Computer Simulation Modeling Using STELLA
to Enhance Investigative Learning in a Biology Curriculum

Steven K. Rice, Grant E. Brown and R. Paul Willing
Department of Biological Sciences
Union College
Schenectady, NY 12308

HOME	PROJECT SUMMARY	INTEGRATING MODELS AND EXPERIMENTS	RESEARCH AND OUTREACH
WHAT IS STELLA?	EXAMPLES FROM UNION	WHAT WE RECOMMEND	CONTACT US

WHAT IS STELLA?

STELLA is a flexible computer modeling package with an easy, intuitive interface that allows users to construct dynamic models that realistically simulate biological systems (visit the High Performance System website for more information).  Given the combination of ease of use and modeling power, the STELLA system is ideal to interface with student investigative experiences.   In its most basic form, modeling in STELLA proceeds in three steps:   constructing a qualitative model, parameterizing it, and exploring the model's dynamics.  Follow this link to view examples of STELLA models.

STELLA presents four model building blocks that are used in the modeling process: stocks, flows, connectors, and converters. 

Steps to Constructing a STELLA Model:

1. Constructing a Model:   To build a qualitative model, modelers first define stocks.  Stocks represent anything that can accumulate or change in number (populations, biomass, nutrient content, water, etc...) and are related to the biological question of interest (i.e., represent the important dependent variables and other variables that influence them).  In addition to tangible, physical accumulations, stocks can represent degrees of non-physical entities such as knowledge.  Next, users construct links to variables that affect the size of the stocks.  These are usually direct inputs or outputs modeled using flows.  For example in a population, births would represent a flow into the population.  The magnitude of these flows can be adjusted by converters using links or be affected by the size of stocks in a density-dependent manner. During this first modeling step, students are forced to consider what the essential elements of the biological problem are, and how they qualitatively influence one another.  Their resulting models are not unlike concept maps.

2. Parameterizing the Model:   During the second modeling step, students quantify the relationships among elements in their model.  STELLA allows both linear and non-linear relationships to be expressed.  Once again, students need to apply their understanding of the biological problem to assist in this process.

3. Exploring Model Dyanamics:  The last step of the modeling exercise is to explore the model output.  Modelers generate output in tabular and/or graphical form to explore quantitative or qualitative outcomes.   Also, modelers can manipulate parameters easily and perform sensitivity analysis.  

We employ STELLA modeling in combination with investigative exercises and experiments.  The modeling helps students develop hypotheses, explore predictions, summarize experimental results, and extend their results to novel scenarios.  STELLA provides the flexibility to allow students to model a variety of experimental systems and the power to provide for meaningful outcomes that relate to specific biological content.