Skip to main content

Enabling Modeling and Simulation-Based Science in theClassroom:Integrating agent-based models, real world sensing andcollaborative networks


This project is for a Full Research and Development project addressing Challenge Two, "How can all students be assured the opportunity to learn significant STEM content?" The project goal is to leverage the power of three computation-based technologies for STEM education by integrating them into curricular modules for use in middle and secondary schools. The three technologies are agent-based modeling, real world sensing and collaborative classroom networks. While each of these technologies separately has been deployed in classrooms to positive effect, they have not yet been united in curricular modules to reach their maximum combined impact and scale. It is our hypothesis that such modules will provide more powerful and comprehensive learning for students, and that a single platform combining the three technologies can accelerate scaled penetration of important STEM learning experiences into middle and secondary curricula.

One argument for these technologies' place in the classroom derives from their increasing importance in scientific practice, to address complex and interacting systems that can be difficult to capture with traditional methods. Agent-based models enable scientists to capture system behavior by "growing it" (Epstein & Axtell, 1996; Wilensky, 2001), modeling the behavior of system elements with simple rules. Science has also taken advantage of advances in sensor technologies and the ability to capture real time data from many types of sources. Computational interfaces for data capture have enabled scientists to assemble large data sets which can often be displayed and analyzed in real-time or stored for subsequent analysis. Finally, collaborative network technologies have enabled scientists to engage in both real-time and asynchronous collaborations, to merge results from their data sets, to create annotated artifacts, to share and compare alternative models and to form communities around both applied and theoretical problems.

It is our belief that as much as these three technologies have transformed science practice, they have an even greater potential to transform STEM education. One major advantage of agent-based models, for example, is that they are accessible to learners without the need for great mathematical sophistication. Adding in real-world sensed data enhances the models believability to students and also spurs them to move beyond simply consuming existing models, to produce creations of their own as they adjust models to fit the data they gather. This challenges students to develop models and arguments that work in real-world engineering contexts Finally, adding in networks enables students to engage in whole-class "participatory simulations" (Wilensky & Stroup, 2000) wherein students individually control elements of a simulation, experience the simulation together, and discuss interpretations. While many successful implementations of these individual technologies have been recorded, a scarcity of classroom-ready units demonstrating their integration has prevented them from achieving a level of impact on classroom practice commensurate with their collective potential. This project seeks to close that gap.

Our project will proceed in two phases. In the first two years, we will conduct design experiments in middle and high school classrooms in order to iteratively refine our curricular modules and, in parallel, our assessment items for phase two. In years three and four, we will create matched modules to those created in phase one wherein all activities are ABM only. We will then implement both sets of units in matched classrooms and compare the results through analysis of pre- and post-assessments, student artifacts, keystroke logs and videotaped semi-structured interviews with samples of students and teachers.

Intellectual Merit: Agent-based approaches to the study of complex phenomena in themselves have transformative potential for the enterprise of STEM education, restructuring not only how certain concepts are learned, but also when and by whom. Participatory simulations and sensor-engaged activities add value, offering additional access and power by grounding student experiences in the richness of their social and physical surround. Gaining a better understanding of the learning design frameworks that guide the balance and interplay between these approaches will have a far-reaching impact on the creation of new pathways for learning, across STEM disciplines. Refining the technology platform for such activities will open the door to applying these frameworks to a wide range of STEM education contexts.

Broader Impact: The curricular modules developed in the project will be designed as classroom-ready units unifying the best of the three approaches, and they have a clear path to wide dissemination through the NetLogo distribution and Inquire Learning. The growing proliferation of network-ready personal computing devices radically extends classroom access to these ABMs and PSAs, as does the continued spread of TI-Navigator. With the GoGo board, very low cost provides broad access to simulation-based learning grounded in data collected by students themselves. Such approaches have been shown to especially serve underrepresented groups. The project promises to impact graduate training in the three PI institutions and to strengthen existing ties among them.