Thanks for taking a look at our video. The good news is that our technology partners, Mursion, have a number of simulated classrooms that they have developed (including the one for this project that is shown in the video) and do demonstrations with faculty so they can get a firsthand experience with the student avatars and see how the simulation works. In the demo, you can talk to the students and see how they respond in real-time during the interaction --it's quite amazing and feels very authentic once you start talking to them. If you go to Mursion's website (https://www.mursion.com/request-demo/), you can click a button to request a demo and someone from their end will reach out to you directly to schedule the demo.

Thanks, Jamie!

Julie, We would be happy to share our work and immerse you in a simulation.  We have an upper elementary class, middle school class, and education leadership scenarios.  We are currently partnered with about 65+ institutions for teacher education, and there may be partners in your area who could show you their work. It is very interactive, and there are so many application - we encourage you to explore!

Thank for the details and link to Mursion. Much appreciated!

Thanks for your questions Mac. This simulated classroom is definitely something that others can use. In fact, over 65 teacher education programs across the U.S. have used these types of simulations to help their teachers learn. If you are interested in learning more, I'd encourage you to reach out to our technology partners, Mursion, to schedule a demo (https://www.mursion.com/request-demo/). Mursion also has a bank of various performance tasks that they and their collaborators have developed over time -- across different grade levels and content areas and for various purposes -- so there is a lot of flexibility in terms of using this simulated classroom. Mursion also has a number of different classrooms (e.g. elementary and middle school) and avatars (both students and parent avatars). On this project, creating the new elementary classroom with the five student avatars took about 9 months in total -- working collaboratively with Mursion on the design and iterative testing of the interface. In terms of designing the actual science and math performance tasks, we used a design based iterative approach whereby we piloted the tasks a couple of times and refined it based on feedback from preservice elementary teachers, external consultants, and our advisory board. We also developed a set of training materials to support the interactor (the "human in the loop" who is behind the scenes and is acting and responding as all five student avatars in real time during the interaction) in delivering a standardized experience for the teachers during the interaction. Our process for creating each new task and interactor training materials occurred over a few months, however, I know that Mursion has created other types of custom scenarios in as little as a couple hours -- it just depends on what you want, the focus of the task, and the content and teaching competencies that it is designed to address. But all of the components are definitely able to be customized and you can create new tasks based on different content for use in the simulated classroom.

Thanks, Heather!

Mac, Heather and Jamie's project is highly customized, and the system allows for all levels of customization. It can be used across content areas, so we design simulations in science, mathematics, English language arts, social studies, and even educational leadership.  This project is unique because we are looking at simulating discussions and focusing on providing opportunities to increase candidates' ability to facilitate discussion across content areas.  If you are interested in experiencing a simulation, please reach out to us at Mursion. 

Thanks for watching our video. For this particular NSF project, we will only be following the pre-service teachers during their science or mathematics methods course in their teacher education program. However, in the future, we would definitely like to follow pre-service teachers who are using the simulation back into their classrooms with real students to learn how they are applying what they learned in the simulated classroom to their work in real classrooms and whether and in what ways that impacts their students' learning. We think that is an important and vital next step in this research!

Indeed, a huge part of our work this year has been focused on training raters to score and provide feedback on performances. We have them do these activities in tandem, but only the written feedback goes to the preservice teachers. The scores go to the teacher educator and of course we will be analyzing them, but based on preliminary evidence it seems like there is considerable improvement, but still significant room for growth, in the skills targeted in the feedback over the course of the semester. Our rubric focuses on five elements: elicitation of student thinking, coherence and connectedness of the discussion, encouragement of direct student-to-student dialogue, content understanding, and focus on argumentation. It is definitely the case, based on what the preservice teachers have told us, that the areas we provide feedback on  are helping them focus as they move from one performance task to the next.

I wish it had been available to me too! Its a good question... and probably mostly technical so I'll let Carrie weigh in, but I think the short answer is not quite now but in the future maybe. Right now a match has to be made between the person on the back end who knows the task and the person who wants to engage in that task, so you can access it whenever you want as long as you plan ahead and schedule it, which is probably not quite the on demand vision you had in mind:) This is probably built into the human in the loop model to some extent, and although you can imagine ways that might work to scale efficiently, its tricky in the details. Will AI eventually replace the human in the loop? I can't call that one at this point. Maybe. (If I'm being honest, it already fools me online all the time in ways that I might be surprised by if I really understood what was happening.) Its harder to see it happening with something like teaching, but I have to think things that seem impossible now will be easy at some point. And I'd bet that the work we are doing now to define the parameters for the human interactors could be used to ground that kind of development effort.

The work being done on this project could eventually be used to support in service teachers to facilitate classroom discussions.  Currently K12 districts can change the profile and look of the avatars, much like was done in this project, to give teachers a range of personalities and learning profiles to interact with.  

We haven't been studying relationships between the avatar characteristics and the learning experience per se... I do feel like the avatars whose student work (for a given task) evinces the clearest misunderstandings of content get the most attention, and we've made some efforts across tasks to make sure its a different mix of content understandings, partial understandings, and misunderstandings each time so that no one student becomes a consistent weaker or stronger student.

While we aren't working with inservice teachers on this project, I think there's a lot of potential there. Reflecting on my own work as a teacher, I think providing the opportunity for an experienced teacher to try out a new approach in a safe "play space" could be really powerful in addressing the tension between wanting to try new things and grow professionally and knowing that you're more likely to do something more familiar well in the short term.  I also think there's a lot of open questions about how this might or might not work differently with more experienced teachers. For example, they might be more sensitive to small divergences from authentic student approaches because they know more about what is realistic.  I would also anticipate that their more developed teaching approaches might require that we build out our sim specialists training in different ways.

Jamie has related project work with experienced teachers and I'll let her speak to that.

I'm curious what comes to mind for others on this point.

Frank, Your questions are top of mind for us at Mursion. Currently the system allows people to choose social and cultural profiles, create their own profiles, and design social interactions.  There are huge implications for every decision and as we move forward designing scenarios and interactions, and we are cognizant of the balance between providing an authentic environment, while at the same time not essentializing children.  Our mandate is to be aware of our bias as designers, learners' bias, and to be open to exploration of how the scenarios we are designing may impact learners. It is complex, but important to keep front and center. Thank you for your thoughtful commentary!