This proposal explores the potential of "Agent-Based Models" to assist learners to acquire environmental science concepts targeted in forthcoming Advanced Placement test standards. It will also help learners acquire a better understanding of complex systems and to shift their attitudes towards the use of scientific models. The investigators frame the research in a simulated scenario where "green" infrastructure is integrated into urban environments, and they propose how to use a new user interface strategy ("Paper-to-Parameters") that promises unique approaches for understanding the spatial and scalar relationships between simulation elements. The project will develop an assessment tool to obtain a picture of prior understandings and attitudes held by learners in different populations (high school, undergraduate, and graduate students and experts); it will conduct an exploratory trial of the Agent-Based learning intervention to investigate the impacts on cognition and attitudes of undergraduate students; and will investigate how selected user interface features facilitate specific spatial and scalar understandings.
The assessment will allow the investigators to describe understanding and attitudes across populations with differing levels of expertise and will provide a baseline for measuring the real impact of the intervention and informing the design of future interventions. The exploratory lesson and targeted experiments will explore the connection between specific features of the computer-based tool and changes in learner understanding of selected AP Environmental Science and complex system concepts, and in learner attitudes towards models.
The main purpose of this research was to explore how complex systems reasoning could be introduced to environmental science learners in a low-cost fashion, and how such learning could be measured. Emerging educational standards emphasize the importance of complex systems thinking across all the sciences, and particularly in environmental science. Environmental science offers unique challenges because much of the phenomena of interest involve understanding how the spatial locations of human and natural resources affect one another and can result in emergent outcomes. The challenge being faced in the field (e.g., in stakeholder planning meetings) is that far too few people are prepared by their science education to engage in the kinds of reasoning GI integration requires: success is not just a matter of selecting the right kind or right amount of GI, it is highly dependent on where GI elements are placed, relative to the human infrastructure, natural features, and other GI elements. There are precious few opportunities for learners to engage in a problem space with spatial challenges, emergent outcomes, and no single "correct" answer. This project thus set out to: 1) Design and build a low-cost tangible interface to a simulation (using paper maps and paper or balsa wood "game pieces", read by webcams as input to the simulation) to expand the number of learners who could participate when only a single computer was available. 2) Study the utility of using a low-cost tangible interface. Specifically: a) Would learners be able to attain the same performance as they would with an equivalent desktop interface? b) Would learners accrue any additional collaboration benefits from working in groups with a tangible interface? (Such interfaces are theorized to better support collaborative tasks) 3) Review the applicability of existing approaches for assessing the learning critical to comprehending a spatial complex systems simulation. For the problem space we were working with, it is essential for learners to develop a sensitivity to spatial patterns (specifically, univariate and bivariate spatial patterns) and to develop a comprehension of core complex systems concepts, like the notion of emergent dynamic, emergent outcomes. 4) Develop new approaches for detecting learner growth in spatial reasoning and complex systems understanding (this goal was added partway through the project, as we began to realize that existing assessment approaches, per major goal 3, were not adequate for this problem space). We were able to: 1) Construct a low-cost simulation interface and associated complex systems simulation. The complex systems simulation was developed with NetLogo, an existing platform for building complex system simulations, designed for use by children but often used by urban planners (ccl.northwestern.edu/netlogo). The tangible interface’s current design makes use of paper maps that represent different neighborhood regions, wooden tokens that represent different green infrastructure elements (swales, rain barrels, semipermeable pavement, and green roofs). Users place these tiles on the paper map, and either a webcam connected to a laptop or an iPad’s camera is used to detect the placement of the tokens on the map, and relay those locations to the simulation to run. The portability of this system allows it to be used in learning scenarios with limited technology: single computer classrooms (the most common type found in Chicago public schools) or neighborhood community centers (which may have no computing technology). 2) Study the effect of the tangible simulation interface on collaboration and learning. We found that when compared to desktop versions of the same simulation, dyads who did not know each other strongly preferred the tangible interface, as compared to dyads who rated each other as friends, suggesting that the partner-monitoring affordances of tangible interfaces may be most useful when partners do not yet have an established working relationship. We also found that users generally preferred the tangible user interface over a desktop interface in terms of its enjoyability, usability, support for exploration, and support of cooperation 3 & 4) We reviewed existing approaches to assessing spatial reasoning and complex system understanding, and developed new approaches to address aspects of the problem space that existing assessments did not measure well. Specifically, we developed two new research methodologies (a card-sorting task and a gesture-coding approach) to better characterize the state of learner understanding of spatial relationships and complex systems concepts when they lacked the proper vocabulary to describe their understanding verbally. We also developed analytic software that applied spatial analytic approaches derived from ecology, which offers the promise of allowing us to diagnose learners’ spatial strategy changes in real-time (and thus offer just-in-time feedback on their strategies). Our educational simulation is currently being tested as a possible activity to help stakeholders make better decisions about how to integrate green infrastructure into their neighborhoods. If successful, our activity has the potential to positively impact the daily lives of thousands of urban residents.
