System Dynamics (to be abbreviated SD): System dynamics is a mathematical modelbuilding methodology for studying complex system behavior. By analyzing feedback loop interaction one can gain a deeper understanding of the causes of systemic behavior, see how delays can create instability in a system, identify leverage points, and test policies that might mitigate undesirable behaviors or unintended consequences.
Modeling has been used in science extensively for years. System Dynamics modeling provides a view of the structure of the system that is more revealing than many other types of modeling. Because the software is visual, because full words or phrases can be used to identify the individual icons that represent a component in the model structure, because dependencies of one part upon another can be explicitly displayed, much more information is provided to the learner. The building of models is an ACTIVE process for the students. They must understand why each component is necessary for the system to operate, how the components are connected, and the role each component has in controlling the behavior of the system. They construct the small models and/or enhance/modify midsized models from a smaller core model.
Within the different science areas students study many systems that are governed by feedback mechanisms. SD modeling has as its fundamental approach the study of how feedback affects the behavior of complex systems. It is a valuable ‘new’ tool to enhance learning in the science classroom.
There are numerous small System Dynamics models appropriate for use in high school science. Some SD science models are embedded within the lessons found in the curriculum resources and links of this website. 
CROSSCUTTING CONCEPTS
(Modelbuilding is a constructing endeavor. Students gather data, formulate an hypothesis, build a model to demonstrate that they understand the underlying structure at work in the system, and test their hypothesis on the model, using the data to help guide the testing process.)
Patterns
 Different patterns may be observed at each of the scales at which a system is studied and can provide evidence for causality in explanations of phenomena.
The SD modeling software allows both simulation time and range to be altered to reveal broad or detailed patterns in output. Causality is core to the design of the structure of an SD model.
 Patterns of performance of designed systems can be analyzed and interpreted to reengineer and improve the system.
Analyzing patterns of system performance is central to SD modeling, both to help students understand the core concepts being studied and to try to improve system behavior, if desired.
 Mathematical representations are needed to identify some patterns.
The SD diagram represents the mathematical structure of the problem and produces the patterns observed in the graphs. These structures are analyzed to explain the behavior patterns.
 Empirical evidence is needed to identify patterns.
Using data capture as a precursor to model development is advised, when applicable, so students have a sense of the behavior pattern the model should exhibit.
Cause & Effect
 Empirical evidence is required to differentiate between cause and correlation and make claims about specific causes and effects.
SD diagrams are drawn to exhibit cause and effect relationships. Empirical evidence can be used to help design the SD diagram to determine if the diagram behavior is designed correctly – i.e. that the cause and effect structure is captured by the model
 Cause and effect relationships can be suggested and predicted for complex natural and human designed systems by examining what is known about smaller scale mechanisms within the system.
Central to SD modeling. Students study smaller generic structures that are used as part of larger models. They recognize those structures and their typical behavior. Explanation of the feedback connections between internal variables and internal generic structures is part of the analysis that is expected in SD modeling.
 Systems can be designed to cause a desired effect.
Central to SD modeling.
 Changes in systems may have various causes that may not have equal effects.
Identifying the strong and weak components in the causal stock/flow diagram is part of the SD analysis. Transfer of feedback loop dominance is also studied.
Scale, proportion, and quantity
 Some systems can only be studied indirectly as they are too small, too large, too fast, or too slow to observe directly.
An important reason to create any type of model (SD is no exception)
 Patterns observable at one scale may not be observable or exist at other scales.
This can have more than one interpretation for SD. There is a high level Module causal design that can give high level segment interaction and feedback analysis, and each module can contain its own detailed stock/flow diagram which can show the behavior of that particular module, isolated from the rest. Also, graphical scales can be adapted to show general trend or more detailed model behavior.
 Using the concept of orders of magnitude allows one to understand how a model at one scale relates to a model at another scale.
In SD the structure is the core concept driving the behavior. The application of that structure to, say a cellular level, or an ecosystem level, would be essentially the same if the same dynamic is at play.
 Algebraic thinking is used to examine scientific data and predict the effect of a change in one variable on another (e.g., linear growth vs. exponential growth).
A core part of SD modeling. The study of rates of change and how those rates affect the behavior of the system is central to SD analysis.


CROSSCUTTING CONCEPTS
Systems and Systems Models
 Systems can be designed to do specific tasks.
Applies directly to SD modeling.
 When investigating or describing a system, the boundaries and initial conditions of the system need to be defined and their inputs and outputs analyzed and described using models.
Applies directly to SD modeling. Students are asked to determine the boundary of the model by deciding what variables to include. The student designs the model diagram, defines the initial conditions, simulates the model, and analyzes the behavior – reconciling the structure with the output of the model.
 Models (e.g., physical, mathematical, computer models) can be used to simulate systems and interactions — including energy, matter, and information flows — within and between systems at different scales.
Applies directly to SD modeling. SD modeling is about trying to capture the core structures that are important to reproduce the typical behavior of a system under study. The problem can include multiple systems, which can be built in a modular fashion, and connected, to display the larger (macro) system behavior.
 Models can be used to predict the behavior of a system, but these predictions have limited precision and reliability due to the assumptions and approximations inherent in models.
Applies directly to SD modeling. Students are expected to state their assumptions as they design their models, and recall those assumptions when they explain the behavior of the model they have designed.
Structure and Function
 Investigating or designing new systems or structures requires a detailed examination of the properties of different materials, the structures of different components, and connections of components to reveal its function and/or solve a problem.
Structure determines behavior is a core concept in SD modeling.
 The functions and properties of natural and designed objects and systems can be inferred from their overall structure, the way their components are shaped and used, and the molecular substructures of its various materials.
Structure determines behavior is a core concept in SD modeling. The model can be designed to show the relationship between structure and function at various appropriate scales.
Stability and Change
 Much of science deals with constructing explanations of how things change and how they remain stable.
A prime use of SD modeling is to aid in communicating how and why systems change over time.
 Change and rates of change can be quantified and modeled over very short or very long periods of time. Some system changes are irreversible.
Core to SD modeling. Every model includes a rate of change and an accumulation. The model can be simulated for thousands of time units or fractions of time units.
 Feedback (negative or positive) can stabilize or destabilize a system.
Core to SD modeling. Feedback analysis is the cornerstone of SD analysis.
 Systems can be designed for greater or lesser stability.
Part of SD modeling and analysis.
