httphfs.sagepub.comErgonomics Societyof the Human F.docx

http://hfs.sagepub.com/
Ergonomics Society
of the Human Factors and
Human Factors: The Journal
http://hfs.sagepub.com/content/52/2/295
The online version of this article can be found at:
DOI: 10.1177/0018720810371689
published online 23 July 2010
2010 52: 295 originallyHuman Factors: The Journal of the
Human Factors and Ergonomics Society
Jamie C. Gorman, Nancy J. Cooke and Polemnia G. Amazeen
Training Adaptive Teams
Published by:
http://www.sagepublications.com
On behalf of:
Human Factors and Ergonomics Society
can be found at:Society
Human Factors: The Journal of the Human Factors and

ErgonomicsAdditional services and information for
http://hfs.sagepub.com/cgi/alertsEmail Alerts:
http://hfs.sagepub.com/subscriptionsSubscriptions:
http://www.sagepub.com/journalsReprints.navReprints:
http://www.sagepub.com/journalsPermissions.navPermissions:
http://hfs.sagepub.com/content/52/2/295.refs.htmlCitations:
at LIBERTY UNIV LIBRARY on August 17,
2012hfs.sagepub.comDownloaded from
http://hfs.sagepub.com/content/52/2/295
http://www.hfes.org
http://hfs.sagepub.com/cgi/alerts
http://hfs.sagepub.com/subscriptions
http://www.sagepub.com/journalsReprints.nav
http://www.sagepub.com/journalsPermissions.nav
http://hfs.sagepub.com/content/52/2/295.refs.html
What is This?

- Jul 23, 2010 OnlineFirst Version of Record
- Sep 14, 2010Version of Record >>
http://hfs.sagepub.com/content/52/2/295.full.pdf
http://hfs.sagepub.com/content/early/2010/07/21/001872081037
1689.full.pdf
http://online.sagepub.com/site/sphelp/vorhelp.xhtml
Training Adaptive Teams
Jamie C. Gorman and Nancy J. Cooke, Arizona State
University–Polytechnic, Mesa,
Arizona, and Polemnia G. Amazeen, Arizona State University,
Tempe, Arizona
Objective: We report an experiment in which three training
approaches are compared
with the goal of training adaptive teams. Background: Cross-
training is an established
method in which team members are trained with the goal of
building shared knowledge.
Perturbation training is a new method in which team
interactions are constrained to pro-
vide new coordination experiences during task acquisition.
These two approaches, and a
more traditional procedural approach, are compared. Method:
Assigned to three training
conditions were 26 teams. Teams flew nine simulated
uninhabited air vehicle missions;

three were critical tests of the team’s ability to adapt to novel
situations. Team perfor-
mance, response time to novel events, and shared knowledge
were measured. Results:
Perturbation-trained teams significantly outperformed teams in
the other conditions in
two out of three critical test missions. Cross-training resulted in
significant increases in
shared teamwork knowledge and highest mean performance in
one critical test.
Procedural training led to the least adaptive teams. Conclusion:
Perturbation training
allows teams to match coordination variability during training
to demands for coordina-
tion variability during posttraining performance. Although
cross-training has adaptive
benefits, it is suggested that process-oriented approaches, such
as perturbation training,
can lead to more adaptive teams. Application: Perturbation
training is amenable to sim-
ulation-based training, where perturbations provide interaction
experiences that teams
can transfer to novel, real-world situations.
Address correspondence to Jamie C. Gorman, Cognitive
Engineering Research Institute, 5810 S. Sossaman Rd.,
Ste. 106, Mesa, AZ 85212; [email protected] HUMAN
FACTORS, Vol. 52, No. 2, April 2010, pp. 295–307.
DOI: 10.1177/0018720810371689. Copyright © 2010, Human
Factors and Ergonomics Society.
INTRODUCTION
In settings ranging from business and manu-
facturing to military and medical operations,
there are many tasks that are too cognitively

demanding to be performed by individuals
working alone. An example is a surgical task,
which requires a set of highly trained individuals,
including two surgeons, an anesthesiologist
and two nurses, each of whom brings different
cognitive capabilities to the team. But it is not
enough to bring together a set of highly trained
individuals. To function as a team, individuals
must coordinate their activities. Adaptive teams
have the ability to coordinate their activities not
only under routine conditions but also under
novel conditions for which they have not been
explicitly trained. Adaptation is the altering of
structure in accordance with changes in the
environment. Because they have the ability to
change their interactions to match the changing
demands of the environment, adaptive teams
can perform at a high level under novel task
conditions.
A number of relatively recent tragic system
failures can be at least partially attributed to
poor coordination of a team-level response to
environmental uncertainty. System failures attrib-
utable to poor team skills at Three Mile Island
and Chernobyl (Gaddy & Wachtel, 1992), social
pathogens behind the 1986 launch decision of
the space shuttle Challenger (Vaughan, 1996),
and lack of communication in the Operation
Provide Comfort friendly fire incident (Gorman,
Cooke, & Winner, 2006; Snook, 2002) each
implicate, in different ways, deficiencies in inter-
action and coordination that result in a failure to
adapt to changes in the task environment. These
incidents, and others like them, highlight the

need for training that addresses limitations and
SPECIAL ISSUE
296 April 2010 - Human Factors
deficiencies at the team level in responding to
novel patterns of events and threats.
Approaches to Training Adaptive Teams
A challenging problem for training team
cognition (i.e., training teams to decide, plan,
think, and act as an integrated unit; Cooke,
Gorman, & Winner, 2007) is how to balance
training for high performance under routine
task conditions with training to adapt to novel
task demands (Marks, Zaccaro, & Mathieu,
2000; Stachowski, Kaplan, & Waller, 2009).
These training goals can be approached with
varying theoretical motives. In this article, we
report an experiment in which three training
approaches, each with a different underlying
theoretical motive, were investigated with the
goal of training teams that perform at a high
level under novel task conditions. The training
approaches include cross-training, procedural
training, and perturbation training.
Cross-training. In cross-training, team mem-

bers are trained on each other’s roles and respon-
sibilities (e.g., Blickensderfer, Cannon-Bowers,
& Salas, 1998). Cross-training is theoretically
aligned with the idea that team cognition is the
shared knowledge of the team members and is
found widely in the team training literature (see
Salas et al., 2008, and Salas, Nichols, & Driskell,
2007, for recent meta-analyses). The goal of
cross-training is the development of shared, or
interpositional, knowledge (Cannon-Bowers,
Salas, Blickensderfer, & Bowers, 1998; Cooke
et al., 2003; Volpe, Cannon-Bowers, Salas, &
Spector, 1996). Positional clarification (receiv-
ing information on other roles), positional mod-
eling (observing other roles), and positional
rotation (firsthand experience performing dif-
ferent roles) (Blickensderfer et al., 1998), which
are types of cross-training, have been effective
in the development of shared knowledge, ulti-
mately improving coordination and team perfor-
mance (e.g., Marks, Sabella, Burke, & Zacarro,
2002). Cross-training has a firm empirical
grounding and a record of success in the team
training literature, making it a good point of com-
parison with the other approaches in this study.
One of the potential benefits of cross-training
for shared knowledge is a high level of team
performance under stress (high workload, time
pressure). Drawing on shared knowledge, team
members anticipate each other’s needs to com-
municate efficiently, or coordinate implicitly,
under stress (Cannon-Bowers et al., 1998; Entin
& Serfaty, 1999; Stout, Cannon-Bowers, Salas, &
Milanovich, 1999). It is thought that shared

expectations, resulting from the development of
shared knowledge, allow team members to gen-
erate predictions for appropriate behavior under
novel conditions, enabling them to quickly adapt
to the changing demands of the task environ-
ment (Fiore, Salas, & Cannon-Bowers, 2001). A
possible drawback of a shared set of expecta-
tions, however, is the habituation of team member
interaction, which could result in dysfunctional
consequences if the situation is highly novel
(e.g., Gersick & Hackman, 1990; Gorman,
Cooke, & Winner, 2006).
Whereas cross-training is feasible for rela-
tively small, homogeneously skilled teams, it
can become impractical as teams grow in diver-
sity and size. For example, it would be impracti-
cal to cross-train the surgeon and nurse positions
of an emergency room team (Cooke et al., 2003;
Marks et al., 2002). Also, cross-training may
negatively impact individual-level performance
due to the demands of training for multiple team
member roles (Cannon-Bowers et al., 1998),
which is problematic as teams grow in size.
Although there are adaptive benefits of cross-
training, there are practical limitations to its
applicability.
Procedural training. Procedural training is a
form of process training in which operators in
complex systems are positively reinforced (through
feedback) to follow a standard sequence of actions
(a procedure) each time a particular stimulus is
encountered. The assumption behind procedural
training is that if the procedure is always fol-
lowed, then errors resulting from human inter-

action will be reduced and performance will be
enhanced, particularly under conditions of stress
and high workload (e.g., Hockey, Sauer, &
Wastell, 2007; Sauer, Burkholter, Kluge,
Ritzmann, & Schuler, 2008). Procedural train-
ing is widely used in aviation, military, medical,
manufacturing, and business settings, in which
deviations from complicated procedures can be
catastrophic. The prevalence of procedural train-
ing for coordination in highly critical team tasks
Training adapTive Teams 297
(e.g., emergency response; Ford & Schmidt,
2000; Stachowski et al., 2009) make it a good
point of comparison for the other training meth-
ods in this study.
Procedural training is compatible with the
concept of “overlearning”: continuation of prac-
tice beyond mastery that leads to automatic
responding. Drilling a standard team interaction
pattern, for a specific class of event, over the
entirety of training can lead to an automatic
response that a team can rely on under stress.
The goal of procedural training, as operational-
ized in the current study, is to overlearn a team
coordination procedure. Ideally, due to over-
learning, procedurally trained teams perform
under stress by automatically (reflexively)

reacting with an a priori coordinated response.
Procedural training does not impose the
practical limitations of cross-training but may
limit a team’s ability to transfer training to novel
situations. Similar to the concept of a “set effect”
(Luchins, 1942), procedural training may set teams
up to coordinate in a routine fashion under a
novel condition. We argue, therefore, that like
habituation (Gersick & Hackman, 1990), rigid
proceedure-following during task acquisition
can lead to poor performance when posttraining
conditions do not match training conditions.
Perturbation training. Perturbation training
is a form of process training introduced in this
study. Adopted from the dynamic systems liter-
ature, a perturbation is an extrinsic application
of force that briefly disrupts a dynamic process,
forcing the reacquisition of a new stable trajec-
tory, and is typically used to probe the stability
of that process (Gorman, Amazeen, & Cooke,
in press). The concept of perturbation can be
applied to team training by disrupting standard
coordination procedures multiple times during
task acquisition, forcing teams to coordinate in
novel ways to achieve their objective. Unlike
training in which the situation or objectives are
varied (e.g., training for low- vs. high-frequency
circumstances), in perturbation training, critical
coordination links are disrupted while the team
objective remains constant. The goal of pertur-
bation training is to counteract habituation and
procedural rigidity associated with team interac-
tions—possible outcomes of cross-training and
procedural training, respectively—allowing teams

to acquire flexible interaction processes that
will transfer to novel task conditions.
Perturbation training is theoretically inspired
by findings in the motor- and verbal-learning
literatures that suggest that introducing difficul-
ties for the learner, such as practice condition
variability, facilitates performance under novel
posttraining conditions (Schmidt & Bjork,
1992). Perturbation training thus shares some
features of motor schema theory (Schmidt,
1975) and desirable difficulties (Bjork, 1994)
but for coordination and for teams. According
to motor schema theory, varying the conditions
of practice during motor skill acquisition
enhances the “rules” that relate movements to
external task demands. In verbal learning, desir-
able difficulties are unpredictable and variable
conditions of practice that cause difficulty for
the learner but ultimately enhance the transfer
of concepts to new contexts. Whereas those
approaches employ equally probable but ran-
domly varying training conditions to introduce
practice condition variability, perturbation
training employs abrupt but focused disruptions
to team coordination.
Bjork (1994) argued that varying practice
conditions exercises more elaborate encoding and
retrieval processes needed in the posttraining
environment. Perturbation training extends this
idea to team process: When coordination is per-
turbed, all team member interactions (not just
those directly affected by the perturbation) must
readjust to accommodate the perturbation in

such a manner that the team objective is never-
theless met (Turvey, 1990). We suggest that simi-
lar to the effects of practice condition variability,
perturbations exercise the team processes needed
to adapt in the posttraining environment.
A major limitation of perturbation training is
that it has not previously been applied and its effec-
tiveness is unknown. An experiment described by
Gorman, Cooke, Pedersen, et al. (2006) provided
some empirical grounding for perturbation
training. Teams were initially trained and per-
formed a repetitive command-and-control task
during a 3-hr experimental session. Participants
returned for a second session after a retention
interval, after which they were either intact
(kept the same team members) or mixed (same
role on the team but different team members).
As expected, intact teams outperformed mixed
teams under routine conditions, but the effect
was short-lived. Mixed teams, however, per-
formed better on tests of situation awareness,
had higher process ratings (Gorman, Cooke,
Pedersen, et al., 2006), and had more flexible
coordination dynamics (Gorman et al., in press).
Those benefits were not attributable to increased
shared knowledge or procedural rigidity but to

increased variation in interaction experience.
In some sense, mixing up the team members
perturbed rigid coordination patterns, ultimately
leading to a more flexible and adaptive team.
Perturbation training, as operationalized in the
current study, does not involve mixing team
members but was designed have the same effect.
By disrupting standard coordination procedures
during task acquisition, perturbation training
increases interaction experience in intact
command-and-control teams.
The Current Study
We compared the three different training
approaches in an uninhabited air vehicle (UAV)
simulator with the goal of producing teams that
perform at a high level under novel task condi-
tions and that respond rapidly to novel events. In
the UAV task, three team members (navigator,
photographer, and pilot) coordinate to take pic-
tures of stationary ground targets. Three training
protocols were developed for cross-training,
perturbation training, and procedural training of
UAV teams. The following hypotheses are
based on prior results and existing literature.
Hypothesis 1: By focusing on introducing varied inter-
action experiences during task acquisition, perturba-
tion training will result in performance scores and
response times to novel events that are as good as or
better than cross-training and superior to procedural
training.
Hypothesis 2: Because of its focus on training team

members to know each other’s roles and responsi-
bilities, cross-training will result in higher levels of
shared knowledge compared with both procedural
and perturbation training.
Hypothesis 3: By training teams to rigidly follow a
procedure, procedural training will result in the
least adaptive teams (i.e., poor performance and
slow response to novel events) compared with both
perturbation and cross-training.
METHOD
Participants
We recruited 32 three-person teams (96 par-
ticipants) for participation from Mesa, Arizona,
and surrounding areas. The team members had
no prior experience working together. Participants
ranged in age from 18 to 54 (M = 28), and
71 were male. The experiment occurred during
two 3- to 4-hr sessions. Because of scheduling
conflicts for Session 2, a total of 26 teams
(78 participants) completed both experimental
sessions. Participants were paid $10 per hour,
and each member of the highest-performing
team received a $100 bonus.
Materials and Apparatus
The experiment was conducted in a UAV
synthetic task environment (UAV-STE) for
teams (Cooke & Shope, 2005). Each of the
three team members was seated at a workstation
in front of three computer monitors with a key-
board and a mouse. To interact, team members

wore aviation-quality headsets and communi-
cated by holding down push-to-talk buttons.
The workstations were located in the same room,
configured in a U shape with team members
backs to each other. With the team members
donning headsets, the UAV-STE did not afford
face-to-face interaction.
The team’s task was to take reconnaissance
photographs of stationary ground targets during
a series of nine 40-min missions divided across
two experimental sessions. There were 11 to
12 targets per mission except for one high-
workload mission that had 20 targets. The three
team member roles—navigator, photographer,
and pilot—were each associated with different,
yet interdependent tasks, information resources,
and needs.
Measures
Team performance. Performance was mea-
sured for each UAV-STE mission as the weighted
composite of several team-level mission param-
eters, including number of missed targets, time to
process targets, and time spent with unaddressed
warnings and alarms. Cooke, Gorman, Pedersen,
et al. (2007) report the parameter weights, which
were established in previous experiments to
maximize score sensitivity. Teams started each

mission with a performance score of 1,000, and
points were subtracted on the basis of final values
of the mission parameters. This team perfor-
mance score has been validated against other
measures of team process and performance
(Cooke, Gorman, Duran, & Taylor, 2007; Cooke,
Gorman, Pedersen, et al., 2007).
Response time to novel events. Novel events
were introduced within UAV missions by intro-
ducing roadblocks. Roadblocks are novel
changes in the task environment that have to be
jointly recognized by two or more team members
who take action to overcome them (e.g., a new
target is introduced, equipment fails, an enemy
threat appears; Cooke, Gorman, & Rowe, 2009;
Gorman, Cooke, & Winner, 2006). Time to over-
come roadblocks, defined as the time from the
initiation of the roadblock to the time that action
is taken that overcomes the roadblock, was the
measure of response time to novel events.
Interpositional taskwork knowledge. This
measure assessed a team’s average knowledge
of the taskwork associated with the other two
roles. To measure taskwork knowledge, related-
ness ratings (1 = completely unrelated to 5 =
completely related) were elicited for 55 pairs of
concepts from the UAV task (e.g., airspeed,
altitude). Individual team member ratings
were analyzed using the Pathfinder algorithm
(Schvaneveldt, 1990), which translates related-
ness ratings across pairs of concepts into a graph-

ical network representation of conceptual
interrelatedness.
Individual networks were compared with
expert role referent networks. The referents were
derived empirically from the top five individual
performers at each role in previous UAV-STE
experiments (Cooke, Gorman, Pedersen, et al.,
2007). Each team member was scored against the
other two role referents on the basis of the pro-
portion of shared links (0 = no similarity to 1 =
exactly similar). Team-level interpositional task-
work knowledge was taken as the average of
these two scores across each of the three team
members. Scores closer to 1 indicated a higher
level of interpositional taskwork knowledge
across team members.
Interpositional teamwork knowledge. This
measure assessed a team’s average knowledge of
the teamwork associated with the other two roles.
Interpositional teamwork knowledge was elicited
with the use of a questionnaire that consisted of 16
items related to which communications were nec-
essary to achieve a given scenario goal (e.g., “For a
priority target, must the photographer communi-
cate camera settings to the navigator, the pilot, or
both?”). Items that were necessary had to be indi-
cated by individual team members using check
marks. To calculate teamwork knowledge, indi-
vidual responses were compared with role-specific
answer keys that were generated by experimenters
familiar with the task, and points were awarded
for correct answers (Cooke, Gorman, Pedersen,
et al., 2007).

To measure interpositional teamwork knowl-
edge, each team member was scored on the basis
of the proportion correct relative to the answer
key for each of the other two roles. Interpositional
teamwork knowledge was calculated as the aver-
age number of these two scores across the three
team members. Scores closer to 1 indicated a
higher level of interpositional teamwork knowl-
edge across team members.
Procedure
When participants arrived for the first session,
they were randomly assigned to a team member
role and the team was assigned to one of the
three training conditions. Participants received
approximately 1 hr 45 min of training via three
PowerPoint training modules and a hands-on
training mission. The first two PowerPoint
modules were identical for all training condi-
tions and covered the general task and interface.
The third module and the hands-on training mis-
sion differed on the basis of training condition.
(Procedures for each training condition are
described in the following section.)
Teams then completed Missions 1 through 5.
Knowledge measures were taken after Mission
1. Missions 2 through 4 were condition-specific
training missions (Table 1). The first roadblock
was introduced during Mission 5, the first post-
training mission. The roadblock consisted of
cutting communication from the navigator to
the pilot for 5 min, during which teams had
to reroute navigator-to-pilot communications

through the photographer to overcome the
roadblock. Mission 5 was the first of three criti-
cal missions that tested the teams’ ability to per-
form under novel conditions. The completion of
Mission 5 concluded the first session.
Teams returned after 8 to 10 weeks for the
second session. All participants received refresher
training on the software interfaces, after which
teams completed four additional UAV-STE mis-
sions. Roadblocks were introduced during each
mission. Mission 6 was the second critical mis-
sion, which tested retention of team skill after
the break and included the second roadblock
(a disguised target was hidden on the navigator
map and photographer target list; teams had to
recognize and photograph the target to over-
come the roadblock). Knowledge was measured
for a second time after Mission 6 as a test of
knowledge retention.
Teams then completed their final three mis-
sions (Missions 7 through 9). Mission 9 was the
high-workload mission, in which the rate of tar-
gets per minute was almost doubled from 0.28
to 0.5 and teams were exposed to a two-part
roadblock (communication channel cut for 5 min
from pilot to navigator and from navigator to

pilot; teams had to reroute their communica-
tions through open channels to overcome the
roadblock). Mission 9 was the third of the three
critical test missions.
Training Procedure
Cross-training. For the third PowerPoint train-
ing module, team members in the cross-training
condition received training on the other two roles
(positional clarification). Teams then completed a
short training mission followed by approximately
15 min of hands-on experience performing all
team member roles (positional rotation). After
Missions 2 through 4, teams in the cross-training
condition were prompted to discuss how well they
performed and to plan for the next mission.
Procedural training. For the third
PowerPoint training module, teams in the
procedural training condition received train-
ing on the standard UAV-STE target photo-
graphing procedure: (a) The navigator provides
target information to the pilot, (b) the pilot
and photographer negotiate altitude and air-
speed for that target, and (c) the photogra-
pher provides feedback on the status of the
target photograph (Figure 1). Teams then
completed a short training mission followed
by approximately 15 min of hands-on train-
ing using the target photographing proce-
dure. After Missions 2 through 4, teams in the
procedural condition received experimenter
feedback on deviations from the standard
procedure. During training, team members
in the procedural condition were provided

with a hard copy of the target photographing
procedure.
Perturbation training. Teams in the pertur-
bation training condition received filler
PowerPoint training on the history and current
uses of UAVs. Teams then completed a short
training mission followed by approximately
15 min of communications system testing in
which they identified the source of static in the
UAV-STE communication system (i.e., which
push-to-talk button was emanating static).
This training exercise provided experience on
the use of multiple communication paths.
During Missions 2 through 4, teams in the per-
turbation condition received perturbations to
the target-photographing procedure (Table 2)
as they attempted to photograph targets.
Perturbations were less general than roadblocks
and forced teams to adjust specific interactions
TABLE 1: Experimental Procedure
Session 1 Session 2
Initial participant training Refresher training
Mission 1 Mission 6 (retention test and second roadblock)C
Knowledge measures Knowledge measures
Mission 2T Mission 7
Mission 3T Mission 8
Mission 4T Mission 9 (high workload and two-part roadblock)C
Mission 5 (first roadblock)C Debriefing
Note. T = condition-specific training mission; C = critical test

mission.
relative to the information-negotiation-feedback
procedure (Figure 1).
RESULTS
Of the 26 teams that completed the experi-
ment, there were 10 teams in the procedural
condition and 8 teams each in the cross-train-
ing and perturbation conditions. Previous
experiments in the UAV-STE exhibited low
between-subjects power with a = .05 (M = .11,
SD = .05) on tests of team performance due to
small sample size. To increase statistical
power, a significance level of a = .10 was
used. For planned critical test mission com-
parisons, two conditions were pooled to form a
comparison against a single condition. These
planned comparisons also served to increase
power.
Team Performance
Team performance results are summarized in
Table 3 and graphed in Figure 2. Team perfor-
mance was analyzed using a 3 (training) × 9
(mission) mixed ANOVA. The main effect of

mission was significant, F(8, 184) = 22.14, p <
.001, MSE = 3535.25, η2 = .53. No other effects
in the omnibus test were significant. A repeated
contrast on mission revealed that performance
increased significantly during initial performance
acquisition until Mission 4. There was a signifi-
cant drop in performance at Mission 6, after the
retention interval. Performance then improved
significantly as teams reacquired the task, until
Mission 8. Task reacquisition was followed by a
significant drop in performance at Mission 9, the
high-workload mission. These results reinforce
Missions 5, 6, and 9 as critical missions.
Figure 1. Standard photographing procedure for the uninhabited
air vehicle synthetic task environment.
TABLE 2: Perturbations to the Standard Uninhabited Air
Vehicle Synthetic Task Environment Target-
Photographing Procedure Used for Perturbation Training During
Missions 2 Through 4
Link in the
Procedure Perturbation Method of Introducing Perturbation
When Introduced
Information Photographer must
provide target
information to pilot
Experimenter calls in new target restrictions
to photographer and disables camera until
restrictions are communicated to pilot
Once in Mission 2
Once in Mission 3

Twice in Mission 4
Negotiation Navigator/pilot
must negotiate
airspeed/altitude
Experimenter calls in new airspeed/altitude
of current target to navigator
Once in Mission 2
Once in Mission 3
Twice in Mission 4
Feedback Photographer does
not provide feedback
to navigator and pilot
Experimenter calls in status of target photo
to navigator and pilot and cuts all
photographer communications
Once in Mission 2
Twice in Mission 3
Twice in Mission 4
Planned comparisons were performed for
each of the three critical test missions to address
which training condition resulted in the highest

performance under novel conditions. As shown
in Figure 2, perturbation-trained teams exhib-
ited better mean performance in two of the three
critical test missions (Missions 5 and 9),
whereas cross-trained teams exhibited better
mean performance in one of the critical test
missions (Mission 6). Performance of the
perturbation-trained teams at Mission 5 (M =
500.37, SD = 50.93) was significantly better than
the other two conditions (M = 459.31, SD =
54.91), F(1, 24) = 3.23, p = .085, MSE = 2892.21,
η2 = .12. Performance of perturbation-trained
teams at Mission 9 (M = 442.24, SD = 36.83) was
also significantly better than the other two condi-
tions (M = 379.62, SD = 60.00), F(1, 24) = 7.37,
p = .012, MSE = 2945.49, η2 = .24. The cross-
training performance advantage at Mission 6 was
not significant, F(1, 24) = .76, p = .392, MSE =
8379.75, η2 = .03.
TABLE 3: Mean Team Performance by Training Condition
Mission Cross-Trained Perturbation Procedural
1 345.04 (65.80) 342.78 (54.23) 316.92 (78.88)
2 383.18 (72.89) 409.33 (80.94) 373.90 (65.65)
3 422.58 (74.39) 463.39 (80.69) 439.92 (54.71)
4 450.54 (77.71) 483.76 (59.83) 447.83 (54.26)
5 446.40 (64.41) 500.37 (50.93)* 469.63 (46.92)
6 435.99 (54.48) 380.30 (166.10) 383.76 (100.91)
7 477.79 (77.32) 471.77 (75.38) 421.38 (86.88)
8 513.25 (70.94) 547.06 (47.86) 502.47 (58.60)
9 389.13 (76.39) 442.24 (36.83)* 372.02 (46.00)
Note. Standard deviations in parentheses.

*p < .10.
300
350
400
450
500
550
600
1 3 4 5 6 7 8 92
Mission
T
e
a
m
P
e
rf
o
rm
a
n

c
e
Cross-Training
Procedural
Perturbation
Acquisition
Reacquisition
Critical Test 1
First Roadblock
Critical Test 2
Retention
Critical Test 3
High Workload
Figure 2. Team performance for each training condition across
missions. Adapted from Gorman et al. (2007).
Response Time to Novel Events

Time to overcome roadblocks was analyzed
with a 3 (training) × 3 (critical mission) mixed
ANOVA. One observation was missing from the
cross-training condition. There was a significant
main effect of mission, F(1.53, 33.59) = 119.82,
p < .001, MSE = 16018.63, η2 = .85 (Greenhouse-
Geisser correction used). The significant mission
effect was attributed to differences in the diffi-
culty of roadblocks. Therefore, no further analy-
ses were performed to isolate that effect. The
main effect of training was also significant, F(2,
22) = 3.53, p = .047, MSE = 19460.72, η2 = .24
(Figure 3).
The planned comparisons at the critical mis-
sions revealed that procedural-trained teams
were significantly slower to overcome road-
blocks (M = 218.70, SD = 94.34) than were
teams in the other two conditions (M = 146.67,
SD = 103.52) at Mission 5, F(1, 23) = 3.11, p =
.091, MSE = 10005.89, η2 = .12. Teams with
procedural training were also significantly
slower (M = 656.00, SD = 191.01) than those in
the other two conditions (M = 533.07, SD =
98.35) at Mission 6, F(1, 23) = 4.50, p = .045,
MSE = 20164.71, η2 = .16. The same compari-
son at Mission 9 was not significant. The
Training × Critical Mission interaction was not
significant. Analysis of the measure for time to
overcome roadblock in the noncritical missions
(i.e., Missions 7 and 8) did not reveal any sig-
nificant differences.
Interpositional Knowledge
Interpositional teamwork and taskwork

knowledge results are summarized in Table 4.
Interpositional teamwork and taskwork knowl-
edge were separately analyzed with 3 (training) ×
2 (session) mixed ANOVAs. The taskwork
ANOVA did not reveal any significant differ-
ences. There was a significant interaction effect
for interpositional teamwork knowledge,
F(2, 23) = 2.70, p = .089, MSE = .01, η2 = .19
(Figure 4). Pooled comparisons revealed that
cross-training (M = .87, SD = .07) led to signifi-
cantly higher interpositional teamwork knowl-
edge compared with the other two conditions
(M = .78, SD = .08) at Session 2, F(1, 24) = 7.04,
p = .014, MSE = .01, η2 = .23. The same compari-
son for Session 1 was not significant. The main
effect of session for interpositional teamwork
knowledge was also significant, F(1, 23) = 6.24,
p = .02, MSE = .01, η2 = .21. Although teams in
all training conditions exhibited some increase in
interpositional teamwork knowledge across ses-
sions, teams in the cross-training condition
exhibited a significantly greater increase.
DISCUSSION
Perturbation-trained teams significantly out-
performed teams in the other conditions in two
out of three critical test missions, and their
response times to overcome novel roadblock
events were roughly equivalent to cross-trained
teams. These results lend support to our first
hypothesis that perturbation training leads to
high performance under novel conditions.
The results suggest that something similar to
the effects of practice condition variability

(Schmidt & Bjork, 1992) contributed to transfer
at the team level: Perturbation training allowed
teams to generalize performance to novel con-
ditions by forcing the teams to coordinate in
new ways during task acquisition. However,
whereas practice condition variability provides
a range of task conditions for the individual
learner, perturbation training induced coordina-
tion variability across team members during
repetitions of the same task. By training teams
to formulate and test new solutions to the prob-
lem of coordinating ground targets during task
acquisition, perturbation training actively engaged
0
100
200
300
400
500
600
700
800
5 (First Roadblock)
Mission

T
im
e
t
o
O
v
e
rc
o
m
e
R
o
a
d
b
lo
c
k
(
s
e
c

o
n
d
s
)
Cross-Trained
Perturbation
Procedural
9 (High Workload)6 (Retention)
Figure 3. Time to overcome roadblocks by training
condition across the critical missions (error bars repre-
sent 90% confidence intervals).
team processes that are needed to adapt to
novel, but related, coordination problems in the
posttraining environment. Because perturbation
training builds on prior novelty, it may also
have allowed teams to develop within a rich
experiential learning environment (Kolb, 1984).
Our second hypothesis was that cross-training
would result in the highest levels of shared
knowledge but that this would not necessarily

result in the best performance under novel task
conditions or the fastest response times to novel
events. Support for that hypothesis was mixed.
Cross-training resulted in greater shared team-
work knowledge in the second session but not
in the first session. This is not a surprise given
that the majority of condition-specific training
took place after the first knowledge measure-
ment session. However, shared taskwork
knowledge did not change across experimental
sessions, regardless of training condition. This
may suggest a ceiling effect, such that task-
related concept relatedness (e.g., the association
between airspeed and altitude) does not change
after initial participant training.
Cross-training also resulted in highest mean
performance at one of the critical missions (the
retention test) and faster response times for
overcoming roadblocks, although those differ-
ences were not significant. It is possible that
with a larger sample size, or a less variable task
environment, cross-training would have resulted
in significant advantages. Marks et al. (2000) found
that development of a shared mental model pre-
dicted performance under novel task conditions
better than under routine task conditions. The
retention and roadblock tests are novel condi-
tions unique to our experiment, however, and
further empirical work is needed to better under-
stand the benefits of cross-training and shared
knowledge under these conditions.
The results support our third hypothesis that
procedural training should result in the least

adaptive teams. Procedural training is arguably
the most prevalent form of training for coordi-
nating highly critical team tasks, but its utility
for training adaptive teams has been increas-
ingly called into question (e.g., Ford & Schmidt,
2000; Grote, Kolbe, Zala-Mezö, Bienefeld-Seall,
& Künzle, 2010; Stachowski et al., 2009).
Procedural training need not be limited to a
single, standardized coordination process; assum-
ing that the space of possible future events is
finite, procedures can be scripted for a variety
of foreseeable contingencies. Nevertheless,
given the current results, we argue that teams
trained to automatically follow a standardized
coordination procedure become rigid and slow
to adapt to novel changes in highly dynamic
task environments. Whereas proceduralization
may be good for unchanging and foreseeable
aspects of a task, in training adaptive teams,
TABLE 4: Mean Interpositional Teamwork and Taskwork
Knowledge by Training Condition
Measure Cross-Trained Perturbation Procedural
Teamwork
Session 1 .73 (.12) .78 (.12) .74 (.11)
Session 2 .87 (.07)* .79 (.06) .77 (.09)
Taskwork
Session 1 .47 (.04) .46 (.05) .48 (.05)
Session 2 .47 (.01) .48 (.04) .48 (.03)
Note. Standard deviations in parentheses.
*p < .10.

0.5
0.6
0.7
0.8
0.9
1
Session 1
In
te
rp
o
s
it
io
n
a
l
T
e
a
m
w

o
rk
K
n
o
w
le
d
g
e Cross-Trained
Perturbation
Procedural
Session 2
Figure 4. Interpositional teamwork knowledge by
training condition across knowledge measurement
sessions (error bars represent 90% confidence intervals).
there should be a match between interaction
variability and the changing dynamics of the
task environment.

We examined team performance under three
critical situations: adaptation to a novel event
(roadblock), after a retention interval, and under
high workload. This is not an exhaustive list of
possibilities. There are many forms of adaptation
for which a team could be trained (e.g., role struc-
ture adaptation; Lepine, 2005). The current results
are intuitively plausible, however, given the nature
of mechanisms of team adaptation currently found
in the adaptive team literature, and extend the
idea of process-based adaptability training. The
building and maintenance of shared mental
models are thought to support team adaptation
(Burke, Stagl, Salas, Pierce, & Kendall, 2006;
Stout et al., 1999; Waller, Gupta, & Giambatista,
2004). Indeed, cross-training was successful in
building shared teamwork knowledge, and cross-
trained teams exhibited potential for high per-
formance under novel conditions. Parallel to the
motivation for perturbation training, however,
teams directly adapt via flexible interaction pro-
cesses (Gorman et al., in press; Manser, Harrison,
Gaba, & Howard, 2009; Stachowski et al.,
2009; Waller, 1999). Kozlowski, Gully, Nason,
and Smith (1999) suggested that teams adapt by
selecting an appropriate form of interaction from
a preexisting repertoire or by creating a new
form. Approaches like perturbation training
have the potential to broaden a team’s interac-
tion repertoire not by prescribing preexisting
forms of coordination but by allowing teams to
exercise bottom-up organization of new coordi-
nation links.
What are the practical implications for train-
ing adaptive teams, and how can we apply

perturbation training? Simulation-based team
training (Dorsey et al., 2009) would allow for
the design of perturbations that focus on specific
events, times, or interactions (Gorman, Cooke, &
Duran, 2009). Simulation-based training can
emphasize physical (equipment) fidelity or cog-
nitive fidelity (how well the simulation exercises
psychological processes required for that task;
Goettle, Ashworth, & Chaiken, 2007). For per-
turbation training, cognitive fidelity should be
emphasized in order to exercise the team
interaction processes needed for the real-world
task (Bowers & Jentsch, 2001). Another con-
cern is the specifics of introducing perturbations:
when, how many, what kind, and how often?
Simulation-based training would be the ideal
venue for perturbation training, and although
approaches such as crew resource management
(see Salas, Wilson, Burke, & Wightman, 2006)
may use simulators to train for rare or novel
events, our results suggest that more thought
and research should go into identifying the
types of team interaction experiences needed
and the ideal timing of those experiences.
What are the implications of the varying the-
oretical training motives—shared knowledge,
proceduralization, flexible interactions—for
team cognition? Prevalent in the team cognition
literature is a distinction between knowledge
and process and which contributes most to team
effectiveness (e.g., Cooke, Gorman, & Winner,
2007). We submit that cross-training most
directly impacts knowledge, that perturbation
training most directly impacts process, and that

procedural training may have little impact on
either. The current study is not an unequivocal
test of knowledge versus process accounts
of team cognition, nor is it an exhaustive sam-
pling of variations on procedural, perturbation,
or cross-training in a variety of contexts.
Nonetheless, the results do suggest that training
focused on process may contribute something
to team effectiveness that a knowledge-focused
approach does not.
CONCLUSION
The details of team adaptation are not speci-
fied at the outset of a novel event. The details
accrue gradually, during the process of adapta-
tion, and there lies the problem for training
adaptive teams: They must be able to decide,
plan, think, and act under conditions never expe-
rienced. Adaptation is the altering of structure
in accordance with environmental change and,
under many circumstances, is not a purely top-
down, knowledge-driven process. Teams should
be provided opportunities to exercise adaptive
competency using not only top-down (knowledge-
focused) training but also bottom-up (process-
oriented) training. Perturbing coordination as
team members interact is one means of eliciting
the bottom-up, process-oriented flexibility that

teams need in order to adapt. Future research
should continue to explore mechanisms of flex-
ible team interaction and how teams use them to
adapt to the pressures of highly dynamic,
high-stakes work environments.
ACKNOWLEDGMENTS
This research was funded by Air Force Office
of Scientific Research Grant FA9550-04-1-0234
and Air Force Research Laboratory Grant
FA8650-04-6442; additional support came from
National Science Foundation Grant BCS 0447039.
The authors would like to thank the following
individuals who contributed to this research:
Dee Andrews, Christy Caballero, Olena Connor,
Jasmine Duran, Preston Kiekel, Harry Pedersen,
Steven Shope, Amanda Taylor, and Jennifer
Winner. Some team performance results were
previously reported in a technical report (Cooke,
Gorman, Pederson, et al., 2007) and a confer-
ence proceeding paper (Gorman et al., 2007).
REFERENCES
Bjork, R. A. (1994). Memory and metamemory considerations in
the training of human beings. In J. Metcalfe & A. Shimamura
(Eds.), Metacognition: Knowing about knowing (pp. 185–205).
Cambridge, MA: MIT Press.
Blickensderfer, E., Cannon-Bowers, J. A., & Salas, E. (1998).
Cross-training and team performance. In J. A. Cannon-Bowers
& E. Salas (Eds.), Making decisions under stress: Implications
for individual and team training (pp. 299–311). Washington,

DC:
American Psychological Association.
Bowers, C. A. & Jentsch, F. (2001). Use of commercial, off-the-
shelf, simulations for team research. In E. Salas (Ed.),
Advances
in human performance and cognitive engineering research
(Vol. 1., pp. 293–317). Amsterdam, Netherlands: Elsevier
Science.
Burke, C. S., Stagl, K. C., Salas, E., Pierce, L., & Kendall, D.
(2006). Understanding team adaptation: A conceptual analysis
and model. Journal of Applied Psychology, 91, 1189–1207.
Cannon-Bowers, J. A., Salas, E., Blickensderfer, E., & Bowers,
C. A.
(1998). The impact of cross-training and workload on team
functioning: A replication and extension of initial findings.
Human Factors, 40, 92–101.
Cannon-Bowers, J. A., Salas, E., & Converse, S. (1993). Shared
mental models in expert team decision making. In N. J.
Castellan
(Ed.), Individual and group decision making (pp. 221–246).
Hillsdale, NJ: Erlbaum.
Cooke, N. J., Gorman, J. C., Duran, J. L., & Taylor, A. R.
(2007).
Team cognition in experienced command-and-control teams.
Journal of Experimental Psychology: Applied, 13, 146–157.
Cooke, N. J., Gorman, J. C., Pedersen, H. K., Winner, J., Duran,
J.,
Taylor, A., Amazeen P. G., Andrews, D., & Rowe, L. (2007).
Acquisition and retention of team coordination in command-
and-control (Technical Report for AFOSR Grant FA9550-04-

1-0234 and AFRL Award No. FA8650-04-6442). Washington,
DC: Department of Defense.
Cooke, N. J., Gorman, J. C., & Rowe, L. J. (2009). An
ecological
perspective on team cognition. In E. Salas, J. Goodwin, &
C. S. Burke (Eds.), Team effectiveness in complex
organizations:
Cross-disciplinary perspectives and approaches (pp. 157–182).
New York, NY: Taylor and Francis.
Cooke, N. J., Gorman, J. C., & Winner, J. L. (2007). Team
cogni-
tion. In F. Durso, R. Nickerson, S. Dumais, S. Lewandowsky,
& T. Perfect (Eds.), Handbook of applied cognition (2nd ed.,
pp. 239–268). New York, NY: Wiley.
Cooke, N. J., Kiekel, P. A., Salas, E., Stout, R. J., Bowers, C.,
& Cannon-Bowers, J. (2003). Measuring team knowledge:
A window to the cognitive underpinnings of team perfor-
mance. Group Dynamics: Theory, Research and Practice, 7,
179–199.
Cooke, N. J., & Shope, S. M. (2005). Synthetic task
environments
for teams: CERTT’s UAV-STE. In N. Stanton, A. Hedge, K.
Brookhuis, E. Salas & H. Hendrick (Eds.), Handbook of
Human Factors and Ergonomics Methods (pp. 46-41-46-46).
Boca Raton, FL: CRC Press.
Dorsey, D., Russell, S., Keil, C., Campbell, G., Van Buskirk,
W., &
Schuck, P. (2009). Measuring teams in action: Automated per-
formance measurement and feedback in simulation-based
training. In E. Salas, G. F. Goodwin, & C. S. Burke (Eds.),
Team

effectiveness in complex organizations: Cross-disciplinary
perspectives and approaches (pp. 351–381). New York, NY:
Taylor and Francis.
Entin, E. E. & Serfaty, D. (1999). Adaptive team coordination.
Human Factors, 41, 312-325.
Fiore, S. M., Salas, E., & Cannon-Bowers, J. A. (2001). Group
dynamics and shared mental model development. In M. London
(Ed.), How people evaluate others in organizations (pp. 309–
336). Mahwah, NJ: Erlbaum.
Ford, J. K., & Schmidt, A. M. (2000). Emergency response
training:
Strategies for enhancing real-world performance. Journal of
Hazardous Materials, 75, 195–215.
Gaddy, C. D., & Wachtel, J. A. (1992). Team skills training in
nuclear power plant operations. In R. W. Swezey & E. Salas
(Eds.), Teams: Their training and performance (pp. 379–396).
Norwood, NJ: Ablex.
Gersick, C. J. G., & Hackman, J. R. (1990). Habitual routines in
task-performing groups. Organizational Behavior and Human
Decision Processes, 47, 65–97.
Goettle, B. P., Ashworth, A. R. S., III, & Chaiken, S. R.
(2007). Advanced distributed learning for team training in
command and control applications. In S. M. Fiore & E. Salas
(Eds.), Toward a science of distributed learning (pp. 93–117).
Washington, DC: American Psychological Association.
Gorman, J. C., Amazeen, P. G., & Cooke, N. J. (in press). Team
coordination dynamics. In Nonlinear dynamics, psychology
and life sciences.

Gorman, J. C., Cooke, N. J., Amazeen, P. G., Winner, J. L.,
Duran, J. L., Pedersen, H. K., & Taylor, A. R. (2007).
Knowledge training versus process training: The effects of
training protocol on team coordination and performance. In
Proceedings of the Human Factors and Ergonomics Society
50th Annual Meeting (pp. 382–387). Santa Monica, CA:
Human Factors and Ergonomics Society.
Gorman, J. C., Cooke, N. J., & Duran, J. L. (2009).
Development
of simulated team environments for measuring team cognition
and performance. In D. Schmorrow, J. Cohen, & D. Nicholson
(Eds.), PSI handbook of virtual environments for training and
education (pp. 347–361). Westport, CT: Praeger Security
International.
Gorman, J. C., Cooke, N. J., Pedersen, H. K., Winner, J. L.,
Andrews, D., & Amazeen, P. G. (2006). Changes in team com-
position after a break: Building adaptive command-and-control
teams. In Proceedings of the Human Factors and Ergonomics
Society 49th Annual Meeting (pp. 487–491). Santa Monica,
CA: Human Factors and Ergonomics Society.
Gorman, J. C., Cooke, N. J., & Winner, J. L. (2006). Measuring
team situation awareness in decentralized command and con-
trol environments. Ergonomics, 49, 1312–1325.
Grote, G., Kolbe, M., Zala-Mezö, E., Bienefeld-Seall, N., &

Künzle, B. (2010). Adaptive coordination and heedfulness
make better cockpit crews. Ergonomics, 53, 211–228.
Hockey, G. R. J., Sauer, J., & Wastell, D. G. (2007).
Adaptability of
training in simulated process control: Knowledge- versus rule-
based guidance under task changes and environmental stress.
Human Factors, 49, 158–174.
Kolb, D. A. (1984). Experiential learning: Experience as the
source of learning and development. Englewood Cliffs, NJ:
Prentice Hall.
Kozlowski, S. W. J., Gully, S. M., Nason, E. R., & Smith, E. M.
(1999). Developing adaptive teams: A theory of compilation
and performance across levels and time. In D. R. Ilgen &
E. D. Pulakos (Eds.), The changing nature of work and perfor-
mance: Implications for staffing, personnel actions, and devel-
opment (pp. 240–292). San Francisco, CA: Jossey-Bass.
Lepine, J. A. (2005). Adaptation of teams in response to
unforeseen
change: Effects of goal difficulty and team composition in
terms of cognitive ability and goal orientation. Journal of
Applied Psychology, 90, 1153–1167.
Luchins, A. S. (1942). Mechanization in problem solving.
Psychological Monographs, 54(248).
Manser, T., Harrison, T. K., Gaba, D. M., & Howard, S. K.
(2009).
Coordination patterns related to high clinical performance in a
simulated anesthetic crisis. International Anesthesia Research
Society, 108, 1606–1615.
Marks, M. A., Sabella, M. J., Burke, C. S., & Zaccaro, S. J.

(2002).
The impact of cross-training on team effectiveness. Journal of
Marks, M. A., Zaccaro, S. J., & Mathieu, J. E. (2000).
Performance
implications of leader briefings and team-interaction training
for team adaptation to novel environments. Journal of Applied
Psychology, 85, 971–986.
Sauer, J., Burkolter, D., Kluge, A., Ritzmann, S., & Schuler, K.
(2008). The effects of heuristic rule training on operator
performance in a simulated process control environment.
Ergonomics, 51, 953–967.
Salas, E., DiazGranados, D., Klein, C., Burke, C. S., Stagl, K.
C.,
Goodwin, G. F., & Halpin, S. M. (2008). Does team training
improve team performance? A meta-analysis. Human Factors,
50, 903–933.
Salas, E., Nichols, D., & Driskell, J. E. (2007). Testing three
team
training strategies in intact teams: A meta-analysis. Small
Group Research, 38, 471–488.
Salas, E., Wilson, K. A., Burke, C. S., & Wightman, D. C.
(2006).
Does crew resource management training work? An update, an
extension, and some critical needs. Human Factors, 48,
392–412.
Schmidt, R. A. (1975). A schema theory of discrete motor skill
learning. Psychological Review, 82, 225–260.
Schmidt, R. A., & Bjork, R. A. (1992). New conceptualizations

of
practice: Common principles in three paradigms suggest new
concepts for training. Psychological Science, 3, 207–217.
Schvaneveldt, R. W. (1990). Pathfinder associative networks:
Studies in knowledge organization. Norwood, NJ: Ablex.
Snook, S. A. (2002). Friendly fire: The accidental shootdown of
U.S. black hawks over northern Iraq. Princeton, NJ: Princeton
University Press.
Stachowski, A. A., Kaplan, S. A., & Waller, M. J. (2009). The
ben-
efits of flexible team interaction during crises. Journal of
Stout, R. J., Cannon-Bowers, J. A., Salas, E., & Milanovich, D.
M.
(1999). Planning, shared mental models, and coordinated per-
formance: An empirical link is established. Human Factors,
41, 61–71.
Turvey, M. T. (1990). Coordination. American Psychologist, 45,
938–953.
Vaughan, D. (1996). The Challenger launch decision: Risky
tech-
nology, culture, and deviance at NASA. Chicago, IL: University
of Chicago Press.
Volpe, C. E., Cannon-Bowers, J. A., Salas, E., & Spector, P. E.
(1996). The impact of cross-training on team functioning: An
empirical investigation. Human Factors, 38, 87-100.
Waller, M. J. (1999). The timing of adaptive group responses to
nonroutine events. Academy of Management Journal, 42,

127–137.
Waller, M. J., Gupta, N., & Giambatista, R. C. (2004). Effects
of
adaptive behaviors and shared mental models on control crew
performance. Management Science, 50, 1534–1544.
Jamie C. Gorman received his PhD in cognitive
psychology from New Mexico State University,
Las Cruces, in 2006 and is a postdoctoral research
associate at Arizona State University–Polytechnic
and the Cognitive Engineering Research Institute in
Mesa, Arizona. Beginning August 2010, he will be
an assistant professor in the Human Factors Program
at Texas Tech University.
Nancy J. Cooke received her PhD in cognitive
psychology from New Mexico State University,
Las Cruces, in 1987 and is a professor in applied
psychology at Arizona State University–Polytechnic
and science director of the Cognitive Engineering
Research Institute in Mesa, Arizona.
Polemnia G. Amazeen received her PhD in experi-
mental psychology from the University of Connecticut,
Storrs, in 1996 and is an associate professor in psy-
chology at Arizona State University, Tempe, and a fac-
ulty research associate at the Cognitive Engineering
Research Institute in Mesa, Arizona.
Date received: May 13, 2009
Date accepted: March 29, 2010

http://sgr.sagepub.com/
Small Group Research
http://sgr.sagepub.com/content/40/2/181
The online version of this article can be found at:
DOI: 10.1177/1046496408328821
2009 40: 181 originally published online 6 January 2009Small
Group Research
Burke, Rebecca Lyons and Gerald F. Goodwin
Cameron Klein, Deborah DiazGranados, Eduardo Salas, Huy Le,
C. Shawn
Does Team Building Work?
Published by:
can be found at:Small Group ResearchAdditional services and
information for
http://sgr.sagepub.com/cgi/alertsEmail Alerts:

http://sgr.sagepub.com/subscriptionsSubscriptions:
http://www.sagepub.com/journalsReprints.navReprints:
http://www.sagepub.com/journalsPermissions.navPermissions:
http://sgr.sagepub.com/content/40/2/181.refs.htmlCitations:
What is This?
- Jan 6, 2009 OnlineFirst Version of Record
- Mar 10, 2009Version of Record >>
2012sgr.sagepub.comDownloaded from
http://sgr.sagepub.com/content/40/2/181
http://sgr.sagepub.com/cgi/alerts
http://sgr.sagepub.com/subscriptions
http://www.sagepub.com/journalsReprints.nav
http://www.sagepub.com/journalsPermissions.nav
http://sgr.sagepub.com/content/40/2/181.refs.html
http://sgr.sagepub.com/content/40/2/181.full.pdf
http://sgr.sagepub.com/content/early/2009/01/06/104649640832
8821.full.pdf
http://online.sagepub.com/site/sphelp/vorhelp.xhtml

Does Team Building Work?
Cameron Klein
Deborah DiazGranados
Eduardo Salas
Huy Le
C. Shawn Burke
Rebecca Lyons
University of Central Florida
Gerald F. Goodwin
Army Research Institute
This research reports the results of a comprehensive
investigation into the
effectiveness of team building. The article serves to update and
extend Salas,
Rozell, Mullen, and Driskell’s (1999) team-building meta-
analysis by assess-
ing a larger database and examining a broader set of outcomes.
Our study
considers the impact of four specific team-building components
(goal setting,
interpersonal relations, problem solving, and role clarification)
on cognitive,
affective, process, and performance outcomes. Results (based on
60 correla-
tions) suggest that team building has a positive moderate effect
across all
team outcomes. In terms of specific outcomes, team building
was most
strongly related to affective and process outcomes. Results are
also presented
on the differential effectiveness of team building based upon the
team size.

Keywords: team building; team performance; team development
Teams of people working together for a common cause touch all
our lives.
From everyday activities like air travel, fire fighting, and
running the United
Way drive to amazing feats of human accomplishment like
climbing Mt.
Everest and reaching for the stars, teams are at the center of
how work gets
done in modern life.
Kozlowski & Ilgen, 2006, p. 78
This quote, from a recent review of work-team effectiveness,
exemplifies
the central role that teams play in our lives. Although many
labels have
been applied to team-based forms of organizing (i.e., crews,
teams, groups,
and collectives), these entities are essential to the
accomplishment of orga-
nizational goals. Indeed, there is ample support in the literature
for the
Small Group Research
Volume 40 Number 2
April 2009 181-222
© 2009 SAGE Publications
10.1177/1046496408328821
http://sgr.sagepub.com
hosted at
http://online.sagepub.com

181
contention that team-based forms of organizing are beneficial
both to
organizations and to individuals. For example, Applebaum and
Batt (1994)
reviewed 12 large-scale surveys and 185 case studies of
managerial prac-
tices and concluded that team-based work leads to
improvements in orga-
nizational performance on measures of both efficiency and
quality. They
argued that team-based systems benefit workers because of the
higher like-
lihood of job enhancement, autonomy, and skill development
associated
with these systems. However, the simple existence of a team-
based orga-
nizing structure is not enough to ensure that positive outcomes
will result.
Teams must be nurtured, supported, and developed.
The Motivation for Understanding the
Efficacy of Team Building
There are three motivations for understanding the efficacy of
team-
building interventions in organizations. First, team building is
one of the

most commonly applied group development interventions in
organizations
today. It is widely used and comes in many forms, including
outdoor expe-
riential activities and indoor group process discussions.
However, no one is
quite sure how and why these interventions work, or if they
even work at
all. Considering the vast sum of money directed toward the
development of
teams in organizations, it is important that practitioners (and
researchers)
gain a better understanding of the effectiveness and boundary
conditions of
team building.
Second, as there are many options available to organizations in
the
pursuit of improved teamwork, it is important to determine
whether team
building is a worthy choice. Some of these interventions are
organizational
182 Small Group Research
Authors’ Note: This work was partially supported by funding
from the U.S. Army Research
Institute for the Behavioral and Social Sciences (Contract
W74V8H-04-C-0025) and grant
SES0527675 from the National Science Foundation, awarded to
Glenn Harrison, Stephen M.
Fiore, Charlie Hughes, and Eduardo Salas. All opinions
expressed in this article are those of
the authors and do not necessarily reflect the official opinion or
position of the University of
Central Florida, the Department of the Army, the Department of

Defense, or the National
Science Foundation. Portions of the article were presented at
the 2nd Annual Conference of
the Interdisciplinary Network for Group Research (INGRoup)
held at Michigan State
University, in Lansing, Michigan. Correspondence concerning
this article should be addressed
to: Dr. Eduardo Salas, Institute for Simulation & Training,
University of Central Florida, 3100
Technology Parkway, Suite 132, Orlando, FL 32826, USA. Tel:
(407) 882-1325; fax: (407)
882-1550; e-mail: [email protected]
or structural in nature and do not specifically target team
member inter-
actions (e.g., job redesign, selection systems, and group
incentive and
performance management programs). In contrast, team-
development inter-
ventions—those whose quest is to directly impact the
functioning and
effectiveness of work teams—provide the focus for the current
research
integration. These interventions, when properly conducted, can
have a
positive impact on organizations. Consider a research study
conducted by
Macy and Izumi (1993), who analyzed 131 studies of
organizational
change. They found that interventions with the largest effects
upon finan-

cial measures of organizational performance were team-
development
interventions. That is, of all organizational interventions, those
that focus
on team development had the largest effect on measures of
financial
performance.
Third, beyond financial performance, it is widely understood
that team
developmental interventions are key mechanisms that may be
used to facil-
itate team effectiveness (Noe, 2002). Therefore, it is important
to under-
stand how these interventions are most effective. At a general
level,
team-development interventions may include some form of
team-training
or team-building activities. Although both types of team-
development inter-
ventions are designed to improve team functioning and
effectiveness (espe-
cially when the science of individual and team training is
utilized; Salas &
Cannon-Bowers, 1997), team training and team building differ
in important
ways (Tannenbaum, Beard, & Salas, 1992). Team training is
skill-focused
(i.e., it is focused on gaining specific competencies), typically
includes a
practice component, and is done in context. It is generally
formal and
systematic. Team building, on the other hand, does not target
skill-based
competencies, is not systematic in nature, and is typically done
in settings

that do not approximate the actual performance environment.
For our pur-
poses, we define team building as a class of formal and informal
team-level
interventions that focus on improving social relations and
clarifying roles,
as well as solving task and interpersonal problems that affect
team func-
tioning. Team building works by assisting individuals and
groups to exam-
ine, diagnose, and act upon their behavior and interpersonal
relationships
(Schein, 1969, 1999).
In light of the above, this article investigates the efficacy of
team building.
We begin with a brief review of conceptual issues and
methodological issues
in team building. Next, the hypotheses for this research are
presented.
Finally, the results of several meta-analytic integrations are
discussed. First,
however, is the rationale for the present research.
Klein et al. / Team Building 183
Understanding the Need for an Update
It is an unfortunate indictment of the literature and practice in
this area

that we are still searching for answers to questions posed by
Beer (1976)
and Salas, Rozell, Mullen, and Driskell (1999). Namely, does
team build-
ing result in positive outcomes? Why? Under what conditions?
Upon a
careful review of the extent literature, it is clear that these
questions need to
be examined more closely to clarify our understanding of the
effectiveness
and boundary conditions of team building. Thus, this article
reports the
findings from a series of meta-analytic investigations of
research on the
efficacy of team-building interventions to update and extend the
current
state of knowledge in the team-building domain. In total, this
research will
examine three moderator variables in addition to making an
overall assess-
ment of the efficacy of team-building interventions. The current
replication
is meant to be systematic, rather than direct in nature (e.g.,
Aronson,
Ellsworth, Carlsmith, & Gonzales, 1990). That is, rather than
replicating
the exact analyses from the study by Salas and colleagues, the
current
research will replicate and extend the earlier study in an effort
to resolve
ambiguities and provide new insights for organizational
stakeholders and
academicians alike concerning the effectiveness of team
building.
Although the empirical evidence on team-building interventions

is lim-
ited, a critical investigation of the available literature is
warranted for sev-
eral reasons. First, since the 1990s there has been an increasing
incursion
of team-building interventions in organizations. Some of the
most recent
trends in team building have taken these interventions into the
kitchen and
even the wilderness. Second, many practitioners feel that these
interven-
tions are useful. However, a more careful exploration of the
utilities and
strengths of these interventions would benefit practitioners now
and would
benefit the development of these interventions in the long run.
Finally, there
are enough data available to form estimates regarding some of
the variables
that may moderate the impact a team-building intervention may
have on
team outcomes.
There is a general consensus for the idea that the science of
team train-
ing can be applied to enhance team functioning across
organizations (e.g.,
Kozlowski & Ilgen, 2006; Salas & Cannon-Bowers, 1997). At
the same
time, reviews of team effectiveness have noted the inconsistent
findings for
the effectiveness of team-building interventions. Despite these
inconsistent
findings, researchers such as Kozlowski and Ilgen have
acknowledged the
potential that these interventions may have on shaping team

development
and on improving team effectiveness. Given the increasing
frequency of
team-building interventions, if researchers do not catch up and
start doing
this research, an important window might be missed to
constructively shape
the important practice in this area. Thus, examining team
building further—
a topic that has not received recent sufficient attention in the
literature—is
critical. In the next section, a number of conceptual and
methodological
issues in team building are discussed.
Team Building
Conceptual Issues in Team Building
Originally designed as a group process intervention (e.g.,
Schein, 1969,
1999) for improving interpersonal relations and social
interactions, team
building has evolved to also include a concern for achieving
results, meet-
ing goals, and accomplishing tasks (Payne, 2001). In the late
1990s, Salas

and colleagues (1999) described team-building interventions as
extremely
popular and common. According to Beer (1976), there are four
basic
approaches to team building, including: (a) a goal-setting,
problem-solving
model; (b) an interpersonal model; (c) a role model; and (d) the
Managerial
Grid (Blake & Mouton, 1964) model. However, this initial
conceptualiza-
tion has since been reconsidered.
Refinements to this four-pronged system began with the
Managerial Grid
model being dropped as a distinct team-building approach. In
addition to
dropping the Managerial Grid model, modern conceptualizations
have sep-
arated the goal-setting and problem-solving approaches, using
Buller’s
(1986) problem-solving component as a distinct approach. As
discussed by
Buller, team-building models rarely exist in pure form. That is,
the interven-
tions reported in the literature usually involve elements from
several or all of
the models. As a result, he proposed a general problem-solving
model that
follows Dyer’s (1977) problem-solving framework. This model
incorporates
a focus on task or interpersonal issues, goal setting, and role
clarification,
depending on the nature of the specific problems identified for
the group
under investigation. Moreover, the problem-solving approach to
team build-

ing is said to subsume each of Beer’s (1976) components, and,
as perhaps
evident by its title, emphasizes the identification of major
problems in the
team. All told, these modifications have added clarity to the
investigation
and implementation of team building. Unfortunately, this
newfound concep-
tual clarity may not have come soon enough for many
investigators who had
previously sought to assess the efficacy of team building.
Perhaps because of the conceptual confusion that existed in this
area,
many reviews of team building did not include the same
articles. For example
two noteworthy reviews (DeMeuse & Liebowitz, 1981;
Woodman &
Sherwood, 1980) had only 14 studies in common out of a total
of 66 studies
reviewed. A subsequent review by Buller (1986) included only 9
studies, 3 in
common with Woodman and Sherwood (1980) and 6 in common
with
DeMeuse and Liebowitz. One reason for the apparent lack of
consistency in
the articles chosen for reviews is that team building has been an
ill-defined

concept (Buller, 1986). At the time of these early reviews there
was no
agreed-upon operational definition of the intervention. Taken
together, there
is now a consensus position that there exist four distinct models
of team
building. Although combinations of these approaches are
common, the
models include goal-setting, developing interpersonal relations,
clarifying
roles, and creating additional capacity for problem solving
(Beer, 1976;
Buller, 1986; Dyer, 1987; Salas et al., 1999). Table 1 describes
each of the
four models or components of team building in more detail.
Despite the recently established consensus on team-building
compo-
nents, there have been other problematic issues that have
persisted for
researchers and practitioners of this topic. Specifically, many
early efforts
were plagued by a number of methodological issues. A few of
these are dis-
cussed in the next section.
Methodological Issues in Team Building
Early reviews of team building described both a lack of
extensive research
on the issue and trepidation concerning the methodological
rigor of published
studies (Buller, 1986; DeMeuse & Liebowitz, 1981;
Tannenbaum et al.,
1992; Woodman & Sherwood, 1980). That is, even if one could
get beyond

the disagreement concerning operational definitions of team
building,
much of the previous team-building research is characterized by
method-
ological flaws (i.e., study design and measurement issues). For
example,
Buller (1982) found that more than half of the reported team-
building studies
employed pre-experimental designs—designs that do not allow
for causal
inferences. Also problematic, there was often no attempt to
disentangle the
effects of team building from other interventions that may be in
process
within an organization. Although this can be a common problem
for field
research in general, it has been particularly problematic in the
team-building
domain.
Tannenbaum and colleagues (1992) highlighted another
methodological
flaw in the team-building research. Specifically, they pointed
out that there
Table 1

Models/Components of Team Building
Salas, Rozell, Mullen, Salas, Priest, &
Component & Driskell, 1999 DeRouin, 2005
Goal setting Emphasis: Setting objectives Designed to
strengthen team member
and development of motivation to achieve team goals
individual and team goals. and objectives.
Team members: Become involved By identifying specific
outcome
in action planning to identify levels, teams can determine what
ways to achieve goals. future resources are needed.
Individual characteristics (e.g., team
member motivation) can also be
altered by use of this intervention.
Interpersonal Emphasis: Increasing teamwork Based on the
assumption that teams
relations skills (i.e., mutual supportiveness, with fewer
interpersonal conflicts
communication, and sharing function more effectively than
of feelings). teams with greater numbers of
Team members: Develop trust in interpersonal conflicts.
one another and confidence Requires the use of a facilitator to
in the team. develop mutual trust and open
communication between team
members.
As team members achieve higher
levels of trust, cooperation, and

cohesiveness, team characteristics
can be changed as well.
Role Emphasis: Increasing communication Defines the team as
comprising a set
clarification among team members regarding of overlapping
roles.
their respective roles within These overlapping roles are
the team. characterized as the behaviors that
Team members: Improve are expected of each individual
understanding of their own and team member.
others’ respective roles and duties Can be used to improve team
and
within the team. individual characteristics (i.e., by
reducing role ambiguity) and work
structure by negotiating, defining,
and adjusting team member roles.
Problem Emphasis: Identifying Buller’s (1986) problem-solving
solving major task-related problems component subsumes
aspects from
within the team. all of the components described by
Team members: Become involved Beer (1976).
in action planning, implement Team members practice setting
goals,
solutions to identify problems develop interpersonal relations,
clarify team roles, and work to
(continued)

has been a reliance on measuring the effectiveness of team-
building inter-
ventions with process measures. Although implicitly appealing,
improve-
ments in processes can not always be linked to improvements in
team
performance (e.g., Porras & Wilkens, 1980). For example, team
perfor-
mance is typically fashioned by additional environmental and/or
organiza-
tional characteristics and contingencies that are out of the
volitional control
of team members. Stated differently, team members are active
participants
in the enactment of team processes, but must also interact
within the larger
system to produce more distal performance outputs. Team
processes may
also be impacted by the larger organizational environment, but
not likely to
the same degree as team performance outputs.
As a final methodological concern with previous team-building
research, there has been an overreliance on subjective indicators
of group
or organizational performance criteria as dependent measures
(Tannenbaum
et al., 1992). Though this type of information may be interesting
and rele-
vant to measuring participant satisfaction and other affective

outcomes that
may be impacted by team-building interventions, it is often not
concrete
enough to allow for accurate predictions of the performance
outcomes of
team building.
A Recent Advancement in Team-Building Research
Despite the conceptual and methodological issues associated
with eval-
uations of team building, one recent effort has represented
advancement
over previous reviews by empirically investigating the
effectiveness of
these interventions. Specifically, Salas and colleagues (1999)
responded to
a call by Buller (1986) to use the primary focus of the
intervention
(i.e., goal setting, interpersonal relations, role clarification, and
problem
solving) as a potential moderating variable. The results of their
study failed
Table 1 (continued)
Salas, Rozell, Mullen, Salas, Priest, &
Component & Driskell, 1999 DeRouin, 2005
and to evaluate those improve organizational
solutions. characteristics through
problem-solving tasks.
Can have the added benefit of

enhancing critical-thinking skills.
to indicate a relationship between the combined set of team-
building inter-
ventions and team performance (r = .01, k = 16 effect sizes).
Moreover, of
the four components of team building, only role clarification
proved to be
effective, as judged by both objective (r = .71) and subjective (r
= .75)
accounts of performance. Interestingly, for objective and
subjective mea-
sures there was a nonsignificant effect of goal setting (r = −.06
and −.11,
respectively), interpersonal relations (r = −.38 and −.04,
respectively), and
problem solving (r = −.31 and .09, respectively). Also examined
in this
study were the potential moderating influences of the source of
the criterion
measurement of performance (i.e., objective vs. subjective),
team size, and
training duration. For objective measures of performance, there
was no evi-
dence of a relationship between team building and performance
(r = −.04,
k = 8); for subjective measures, there was a small positive
relationship
between team building and performance (r = .14, k = 8).

Concerning team
size, the results of their study suggested that the effects of team
building on
performance decreased as a function of the size of the team (r =
−.34).
Finally, there was a slight tendency for the effects of team
building to
decrease as a function of the duration of the intervention (r =
−.20).
Taken together, this study enhanced our understanding of the
efficacy of
team building. However, there were a number of limitations
associated with
it—limitations that now necessitate the need for additional
inquiry. For
example, the amount of data analyzed in this study was
relatively modest.
Specifically, the research findings presented were based on only
16 effect
sizes, and thus, it is difficult to discern, with any degree of
confidence, the
actual effectiveness of these interventions. Moreover, the
findings derived
from the Salas and colleagues’ (1999) study did not necessarily
reflect the
findings from existing narrative reviews. Finally, their study
left many
questions unanswered regarding the potential moderating impact
of other
relevant variables. As an organizing tool provided to summarize
the litera-
ture in this area, Table 2 provides a summary of five previous
investigations
into the efficacy of team building, and includes the number of
articles

reviewed, the years spanned, and other noteworthy features.
In summary the current research was initiated to provide an
updated
meta-analysis of the team-building literature. Phrased in the
form of ques-
tions, an examination of these issues will help clarify our
understanding of
the effectiveness of team building and lead to hypothesized
relationships
involving the effectiveness of team building, including an
investigation of
specific moderators. A model that serves to graphically
illustrate these
hypotheses is presented in Figure 1.
Table 2
Summary of Previous Team-Building Reviews
Study Summary of Noteworthy
Review Details Findings Features
Woodman & 30 articles TB elicits positive affective
Distinguished between team
Sherwood reactions. development and T-group
(1980) Qualitative The linkage between TB and or sensitivity

training.
data work group performance Made subjective assessments
remains largely unsubstantiated. of the internal validity
Years spanned: TB is more commonly of the studies reviewed.
1964–1978 conducted with management
teams than groups lower
in organizational hierarchies.
TB is more commonly conducted
with intact and established
work teams than new groups.
Affective reactions as dependent
measures are used more often
than objective performance data.
DeMeuse & 36 articles TB is described as having great Coded
studies according to
Liebowitz promise for improving employee research design,
sample size,
(1981) Qualitative attitudes, perceptions, behaviors, multiple
dependent variables,
data and organizational effectiveness. and duration of the TB
Eighty-seven percent of the 68 intervention.
Years spanned: evaluations indicated positive results.a
1962–1980 Due to the lack of rigorous research
designs, firm conclusions
concerning the effectiveness of these
interventions could not be made.

Buller 9 articles TB must be more carefully defined. Presented
and described a
(1986) More rigorous experimental general problem solving
Qualitative designs should be used. approachto TB that added to
data There are numerous methodological Beer’s (1976)
flaws in previous TB studies. four-component model.
Years spanned: Therefore a clear assessment of the Argued that
Beer’s
1964–1981 TB and task performance classification is difficult
to
relationship had yet to emerge. use in practice because TB
programs usually involve
elements from each of the
models.
Tannenbaum, 17 articles The quantity of TB research decreased;
Presented a comprehensive
Beard, & however, the quality had improved. model of team
effectiveness
Salas Qualitative Most studies used multiple components that
continues to influence
(1992) data in their TB interventions. research and theorizing
More researchers began using in this field.
Years spanned: behavioral and objective measures.
1980–1988 Presented evidence to cast doubt the
connection between process and
performance; TB is effective, but
for only perceptions and attitudes.
(continued)

Table 2 (continued)
Study Summary of Noteworthy
Review Details Findings Features
Post-intervention strategies may be a
key mechanism to ensure the
long-term effectiveness of these
interventions.
Salas, Rozell, 11 articles No significant effect of TB on First
known attempt to
Mullen, & performance. empirically summarize the
Driskell Quantitative A nonsignificant tendency for TB to
effectiveness of TB.
(1999) data result in lower performance when Assessed a
number of
measured objectively, but increase moderators, including team
Years spanned: performance when subjective building
component, team
1965–1990 measures were used. size, training duration, and
The role clarification component was type of performance
measure
more likely to increase performance.b used (i.e., objective

The effects of TB decreased as the size versus subjective).
of the team increased.
The effects of TB decreased as the
duration of the intervention increased.
Note: TB = team building.
a. Many of the 36 studies reported evaluations of multiple
dependent variables.
b. This result should be interpreted with caution as it was based
on a small number of effect sizes (it’s dif-
ficult to determine, but likely only three or four effect sizes
were used in this calculation).
Hypotheses
Is Team Building Effective?
We agree with Salas and colleagues (1999) that previous
narrative
reviews have frequently expressed the benefits that can result
from team
building (e.g., Buller, 1986; DeMeuse & Liebowitz, 1981;
Sundstrom,
DeMeuse, & Futrell, 1990, Tannenbaum et al., 1992; Woodman
&
Sherwood, 1980); however, there has been a lack of definitive,
compelling
evidence concerning the positive effects of team building on
team performance.
Specifically, previous qualitative reviews of the team-building
domain have
concluded that evidence of an effect of team building on
performance was
“inconclusive” (Buller, 1986), “unsubstantiated” (Woodman &

Sherwood,
1980), “equivocal” (Tannenbaum et al., 1992), and “mixed”
(Sundstrom
et al., 1990). Meta-analytic results from one study have
suggested there is
no overall effect of team building on team performance (Salas
et al., 1999).
However, there was support for the role-clarification component
of team
building. Theoretically, one would assume that an intervention
focused on
Figure 1
Theoretical Model Depicting Study Hypotheses
improving team functioning would result in positive (as opposed
to nega-
tive) outcomes. In addition, the moderate support suggested by
several
narrative reviews (e.g., Tannenbaum et al.) leads us to predict a
positive
overall effect of team building on team functioning. Thus, we
present our
first hypothesis:
Hypothesis 1: Team building interventions will result in
enhanced team out-
comes.

The result from testing this hypothesis will provide for a
baseline judg-
ment concerning the efficacy of this particular form of team-
development
intervention. The omnibus test will therefore allow for an
overall assess-
ment of the efficacy of team building, with all independent
outcomes from
primary studies combined for this analysis.
Is Team Building More Effective for Some
Outcomes Than Others?
This framing question asks whether the combined set of team-
building
interventions is shown to be more useful for improving certain
team outcomes
than others. Specifically, does team building work better for
cognitive
outcomes (e.g., declarative knowledge of teamwork
competencies), team
member affective outcomes (e.g., trust, team potency), team
processes
(e.g., coordination, communication), or team performance
outcomes (e.g.,
volume of sales, productivity measures)? This division of team-
building
outcomes is similar to the commonly discussed cognitive,

affective, and
skill-based breakdown of general training outcomes (e.g.,
Kraiger, Ford, &
Salas, 1993). However, for the current research, skill-based
outcomes are
further divided into two additional categories—team processes
and more
performance-related or productivity-related outcomes.
The division and examination of team-building effectiveness
based on
specific outcomes is intended to help clarify the often disparate
results
seen in the literature on these interventions. For example,
Woodman and
Sherwood’s (1980) qualitative review of team building
concluded that it
was only useful for facilitating affective outcomes, not team
performance.
Equally pejorative to the position that team building results in
improved
team performance was the conclusion provided by Buller
(1986), who upon
reviewing team-building studies conducted through 1980,
suggested that
the relationship between team building and performance was
inconclusive.
In yet another review that added to the opaque nature of the
efficacy of team
building, DeMeuse and Liebowitz (1981) accurately noted that
most of the
early research on team building relied almost exclusively on
perceptual
ratings of the dependent variables being studied.
More recent reviewers of the efficacy of team building have

reported
somewhat different (i.e., more positive) conclusions. For
example, the inte-
gration of team-building research reported by Salas and
colleagues (1999)
has served, in many ways, as the focal point for the current
research. And,
although there was no significant overall effect of team building
in their
research, there was a small, yet significant, tendency for team
building to
increase performance when criteria were assessed with
subjective measures.
In another investigation into the efficacy of team building,
Tannenbaum and
colleagues (1992) reviewed team-building research conducted in
the 1980s
and found support for these interventions, especially when the
outcome of
interest was limited to team member perceptions or attitudes.
Finally,
Svyantek, Goodman, Benz, and Gard (1999) found that team
building posi-
tively impacted work-group productivity. Specifically, the
largest impact of
team building was on productivity measures of cost-
effectiveness, with a
smaller influence on quantity and quality measures.
In general, previous research that has investigated the
effectiveness of
team building for improving specific outcomes has been
equivocal at best,
especially considering performance or productivity measures.
At the same
time, there has been perhaps the greatest support for the

efficacy of team
building for improving affective or attitudinal outcomes.
Finally, to our
knowledge there have been no reviews citing the correlation
between team-
building interventions and improvements in cognitive outcomes.
The mixed results found in the literature concerning the effect
of team-
building interventions on team outcomes lead us to believe that
moderators
may exist. Therefore, team building impacts certain outcomes
more than
others. Based on the general characteristics of team building,
(e.g., that
team building develops interpersonal relations, mutual trust, and
open com-
munication between team members), it is likely that team
building will
have a greater effect on affective outcomes than any on other
type of out-
comes. As a process intervention, it makes theoretical sense that
team
building would result in enhanced team member affective
outcomes.
Finally, although the findings of more distal reviews of team
building were

taken into consideration, a decision was made to place more
credence on
recent research, which more often reported positive findings.
These find-
ings, along with a complete consideration of the expected
benefits of team
building, led us to Hypotheses 2a and 2b:
Hypothesis 2a: Team building interventions will result in
improved outcomes
across each of the four outcome types.
Hypothesis 2b: Team building will be most effective for
improving affective
outcomes.
Does the Focus of Team Building Moderate Its Effectiveness?
It is certainly important to be able to identify whether team
building
results in enhanced team functioning, but it is perhaps more
informative to
know which forms or components of team building are most
effective.
Advancements in theory in the last quarter century (e.g., Buller,
1986; Salas
et al., 1999; Salas, Priest, & DeRouin, 2005) have allowed for
the parcel-
ing of team-building interventions into four distinct foci (see
Table 1).
Unfortunately, the single existing empirical integration on the
relative effi-
cacy of different components of team building (i.e., Salas et al.)
did not
provide encouraging results. Combining subjective and
objective estimates,

there was a slight (nonsignificant) tendency for goal-setting (r =
−.16),
interpersonal relations (r = −.06), and problem-solving (r =
−.05) models
of team building to result in decreased performance. It was only
the
role-clarification component that appeared to be effective for
improving
performance (r = .76). However, upon examination of their data
set, it
appears this positive finding for role clarification was supported
by the
results from only three studies and three effect sizes. Thus,
caution and further
investigation are warranted before concluding that the role-
clarification
component of team building is superior to the other forms.
Nonetheless, because the role-clarification component of team
building
is designed to relieve role stress as created by role ambiguity or
role con-
flict, it is reasonable to anticipate that a significant
improvement in team
functioning should result. In addition, the role-clarification
component of
team building emphasizes communication among team members,

and
thus it is likely that an increase in the level and quality of
communication
between team members will impact their effectiveness. Unlike
the other
forms of team building, improvements in role clarity and
communication
are expected to produce more lasting benefits in terms of team
function-
ing (as assessed through an analysis of the combined set of team
outcomes).
Combining this theoretical rationale with the preliminary
findings of
Salas and colleagues (1999) concerning the role-clarification
component
of team building, it is our belief that a team-building
intervention that
utilizes a role-clarification focus will provide the most benefit
to team
functioning. At the same time, it is our view that any carefully
thought out
team-building intervention should have at least some positive
impact on
team members. Thus, we also expect that team building that
focuses on
interpersonal relations, goal setting, or problem solving will
also prove
useful, at least for the short-term benefits that are typically
assessed.
Taking these dual considerations into perspective, we present
our next set
of hypotheses:
Hypothesis 3a: Each of the four components of team building
will demon-
strate a moderate level of effectiveness for improving team

functioning.
Hypothesis 3b: The role clarification component of team
building will be
most effective for improving team functioning.
Is the Effectiveness of Team Building
Moderated by Team Size?
This research will also investigate whether the effectiveness of
team
building is moderated by the size of the team. At a basic level,
“the resources
available on a team result from how many people are on it”
(Hambrick &
D’Aveni, 1992, p. 1449). Sundstrom, McIntyre, Halfhill, and
Richards’s
(2000) review of 83 field studies and experiments conducted
with work
groups suggested that the average group size was 11 members.
Moreover,
Halebian and Finkelstein (1993) have suggested that team size
is synony-
mous with cognitive capability. Providing support for this
assertion, Bantel
and Jackson (1989) found that larger teams generally have a
greater reservoir

of cognitive resources than smaller teams. There is little doubt
that there are
benefits to having medium- to large-sized teams available to
perform work
tasks rather than smaller teams. However, research has also
found that larger
teams can facilitate the enactment of other, less desirable, group
phenome-
non, including the participation leadership effect, the in-group
bias effect, the
cohesiveness performance effect, and the well-known
groupthink effect (cf.
Mullen, Anthony, Salas, & Driskell, 1994; Mullen, Brown, &
Smith, 1992;
Mullen & Copper, 1994; Mullen, Salas, & Driskell, 1989).
From other research, we know that information pooling is
critical to team-
based decision making. That is, the size of a team may impact
the effectiveness
with which a team pools common and unique information. For
example, Stasser,
Vaughan, and Stewart (2000) discussed the tendency for group
members to dis-
cuss shared information rather than the unique information that
is held by team
members—an issue that becomes more problematic as teams
increase in size.
This lack of attention to unique information held by individual
team members
can lead to mal-informed decisions and occasionally even
detrimental results.
Salas and colleagues (1999) presented the only existing research
integration
that has assessed the relative efficacy of team building for

httphfs.sagepub.comErgonomics Societyof the Human F.docx

httphfs.sagepub.comErgonomics Societyof the Human F.docx

Recommended

Recommended

More Related Content

Similar to httphfs.sagepub.comErgonomics Societyof the Human F.docx

Similar to httphfs.sagepub.comErgonomics Societyof the Human F.docx (20)

More from AASTHA76

More from AASTHA76 (20)

Recently uploaded

Recently uploaded (20)

httphfs.sagepub.comErgonomics Societyof the Human F.docx