Wilensky, Brady, & Horn (2014) resonates with Papert and Resnick's argument that students are typically consumers, rather than creators or users, of technology. From this perspective, it seems like one powerful idea for science is that computers can be used as tools for learning, expression, problem solving, and exploration. This idea is powerful in that it can connect disciplines and practices in science, is rooted in students' experiences with computer use and problem solving, and is useful in a wide range of contexts. However, it seems like Wilensky and colleagues believe that this idea comes for free as students engage in curricula centered around agent-based modeling of emergent phenomena.
Wilensky and colleagues (2014, 1999) would argue that the idea of "micro" and emergent "macro" levels is powerful in that it can be used to make a range of concepts (ideal gas laws, predation, firefly behavior, global warming) accessible to students while leveraging their intuitive understandings about agent-level behaviors. They would also argue that this idea is especially powerful in the context of netLogo because papers have been published using netLogo research, suggesting that netLogo has a high ceiling that makes it useful in K-12 classrooms and beyond. I am a little more skeptical of this claim, because I think that to authentically participate in the scientific community students would need to work in a more popular/adult language, if only so that they can more easily collaborate with colleagues. Still, the ideas of thinking in levels, randomness, and emergence could be applied to a range of concepts and programming languages, whether developed in the context of netLogo or scratch.
These ideas, however, do not necessarily support students in seeing models as tools that can solve real-world problems, particularly in a typical K-12 context (i.e., greater than 1-6 students with a traditional teacher rather than teacher-researcher). What work needs to be done on the part of the teacher to make these models feel real-world in a classroom setting? How do we go from ABMs that are descriptive to ABMs that are constructive, investigative tools? These questions come from my experiences with agent-based models in middle school classrooms, but are mirrored in Simpson et al. (students observing, creating a specific and narrow narrative model, and being "done" when it sufficiently mimics the observation). While students were able to use diagrammatic and physical models to test and generate new ideas, their computational models tended to serve a different purpose. Often, students gained a deeper, more concrete understanding of the events of a phenomenon as they tried to illustrate that phenomenon with a computational model. However, this seems more like using models as descriptive tools rather than "powerful" tools that are generative, connected, and useful beyond reviewing or examining content. This might be a question of goals - do we want kids to adopt and develop scientific practices around computational modeling so they can deeply understand known, but typically intuitive phenomena, or so that they have a tool for iterative and generative knowledge building?
**Powerful modeling ideas for working with USN:
Very observable:
-Ants (ecology)
-Greenhouse/methane (chem)
Less observable:
-Optimization problems (how can we make the robot quickly clean the room, etc.)
-Immune system/natural selection (bio, but also easy to make into a game)
Elaborating on ideas for USN:
Ants are a powerful idea because their foraging patterns are used to help scientists understand a range of phenomena, from brains to the world wide web. Ants naturally employ very efficient search algorithms in their foraging patterns. They also are all over the place. It is (relatively) easy to build an ant habitat for this context, video ants so that kids can break down the behavior patterns, and have them create computational models to represent the search patterns. I think it is powerful because it is rooted in things familiar to kids (ants, movement, drive for food and survival), because it has powerful use (i.e. optimizing networks or searches), and because it can connect to other domains (computer science, microbio- cancer sometimes behaves like ants, etc.)
Natural selection is powerful because it draws on similar intuitive understandings, but can be applied to a range of contexts as small as cells and and as large as species.
I suggests greenhouse gasses because it is compelling in terms of impact, and optimization problems because they have a wide range of uses, but I have less experience implementing projects around those ideas.
Wilensky and colleagues (2014, 1999) would argue that the idea of "micro" and emergent "macro" levels is powerful in that it can be used to make a range of concepts (ideal gas laws, predation, firefly behavior, global warming) accessible to students while leveraging their intuitive understandings about agent-level behaviors. They would also argue that this idea is especially powerful in the context of netLogo because papers have been published using netLogo research, suggesting that netLogo has a high ceiling that makes it useful in K-12 classrooms and beyond. I am a little more skeptical of this claim, because I think that to authentically participate in the scientific community students would need to work in a more popular/adult language, if only so that they can more easily collaborate with colleagues. Still, the ideas of thinking in levels, randomness, and emergence could be applied to a range of concepts and programming languages, whether developed in the context of netLogo or scratch.
These ideas, however, do not necessarily support students in seeing models as tools that can solve real-world problems, particularly in a typical K-12 context (i.e., greater than 1-6 students with a traditional teacher rather than teacher-researcher). What work needs to be done on the part of the teacher to make these models feel real-world in a classroom setting? How do we go from ABMs that are descriptive to ABMs that are constructive, investigative tools? These questions come from my experiences with agent-based models in middle school classrooms, but are mirrored in Simpson et al. (students observing, creating a specific and narrow narrative model, and being "done" when it sufficiently mimics the observation). While students were able to use diagrammatic and physical models to test and generate new ideas, their computational models tended to serve a different purpose. Often, students gained a deeper, more concrete understanding of the events of a phenomenon as they tried to illustrate that phenomenon with a computational model. However, this seems more like using models as descriptive tools rather than "powerful" tools that are generative, connected, and useful beyond reviewing or examining content. This might be a question of goals - do we want kids to adopt and develop scientific practices around computational modeling so they can deeply understand known, but typically intuitive phenomena, or so that they have a tool for iterative and generative knowledge building?
**Powerful modeling ideas for working with USN:
Very observable:
-Ants (ecology)
-Greenhouse/methane (chem)
Less observable:
-Optimization problems (how can we make the robot quickly clean the room, etc.)
-Immune system/natural selection (bio, but also easy to make into a game)
Elaborating on ideas for USN:
Ants are a powerful idea because their foraging patterns are used to help scientists understand a range of phenomena, from brains to the world wide web. Ants naturally employ very efficient search algorithms in their foraging patterns. They also are all over the place. It is (relatively) easy to build an ant habitat for this context, video ants so that kids can break down the behavior patterns, and have them create computational models to represent the search patterns. I think it is powerful because it is rooted in things familiar to kids (ants, movement, drive for food and survival), because it has powerful use (i.e. optimizing networks or searches), and because it can connect to other domains (computer science, microbio- cancer sometimes behaves like ants, etc.)
Natural selection is powerful because it draws on similar intuitive understandings, but can be applied to a range of contexts as small as cells and and as large as species.
I suggests greenhouse gasses because it is compelling in terms of impact, and optimization problems because they have a wide range of uses, but I have less experience implementing projects around those ideas.
