A Methodological Framework for Integrating
Systems Thinking and System Dynamics
Robert Y. Cavana Kambiz E. Maani
Victoria University of Wellington The University of Auckland
PO Box 600, Wellington, New Zealand Private Bag, Auckland, New Zealand
Tel: +64 4 463 5137; Fax: +64 4 463 5253 Tel: +64 9 373 7599 Ext 8813
Email: bob.cavana@vuw.ac.nz Email: k.maani@ auckland.ac.nz
Abstract
This paper discusses a methodological framework for integrating the fields of systems
thinking and system dynamics. A discussion of the four levels of thinking and their
implications for organisations and management is followed by the core of the paper, which
presents a methodology for systems thinking and modelling. The methodology includes five
major phases: problem structuring; causal loop modelling; dynamic modelling; scenario
planning and modelling; and implementation and organisational learning. These phases
follow a rigorous systematic process, each involving a number of steps. Some general and
specific applications of systems thinking are then discussed, and hard and soft modelling
approaches are compared and contrasted. Finally, some cases are provided to illustrate
applications of the systems thinking and modelling methodology.
Introduction
Daily, we are exposed to information from a multitude of sources: the news media,
newspapers, radio, TV, and the Internet. Generally this kind of information reports events
what happened, where, when, how, who was involved, etc. This is a snapshot view of the
world because this level of information is very shallow; the reports only touch the surface of
what actually happened. For example, the stock market information that is reported daily
gives a snapshot of the day’s activities. It tells us whether stocks, on average, went up or
down (often the index goes both up and down within one day) and by how much. We also get
information on the volume of shares traded, the dollar value of stocks traded (capital
tumover) and much more. All of this information is at event level.
Sometimes there is commentary about a news item or an issue, and this allows one to look
back and examine the trends and patterns of events and data. This provides a richer picture of
reality and gives more insight into the ‘story’. In the stock market example, this means
looking at the trends over past months or years, observing the fluctuations and trying to
explain what caused ‘pulses’ in the system - for example, news of a merger, a quarterly
economic report or a political scandal.
However, it is rare to see a study of how such trends and patterns relate to and affect one
another. This represents a much deeper level of thinking that can show how the interplay of
different factors brings about the outcomes that we observe. In the stock market example, this
would mean trying to relate a host of factors that systemically cause the fluctuating patterns.
These factors could be economic, social, political or structural. The critical thing at this level
of thinking is to understand how these factors interact.
Furthermore, there is yet another, much deeper level of thinking that hardly ever comes to the
surface. This represents the ‘mental models’ of individuals and organisations that influence
why things should/do or should not/do not work. Mental models are based on the beliefs,
values, and assumptions that we (privately) hold, and underlie our reasons for doing things
the way we do them. Harvard educationalist Chris Argyris (1990, pp.25-27) calls these the
“undiscussables’.
The four levels of thinking described above are depicted in Figure 1. This figure uses the
analogy of an iceberg, where the event level of thinking is represented by the tip and yet most
of us are satisfied with this level. However, this level is clearly completely inadequate for
understanding the dynamics underlying change and complexity in the world around us. This
paper outlines a systems thinking and modelling (ST&M) methodology, which is fully
described in a recent book by Maani and Cavana (2000) on ‘Systems thinking and modelling:
understanding change and complexity’.
Figure 1 Four levels of thinking
og
“
| HAVE A PROBLEM USING HIERARCHICAL |
MODELS TO UNDERSTAND SYSTEMIC THINKING.
Source: Cartoon provided by Jock Macneish in Maani and Cavana (2000, p13).
Systems methodology
Systems methodology or the systems approach refers to a set of conceptual and analytical
methods used for systems thinking and modelling. The general methodological approach
towards systems thinking and modelling used in this paper is based on the system dynamics
method. The field of system dynamics was developed by Jay Forrester (1961) and others at
the Massachusetts Institute of Technology in the late 1950s, based on developments
following World War II in:
¢ the theory of information feedback systems;
¢ the understanding of decision-making processes;
* the use of mathematical models to simulate complex systems; and
* the development of high-speed electronic digital computers as a means of simulating
mathematical models.
Many other people have contributed to the development of systems thinking and system
dynamics including Coyle (1977, 1996), Randers (1980), Richardson and Pugh (1981),
Roberts et al. (1983), Senge (1990), Wolstenholme (1990), Richardson (1991), Mohapatra et
al. (1994), Morecroft and Sterman (1994), Vennix (1996), Richmond and Petersen (1997),
Sterman (2000), and many others! However, several authors have provided definitions of the
system dynamics methodology, but we consider the one recently provided by Eric
Wolstenholme (1997) as most appropriate. Wolstenholme’s description of the scope of
system dynamics is set out below.
What: A rigorous way to help thinking, visualising, sharing, and communication of
the future evolution of complex organisations and issues over time;
Why: — for the purpose of solving problems and creating more robust designs, which
minimise the likelihood of unpleasant surprises and unintended consequences;
How: _ by creating operational maps and simulation models which externalise mental
models and capture the interrelationships of physical and behavioural
processes, organisational boundaries, policies, information feedback and time
delays; and by using these architectures to test the holistic outcomes of
alternative plans and ideas;
Within: a framework which respects and fosters the needs and values of awareness,
openness, responsibility and equality of individuals and teams.
(Wolstenholme, 1997)
The development of a systems thinking and modelling (Maani & Cavana, 2000) intervention
involves five major phases:
problem structuring;
causal loop modelling;
dynamic modelling;
scenario planning and modelling;
implementation and organisational learning (learning lab).
OPRWONE
These phases follow a process, each involving a number of steps, as outlined in Table 1.
However, it must be emphasised that a ST&M intervention does not require all phases to be
undertaken, nor does each phase require all the steps listed in Table 1. Rather, these phases
and steps are presented as guidelines, and which phases and steps are included in a particular
ST&M intervention depends on the issues or problems that have generated the systems
enquiry and the degree of effort that the organisation is prepared to commit to the
intervention.
Table 2.1: Systems Thinking & Modelling Process
Phases Steps
1, Problem 1. Identify problems or issues of concern to management
structuring 2. Collect preliminary information & data
2. Causal Loop
modelling
Identify main variables
Prepare behaviour over time graphs (reference mode)
Develop causal loop diagrams (influence diagrams)
Analyse loop behaviour over time
Identify system archetypes
Identify key leverage points
Develop intervention strategies
Noo Pwhr
3. Dynamic
modelling
Develop a systems map or rich picture
Define variable types and construct stock-flow diagrams
Collect detailed information and data
Develop a simulation model
Simulate steady-state / stability conditions
Reproduce reference mode behaviour (base case)
Validate the model
Perform sensitivity analysis
Design & analyse policies
0. Develop & test strategies
BSON OO Pwryr
4, Scenario planning
and modelling
Plan general scope of scenarios
Identify key drivers of change & keynote uncertainties
Construct forced & learning scenarios
Simulate scenarios with the model
Evaluate the robustness of the policies and strategies
oP wh
5. Implementation 1. Prepare a report and presentation to management
and organisational | 2. Communicate results and insights of proposed
learning intervention to stakeholders
3. Develop a microworld and learning lab based on the
simulation model
4. Use learning lab to examine mental models and facilitate
learning in the organisation
Source: Maani and Cavana, 2000, Table 2.1, p16.
Phases of the methodology
Figure 2 shows the progression of the phases. As mentioned earlier, although these phases
can be used individually, their cumulative use adds more value and power to the
investigation. These phases are described below.
Figure 2 Phases of the systems thinking and modelling methodology
Scenario planning and
modelling
Problem Causal loop >
structuring modelling Implementation and
organisational learning
Problem structuring
In this phase, the situation or issue at hand is defined and the scope and boundaries of the
study are identified. This is the common first step in most problem-solving approaches. The
problem structuring phase consists of the following steps:
Dynamic
modelling
(1) Identification of the problem area or policy issues of concern to management. This
step requires that we clearly establish the objectives, taking into account multiple
stakeholders and perspectives.
(2) Collection of preliminary information and data including media reports, historical
and statistical records, policy documents, previous studies, and stakeholder
interviews.
Causal loop modelling
During this phase, conceptual models of the problem, known as causal loop diagrams (CLDs)
will be created. This is a major component and the most commonly used part of the systems
thinking approach. The following steps are used in causal loop modelling:
(1) Identify main (key) variables.
2) Draw behaviour over time charts (or reference modes) for the main variables.
(3) Develop causal loop diagrams (influence diagrams) to illustrate the relationships
among the variables.
Discuss behaviour over time of the dynamics implied by the causal loop diagrams.
Identify system archetypes that would describe high-level causal patterns.
Identify key leverage points.
Develop intervention strategies.
Dynamic modelling
This phase follows the causal loop modelling phase. Although it is possible to go into this
phase directly after problem structuring, performing the causal loop modelling phase first will
enhance the conceptual rigour and learning power of the systems approach. The completeness
and wider insights of systems thinking is generally absent from other simulation modelling
approaches, where causal loop modelling does not play a part. The following steps are
generally followed in the dynamic modelling phase:
(1)
(2)
(6)
(7)
(8)
(9)
(10)
Develop a high-level map or systems diagram showing the main sectors of a
potential simulation model, or a ‘rich picture’ of the main variables and issues
involved in the system of interest.
Define variable types (e.g. stocks, flows, converters, etc.) and construct stock flow
diagrams for different sectors of the model.
Collect detailed, relevant data including media reports, historical and statistical
records, policy documents, previous studies, and stakeholder interviews.
Construct a computer simulation model based on the causal loop diagrams or stock-
flow diagrams. Identify the initial values for the stocks (levels), parameter values for
the relationships, and the structural relationships between the variables using
constants, graphical relationships and mathematical functions where appropriate.
This stage involves using specialised computer packages like STELLA, ithink,
POWERSIM, DY NAMO, DYSMAP, COSMIC or VENSIM.
Simulate the model over time. Select the initial value for the beginning of the
simulation run, specify the unit of time for the simulation (e.g. hour, day, week,
month, year, etc.). Select the simulation interval (DT) (e.g. 0.25, 0.5, 1.0) and the
time horizon for the simulation run (ie. the length of the simulation). Simulate
model stability by generating steady state conditions.
Produce graphical and tabular output for the base case of the model. This can be
produced using any of the computer packages mentioned above. Compare model
behaviour with historical trends or hypothesised reference modes (behaviour over
time charts).
Verify model equations, parameters and boundaries, and validate the model’s
behaviour over time. Carefully inspect the graphical and tabular output generated by
the model.
Perform sensitivity tests to gauge the sensitivity of model parameters and initial
values. Identify areas of greatest improvement (key leverage points) in the system.
Design and test policies with the model, to address the issues of concern to
management and to look for system improvement.
Develop and test strategies (ie. combinations of functional policies, for example
operations, marketing, finance, human resources, etc.).
Scenario planning and modelling
In this phase, various policies and strategies are postulated and tested. Here ‘policy’ refers to
changes to a single internal variable such as hiring, quality, or price. Strategy is the
combination of a set of polices and as such deals with internal or controllable changes. When
these strategies are tested under varying external conditions, this is referred to as scenario
modelling:
(1)
(2)
(3)
Develop general scope, time frame and boundaries of external environment for
scenarios. Prepare stories of possible futures or theme scenarios.
Identify key drivers of change, uncertainties and factors that could have a significant
impact on the decisions, policies and strategies being evaluated. Determine ranges
for external parameters and graphs.
Construct forced scenarios by placing all the positive outcomes in an optimistic
scenario and all the negative scenarios in a pessimistic scenario. Check the forced
scenarios for internal consistency. Modify these scenarios as learning scenarios
(based on Schoemaker, 1995).
(4) Simulate the scenarios (either the individual scenarios varying the key uncertainties
or the learning scenarios) with the model. Redesign scenarios if necessary.
(5) Evaluate the performance of the policies and strategies with the model for each
scenario. Assess the performance against a range of relevant performance measures
for overall robustness. Select the policies or strategies that meet management's
objectives for the investigation.
Implementation and organisational learning
One of the most beneficial and enduring outcomes of systems thinking and modelling is
organisational and team learning. Once simulation models have been developed, they can be
enhanced by extending them into a microworld. Microworlds (also known as management
flight simulators) provide an interactive and user-friendly interface for managers to
experiment with the model. The learning laboratory uses microworlds in a structured process,
akin to a scientific environment, to test hypotheses and mental models designed to create
individual and group learning. The following steps summarise this phase:
(1) Prepare a report and presentation to the management team and other stakeholders.
This should document the background and development of the systems thinking
project, the challenges faced and lessons learned.
(2) Communicate results and insights of the study and the reasons for the proposed
intervention to all stakeholders.
(3) Develop a microworld and design a learning lab for the simulation model. This
involves adding necessary features (ie. from computer software) to convert the
simulation model into an interactive and user-friendly microworld. Then design a
learning lab process for the microworld.
(4) Use the learning lab process to diffuse and facilitate learning in the organisation.
Systems thinking and modelling applications
The systems thinking and modelling methodology outlined above has a wide range of general
and specific applications. The general applications are:
design of new systems;
Te-engineering or improvement of existing systems;
prediction of behaviour of complex systems under varying conditions;
understanding the interaction of component sub-systems;
strategy development and testing;
scenario modelling and testing;
group and organisational learning.
The specific applications of systems thinking cover both strategic and functional aspects of
business and organisations. Some of these are outlined below.
Strategy and policy
Systems thinking is widely used for strategy formulation and testing. This occurs at the level
of government and industry (e.g. health care, communication, regulation, etc.) as well as at
the organisational level (e.g. marketing, production, human resources, finance and their
interfaces). Systems thinking highlights the following areas of strategy, which are often
ignored or missed by other methodologies:
* internal contradictions in a strategy;
¢ hidden strategic opportunities;
* untapped strategic leverages.
Operations and design
Systems thinking also has widespread applications in operations and design. Traditionally,
manufacturing systems have been a prominent area of application. Service industries such as
health care, communications and logistics are the upcoming areas that readily lend
themselves to the application of systems thinking and modelling. Some of the specific
applications are:
* new product and service development;
* supply-chain management;
* enterprise resource planning (ERP);
* network design and management.
Functional modelling
In addition to the areas mentioned above, the systems thinking and modelling methodology
can be used to model functional areas such as finance, marketing, information technology and
human resource management. In Maani and Cavana (2000) we discuss these applications and
illustrate how to integrate them using systems models.
Hard and soft modelling
As the terms model and modelling are frequently used, it is important to clarify their meaning
in this context. A model is defined as being a representation of the real world. Models can
take on different forms, physical, analog, digital (computer), mathematical, and so on. This
sense of the word model is the more traditional one and is sometimes referred to as
quantitative or ‘hard’. More recently, the concept of soft modelling has been developed by
Checkland (1981) and others. Soft modelling refers to conceptual and contextual approaches
that tend to be more realistic, pluralistic and holistic than ‘hard’ models. Hard and soft
models are sometimes referred to as ‘quantitative’ and ‘qualitative’, respectively. The
differences between the hard and soft approaches are summarised in Table 2.
Table 2 Hard versus soft approaches
Hard approaches Soft approaches
Model definition A representation of the real A way of generating debate
world and insight about the real
world
Problem definition Clear and single dimensional Ambiguous and multi-
(single objective) dimensional (multiple
objectives)
People and organisation | Not taken into account Are integral parts of the
model
Data Quantitative Qualitative
Goal Solution and optimisation Insight and learning
Outcome Product or recommendation Progress through group
learning
Source: Maani & Cavana (2000, p21) adapted from Pidd (1996, p121)
The systems thinking and modelling methodology presented in this paper covers both hard
and soft approaches, because we regard these approaches as complementary and mutually
reinforcing. Systems thinking tends to fall in the category of soft approaches, while dynamic
modelling gravitates toward the category of hard modelling.
Systems thinking and modelling cases
A number of systems thinking and modelling (ST&M) cases are fully discussed in Maani and
Cavana (2000). However, only a few of them will be briefly outlined here to illustrate the
various ways the ST&M methodology can be utilised, depending on the specific systems
intervention. Table 3 summarises the ST&M steps used in each case. A tick indicates that
the step was used. However, it should be re-emphasised that although the steps appear in a
linear fashion, in some cases an earlier step in the Table is carried out at a later stage in the
investigation. For example in the Beer case (case 4), a causal loop diagram was constructed
after a simulation model had been developed to help explain the dynamic behaviour that the
model generated.
However, in most cases the steps are followed in a ‘sequential’ fashion. Interventions that
required a ‘soft’ systems approach (ie cases 1, 2 & 3) tended to draw upon phases 1 and 2 of
the ST&M process. However, in cases that developed into ‘hard’ modelling projects tended
to utilise the later stages of the ST&M process. In addition, the telecommunications business
unit case (Case 5) utlised all phases of the process.
Table 3: Applications of the Systems Thinking & Modelling Process
Case | Case | Case | Case | Case
Phases & Steps 1 2 3 4 5
1. Problem structuring
1. Identify problems or issues of concern to management
2. Collect preliminary information & data
as
as
Sy
<
as
2. Causal Loop modelling
Identify main variables
Prepare behaviour over time graphs (reference mode)
Develop causal loop diagrams (influence diagrams)
Analyse loop behaviour over time
Identify system archetypes
Identify key leverage points
Develop intervention strategies
SES SR whip
es Sk SSS
2.35. 2S
assis
3. Dynamic modelling
Develop a systems map or rich picture
Define variable types and construct stock-flow diagrams
Collect detailed information and data
Develop a simulation model
Simulate steady-state / stability conditions
Reproduce reference mode behaviour (base case)
Validate the model
Perform sensitivity analysis
. Design & analyse policies
0. Develop & test strategies
BOON oURWNS
<<essdsdc<
<<<<<4446<5
4. Scenario planning and modelling
Plan general scope of scenarios
Identify key drivers of change & keynote uncertainties
Construct forced & learning scenarios
Simulate scenarios with the model
Evaluate the robustness of the policies and strategies
Sh oO i
<a<ss8
5. Implementation and organisational learning
1. Prepare a report and presentation to management Vv Vv Vv
2. Communicate results and insights of proposed Vv Vv
intervention to stakeholders
3. Develop a microworld and learning lab based on the Vv
simulation model
4. Use learning lab to examine mental models and facilitate
learning in the organisation
a
Case 1: Public health reform - the case of New Zealand
In July 1993, New Zealand’s health system fundamentally changed, splitting the health care
provider from the purchaser, whereas previously Area Health Boards had assumed both these
roles. The providers, Crown Health Enterprises, were in future to compete amongst
themselves for funding from the purchasers, Regional Health Authorities. Private sector
managers were brought in to run the hospitals in a business-like manner, increasing
efficiency, and even returning a profit to the Government. The objectives of the reforms
(Upton, 1991) were:
* to reduce hospital waiting times;
* to improve access for all New Zealanders to an effective, fair and affordable health care
system;
* to emphasise health promotion and illness prevention.
However, in reality and despite increased government spending on surgery, waiting list
numbers have soared. This case study used a systems thinking approach to investigate the
effects of the reform, and to determine whether the new system was consistent with
government's stated objectives. The study revealed a number of inconsistencies and gaps in
current policies and proposed intervention strategies for reversing the adverse trends.
Case 2: Lowering the legal drinking age!
Alcohol is a key feature of New Zealanders’ social and sporting activities. Eighty-nine per
cent of men and 85% of women consume alcohol regularly. When this case study was
undertaken, the legislation stated that ‘... to legally consume alcohol in New Zealand an
individual must be 20 years or older. An individual can also consume alcohol if between the
ages of 18-20 and consuming food or accompanied by a parent/guardian.’
Various members of the Parliament were pushing for changes to these laws. A number of
changes have been proposed, including Sunday liquor sales and permitting supermarkets to
sell liquor. One proposed change that caused considerable debate within the government and
in public circles, was the idea of lowering the legal drinking age from 20 to 18 years.
The use of systems thinking methods provided a much clearer understanding of the complex
issue of lowering the legal drinking age and its effects, implications, and unintended
consequences. The study concluded that the lowering of legal drinking age would push the
actual drinking threshold downward and would, in effect, create a new ‘under-age’ drinking
group. Only a few months after the introduction of the new policy, this 'unseen' consequence
is already occurring which has caused surprise and disappointment by the policy makers and
the public alike.
Case 3: Drivers of Quality in Health Services
This case describes how an exploratory project at the New Zealand Ministry of Health, using
a qualitative system dynamics approach to identify the factors that interact to drive quality in
the health and disability sector, revealed sound evidence for the much cited different
worldviews (mental models) of medical/health clinicians and policy managers. (Cavana et al.,
1999)
Case 4: Mainland Beer Distribution Model
This case illustrates model conceptualisation, model construction, and an introduction to
policy analysis. It is based on the famous beer game first developed in the 1960s at the
Massachusetts Institute of Technology's Sloan School of Management (see Senge, 1990).
The version of the beer distribution system presented here is derived from Clark (1988).
Details of this case are: The distribution system for Mainland beer consists of a retail store
and brewery; each managed independently under a centralised inventory policy. The case
uses systems thinking and modelling to assist with the design of inventory control policies.
Case 5: Strategy Development for a Telecommunications Business Unit
This case demonstrates all the phases and nearly all the steps in the systems thinking and
modelling process. It is based on a consultancy project for a business unit in the
telecommunications industry in New Zealand (Cavana and Hughes, 1995). However, the
issues, data and names have been changed to preserve client confidentiality. The major
issues dealt with in this case are how to design policies and strategies to help managers’
turnaround a business unit that is experiencing a declining market share and eroding
profitability.
Conclusions
This paper has presented a methodological framework for understanding change and
complexity using a systems thinking and modelling approach. The phases and steps of this
methodology have been outlined, and we have indicated how we have used aspects of this
methodology in specific systems interventions (cases).
The methodology outlined combines a range of soft and hard modelling/systems approaches,
which can be used in various combinations depending on the specific systems intervention
being considered. We expect that this methodology will be further developed and tested with
other applications over time. The full treatment of the methodology and the five cases
outlined in this paper can be found in authors' new book Systems Thinking and Modelling -
Understanding Change and Complexity (Maani & Cavana 2000).
References
Argyris, C. (1990). Overcoming Organizational Defenses - Facilitating Organizational
Learning, Allyn and Bacon, Boston.
Cavana, R.Y, Davies, P.K., Robson, R.M. & Wilson, KJ., (1999) Different world
perspectives of clinicians and policy managers highlighted through a qualitative system
dynamics analysis of quality in health services, System Dynamics Review, 15(3), 331-
340.
Cavana, R.Y. & Hughes, R.D.. (1995). Strategic Modelling for Competitive Advantage.
Proceedings of the 1995 International System Dynamics Conference. Gakushuin
University, Tokyo, Japan. July 30 - Aug 4, 1995 (Volume II, pp. 408-417).
Checkland, P.B. (1981). Systems Thinking, Systems Practice. John Wiley & Sons, Chichester.
Clark, R. (1988). System Dynamics and Modelling, Operations Research Society of America,
Arlington.
Coyle, R.G. (1977). Management System Dynamics, John Wiley & Sons, Chichester.
Coyle, R.G. (1996). System Dynamics Modelling: A Practical Approach, Chapman & Hall,
London.
Forrester, J.W. (1961). Industrial Dynamics, MIT Press, Cambridge, Mass.
Maani, K.E and Cavana, R.Y. (2000). Systems Thinking and Modelling: Understanding
Change and Complexity, Prentice Hall, Auckland.
Mohapatra, P.K.J., Mandal, P. and Bora, M.C. (1994). Introduction to System Dynamics
Modelling, Universities Press (India) Ltd, Hyderabad 500 029.
Morecroft, J.D.W and Sterman, J.D. ed (1994). Modelling for Learning Organizations,
Productivity Press, Oregon.
Pidd, M. (1996) Tools for Thinking, Modelling in Management Science. John Wiley & Sons,
Chichester.
Powersim (1994). User’s Guide and Reference. ModellData AS, Norway.
Randers, J. ed. (1980). Elements of the System Dynamics Method. The MIT Press,
Cambridge, MA.
Richardson, G.P. (1991). Feedback Thought in Social Science and Systems Theory,
University of Pennsylvania Press, Philadelphia.
Richardson, G.P. and Pugh III, A.L. (1981). Introduction to System Dynamics Modelling with
DYNAMO, Productivity Press, Cambridge, Ma.
Richmond, B. and Petersen, S. (1997). An Introduction to Systems Thinking, High
Performance Systems, Hanover.
Roberts, N., Andersen, D.F., Deal, R.M., Grant, M.S. and Schaffer, W.A. (1983).
Introduction to Computer Simulation: a System Dynamics Modelling Approach.
Addison-Wesley Publishing Co., Reading, Mass.
Schoemaker, P.J.H. (1995) Scenario planning: a tool for strategic thinking. Sloan
Management Review, Winter.
Senge, P. (1990). The Fifth Discipline: The Art & Practice of the Learning Organization,
Doubleday Currency, New Y ork.
Sterman, J.D. (2000). Business Dynamics: Systems Thinking and Modelling for a Complex
World. Irwin McGraw-Hill, Boston.
Upton, S. (1991). Your Health and the Public Health. Ministry of Health, Wellington.
Vennix, J.A.M. (1996). Group Model Building: Facilitating Team Learning Using System
Dynamics, John Wiley & Sons, Chichester.
Wolstenholme, E. (1997, 24 October) System dynamics in the elevator (SD1163), e-mail
communication, system-dynamics@ world.std.com.
Wolstenholme, E.F. (1990). System Enquiry: A System Dynamics Approach, John Wiley &
Sons, Chichester.
1 This case was based on a project prepared by David Todd for the graduate course in
systems modelling at the University of Auckland
