtreme teams. Forbes. Retrieved from http://www.
fastcompany.com/38070/xtreme‐teams
DeChurch, L. A., Burke, C. S., Shuffler, M. L., Lyons, R., Doty, D., & Salas, E.
(2011). A historiometric analysis of leadership in mission critical
multiteam environments. The Leadership Quarterly,22, 152–169.
DeChurch, L. A., Hiller, N. J., Murase, T., Doty, D., & Salas, E. (2010).
Leadership across levels: Levels of leaders and their levels of impact.
The Leadership Quarterly,21(6), 1069–1085.
Dietz, A. S., Weaver, S. J., Bedwell, W. L., Sierra, M. J., Salas, E., Smith‐
Jentsch, K., & Fiore, S. M. (2010, March). Stress in long‐duration space-
flight teams: Exploring the complex network of critical antecedents
impacting team processes and performance. Unpublished white paper.
D'Innocenzo, L., Mathieu, J. E., & Kukenberger, M. R. (2016). A meta‐
analysis of different forms of shared leadership–team performance
relations. Journal of Management,42(7), 1964–1991.
Driskell, J. E., & Salas, E. (1991). Group decision making under stress.
Journal of Applied Psychology,76(3), 473–478.
Driskell, J. E., Salas, E., & Johnston, J. (1999). Does stress lead to a loss of
team perspective? Group Dynamics: Theory, Research, and Practice,3(4),
291–302.
Ensley, M. D., Hmieleski, K. M., & Pearce, C. L. (2006). The importance of
vertical and shared leadership within new venture top management
teams: Implications for the performance of startups. Leadership Quar-
terly,17, 217–231.
Fleishman, E. A., Mumford, M. D., Zaccaro, S. J., Levin, K. Y., Korotkin, A. L.,
& Hein, M. B. (1991). Taxonomic efforts in the description of leader
behavior: A synthesis and functional interpretation. Leadership
Quarterly,2(4), 245–287.
Friedrich, T. L., Vessey, W. B., Schuelke, M. J., Ruark, G. A., & Mumford, M.
D. (2009). A framework for understanding collective leadership: The
selective utilization of leader and team expertise within networks.
The Leadership Quarterly,20(6), 933–958.
Gladstein, D. (1984). Groups in context: A model of task group effective-
ness. Administrative Science Quarterly,29, 499–517.
Hackman, J. R. (2002). Leading teams: Setting the stage for great perfor-
mances. Boston: Harvard Business School Press.
Hackman, J. R. (2003). Learning more by crossing levels: Evidence from
airplanes, hospitals, and orchestras. Journal of Organizational Behavior,
24, 905–922.
Hannah, S. T., Campbell, D. J., & Matthews, M. D. (2010). Advancing a
research agenda for leadership in dangerous contexts. Military
Psychology,22(S1), S157–S189.
Hannah, S. T., Uhl‐Bien, M., Avolio, B. J., & Cavarretta, F. L. (2009). A
framework for examining leadership in extreme contexts. The
Leadership Quarterly,20(6), 897–919.
Harrison, D. A., Price, K. H., Gavin, J. H., & Florey, A. T. (2002). Time,
teams, and task performance: Changing effects of surface and deep
BURKE ET AL.13

level diversity on group functioning. Academy of Management Journal,
45(5), 1029–1045.
Janis, I. L. (1982). Groupthink. Boston, MA: Houghton Mifflin.
Keeton, K. E., Schmidt, L. L., Slack, K. J., & Malka, A. A. (2012). The rocket
science of teams. Industrial and Organizational Psychology,5(1), 32–35.
Kozlowski, S. W. J., Gully, S. M., Salas, E., & Cannon‐Bowers, J. A. (1996).
Team leadership and development: Theory, principles, and guidelines
for training leaders and teams. In M. M. Beyerlein, D. A. Johnson, &
S. T. Beyerlein (Eds.), Advances in interdisciplinary studies of work teams:
Team leadership (Vol. 3) (pp. 253–291). Bingley, UK: Emerald Group
Publishing Limited.
Kozlowski, S. W. J., Watola, D. J., Jensen, J. M., Kim, B. H., & Botero, I. C.
(2009). Developing adaptive teams: A theory of dynamic team
leadership. In E. Salas, G. F. Goodwin, & C. S. Burke (Eds.), Team
effectiveness in complex organizations: Cross‐disciplinary perspectives
and approaches (pp. 113–156). Mahwah, NJ: LEA.
Levine, A. (2014). On the edge: Leadership lessons from Mount Everest and
other extreme environments. New York, NY: Grand Central Publishing.
Manz, C. C., & Sims, H. P. (1980). Self‐management as a substitute for lead-
ership: A social learning theory perspective. Academy of Management
Review,5(3), 361–367.
Marks, M. A., Mathieu, J. E., & Zaccaro, S. J. (2001). A temporally based
framework and taxonomy of team processes. Academy of Management
Review,26(3), 356–376.
Mathieu, J., Maynard, M. T., Rapp, T., & Gilson, L. (2008). Team effective-
ness 1997–2007: A review of recent advancements and a glimpse
into the future. Journal of Management,34(3), 410–476.
McGrath, J. E. (1962). Leadership behavior: Requirements for leadership
training. Prepared for U.S. Civil Service Commission Office of Career
Development, Washington, D.C.
Mohammed, S., Hamilton, K., & Lim, A. (2009). The incorporation of time in
team research: Past, current, and future. In E. Salas, G. F. Goodwin, &
C. S. Burke (Eds.), Team effectiveness in complex organizations:
Cross‐disciplinary perspectives and approaches (pp. 321–348). Mahwah,
NJ: Erlbaum.
Morgeson, F. P., DeRue, D. S., & Karam, E. P. (2010). Leadership in teams: A
functional approach to understanding leadership structures and
processes. Journal of Management,36,5–39.
Mulhearn, T., McIntosh, T., Gibson, C., Mumford, M. D., Yammarino, F. J.,
Connelly, S., …Vessey, B. (2016). Leadership for long‐duration space
missions: A shift toward a collective approach. Acta Astronautica,129,
466–476.
Nicolaides, V. C., LaPort, K. A., Chen, T. R., Tomassetti, A. J., Weis, E. J.,
Zaccaro, S. J., & Cortina, J. M. (2014). The shared leadership of teams:
A meta‐analysis of proximal, distal, and moderating relationships. The
Leadership Quarterly,25(5), 923–942.
Pearce, C. L., & Conger, J. A. (2003). Shared leadership: Reframing the hows
and whys of leadership. Thousand Oaks, CA: Sage Publications.
Salas, E., Sims, D., & Burke, C. S. (2005). Is there a ‘Big 5’in teamwork?
Small Group Research,36(5), 555–599.
Simonton, D. (1998). A history of European women's work: 1700 to the
present (p. 269). US: Taylor & Francis.
Simonton, D. K. (2003). Qualitative and quantitative analyses of historical
data. Annual Review of Psychology,54(1), 617–640.
Suedfeld, P. (2012). 2012 behavioral health and performance standing
review panel research Plan Review. Retrieved from https://
humanresearchroadmap.nasa.gov/reviews/BHP%20SRP%
20Research%20Plan%20Review%20(2012).pdf
Vessey, W. B. (2014). Multiteam systems in the spaceflight context:
Current and future challenges. In M. L. Shuffler, R. Rico, & E. Salas
(Eds.), Pushing the boundaries: Multiteam systems in research and practice
(Vol. 16) Research on Managing Groups and Teams (pp. 135–153).
London: Emerald Publishing.
Vessey, W. B., Barrett, J. D., Mumford, M. D., Johnson, G., & Litwiller, B.
(2014). Leadership of highly creative people in highly creative fields:
A historiometric study of scientific leaders. The Leadership Quarterly,
25(4), 672–691.
Wang, D., Waldman, D. A., & Zhang, Z. (2014). A meta‐analysis of shared
leadership and team effectiveness. Journal of Applied Psychology,
99(2), 181–198.
Weick, K., & Sutcliffe, K. (2015). Managing the unexpected: Assuring high
performance in an age of uncertainty. San Francisco, CA: Jossey‐Bass.
Wilson, K. A., Burke, C. S., Priest, H. A., & Salas, E. (2005). Promoting health
care safety through training high reliability teams. Quality and Safety in
Health Care,14, 303–309.
Yukl, G. (2006). Leadership in organizations (6th ed.). Upper Saddle River,
NJ: Prentice Hall.
Zaccaro, S. J., & DeChurch, L. A. (2012). Leadership forms and functions in
multiteam systems. In S. J. Zaccaro, M. A. Marks, & L. A. DeChurch
(Eds.), Multiteam systems: An organization form for dynamic and complex
environments (pp. 253–288). New York, NY: Routledge.
Zaccaro, S. J., Heinen, B., & Shuffler, M. (2009). Team leadership and team
effectiveness. In E. Salas, G. F. Goodwin, & C. S. Burke (Eds.), Team
effectiveness in complex organizations: Cross‐disciplinary perspectives
and approaches (pp. 83–112). New York, NY: Psychology Press.
Zaccaro, S. J., Rittman, A. L., & Marks, M. A. (2001). Team leadership. The
Leadership Quarterly,12, 451–483.
C. Shawn Burke is a Professor (Research) at the Institute for
Simulation and Training, University of Central Florida. Her exper-
tise includes teams, team leadership, team adaptability, team train-
ing, and effectiveness with an emphasis on teams operating in
complex environments. Her work has been funded by NASA,
ARI, NSF, and ONR.
Marissa L. Shuffler is an assistant professor of Industrial/Organi-
zational Psychology at Clemson University with over 10 years of
experience conducting basic and applied research. Dr. Shuffler's
areas of expertise include team and leader training and develop-
ment, multiteam systems, communication, and adaptation, with
an emphasis on high‐risk and complex environments.
Christopher W. Wiese is an assistant professor of Industrial/Orga-
nizational Psychology at Georgia Institute of Technology. His
research investigates how events at work impact well‐being and
how this translates to performance outcomes. In particular, he
seeks to understand the temporal dynamics of well‐being. To fully
understand this phenomenon, Dr. Wiese conducts research at
individual and team levels.
How to cite this article: Burke CS, Shuffler ML, Wiese CW.
Examining the behavioral and structural characteristics of team
leadership in extreme environments. J Organ Behav. 2018;1–15.
https://doi.org/10.1002/job.2290
14 BURKE ET AL.

APPENDIX A
Sample of archival sources reviewed.
Document title Source/Author
Blogs/electronic sources
Astronauts find a coolant on the space station* http://www.universetoday.com/102046/astronauts‐find‐a‐coolant‐leak‐on‐the‐space‐station/
Atkinson
Space station loses contact with NASA mission control* http://www.space.com/19853‐space‐station‐contact‐lost‐nasa.html/ Kramer
Big station can have big malfunctions* Retrieved from http://www.space.com/8867‐big‐space‐station‐big‐malfunctions.html/ Malik
Mars crew guinea pigs suffered insomnia, lethargy http://news.discovery.com/space/mars500‐crewexperiment‐insomnia‐health‐effects‐130116.
htm/ Klotz
http://www.astrobio.net/topic/exploration/moon‐to‐mars/lessons‐from‐mars‐500/
Spacecraft: Manned: Mir: Close calls* http://www.russianspaceweb.com/mir_close_calls.html
Luca Parmitano's blog http://blogs.esa.int/luca‐parmitano/
RED threshold late notice conjunction threat misses
ISS—crew egress
http://www.nasaspaceflight.com/2009/03/threat‐to‐iss‐crew‐soyuz/ Bergin
Russia identifies cause of rocket launch failure http://www.space.com/12779‐russian‐rocket‐failure.html/ Moskowitz
Rocket liftoff aborted a half‐second before launch http://www.npr.org/blogs/thetwo‐way/2012/05/19/153061648/in‐historic‐space‐mission‐
launch‐is‐only‐the‐first‐test/ Farrington
ESA's participation in Mars500* http://www.esa.int/Our_Activities/Human_Spaceflight/Mars500
Mars500 Mission Diary (Video & Text)* http://www.esa.int/Our_Activities/Human_Spaceflight/Mars500/%20Mars500_diary
Books
Dragonfly: NASA and the crisis aboard the MIR* Harper Collins/Burroughs
The worst journey in the world* Picador/Cherry‐Garrad
Homesteading space: The Skylab story* University of Nebraska Press/Garriott, Hitt, Kerwin
Off the planet: Surviving five perilous months aboard the space station Mir* McGraw Hill Professional/Linenger
Shuttle‐Mir: The United States and Russia share history's highest stage* Houston: NASA/Morgan
Team spirit: Life and leadership on one of the world's toughest yacht races* Adlard Coles Nautical/Hall
Ocean warriors: The thrilling story of the 2001/2002 Volvo ocean race around the world* HarperCollins Publishers, Inc./Mundle
One watch at a time: Around the world with Drum on the Whitbread race* W. W. Norton & Company, Inc./Novak
Global challenge: Leadership lessons from “the world's toughest yacht race”* The Book Guild Ltd./Walters, Mackie, Mackie, & Bacon
Fighting finish: The Volvo ocean race, race around the world 2001–2002 Nomad Communications/Jobson
Island of the lost: Shipwrecked at the edge of the world* Algonquin Books/Druett
Fatal North: Adventure survival aboard USS Polaris: 1
st
U.S. expedition North Pole* Signet Book/Henderson
Race for the South Pole: The expedition diaries of Scott and Amundsen* A & C Black/Huntford
Endurance: Shackleton's incredible voyage* McGraw‐Hill/Lansing
Shackleton's way: Leadership lessons from the great Antarctic explorer* Penguin Putnam, Inc./Morell & Capparell
Leading at the edge: Leadership lessons from the extraordinary
saga of Shackleton's Antarctic exploration*
Amacon/Perkins
Journals: Captain Scott's last expedition* OUP Oxford/Scott & Jones
The heart of the Antarctic: Being the story of the British Antarctic expedition 1907–1909* Carroll & Graf Publishers, Inc./Shackleton
Bold endeavors: Lessons from polar and space exploration* Naval Institute Press/Stuster
Scott, Shackleton, and Amundsen: Ambition and tragedy in the Antarctic* Thomson
Ice: Stories of survival from polar exploration* Thunder's Mouth Press, Balliett & Fitzgerald, Inc./Willis
Shackleton's boat journey: The narrative from the Captain of the Endurance* W & J Mackay Limited, Chatham/Worsley
*Sources actually used in the final analysis.
BURKE ET AL.15