Darphin - Modeling and Powerful Ideas

The explanation of thinking in levels helped me better understand the relationship between Piaget's constructivism and Papert's constructionism.  The emergent view of levels focuses on levels that arise from interactions of objects at lower levels - like the traffic jam that emerged from the interactions among the cars (Wilensky, 1999, p. 5).   Papert's theory of learning and teaching emerged from his interactions with Piaget’s theory of knowledge development.  Kafai (2006) clarified the two theories by pointing out, “where constructivism places a primacy on the development of individual and isolated knowledge structures, constructionism focuses on the connected nature of knowledge with its personal and social dimensions” (p. 36).   Maybe we could consider these two theories as levels in our dynamic understanding education. 

When thinking about designing around powerful ideas, we can consider how cognitive tenets of constructionism and the physical tenets of constructivism interact.  Papert’s believes ideas are empowered if they have three components: 1) leverage intuitive ideas 2) are personally meaningful and 3) connect to others (2000, pg. 727).   I agree with Wilensky that this can be done through fostering computational literacy in science classrooms.  Using agent based modeling tools like ToonTalk and StarlLogo are a way of using scientific enquiry to engage students in difficult concepts without allowing algebraic expression to overshadow and disempower ideas.  Students are able to investigate concepts as a professional scientist does.  This brings to mind contextual literacy.  To aid comprehension and analysis we consider the context and read literature written for different purposes very differently.  In classes, teachers cue students to read a letter written by a soldier during WWII “like a historian” or a lab results “like a scientist”.  Computational literacy and agent based models allow students to “act like a scientist” in science classrooms.  I completely agree with Papert’s general thesis that what is good for professionals is good for children (1980, p.30).  I think learning that is authentic to it’s field is more meaningful and might circumvent future transfer of knowledge issues.

 Wilensky and Resnick (1999) and Simpson (2005) support my conclusion that computational literacy can re-empower ideas using Papert’s three components of powerful ideas:

1) Leverage intuitive ideas: Students think about individual creatures. This is more intuitive because students can imagine themselves as individual turtles and think about what they might do. (17) 

2) Connect to others: “students share and discuss not only their current thoughts, difficulties, and conjectures, but working models that instantiate their ideas.” (Simpson, pg. 144)

3) Ideas are personally meaningful: “StarLogo makes systems-related ideas much more accessible to younger students by providing them with a stronger personal connection to the underlying models” and  (16). 

I think students could explore process of how the earth is shaped using computational modeling tools.  It might be helpful to have layers of processes (human, erosion, plate tectonics) to see the levels of interactions.  Students could also use the computer to understand rapid and slow changes.  To make it personal students could create a model for a location of meaning to them.  For some students this might be a southern coastal town with erosion and hurricanes to consider and for another student it might be western city with earthquakes to consider.  I think this could lead to a more robust understanding of processes that shape the earth.  How would exploring the processes of how the earth is shaped on a computer be different than a physical model?

Doherty - Important features of computer modeled systems

As Kafai notes, there is a problematic lack of diversity in professional math and science. This comes as no surprise, however the reason proposed is new to me. That the representations of content and the way in which we manipulate and access that content may specifically reflect the people at the professional level instead of the best or only methods of the discipline. This puts a different spin on the results in Wilensky and Resnick’s work on understanding levels. While they clearly show that the general public, as well as specialized professionals, have difficulty with negotiating the notion of levels within systems, they also show that new representations of those systems aid in understanding. Thus, perhaps with different representation, not only would a wider population have access to thinking in a way that suits them, but the subtle intricacies of  complex content would also be better understood.

In each of the examples suggested by Kafai, Wilensky & Resnick and Simpson, Hoyles & Noss, there are common features that nonetheless are uncommon in traditional instruction. Each feature provides higher levels of student agency in learning as well as flexibility of representation within a structured model.

• Non-static visual representation – Most commonly content is seen without movement or change. Textbook drawings, written explanations or teacher drawn examples may seek to provide a visual representation, yet expect the student to imagine the related actions. Multiple issues arise from this expectation: some students cannot mentally create the moving imagery, there is no way to assess the accuracy of each student’s visualization and the student’s personal assumptions color their mental video. Take for example the traffic jam. Even if the instructor explains the mathematics and explicitly states that the traffic jam will move backward (which of course takes the lesson out of constructionism and into instructionism), it is highly likely given Wilensky et al.’s evidence, that the student will imagine cars being added from behind but the object of the traffic jam moving forward.
• Possibility for student directed input variation – When modeling a situation or system there exist varying factors that influence the outcome of the situation or the behavior of the system. I cannot think of a situation in which all of the factors remain stable in all contexts. For example, gravity might be constant, but the object’s relation to gravity can change. In the collision model, the carts are governed by specific rules of physics, how those rules affect each collision changes based on the mass, velocity, and direction of the carts. By allowing students to change the input values or object states they can explore various cases of a single content idea. The examples given throughout the articles add an important component to this; the students construct each case as it becomes important to them. Choice gives the student agency over their learning, and affords them the opportunity to discover personal questions and inaccuracies in thinking, e.g. the traffic jam, forest fire movement, sheep survival, etc.
• Convincing accuracy through personal construction – Due to the difficulty of understanding the effect of micro and macro actions on a system, as noted in Wilensky & Resnick, scientific results, often through mathematics, can contradict our “common sense.” This can lead to student’s saying “okay, I’ll just believe you” when confronted with a counterintuitive concept.  The student gains no true understanding and possibly thinks it’s not actually, really true because they cannot make sense of it. In these models, there is no “man behind the curtain,” the student either programs the model or they can inspect the existing programming. There is a cyclical process of understanding: disbelieving the results, then checking the code, then trying another case, until they are convinced the model is correct and thus the results must be real. Their learning lives up to their own need for rigor, and is furthered by the time spent figuring out why the code produces the counterintuitive results.

            When thinking about helping the USN students model systems, these elements should be reproduced. Whether or not we create the structure of the program, it’s important that we not limit the scope to our own understandings. I am also reminded that while systems seem complicated in action, the rules governing those actions are very simple. Therefore programming a system is rather easy; it’s interpreting the results that provides challenge. On a personal note, I greatly enjoy the surprising twist that results from this simple-complex dichotomy. It immediately produces questions that are virtually impossible to ignore.

Daily

I have been struggling this week for how to write a blog of some worth to Masters and Ph.D. students. As a rule, I tend to get more than I give in interactions with classmates, and am much more capable of leeching your good ideas than I am at creating my own. Additionally, I didn’t think this week’s readings were that inspiring. As a guy who has spent his fair share of time playing in the dirt, I enjoyed reading about the slime mold, and was challenged by the use of singular/plural pronouns, but that’s pretty much it.

Perhaps surprisingly, I find much of the reading to be intuitive. Should it be surprising that many children are able to think and make connections at the agent level, but struggle to identify patterns and accurately predict outcomes at the system level? In the turtle example, (Kafai, 2006) it is easy to understand how to get one turtle to draw a circle, because the child himself can think through what it would take for him to walk in a circle. Only basic knowledge of what it means to be “a circle” (a 360-degree shape) and control commands for the turtle are required. Understanding how to get a collection of turtles to form themselves into a circle is much more challenging, and it follows that creating the rules for their interactions might be more than the children would be able to do. As someone who was considered a “gifted” student, I acknowledge that I probably have an inaccurate understanding of what a 10-year-old student should be expected to do.

I was very interested in a seemingly tangential quote from the end of the Kafai chapter, “children… often assume that teaching is about asking questions and learning is about giving answers.” This quote agrees with my own indoctrination of what “school” and by extension “learning.” As a non-education professional, and someone who hasn’t entered a K-12 classroom since graduating from a K-12 school, I tend to have a very rigid idea of what these things look like.  While I theoretically can agree that constructionism is probably better than the learning environment I grew up in, I struggle with designing examples unless I see them in the text. What’s more, every time I read about a challenge identified in a study, I catch myself seeing those challenges as proof that this flipped classroom must not work.  The children in the Kafai study, and Keith from Peabody College, might find that they learn more fully in constructionist environments, but building those environments is a bridge too far.

I think that we can use computers to create simple models for any concept involving interactions between objects, which is a fairly vague and open definition. Much like my previous classmates have stated, this includes Physics, Earth Science, Physical Geography. Programs could also be created to model basic concepts in Chemistry and Physiology, but I believe would be too complicated for the value they would provide.   

I look forward to your input!

Eid Carol- Blog 2- Level Analysis and Computational Literacy

Wilensky and Resnick’s reiterate Papert’s emphasis on the role of our language in defining the contours of our thinking.

Languages make a fundamental distinction between the singular and the plural. Indeed, in writing the above description, we needed to decide whether to use the verb “is” or “are” when referring to slime mold. But, as the case of the slime mold shows, the distinction between singular and plural is not as sharp as might first appear… In our view, the very question of “objectness” becomes a question of “levels.” Objects that are viewed singular at one level are best viewed as plural at another  (Wilensky & Resnick, 1999, p. 8, 9).

Our language underlies many of our conceptual misunderstandings of emergent levels since it forces us to submit to binaries when things could, in fact, be much more complex. As described in the above quote, the same object could be viewed as singular or plural depending on our level of analysis. It is as if we have two units of analysis: the singular and the aggregate. The behavior of the aggregate is not a direct reflection of the behavior of its constituent elements. Nevertheless, it could be predicted if analyzed as an “emergent level.” Debugging one’s train of thought is another advantage offered by programming and that is also harder to achieve via our spoken language. This makes clearer that the “way we see the world is greatly influenced by the tools we have at our disposal (p. 14).”

Being able to predict the attributes of the emergent level from the behavior of its constituent elements can have a tremendous impact on the lives of students and people in general. Our daily actions shape our cultures, societies, and environments in ways that are hard to predict without level analysis. Nevertheless, we do in many cases formulate erroneous predictions and act accordingly when there could be better ways to act.

When people see patterns in the world, they tend to assume centralized control even if it doesn’t exist. And when people try to create structures in the world (such as organizations or technological artifacts), they often impose centralized control even if it’s not needed (p. 9).  

Can you imagine how a deeper understanding of patterns that emerge without central control could affect our society? Organizations could then be designed in different ways and people’s relationships may be redefined. For instance, designing less centralized social structures may engender less social hierarchy and privilege. Moreover, as echoed by Wilensky and Resnick, it is hard for us to predict how our individual actions contribute to macro social patterns. Gaining such an understanding may assist us in combating global concerns, such as ecological and ideological threats, by understanding what and how individual actions affect emergent patterns.  

               Finally, I believe that fostering computational literacy could also “empower ideas” since it offers insightful discoveries—is “powerful in its use;” reveals connections that are hard to see otherwise, such as those between different levels or different disciplines (physics and chemistry for example)—is “powerful in its connections;” and relies on learners’ previous knowledge of micro-levels to predict macro-levels—offers syntonic learning (Papert, 2000, p. 727; Wilensky & Resnick, 2000; Wilensky et al., 2014). Consequently, computational literacy could bring the “magical flare” (that I described in my first blog) back to ideas by revealing underlying patterns and connections.  

               My question is whether people who practice level analysis will be able to better predict macro phenomena without using computational literacy tools. In other words, could level analysis become a skill that helps us better analyze social or aggregate phenomena without using computers—is it a transferrable skill? Or will it remain a computer-bound one? 

Amanda, powerful ideas in science

Collisions, like in Simpson et al. (2005), are one example of a potentially powerful idea in science. Children witness different kinds of collisions in everyday life, particularly 1-D collisions or two colliding objects, so the idea of collisions can be powerful in its roots by building on students’ prior experiences. The authors even say that “it is more productive from a pedagogical perspective to focus attention on knowledge that is contiguous with students’ prior understandings” (Simpson et al., 2005, p. 6). Genetics is another example from science that connects to children’s roots. It’s something children witness every day and talk about, like eye color, hair color, skin color, height, weight, gender, etc. Levels, in the way it’s presented in Wilensky & Resnick’s (1999) paper, is another powerful idea because it connects to problems in a variety of situations that involve systems, including lots of things in science.

Doesn’t every idea have the potential to be powerful, depending on how it is used or how a learner is exposed to it? In other words, are there really any ideas that could never be considered powerful? It seems like ideas could be disempowered based on how they’re used or taught, but they could certainly be empowered again if approached in a different way. That’s probably the point, anyway. It’s not whether the idea itself is powerful because it can’t really be powerful by itself, it’s only powerful when you consider the idea in context with the people engaging with the idea, the tools available to explore the idea, etc.

On another note, the notion of reaching “more students more quickly by getting science teachers to add computers into their classes” rather than “developing a nationwide cohort of computer science teachers” is interesting (Wilensky, Brady, Horn, 2014, p. 24). It may be quicker because you don’t have to wait for a bunch of new teachers to be trained and to take the time to restructure the school day to figure out how to fit a separate computing course into the agenda. But is it actually better? They say that they’ve experienced that some teacher training and materials are sufficient for bringing computational modeling into science classrooms, but what about other kinds of computing in other classrooms/subjects? For instance, is it really quicker or even better to train existing math teachers to bring programming into their math classes? Maybe. I would think that to do it well, it requires strong disciplinary knowledge. So good math teachers with strong mathematical understandings and an interest in learning more would probably be able to incorporate computing concepts into their classes faster and better than people trained to be computer science teachers. But it’s a difficult balance to find.  

I think a bigger argument for this strategy of incorporating computing into existing classes is that it exposes more students to computing than if it was a separate, possibly elective course. Students who wouldn’t normally sign up for a computer science course would be exposed to computing concepts, which is really valuable. And it exposes them to computing ideas in a potentially more powerful way because students might make more connections to situations outside of computing, like mathematics or science (whatever course it’s included in), which they might find useful for solving problems that they’re interested in, and which might build on students own knowledge and identity with the discipline of focus.

Bergin, Constructionism in the Classroom

     In "Chapter 3 The Cambridge Handbook of The Learning Sciences" Kafai clearly reviews Papert's theory of Constructionism and clarifies how it differs from Piaget's constructivism. Kafai makes the distinction that constructionism "focuses on the connected nature of knowledge with its personal and social dimensions." as opposed to Piaget who focuses on individual centered development. This becomes important to us as educators and program designers, because it immediately makes a case for computers as an optimal tool for this new type of learning. The style which strays away from the typical acquisition metaphor in which students system gain more knowledge from direct instruction rather than participating in a more authentic way.   

    This is where we begin make the link between Papert's theories and application through, Wilensky and Resnick's proposed modeling method. In the article they discuss how students were able to use the computers as a tool to replicate real life situations, these replicas allowed the students, and even adult researchers to gain different perspective on these scenario, and even promoted productive cognitive dissonance. For example, when students viewed the traffic jam moving backward, it prompted a discussion on the advanced concepts of levels. Additionally, when another student named Benjamin enacted andsimple-minded' ecosystem scenario, with few variables and yet he was able to visualize and, draw new inferences and conclusions about the complex interaction such as predator vs prey relationships.  The key factor in these two programming models is not just that students were able to create a visual aid to solve a single problem they were struggling with but in these and the many other listed examples, the computers promoted metacognition. 

     As Papert had hoped, as we sift through the many example of computer programs designed and/or used by children and adults we see evidence of a shift in thinking. Kafai points out "children learn to articulate procedures, recognize repetition, and 'debug' their own thinking when programs don't run as expected" In the case of Benjamin's ecosystem he was natural using the scientific method without direct instruction, he made an observation, developed a question, formed a hypothesis, ran an experiment, drew a conclusions, and repeated different scenarios to see how the results would be affected without specific prompting from the teacher. The program allowed him to have an authentic learning experience and truly embodied Papert's theory of computers as "objects to with" as opposed to our typical definition as tools or supports.  

The goal is to create similar circumstances for out students at USN:

I think science and math provides the biggest list of topics to work to create

physics models: (laws of motion, force, velocity)

fraction models 

-ecosystem simulation

water cycle

roller coaster simulation

immune system simulation