Archetypal self-reinforcing structures in organizations:
A system dynamics perspective of cognitive, social, and institutional
processes
Anjali Sastry
University of Michigan Business School
701 Tappan Street
Ann Arbor MI USA 48109-1234
(734) 763-1591
masastry@ umich.edu
June 1998
Psychology, sociology, economics, and organizational studies have identified
cognitive, social, and institutional processes that appear to shape a wide range
of social systems. This existing research provides the starting point for an
exploration of specific feedback processes that recur in a variety of settings. Ina
sense, the resulting generic structures offer an alternative to the archetypes from
the systems thinking literature. The present focus is five areas of research on
organizations—leaming by doing, imprinting, escalation of commitment,
institutional isomorphism, and diffusion— that explain how past decisions and
actions are reinforced within organizations and in communities of organizations.
Some two dozen self-reinforcing processes that account for history dependence
in social systems emerge from an analysis of the research in these areas. Such
generic processes grounded in existing theoretical and empirical traditions have
much to offer system dynamicists interested in organizations. The feedback
loops and related bodies of research provide a ready-made collection of valuable
ideas for modelers to draw on in trying to understand dynamics of real-world
organizations. The resources presented here will be useful in building models,
developing links between system dynamics and other forms of research, and
making work in our field (and in others) more cumulative.
Some of the work presented in this paper was carried out jointly with Corinne Coen, to
whom I am grateful for research assistance.
System dynamicists’ search for generic structures has proved challenging. The promise is
that if we could identify the feedback processes responsible for common problems, and
pair system structures and resultant behaviors with each, we would be able to diagnose,
understand, and even remedy the problems we encounter in social systems. Potential
payoffs for practitioners are evident— a well-defined and validated set of generic
structures would serve as building blocks for system dynamics models, and may provide
shortcuts to developing insight into causes, consequences, and treatment of problems. As
Lane and Smart recently reminded us (1996), the field of system dynamics appears to have
been much influenced by Forrester’ s assertion that “probably twenty basic structures
would span 90% of the policy issues that most managers encounter” (Forrester 1980: 18).
No wonder so much attention has been devoted to the search for these fundamental
feedback processes.
Despite the promise offered by generic structures, debate continues about the progress
that system dynamics has made towards this goal (Lane and Smart 1996). Some see the
archetypes popularized by advocates of systems thinking in The Fifth Discipline
publications and practice (Senge 1991; Senge et al 1994), the work of Pegasus
Communications (e.g., Kim 1996), and others (e.g., Bellinger 1998) as important
contributions towards our understanding of generic structures. Taking the form of such
generic counter-intuitive system archetypes as “Shifting the Burden,” “Fixes That
Backfire,” and “Accidental A dversaries” (Senge et al. 1994: 125-150), these feedback
loop structures are promoted as tools for systems thinking that can be applied in a wide
range of organizational settings to both diagnose problems and generate solutions.
Archetypal self-reinforcing structures in organizations 2:
Systems thinking advocates suggest that archetypes are easy to apply: “Y ou start by
making guesses... Some people worry that they will apply the ‘wrong’ archetype,
misdiagnose the problem, and make things worse. In practice, this doesn’t happen”
(Goodman and Kemeny 1994: 121-122). Y et others note the difficulties in applying and
analyzing the implications of systems archetypes (Lane and Smart 1996; Sterman 1994).
One critique of the archetypes approach concerns the limitation of causal loop diagrams as
tools for analysis: loop polarity is difficult to establish via such diagrams alone; and
behavior is notoriously difficult to infer from even a simple set of loops (Morecroft 1982;
Richardson 1986; Sterman 1994).
System dynamics provides solutions to this first set of problems. Alternative diagramming
tools— such as stock and flow maps and policy structure diagrams— are useful in
conceptualizing and representing feedback structure and avoid some of the problems
inherent in causal loop diagrams; and of course the process of building and testing
computer models of the situation of interest allows the researcher or practitioner to
correctly infer how the processes interact and change over time. Fitting a generic model to
a specific dynamic system can help validate the generic model, while also exposing other
features of the situation— such as additional feedback loops— that may have escaped
attention.
Other critiques of the archetypes are linked to apparent ambiguity in the language,
questions about their use and application, and difficulties in formalizing them (Dowling et
al. 1995; Lane and Smart 1996). Particularly contentious is the suggestion that one can
Archetypal self-reinforcing structures in organizations 3
infer structure from behavior and select an appropriate antidote to existing problems, all
without formulating and simulating a model (Sterman 1994; Lane and Smart 1996).
However, as work by members of the Society for Organizational Leaning and other
systems thinkers progresses and the archetypes are refined— as more people follow
Sterman’s advice to use the archetypes as working hypotheses to be tested by computer
simulation— these issues may be resolved.
Y et such an approach taking the systems archetypes as a starting point may miss
significant opportunities to link work in system dynamics with findings from other fields
and to build better models and theories. The archetypes approach may also focus attention
at such a general level of analysis that insight, transfer, and testing are made more difficult
than they would be if mechanisms of change were more specifically described. In addition,
while there appears to be no reason why more traditional computer-based system
dynamics models built for research or practice should not follow Sterman’s advice and
draw on, elaborate, verify, and extend archetypes, we see little evidence that this is
happening. For instance, archetypes are absent in recent presentations of models appearing
in the System Dynamics Review and elsewhere.
Does this mean that the search for common basic feedback processes has been abandoned?
Not according to Lane and Smart, who point to two alternative approaches to generic
structure in system dynamics: canonical situation models, or general models of an
application domain, such as Forrester’ s market growth model (Lane and Smart 1996: 91);
and abstracted micro-structures, more transferable chunks of a simulation model that can
Archetypal self-reinforcing structures in organizations 4
be paired with behavior modes that they generate, such as “overshoot and collapse” (Lane
and Smart 1996: 96). While both approaches have been very productive for the field, they
fall short of providing us with theories that explain and predict the often problematic
behavior of social systems— precisely the area that the systems thinking archetypes
attempt to address. Generic structures that represent cognitive and social processes shown
to operate in a wide variety of settings could help system dynamicists to build better
models, develop hypotheses to test, generate alternative explanations for observed
behavior, and explain macro-level behavior in terms of micro structures. They could also
help link existing and new system dynamics work with research in other fields, while
making our work— as well as that of others— more cumulative, thus addressing system
dynamics’ need for what Richardson (1996) has labeled “a culture of accumulation.”
Luckily, we have much to draw on. Feedback loops are central to processes studied by
sociologists, psychologists, economists, and organizational scholars (Richardson, 1991;
Masuch, 1985). Such processes have been studied with a variety of research methods,
from experiments to case studies; many have long and rich histories, and consequently
have been extensively tested and validated. While the terminology and language employed
within each field and sub-field differ, and while feedback loops may be more or less clearly
articulated in the literature, feedback processes are nevertheless key in a wide range of
fields that study change over time.
Many of the same processes studied by other organizational researchers have been
identified by system dynamicists, and existing system dynamics models already may— and
Archetypal self-reinforcing structures in organizations 5
many do— incorporate elements of the processes studied in other fields. For instance, the
classic self-reinforcing process of leaming by doing appears in many system dynamics
models (e.g., Sterman and Richardson 1985; Sterman, Repenning and Kofman 1997). Y et
the terminology and specific feedback structures chosen to represent the phenomenon
often differ across models, and it is often difficult to establish whether a given model
includes the effect.. As a result, the field may miss opportunities to identify similarities
between apparently different models and application domains and to improve the
formulation and testing of the common model elements. For example, the leaming-by-
doing structure may appear in more models than is immediately evident from the literature
(for instance, the effect may be at work in such models as the Fossil2 energy model (Naill
1992) and Mashayekhi’s solid waste model (1993), although it is not evident from
published papers on these models). Y et despite the relatively simple nature of learning by
doing, a careful consideration of the phenomenon suggests many questions that could lead
modelers to new insights. For instance, economies of scale (present in Mashayekhi'’s
model) may substitute for leaming effects, or may complement them. Leaming effects may
generate advantages for one actor, technology, or organization that are more or less
permanent depending on the appropriability of the leaming— which may itself be affected
by policy choices. The extent to which learning can be transferred depends at least in part
on the extent to which the acquired knowledge resides in individuals (e.g., in their skills)
versus in the organization (e.g, in its operating procedures and routines). Similarly, the
extent to which learning can be encouraged by management also varies across situations.
Developing a body of knowledge about generic structures would help to highlight such
issues for a wide range of system dynamics research.
Archetypal self-reinforcing structures in organizations 6
The aim of this paper is to propose an alternative approach to generic structures: why not
take the large body of existing research on organizations as the starting point, and look for
recurring feedback processes that have already been found in a variety of settings and
explored in the literature? Such research can offer system dynamicists ready-made— and
already tested— feedback loops to choose from, as well as questions to ask, extensions to
consider, and rival hypotheses to explore. Documented case studies provide histories
against which the results of computer simulations can be compared. Perhaps most
importantly, the processes that generate behavior at psychological, interpersonal,
organizational, and institutional levels are described, suggesting variables that are easily
measured and compared. Rather than the general level of analysis at which systems
archetypes operate, the processes explored here work at specific cognitive or social levels.
The resulting causal loop diagrams are resources that any system dynamicist can draw on
to understand the processes uncovered in the course of building models of particular
organizations or social systems, as well as to develop and explore generic structures at
work in social systems.
To bound the present study, we confine attention here to theories that explain why history,
in the form of past events, initial conditions, and early choices, is important to social
systems. History dependence is one important type of counter-intuitive behavior seen in
organizations, since systems that are shaped by their past may resist efforts to change them
and may lock in to apparently sub-optimal states. Self-reinforcement is central to five areas
of organizational research— imprinting, learning by doing, escalation of commitment,
institutional isomorphism, and diffusion. Analyzing the generic processes that shape
Archetypal self-reinforcing structures in organizations 7
behavior in each yields some two dozen positive feedback structures that influence a wide
variety of social systems.
Methods
A natural starting point for finding self-reinforcing processes was to search the existing
published organizational research for references to positive feedback, history dependence,
vicious and virtuous circles, self-reinforcement, and path dependence. Drawing on the
Social Sciences Citation Index, our own reading, indexes of survey books and collected
papers, suggestions from colleagues, the intemet, ABI Inform, and other sources, we
looked for candidate articles, research papers, and books. The multiple search strategies
rapidly converged and the five main areas of research emerged.
The next step was to identify the papers that best captured the key ideas of each set of
theories. In some areas, survey articles were helpful; in others, early foundational articles
were the best source; and in still other areas, more recent publications gave a good
overview of the theory. Treating each paper or set of papers as our data sources, next we
analyzed each text to identify key processes. The result was a set of causal loop diagrams
grounded in the original theoretical and empirical work in each research area.
Interestingly, difficulties in constructing causal loop diagrams from existing research
accounts in some areas of organizational studies may suggest problems with the
underlying theories— in particular, missing or ambiguous feedback structures. For
instance, the recent literature on organizational co-evolution does not present a clearly
identifiable set of reinforcing processes and causal mechanisms, despite the focus on
Archetypal self-reinforcing structures in organizations 8
mutual influence evident in the notion that organizations, technologies, institutions and
belief structures evolve in concert (Garud and Rappa 1994; Nelson 1994; Van de Ven and
Garud 1994).
A second check of the causal loop diagrams was provided by a review of additional
experiments, field studies, and case research as well as theoretical extensions to the
original ideas. In addition, because all the candidate theories are well-known, it has also
been easy to solicit suggestions and criticism from others, which also helped to refine the
analysis.
The five bodies of research that emerged from this analysis address organizational
phenomena that are, at first blush, very different from each other. Explanations for
learning by doing account for the tendency of people and organizations to improve
performance over time. Cumulative experience leads to improvements in average costs,
quality, or error rates. The result of these improvements is higher demand, which in tum
facilitates additional improvements as more experience cumulates (Nelson and Winter
1982; Argote, Beckman, and Epple 1990). A different set of processes is at work when
organizations maintain elements of founders’ decisions, choices and approaches, or
preserve aspects of early strategy or initial conditions, over time. The explanation for
organizational imprinting identifies social and organizational processes responsible for the
persistence of founders’ practices, views, and policies (Kimberly 1979; Schein 1983;
Boeker 1988; Eisenhardt and Schoonhoven 1990), the effects of early environmental
conditions (Stinchcombe 1965), or initial strategy choices (Boeker 1988; Boeker 1989b;
Archetypal self-reinforcing structures in organizations 9
Romanelli 1989). Escalation of commitment has been explained by a variety of
psychological, social, and organizational processes that account for the increasing
difficulty of changing once a course of action has been selected. Ongoing investments of
time and effort into a chosen course of action results— even when it is an apparently poor
choice (Staw and Ross, 1978; Staw, 1981; Staw and Ross, 1987; Staw and Ross, 1989).
Institutional isomorphism explains why organizations grow to resemble others within their
field. This similarity is caused by the spread of common noms and practices, the
mimicking of peers, and the direct control of other, more powerful organizations or
institutions (DiMaggio and Powell, 1983). Innovations may also spread through a
population by processes of diffusion, bandwagons, and fads. Cognitive and social factors
shape the adoption process (Rogers, 1983; Valente, 1995).
Each of the five areas together with a description and diagram of the key feedback loops is
described in the sections that follow. We tum to each of these areas next, and then step
back to assess what we have learned.
Learning by doing
The literature on learning by doing details the processes by which people and
organizations become better at what they are doing as a result of experience (Nelson and
Winter 1982; Argote, Beckman and Epple 1990). Practice in producing a product or in
delivering a service results in greater understanding of the required processes as people
“work out the kinks” and solve problems as they encounter them, learn to coordinate with
others and improve the timing of interactions, discover short cuts, develop supporting
Archetypal self-reinforcing structures in organizations 10
tools and technologies, refine the equipment they use, and evolve and elaborate working
tules, policies, procedures and norms to support production. Organizational leaning of
this sort takes place when unit costs decrease as a function of knowledge acquired through
production (Argote and Epple 1990). It also results when quality improves, or accidents,
defects, or complaints per unit fall, as a product of experience. As cumulative experience
builds, costs decline, efficiency improves, and quality increases, stimulating demand and
consequently production; in turn, activities are repeated and tools used more as experience
builds. The knowledge acquired as a result may be embodied in individuals— for instance
in their skills— or may reside in the organization— for example, in its operating procedures
and routines.
Figure 1 presents a template of the learning by doing process. While the loop represents
price reductions that result from cumulative experience driving down unit costs and hence
generating demand and production, learning by doing may also operate in a similar process
without price changes, via improved quality driving up demand and hence production.
Figure 1: Learning by doing loop
cumulative relevant
a acquired
production
4 unit cost
Archetypal self-reinforcing structures in organizations 11
Constraints to the growth of organizational knowledge exist: the values of experience
gained in production may depreciate rapidly (Argote, Beckman, and Epple, 1990; Argote
and Epple, 1990). In empirical tests of both individual psychology and organizational
processes, learning does not necessarily persist. Forgetting, turnover, the loss of records,
obsolescence of knowledge, and disruption in practice all hasten the loss.
Imprinting
Imprinting, or founding effects, offer a second general type of self-reinforcing mechanism.
This area of organizational research provides empirical evidence of the permanence of
early choices, as well as some explanation of the processes that generate persistence. Over
time, imprinted organizations become increasingly committed to early strategies,
reinvesting their resources in employees with consonant skills, building sets of norms,
practices and beliefs that promote the original vision, and refining policies to support the
goals arising from their original strategies. Three factors shape the imprinted organization:
the founder (Kimberly 1979; Schein 1983; Boeker 1988; Eisenhardt and Schoonhoven
1990); early choices and decisions (Boeker 1988; Boeker 1989b; Romanelli, 1989); and
initial conditions (Stinchcombe 1965). Any of these may leave a lasting impression on the
organization, influencing its strategy, culture, policies, and procedures in later life.
By setting forth expectations, shaping practices, and articulating a vision, founders
establish the norms, practices and beliefs that guide recruitment and socialization and
provide the starting point for further elaboration by organization members (Boeker 1988).
As employees identify with founders, they reinforce the founders’ behaviors by repeating
Archetypal self-reinforcing structures in organizations 12
and elaborating them, generating pressure for the founders to continue. Through the
process of socialization, new employees recreate the founders’ fundamental assumptions,
priorities, and strategies as they learn how to behave and interpret information in the
organization, respond to the reward system, and model themselves after respected figures
(Boeker 1989b). Such effects may persist even after the founder or founding team has left
the organization or when the environment has changed. As the founders’ habits, practices,
and beliefs become institutionalized, reward systems evolve to encourage organizational
members to sustain or repeat consistent actions (Boeker 1989b).
In Figure 2, the term “culture” represents the norms, practices, and beliefs that shape
organizational members’ behaviors and views. The processes involved in the socialization
and the elaboration of culture feedback loop serve as reinforcing mechanisms (Schein
1983; 1985). As Figure 2 shows, as culture is strengthened, selection, recruitment and
socialization of compatible members becomes easier; in tum, as homogeneity increases
and cultural norms are better communicated, the culture is reinforced.
Figure 2: Socialization and culture loop
strength of
culture
+
iG A *
ease of recruitment and
socialization
increase in
culture seem
rn
Archetypal self-reinforcing structures in organizations 13
Once a firm chooses an initial strategy, adopts an organizational form, or elaborates
procedures, it recruits employees with consonant knowledge and skills. Competencies for
alternative practices may be undeveloped (Boeker 1989b). The longer a founder runs an
organization, the more the manner of operating will be taken for granted (Boeker 1989b)
by both employees and potential recruits.
In the realm of strategy, initial choices may be reinforced as further resources are invested
and supporting policies are articulated. Similar processes are at work even when the initial
choices are dictated by the environment instead of being selected by managers. In such a
case, the conditions at founding— what Stinchcombe (1965) calls the “available social
technology”— shape the organization’ s structure, strategic choices, and culture. In either
situation, early choices persist because changing course is costly and investments are
difficult to reverse, early practices quickly become taken for granted, and choices are
elaborated into interconnected systems (Eisenhardt and Schoonhoven 1990), making the
strategy less reversible over time (see Figure 3, depicting strategy irreversibility). The
effect is further compounded when additional costly resources are invested or when
obsolete technology, physical plant, or skills constrain choices (Stinchcombe 1965;
Eisenhardt and Schoonhoven 1990).
Archetypal self-reinforcing structures in organizations 14
Figure 3: Strategy irreversibility loop
investment
-_ in strategy
increase in reversibility
strategy of strategy
investment 8
XN internal .
commitment
to strategy
As a result, the more a firm invests in a strategy, the more difficult it becomes to change
(Boeker 1989b). Information gathering and decision-making evolve to support the
strategy, further reinforcing the internal consensus. Subsequently, other strategic options
are rarely considered. Commitment builds through the investment in facilities, personnel,
practices and policies uniquely suited to that strategy; the strategy becomes less and less
reversible over time as a result. The networks of relationships among members of the firm
influence their ability to perform and the creation of consensus (Boeker 1988), and these
relationships evolve to support the strategy in use. As intemal belief systems and power
distributions develop, intemal commitment to the strategy is reinforced, resulting in the
strengthening of constituencies within the organization who support the status quo.
In a related process, the refinement of strategy affects the elaboration of practices and the
expression of beliefs. Idiosyncratic conceptions of a founder or founding team, driven by
their cognitive biases and personal beliefs, shape a firm’s predilection toward a strategy
(Boeker 1988). Early patterns of activity become “traditionalized” (Stinchcombe 1965) or
routinized (Nelson and Winter 1982), limiting the variations considered for future practice.
Archetypal self-reinforcing structures in organizations 15
Over time, strategy development becomes increasingly entwined with the belief system and
traditional practices of the firm, rather than changing in response to environmental
demands.
This process is represented in a reinforcing loop termed values elaboration. As the
organization’ s strategy develops, supporting connections with other aspects of the
organization are reinforced, lending increasing support to the strategy and resulting in
more effort being directed to its development and refinement.
Figure 4: Values elaboration loop
development and
+ refinement of
strategy
A steele
interconnections with
investment organizational
in strategy practices, consistency
Bi belief systems
+
In addition to the growing consistency of strategy and beliefs, planning and elaboration of
procedures replace a founder's vision and invention (Mintzberg and Waters 1982). Once
the founder departs from the firm, employees extrapolate appropriate behaviors from past
practice or from policy, reinforcing the status quo. Where the founders may have felt free
to change their vision at will, subsequent planners feel constrained to honor the plan that
the founders articulated.
Archetypal self-reinforcing structures in organizations 16
New organizations often face a crisis of legitimacy with external constituencies (DiMaggio
and Powell 1983; Hannan and Freeman 1984). To address this problem, the organization
makes increasing commitments to outside social forces (Stinchcombe 1965). Ideologies
are elaborated to justify and sustain stable relations with external constituencies. New
firms lack resources to shape the market and competitive conditions, but instead depend
on legitimacy to support their demands and claims on others. An important antecedent of
this external validation is the organization’ s apparent commitment to a course of action.
The result of the effort to maintain the appearance of commitment is the increased support
of these external parties to the original strategy. Thus, outside parties tend to invest more
in organizations that appear validated; and apparent persistence— which is itself the result
of the organization’s making a visible commitment to a strategy— is the key driver of that
validation. Once more is invested in a given strategic approach, the organization’ s
commitment to the strategy is in even greater evidence, resulting in reinforcement of
visible commitment
Figure 5: Visible commitment loop
investment
a strategy »;
extemal visible
validation commitment
to strate
xe °
apparent
persistence +
Archetypal self-reinforcing structures in organizations 17
The final process is one named resource accunulation. It represents the simple principle
that as the organization grows in size, it generates greater resources capable of further
supporting its growth (Eisenhardt and Schoonhoven 1990). For instance, a small
organization may not be able to invest in a customized training programs, whereas a larger
organization may be able to afford this more efficient method of training and socializing
employees. The investment in the training program thus both saves money and increases
performance, supporting further growth of the organization.
Figure 6: Resource accumulation loop
organization
[( growth rate \,
investment organization
in growth size
xX aay;
resources
What constrains these positive feedback loops? Empirical studies show that the effects of
imprinting tend to decline with time (Romanelli 1992). Over time, these processes
encounter diminishing retums: once an organization has become very effective in
socializing entrants, homogeneity of its members will increase at a diminishing rate as the
addition of one more similar member does little to change the organization’ s overall level
of homogeneity. In addition, later events may prevent the perpetuation of initial strategies,
alter norms or the balance of power across functions, disrupt commitment, or cut the flow
of resources. The primary force that causes this disturbance is a drastic change in level of
Archetypal self-reinforcing structures in organizations 18
performance (Boeker 1989b). Performance failures, in particular, may force a fundamental
reordering of activities (Tushman and Romanelli 1985). In addition to market demands,
changes in corporate legitimacy may induce organizational change (Aldrich 1979). The
effect of initial conditions is also diminished by the departure of the founder or founding
team and thus their reinforcing influence (Boeker 1989b; 1989a).
Escalation of commitment
The research on escalation of commitment explains why people continue to invest time,
money, and attention in losing courses of action (Staw 1976; 1981; Staw and Ross 1978;
1987; 1989). Contributing processes appear at multiple levels of analysis— psychological,
social, and organizational— and different processes may dominate at different points in the
organization’ s experience. Early, apparently “rational” choices may subsequently be
reinforced because the decision-makers’ aversion to change, the irreversibility of
supporting investments, and organizational routinization, among other factors.
The story begins when organizational memhers choose one course of action over others.
In allocating resources, time, effort, or money, people may make their initial choice for
apparently rational reasons, such as cost-effectiveness or technical superiority. The very
act of undertaking the actions, however, may generate other reasons for continuing its
practice; the action or project can begin to generate a logic of its own that goes beyond
the initial rationale. Some reasons for persisting are objectively rational: early losses may
be expected if payoffs are known to follow some time after the initial investments; closing
Archetypal self-reinforcing structures in organizations 19
costs may be substantial. Y et other contributors to escalating commitment are the
psychological, social and political processes explored here.
If the choice entails an irreversible investment, decision-makers may consider their sunk
costs to be evidence that the action should be continued. This “sunk cost effect” results in
further investment which is justified not in terms of the initial rationale but because so
much has already been invested (Staw and Ross 1989). In this process, the sunk cost itself
is seen by the decision maker as a manifestation of commitment to the course of action,
further reinforcing the choice, action, or investment and thus generating a sunk cost loop.
Figure 7: Sunk cost loop
commitment to
course of action
oo
+ action
time and "o
resources
+
invested in
action
A similar process is at work when individuals frame later choices as opportunities to
remedy a past loss by making up for the cumulative investments already made (Staw and
Ross 1989). If an action is not successful, the greater the investment, the greater the
apparent loss entailed. Because of the common tendency to avoid loss (what psychologists
call “loss aversion”), this cumulative investment— or apparent loss— can generate
Archetypal self-reinforcing structures in organizations 20
increased commitment to undertaking the action in the hopes of recovering past losses.
This process is represented below in the framing bias loop.
Figure 8: Framing bias loop
apparent loss
+H entailed in course
of action
cumulative
investment cf
inaction +
commitment to
+ course of
action
action
Alongside these two errors in interpreting information are processes that operate at the
individual psychological level. For instance, the framing error is exacerbated by a second
process when the person making subsequent investment decisions also made the initial
choices: when they are responsible for past decisions, people are motivated to reinforce
them so as to feel that they have “done the right thing” and therefore rationalize past
actions (Staw and Ross 1989). We label this the self-justification process.
Archetypal self-reinforcing structures in organizations Pah
Figure 9: Self-justification loop
+ commitment toa
course of action
.\
responsibility
for action
action
+ %
+
Contributing to this process is the tendency for decision makers to pay attention to data
that supports (or is at least consistent with) their past decisions and to downplay
conflicting information. This confirmatory bias may skew the basis on which subsequent
decisions are made (Staw and Ross 1989); conflicting information is simply ignored and
alternative courses of action therefore not considered.
Figure 10: Confirmatory bias loop
commitment to a
course of action
+
+
confirmatory bias
in in data
action
tS apparent ca
of action
Passive self-inference may also generate escalation of commitment (Staw and Ross 1989).
Because people begin to define themselves in terms of the actions they take, an initial
choice may be repeated as the decision makers form personal beliefs that define
Archetypal self-reinforcing structures in organizations 22
themselves in terms of what they do. Self-inference is more pronounced when the action is
repeated, important, voluntary, public, difficult to reverse, and easy to interpret.
Figure 11: Self-inference loop
+
commitment
to action
importance of
action in self- oH
definition
+
action *
At the interpersonal level, external justification may be important. Decision makers’ fears
of others inferring that they have made mistakes in the past may result in their reinforcing
their initial decisions (Staw and Ross 1989). We label this the face-saving process.
Similarly, extemal binding occurs when observers both form inferences about why the
decision makers are undertaking the action and link those inferences to the decision
makers’ social identity (Staw and Ross 1989). As a result, observers identify the actors
with the action, increasing the costs of withdrawal. We label this the external binding
loop.
Archetypal self-reinforcing structures in organizations 23
Figure 12: Face-saving loop
fo ™
fon
desire to avoid to action
inference that past
action is
inappropriate
CO, action
Figure 13: External binding loop
+
—-
social identification nme
of actor with action a action,
+
+
action
A third process at work in the social arena rewards decision makers for persistence (Staw
and Ross 1989). Observers value the unswerving commitment manifested by decision
makers who stay the course; interestingly, those who persevere in the face of bad
outcomes may be revered more than those who adjust their actions to account for results
(p. 218). The increased value placed on persistence leads to greater commitment to
sustaining or repeating the action. We label this process perseverance.
Archetypal self-reinforcing structures in organizations
24
Figure 14: Perseverance loop
a
aaotient commitment
pa to action
persistence
+
+
cumulative action
experience
with action
+
In addition to the increased perceived value related to persistence, as an early choice is
first implemented within the organization, its members develop rules, policies, and
procedures that are consistent with the selected course of action. Once established and
elaborated upon, these structures filter information and make any change necessarily slow
and difficult. As rules become elaborated, consideration of further alternatives declines,
leading to increased commitment to the original choice (Staw and Ross 1989). Thus the
processes by which organizational action becomes routinized lead to an institutional
inertia loop.
Figure 15: Inertia loop
ty» commitment
fo to action
consideration of
altematives
+ G FS
action
elaboration of rules,
policies, and
procedures
Archetypal self-reinforcing structures in organizations
25
At the intra- or inter-organizational level, the behavior of political allies also shapes
commitment to an action (Staw and Ross 1989). As decision-makers invest time and
resources in an action, allies invest in consistent or complementary structures. The allies’
investments create both political and social costs for withdrawing from an early choice.
We label this the political and social commitment loops.
Figure 16: Political and social commitment loops
litical costs
of a
ao costs
of withdrawal actors’ commitment
allies' commitment to j.
to action
: js
action
allies' investment in
secant
While the Staw and Ross (1989) paper both summarized past work and provided a
foundation for future research on escalation of commitment, we see further development
of these ideas in subsequent work. The research on escalation of commitment includes
experimental (Staw 1976; Garland, Sandefur, and Rogers 1990; Heath 1995) and other
empirical studies (Arkes and Blumer 1985; Ross and Staw 1993; Staw and Hoang 1995)
as well as attempts to integrate conflicting theories (Brockner 1992). In recent
investigations of the limitations of the theory— and in particular of the effects of task
design on the tendency of people to continue investing in failing courses of action— we
find a debate about the relative importance of the roles of the justification processes versus
Archetypal self-reinforcing structures in organizations 26
the cognitive ones. Y et we observe general agreement that the processes described by
Staw and Ross are those that generate escalation of commitment.
Institutional isomorphism
Different mechanisms for organizational history dependence are provided by institutional
theory. Faced with uncertainty about how to solve a business or managerial problem,
organizational members tum to the practices, structures, and solutions to common
problems codified by their professional community, practiced by other organizations, or
mandated by more powerful institutions (DiMaggio and Powell 1983).
By adopting professionally sanctioned noms and practices, individuals and the
organizations to which they belong develop legitimacy. They also grow to resemble others
in their field. The result is normative isomorphism, or homogeneity in organizational
structure, culture, or output resulting from the diffusion of occupational standards and
practices. Professional practices spread in this manner include decision-making
approaches, acceptable solutions to common problems, and norms for personal behavior,
dress, and verbal communication (DiMaggio and Powell 1983). The process by which
members of an occupation establish and propagate shared work standards, known as
professionalization, is the primary vehicle of normative isomorphism.
Professional education is one key source of these occupational practices; social networks
that cross organizations also propagate normative rules. On-the-job training, consultant
relationships, industry associations, professional school networks, and trade magazines
Archetypal self-reinforcing structures in organizations 27
serve as mechanisms for further reinforcing normative isomorphism. Organizations
contribute to the process when they filter recruits or allocate promotions to favor those
from within their industry, from a small number of schools, from an established
background or with a set of defined skills (DiMaggio and Powell 1983).
At the individual level, professionalization spreads when organization members become
socialized into the norms of a profession (DiMaggio and Powell 1983). As individuals
limit their behaviors and choices to those consistent with the standards of their chosen
profession, they narrow the number of altemative practices from which they draw
solutions to organizational problems. As a result, such practices become further
professionalized, making them easier to codify, communicate, and teach to others.
Figure 19: Professionalization loop
professionalization
(" ay
codification
of practices
variance in
=
practices
~
A second type of institutionalization involves learning from others. Copying can be an
efficient means of addressing uncertainty. When an organization confronts technological
uncertainty, ambiguous goals, unclear problems and solutions, or difficulty in establishing
legitimation with others, its members may decide to adopt the practices or structures of
Archetypal self-reinforcing structures in organizations 28
another organization. In addition, powerful stakeholders may encourage an organization
to emulate practices of others in its field. The result is mimetic isomorphism, or a similarity
of organizational structures within an organizational field that results when organizations
model themselves after others(DiMaggio and Powell 1983).
Figure 18: Mimesis loop
modeling after selected
organizational form
en
likelihood that selected +
organizational form is
modeled prevalence of
organizational
KR form
Organizations become increasingly similar through the following process: First, the
prevalence of a particular organizational form drives the odds of its being selected as a
model. Models may be selected on the basis of perceived legitimacy, apparent success, or
convenience; the more common a form, the more likely it will be selected as a model. The
more often it is copied, the greater its frequency in the population. The loop representing
this process is labeled mimesis.
A third feedback process is associated with institutional isomorphism. Organizations are
subject to common coercive forces from other, more powerful organizations and from the
social system on which they are dependent. Such coercion takes the form of government
mandates, persuasion or social pressure, legal constraints, technical requirements,
Archetypal self-reinforcing structures in organizations 29
monopolistic control of resources, centralized banking practices, or the more subile
demands for ceremonial roles required for interacting with hierarchical organizations
(DiMaggio and Powell 1983). As the coercive power of institutions that influence others
increases, so does their ability to achieve compliance; increased compliance increases
dependence of others on coercers, thereby reinforcing the coercive power, as the coercive
power loop illustrates.
Figure 19: Coercive power loop
coercer’ s
= power
increase in
coercer’s power
Kom
KK tonmtnes
Diffusion, bandwagons, and fads
The literature on diffusion describes the process by which new ideas, technologies, or
practices spread through a population as new adopters imitate, learn from, or otherwise
glean information from others. Research on this topic examines how potential adopters
select others as sources of information in deciding whether to take up the innovation.
Bandwagons represent a special case of the diffusion processes, in which the number of
other adopters of an innovation is the key determinant of the decision to adopt (Rogers
1983; Bikhchandani, Hirshleifer, and Welch 1992; Abrahamson and Rosenkopf 1993;
Archetypal self-reinforcing structures in organizations 30
Valente 1995; Abrahamson 1996). People or organizations may decide to adopt if they
perceive pressure from peers or the public. Adopting the innovation may help them to gain
legitimacy— to be seen doing the right thing— or to catch up with competitors.
Diffusion may operate through social contagion or persuasion. Potential adopters rely on
the data from others selected as sources of information because of their physical
proximity, social stature, social similarity, or frequency of communication. The rate of
diffusion depends on a number of factors, including an individual or firm’s inclination to
adopt the innovation and the number of others who have already adopted it. The number
of adopters convinces potential adopters of an innovation’ s value through three distinct
processes: strength of word of mouth, apparent legitimacy, and positive network
extemalities.
Word of mouth may alter the attractiveness of an innovation by making the potential
adopter aware of the innovation and by permitting a vicarious trial (Burt 1987). Rather
than having too little information available, most modem innovators have too much
information. Identifying trustworthy others allows the potential adopter to cut search time
(Coleman, Katz, and Menzel 1982). The knowledge that trusted others are considering an
innovation is treated as an indication of its value. The more trustworthy or admired others
who have adopted the innovation, the more likely that a potential adopter will hear of an
innovation, including its consequences, through these informal channels. The intensity and
quality of rhetoric promoting the innovation promotes conversion. For example, in a
classic study of tetracycline use (Coleman, Katz, and Menzel 1982), the average
Archetypal self-reinforcing structures in organizations 31
physician's preference for the new drug and the number of doctors who had already
adopted the drug determined the diffusion rate; later studies added differing preferences
and the non-uniform availability of information across social networks (Burt 1987). A
system dynamics study of medical diffusion endogenized both actual and perceived
performance (Homer 1987).
In the word of mouth loop, the more adopters there are, the stronger the word-of-mouth
effect, and the more attractive the innovation. The process is self-reinforcing because the
greater number of adopters that result further reinforces the strength of the word-of-
mouth effect.
Figure 20: Word of mouth loop
number of
/ adopters
C r | +
rate of
adoption strength of
word of mouth
-
attractiveness
of innovation” +
Discussing an innovation with others allows a potential adopter to reach a normative
understanding of the costs and benefits of adoption to people or firms in their social role
(Burt 1987). A person risks embarrassment or sanctioning for being the last to adopt a
practice that has become a feature of occupying a status. Potential resources or
relationships may be withheld for failure to operate similarly to equivalent others.
Archetypal self-reinforcing structures in organizations 32
Legitimacy is enhanced by being perceived as complying with standard practice as defined
by the actions of competitors. A doption is likely to be a weighted function of the
innovation’s objective merits and other's social behavior. The less certain a firm or
individual is of the objective merits, the more heavily they weigh social information
(Abrahamson and Rosenkopf 1993). The result is the positive feedback process
represented in the legitimacy of adopting loop.
Figure 21: Legitimacy of adopting loop
number of
fo adopters ‘\
be
rate of
‘ apparent
adoption G¢ A legitimacy
+
ittractiveness,
of innovation
Separately from social expenses, network externalities may drive up the rewards or cut the
costs of adopting an innovation. Externalities— for example, production volume,
availability of information or support, ease of access— that are a function of the number of
current adopters affect the retums to following an innovation. Positive externalities
escalate retums non-monotonically. For example, early entry into microprocessor
production communities, which eventually had many participants, increased the market
share of early adopters (Wade 1995). Communities with higher sales early on attract
secondary sources which increase production capacity and make the market more
Archetypal self-reinforcing structures in organizations 33
attractive (p. 129). Support for complementary products increases the market for the
original product and may discourage competitors (p. 130).
Figure 22: Network externalities loop
ge adopters
rate of +
adoption (E positive network
in externalities
attractiveness of tine
innovation, poe
W__ adopting
+ innovation
Several factors can stop diffusion. The most obvious is the saturation effect: once most
people have adopted the innovation, fewer are left to adopt it, and the rate of adoption
necessarily slows. Sometimes people or firms reject an innovation because so many others
have adopted it (see Abrahamson and Rosenkopf 1993 on counter-bandwagons). This
snob effect may result in the introduction of fashion replacements. Alternatively, the
arrival of a little information to some or all potential adopters or the possibility of a change
in perceived value of an innovation may end the diffusion process (Bikhchandani,
Hirshleifer, and Welch 1992). Just as localized patterns of conformity may start a diffusion
chain, local patterns of dissent may break it.
Archetypal self-reinforcing structures in organizations 34
Discussion: Comparing processes
What do we leam from mapping the positive feedback processes at work in a variety of
organization theories? One idea unexplored in most of the existing theoretical work is that
the processes may operate at a variety of levels of analysis. For instance, learning by doing
may operate at the individual, group, organizational, institutional field, and societal levels,
although the mechanisms may vary— clearly the social processes that strengthen
organizational learning by doing cannot operate at the individual level, and leaming at the
institutional field level cannot rely on the same processes postulated for learning within
organizations, since the mechanisms of selection and control within an organization are
likely to be stronger than those within an industry. Another example is from the theory of
institutional isomorphism: the normative, mimetic, and coercive processes that result in
organizations within an industry growing more similar over time may well operate within
organizations as individuals facing uncertainty about how to proceed draw on normative
professional practices, mimic peers, or submit to the power of others and adapt their
behavior accordingly. Further development and testing of these ideas will be instrumental
in helping organizational scholars to build integrative, multi-level theories. It will also help
to sharpen our understanding of the mechanisms by which such processes operate in
organizations and organizational communities.
The juxtaposition of positive feedback processes is also suggestive of new research ideas.
Taken together, the twenty-two feedback loops presented here describe a wide range of
ways in which actions and decisions are reinforced in organizations and communities of
Archetypal self-reinforcing structures in organizations 35
organizations. Assembling and comparing them suggests several avenues of research that
could help us better understand how organizations change— or do not— over time.
In some situations, the feedback structures may interact with each other. For instance,
once information-processing, psychological and interpersonal processes that generate
escalation of commitment result in an organization’ s persisting with a course of action,
social processes at work in diffusion and isomorphism could encourage other
organizations to adopt the same course of action. Thus one set of processes may take over
from another when a critical threshold has been reached— in this case, once the
organization that is the early adopter achieves legitimacy as a result of its perseverance, it
may be selected as a model to emulate by other organizations.
Another promising idea is that processes at one level may shape (or even generate) those
at another level. For instance, norms and practices may spread throughout an organization
as members uncertain of how to proceed emulate others. The sharing of norms may help
further both learning by doing and the spread of culture and strengthening of socialization.
Thus the leaming-by-doing and culture-and-socialization feedback structures may rely on
underlying processes of emulation described by mimetic isomorphism.
A further research suggestion concems the association of processes. Once one self-
reinforcing feedback structure is observed, researchers and practitioners may draw on the
present collection of reinforcing processes to predict side-effects and additional feedback
processes that may be associated with it. The present analysis also highlights links between
Archetypal self-reinforcing structures in organizations 36
related feedback structures, suggesting that when one effect is discovered, others may be
associated with it. Learning by doing is linked with institutional inertia, for example: as
organizational members become better at implementing a chosen course of action,
consistent organizational norms, policies, and procedures evolve in concert— this is part of
the learning by doing process described above. As a result, the organization’s ability to
implement alternatives falls because the norms, policies, and procedures required to
support alternative courses of action are atrophied or never develop, reflecting the inertia
loop from escalation of commitment research. Thus the basic process of elaborating rules
and noms to support one way of doing things results in both learning and inertia.
In addition, with more systematic exploration of the theories laid out here (and in
particular with the formal modeling of processes posited to explain a given phenomenon),
some feedback structures may be ruled out as explanations for the observed dynamics— a
single reinforcing process may be enough to account for the observed history dependence.
For instance, while the research on escalation of commitment proposes ten reinforcing
processes to account for the tendency to persist with a losing course of action, only some
of those processes may be present in a given situation.
While the present exercise of mapping and comparing reinforcing feedback structures in a
variety of theories has generated intriguing proposals that may help explain change over
time (or the lack of it) in organizations, our analysis also suggests ways in which system
dynamics can make unique contributions to our understanding of organizations.
Archetypal self-reinforcing structures in organizations 37
In particular, formulating simulation models to explore these processes will expose
questions important in understanding how each operates and the situations in which each
applies. What are the initial conditions that shape the system when the reinforcing process
is getting established? What are the constraints that limit reinforcement? In other words,
what are the negative feedback loops that balance the system? Could other processes act
as substitutes? How do processes at one level—e.g., cognitive— interact with or generate
those at another—e.g., organizational?
Testing and extending the feedback processes presented here will help the field of system
dynamics identify a set of fundamental feedback structures that recur in a wide variety of
social systems. Assembling and validating archetypal organizational processes is an
important step in building a greater “culture of accumulation” (Richardson 1996). Because
these processes operate at specific cognitive, psychological, social, and institutional levels,
and because they are linked to theoretical an empirical research introduced here and
documented in more detail elsewhere (see Sastry and Coen 1998 for a discussion of the
related literature and a typology of the processes), they may offer a starting point for
understanding the system dynamics of organizations that is more useful than the
archetypes presented in the systems thinking literature.
The system dynamics approach may also help build understanding in the five areas of
research identified above. As others extend and formalize the feedback loops presented
here, the processes at work will be clarified and ambiguities and inconsistencies in the
original research exposed. Building simulation models that incorporate the reinforcing
Archetypal self-reinforcing structures in organizations 38
loops will validate the underlying theory in ways that complement traditional methods of
other fields; and documenting other, interacting loops may help to extend the theories
which often fail to explain the limits to self-reinforcing processes in organizations. The
result will be an increase in the quality of organizational theory, better linkages between
research in system dynamics and more traditional areas of organizational studies, and
greater insight into the often vexing problem of resistance to change in organizations.
Archetypal self-reinforcing structures in organizations 39
References
Abrahamson, E. 1996. Management fashion. Academy of Management Review 21 (1):
254-285.
Abrahamson, E., and L. Rosenkopf. 1993. Institutional and competitive bandwagons:
Using mathematical modeling as a tool to explore innovation diffusion. Academy of
Managenent Review 18 (3): 487-517.
Aldrich, H. 1979. Organizations and environments. Englewood Cliffs, NJ: Prentice Hall.
Argote, L., and D. Epple. 1990. Learning curves in manufacturing. Science 247: 920-924.
Argote, L., S. L. Beckman, and D. Epple. 1990. The persistence and transfer of learning in
industrial settings. Management Science 36: 140-154.
Arkes, H. R., and C. Blumer. 1985. The psychology of sunk cost. Organizational
Behavior and Human Decision Processes 35: 124-140.
Bellinger, G. 1998. Archetypes: Interaction structures of the universe. W orld wide web
site: www. radix.net/crbnblu/musings/arch/arch.htm.
Bikhchandani, S., D. H. Hirshleifer, and I. W. Welch. 1992. A theory of fads, fashion,
custom, and cultural change as informational cascades. Journal of Political Economy 100
(51): 992-1026.
Boeker, W. 1988. Organizational origins: Entrepreneurial and environmental imprinting at
the time of founding. In Ecological Models of Organizations, ed. G. Carroll (pp. 33-51).
Cambridge, MA: Ballinger Publishing.
Boeker, W. 1989a. The development and institutionalization of sub-unit power.
Administrative Science Quarterly 34: 388-410.
Boeker, W. 1989b. Strategic change: The effects of founding and history. Academy of
Managerrent J ournal 32 (3): 489-515.
Archetypal self-reinforcing structures in organizations 40
Brockner, J. 1992. The escalation of commitment to a failing course of action: Toward
theoretical progress. Academy of Management Review 17 (1): 39-61.
Burt, R. S. 1987. Social contagion and innovation: Cohesion versus structural
equivalence. American J ournal of Sociology 92 (May): 1287-1335.
Coleman, J. S., E. Katz, and H. Menzel. 1982. Medical Innovation: A Diffusion Study.
Indianapolis: Bobb-Memill.
DiMaggio, P. J., and W. W. Powell. 1983. The iron cage revisited: Institutional
isomorphism and collective rationality in organizational fields. American Sociology Review
48: 147-160.
Dowling, A. M., R. H. MacDonald, and G. P. Richardson. 1995. Simulation of systems
archetypes. Proceedings of the 1995 International System Dynamics Conference, Tokyo.
Lincoln MA: System Dynamics Society.
Eisenhardt, K. M., and C. B. Schoonhoven. 1990. Organizational growth: Linking
founding team, strategy, environment, and growth among U.S. semiconductor ventures,
1978-1988. Administrative Science Quarterly 35: 504-529.
Forrester, J. W. 1980. System Dynamics— Future opportunities. In System Dynamics, ed.
Legasto, A. A., J. W. Forrester, and J. M. Lyneis, TIMS Studies in Management Science,
vol. 14. New Y ork: North-Holland.
Garland, H. 1990. Throwing good money after bad: The effect of sunk costs on the
decision to escalate commitment to an ongoing project. Journal of Applied Psychology
75: 728-731,
Garud, R., and M. A. Rappa. 1994. A socio-cognitive model of technology evolution: The
case of cochlear implants. Organization Science 5 (3): 344-362.
Archetypal self-reinforcing structures in organizations 41
Goodman, M., and J. Kemeny 1994. Applying an archetype. In The Fifth Discipline
Fieldbook: Strategies and Tools for Building a Learning Organization, ed. P. M. Senge,
A. Kleiner, C. Roberts, R. B. Ross, and B. J. Smith: 121-125. New Y ork: Doubleday.
Hannan, M. T., andJ. Freeman. 1984. Structural inertia and organizational change.
Aterican Sociology Review 49: 149-164.
Heath, C. 1995. Escalation and de-escalation of commitment in response to sunk costs:
The role of budgeting in mental accounting. Organizational Behavior and Human
Decision Processes 62 (1): 38-54.
Homer, J. B. 1987. A diffusion model with application to evolving medical technologies.
Technological Forecasting and Social Change 31: 197-218.
Kim, D. H. 1996. Systems archetypes: Diagnosing systemic issues and designing high-
leverage interventions (reprinted from Toolbox Reprint Series: Systems Archetypes,
Cambridge MA: Pegasus Communications). In Modelling for Management 1: Simulation
in Support of Systems Thinking, ed. G. P. Richardson: 153-164. Aldershot UK:
Dartmouth.
Kimberly, J. 1979. Issues in the creation of organizations: initiation, innovation and
institutionalization. Academy of Management J ournal 22 (4): 37-457.
Lane, D. C., and C. Smart. 1996. Reinterpreting ‘generic structure': evolution, application
and limitations of a concept. System Dynamics Review 12 (2): 87-120.
Mashayekhi, A. N. 1993. Transition in the New Y ork State solid waste system: a dynamic
analysis. System Dynamics Review 9 (1): 23-47.
Masuch, M. 1985. Vicious circles in organizations. Administrative Science Quarterly 30:
14-33.
Mintzberg, H., and J. Waters. 1982. Tracking strategy in an entrepreneurial firm. Academy
of Management 25 (3): 465-499.
Archetypal self-reinforcing structures in organizations 42
Morecroft, J. D. W. 1982. A critical review of diagramming tools for conceptualizing
feedback system models. Dynamica 8 (1): 20-29.
Naill, R. F. 1992. A system dynamics model for national energy policy planning. System
Dynamics Review8 (1): 1-19.
Nelson, R. B. 1994. The co-evolution of technology, industrial structure, and supporting
institutions. Industrial and Corporate Change 3 (1): 47-63.
Nelson, R. R., and S. G. Winter. 1982. An Evolutionary Theory of Economic Change.
Cambridge, MA: Harvard University Press.
Richardson, G. P. 1986. Problems with causal-loop diagrams. System Dynamics Review 2
(2): 158-170.
Richardson, G. P. 1991. Feedback Thought in Social Science and Systems Theory.
Philadelphia: University of Pennsylvania.
Richardson, G. P. 1996. Problems for the future of system dynamics. System Dynamics
Review 12 (2): 141-157.
Rogers, E. M. 1983. Diffusion of Innovations. New Y ork: Free Press.
Romanelli, E. 1989. Environments and strategies of organizational start-up: Effects on
early survival. Administrative Science Quarterly 34 (3): 369-387.
Ross, J., and B. M. Staw. 1993. Organizational escalation and exit: Lessons from the
Shoreham nuclear power plant. Academy of Management J ournal 36 (4): 701-732.
Sastry, M. A. and C. Coen. 1998. Tracing the hand of the past: Towards an integrated
theory of how history matters in organizations. Working paper. University of Michigan
Business School.
Schein, E. H. 1983. The role of the founder in creating organizational culture.
Organizational Dynamics 12: 13-28.
Archetypal self-reinforcing structures in organizations 43
Senge, P. M. 1990. The Fifth Discipline: The Art and Practice of the Learning
Organization. New Y ork: Doubleday.
Senge, P. M., A. Kleiner, C. Roberts, R. B. Ross, and B. J. Smith. 1994. The Fifth
Discipline Fieldbook: Strategies and Tools for Building a Learning Organization. New
Y ork: Doubleday.
Staw, B. M. 1976. Knee deep in the big muddy: A study of escalating commitment to a
chosen course of action. Organizational Behavior and Human Performance 16: 27-44.
Staw, B. M. 1981. The escalation of commitment to a course of action. Academy of
Management Review 6: 577-587.
Staw, B. M., and H. Hoang. 1995. Sunk costs in the NBA: Why draft order affects
playing time and survival in professional basketball. Administrative Science Quarterly 40:
474-494.
Staw, B. M., and J. Ross 1987. Behavior in escalation situations: Antecedents, prototypes,
and solutions. In Research in Organizational Behavior, ed. B. M. Staw and L. L.
Cummings: 39-78. Greenwich, CT: JAI Press.
Staw, B. M., andJ. Ross. 1978. Commitment to a policy decision: A multi-theoretical
perspective. Administrative Science Quarterly 23: 40-64.
Staw, B. M., and J. Ross. 1989. Understanding behavior in escalation situations. Science
246: 216-220.
Sterman, J. 1994. Beyond training wheels. In The Fifth Discipline Fieldbook: Strategies
and Tools for Building a Learning Organization, ed. P. M. Senge, A. Kleiner, C. Roberts,
R. B. Ross, and B. J. Smith: 177-183. New Y ork: Doubleday.
Sterman, J. D., and G. P. Richardson. 1985. An experiment to evaluate methods for
estimating fossil fuel resources. Journal of Forecasting 4: 197-226.
Archetypal self-reinforcing structures in organizations 44
Sterman, J. D., N. P. Repenning, and F. Kofman. 1997. Unanticipated side effects of
sucessful quality programs: Exploring a paradox of organizational improvement.
Management Science 43 (4): 503-521.
Stinchcombe, A. L. 1965. Social structure and organizations. In Handbook of
Organizations, ed. J. G. March: 153-193. Chicago, IL: Rand McNally.
Tushman, M. L., and E. Romanelli 1985. Organization evolution: A metamorphosis model
of convergence and reorientation. In Research in Organizational Behavior, ed. B. M.
Staw and L. L. Cummings: 171-222. Greenwich, CT: JAI Press.
Valente, T. W. 1995. Network Models of the Diffusion of Innovations. Cresskill NJ:
Hampton Press.
Van de Ven, A., and R. Garud 1994. The coevolution of technical and institutional events
in the development of an innovation. In The Evolutionary Dynamics of Organizations, ed.
J. A.C. Baum and J. Singh: 425-443. Oxford: Oxford University Press.
Wade, J. 1995. Dynamics of organizational communities and technological bandwagons:
An empirical investigation of community evolution in the microprocessor market.
Strategic Management Journal 16: 111-133.
Archetypal self-reinforcing structures in organizations 45
CC BY-NC-SA 4.0