You are on page 1of 18

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/7600996

A Multilevel Examination of the Relationships Among Training Outcomes,


Mediating Regulatory Processes, and Adaptive Performance.

Article  in  Journal of Applied Psychology · October 2005


DOI: 10.1037/0021-9010.90.5.827 · Source: PubMed

CITATIONS READS
156 2,139

3 authors, including:

Gilad Chen Craig Wallace


University of Maryland, College Park University of Denver
34 PUBLICATIONS   9,260 CITATIONS    46 PUBLICATIONS   3,510 CITATIONS   

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Gilad Chen on 24 May 2015.

The user has requested enhancement of the downloaded file.


Journal of Applied Psychology Copyright 2005 by the American Psychological Association
2005, Vol. 90, No. 5, 827– 841 0021-9010/05/$12.00 DOI: 10.1037/0021-9010.90.5.827

A Multilevel Examination of the Relationships Among Training Outcomes,


Mediating Regulatory Processes, and Adaptive Performance

Gilad Chen Brian Thomas


Texas A&M University Georgia Institute of Technology

J. Craig Wallace
Tulane University

This study examined whether cognitive, affective–motivational, and behavioral training outcomes relate
to posttraining regulatory processes and adaptive performance similarly at the individual and team levels
of analysis. Longitudinal data were collected from 156 individuals composing 78 teams who were trained
on and then performed a simulated flight task. Results showed that posttraining regulation processes
related similarly to adaptive performance across levels. Also, regulation processes fully mediated the
influences of self- and collective efficacy beliefs on individual and team adaptive performance. Finally,
knowledge and skill more strongly and directly related to adaptive performance at the individual than the
team level of analysis. Implications to theory and practice, limitations, and future directions are
discussed.

Changes in the nature of work, such as rapid technological Bowers, 2001) and work teams (e.g., Kozlowski & Bell, 2003), we
innovations and the increasing reliance on interdependent work still know relatively little about how individual- and team-level
teams, require employees to be highly adaptable (Ilgen & Pulakos, learning outcomes translate simultaneously into adaptive posttrain-
1999; Pulakos, Arad, Donovan, & Plamondon, 2000). This new ing performance in work teams. We are aware of no multilevel
reality has created a greater need for improving training effective- studies in which both training and transfer outcomes were exam-
ness in organizations. In particular, it is critical that we gain a ined simultaneously at the individual and team levels of analysis,
better understanding of which learning outcomes developed during and only a few individual-level studies (e.g., Smith-Jentsch, Salas,
training (e.g., knowledge, skills, and other characteristics) are most & Brannick, 2001) have considered contextual influences of teams
likely to enable employees to effectively adapt to turbulent and on individual-level transfer. Furthermore, when team researchers
unpredictable work environments (Ford, Smith, Weissbein, Gully, have studied training and transfer the focus has been on team-level
& Salas, 1998; Kozlowski et al., 2001; Salas & Cannon-Bowers, predictors, processes, and outcomes and has largely ignored
2001). Moreover, it is important to consider whether and how individual-level relationships (Chen & Bliese, 2002; Kozlowski &
training interventions can enable effective transfer in interdepen- Bell, 2003). Accordingly, the purpose of the present research was
dent work teams, on which adaptability is crucial at both the to delineate and test a multilevel model relating training outcomes
individual and team levels (Kozlowski & Bell, 2003; Kozlowski & to training transfer at both the individual and team levels. To do so,
Salas, 1997). we integrated the individual-level and team-level literatures on
Unfortunately, despite progress in the literatures on training and training and posttraining transfer performance, as well as self-
transfer (e.g., Colquitt, LePine, & Noe, 2000; Salas & Cannon- regulation theory and theories of team processes. Thus, the present
research provides both vertical (i.e., multilevel) and horizontal
(i.e., within level) integrations and extensions of the extant train-
Gilad Chen, Department of Management, Mays Business School, Texas
ing, work teams, and motivation literatures.
A&M University; Brian Thomas, School of Psychology, Georgia Institute One type of work team in which adaptability is particularly
of Technology; J. Craig Wallace, Department of Psychology, Tulane crucial is the action team (Kozlowski, Gully, Nason, & Smith,
University. 1999), defined as “highly skilled specialist teams cooperating in
Data collection for this study was conducted while Gilad Chen and J. brief performance events that require improvisation in unpredict-
Craig Wallace were at the School of Psychology, Georgia Institute of able circumstances” (Sundstrom, De Meuse, & Futrell, 1990, p.
Technology. An early version of this article was presented at the August 121). Examples of such teams include sports teams, Special Forces
2003 63rd Annual Meeting of the Academy of Management, Seattle, WA. combat teams, search and rescue teams, and flight crews. Training
We thank Michelle Marks and John Mathieu for their help during early transfer is particularly challenging in such teams, because of the
phases of the project, Dov Eden and Steve Kozlowski for providing helpful
novel, complex, dynamic, and unpredictable nature of their work
and insightful comments on drafts of the article, and Erin Page for her
assistance in data collection.
environment (Marks, Sabella, Burke, & Zaccaro, 2002; Marks,
Correspondence concerning this article should be addressed to Gilad Zaccaro & Mathieu, 2000). Such teams can rarely be trained to
Chen, Department of Management, Mays Business School, Texas A&M perform effectively in all possible environments and situations.
University, College Station, TX 77843-4221. E-mail: gilad.chen@tamu Instead, training programs directed at action teams should focus on
.edu developing certain knowledge, skills, and other characteristics that

827
828 CHEN, THOMAS, AND WALLACE

enable such teams to be highly adaptable (Kozlowski et al., 1999; acquired during training to effectively meet novel, difficult, and
Marks et al., 2000, 2002). Therefore, the present research focused complex situations (Schmidt & Bjork, 1992). Following the work
on action teams and their individual members. of Ford et al. (1998) and Kozlowski et al. (2001), in the present
research we examined the transfer criterion that represents adap-
Conceptual Foundations and Framework tive performance, or the extent to which trainees perform effec-
tively in complex and novel situations following training. It is
We developed a framework for studying training transfer in important to note, we considered both individual and team adap-
teams, according to which training and/or learning outcomes relate tive performance. Team adaptive performance reflects the extent to
to training transfer performance at least partially through posttrain- which the team meets its objectives during a transfer performance
ing regulation processes involving the reactions and actions of
episode, whereas individual adaptive performance reflects the ex-
trainees in the pursuit of posttraining task goals. Figure 1 delin-
tent to which each member effectively executes his or her role in
eates a multilevel model derived from the framework, which
the team during the transfer episode.
includes similar multilevel constructs (i.e., constructs that share
Trainees in our study learned to operate a simulated attack
similar meanings across levels of analysis; Chen, Mathieu, &
helicopter and were then asked to apply the tasks and procedures
Bliese, 2004), as well as comparable individual- and team-level
learned during training in a more novel and complex environment.
relationships among these constructs (Kozlowski & Klein, 2000).
Specifically, the transfer task involved higher component com-
Below, we delineate the constructs in the model and discuss the
plexity than the tasks trainees were exposed to during training—
expected relationships among them.
that is, it involved a greater number of distinct acts and required
handling more informational cues when performing the task
Training Transfer and Adaptive Performance (Wood, 1986). According to Smith et al. (1997), this characterizes
Training transfer is usually defined in terms of the generaliza- a relatively weak form of adaptation process, in which people are
tion and maintenance of knowledge and skills acquired during required to apply the same methods learned during training to a
training (Baldwin & Ford, 1988; Salas & Cannon-Bowers, 2001). new exemplar (e.g., learning to operate a car off-road or under
More recently, several researchers have argued that the generali- severe weather after being accustomed to operating it on highways
zation component of training transfer is more critical when con- in good weather). We focused on this level of adaptability because
sidering complex, unpredictable, and dynamic posttraining envi- participants in this study lacked the level of expertise needed to
ronments (Ford et al., 1998; Kozlowski et al., 2001). In its core, the perform more complex forms of adaptability, and asking them to
generalization component of training transfer involves adaptive apply the tasks they learned in training in a more complex and
expertise (Smith, Ford, & Kozlowski, 1997). Adaptive expertise is novel environment still required them to be highly adaptable (cf.
the capability to modify knowledge, skill, and other characteristics Smith et al., 1997). In addition to being more complex relative to

Figure 1. A multilevel model of training outcomes and posttraining adaptive performance.


MULTILEVEL TRAINING TRANSFER 829

training missions, the transfer performance episode in our study fully certain trained tasks; e.g., Colquitt et al., 2000; Ford et al.,
required participants to handle three specific aspects of adaptabil- 1998; Kozlowski et al., 2001).
ity discussed by Pulakos et al. (2000): (a) handle emergencies or Recently, two individual-level studies (Ford et al., 1998; Koz-
crisis situation (e.g., quickly analyze plausible strategies when lowski et al., 2001) tested the extent to which different training
realizing ammunition is lower than expected or when mission time outcomes uniquely predict posttraining adaptive performance.
limit neared), (b) handle work stress (e.g., maintain composure These studies used a computer-based individual decision-making
while escaping antiaircraft artillery), and (c) deal with uncertain task, which required individuals to work on transfer tasks in more
and unpredictable work situations (e.g., adjust mission plans when novel and complex environments than the training environment. In
realizing the aircraft was damaged or after certain ammunition had both studies, declarative knowledge, self-efficacy, and skill ac-
depleted). quired during training uniquely and positively predicted posttrain-
Both individual-level (e.g., Ford et al., 1998; Kozlowski et al., ing adaptive performance, even when individual differences (aca-
2001) and team-level (e.g., Marks et al., 2000, 2002) studies have demic ability and goal orientation) and various training and
examined adaptive performance. However, most research to date learning strategies (e.g., self-regulation activity during learning,
has studied either individual adaptive performance or team adap- extent to which the training program was composed of identical
tive performance (cf. Chen et al., 2002). Studying adaptive post- elements) were controlled. These results were also corroborated by
training performance at both the individual and team levels simul- a recent meta-analysis conducted by Colquitt et al. (2000). Thus,
taneously can uncover similar and distinct predictors of adaptive there is considerable evidence that cognitive, affective–
performance at different levels. If similar training-related predic- motivational, and behavioral aspects of learning all contribute to
tors account for adaptive performance at both levels, it would individual-level adaptive performance. However, the specific
suggest team training programs should attempt to develop similar mechanisms linking training outcomes and individual adaptive
knowledge, skills, and other characteristics at both levels. In con- performance have remained largely unexplored.
trast, finding differences across levels may suggest team training At the team level, training researchers have focused primarily on
programs should focus on developing unique or complementary shared knowledge (or shared mental models) and collective effi-
individual and team competencies. cacy as predictors of posttraining team adaptive performance (e.g.,
It is interesting that individual- and team-level frameworks for Marks, 1999; Marks et al., 2000, 2002; Mathieu, Heffner, Good-
studying training transfer are both similar and complementary. At win, Salas, & Cannon-Bowers, 2000). Shared mental models refer
the individual level, Baldwin and Ford (1988) have proposed that to shared knowledge organization and/or structure among team
trainee characteristics (e.g., ability, personality), the work environ- members (Klimoski & Mohammed, 1994). Collective efficacy is a
ment (e.g., supervisory support), and learning outcomes help pre- group- or team-level analog of self-efficacy, reflecting “a group’s
dict individual differences in transfer performance. At the team shared belief in its conjoint capabilities to organize and execute the
level of analysis, the input–process– output (IPO) framework of courses of action required to produce given levels of attainments”
team effectiveness (Hackman, 1987; Marks, Mathieu, & Zaccaro, (Bandura, 1997, p. 477). However, team research has shown that
2001) suggests that team training, as well as other “inputs” (e.g., team processes, such as coordination and backup behaviors, fully
team composition), contribute to team effectiveness, albeit through mediate the impact of training outcomes on adaptive performance
their impacts on team processes. There are two key differences (e.g., Marks et al., 2000, 2002; Mathieu et al., 2000). Also, team
between individual- and team-level frameworks of training trans- studies generally have focused on one type of training outcome,
fer. First, the team-level IPO framework assumes that team pro- but have yet to consider multiple training outcomes in combina-
cesses, such as coordination, explain how training outcomes affect tion. Nonetheless, Salas and Cannon-Bowers (1997) have pro-
team transfer performance, whereas Baldwin and Ford’s posed that the training outcome taxonomy developed by Kraiger et
individual-level framework does not consider the possibility that al. (1993) should also apply to teams. Specifically, in addition to
posttraining self-regulation processes could help explain the influ- shared mental models and collective efficacy, role-specific knowl-
ences of training outcomes on transfer of training performance. edge of team members and the teams’ collective skill should
Second, as we summarize below, the role of different types of contribute to team adaptive performance.
training outcomes in training transfer is considered more thor- In this study, individual-level training outcomes included role
oughly in individual-level models of training and transfer than in knowledge (i.e., knowledge regarding one’s role in the team),
team-level models. self-efficacy (i.e., one’s belief in his or her capability to perform
his or her particular role in the team), and individual skill (i.e.,
Training Outcomes one’s demonstrated capability to perform his or her particular
role). At the team level, training outcomes included team knowl-
Kraiger, Ford, and Salas (1993) have proposed a training out- edge, collective efficacy, and team skill. We conceptualized team
comes taxonomy, involving three types of learning outcomes: (a) knowledge as a compositional variable based on the aggregation of
knowledge (i.e., cognitive learning outcomes), (b) skills (i.e., be- role knowledge to the team level—that is, it involves the extent to
havioral learning outcomes), and (c) affect (i.e., affective and which teams are composed of individuals who are knowledgeable
motivational learning outcomes). In line with Kraiger et al.’s about their individual roles. Bliese (2000) has noted that aggre-
taxonomy, three of the most widely studied individual-level train- gating an individual construct to the group level results in a new
ing outcomes are declarative and procedural knowledge (i.e., in- construct that is both similar to and different from the individual-
formation about what and how to do certain trained tasks), skill level construct (see also Chen et al., 2004). Thus, role and team
acquisition (i.e., capability to actually do certain trained tasks), and knowledge are not entirely isomorphic, yet both capture cognitive
self-efficacy (i.e., belief in one’s capability to accomplish success- outcomes of training.
830 CHEN, THOMAS, AND WALLACE

Collective efficacy, on the other hand, is based on a referent- action processes, respectively. Still, goal choice and goal striving
shift composition model (Chan, 1998) in that it involves a shared are inherently individual-level constructs, reflecting the amount of
belief among team members regarding their capability to perform effort individual team members expend when working with other
team, not individual, tasks. Research has suggested that self- and team members on developing and executing team plans. Goal
collective efficacy beliefs are isomorphic constructs— both reflect choice and goal striving activities facilitate effective team regula-
perceived competence, and both are hypothesized to drive perfor- tion, but they also enhance individual performance because mem-
mance through motivational processes (Chen & Bliese, 2002; bers’ roles are highly interdependent and development and execu-
Chen et al., 2002, 2004). Finally, team skill is similar to individual tion of effective team plans can aid members’ own performance. In
skill, as it captures the demonstrated collective capability of the contrast, team transition and action processes reflect team-level
team to perform its team tasks. In particular, both individuals and processes (i.e., what the team as a collective does), rather than
teams can differ on how well they execute trained tasks, and team individual-level contributions to these processes (i.e., what indi-
training programs should develop such competencies at both the vidual members do).
individual and team levels of analysis (Kozlowski et al., 1999). Both individual- and team-level models of regulation processes
explicate that over time individuals and teams perform multiple
Regulation Processes performance episodes (e.g., Marks et al., 2001; Mathieu & Button,
1992). For action teams, the notion of recurring performance
The work motivation literature has identified two main self- episodes is particularly explicit, as they perform collectively over
regulation or motivational processes: goal choice and goal striving multiple and clearly differentiated performance episodes (e.g.,
(e.g., Kanfer, 1990; Mitchell & Daniels, 2003). These self- combat flight missions, surgeries, rescue operations). What is
regulation processes involve how people act and react in the important is that goal choice and transition processes most often
pursuit of goals. Specifically, goal choice involves the process of take place prior to and/or in between performance episodes,
deciding where and how to allocate task-related effort, whereas whereas goal striving and action processes occur during perfor-
goal striving has to do with actually allocating and sustaining mance episodes. Thus, we refer to transition and goal choice
effort in the pursuit of goal accomplishment. Likewise, the liter- processes as between-episodes processes and to action and goal
ature on work teams has identified team-level processes that are striving processes as within-episode processes. Research on action
highly similar to the individual-level goal choice and goal striving teams has shown that cognitive and affective–motivational learn-
processes, namely action and transition processes (Marks et al., ing outcomes affect adaptive performance following training
2001). Team transition processes involve the following interde- through within-episode team processes, such as coordination and
pendent team activities: (a) mission analysis (i.e., interpretation backup behaviors (e.g., Marks, 1999; Marks et al., 2000, 2002;
and evaluation of the team’s mission or task), (b) goal specification Mathieu et al., 2000). However, we are aware of no research to
(i.e., identification and prioritization of team goals and subgoals), date that has examined the extent to which both within-episode and
and (c) strategy formulation and planning (i.e., development of between-episodes processes mediate the effects of different team
particular courses of action for goal accomplishment). These tran- training outcomes on team adaptive performance. Moreover,
sition activities are highly consistent with the individual-level goal individual-level research has yet to examine the possible mediating
choice process. On the other hand, team action processes include regulation processes linking training outcomes and adaptive
four different interdependent team activities, which are consistent performance.
with the goal striving process: (a) monitoring progress toward
Hypotheses
goals (i.e., assessing how the team does relative to its mission/task
goals), (b) system monitoring (i.e., tracking material resources and Following Figure 1 and building on the discussion above, we
environmental conditions as they relate to mission accomplish- delineated specific hypotheses pertaining to the relationships
ment), (c) team monitoring and backup behaviors (i.e., assisting among training outcomes, posttraining regulation processes, and
team members in performing their task roles), and (d) coordination adaptive performance at both the individual and team levels of
(i.e., orchestrating the sequence and timing of interdependent actions). analysis. Given that several of the constructs in the model are not
A recent study by DeShon, Kozlowski, Schmidt, Milner, and entirely parallel (i.e., are similar, yet not isomorphic), we do not
Weichmann (2004) found evidence that the content and function of position Figure 1 as a truly homologous model (cf. Chen, Bliese,
self- and team regulation processes can be quite similar during & Mathieu, in press; Kozlowski & Klein, 2000). Nonetheless, the
learning. However, it is still unclear how self-regulation processes constructs in the model are sufficiently similar as to make mean-
might operate in teams following training. According to Kozlow- ingful comparisons of relationships across levels. Moreover, test-
ski et al. (1996, 1999), as teams develop, the focus of self- ing the similarity (or dissimilarity) of relationships across levels
regulated activities shifts. During early phases of team develop- can help uncover the extent to which the validity of these con-
ment, team members focus on learning their specific roles, but as structs is similar or different across levels.
individuals compile into mature teams, the focus of individual Following Marks et al. (2001), at both levels of analysis, it was
activities shifts to regulating collective team tasks. Hence, because expected that within-episode processes (i.e., goal striving activities
we focused on posttraining regulation processes, we conceptual- and action processes) would fully mediate the effects of between-
ized goal choice as individual activities directed at helping the episodes processes (i.e., goal choice activities and transition pro-
team choose its goals and goal striving as individual activities cesses). Individual- and team-level research on goal setting, plan-
directed at helping the team accomplish its goals. Given the high ning, and self-regulation has lent support for this theoretical
interdependence level among team members, goal choice and goal prediction, by showing that goals affect task performance through
striving are likely to be tightly coupled with team transition and allocation of on-task effort (e.g., Locke & Latham, 1990; Wein-
MULTILEVEL TRAINING TRANSFER 831

gart, 1992). Moreover, according to Wood and Locke (1990), the However, at the individual level, whereas knowledge and skill can
direct relationship between goal setting and performance dimin- enable individuals to regulate more effectively when performing
ishes as task complexity increases. Thus, given that the transfer complex tasks (e.g., Kanfer & Ackerman, 1989), they also enable
task in the study was quite complex, it was more likely the effects individuals to accomplish complex tasks more efficiently and even
of between-episodes regulation processes (which involve setting automatically without requiring much conscious cognitive effort
and prioritizing goals) would be entirely through within-episode (e.g., Anderson, 1983; Smith et al., 1997). Thus, although knowl-
processes (which involve the actual execution and possible adap- edge and skill are expected to contribute to adaptive performance
tation or modification of goals and strategies). at both the individual and the team levels, they do so more directly
at the individual level than the team level. Indeed, individual-level
Hypothesis 1: At both the individual and team levels, post- research has found that knowledge and skill relate more strongly
training within-episode regulation processes fully mediate the and directly to adaptive performance (e.g., Ford et al., 1998;
relationships between posttraining between-episodes regula- Kozlowski et al., 2001) than similar team-level research has found
tion processes and posttraining adaptive performance. (Marks et al., 2000, 2002; Mathieu et al., 2000). We therefore
expected that knowledge and skill would relate to adaptive per-
The hypothesized model also suggests that self- and collective formance both directly and indirectly (through self-regulation pro-
efficacy beliefs relate similarly to adaptive performance across cesses) at the individual level, but only indirectly (through team
levels. This is consistent with a study conducted by Chen et al. processes) at the team level.
(2002) and with meta-analyses that found that the average (“true”)
correlation between self-efficacy and individual performance (␳ ⫽ Hypothesis 3a: At the individual level, posttraining between-
.38; Stajkovic & Luthans, 1998) is highly similar to the true episodes and within-episode regulation processes partially
collective efficacy–team performance correlation (␳ ⫽ .39; Gully, mediate the relationships of role knowledge and individual
Incalcaterra, Joshi, & Beaubien, 2002). Extending previous work, skill acquired during training with posttraining adaptive
however, one would expect that efficacy beliefs would predict performance.
adaptive performance through the allocation of task-related effort
both prior to and during transfer performance periods. Bandura Hypothesis 3b: At the team level, posttraining between-
(1997) reviewed research that showed that more efficacious indi- episodes and within-episode regulation processes fully medi-
viduals and groups tend to engage in more effective planning (i.e., ate the relationships of team knowledge and team skill ac-
goal choice and transition processes) and to allocate and persist in quired during training with posttraining adaptive performance.
more on-task effort when pursuing goals (i.e., goal striving and
action processes). In the training and transfer literature, several Method
authors have proposed that boosting trainees’ task-related efficacy Sample
beliefs can help ensure that trainees choose to use and generalize
We recruited 156 undergraduates, who performed in 78 two-person
the knowledge and skill acquired in training during transfer peri-
teams, from a large Southeastern university. The sample was 74% men and
ods (e.g., Ford et al., 1998; Kozlowski et al., 2001). 26% women, with an average age of 20 years (SD ⫽ 1.56). All participants
received extra course credit. To motivate transfer and adaptive perfor-
Hypothesis 2: At both the individual and team levels, post- mance, members of the three highest scoring teams each received prizes
training between-episodes and within-episode regulation pro- ($50 gift certificates for members of the best-scoring team, $25 certificates
cesses fully mediate the relationships between efficacy beliefs for members of the second-best-scoring team, and $15 certificates for
acquired during training and posttraining adaptive performance. members of the third-best-scoring team).

Finally, at both levels of analysis, knowledge and skills are Task Apparatus
likely to positively predict regulation processes, at least in part
because they are positively associated with efficacy beliefs (e.g., We used flight simulator task (Longbow2, 1997) based on a low-fidelity
Bandura, 1997; Ford et al., 1998; Kozlowski et al., 2001). Specif- personal computer (PC) as our research platform. The software is a
ically, trainees and teams who acquire higher levels of knowledge computer-generated simulation program of the Apache attack helicopter.
The hardware included an IBM compatible PC equipped with a joystick
and skill during training are likely to be more efficacious and to
and video and audio recording equipment. A physical divider separated
have a stronger basis on which they can plan for and accomplish players, and they communicated via microphone-equipped headsets. This
effectively complex transfer tasks (e.g., Smith et al., 1997). task environment is similar to that used by Marks et al. (2002; Experiment
However, we expected that regulation processes would be more 1). However, their task was a three-player simulation, whereas our task was
likely to mediate the influences of team knowledge and skill on a two-player simulation consisting of a pilot and a gunner. In each team,
team adaptive performance than the influences of individual the pilot was responsible for (a) flying the helicopter, (b) firing weapons,
knowledge and skill on individual adaptive performances. Specif- and (c) escaping enemy antiaircraft fire, whereas the gunner’s roles in-
ically, team knowledge and skill cannot translate into team adap- cluded (a) radar surveillance (e.g., identification of enemy or friendly
targets), (b) weapon selection and management, and (c) system monitoring
tive performance unless they enable team members to collectively
(e.g., monitoring helicopter damage). The task required teams to attack
engage in effective team transition and action processes (cf. Hack-
primary and secondary enemy targets (e.g., tanks, armored vehicles) posi-
man, 1987). According to Kozlowski et al. (1999), more adaptive tioned along prespecified waypoints while avoiding antiaircraft artillery
teams rely on their knowledge and skill to appropriately adjust and protecting friendly targets. The team task was highly interdependent.
their interaction patterns (e.g., their coordination and communica- For instance, to destroy enemy tanks the gunner had to identify the tank
tion patterns) to meet novel and complex environmental demands. correctly as a primary or secondary target, select the appropriate weapon,
832 CHEN, THOMAS, AND WALLACE

and aim the weapon such that the pilot could fire it accurately, whereas the dent and consistent with a conjunctive team task (Steiner, 1972) in that an
pilot had to appropriately position the helicopter and fire the weapon at the unknowledgeable member could have greatly (and adversely) affected
most appropriate time. team performance (cf. Beersma et al., 2003; Moynihan & Peterson, 2001).
Efficacy beliefs. On the basis of a task analysis of our team simulation,
we identified nine key pilot tasks (e.g., maintaining optimal altitude, using
Procedure chain gun), eight key gunner tasks (e.g., identifying and differentiating
Sessions lasted approximately 3 hr and were conducted in three phases: targets, selecting weapons), and six key collective tasks for the team as
(a) role and team training, (b) a training evaluation session, and (c) a whole (e.g., following waypoints, positioning the helicopter for targeting).
transfer-of-training mission. Initially, participants were guided through a Four SMEs reviewed these tasks and sorted the tasks as being either
90-min training program, which followed Kozlowski et al.’s (1999) theory “primarily a pilot task,” “primarily a gunner task,” or “primarily a team
of team compilation. Specifically, participants first received an overview of task.” These tasks were then used in the self-efficacy and collective
the task and watched an automated demonstration of the simulation. They efficacy measures. The pilot and gunner self-efficacy measures asked them
were then guided through a structured hands-on training program that to “rate how confident you are in your ability to successfully and consis-
progressed from individual role training (i.e., pilot or gunner) to team tently accomplish each of the following [pilot or gunner] tasks” (1 ⫽ not
training. Throughout the various stages of training, we ensured that par- at all confident, 5 ⫽ extremely confident). The collective efficacy measure
ticipants could accomplish all critical individual and team tasks, which asked both the pilot and gunner to “rate how confident your team is in its
were chosen on the basis of a thorough task analysis (cf. Tesluk, Zaccaro, ability to successfully and consistently accomplish each of the following
Marks, & Mathieu, 1997). When participants failed to accomplish a par- team tasks” (1 ⫽ not at all confident, 5 ⫽ extremely confident). The
ticular task, the experimenters coached them until they were able to internal consistency reliability (alpha) estimates were .57 for pilot self-
accomplish it before moving to training the next task. Although this efficacy, .74 for gunner self-efficacy, and .72 for collective efficacy. The
procedure ensured the participants reached a minimal level of task com- rather low reliability estimates were expected given the heterogeneous
petency needed for them to perform their individual roles and the team nature of the tasks; however, the low reliability may have attenuated
task, there were differences in level of individual and team task mastery, relationships involving efficacy beliefs.
which we captured using measures of individual and team skills. The For each team, average collective efficacy responses of the pilot and the
training phase ended with teams playing a practice team mission, on the gunner were aggregated to the team level. Justifying the aggregation of this
basis of which individual and team skills were recorded. referent-shift compositional model measure, we found that intermember
At the conclusion of training, participants completed a survey containing agreement—average rwg(j), using a uniform expected variance distribu-
various training evaluation measures, including role-specific knowledge
tion—was .88, indicating that members shared their collective efficacy
tests and self- and collective efficacy measures. Finally, following a short
beliefs (James, Demaree, & Wolf, 1984). Moreover, intraclass correlations
break, teams completed a transfer performance mission. The transfer
(ICCs) provided evidence for sufficient intermember reliability, ICC(1) ⫽
mission began by providing the pilot and gunner two separate “intelligence
.49, ICC(2) ⫽ .66; F(77, 78) ⫽ 2.95, p ⬍ .05. ICC(1) indicates the
reports,” which contained complementary pieces of information regarding
percentage of variance in ratings due to team membership, whereas ICC(2)
the mission (e.g., location and type of various enemy and friendly targets,
indicates the reliability of differences between team means (Bliese, 2000).
timing of movement of the various targets during the mission, information
about terrain and antiaircraft artillery.). Using the reports, teams were Skill and adaptive performance. On the basis of the task analysis we
instructed to take 10 min to plan for their mission. At the conclusion of the conducted, we developed Behaviorally Anchored Rating Scales (BARS)
planning session, we measured individual goal choice activities and team for each position (pilot and gunner; 1 ⫽ hardly any skill, 5 ⫽ complete
transition processes. Following the administration of these short process skill). The pilot BARS included the following dimensions: (a) flying the
measures, the teams “flew” their 15-min transfer mission, at the end of helicopter, (b) firing weapons, and (c) escaping enemy antiaircraft fire. The
which we obtained measures of individual goal striving activities, team gunner BARS consisted of: (a) radar surveillance, (b) weapon selection and
action processes, and individual and team adaptive performance. It is management, and (c) system monitoring. These BARS measures were
important to note that teams were trained and practiced in the same similar to those used in Chen et al. (2002, Lab Sample). Three SMEs were
environment throughout training and that environment was substantially trained to use the BARS and reached high levels of agreement, average
less complex than the transfer performance environment—it included rwg(j) ⫽ .94 and .89) and adequate levels of reliability for pilot and gunner
fewer and less aggressive antiaircraft artillery, fewer targets, shorter way- ratings, respectively, ICC(1) ⫽ .74, .73, ICC(2) ⫽ .90, .89; F(4, 10) ⫽
point routes, and easier terrain. 9.73, 9.20, p ⬍ .05. Besides the five study sessions used to evaluate
interrater agreement and reliability, one SME rated pilot and gunner
performance in each study session. Confirmatory factor analyses (n ⫽ 78),1
Measures conducted separately on ratings rendered following the practice and trans-
Knowledge. Separate 10-item multiple-choice tests were developed to fer missions, confirmed two dimensions corresponding to pilot perfor-
assess pilot and gunner role knowledge. Consistent with Ford et al. (1998) mance (␣s ⫽ .87 and .85 for the practice and transfer missions) and gunner
and Kozlowski et al. (2001), items on these tests covered both declarative performance (␣s ⫽ .75 and .78 for the practice and transfer missions).
knowledge (i.e., knowledge about key aspects of their roles and equipment) Performance ratings following the practice mission were used as individual
and procedural and strategic knowledge (i.e., knowledge about how best to skill measures, whereas ratings obtained following the transfer mission
approach certain role-related situations, such as which weapon the gunner were used as measures of individual adaptive performance.
should select when approaching certain targets or what tactics the pilot
should use when eliminating specific targets). The pilot and gunner com-
1
pleted only the test for their respective role. Four subject matter experts All confirmatory factor analyses we report herein were conducted
(SMEs) reviewed an initial version of the test, and based on their com- using LISREL 8 (Jöreskog & Sörbom, 1993). In each analysis, the fit of a
ments, the tests were modified by the authors and later judged by the SMEs theoretical measurement model (e.g., a two-factor model) was compared
to capture the main content domains of the roles. Participants were given with the fit of a single-factor alternative model. Support for the measures
1 point for each correct answer; scores on the tests could range between 0 was based on finding that the theoretical model fit the data well and
and 10. We aggregated knowledge to the team level using the minimum significantly better than the alternative models as well as on detecting high
(i.e., worst) individual score in teams, as the task was highly interdepen- (⬎.70) and significant factor loadings.
MULTILEVEL TRAINING TRANSFER 833

Team scores on the practice mission were used to index team skill, important to ensure the estimates obtained in the analyses were compara-
whereas team scores on the transfer mission measured team adaptive ble. Therefore, because different metrics were used in some of the
performance. In both practice and transfer missions, team performance individual- and team-level measures, we standardized all measures within
score was based on the extent to which the team (a) eliminated primary their respective levels, such that parameter estimates essentially reflected
enemy targets (worth 20 points each), (b) eliminated secondary enemy comparable standardized (i.e., “beta”) coefficients (see Chen et al., 2002, in
targets (worth 10 points each), (c) remained alive (scored as two times the press).
actual minutes the team survived during the mission), (d) remained un-
damaged at the end of the mission (undamaged and alive ⫽ 30 points, Results
damaged and alive ⫽ 20 points, destroyed ⫽ 0 points), and (e) avoided
elimination of friendly targets (20 points were subtracted for each elimi- The individual- and team-level correlations are provided in
nated friendly target). Scores were obtained from a computer screen Table 1. As expected, skill measures were related to efficacy
following the mission; maximum possible score was 260 points in the beliefs at both the individual and team levels. However, whereas
practice mission and 410 points in the transfer mission. role knowledge significantly related to individual skill, team
Regulation processes. We developed five items to measure individual
knowledge did not relate to team skill. Also, although knowledge
goal choice activities and 10 items to measure individual goal striving
activities (the items are listed in the Appendix). The items were based on
related similarly to efficacy beliefs at both levels, the relationship
our conceptualization of goal choice and goal striving as individual activ- was significant at the individual level, but not at the team level. In
ities directed at assisting the team to choose and execute its goals and use addition, the correlations of efficacy beliefs with adaptive perfor-
examples of transition and action activities provided by Marks et al. mance were highly similar across levels, whereas the correlations
(2001). Following the planning mission and prior to performing the transfer of knowledge and skill with adaptive performance were substan-
mission, we presented participants with the goal choice items and asked tially larger at the individual level than at the team level. Finally,
them to “rate the extent to which you personally engaged in each of the the correlations of within-episode (i.e., goal striving and action)
following behaviors during this last planning session,” and following the processes with adaptive performance were significant, whereas the
transfer mission, we gave them the goal striving items and asked them to correlations of between- episodes processes (goal choice and tran-
“rate the extent to which you personally engaged in each of the following
sition processes) with adaptive performance were lower and only
behaviors during this last flight mission” (1 ⫽ never, 5 ⫽ constantly).
Confirmatory factor analyses in LISREL (n ⫽ 156) confirmed that the
significant at the individual level.
items yielded two factors corresponding to goal choice (␣ ⫽ .73) and goal
striving (␣ ⫽ .81). Hypotheses Tests
Using BARS developed by Cobb and Mathieu (2003), SMEs rated the
extent to which the teams effectively executed the three transition pro- We first examined the extent to which the independent variables
cesses (i.e., mission analysis, goal specification, and strategy formulation (i.e., training outcomes) predicted the mediators (i.e., posttraining
and planning) and four action processes (i.e., monitoring progress toward regulation processes). Table 2 summarizes the individual-level
goals, systems monitoring, team monitoring and backup behavior, and random coefficient modeling analyses of goal choice and team-
coordination) delineated by Marks et al. (2001; 1 ⫽ hardly any skill, 5 ⫽ level ordinary least squares regression analyses of transition pro-
complete skill). In all but five study sessions, one of three SMEs rated cesses. Results indicated that self-efficacy only marginally ( p ⬍
transition and action processes. Data obtained in five study sessions in .10) predicted goal choice activities, whereas collective efficacy
which all three SMEs rated the team processes indicated sufficient levels of
significantly predicted transition processes, and knowledge and
interrater agreement and reliability. Specifically, for transition processes
skill did not predict goal choice activities or transition processes.
ratings, average rwg(j) ⫽ .89, ICC(1) ⫽ .33; F(4, 10) ⫽ 2.50, p ⫽ .10, and
ICC(2) ⫽ .60. For action processes ratings, average rwg(j) ⫽ .97, ICC(1) ⫽ However, when regressing goal choice activities only on self-
.72; F(4, 10) ⫽ 8.65, p ⬍ .05, and ICC(2) ⫽ .88. Confirmatory factor efficacy, self-efficacy was a significant predictor, estimate ⫽ .18,
analyses in LISREL (n ⫽ 78) supported the distinctiveness of transition t(77) ⫽ 2.30, p ⬍ .05. Thus, between-episodes processes were
processes (␣ ⫽ .89) from action processes (␣ ⫽ .89). Further supporting only likely to mediate the relationships between efficacy beliefs
the validity of the regulation measures, average reports of individual goal and adaptive performance.
choice activities in teams correlated .34 ( p ⬍ .05) with SME ratings of Results from individual- and team-level analyses of within-
team transition processes, and average reports of individual goal striving episode processes are summarized in Table 3. The individual-level
activities in teams correlated .38 ( p ⬍ .05) with SME ratings of team action analyses revealed that self-efficacy and goal choice activities, but
processes.
not role knowledge and individual skill, uniquely and positively
predicted goal striving activities. Also, the introduction of goal
Analyses choice into the equation in a second step reduced only slightly the
self-efficacy estimate. The first step of the team-level analyses
Analyses were conducted using the Nonlinear and Linear Mixed Effects showed that collective efficacy and team knowledge, but not team
program for S-PLUS and R (Pinheiro & Bates, 2000). We estimated skill, uniquely and positively predicted action processes. The ad-
individual-level relationships using random coefficient modeling, which dition of transition processes at the second step reduced the influ-
provides the correct parameter estimates and significance tests for the ences of both team knowledge and collective efficacy on action
nested individual-level data in our study (Bliese, 2002), and team-level processes, but collective efficacy remained a significant predictor.
relationships using ordinary least squares regression. At each level, we
Thus, the results summarized in Tables 2 and 3 suggest that, at
conducted hierarchical mediated regression analyses, following the ap-
proach proposed by Kenny, Kashy, and Bolger (1998). We also tested
both levels of analysis, (a) efficacy beliefs predicted within-
mediated (indirect) effects using Sobel’s (1982) test, which MacKinnon, episode processes more strongly than between-episodes processes,
Lockwood, Hoffman, West, and Sheets (2002) and Shrout and Bolger (b) between-episodes processes predicted within-episode pro-
(2002) suggest provides a more powerful test for mediation. Given that we cesses, (c) between-episodes processes partially mediated the in-
were interested in comparing individual- and team-level results, it was fluences of efficacy beliefs on within-episode processes, and (d)
834 CHEN, THOMAS, AND WALLACE

Table 1
Individual-Level and Team-Level Descriptive Statistics and Correlations

Variable M SD 1 2 3 4 5 6

Individual level (n ⫽ 156)


a

1. Role knowledge 0.73 0.14 —


2. Self-efficacy 3.64 0.53 .14* —
3. Individual skill 2.78 0.88 .27** .20** —
4. Goal choice activities 3.35 0.69 .09 .14* ⫺.03 —
5. Goal striving activities 3.45 0.63 .07 .45** .17* .33** —
6. Individual adaptive performance 2.78 0.88 .41** .30** .62** .13* .31** —
Team level (n ⫽ 78)
1. Team knowledge 0.66 0.13 —
2. Collective efficacy 3.53 0.52 .15 —
3. Team skill 110.79 84.91 .22 .61** —
4. Transition processes 3.27 0.90 .19 .34** .20 —
5. Action processes 3.09 0.80 .27* .46** .20 .51** —
6. Team adaptive performance 162.05 90.86 .20 .34** .19 .10 .51** —
a
Within-team (individual-level) correlations are reported.
* p ⬍ .05. ** p ⬍ .01.

knowledge and skills generally did not predict posttraining regu- .05, and, likewise, the indirect effect of transition processes on
lation processes (with the exception of a marginal influence of team performance through action processes (.20) was statistically
team knowledge on team action processes). significant, z ⫽ 2.72, p ⬍ .05. Thus, these results provided support
Recall that Hypothesis 1 suggests that within-episode processes for Hypothesis 1.
mediate the effects of between-episodes processes on performance Hypothesis 2 states that efficacy beliefs relate to adaptive per-
at both levels of analysis. Table 4 provides the individual- and formance through between-episodes and within-episode processes
team-level analyses of adaptive performance. In the second step of at both levels. As shown in Table 4, efficacy beliefs related to
the analyses, neither goal choice activities nor transition processes performance over and above knowledge and skill at the team but
predicted performance. Further, at both levels, within-episode reg- not the individual level. Table 4 also suggests that the unique
ulation processes (i.e., goal striving activities and action processes) influences of collective efficacy beliefs on performance dropped
positively predicted adaptive performance (see Step 3 in the anal-
once introducing the action (within-episode) regulation processes
yses). The significant negative influence of transition processes on
(Step 3), but not when introducing transition (between-episodes)
team adaptive performance in the final step of the team-level
processes (Step 2). These results suggest that, at the team level,
analyses seems to be due to suppression given transition processes
within-episode processes, more so than between-episodes pro-
related to action processes but not team performance (see Table 1).
Further, Sobel’s (1982) test indicated that the indirect effect of cesses, mediated the relationships between collective efficacy be-
goal choice activities on individual performance through goal liefs and performance.
striving activities (.05) was statistically significant, z ⫽ 2.21, p ⬍ Although the results summarized in Table 4 found evidence that
within-episode regulation processes mediated the efficacy—per-
formance relationship at the team but not the individual level, there
Table 2 were significant indirect effects for efficacy beliefs on adaptive
Individual-Level and Team-Level Analyses of Between-Episodes performance through within-episode processes at both the individ-
Processes ual level (indirect effect of self-efficacy through goal striving
activities ⫽ .04, z ⫽ 2.05, p ⬍ .05) and the team level (indirect
Parameter effect of collective efficacy through action processes ⫽ .22, z ⫽
Level and model estimate SE df t R2
2.59, p ⬍ .05). In contrast, the indirect effects of efficacy beliefs
Individual-level resultsa on adaptive performance through between-episodes processes
Role knowledge .05 .08 75 0.58 were nonsignificant at both the individual level (indirect effect of
Self-efficacy .15 .08 75 1.76
Individual skill .09 .09 75 1.05 .00
self-efficacy through goal choice activities ⫽ .01, z ⫽ 1.29, ns)
Team-level resultsb and the team level (indirect effect of collective efficacy through
Team knowledge .15 .11 74 1.33 transition processes ⫽ ⫺.01, z ⫽ ⫺0.31, ns). The apparent incon-
Collective efficacy .35* .14 74 2.53 sistent findings summarized above with respect to the mediated
Team skill ⫺.04 .14 74 ⫺0.29 .14*
efficacy—performance relationship at the individual level are in
Note. Individual-level results are from a random coefficient modeling line with Monte Carlo studies conducted by MacKinnon et al.
analysis (n ⫽ 156); team-level results are from an ordinary least squares (2002), which showed that tests of indirect effects are more pow-
regression analysis (n ⫽ 78).
a erful in detecting mediation than Kenny et al.’s (1998) hierarchical
Dependent variable is goal choice activities. b Dependent variable is
transition processes. regression mediation tests. In sum, these results partially supported
* p ⬍ .05. ** p ⬍ .01. Hypothesis 2, as the relationships of efficacy beliefs with adaptive
MULTILEVEL TRAINING TRANSFER 835

Table 3
Individual-Level and Team-Level Analyses of Within-Episode Processes

Parameter
Level and step estimate SE df t R2 ⌬R2

Individual-level resultsa
Step 1
Role knowledge .02 .08 75 0.29
Self-efficacy .33** .08 75 4.27
Individual skill .09 .09 75 1.08 .18** .18**
Step 2
Role knowledge .01 .07 74 0.10
Self-efficacy .27** .07 74 3.71
Individual skill .06 .08 74 0.79
Goal choice activities .36** .07 74 5.01 .24** .06**
Team-level resultsb
Step 1
Team knowledge .23* .10 74 2.24
Collective efficacy .54** .13 74 4.28
Team skill ⫺.18 .13 74 ⫺1.45 .27** .27**
Step 2
Team knowledge .17 .10 73 1.83
Collective efficacy .41** .12 73 3.39
Team skill ⫺.17 .12 73 ⫺1.45
Transition processes .37** .10 73 3.78 .39** .12**

Note. Individual-level results are from a random coefficient modeling analysis (n ⫽ 156); team-level results
are from an ordinary least squares regression analysis (n ⫽ 78).
a
Dependent variable is goal striving activities. b Dependent variable is action processes.
* p ⬍ .05. ** p ⬍ .01.

performance were mediated through within-episode processes but turn related to team adaptive performance through team action
not between-episodes processes, and did so more clearly at the processes. Indeed, the indirect effect of team skill on action pro-
team than the individual level of analysis. cesses through collective efficacy (.25) was significant, z ⫽ 3.02,
Hypothesis 3a states that, at the individual-level of analysis, p ⬍ .05, and the indirect effect of team skill on team performance
goal choice and goal striving only partially mediate the relation- through collective efficacy and action processes (.16) was margin-
ships of role knowledge and individual skill with adaptive perfor- ally significant, z ⫽ 1.68, p ⬍ .10. Thus, although not hypothe-
mance. As shown in Tables 2 and 3, individual knowledge and sized, team skill had an even less direct relationship with team
skill did not predict the individual regulation processes, precluding performance through collective efficacy and team action
possible indirect effects. However, role knowledge and individual processes.
skill uniquely and positively predicted individual adaptive perfor-
mance, even when including self-regulation processes in the model
Discussion
(see Table 4). These results provided partial support for Hypoth-
esis 3a: Role knowledge and individual skill directly and uniquely This is the first study attempting to test the relationships among
predicted adaptive performance, but self-regulation processes did multiple training outcomes, posttraining regulation processes, and
not partially mediate these relationships. adaptive performance at the individual and team levels of analysis
Finally, according to Hypothesis 3b, team knowledge and team simultaneously. Figure 2 provides a summary of the results across
skill relate to team adaptive performance entirely through all analyses. The results indicated both similar and different
between-episodes and within-episode team processes. As ex- patterns of relationships across levels of analysis, with impor-
pected, team knowledge and skill did not directly predict team tant implications for the training, motivation, and work team
performance (see Table 4, Step 1). However, team knowledge did literatures.
not relate to team transition processes and only marginally ( p ⬍
.10) related to team action processes, and team skill did not relate
to either team transition or action processes (Tables 2 & 3). Also, Theoretical and Practical Implications
the indirect effect of team knowledge on team performance In line with Kraiger et al. (1993) and Salas and Cannon-Bowers
through team action processes (.12) was statistically significant, (1997), our results support the notion that multiple aspects of
z ⫽ 1.97, p ⬍ .05. Thus, Hypothesis 3b was only weakly sup- learning are critical at both the individual and team levels. As such,
ported, in that team action processes mediated the team training programs directed at developing adaptive teams should
knowledge–team performance relationship. ensure that both individual team members and the team as a whole
acquire the knowledge, skills, and efficacy beliefs that enable
Additional Analyses adaptability. However, the relative importance of training out-
Although team skill did not directly relate to any of the pro- comes differs across levels. Specifically, knowledge and skill
cesses, it did relate significantly to collective efficacy, which in directly enable individual adaptive performance but, in accordance
836 CHEN, THOMAS, AND WALLACE

Table 4
Individual-Level and Team-Level Analyses of Adaptive Performance

Parameter
Level and step estimate SE df t R2 ⌬R2

Individual-level resultsa
Step 1
Role knowledge .15** .05 75 2.88
Self-efficacy .08 .05 75 1.52
Individual skill .57** .07 75 8.62 .46** .46**
Step 2
Role knowledge .14* .05 74 2.76
Self-efficacy .06 .05 74 1.23
Individual skill .57** .07 74 8.67
Goal choice activities .10 .05 74 1.92 .46** .00
Step 3
Role knowledge .14* .05 73 2.82
Self-efficacy .02 .05 73 0.30
Individual skill .55** .06 73 8.57
Goal choice activities .05 .05 73 0.98
Goal striving activities .15* .06 73 2.50 .47** .01
Team-level resultsb
Step 1
Team knowledge .16 .11 74 1.41
Collective efficacy .35* .14 74 2.57
Team skill ⫺.06 .14 74 ⫺0.46 .14* .14*
Step 2
Team knowledge .16 .11 73 1.44
Collective efficacy .36* .14 73 2.54
Team skill ⫺.06 .14 73 ⫺0.47
Transition processes ⫺.04 .12 73 ⫺0.34 .14* .00
Step 3
Team knowledge .07 .10 72 0.66
Collective efficacy .15 .14 72 1.04
Team skill .03 .13 72 0.20
Transition processes ⫺.24* .12 72 ⫺2.08
Action processes .54** .13 72 4.30 .31** .17**

Note. Individual-level results are from a random coefficient modeling analysis (n ⫽ 156); team-level results
are from an ordinary least squares regression analysis (n ⫽ 78).
a
Dependent variable is individual performance. b Dependent variable is team performance.
* p ⬍ .05. ** p ⬍ .01.

with the IPO framework, more weakly and indirectly enable team longer period of time, allowing the teams to master their task
adaptability, through their relationships with collective efficacy better. Second, we used different measurement approaches to
and team regulation processes. In contrast, whereas the bivariate capture knowledge, skill, and adaptive performance across levels,
relationship of posttraining efficacy beliefs with adaptive perfor- and these differences might have accounted (at least in part) for the
mance was similar across levels, the importance of efficacy beliefs different findings across levels. However, it is worth noting that
relative to knowledge and skill was greater at the team level of several similar relationships were still detected with other con-
analysis, in which collective efficacy strongly related to posttrain- structs (e.g., efficacy beliefs, within-episode processes, and per-
ing regulation processes and helped link team skill to team regu- formance) that were measured differently across levels. Undoubt-
lation processes. edly, more research is needed to test whether knowledge and skill
The inconsistent findings across levels pertaining to the influ- relate differently to adaptive performance across levels irrespec-
ences of knowledge and skill on adaptive performance were ex- tive of the level of task mastery and measurement approaches.
pected on the basis of previous team and training literatures. The results also extended theories of motivation. Specifically,
However, there are two additional plausible explanations for the consistent with DeShon et al. (2004), there were similarities in the
different findings detected across levels. First, it is possible that the relationships among efficacy beliefs, regulation processes, and
level of team task mastery at the end of training was inferior to the performance across levels, despite the use of vastly different
level of individual task mastery, which resulted in the lower measurement approaches at each level. At both levels, efficacy
team-level relationships between the knowledge and skill mea- beliefs and between-episodes processes related to adaptive perfor-
sures and team adaptive performance (cf. Smith et al., 1997). mance through within-episode regulation processes. These results
Indeed, all else being equal, it takes longer to compile expert teams are consistent with social– cognitive theory (Bandura, 1997) and
than expert individuals (Kozlowski et al., 1999), and so results goal setting theory (Locke & Latham, 1990), according to which
could have differed if the training program was spread over a efficacy beliefs, goals, and strategies translate into performance
MULTILEVEL TRAINING TRANSFER 837

Figure 2. Summary of results from hypotheses tests. Arrows indicate hypothesized paths supported in the study.

through effort and persistence at both the individual and team vational processes as motivational interventions directed at teams.
levels. However, it is still unclear whether the nature of relation- In particular, boosting self-efficacy beliefs is likely to improve
ships among efficacy beliefs, effort, and performance differ over individual motivation and performance much like boosting teams’
time, as suggested by recent research conducted by Vancouver, collective efficacy is likely to improve team motivation and per-
Thompson, and Williams (2000). For instance, it is possible that formance. In addition, the finding that posttraining motivational
over time successful individuals and teams become overconfident, processes helped explain (at least in part) the relationships between
and therefore their efficacy beliefs no longer lead to higher sub- training outcomes and adaptive performance suggests that post-
sequent effort and performance. training interventions directed at facilitating training transfer in
Nonetheless, in line with DeShon et al. (2004) and Marks et al. teams should attempt to impact these processes. For instance, to
(2001), the results provide initial support for the generalizability of facilitate adaptive performance, team leaders should ensure that
the goal choice and goal striving concepts to the team level of members of their team engage in effective regulation processes
analysis, thus allowing greater integration between theories of (e.g., by helping the team develop and monitor the attainment of
individual and team motivation. It is important to note that DeShon performance strategies), particularly when the team performs in
et al. (2004) found similar relationships among self- and team complex and dynamic environments (see Kozlowski et al., 1996).
regulation processes during training, whereas our study replicated However, factors contributing to self-efficacy may not necessarily
their findings in a posttraining environment. In that sense, these affect collective efficacy, and vice versa. Indeed, Chen and Bliese
two studies have taken initial steps toward redressing the paucity (2002) found that leadership climate more strongly predicts col-
of multilevel research on motivation in and of teams (cf. Kozlow- lective efficacy than self-efficacy, whereas experience, psycholog-
ski & Bell, 2003). However, additional theory development and ical strain, and role clarity more strongly predict self-efficacy than
research are needed to uncover whether and why motivational collective efficacy. As such, researchers should strive to uncover
processes are similar or different across levels. For instance, fol- both unique and similar means of enabling teams and individual
lowing Kozlowski et al.’s (1999) theory of team compilation, it team members to regulate effectively in complex and dynamic
would be interesting to study whether, as work teams develop, the environments.
focus of self-regulated activities actually shifts from individual to
team processes and outcomes, and whether individually focused Limitations and Future Research
versus team-focused regulation activities contribute to individual
effectiveness differently at different stages of team development Several limitations of this research should be noted. First, this
(e.g., in learning vs. transfer). study was conducted in a laboratory, which may limit the gener-
Practically speaking, these results suggest that motivational alizability of the findings. Furthermore, the results may not gen-
interventions directed at individuals should target the same moti- eralize to teams with more than two members. Generalizability is
838 CHEN, THOMAS, AND WALLACE

also limited in light of the focus on one particular type of team can claim measures in this study were isomorphic, and therefore
(i.e., the action team) and given that we only focused on one we could not appropriately test whether relationships were homol-
transfer performance episode, which was close in time to the end ogous across levels of analysis (cf. Chen et al., 2002, in press).
of training (i.e., near transfer). However, our aim was primarily to Clearly, more research is needed to uncover whether different
test a theoretical multilevel nomological network, as opposed to measurement approaches can capture the same construct across
modeling an actual real-world environment (Driskell & Salas, levels. For example, researchers should examine whether measur-
1992). Furthermore, following Marks (2000) recommendations, ing regulation processes using self-report measures at both levels
we took several steps to increase the realism and external validity (e.g., as done in DeShon et al.’s, 2004, study) yields different
of our study. Specifically, the simulation we used mimicked the results as compared with measuring processes using self-report at
reality of actual flight teams by (a) ensuring that participants the individual level and expert ratings at the team level (as was
actually believed they were a part of an interdependent team, (b) done in the present study). We believe tests of homology is a new
designing positions and interpositional relationships that closely and exciting line of multilevel research, but the conceptual and
matched real-world flight teams, and (c) creating an environment measurement preconditions that need to be met before testing such
that required the same theoretical processes that drive effective- multilevel models are still evolving (Chen et al., in press, 2004).
ness in real-world settings. Moreover, the transfer episode was Meanwhile, along with additional multilevel research in the areas
developed to capture key theoretical aspects underlying adaptive of adaptability, motivation, and training, researchers need to ad-
performance in work environments. vance stronger theories regarding whether constructs and relation-
Second, although we captured cognitive, affective–motivational, ships among them are similar or different across levels.
and behavioral aspects of learning, we did not measure all possible Future research should attempt to replicate our findings using
aspects of these important training outcomes. In particular, we did different types of work teams (e.g., project, production, or man-
not include measures of knowledge structures, or mental models, agement teams) of different sizes and in different contexts. It
which have been proposed to be important training outcomes would be interesting to test, for instance, whether training out-
driving individual and team adaptive performance (e.g., Kozlowski comes and regulation processes account similarly for individual
et al., 2001; Kraiger et al., 1993; Marks et al., 2000, 2002; Mathieu and team adaptive performance in contexts of varying degrees of
et al., 2000). However, researchers have focused on somewhat complexity, novelty, and expertise. For instance, Kozlowski et al.
different mental models constructs at the individual and team (2001) showed that self-efficacy is more predictive of individual
levels of analysis. At the individual level, mental models involve performance in complex, as opposed to simple, task environments.
knowledge coherence, or the extent to which the relatedness Moreover, it is important to test whether our findings generalize to
among pairs of concepts reflects a consistent structure (e.g., Koz- teams with higher levels of expertise, who perform more complex
lowski et al., 2001). On the other hand, team mental models reflect and ill-defined tasks that require higher levels of adaptability than
the intermember similarity or sharedenss in knowledge structures the level examined in this research (see Smith et al., 1997), as well
(i.e., a dispersion construct; cf. Chan, 1998). Several authors have as to intact teams, such as Special Forces military teams or
noted that dispersion team-level constructs, such as shared or team top-management teams. Studying such teams requires a shift in
mental models, capture qualitatively different constructs from their focus from near to far transfer, because these teams perform more
individual-level analogues (Chen et al., 2004; Kozlowski & Klein, complex tasks than the one examined in our study, and do so over
2000). It is interesting though, despite the fact that we operation- repeated performance episodes beyond training (Sundstrom et al.,
alized team knowledge differently from the team mental models 1990). To ensure that they capture complex forms of adaptation,
literature (i.e., as an additive, as opposed to dispersion, composi- researchers can directly measure adaptability requirements (see
tional model), that the obtained relationships of team knowledge Pulakos et al., 2000). However, researchers should also consider
with action processes and team performance closely matched the adaptability requirements relative to the level of employee and
relationships obtained between team mental models and these team expertise, as higher levels of expertise can enable individuals
outcomes in previous research (cf. Marks et al., 2000, 2002; and teams to adapt more effectively to higher levels of task
Mathieu et al., 2000). complexity and novelty (Smith et al., 1997). Thus, lower levels of
Finally, the use of different measurement approaches to capture task complexity may require novices, but not experts, to be highly
the focal constructs across levels was a limitation, particularly with adaptive.
respect to our ability to appropriately test homology (i.e., equiva- Researchers should also examine whether leadership climate
lent relationships) across levels. For instance, some might argue and environmental complexity may impact the extent to which
that performance ratings used to capture individual adaptive per- certain training outcomes enable individuals and teams to be
formance may not capture the same construct as does objective adaptable (Kozlowski et al., 1996; Marks et al., 2000; Smith-
performance output used to measure team adaptive performance. Jentsch et al., 2001). Moreover, although we did not consider their
Likewise, it is unclear whether self-ratings of self-regulation pro- impact in this study, other researchers have suggested that indi-
cesses are equivalent to expert ratings of team regulation pro- vidual differences (e.g., openness to experience, ability) and team
cesses. On the one hand, the use of different measurement ap- composition of these individual differences also enable adaptabil-
proaches across levels helps eliminate common method as an ity at the individual and team levels of analysis (e.g., LePine, 2003;
explanation for the similarity in findings. Also, research has sug- LePine, Colquitt, & Erez, 2000). Such research should further
gested that objective and subjective measures of performance often examine the extent to which posttraining regulation processes help
yield convergence validity (Bommer, Johnson, Rich, Podsakoff, & explain why and how leadership, individual differences, team
MacKenzie, 1995; Wall et al., 2004). On the other hand, the use composition, and other inputs enhance or debilitate training trans-
different measurement approaches limited the extent to which we fer and whether the same factors affect the relationships between
MULTILEVEL TRAINING TRANSFER 839

training and transfer outcomes at the individual and team levels of R. W. Swezey & E. Salas (Eds.), Teams: Their training and performance
analysis. Ultimately, multilevel research that focuses on similari- (pp. 101–124). Norwood, NJ: Ablex.
ties and differences across levels of analysis can improve our Ford, J. K., Smith, E. M., Weissbein, D. A., Gully, S. M., & Salas, E.
understanding of the unique and similar factors and processes that (1998). Relationships of goal orientation, metacognitive activity, and
enable individual team members and teams as a whole to be highly practice strategies with learning outcomes and transfer. Journal of Ap-
plied Psychology, 83, 218 –233.
adaptive, as well as what managers, trainers, and consultants can
Gully, S. M., Incalcaterra, K. A., Joshi, A., & Beaubien, J. M. (2002). A
do to help them achieve it. meta-analysis of team-efficacy, potency, and performance: Interdepen-
dence and level of analysis as moderators of observed relationships.
References Journal of Applied Psychology, 87, 819 – 832.
Hackman, J. R. (1987). The design of work teams. In J. W. Lorsch (Ed.),
Anderson, J. R. (1983). The architecture of cognition. Cambridge, MA: Handbook of organizational behavior (pp. 315–342). Englewood Cliffs,
Harvard University Press. NJ: Prentice Hall.
Baldwin, T., & Ford, J. K. (1988). Transfer of training: A review and Ilgen, D. R., & Pulakos, E. D. (1999). Introduction: Employee performance
directions for future research. Personnel Psychology, 41, 63–105. in today’s organization. In D. R. Ilgen & E. D. Pulakos (Eds.), The
Bandura, A. (1997). Self-efficacy: The exercise of control. New York: changing nature of performance: Implications for staffing, motivation,
Freeman. and development (pp. 1–18). San Francisco: Jossey-Bass.
Beersma, B., Hollenbeck, J. R., Humphrey, S. E., Moon, H., Conlon, D. E., James, L. R., Demaree, R. G., & Wolf, G. (1984). Estimating within-group
& Ilgen, D. R. (2003). Cooperation, competition, and team performance: interrater reliability with and without response bias. Journal of Applied
Toward a contingency approach. Academy of Management Journal, 46,
Psychology, 69, 85–98.
572-590.
Jöreskog, K., & Sörbom, D. (1993). LISREL 8: Structural equation mod-
Bliese, P. D. (2000). Within-group agreement, non-independence, and
eling with the SIMPLIS command language. Hillsdale, NJ: Erlbaum.
reliability: Implications for data aggregation and analyses. In K. J. Klein
Kanfer, R. (1990). Motivation theory and industrial and organizational
& S. W. J. Kozlowski (Eds.), Multilevel theory, research, and methods
psychology. In M. D. Dunnette & L. M. Hough (Eds.), Handbook of
in organizations: Foundations, extensions, and new directions (pp. 349 –
industrial and organizational psychology (2nd ed., Vol. 1, pp. 75–170).
381). San Francisco: Jossey-Bass.
Palo Alto, CA: Consulting Psychologists Press.
Bliese, P. D. (2002). Using multilevel random coefficient modeling in
Kanfer, R., & Ackerman, P. L. (1989). Motivation and cognitive abilities:
organizational research. In F. Drasgow & N. Schmitt (Eds.), Advances in
An integrative/aptitude-treatment interaction approach to skill acquisi-
measurement and data analysis (pp. 401– 445). San Francisco: Jossey-
tion [Monograph]. Journal of Applied Psychology, 74, 657– 689.
Bass.
Kenny, D. A., Kashy, D. A., & Bolger, N. (1998). Data analysis in social
Bommer, W. H., Johnson, J., Rich, G. A., Podsakoff, P. M., & MacKenzie,
psychology. In D. T. Gilbert, S. T. Fiske, & G. Lindzey, The handbook
S. B. (1995). On the interchangeability of objective and subjective
of social psychology (4th ed., pp. 233–265). Boston: McGraw-Hill.
measures of employee performance: A meta-analysis. Personnel Psy-
Klimoski, R. J., & Mohammed, S. (1994). Team mental model: Construct
chology, 48, 587– 605.
or metaphor? Journal of Management, 20, 403– 437.
Chan, D. (1998). Functional relations among constructs in the same content
domain at different levels of analysis: A typology of composition mod- Kozlowski, S. W. J., & Bell, B. S. (2003). Work groups and teams in
els. Journal of Applied Psychology, 83, 234 –246. organizations. In W. C. Borman, D. R. Ilgen, & R. J. Klimoski (Eds.),
Chen, G., & Bliese, P. D. (2002). The role of different levels of leadership Comprehensive handbook of psychology: Vol. 12: Industrial and orga-
in predicting self and collective efficacy: Evidence for discontinuity. nizational psychology (pp. 333–375). New York: Wiley.
Journal of Applied Psychology, 87, 549 –556. Kozlowski, S. W. J., Gully, S. M., Brown, K. G., Salas, E., Smith, E. M.,
Chen, G., Bliese, P. D., & Mathieu, J. E. (in press). Conceptual framework & Nason, E. R. (2001). Effects of training goals and goal orientation
and statistical procedures for delineating and testing multilevel theories traits on multidimensional training outcomes and performance adapt-
of homology. Organizational Research Methods. ability. Organizational Behavior and Human Decision Processes, 85,
Chen, G., Mathieu, J. E., & Bliese, P. D. (2004). A framework for 1–31.
conducting multilevel construct validation. In F. J. Yammarino & F. Kozlowski, S. W. J., Gully, S. M., McHugh, P. P., Salas, E., & Cannon-
Dansereau (Eds.), Research in multilevel issues: Multilevel issues in Bowers, J. A. (1996). A dynamic theory of leadership and team effec-
organizational behavior and processes (Vol. 3; pp. 273–303). Oxford, tiveness: Developmental and task contingent leader roles. Research in
United Kingdom: Elsevier. Personnel and Human Resources Management, 14, 253–305.
Chen, G., Webber, S. S., Bliese, P. D., Mathieu, J. E., Payne, S. C., Born, Kozlowski, S. W. J., Gully, S. M., Nason, E. R., & Smith, E. M. (1999).
D. H., & Zaccaro, S. J. (2002). Simultaneous examination of the ante- Developing adaptive teams: A theory of compilation and performance
cedents and consequences of efficacy beliefs at multiple levels of anal- across levels and time. In D. R. Ilgen & E. D. Pulakos (Eds.), The
ysis. Human Performance, 15, 381– 409. changing nature of work performance: Implications for staffing, per-
Cobb, M. G., & Mathieu, J. E. (2003, April). The impact of training and sonnel actions, and development (pp. 240 –292). San Francisco: Jossey-
environmental complexity on the effectiveness of multi-team systems. Bass.
Paper presented at the 18th Annual Meeting of the Society for Industrial Kozlowski, S. W. J., & Klein, K. J. (2000). A multilevel approach to theory
and Organizational Psychology, Orlando, FL. and research in organizations: Contextual, temporal, and emergent pro-
Colquitt, J. A., LePine, J. A., & Noe, R. A. (2000). Toward an integrative cesses. In K. J. Klein & S. W. J. Kozlowski (Eds.), Multilevel theory,
theory of training motivation: A meta-analytic path analysis of 20 years research, and methods in organizations: Foundations, extensions, and
of research. Journal of Applied Psychology, 85, 678 –707. new directions (pp. 3–90). San Francisco: Jossey-Bass.
DeShon, R. P., Kozlowski, S. W. J., Schmidt, A. M., Milner, K. R., & Kozlowski, S. W. J., & Salas, E. (1997). An organizational systems
Weichmann, D. (2004). A multiple goal, multilevel model of feedback approach for the implementation and transfer of training. In J. K. Ford,
effects on the regulation of individual and team performance in training. S. W. J. Kozlowski, K. Kraiger, E. Salas, & M. Teachout (Eds.),
Journal of Applied Psychology, 89, 1035–1056. Improving training effectiveness in work organizations (pp. 247–287).
Driskell, J. E., & Salas, E. (1992). Can you study real teams in contrived Mahwah, NJ: Erlbaum.
settings? The value of small group research to understanding teams. In Kraiger, K., Ford, J. K., & Salas, E. (1993). Application of cognitive,
840 CHEN, THOMAS, AND WALLACE

skill-based, and affective theories of learning outcomes to new methods for the 21st century technology: Applications of psychological research
of training evaluation. Journal of Applied Psychology, 78, 311–328. (pp. 249 –279). Washington, DC: American Psychological Association.
LePine, J. A. (2003). Team adaptation and post change performance: Salas, E., & Cannon-Bowers, J. A. (2001). The science of training: A
Effects of team composition in terms of members’ cognitive ability and decade of progress. Annual Review of Psychology, 52, 471– 499.
personality. Journal of Applied Psychology, 88, 27–39. Schmidt, R. A., & Bjork, R. A. (1992). New conceptualization of practice:
LePine, J. A., Colquitt, J. A., & Erez, A. (2000). Adaptability to changing Common principles in three paradigms suggest new concepts for train-
task contexts: Effects of general cognitive ability, conscientiousness, and ing. Psychological Science, 3, 207–217.
openness to experience. Personnel Psychology, 53, 563–593. Shrout, P. E., & Bolger, N. (2002). Mediation in experimental and non-
Locke, E. A., & Latham, G. P. (1990). A theory of goal setting and task experimental studies: New procedures and recommendations. Psycho-
performance. Englewood Cliffs, NJ: Prentice Hall. logical Methods, 7, 422– 445.
Longbow2. (1997). [Computer software]. Coulsdon, England: Janes Com- Smith, E. M., Ford, J. K., & Kozlowski, S. W. J. (1997). Building adaptive
bat Simulations. expertise: Implications for training design. In M. A. Quiñones & A.
MacKinnon, D. P., Lockwood, C. M., Hoffman, J. M., West, S. G., & Ehrenstein (Eds.), Training for the 21st century technology: Applica-
Sheets, V. (2002). A comparison of methods to test mediation and other tions of psychological research (pp. 89 –118). Washington, DC: Amer-
intervening variable effects. Psychological Methods, 7, 83–104. ican Psychological Association.
Marks, M. (1999). A test of the impact of collective efficacy in routine and Smith-Jentsch, K. A., Salas, E., & Brannick, M. T. (2001). To transfer or
novel performance environments. Human Performance, 12, 295–309. not to transfer? Investigating the combined effects of trainee character-
Marks, M. A. (2000). A critical analysis of computer simulations for istics, team leader support, and team climate. Journal of Applied Psy-
conducting team research. Small Group Research, 31, 653– 675. chology, 86, 279 –292.
Marks, M. A., Mathieu, J. E., & Zaccaro, S. J. (2001). A conceptual
Sobel, M. E. (1982). Asymptotic confidence intervals for indirect effects in
framework and taxonomy of team processes. Academy of Management
structural equation models. In S. Leinhardt (Ed.), Sociological method-
Review, 26, 356 –376.
ology (pp. 290 –312). Washington, DC: American Sociological
Marks, M. A., Sabella, M. J., Burke, C. S., & Zaccaro, S. J. (2002). The
Association.
impact of cross-training on team effectiveness. Journal of Applied Psy-
Stajkovic, A. D., & Luthans, F. (1998). Self-efficacy and work-related
chology, 87, 3–13.
performance: A meta-analysis. Psychological Bulletin, 124, 240 –261.
Marks, M. A., Zaccaro, S. J., & Mathieu, J. E. (2000). Performance
Steiner, I. D. (1972). Group process and productivity. New York: Aca-
implications of leader briefings and team-interaction training for team
demic Press
adaptation to novel environments. Journal of Applied Psychology, 85,
Sundstrom, E., De Meuse, K. P., & Futrell, D. (1990). Work teams:
971–986.
Applications and effectiveness. American Psychologist, 15, 120 –133.
Mathieu, J. E., & Button, S. B. (1992). An examination of the relative
impact of normative information and self-efficacy on personal goals and Tesluk, P., Zaccaro, S. J., Marks, M. A., & Mathieu, J. E. (1997). Task and
performance over time. Journal of Applied Social Psychology, 22, aggregation issues in analysis and assessment of team performance. In
1758 –1775. M. T. Brannick, E. Salas, & C. Prince (Eds.), Team performance
Mathieu, J. E., Heffner, T. S., Goodwin, G. F., Salas, E., & Cannon- assessment and measurement: Theory, methods, and applications (pp.
Bowers, J. A. (2000). The influence of shared mental models on team 197–224). Mahwah, NJ: Erlbaum.
process and performance. Journal of Applied Psychology, 85, 273–283. Vancouver, J. B., Thompson, C. M., & Williams, A. A. (2000). The
Mitchell, T. R., & Daniels, E. (2003). Motivation. In W. C. Borman, D. R. changing signs in the relationships between self-efficacy, personal goals,
Ilgen, & R. J. Klimoski (Eds.), Comprehensive handbook of psychology: and performance. Journal of Applied Psychology, 86, 605– 620.
Vol. 12. Industrial and organizational psychology (pp. 225–254). New Wall, T. D., Michie, J., Patterson, M., Wood, S. J., Sheehan, M., Clegg,
York: Wiley. C. W., & West, M. (2004). On the validity of subjective measures of
Moynihan, L. M., & Peterson, R. S. (2001). A contingent configuration company performance. Personnel Psychology, 57, 95–118.
approach to understanding the role of personality in organizational Weingart, L. R. (1992). Impact of group goals, task component complexity,
groups. Research in Organizational Behavior, 23, 327–378. effort, and planning on group performance. Journal of Applied Psychol-
Pinheiro, J. C., & Bates, D. M. (2000). Mixed-effects models in S and ogy, 77, 682– 693.
S-PLUS. New York: Springer-Verlag. Wood, R. E. (1986). Task complexity: Definition of the construct. Orga-
Pulakos, E. D., Arad, S., Donovan, M. A., & Plamondon, K. E. (2000). nizational Behavior and Human Decision Processes, 37, 60 – 82.
Adaptability in the workplace: Development of a taxonomy of adaptive Wood, R. E., & Locke, E. A. (1990). Goal setting and strategy effects on
performance. Journal of Applied Psychology, 85, 612– 624. complex tasks. In B. Staw & L. L. Cummings (Eds.), Research in
Salas, E., & Cannon-Bowers, J. A. (1997). Methods, tools, and strategies organizational behavior (Vol. 12, pp. 73–110). Greenwich, CT: JAI
for team training. In M. A. Quiñones & A. Ehrenstein (Eds.), Training Press.
MULTILEVEL TRAINING TRANSFER 841

Appendix

Measures of Individual Goal Choice and Goal Striving 3. I monitored the physical condition of our helicopter
Activities
4. I focused my attention on how many weapons we had left
Goal Choice Items:
5. I monitored the approaching primary and secondary targets
1. I tried to better understand the main objectives of the upcoming
mission 6. I provided verbal feedback to my flight partner

2. I identified specific mission goals for our team to accomplish 7. I coached my flight partner to help him/her accomplish his/her
tasks
3. I prioritized the importance of different goals for our team
8. I timed my own actions to correspond with my flight partner’s
4. I formulated strategies for accomplishing our goals actions
5. I developed alternative courses of action for accomplishing our
9. I focused on coordinating well with my flight partner
goals
10. I communicated my actions to my flight partner when it was
Goal Striving Items:
necessary
1. I focused on how well our team progressed toward accomplish-
ing our goals
Received August 31, 2003
2. I paid close attention to what I needed to do at each stage of the Revision received August 22, 2004
mission Accepted September 15, 2004 䡲
View publication stats

You might also like