DEVELOPING A THEORY OF
SERVICE QUALITY/SERVICE CAPACITY INTERACTION
Peter M. Senge and Rogelio Oliva
Center for Organizational Learning, MIT
Room E40-294, 77 Massachusetts Ave.
Cambridge, MA 02139, U.S.A.
ABSTRACT
Service quality cannot be measured and tested in’as straight forward’a manner as in
manufacturing. This biases service businesses to focusing on keeping measurable
variables—typically, expenses and work flows—in control, while underinvesting in the
intangibles of service capacity and service quality. In the long-term, results can be
mediocre levels of service quality, poor customer satisfaction, high turnover of service
personnel, and, ultimately, higher total costs. In this paper we will present an emerging
theory of interactions between Service Quality and Service Capacity, relate this theory to
past research in both the System Dynamics and Total Quality Management traditions,
and outline ongoing empirical testing of the theory.
FOCUS OF THE RESEARCH
At the Center for Organizational Learning, we are engaged in developing new tools and methods to help
organizations develop systems thinking and related learning capabilities. Present research is focused in
three broad substantive areas: managing new product development efforts, understanding determinants
of cycle time in complex supply chains, and managing quality in service businesses and developing and
implementing “value-adding services" strategies.
In each of these areas, better theories are needed to guide development of better management tools and
methods. Each of the areas represents a complex dynamic context in which key variables are
interrelated through multiple feedback interactions. Delays, nonlinearities, and non-obvious dynamics
confound. the efforts. of decision makers to bring about improvement through typical management
interventions. Without well-developed theories that reveal the dynamic structures underlying problems,
managers are prone to focus on problem symptoms rather than areas of high leverage. Without explicit
theories, different managers will be guided by different mental models in their improvement efforts.
Without explicit theories that can be continuously tested and improved, it is unlikely that interventions
will produce new understandings that can be captured and serve as a starting point for further learning.
Current research at the Center for Organizational Learning is focused on developing better theories in
each of the three subject areas mentioned above. Specifically, we are developing "system dynamics"
computer simulation models and interactive "management flight simulators" embodying generic
theories of the dynamics of new. product development, complex, supply chains, and service quality.
These generic theories are being tested in multiple organizational settings as part of the collaborative
efforts of the Center with its organization partners. Eventually, these simulators and underlying
dynamic models will be disseminated widely to aid continuing research and education.
. The purpose of this paper is to lay out a» emerging theory of interactions between service quality and
service capacity, relate this theory to past research, and outline the empirical tests we plan to conduct of
the theory. A subsequent paper will lay out a related theory of implementing value-added services
strategies, which is guiding other aspects of the service quality projects. ~
HISTORICAL DEVELOPMENT OF THE THEORY
The theory of service quality-service capacity interactions that is guiding present research has been
developed over several years. The first articulation of the theory emerged in the context of a multiple-
476 SYSTEM DYNAMICS '93
year study with a leading property and liability insurance company. That study!, which resulted: in a
“claims management learning laboratory" that operated for several years within the firm, focused
attention on the rising costs of claims settlements and litigation and related costs, and the declining
overall financial health of the industry. These costs trends have:been- developing for many years.
Within the industry, rising settlement and litigation costs are often blamed on external factors such as
the high number of lawyers in the US, increasing litigiousness of society, the tendency for juries to side
with victims rather than 'big business’ insurers, and increasing risks born of technological complexity
(such as toxic waste). Our study illuminated internal sources of the problems.
During the last 50 years there has been a rising trend in "loss ratios" (settlement costs and litigation
costs relative to premiums) and a falling trend in "expense ratios" for the whole insurance industry
(Moissis, 1989). One might interpret the falling expense ratios as evidence of increasing productivity
and management innovation. Our work suggests something different. While not denying that there
have been some increases in productivity gained from technological innovation in underwriting, claims
processing, and customer service, the central hypothesis that has emerged from our work is that the
rising settlement costs and the falling expense costs are causally related: there has been a long term
trend of underinvestment in service capacity that has resulted in erosion of quality of investigation,
negotiation, and customer service, resulting in rising costs of settlement and litigation. Moreover, the
savings in expenses have been more than offset by the increases in costs of poor quality. The
consequent long-term increase in total costs and erosion in profitability have led to increasing focus on
expense control and "productivity" (normally defined as customers served per service person), thereby
reinforcing underinvestment in service capacity.
A good theory, according to the system dynamics paradigm, links observable "macro", i.e., system-
wide, patterns of behavior to "micro" decision making. Our first efforts to develop the above
hypothesis focused on showing how the underinvestment dynamic could emerge from interactions
among goals, norms, performance measures, and pressures that managers in the insurance industry
could identify in their own experience. A team led by the Vice President of Claims in the sponsor
company worked with us to develop a system dynamics model showing how established management
practices and policies could produce underinvestment and rising total costs. The process ‘whereby the
initial model was developed is described by Senge (1990a, p 216-217):
“The key to the hypothesis lay in distinguishing two classes of performance measures:
"production standards" and "fuzzy standards." Production standards are measures such as
“production ratio" and "pending ratio," which indicate whether current claims pending are
settled at a rate commensurate with the inflow of incoming claims. The production standards
are relatively easy to measure, are understood by everyone in the business, and send out clear
immediate warning signals when they become out of balance. The fuzzy standards include
quality of investigation, file quality, effective oversight of litigation.and subrogation
(recovery of costs from other insurers), and service quality. The fuzzy standards are difficult
to measure. Though there is widespread appreciation that the fuzzy standards are important,
the team felt that there is usually considerable uncertainty as to how well a claims office is
doing on the "fuzzies." Because they are easier to measure, the team felt that there were
natural pressures to manage by the production measures. As the vice president put it,"In this
business there are lots of ways to look good without being good."
Initial simulation tests of the service quality-service capacity model showed how focusing on
production standards could be problematic under times of stress. In particular, two simulation tests
showed that it was impossible to distinguish two different adjustment mechanisms that might operate in
response to an increase in incoming claims if one focused only on production measures. In one case,
production measures readjust to acceptable balances because of increasing adjuster capacity. In the
other case, they readjust because of eroding fuzzy standards. Thus, if management tracks only the
production measures, it is impossible to know what is going on at a deeper level: desired levels may be
1 Complementary accounts of this study can be found in (Kim,.1991; Senge, 1990a; Senge, 1990b;
Senge & Lannon, 1990; Senge & Sterman, 1992)
SYSTEM DYNAMICS '93 477
maintained only because. of eroding quality. This model behavior reflected the vice president's
statement that it is "easy to look good without being good." In a simulation where incoming claims
grow steadily, there is a rising volume of claims settled per adjuster along with a steady decline in fuzzy
standards, a behavior pattern which matches qualitatively the historical pattern of falling expense ratios
and rising loss ratios (Senge, 1990a, pp. 220-227).
The basic formulation for the insurance industry is applicable to a set of service settings where
providing the service involves highly intangible actions like building trust with the customer, where
professional skills and considerable experience are required to successfully provide the service, and
where complex personal.interactions between server and customer are involved. In the four years since
the original theory of capacity-quality-cost interactions was developed for insurance application, the
basic model has been used in workshops for hundreds of managers from diverse service industries.
This has led to a recasting of the original model as a generic theory. The full theory can be summarized
by the following propositions: 2
1. In service businesses it is always difficult to measure quality because it is intangible and subjective.
2. Consequently, there is a tendency to manage service businesses by what is more measurable—
notably, expenses and production figures.
3. This leads to a systematic bias toward underinvestment in "service capacity"-the ability to provide
services at a given quality level, which is a function of number of people, experience levels, skills,
and supporting infrastructure. Decisionmakers tend to assess whether or not capacity is adequate
based on expenses and production figures, which may be unrelated to service quality.
4. The consequence of underinvestment is low levels of service relative to what is possible, high costs
of poor quality (e.g., rework), low customer loyalty, high turnover. of service personnel,-and
mediocre financial performance.
5. Underinvestment in service capacity is frequently masked by eroding operating standards, so that
serving people and customers come to expect mediocre service and justify current performance
based on past performance, rather than on absolute standards or goals.
6. As entire. industries become locked in cycles of underinvestment and eroding standards, industry
norms reinforcing expense control and "productivity" become more and more influential in shaping
individual firm decisions.
The generic theory of service quality and service capacity. has been elaborated in a system dynamics
model. An overview of the model and the essential feedback relationships expressing the theory are
presented below.
SYSTEM DYNAMICS ARTICULATION OF THE GENERIC THEORY
The Service Quality/Service Capacity model simulates a
“Service Center” where customers enter the system and,
after a waiting-time, are served by the Center’s employees.
Service capacity -- i.e., service personnel, years of
experience, skill and motivation -- is required to provide that
service; the desired amount of capacity is determined by the
desired level of quality, and the desired throughput of the
Service Center. If a particular request is not satisfied to the
customers’ standards, it comes back into the Service Backlog
and has to be reprocessed as rework.
To understand the explicit, formulations of the Service
Quality/Service Capacity model, it is best to differentiate
four subsystems in the model: Service Capacity, Service
2 The following presentation is based on the “Service Quality Management. Flight Simulator
Facilitators’ Training Guide”, prepared by Rogelio Oliva (1992), available from the MIT
Organizational Learning Center.
478 SYSTEM DYNAMICS '93.
Backlog, Quality/Time Pressure and
Market Response. Figure No. 1
shows these subsystems and the main
relationships among them. Each of the
subsystems will be explained with
more detail below.
‘apacit tem
The capacity sector assumes a chain of
experience structure for the
development of capacity. Personnel
hired recently (Rookies) will not be as
effective as Sr. Personnel until they go
through a training and experience-
Effective Tumover Rookies’ Tumover
Training Time
Sr Personnel Tumover
Figure No 2
building period (see Fig. No. 2). Furthermore, the training of Rookies will demand some time from the
Sr. Personnel, reducing further the total capacity of the system. .
Personnel turnover is determined by a normal turnover rate that is further modified by two factors: a) a
Time
(~ Pressure ‘
Work
Effective
Time
i ae “y
3
Capacity R a, Employee's
Index Perception of
Service Quality
‘Turnover
Figure No. 3
Burnout Index that is the accumulation of stress
due to the work intensity(Homer, 1985), and b)
the Employees Perception of Service Quality.
The Human Resources literature has extensive
research that shows that employees will support
more pressure and develop greater loyalty to the
organization if they perceive a high service
quality (Schneider, 1991; Schneider, Parkington,
& Buxton, 1980; Tornow, 1991). The
relationships between these indicators can be seen
in Fig. No 3.
The current version of the simulator does not
include the impact of macroeconomic indicators,
i.e., unemployment, in the turnover rate. In the
simulator, the Hiring rate is left as a decision for
the user of the simulator although is can be set to
keep a constant head-count. In the full simulation
model, hiring rates are determined by customer flows and work backlog pressures (Senge 1990a).
jervice lo; si
The Service Backlog Subsystem keeps track of all the customer orders as they flow through the Service
Center. Two stocks—Service Backlog and Undiscovered Rework- are used to model those dynamics.
Figure No. 4 shows the main variables of this sector and their relationship.
One of the strongest assumptions in this particular model is that to deliver quality more time is required
from the service provider. This implies that the Work Completed is reduced if the Actual Quality is
increased. This was clearly the case for the claims adjusters because in order to do a better
investigation, and keep more complete and accurate records, they had to spend more time with each
claim. It is generally the case in service settings where customer interactions are important and
perceived service quality suffers if customers feel rushed. In such settings, even though technology
may improve efficiency in some aspects of the service provision, it cannot substitute for human contact,
The implications of this assumption are going to be explored in the next section.
According to this diagram, the greater the Time Pressure, the lower the Actual Quality —employees
will not have time to inspect their work or to respond to specific queries. This, assuming a constant
work completed, will cause a higher rate of Mistakes that, after a period, are discovered and come back
to the system as Rework, increasing the Service Backlog. A higher Service Backlog will be translated
SYSTEM DYNAMICS '93 479
into higher Production Goal as
management seeks to reduce Customer____y, Service
backlogs, and higher Time Orders — Ss Backlog’
Pressure, thus reinforcing
poorer quality. The way the ae
system is formulated the Be ria
balancing loop from Production
Goal to Work Completed is SpP roi ction
always ‘stronger than the {
reinforcing loop. However, it is
this structure that makes the
aii,
Work s pe
ep fommleter
pursuit of higher quality an R
attractive goal, as it becomes a be
self-reinforcing virtuous cycle Actual
that increases productivity and Required Quality:
reduces cost (Deming, 1982). 4 ~
Effective 8 ‘Time 5 A
Quality/Time Pressure Time ——<p mployee's
Subsystem Acailable Pressure Perception of
Service Quality
This subsystem does not =
represent a physical department Figure No. 4
of the Service Center, nor a
specific function of it. The Quality/Time Pressure Subsystem captures the tradeoffs and relationships
between quality and time pressure. The speed at which a quality improvement process will achieve its
objectives, and the impacts on quality of an increase of demand will be determined by this subsystem.
To capture the dynamics of the service quality, it was necessary to define three different indicators for
it.
Quality Goal: Reflects the management's desired level of quality. The Quality Goal is translated into
pressure to modify the Quality Standard.
Quality Standard: The current Quality Standard is the employees’ and management's perception of
what is acceptable service quality, i.e, the quality level that the employees would perform under
normal Time Pressure and Work Intensity.
Actual Quality: Actual Quality is the level of quality that the customers are receiving. The Actual
Quality is based on the Quality Standard and Time Pressure.
The balancing loop in Figure No. 5 represents the
relationships between quality and workload. The
higher the Quality Standard is set, the more
Effective Time Required to perform the job (more
inspections, more quality improvement activities or
Quality__gy Quality
Goal B rN
s
Mi Actual more time spent in each transaction). This will,
Biers B Quajty assuming a constant Productive Time Available,
. re increase the Time Pressure. Because of hastiness
Required induced by the Time Pressure, errors could be
. expected thus having a negative impact on the
s Time Actual Quality achieved, If the Actual Quality
Pressure level is lower than the Quality Standard for.a long
Figure No, 5 period, the standard. then will begin to deteriorate
as employees ‘get used’ to the lower quality level.
The reinforcing loop between Quality Standard and Actual Quality has different strengths depending on
whether it is functioning as a vicious or a virtuous cycle. If the Actual Quality is lower than the Quality
Standard and there is no pressure from the Quality Goal, the Quality Standard will begin to drift
downward causing the Actual Quality to drop-even under constant Time Pressure. On the other hand,
480 SYSTEM DYNAMICS '93
if the Quality Standard is set at a value Internal Market Indicator | Perception
where the Time Pressure is less than variable Delay
normal, the Actual Quality achieved will be
higher that the Quality Standard. This does Work Backlog, ‘Average, Wain 4
not mean that the Quality Standard will Actual Quality | Castor - r é
increase with the Actual Quality, unless, of ctual Quality a Servic ratty
course, there is some pressure from the - of Service Quality
* Time Pressure | Sales Effort from the 1
Quality Goal. .
Service Personnel
Market Response Subsystem Table No. 1
The Relative Attractiveness of the service
provided by the Service Center is determined by three market. indicators that are linked, through
perception delays and biases, to internal variables of the Service Center (see Table No. 1). All these
indicators form negative loops regulating the Customer Orders through the Relative Attractiveness of
the service.
COMPARISON TO OTHER MODELS AND THEORIES OF SERVICE QUALITY
As the Service sector is becoming increasingly important in the US economy (Cohen & Zysman, 1987;
Quinn & Gagnon, 1986), more time has been spend in trying to develop guidelines to manage services.
In this section we will identify how our dynamic theory relates to other emerging service quality models
and theories.
Much of the research’effort on the service industry, specially in the marketing and operations
management literature, has focused in determining the nature of service and identifying the different
kinds of settings under which services are provided. There are many classifications of services available
(Haywood-Farmer, 1988; Lovelock, 1983; Schmenner, 1986), each focusing in additional distinctions
in order to define strategic or operational guidelines for the service managers. It is difficult, however, to
build a consistent view of service quality from all those models and classifications. Perhaps, the
important lesson from this proliferation of service models is that the determinants of service quality will
vary with the type of business studied. The different classifications of services, are however, useful to
explore the limitations and shortcomings of the model presented here.
a) As discussed above, the current version of the model assumes that an increase in Actual Quality
will necessarily imply a lower rate of Work Completed. It is, however, difficult to make the
generalization of this relationship since one of the strongest argument in the Total Quality
Management literature is that total cost is reduced through better quality (Crosby, 1979; Deming,
1982; Juran, Gryna, & Bingham, 1974). The model, as it stands, does reflect some reduction of
total cost if Actual Quality is improved—through reduced Rework and lower Personnel Turnover.
The model, however, is not capable of capturing the increase on productivity through the
standardization process that might emerge after a technological or process improvement of quality.
The current formulation of the theory takes the only indicator of capacity to be personnel and their
effectiveness. It is not possible to capture other ‘capacity increasing’ investments. This particular
conceptualization of capacity limits the learning curve to enhanced productivity through work
experience.
b)
Although formulations to deal with those limitations can easily be incorporated into the model, the fact
that we have not done it limits the settings in which the model can be used. The model is currently best
suited to represent service centers where the service interaction is complex, the average waiting time is
in the order of days or weeks, not hours or minutes, and the service provided depends on highly trained
personnel.
Regardless of the diversity of service settings, there are a core of concepts widely accepted as intrinsic
to ‘ices (Chase, 1981). Among: those concepts is the recognition that services are intangible
activities, thus the difficulty of measuring their quality. This recognition of intangibility has brought up
the recognition that service quality perceptions result from a comparison of customer expectations with
SYSTEM DYNAMICS '93 481
service performance (Maister, 1984). The most articulated model of this perspective (Parasuraman,
Zeithaml, & Berry, 1985) argues that the difference between the customer expectations and the actual
service provided cannot be managed directly but through other “gaps,” or discrepancies, between
expectations and performance that occur in organizations. Figure No. 6 is the graphical representation
of these.gaps.
Gap 1 the difference between what consumers expect and what management perceives them
to expect,
Gap 2..the. difference between management’s perceptions of consumer expectations (Quality
Goal in our model) and actual service quality specifications (Quality Standard); ;
Gap 3 the difference between service quality specifications and the service actually
delivered (Actual Quality),
Gap 4 the difference between service delivery and what is communicated about the service
to consumers, an
Gap 5 the difference between the customers’ perception and their expectations of the
service,
‘Word of Mouth [Personal Needs] [Past Experience ] Our model, as it stands, assumes
co perfect knowledge of the
customer and only explores the
dynamics of the internal
organization—Gaps 2 and 3
CONSUMER (shaded area in Figure No. 6).
RS Reeceenens We believe that these dynamics
MARKETER Coie are taken-for-granted in most of
to Consumers the Total Quality Management
ry literature and it is worth
exploring them because they
have first order effects in
practical managerial settings.
We believe that our theory,
although limited in the scope of
total quality, addresses
significant interdependencies that
constitute a meaningful "internal
perspective” on service quality.
Figure No. 6
In relation to other studies of service quality, this theory has certain distinct advantages, namely
1. -Endogenous theory of dynamics: the theory, when simulated, recreates a pattern of
underinvestment, eroding quality, increasing costs of poor quality and deteriorating overall
financial performance that occurs in many real service industries.
2. All assumptions are explicit: true for other mathematical models of services, but most of these
are not fully dynamic.
STRATEGY FOR BUILDING CONFIDENCE IN THE THEORY
The generic theory of service quality-service capacity interactions is in a formative stage. The
insurance case provided one context for development and testing. The results were sufficiently
encouraging to pursue other contexts. In those settings we will endeavor to build on and go further in
testing the theory than was feasible in the first setting.
Understanding a system dynamics theory requires an. intuitive appreciation of “its dynamics." The
model is an artifact that expresses a theory in an interactive form. The theory can therefore only be
fully understood through interacting with it. This is why actual simulation experience is so important to
482 SYSTEM DYNAMICS '93
the validation process. Building confidence in the theory is a continual process of deepening
understanding of its dynamics and checking against experience in the system being studied, often
against the experiences of people who live in the system.
This is why the "validation process" cannot be reduced to a few simple tests or summary presentations
of how the model behaves. Rather it is a never-ending process of performing increasingly diverse tests
(Forrester & Senge, 1980; Sterman, 1988). For example, for the claims management simulator
described above, there are basic."model behavior tests" that seek-to determine that the model can
reproduce salient historical behavior patters like rising loss ratios, without the aid of external inputs that
might predetermine such results. As illustrated, the original model passed such tests. But, validation
also requires "policy testing" to ascertain how the model responds to different changes in management
policies. For example, total costs in the model do not improve simply by increasing quality goals. In
fact, elevating quality goals alone can be counterproductive, leading to overwork, burnout, increased
turnover and increasing delays in customer service.
The basic strategy for continuing to test the generic service capacity-service quality theory:is to use it as
a basis for theory development and intervention in a variety of service companies participating in the
Center for Organizational Learning. In each case, the basic precepts and interdependencies
incorporated into the theory will be held up against the specific world of a particular group of managers
in a particular industry. Developing learning laboratories like that developed in the original insurance
study will guarantee that the theory will come in contact with a large number of managers. Hopefully,
it will also stimulate a variety of organizational experiments based upon its implications, which will
provide further data regarding the model's usefulness and reveal limitations and flaws that can lead to
further improvement.
The general approach that we have decided to take is to explore the ‘range of usefulness of the model.”
This might be done by testing the overall behavior and implications with the managers, and verify the
transferability of the model (assumptions) through several different service settings in an empirical way.
The overall objectives of the empirical testing are to
1. gain more experience in the practical utilization of the theory
2. gain better understanding of the theory's strengths and limitations
3. provide stimulus to develop new and better theories and better practical tools.
The range of activities now planned are summarized below. As with all validation strategies, each
particular type of test has its own limitations, which are also briefly summarized.
1. Organize conceptualizing sessions where managers can articulate their perceptions of service-
quality-service capacity interactions in feedback terms that can be related to the theory.
Limitation: can get a wide variety of inputs based on very different mental models;
commensurability with core theory can be challenging '
2. Collect data in companies that might assist in further testing of overall patterns of model
behavior.
Limitation: available numerical data are often limited
3. Develop Learning Laboratory in each company, observe decision making behavior within
these labs, and relate to decision rules incorporated into full simulation model.
Limitation: do people behave in experimental setting in a way that is representative of their
behavior in real decision making settings’?
4. Test policy implications of theory through working with managers to translate implications,
implement changes, and study the consequences of those changes.
Limitation: messy "data" that comes from interventions and assessing the consequences of
interventions in the presence of confounding factors and long delays
Ultimately, this leads to an approach to validation which puts theory testing within the context of
organizational learning. The roots of the term "valid" are in law and have to do with the logical
consistency and usefulness of an argument, and not contradicting known facts. Thus, it makes very
good sense to pursue a validation strategy of having managers interact with a model, deepen their
intuition about its dynamics, draw conclusions about possible implications for their management
SYSTEM DYNAMICS '93, 483
practice, undertake experiments to test out those conclusions in the real system, and feed back new
insights that enrich the basic theory. Kofman (1992) has articulated the overall organizational learning
process in terms of observation, assessment, design and implementation or intervention. We believe
that this process can also serve to test theories guiding that learning,
In this-context, the system dynamics theory is a tool to foster new, more insightful "observations" based
on.a more systemic interpretation of the “data of experience." It is also to be used in assessing likely
effects of alternative actions that might be taken to improve system performance and in designing
specific interventions. These actions usually take the form of new organizational structures, rewards,
performance measures or other changes in operating policies. The merits of the theory are then
ascertained through studying the effects of an actual implementation of new policies.
To illustrate, the original claims management learning laboratory was not intended to dictate particular
changes that managers should make but rather to raise important questions about interdependencies
among costs, quality, and production. The intention was to send managers back into their field
operations strategizing about how they might improve total costs.. For example, one participant
reported:
“When I came back from the learning laboratory, I had a much better understanding of what
the important issues were. Before the lab, I would have said that lack of quality was the only
important factor. After the lab, it was obvious to me that productivity was also a key issue.
So I restructured some units to enhance their ability to settle claims. After I saw dramatic
increases in productivity [in the real organization], I applied pressure to.improve quality—and.
Thave seen a difference." (Bergin & Prusco, 1990).
REFERENCES
Bergin, R. and G. Prusco. 1990. The Learning Laboratory. The Health Care Forum Journal, Vol. 33
(2): 32-36.
Chase, R. B. 1981. The Customer Contact Approach to Services: Theoretical Bases and Practical
Extensions. Operations Research, Vol. 29 (4): 698-706.
Cohen, S. S, and J. Zysman. 1987. Manufacturing Matters: The Myth of a Post-Industrial Economy .
New York: Basic Books, Inc.
Crosby, P. B. 1979. Quality is Free . New York: New American Library.
Deming, W. E. 1982. Out of the Crisis . Cambridge, MA: MIT. Center for Advanced Engineering
Study.
Forrester, J. W. and P. M. Senge. 1980. Tests for Building Confidence in System Dynamics Models.
TIMS Studies in the Management Sciences, Vol. 14 (1980): 209-228.
Haywood-Farmer, J. 1988. A Conceptual Model of Service Quality. International Journal of
Operations and Production Management, Vol. 8 (6): 19-29.
Homer, J. B. 1985. Worker Burnout: A Dynamic Model with Implications for Prevention and Control.
System Dynamics Review, Vol. 1 (1-Summer 1985): 42-62.
Juran, J. M., F. M. J. Gryna and R. S. Bingham (Ed.). (1974). Quality Control Handbook. Third
Edition. New York: McGraw-Hill Book Company.
Kim, D. 1991. Systemic Quality Management: Improving the Quality of Doing and Thinking. The
Systems Thinker, Vol. 2 (7): 1-4.
Kofman, F. 1992. Lecture Slides. MIT Sloan School of Management. Cambridge, MA
Lovelock, C. H. 1983. Classifying Services to Gain Strategic Marketing Insights. Journal of Marketing,
Vol. 47 (Summer 1983): 9-20.
Maister, D. H. 1984. Quality Work Doesn’t Mean Quality Service. The American Lawyer, Vol. April
(1984): 6-8.
484 SYSTEM DYNAMICS '93
Moissis, A. A. (1989) Decision Making in the Insurance Industry: A Dynamic Simulation Model and
Experimental Results. Master’ s Thesis, MIT Sloan School of Management.
Oliva, R. 1992. Service Quality Management. Flight Simulator Facilitator’s Training Guide. Center for
Organizational Learning. MIT Sloan School of Management
Parasuraman, A., V. A. Zeithaml and L. L. Berry. 1985. A Conceptual Model of Service Quality and its
Implications for Further Research. Journal of Marketing, Vol. 49 (Fall): 41-50.
Quinn, J. B. and C. E. Gagnon. 1986, Will Services Follow Manufacturing into Decline? Harvard
Business Review, Vol. 64 (6): 95-103.
Schmenner, R, W. 1986. How Can Service Business Survive and Prosper? Sloan Management Review,
Vol. 27 (3): 21-32.
Schneider, B. 1991. Service Quality and Profits: Can you have your cake and eat it, Too? Human
Resource Planning, Vol. 14 (2): 151-157.
Schneider, B., J. J. Parkington and V. M. Buxton. 1980. Employee and Customer Perceptions of Service
in Banks. Administrative Science Quarterly, Vol. 25 (1980): 252-267.
Senge, P.M. 1990a. Catalyzing Systems Thinking within Organizations. In F. Masaryk (Ed.), Advances
in Organizational Development (pp. 197-246). Norwood, NJ: Ablex.
Senge, P. M. 1990b. The Fifth Discipline: The Art and Practice of the Learning Organization . New
York: Doubleday Currency.
Senge, P. M. and C. Lannon. 1990. Managerial Microworlds. Technology Review, Vol. (July 1990):
63-68.
Senge, P. M. and J. D. Sterman. 1992. Systems Thinking and Organizational Learning: Acting Locally
and Thinking Globally in the Organization of the Future. European Journal of Operational
Research, Vol. 59 (1992):.137-150.
Sterman, J. D. 1988. A Skeptic’s Guide to Computer Models. In L. Grant (Ed.), Foresight and National
Decision Lanham, MD: University Press of America.
Tornow. 1991. Service Quality and Organizational Effectiveness. Comments from the Guest Editor.
Human Resource Planning, Vol. 14 (2): 86-88.
SYSTEM DYNAMICS '93 485
