Engaging Science Students with STELLA

The complex, elegant systems and cycles that make up the human body and the natural world are fascinating. Using lectures, blackboard diagrams, or even animations to explain those systems and cycles can make them incomprehensible and, well, boring. That’s why Jon Darkow uses Stella, Systems Thinking software from isee systems, to teach biology, anatomy and physiology, and environmental science to students at Seneca High School in rural, northwest Ohio.

By the end of the year students had become better thinkers. They could make their own data tables and graphs and choose their own variables

“I started using modeling software in graduate school at Bowling Green University,” says Darkow. “It was an organic learning process – I was self-taught – and I tried to integrate a couple of applications into my high school classes. Through the Creative Learning Exchange, I heard that Stella was easier to use and discovered that the last protocol in a curriculum called Watershed Dynamics used a Stella model to make decisions on water use. I thought, ‘I have to get this software.’”

Darkow wrote a grant and secured a district-wide site license. That was four years ago. Since then he’s developed models that allow students to learn about physiological, biological, ecological and environmental systems through experimentation and discovery.

When tested on cellular respiration, they did better than students who had been taught through animations

One of Darkow’s models simulates the cellular respiration process. “Cellular respiration is very complicated, very dynamic and there are lots of interactions,” says Darkow. “We use the model to answer the question ‘How do you get energy from a glucose molecule to make adenosine triphosphate, ATP?’ The model makes it easier for students to see how the process works over time.”

Students don’t just passively watch the model. They manipulate four sliders to test the influence of oxygen, temperature, pH, and other chemical variables on ATP production at every phase of the cycle. “When students play with the model they see that the cycle works better with a lower pH level, for example,” says Darkow.

There are several indications that Darkow’s cellular respiration model is an effective teaching tool. “At the beginning of the year, I chose the data table they would use for variable analysis,” says Darkow. “I’d prescribe the X-axis and the Y-axis. By the end of the year students had become better thinkers. They could make their own data tables and graphs and choose their own variables. And when tested on cellular respiration, they did better than students who had been taught through animations.”

Inspired by the Ohio Teacher Showcase that highlights the work of teachers engaged in project-based education, Darkow created a template for a natural selection model. “I got the idea from Barry Richmond’s book,” says Darkow. “My students have to adapt the model to their own study of natural selection. One student looked at the evolution of lactose-persistency in herding communities. Another is looking at sickle cell anemia in areas with malaria.”

The model forces them to think about how variables will increase or decrease in certain conditions

Before modeling natural selection, students spend a lot of time choosing the variables they’ll test. “The model forces them to think about how variables will increase or decrease in certain conditions,” says Darkow. “Students have to engage in trial and error. They see a screwy thing happen, develop a hypothesis, test the hypothesis, make revisions, and test again.”

This spring, the top six natural selection models from Darkow’s class were presented to professors from Heidelberg College. “I’ve been really impressed with their work. By researching an HIV-resistant gene in places where there is a high rate of HIV infection, one group even managed to find the intersection of biology, ecology, and social dynamics.”

Sliders and comparative graphs allows students to discover small manipulations that make big changes

Darkow does teach some AP-level classes but he’s quick to point out that he’s not always putting Stella in the hands of academic high-fliers. “I teach regular rural kids,” he says. “I had one student who wasn’t that into class – he doodled a lot,” says Darkow. “But he just took charge with his model and even went around to help other groups. He’s like me, not that interested in details, but loves testing things over and over to see what changed. With Stella, he just blossomed.”

His and his students’ appreciation for variable testing have made Stella’s sliders and graphs two of Darkow’s favorite features. “Sliders and comparative graphs allows students to discover small manipulations that make big changes,” says Darkow. Other features have made it easier for Darkow to model specific systems. “The Random feature is very useful in AP biology classes where students have to understand natural variation. I was having trouble getting genetic drift and natural selection ideas into models and the built-in binomial function helped me figure that out.”

Stella puts computing back in the computer... I was able to teach about population growth by running a million generations during a one-period class

Stella’s ability to handle large amounts of data also adds to its capacity as an educational tool. “Stella puts computing back in the computer,” says Darkow. “Crunching numbers is something computers are really good at. I was able to teach about population growth by running a million generations during a one-period class. I couldn’t do that without Stella and a computer.”

NetSim has made it easy for Darkow’s students to play with models outside of class and for him to share his work with thousands of peers. “My enzyme simulation has been run over 4,000 times,” says Darkow. “I link to it from the College Board AP Biology site. The feedback has been very positive, everything from ‘This is great’ to detailed questions from teachers who want to be sure they’re on the right track.”

As happy as he is to share his models with other teachers, Darkow hopes that more and more teachers start using Stella on their own. “There is a definite learning curve but the more you use Stella, the more confident you become,” says Darkow. “I’m doing more this year than ever before and I’m more comfortable making models that are realistic.”

I just love the creativity Stella allows. As an educational tool, it’s awesome

“The best way to learn about modeling is to start with something simple and then experiment with more and more complexity,” says Darkow. “The more you play, the more interesting and fun it gets. It’s like a game.” And, just like science, Darkow finds that modeling is best learned through trial and error.

He’s experimenting with more complexity himself. “I’m thinking about where I’m headed next with teaching about environmental problems and looking at leverage points,” says Darkow. “I just love the creativity Stella allows. As an educational tool, it’s awesome.”

iThink dynamic modeling improves freight and passenger transportation systems on Latin America’s borders

For developing countries, building international trade capabilities is often the key to economic success. Trade opens new markets and allows small producers to grow. It invites tourism and job opportunities for migrants and immigrants who will contribute their earnings to local economies. Implementing systems that control efficient trade is often the challenge that gets between trade aspirations and actual economic gains.

ICCs aim to improve the operational efficiency of border crossings and ensure the integrity of control processes for both freight and people

“Since 1959, the Transport Division of the Inter-American Development Bank (IADB) has worked to improve the fluidity and control of freight and passenger processes at borders,” says Graciela Sicra, Principal of OMN Systemic Strategy, an isee systems consulting and training partner. “Through several projects, it developed the concept of Integrated Control Centers (ICC). Specializing in freight and passenger systems, each ICC functions as a binational city; within its boundaries, there is free movement of people and goods.”

ICCs aim to improve the operational efficiency of border crossings and ensure the integrity of control processes for both freight and people. To increase efficiency, ICCs have to decrease the unplanned down time that is often the result of private agent staffing practices or unanticipated flows of people and freight. Decreasing downtime depends on a thorough understanding of people and freight transit flows and the ability to continually optimize control processes.

Using the iThink models, analysts interact with complex systems and explore, design, and test management policies and processes

Dynamic Modeling

To get a deeper understanding of the flows moving in and out of their transit systems and the relationship of those flows to staffing levels, ICC project experts turned to iThink models developed by OMN Systemic Strategy. “Using the iThink models, analysts interact with complex systems and explore, design, and test management policies and processes,” says Sicra. Working with ICC analysts, OMN developed iThink management flight simulators for all ICC projects in Latin America at Southern Cone country borders, as well as borders with Peru, Ecuador, Columbia, Panama, and Costa Rica.

Each project uses two models, one for passengers and one for freight, that support both the initial planning and design of each center and its operations after construction. At the planning and design stage, the simulator provides a solid base for sizing parking facilities, passenger and freight lines, and control posts using actual field data and projections that look out 20 years. After the ICC has opened, the simulator provides a consistent, shared method for analysts and decision makers to determine if operational targets will be achieved and to test additional actions that will be required.

By using models, ICC analysts are able to demonstrate how the concept works and its benefits to the countries and cities involved

Each simulator captures the processes observed at existing sites, thus allowing scenario analyses that compress time and space in a framework that emphasizes:

• Transparency: The structure of the model is available and the results are displayed numerically and graphically
• Flexibility: Models support a wide variety of complex scenarios. The data can be exported to an Excel spreadsheet
• Consistency: The behavior of the model is consistent with observed data from existing sites
• Accessibility: The users interact through a user-friendly interface on laptops in real time
• Robustness: The models capture the operational details specific to each user category and vehicle

By capturing key elements of the ICC operation, the simulator allows rapid experimentation with facility-routing proposals and the future operation of the complex. “ICCs impose many changes at border crossings,” says Sicra. “By using models, ICC analysts are able to demonstrate how the concept works and its benefits to the countries and cities involved.”

Simulator User Interface

The custom-designed simulator user interface makes it possible for analysts who have no training or experience in systems dynamics or modeling to interact with passenger and freight control processes. The impact of their decisions are displayed using indicators, tables, and charts. By changing the operational parameters, analysts can verify the validity of their proposals under a variety of scenarios.

Using actual data and projections, it determines the number of entry lines, control booths, and physical review docks needed to process each type of freight transit

For example, analysts optimize the passenger control process by modeling flows of people in each direction through the ICC’s control facilities, looking at both mode of transportation (foot, car, or bus) and classification of person (tourist, immigrant, or citizen of neighboring country).

A setup and analysis screen provides a quick overview of ICC operations and is sensitive to changes in demand (passenger flow over time) scenarios. Analysts can test and determine the processes and facilities (number of booths, scanners, parking places, etc.) needed to accommodate any number of passengers without causing downtime and/or long queues.

A separate passenger interface allows analysts to vary the number of cars and buses moving in each direction, the number of passengers per vehicle, and the types of passengers moving in each direction to understand how queues build and wait times change. They can then test operational and staffing level remedies.

By opening the appropriate number of lines and managing staffing levels, ICCs are avoiding bottlenecks and reducing queues

The simulator also includes the control process for both local and international freight vehicles. Using actual data and projections, it determines the number of entry lines, control booths, and physical review docks needed to process each type of freight transit. Analysts can also experiment to determine how staffing levels influence queue times.

Using controls on the Expedited Dispatch Channels screen, analysts can set the average time needed for each freight process per line. The simulator then displays the status of each process along with the number of vehicles in the queue, the maximum waiting time, and the maximum vehicle load each hour.

Modeling insights and benefits

Using this simulator, ICC analysts have moved beyond reporting on transit processes to experimenting with modifications to optimize each process and overall operations. “Managers are determining resource allocations according to traffic flows expected on each shift over a 24-hour period,” says Sirca. “By opening the appropriate number of lines and managing staffing levels, ICCs are avoiding bottlenecks and reducing queues.”

The model was built with sufficient flexibility and data to support even extreme scenarios for up to 20 years

While the simulator has led to changes in ICC operations, Sirca doesn’t anticipate changes to the model itself. “The model was built with sufficient flexibility and data to support even extreme scenarios for up to 20 years. If and when it’s needed, we’ll be able to change the model to align with any modifications to the ICC concept, its basic processes, or its layout.”

Community Model Building

On Thursday May 21st, we had our first community model building session! For this pioneer event, we invited people to contribute in a discussion and model building session on the drought currently affecting California. We were absolutely overwhelmed with the incredible response we received for this community model building session! Over 160 people from all over the world registered for this event and so many participated in the discussion it was a challenge for us to keep up with their suggestions.

a model that showed the relationships between water supply, agricultural crop demand, crop price, and population

Based on the feedback from the participants, we focused in on water usage and agriculture. In our one-hour discussion with this community, we not only identified key variables that affected the focus area, but also discussed the variable behavior and created an interesting causal loop diagram showing largely balancing feedback loops. While the group discussion was active, co-president Bob Eberlein was actively modeling the comments and suggestions we received from the community, resulting in a model that showed the relationships between water supply, agricultural crop demand, crop price, and population. The recording and results for this community build session will be available for viewing on June 9th, so be on the look out!

We were so pleased with the participation from the community and we look forward to continuing these community model building sessions in the future! If there is a particular complex issue you would like to see modeled as a community please let us know!

Advanced Modeling Series

Version 10.1 comes with exciting new features and yet remains the trusted modeling software that you are used to. Stella Live allows you to see modeling results of stocks and flows directly in the diagram. They are instantly updated as you dial up or down constants in the model. Cycle time, allowing you to time stamp and track material in your model, and Sketchable Graphs, enabling you to sketch expected behavior, have been reintroduced.

Stella Professional is the software that you know and love – with the new professional look and functionality that you have been waiting for!

New Sensitivity analysis is now available for Monte Carlo simulations. Now table settings can vary from table to table.

Stella Professional - The next generation in modeling

Stella Professional has been rebuilt from the ground up and takes your modeling experience to a higher level. Now you can build models even more efficiently and get a fresh professional look. Models can be kept clean and organized by resizing objects, using smooth Bezier connectors, and by allowing for unlimited control over color, font, and title alignment. Tool bars can be detached and docked to wherever you need them to improve your work flow. A new results panel enhances the effectiveness of your model analysis. Equation and unit errors are summarized in a list so that you can solve any problems more quickly. Stella Professional is the software that you know and love – with the new professional look and functionality that you have been waiting for!

Which software will fit your needs best?

Stella Professional makes it even easier to build and learn from your models. Currently, it comes without an interface layer. To present and publish your models, you can easily open any Stella Professional file in STELLA / iThink 10.1.