VanLehn

What kind of models or activity ideas from Van Lehn do you think would be important to incorporate when having students model Zika?
First of all, as VanLehn clearly outlined, it is very important to have students engage in model construction rather than simply model exploration. Construction involves actively "debugging" the model and becoming familiar with its language, whereas exploring a model does not make a student engage with it as deeply. 
A good model that could be incorporated into a Zika Virus lesson plan would involve scaffolding on the part of the teacher, or "sophisticated systems that give feedback and hints to the students" (page 20). That is, while the student is creating their original model, the teacher can tutor, clarify the language, and introduce more complex problems to solve. For example, this model could be  a qualitative systems-constraint model, such as a node-link diagram constructed using a program like Betty's brain. Each box or variable is given certain parameters, and the diagram can answer questions such as "If we increase the variable of standing water, what will happen to the variable of infected humans that travel?" Students can construct their model and assess how well it answers questions of varying complexity, and make adjustments as necessary. Scaffolding comes into play when the program being used asks students thought-provoking questions about their model, or the teacher clarifies the meaning and purpose of certain variables. 
In a qualitative systems constraint model, it is essential for students to construct their own models and maybe compare them with those of their peers. If the same model of the Zika virus were created by the instructor and simply presented to the student to figure out, the student might not grasp the processes of deciding which variables are important, and the language of the program, and would not be able to create similar models in the future. VanLehn made a very good point that although model construction may cost more- in time and money- on the outset, the gains made for the students are well worth it in the long run. 
Of course, another model proposed by VanLehn would be an agent-based model, using a program such as NetLogo or StarLogo Nova. What stood out to me in the VanLehn piece that would be integral to creating an effective Zika virus model is whether or not the model is executable. Does the model predict the behavior of the system? Moreover, can the student'smodel make such predictions without making false assumptions about natural events? In other words, in an agent-based model, students should strive to create predictive models while also examining/understanding the "underlying mechanisms" that go into the model rather than "tweaking" the model so that it gives them an answer they would like to see (page 17). 
My main takeaway from VanLehn is that the active construction of models with interacting variables can be a very enriching experience for students is they engage in thoughtful practices and are assisted by various scaffolding techniques by both the instructor and the program or language chosen by the instructor.