Supporting Material is available for this work. For more information, follow the link from
the Table of Contents to "Accessing Supporting Material".
Table of Contents
Go Back
UNDERSTANDING THE LEARNING PROCESS IN
WORK GROUPS
Elaine Lizeo
Fundagao Gettilio Vargas - Escola de Administragao de Empresas de Sao Paulo
Avenida Nove de Julho 2029 - Sao Paulo - SP
Brazil - 01313-902
elizeo@gvmail.br
ABSTRACT
Drawing on theories of organizational learning, group learning, group dynamics and
effectiveness, and using the system dynamics approach in three case studies of work groups,
this paper seeks to improve our understanding of the factors influencing group learning and
effectiveness in organizations. I propose that the relationship between group effectiveness and
ineffectiveness is not symmetrical and is characterized by an amplification phenomenon.
Groups that present a low level of learning tend to enter into a vicious cycle of low
effectiveness or to stagnate, while groups that present a high level of learning are able to
identify sources of ineffectiveness, manage limitations and implement the necessary changes
to recover or to elevate their level of effectiveness. Regarding work groups as complex social
systems, I suggest that the explanation of the level of learning and effectiveness of a group lies
in the interrelations of structural, cognitive and behavioral factors in interpersonal contexts.
Key words: group learning, group dynamics, group effectiveness, leadership, organizational
learning, system dynamics.
1. INTRODUCTION
Work groups have become a significant element of organizational effectiveness. To increase
their efficiency, flexibility and responsiveness, organizations are relying more and more on
work groups. Groups not only outperform individuals, especially when performance requires
multiple skills and experiences, but also represent a superior means to support broad-based
changes necessary for high-performing organizations (Katzenbach & Smith, 1993). Groups
can solve organizational problems of control and coordination, make decisions, and achieve
results that otherwise could not be achieved by individuals working independently
(McGregor, 1960).
This increasing reliance on work groups creates the need for them to be more creative,
innovative and efficient. To accomplish these aims, groups must be able to continuously learn
and improve. Learning enables groups to acquire new skills, improve processes, find new
ways of working, and enhance their decision-making and problem-solving skills. Group
learning is regarded in this study as a process of reflection and action, in which the ultimate
result is the creation of knowledge and insights.
Drawing on theories of organizational learning, group learning, group dynamics and
effectiveness, and using the feedback process approach of system dynamics, this paper seeks
to improve our understanding of the factors influencing group learning and effectiveness in
organizations. Improving our understanding of existing explanations for organizational
phenomena helps bridge the theory-practice gap long discussed (Pfeffer, 1993; Sutton &
Staw, 1995; Weick, 1989). The feedback process approach has been used in organizational
theory to understand a large number of social processes including organizational failure (Hall,
1976), organizational learning (Senge, 1990), improvement programs (Sterman, Repenning &
Kofman, 1997), and complex organizational processes (Weick, 1979). Although past studies
on group learning have identified feedback relations inherent to the learning process (Crossan,
Lane & White, 1999), only a few have used a systemic approach to model this phenomenon
(Lizeo, 2000). By showing how proposed causal relationships generate patterns of behavior,
dynamic models that capture non-linear, interdependent relationships, allow researchers to
explore and interpret complex organizational theories (Sastry, 1995).
The objective of this paper is to investigate the conditions under which some groups
become high-learning, highly-effective while others are caught into a cycle of low-learning,
low-effectiveness. Like natural systems, social systems (e.g. groups) are subject to self-
fulfilling spirals created by reinforcing feedback processes, in which the output amplifies
itself leading to exponential growth or decay (Levinthal & March, 1993; Sterman, 2000) —
e.g. population growth. Hackman (1990) suggests that work groups function as amplifiers
such that both effective and ineffective performance may be characterized as self-fulfilling
phenomena.
This paper proposes that the relationship between group effectiveness and
ineffectiveness is not symmetrical and is characterized by an amplification phenomenon.
Groups that present a low level of learning tend to enter into a vicious cycle of low
effectiveness or to stagnate, unable to alter their situation without adequate intervention. On
the other hand, groups that present a high level of learning are able to identify sources of
ineffectiveness, manage limitations and solve problems that come up along the course of their
task execution, and then, implement the necessary modifications to recover or to elevate their
level of effectiveness. In a study of group performance, Edmondson (1999) found that low-
learning groups, even with structural support initially in place, were vulnerable to a self-
sealing pattern, making it difficult for the group to escape the low-learning condition.
Whereas, high-learning groups through proactive learning behaviors were able to overcome
obstacles created by poor group design, low level of task interdependence, or lack of
resources, and thus correct their situation or course of action. This paper argues that the
asymmetry patterns of group ineffectiveness arise from infrequent learning behaviors,
undetected errors and misperceptions. Competency traps, limits to growth and delays in the
system are other factors predicted by this study to contribute to the asymmetrical
phenomenon.
The paper is organized as follows: in the remainder of this section, I outline the
problem towards which the modeling effort is directed and define group learning. Section two
presents the main dynamic hypotheses of the study. In section three, using causal loop
diagrams, I present three short case studies in order to explore the dynamic asymmetrical
phenomenon in group learning and effectiveness. Section four offers reflections on the cases
and implications for group learning and effectiveness, while section five contains concluding
thoughts.
The Problem: Why do some work groups present a high level of learning and effectiveness
while others are caught into a low level of learning and effectiveness?
Learning enables a group to find innovative approaches to problems, become more
efficient over time, acquire new skills, and change norms and procedures in response to
external circumstances (Ancona et al., 1996). Senge (1990, p.236) defines group learning as
“the process of aligning and developing the capability of a group to create the results its
members truly desire.” Ancona and her colleagues (1996) refer to group learning as the
group’s ability to survive, improve and adapt to changing circumstances. Edmondson (1999)
regards group learning as an ongoing process of reflection and action, characterized by
activities such as asking questions, seeking feedback, experimenting, reflecting on results, and
discussing errors or unexpected outcomes. For group learning to occur there must be a
complete cycle of reflection and action.
Following Edmondson’s conceptualization (Edmondson, 1999), I define group
learning as a continuous process of reflection and action based on shared understandings and
grounded in experience, resulting in the creation of knowledge and insights. Together, group
members reflect on the results of their actions (data on group progress) through reflective
conversation based on dialogue and discussion. Group members then design an action plan
which they use to make decisions and solve problems (experimentation). Their experiences
create opportunities to monitor results and serve as guides to new experimentations,
completing a cycle of group learning (figure 1).
Group interaction, reflection and action take place somewhere in some environment
that includes both physical and social aspects, which can make a difference in how members
behave and which affect the group’s shared beliefs and perceptions.
data on team dialogue and
e progress “\ discussion
opportunities to reflective
monitor results conversation
experimentation
Environment
Figure 1: Group Learning Cycle — Reflection and Action
Incomplete cycles can occur when a group fails to reflect on its own actions, or when
group members reflect but fail to implement changes following reflection. In either case,
groups are unlikely to create new knowledge or new ways of working that could help the
organization succeed in an uncertain or changing environment. Incomplete learning cycles
possess implications for groups and hence organizations. The quality of collective reflection
suffers and the ability to make sound decisions and to timely implement changes in response
to the environment can be disturbed by interpersonal processes that occur within the context
of work groups (Edmondson, 2002).
2. SOURCES OF EFFECTIVENESS (AND INEFFECTIVENESS) IN GROUPS
The learning process ofa group is influenced by the shared perceptions of its members and by
the group dynamics and functioning. Group learning requires that a group engage in learning
behaviors, such as reflective conversation, experimentation, feedback seeking, and
discussions of errors or unexpected outcomes. Learning behaviors have been studied
individually by psychologists and management researchers and have shown to promote
adaptation to changes and effective performance (Edmondson, 1999). To learn from
experience, including from mistakes, groups must consciously reflect on what has occurred
and on the results of their actions.
Research suggests that interpersonal factors influence the likelihood of people seeking
feedback, discussing mistakes or conducting experiments (Argyris, 1982). Even in well-
designed organizational structures, individuals in groups may fail to engage in learning-
oriented activities as a result of individual perceptions and prevalent social norms
(Edmondson, 1999). In situations of potential embarrassment or threat, individuals tend to
present behaviors that help them preserve a positive social value (Goffman, 1955).
Disagreement, conflict, negative feedback, admission of error, and request for help are
examples of interpersonal facts that can threaten the social image of individuals. Thus, in the
absence of a high degree of interpersonal skills (Argyris, 1993), group members are unlikely
to openly confront error, disagreement, or any kind of feedback that lead to situations of
embarrassment or threat. Furthermore, members of a group can be reluctant to disclose their
own mistakes for fear of penalty even though they are sacrificing the opportunity of getting
help to solve problems or correct mistakes. This body of work suggests a first hypothesis:
H.1 The asymmetrical phenomenon arises from infrequent learning
behaviors, undetected errors and individual perceptions that lead to self-
protection and reluctance to take interpersonal risks.
For learning-oriented behaviors to take place it is necessary that group members feel
safe and confident that they will not be retaliated for speaking up their ideas and discussing
errors openly. Some individuals are highly sensitive to hierarchy and power and, by fear or
deference, they become polite and indirect (Edmondson, 2002). Important issues, but
somewhat risky, such as the validity of a new policy or the implementation of a new process,
are not discussed in group meetings since questioning or substantial suggestion is likely to
encounter opposition. Group members may engage in “abstract conversation” to avoid
excessive personal risk (Argyris, 1993) and the possibility of offending those with power. In
such circumstances, openness and quality of group reflection suffer in the presence of private
concerns regarding power. Members’ perception of power and interpersonal risk at the group
level must be low since the willingness to engage in learning-oriented activities is inversely
related to group perceptions of interpersonal risk. The level of perceived power and
interpersonal risk is a continuous variable, which differs across individuals and groups in
organizations in ways not explained by objective organizational conditions and structures
(Edmondson, 1997). Therefore, when analyzing learning conditions, it is important to
consider the effects of aggregating all group members’ perceptions. The actions of one or
more group members perceiving the situation as unsafe for carrying out learning behaviors
can overwhelm proactive learning behaviors of the other group members. The high level of
perceived power and interpersonal risk prevent group members from engaging in learning-
oriented activities, affecting the group’s ability to discuss results openly and adequately
correct actions or implement new ones. Basing her work on these findings, Edmondson
(1997) suggested a hypothesis that is revisited in this study:
H.2 Groups that present lower aggregate level of perceived power and
interpersonal risk will engage more frequently in learning-oriented
activities than those that present higher aggregate level of perceived
power and interpersonal risk.
These same dynamics that help groups to correct their course of action may well
function as limits to growth (Senge, 1990; Sterman, 2000). In the learning-performance
process, as a group learns, its performance will improve, but there are certain aspects, such as
environmental factors, constraints in the organizational context (e.g. lack of resources, time
constraints and technology) and member’s perceptions, which may limit the engagement of
the group in learning-oriented activities. Based on these arguments, this study suggests a third
hypothesis:
H.3 Limits to growth related to environmental and contextual aspects of
work groups hinder their learning process.
Another factor that keeps groups into the cycle of ineffectiveness is their persistence in
habitual routines, even in the face of external circumstances that require change. Routines in
organizations (and groups) can cause change and learning initiatives to be problematic since
the lessons of experience are generally drawn from a relatively small number of observations
in the complex system of organizations. Besides, interpretations of experience come from
features of individual inference and judgment, which can lead to biases or misinterpretations.
Learning barriers such as “competency traps” thus arise (Levitt & March, 1988). Competency
traps occur when a high level of competence in a particular routine leads a group to favor the
continuity of such a routine even if it is actually inferior to other alternatives available. In this
case, perceived need for change may be low, hindering learning and keeping the group
stagnated.
H.4 Competency traps contribute to keeping the group stagnated in a low
level of learning and effectiveness.
Some results from learning activities may take time to appear, creating likely
perceptions that there is no efficacy in the group’s actions. It may take time to perceive the
performance of the group given measurement and reporting systems (Sterman, 2000). As it
may take time to implement corrective actions after a discrepancy in the group performance is
perceived because of the period required to make the decision. Time delays may be present
between the implementation of a corrective action and its effect on the performance of the
group. Significant time delays in the system can cause corrective actions to continue even
when they are no longer necessary, thus forcing the system to adjust too much and triggering
a new correction in the opposite direction.
H.5 Delays contribute to creating amplification pattern in the group
performance and effectiveness.
In sum, several factors contribute to the asymmetrical differences in group learning
and effectiveness. Even within the same organizational context — organizational culture,
reward system, information systems and resources — groups may present distinct levels of
effectiveness. The many interactions of structural, cognitive and behavioral factors in
interpersonal contexts contribute to the complexity of the group as a social system and give
rise to the asymmetrical phenomenon discussed above.
3. THE UNDERLYING CAUSES OF GROUP LEARNING AND EFFECTIVENESS
The following cases explore the asymmetrical dynamic phenomenon in learning and
effectiveness among three groups observed under the same organizational setting. The cases
also help explore the (dynamic) hypotheses stated in section two through the development of
causal loop diagrams for each group studied. Qualitative research is a useful methodology for
investigating phenomena that are not well understood. The data for the cases were drawn from
a study conducted by Edmondson (2002) on the nature of learning in organizations.
Edmondson conducted an exploratory study of learning processes in twelve organizational
work groups engaged in activities ranging from strategic planning to hands-on manufacturing
of products. From the sample provided by her study, I selected and reviewed three groups
from which I developed three short case studies applying a system dynamics approach. The
criteria for the selection were (1) availability of data that provided enough information for the
causal loop diagrams and for comparisons among the groups, and (2) groups that presented
differences in their quality of learning according to their level of reflection and action —
reflection was identified through behaviors that fostered new insights about performance or
group process (e.g. reflective conversation) and action was identified through behaviors that
involved taking action to test or implement new ideas (e.g. experimentation). Group
(effective) performance was used as a proxy to group effectiveness, in which performance is
considered the degree to which the group output meets or exceeds organizational standards
measured in terms of quality, quantity, timeliness, efficiency and innovation (Hackman,
1987).
The Organizational Setting — the groups are Beanstalk, Radar, and Storage (this paper kept the
original group names provided in Edmondson’s study). Beanstalk and Radar are product
development groups and Storage is a production group (group type did not influence the
selection of the cases presented here, and no consistent patterns of learning and performance
in regard to group type emerged in Edmondson’s study or in this paper).
The groups work in a leading manufacturing company of office furniture, with
approximately 5,000 employees and a history of product and management innovation and
financial success in the United States. Besides the organizational context, the groups share the
10
same environmental factors in which recent significant changes in the industry can threaten
the company’s market leadership. Corporate customers are becoming increasingly price
sensitive, less willing to pay a premium for the high-end image of the company’s product
lines, and expecting more from consultative services. At the time of the study, a new CEO had
been hired from the outside to facilitate a turnaround in order to meet customer’s needs and
keep the organization’s leading position in the market.
Case 1: The Beanstalk Group
Beanstalk is composed of seven members — the group leader (a peer) and six members from
different functions. The goal of the group is to develop a new line of products designed to
extend and enhance an existing product line. Members have relative autonomy to develop
their individual tasks, which present a high level of interdependence, requiring group
members to work together and collaborate.
The observed learning process and performance improvement of the Beanstalk group
is supported by a low level of power and perceived interpersonal risk shared among its
members. Reflective conversation, or the open exchange of ideas and opinions, is a frequent
activity of Beanstalk. The reflective conversation of group members have generated ideas,
plans of action and experimentation, creating opportunities for group members to monitor the
results of their actions and to gather data on the group’s progress, which serve as basis for
further group reflection. These dynamics are represented by the reinforcing loop reflection
and action in the CLD (causal loop diagram) shown in figure 2.
The data on the group’s progress are used to assess the group performance against pre-
defined goals. When a gap between perceived performance and performance goals is
observed, the group tends to investigate the reasons of this gap. To do so, group members
seek feedback and help from the leader or people in other functions — senior management and
customers — that can help them to solve problems, seeking to correct the course of the project.
The group, then, takes corrective actions, which provide them with new data on how the
group is doing. These dynamics are represented in the CLD by the balancing loop
performance (figure 2).
11
Beanstalk’s leader has a participative or democratic style according to the leader
categorization developed by Lewin, Lippitt and White (1939), in which she behaves as a
facilitator, encouraging dialogue and inputs from group members. A supportive leader, she
collaborates to create a safe and trustful environment within the group and seeks to encourage
group members to speak up their opinions and points of view. This supportive behavior of the
leader contributes to decreasing the group’s shared perception of power and interpersonal risk,
fostering the group engagement in learning behaviors such as help and feedback seeking —
dynamics represented by the balancing loop leader influence (coaching) in figure 2.
A low level of the group’s perception of power and interpersonal risk enhances
members’ shared belief that the group is safe for speaking up and challenging one another’s
notions and ideas, facilitating the engagement of the group in dialogue and discussion. The
practice of dialogue and discussion leads group members to reflect upon and improve its
internal process, enhancing group collaboration and communication, and the group ability to
deal with conflicts. Such dynamics are represented by the reinforcing loop internal process in
figure 2.
12
performance
leader willingness
als
, ‘ geet to coach
performance “hy
is lea & in o, ence _ leader collaboration
+ (coaching) and facilitation
perceived help and
p = Ap) feedback seeking
performance 4 TS %
corrective group's perceptions
action of power and
data on group Pe, dialogue and interpersonal risk
fs Progress discussion 2
+ Cc,
opp ortunities to ae i ; group
monitor results ao) OWS: reflective internal collaboration
action conversation process ts
“hea ww
i group process
experimentation improvement
Figure 2: Beanstalk’s CLD — Dynamics of Group Learning and Performance
Beanstalk presents self-correcting dynamics and performance improvement over time
through a mix of balancing and reinforcing loops. As the group completes the cycle of
reflection and action, it is able not only to learn but also to improve its tasks and internal
processes. A lower level of perceived power and interpersonal risk is supported by the
group collaboration as well as by the leader collaboration and facilitation.
13
Case 2: The Radar Group
Like Beanstalk, Radar, a product development group, is composed of seven members — the
group leader (also a peer) and six members from different functions. Radar’s goal is to
develop a product envisioned by senior management to augment an existing product line. The
product development was relatively well defined by senior management, but with no inputs
from group members, who for the most part believed that the project would not work.
Although individual tasks present some level of interdependence, members have little
autonomy to experiment with new ideas and new ways of carrying out their work.
This case shows a group that has some level of reflection (e.g. reflection on the project
development and customer’s needs) but that does so individually. Some members of the group
has developed negative attributions about each other’s motives and became concerned about
getting punished by senior management. Consequently, they did not raise questions about the
feasibility of the project and avoid discussing ideas openly or to offer any feedback that may
be potentially negative. This has resulted in poor quality of reflection and, consequently, in
time wasted in unproductive discussions. The existence of such an atmosphere has hampered
the development of shared understanding about the situation of the group, limiting the
opportunities the group has to monitor its results and gather data on its progress — data that
could form the basis for the group process of reflection and action. These dynamics are
represented by the reinforcing loop reflection and action in the CLD shown in figure 3.
Radar has presented a relatively low level of performance, which contributes to
enlarge the group performance gap. An increased gap on performance should signal the group
the need to seek help and feedback from the group leader and from others outside the group
and to engage in other learning-oriented activities. This way, the group would be able to
design and implement corrective action in order to improve results and close the performance
gap. However, the fact that group members systematically avoid seeking help and feedback
and discussing problems openly has led them to not obtain relevant information early enough
to learn more about customer’s needs and to make decisions timely in the project
development phase. This situation creates a significant delay, which has led the group to
correct its course later than necessary, hindering group performance. Such dynamics are
represented by the balancing loop performance (figure 3).
14
Radar’s leader follows a somewhat autocratic style (Lewin, Lippitt & White, 1939), in
which he holds on the role of spokesperson of the group and makes the decisions. According
to a group member, the leader has a “boss” attitude, controlling the decision making process.
Consequently, the group has perceived an unsafe environment for open communication,
reinforcing the perception that power and influence are concentrated in the hands of the
“leader-boss.” Such a perception can explain the collective behavior of low disposition for
dialogue and productive discussion and for help and feedback seeking, negatively influencing
learning, performance and other aspects of group effectiveness. These dynamics are
represented by the reinforcing loop leader influence (control) in figure 3.
This situation, a result of the group’s perceptions of high level of power and
interpersonal risk, has prevented the group from improving its own internal processes and
hence from increasing its collaboration. These dynamics are represented by the reinforcing
loop internal process in figure 3.
15
performance leader willingness to
goals take control
4s
x An)
performance leaden &
= gap “a be ee leader collaboration
a and facilitation
perceived B help and
performance feedback seeking
ms performance xi } PP tg =
corrective group's perceptions
action of power and
data on group a2 interpersonal risk
progress dialogue and
f° — es
CG,
opportunities to ; group
monitor results yh internal collaboration
B ‘tion & conversation process eu
a action
“Sf,
e group process
experimentation improvement
Figure 3: Radar’s CLD — Dynamics of Group Learning and Performance
The Radar group presents one balancing loop, which does not have strength to
overcome the dynamics of the three reinforcing loops that function in a vicious cycle.
The delay between the time that the group perceives its performance to be lower and
the time that the group takes action is too long (into the project development) to help
counter the behavior of the system.
16
Case 3: The Storage Group
Storage, the production group, is composed of 26 members — a supervisor and 25
production workers, who work in the day shift. The group’s goal is to manufacture storage
and filing systems and its work is organized into specialized subtasks.
The Storage group presents dynamics that distinguish itself from the other two groups.
When facing problems, as an observer reported (Edmondson, 2002), group members neither
try to solve the problems nor do they seek help, they just sit there and do not work. Then, they
group members get to work overtime on weekends to achieve the performance goals.
Technical problems, however, accumulate since the group doesn’t seek help from engineers
or people from other functional areas that could assist them. The less inputs and help group
members seek, the more problems accumulate, signaling a greater need for overtime — a
situation that corroborates the group’s norms and mental models towards overtime, but that
cover the real sources of ineffectiveness, hindering any efforts of improvement. With
overtime, though, group members can decrease the number of production error and technical
problems, improving their perceived performance and contributing to close the performance
gap. These dynamics are represented by the balancing loop performance through overtime in
the group CLD in figure 4.
Overtime work not only helps Storage members reach production output but also
provides them with financial gains — a reinforcing dynamic represented by the reinforcing
loop financial satisfaction (figure 4).
The group seems to prefer to work through overtime to engage in learning-oriented
activities, which would help them to increase their performance and achieve their goals within
the normal working hours. Punctual increases in the group performance result from overtime
work in the short run, confirming the group’s perceived need for overtime. This need for
overtime favors the distribution of the work in subtasks, which thwarts members’
collaboration and hence group process improvement. The less group members collaborate, the
higher they may perceive the level of interpersonal risk, which prevents them from engaging
in learning-oriented behaviors such as dialogue and discussion. The Storage group also avoids
the idea of participating in cross-functional training, restricting the opportunity for
experimentation with new processes and hampering production process improvement, which,
17
in turn, contributes to the persistence of production errors and technical problems. As errors
and problems increases, the need for overtime increases. These dynamics are represented by
the reinforcing loop internal process in the CLD — figure 4.
The higher group members perceive the level of interpersonal risk, the less they will
engage in dialogue and discussion, decreasing the opportunities for reflective conversation.
Such a situation of little, if no, reflection hampers the ability of the group to improve its
internal processes (e.g. communication, conflict resolution), contributing to maintain
members working independently and collaborating less. As group collaboration decreases, the
level of perceived interpersonal risk intensifies, closing a reinforcing loop interpersonal
process, shown in figure 4.
The Storage group does not present any kind of reflective conversation. Neither does
the group leader (or supervisor) try to facilitate dialogue and discussion, following a non-
participative or a /aissez-faire style (Lewin, Lippitt and White, 1939). In fact, Storage’s leader
does not participate or interfere in the group work, contributing to create an atmosphere of
apathy within the group. The loops — reflection and action, leader influence (coaching) and
performance — are represented in the CLD by punctuated gray lines because these structures
are not part of the dynamics of this group.
18
financial gain
Hh
Ax)
financial
satisfaction = + a. ~
aig
4— perceived need Lace ream
overtime + Oo overtime _ performance
goals
Gf cual leader willingness
performance to coach
through overtime
production errors / +4
technical problems pr tata leider betes .
‘ a) 4@ leader influence |eader collaboration
e + (coaching) facili
perceived help and and facilitation
performance feedback seeking
production process 3 performance Q v
improvement ‘oup's perceptions
e corrective e i" panels d
+ R + action oF power an
A data on group = interpersonal risk
perceived eee dialogue and ®
competence progress : :
wv discussion -\.
+ ra
_ + @
opportunities to Ar) +4 " . group
monitor results rellesbon il reflective internal collaboration
e t i rOCess
+" Teton conversation P +
+ Sah
group process
experimentation improvement
Note: Gray links, variables and loops denote that the causal relations are not part of the
dynamics of this group.
Figure 4: Storage’s CLD — Dynamics of Group Learning and Performance
Storage presents dynamics, through a set of reinforcing loops, which keeps the
group in a vicious cycle of low learning, if any, low-performance state.
19
4, REFLECTING ON THE CASES
The causal loop diagrams drawn for each group help us understand the underlying
causes of the dynamics of each group’s internal processes, leader collaboration, learning
(reflection and action) and performance. Next, this study contrasts the dynamics of the three
groups, highlighting the factors that cause similar loops to produce distinct dynamics and
contribute to the differences in learning and effectiveness among these groups.
Reflection and action — the same dynamics that contributes to Beanstalk’s learning
and improvement keeps Radar in a vicious cycle of low reflection, low action. This is so
because of the high level of perceptions held by Radar group members about power and
interpersonal risk, contributing to keep members from questioning ideas and speaking up.
This situation prevents Radar members from engaging in productive conversation and
discussing relevant aspects of their tasks, hindering the implementation of better solutions and
improved product designs. The Storage group does not present such dynamics. It presents, in
turn, the reinforcing loop perceived competence, which represents a belief that the group is
competent, and does not need either training or help from others; they can “fix their problems
their way.” Reflection and experimentation is not part of this group’s day-to-day activities.
Performance — both Beanstalk and Radar present the same performance structure, in
which group performance is guided by the group’s progress towards completing its tasks
adequately in order to achieve their pre-established group goals. Storage, on the other hand,
presents a performance structure in which the work done in overtime has a key role in
explaining the group functioning. One could claim that such a distinction arise from the
difference in group type and group tasks. While Beanstalk’s and Radar’s organizational goal
is to develop new products, requiring innovation (reflection) and change (action) of these two
product development groups, Storage’s organizational goal is to deliver the organization’s
products, which is more related to routine tasks. However, given the development of new
products, the work of the Storage group may require continuous innovation and improvement
in manufacturing processes and product quality, entailing in this case group learning —
reflection and action. The extent to which group type and task type provide more or fewer
opportunities for learning is not object of the present study. Such an issue can be investigated
in details in future work.
20
Internal process — the internal processes of the three groups are represented by a
reinforcing loop. However, each one has distinct dynamics. Beanstalk’s internal process is
enhanced by learning-oriented activities (e.g. dialogue and discussion), which provide
opportunities for group members to reflect on their own group functioning and improve group
collaboration. Radar’s internal process suffers because of the group reluctance to discuss ideas
openly, hindering the opportunities that the group has to reflect on their internal processes and
improve them if necessary. The dynamics of Storage’s internal processes serve to reinforce
the status quo — no need for dialogue, reflection or any kind of improvement.
Leader influence (coaching and control) — when Beanstalk’s performance suffers
for some reason, the group leader experiences an increase in her willingness to coach the
group, helping restore or improve the group performance. This pattern of behavior happens
because of the balancing structure of the leader coaching loop that the group presents. The
behavior of Radar’s leader when the group faces a decrease in its performance, however, is to
take control over the group. The leader of Radar shows governing values related to what
Argyris and Schén (1974) call “Model I theories-in-use”. The leader stays in control by
advocating his views and not inquiring into group members’ views, generating defensiveness
and self-fulfilling prophesies — such a behavior only contributes to reinforce the perceptions
of the group about the leader (as autocratic) and the group atmosphere (as unsafe for risk
taking). In the Storage case, the leader has little (if any) influence on group members and
group internal process. Therefore, this group does not present any dynamics related to the
variables that represent leader behaviors.
The causal loop diagram of the Storage group presents two reinforcing loops —
financial satisfaction and perceived competence — and a balancing loop — performance
through overtime — that are not common to the other two groups. These structures arise
because of the group dynamics and shared beliefs about group goals, tasks and environment,
and work in conjunction to keep the group in a vicious cycle. A balancing force that could
take the group away from this situation — leader influence (coaching) — is not part of the
structure of this group.
It might be the case that the dynamics presented by one group can be more prevalent
than those presented by the others. For example, Beanstalk group may have a perceived
21
competence loop in its structure but such a loop is too weak to arise or to change the patterns
of behavior in the group.
According to the causal loop diagrams, table 1 summarizes some of the variables that
demonstrated to be relevant in the learning process of work groups and relates them to the
hypotheses offered in the section two of this paper.
Perceived need for
overtime
Time to perceive results
(delays)
Need to invest in learning
behaviors
Time invested in routine
processes and procedures
Engagement in learning
behaviors
Variable level | Dynamic Result level | Hypothesis
Atmosphere of safety and 1
Group’s perceptions of trust
power and interpersonal Tt 1 and2
risk Engagement in learning
behaviors t
Level of perceived 1
Leader collaboration and t interpersonal risk 2
facilitation Atmosphere of safety and
trust T
3 and 4
Note: f - means an increase; | - means a decrease
Table 1: Relationships of factors affecting learning and performance in groups
22
Implications for Learning and Effectiveness of Work Groups
The enthusiasm or apathy of a group in dedicating itself to learning oriented activities is
positively related to group members’ perception of power and interpersonal risk present in the
group atmosphere. For individuals to participate in learning-oriented activities, they need to
feel safe and comfortable and to expect that they will not be exposed to situations that
threaten their self-esteem (Argyris & Schén, 1974). Group members need to perceive the
possibility of analyzing data on group performance, seeking feedback, and discussing
mistakes without losing their identity or integrity, thereby allowing them to admit
disconfirming data rather than denying it (Schein, 1992). As group members increasingly
engage in learning behaviors, they interact and collaborate more, giving rise to a sense of
openness and trust within the group. In turn, the increasing openness and the interpersonal
trust, including respect for competence, promote the development of trusting interpersonal
relations of group members, fostering productive communication (McGrath, 1984),
collaboration and cohesiveness. As members develop interpersonal relations the group tends
to cohere more and develop a sense of safety (Kahn, 1990).
The leader of the group can promote group effectiveness by helping group members
learn to work interdependently, building commitment to the group and its tasks, and
coordinating the implementation of performance plans (Hackman, 1990). Group leader
coaching is likely to influence group psycho-social traits once a supportive, non-defensive
leader collaborates to create a safe and trustful environment within the group (Edmondson,
1999). A leader who is available to discuss ideas and help group members, who provides
constructive feedback, and who are open to inputs tends to create an atmosphere of trust and
safety. A leader, who acts in autocratic or punitive ways, creates an atmosphere in which
group members will be reluctant to face the interpersonal risk involved in such learning
behaviors as discussing errors and seeking feedback, as it was the case of the Radar group.
Once the leader contributes to creating a safe and trustful environment, the level of perceived
power and interpersonal risk of the group decreases and its members become more
comfortable to engage in reflective conversation, experiment with new ideas and look for
better ways to work together, hence improving group overall performance and effectiveness.
23
Any of the structures in the causal loop diagram of each group can present significant
delays that can produce misperceptions and mistaken analysis on the system functioning.
While some dynamics are easily observed in a short period of time, others can take
considerable time to become noticeable. Time delays may be especially salient in the case of
the implementation of an action and the perception of its effect on the system and to perceive
the benefits of learning or to build trustful relations. Therefore, time is and important factor in
the learning process and as such must be considered when analyzing or implementing
initiatives of group learning.
It is important to notice that time availability is a relevant element to learning
activities. In situations of scarcity of time and in the face of pressure to increase performance,
groups may stop evaluating the results of previous actions, discussing problems, seeking
feedback, or experimenting with new ideas. This type of attitude harms the group
effectiveness, since continuous learning process helps groups to understand the environment,
their situation, their customers’ needs, and to improve their decision making process and
group dynamics. If groups do not have enough time to explore and experiment, they can lose
the opportunity to learn something new and will have to content themselves with the old
thinking (Dixon, 1994).
5. CONCLUSIONS
Work groups are complex social systems (Lewin, 1948) and a systemic perspective is
valuable in the study of group learning because the existing cause-and-effect relations in
groups and organizations are complex and form causal structures that sometimes neutralize
the influence of a variable on another, and sometimes amplify the effect of a variable on
another (Weick, 1979). These relations offer significant explanations on the variations in the
quality of learning and the level of effectiveness of groups in organizations.
The analysis developed in this paper showed the importance of studying group as an
intact system that possesses specific characteristics resulting from particular dynamic
interrelations of factors, such as group internal processes, leader behavior and perceptions
24
shared by group members, which affect the group over time. Psycho-social factors have a
great impact on the quality of learning and level of effectiveness of work groups. In face-to-
face situations, individuals make sense of their environment and use their understanding to
make implicit decisions about what they (and the group) should do, how they should behave,
and what they should say. Group member’s attitudes affect the outcomes of the group and
many times lead to interpersonal contexts that create barriers to learning. When power
differences produce perceptions of interpersonal risk, interpretive leaning processes lead to
self-protection strategies.
The causal loop diagrams, or modes of behavior, developed in this paper demonstrated
to be powerful tools for exploring the conditions under which some groups fail to learn and
improve overtime. They also can help managers, group leaders and group members to
understand the situation in which the group finds itself and the underlying causes that
generate the patterns of behavior observed in the group learning and performance. By
understanding the causes of inefficiencies, stakeholders can implement corrective actions and
help the group to improve not only its learning process but also its effectiveness.
A next step to advance this study is the construction of a simulation model in order to
extend the understanding of the patterns of behaviors of the feedback structures identified. A
simulation model will assist in the investigation of the relevance of each factor that influences
the learning process and will allow the simulation of possible scenarios in relation to these
factors and the circumstances of the group.
REFERENCES
Ancona, D. G., Kochan, T., Scully, M., Van Maanen, J., Westeney, E. D. (1996). Managing
for the Future - Organizational Behavior and Processes. Modules 3, 5, and 6. South-
Western College Publishing.
Argyris, C. (1982). Reasoning, Learning and Action: Individual and Organizational. San
Francisco: Jossey-Bass.
25
Argyris, C. (1993). Knowledge for Action: A Guide to Overcoming Barriers to
Organizational Change. San Francisco: Jossey-Bass.
Argyris, C. and Schén, D. (1974). Theory in Practice. San Francisco: Jossey-Bass.
Crossan, M.M.; Lane, H.W.; White, R.E. (1999). An organizational learning framework: from
intuition to institution. Academy of Management Review, v.24, n.3, p.522-537.
Dixon, N.M. (1994). The Organizational Learning Cycle: How We Can Learn Collectively.
London; New York: McGraw-Hill
Edmondson, A.C. (1997). Asymmetrical amplification in social systems: the “poor get
poorer” but the rich don’t get richer in work teams. Working paper, HBS.
Edmondson, A.C. (1999). Psychological safety and learning behavior in work teams.
Administrative Science Quarterly, 44, 350-383.
Edmondson, A.C. (2002). The local and variegated nature of learning in organizations: a
group-level perspective. Organization Science, Vol.13, No. 2, March-April, pp.128-
146.
Goffman, E. (1955). On face-work: an analysis of ritual elements in social interaction.
Psychiatry, 81: 213-231.
Hackman, J. R. (1987). The Design of Work Teams. In J. Lorsch (Ed.), Handbook of
Organizational Behavior, 315-342. Englewood Cliffs, NJ: Prentice Hall.
Hackman, J.R. (1990). Groups that Work (and Those that Don’t). CA: Jossey-Bass.
Hall, R. (1976). A system pathology of an organization: the rise and fall of the old Saturday
Evening Post. Administrative Science Quarterly, 21: 185-211.
Kahn, W.A. (1990). Psychological conditions of personal engagement and disengagement at
work. Academy of Management Journal, 33, 4, 692-724.
Katzenbach, J.R. and Smith, D.K. (1993). The Wisdom of Teams. MA: Harvard Business
School Press.
Levinthal, D. and March, J.G. (1981). A model of adaptive organizational search. Journal of
Economic Behavior and Organizations, (2) 307-333.
26
Levitt, B. and March, J.G. (1988). Organizational learning. Annual Review of Sociology, 14,
319-340.
Lewin, K. (1948). Resolving Social Conflicts: Selected Papers on Group Dynamics. New
York: Harper and Row.
Lewin, K.; Lippitt, R.; White, R.K. (1939). Patterns of aggressive behavior in experimentally
created social climates. Journal of Social Psychology, v. 10, p.271-279.
Lizeo, E. (2000). The Influences of Learning Behavior on the Performance of Work Team — A
System Dynamics Approach. Master’s Thesis, MIT. Cambridge, MA.
http://theses.mit.edu/Dienst/UI/2.0/Composite/0018.mit.theses%2f2000-25/1?nsections=4
McGrath, J.E. (1984). Groups: Interaction and Performance. Englewood Cliffs, NJ: Prentice-
Hall.
McGregor, D.M. (1960) The Human Side of Enterprise. New York, McGraw-Hill.
Pfeffer, J. (1993). Barriers to the advance of organizational science: paradigm development as
a dependent variable. Academy of Management Review, 18, 599-620.
Sastry, M.A. (1995). Time and Tide in Organizations — Simulating Change Processes in
Adaptive, Punctuated, and Ecological Theories of Organizational Evolution. Ph.D.
Dissertation. MIT, Cambridge, MA.
Schein, E.H. (1992). Organizational Culture and Leadership. 2™ Ed. San Francisco, CA:
Jossey Bass.
Senge, P.M. (1990). The Fifth Discipline: The Art and Practice of the Learning Organization.
NY: Doubleday.
Sterman, J.D. (2000). Business Dynamics: Systems Thinking and Modeling for a Complex
World. IL: Irwin/McGraw Hill.
Sterman, J.D, Repenning, N. and Kofman, F. (1997). unanticipated side effects of successful
quality programs: exploring a paradox of organizational improvement. Management
Science, 43, 503-521.
Sutton, R.I. and Staw, B.M. (1995). What theory is not. Administrative Science Quarterly,
v.40, p.371-384.
27
Weick, K.E. (1979). The Social Psychology of Organizing. 2™ Ed. New York: McGraw-Hill.
Weick, K.E. (1989). Theory construction as disciplined imagination. Academy of
Management Review, 14, 516-531.
Back to the Top
