Innovations in STEM Education: Technology to Support Students with Autism

Brought to you by the U.S. Dept. of Education’s Offices of Special Education Programs and STEM Initiatives. Presenters shared their leading-edge research and experience in developing technology supports to give students with autism access to STEM curricula and activities. (Get the slides)

Transcript: 

- [Russell] Welcome everyone. My name is Russell Shilling from the, I'm the executive director for STEM initiatives at the U.S. Department of Education, and on behalf of the Office of Special Education Programs and my Office of STEM Initiatives as well as the Center for Technology and Disability, I'd like to welcome everyone who's phoned in today for our Innovations in STEM Education, Technology to Support Students with Autism webinar. Before we get started, I would like to ask, please do not put us on hold today. Although we enjoy music, we do not enjoy muzak, so please if you have to put us on hold, please just dial back in instead. To get us going, I'd like to say that the STEM office here is two years old, and our mission has been from the beginning to improve STEM education for all students, and that includes students with disabilities and students on the autism spectrum. I have a personal interest in this area since I have sons, fifteen and seventeen years old, who are both on the spectrum. Before I got here to the department, I was at DARPA with the DoD, and we developed some technology out of our programs there that are now being used for job interview training systems for students with autism, being used by the Dan Marino Foundation and USC. And we also started having conversations at that time about what we needed to be doing to design software better for students with autism, and we discovered there weren't a lot of answers to that. So one of the reasons we put this together today was to actually start this conversation up, and we really think it'll be an ongoing conversation. So not only will we be discussing universal design for learning, we'll also discuss some of the approaches to design software better for kids with autism and then also how to use technology for students with autism to better teach STEM topics. So with that, I will introduce our four speakers because I'm far more uninteresting than they are. First, we have Jeff Munson from the University of Washington whose research deals with the prognosis and interventions for autism, and he'll talk to us a bit about the use of games and some other issues in STEM education. Matt Marino from the University of Central Florida is a STEM technology specialist for kids with and without disabilities. Maya Israel from the University of Illinois Champaign-Urbana will discuss a bit about universal design for learning, computational thinking, and computer science. Amelia Moody from the University of North Carolina at Wilmington will not only deal with robotics but talk a little bit about her expertise in using technology to support children with disabilities. Before I turn it over to Jeff, I'd also like to remind everyone that at the end of this, we will be presenting a questionnaire we would love everyone to fill out, giving us feedback about how the webinar went today and what we could do better maybe next time. And with that, I'll turn it over to our first speaker, Doctor Jeff Munson, Jeff.

- [Jeff] Great, thank you, Russell. It's a pleasure to be here with everybody, and to kickoff this event, I'm a child clinical psychologist, and like Russell said, been involved in research not only in treatment in following the course of autism but also in our understanding of brain development and genetics. So it's covered a wide range of things, and similarly, in thinking about technology and STEM education that support students with ASD, I'm gonna kind of take a broad view and review some specific examples that, at my institution, that we've been working on. So we're gonna cover the basics really quickly. I'm gonna talk about, well, what is technology and why is this so critical for students with ASD, and finally, like I said, three examples that I've been involved with. So first the basics, ASD is a neurodevelopmental disorder. Symptoms are present very early in life, certainly by age three, and diagnosis can happen even by age two by, you know, a trained professional in the area. And the deficits are characterized first by social communication and social interaction difficulties. These folks don't engage with other people as readily or as easily and have sometimes less interest in doing so. Secondly, there's restricted and repetitive patterns of behavior. This could be physical movements like rocking or repetitive hand gestures as well as limited and focused interests that are often atypical and get in the way of their interacting with others. The current estimates of prevalence are one in 68, and it's far more common in boys, where you can see it's one in 42 in boys but only one in 189 in girls. So it's a very common disorder, but it occurs across all racial, ethnic, and socioeconomic groups. And at this point, we have no definite diagnostic markers or biological or medical tests that we use to identify autism. It's the onset of symptoms is what, you know, tips us off in terms of that the child's development is problematic. This slide is busy, but the take home here is that just in special education across, you know, the decade from 2000 to 2010, been a huge increase in autism. Corresponding, there's been a lower, you know, a reduction in students classified with intellectual disabilities, so the, sort of the makeup of the special education population has been changing, but some of this is, you know, moving categorizations and focus in this population. So technology, it seems like maybe this is an obvious thing, but I was curious, when I looked at the definition, some of the details here. What I want to emphasize is technology's not a screen. It's not a computer, and batteries aren't required. Rather, technology is the practical application of knowledge or a means of accomplishing a task using knowledge, and that's just straight from the Merriam-Webster Dictionary. And this is important because we gotta think broadly. I hope we can think broadly about the application of technology and that it is more than just screens and computers, although those play a central and important role. So this is just a picture to show the complexity that all students with ASD face. They live in families, and they have relationships with peers. They're in schools. They have different medical and clinical settings. Knowledge transfer and information transfer is a challenge in this and that all of these are embedded in the broader community and sort of our economic and business world where we want to see these kids thrive and develop. And my focus is gonna be on how we've leveraged technology to improve some of the information flow in this arenas. So first, we've got a pencil. The pencil, humble as it is, is an incredible piece of technology. Jean Piaget, the famous developmental psychologist, wrote, "I could not think without writing," and clearly, recording accurate information is a critical task in educational and treatment planning and evaluation. So it's not easy though. Any of us involved in the process know how challenging this can be. So we think about technology; what can it do for us? Let's imagine, pencil 2.0, the latest and greatest. This doctor in the cartoon says, "Good news. "When we transferred your medical history "to our digital system, it suddenly became legible." So the magic of a technology, we hope, can simplify and improve our lives and especially in information management is a key part. The one project that we've done here at the University of Washington was taking our five week, full-time summer treatment program. It's essentially a summer camp for kids on the spectrum. We had 90 kids, each with individualized goals, and 40 intervent, sorry, 40 intervention staff, and traditionally, we've had a paper-based data collection system where we record lots of details of a child's behavior and how they're progressing in treatment goals. But this year, transitioned to a tablet-based system developed by startup company, Experiad, with a small business innovative research grant that they received from the National Institution of Health. This Motivity program is simply a replacement for this paper-based system where 25 behaviors are tracked continuously and tied to the program's token economy, system of rewards, and a way to track progress students have towards their goal. It recorded 70,000 behaviors and saved four hours per person per week with this system, and the immediate benefit was that staff found they had more time to actually discuss treatment progress and goals, and it was based on more comprehensive and better information. Here's just a picture of it in use on a field as kids play games, zoom in on some controls, and summary reports and data visualization is just part of the package. So this freed up clinicians' time. But remember, this is information that they themselves created and had created in the past, just vastly improved its organization and availability. The second project focused on the effect that we've seen in infant development, that there's an increased presence of atypical or stereotyped motor mannerisms in babies who go on to develop autism. This is so expensive and labor-intense. So this project seeks to take what a baby book where we, well commonly, as parents and interested grandparents and folks, record the development of progress of kids. Here we're using sort of computer-based machine learning to track babies' movements and to record it making a library of, a library of movements that can be early markers of potential conditions like autism and others where atypical motor development is characteristic. Here, we've got information that we know we want to record, but we're gonna employ technology that will allow us to record it, where otherwise, it doesn't really exist. So the third project focuses on what do we do when we don't have information or we don't even know what we want to record? Here and just pause and think about not just the pros but what are the cons of technology? Is it saving us or failing us? And I found over this past year just two examples that illustrate an optimistic view and a pessimistic view. Saw this ad in a airport. How do you turn 3.6 billion medical data points into a miracle? It goes on to say, "It's simple. "The answer is SAP HANA," a company that creates and builds information management systems. Very optimistic, it's amazing, a miracle, and not only a miracle, but this one's a simple miracle. Everything sounds good. There is a pessimistic view though, on the other hand. So this is from the MIT Technology Review magazine, on the cover. It said, "You promised me Mars Colonies. "Instead, I got Facebook." A little bit let down. Mars Colonies sounds so great, sounds cool, but this person here was disappointed. What was interesting to me was how these publishers communicated this optimism and this pessimism about technology. They did it with a human face. Here, we see this young woman who's clearly optimistic and happy, so much that she can even close her eyes, and she doesn't even need to look ahead. Just sit back and enjoy the journey, whereas this gentleman, This is Buzz Aldrin, the astronaut. He's certainly serious, weary, maybe skeptical. You can see these are the two images they saw. He says, "You promised me Mars Colonies. "Instead, I got Facebook." Not too pleased with this situation. What's critical though is the subtitle, which you probably can't read, but it says, "We stopped solving big problems." And it says, "Meet the technologists "who refuse to give up." My work in studying children with autism, the big problem has been those who don't have much spoken language or even any communication at all. So there's a story about, no, a story about a situation, I observed a child who was given this task, to solve where does this object go belong, in this pile or this pile? That's something I'm confident everyone listening knows the answer to, but this was a legitimate and reasonable task for this 10-year-old boy who was very limited in his problem solving and communication abilities. My hope is that technology we can use to help shine on a light on this situation such that we can better understand what competencies these individuals have. What struck me, this is in the plan was, who did get this correct, how would that success impact other people and, this much, I think technology is a way to change, some of that, and these applications are truly unique opportunity to serve these folks with severe communication impairments that we haven't maybe thought about. This just illustrates what we've done in our group, showing about 20% of kids. This is across like a 10 year time span where there's almost no gain in their communication skills. This is something that is a huge challenge to families and clinicians, educators, and we just don't know very much when someone can't communicate well with us. So to address this, I've been starting a process of building games for these students who are very minimally verbal. This is just a little screenshot of a game where the students could pop a balloon as they see themselves through using the Microsoft Kinect. We'd color their hands so that you have to match the color of the balloon that they're popping. Some advantages here, that it's engaging, we can repeat these, we can allow these students to explore and see what questions they ask. Typical assessment settings are just often adults asking questions, questions, questions. Let's see how do these kids explore. Another key is that we have the ability with technology to amplify what may be a really limited skillset such that categorizing a spoon and a fork and placed in a context of a game can actually make a difference and be a positive contribution to other kids in the classroom. We also need to think importantly about designing folks with such limitations aren't gonna be walking into a room, necessarily on their own. Focused on learning and the achievements students make, but like the graph we saw earlier, those are students who have been failing repeatedly, maybe for a decade, and families and those caring for them are really faced with a super challenging situation. These plots are just busy and small, but I just want to highlight the fact that the students playing this balloon pop game, on the bottom row, C and D, these are kids who are swaying, and imagine just swaying your arms back and forth either in a vertical way or a horizontal way, that they're playing this game and solving it even though they're doing it in an incredibly inefficient manner. But this student here, number or letter B, he just swung his arms kind of like making a snow angel, just in front of himself, up over his head, and back down and really didn't engage with the content at all. But in talking with his sister, we saw that he liked to scribble. He liked to color with crayons. So I gave him a tablet, drew some circles myself, encouraged him to fill it in, and he didn't want to mess with that. But he did like to scribble, and he learned how to click an icon and change the color of the crayon. So this is a place where we've recorded a positive, intentional behavior that we can start from. This is something positive, something we can perhaps learn from as he progresses. These little squiggles are just similar notion that he's a gamer; you control the tilt of an airplane. And when he played this game, all he would do is sort of rock back and forth in this similar, you know, pattern. Couldn't play the game, but what can we learn about recording in detail and observing how it is he goes about taking part? Finally, this was an example of again playing this, walk you through these five pictures that were surprising to me when I saw them. So first he starts a level on A. It was a simple task. Lift both hands and you can pop the red and blue balloon. But B has switched the colors, so now the blue and red are on the opposite side. So in the second picture, let's just repeat what worked the last time. Lift your hands up and then nothing happened. That was repeated seven times. You can see on panel B, this student kept trying to pop the balloons with the wrong color. Finally, in number four, he said, "Ah, I figured it out; I've got to switch hands," but he tries a movement that was physically impossible, to reach them both simultaneously. And then only after that failed did he go to number five and realize, oh, let me start with the red hand, and then I'll cross over and start with the blue hand. This sort of in-depth ability to focus on a really simple behavior has a possibility of giving into what is going on with these kids who are really struggling to learn even the basic problem solving and communication tasks. So my hope and excitement about this is that parents, educators, clinicians, the whole gamut need to be talking to each other, transferring this information and informing each other, what are the tasks we need solving and how can technology help us and to avoid Walt Disney's approach. This is a quote from a book about him. He had "a distrust of engineers" because he felt they were people "who designed primarily for themselves "without regard for the intended use of the product." So I don't think we need to, obviously not shun engineers. We need to embrace them and what they have to offer, but we need to make sure they understand the challenges that we face in caring for students with ASD. So in conclusion, I think the tools we have at our disposal are amazing, and I'm looking forward to hearing the rest of the speakers as they talk about their work in supporting students with ASD. And my encouragement is to not leave anyone out, even these folks who haven't learned much over a decade. There's things there that we can use technology to bolster our care and service of them. Thank you.

- [Russell] Thanks, Jeff, and one more thing for the audience. We will be doing our four speakers first, and then we'll have some time at the end where we can open up for panel discussion. So with that, since Matt or Jeff gave us a warning about Disney, we'll be on to Orlando and the University of Central Florida with Matt Marino.

- [Matt] Thanks, Russ. So my name is Matt Marino. Thank you, Jeff, for that very enlightening presentation as well. My name is Matt Marino. I'm an associate professor at the University of Central Florida, and quite honestly, when I started my life as a pathobiologist, I never thought I would be in education because my father was a lifelong special ed. administrator, and I never thought out I would end up there. But now, here I am, giving a talk on students with autism, and I'm gonna focus on students with games today. The target is gonna be a little bit different than the folks that Jeff was talking about who are minimally verbal. The folks that I work with most are in inclusive science classrooms at the middle school level. And so, we're talking about kids in grades six through eight mostly, and now we've expanded some of our research to look at high school and to look at how they can expand their opportunities in STEM, so that's really what I'd like to talk to you about today. First I have to say, thank you to the Department of Ed. and the National Science Foundation for funding the work I'm gonna talk with you about related to games and videos games specifically and students with autism spectrum disorders. And so Jeff already went over the information on this slide. I just want to clarify the PowerPoint presentation should be downloadable to you folks. So all the information that's in our slides, my understanding is you'll have access to. I've been watching the chat box, and I know there were some questions about that, but I think it's been straightened out now. So feel free to go ahead and download the presentation. There are some slides at the end of this presentation that I'm not gonna talk about, but they give you a lot of resources to STEM-specific games, and so that'll be the focus of this talk. So why would we identify students with autism and say that we need to increase their participating in STEM careers? And this is something that we're very interested in at UCS and something I've been working on for the past several years. And so the students with autism who I work with have very sustained and hypersensitive attention to detail. Right, they're meticulous about the work that they do. They absolutely love the gaming environment, and they have incredibly honed skills related to their systematic approach to gameplay. And so, we started thinking, how can we harness that and use it, in a way that, clearly, if we look at the statistics, students with disabilities are not doing well in science, and so on this slide, you can see with the 2009 standards which we, we now don't have any test results yet to see how the kids are doing. You'll see I've entered the basic scores for 2009 and 2011, students with disabilities, and this is all students with disabilities, represented here. And so if you look at them compared to their peers who do not have disabilities, you see they score much more often at the below basic level than any other level. So that's inexcusable; we should do something about that. And so I started thinking about the kids I had in my classroom who had spectrum disorders, and they loved to play video games, right, whether it be Minecraft or any of the other games. And it's really interesting if you can see this slide and you look at the date, it's from 2008, and then when you look at the recent statistics, you'll see it really hasn't changed much. So there were a couple of recent studies, one out of the Pew Research Center and another one from a blog that was published by bigfishgames.com, and they looked at how often kids played games, and they found that there's still a lot of kids who are interested in playing games. And so video games help students with ASD if they follow universal design for learning. I'm gonna kind of kickoff this discussion of UDL, and Maya Israel's gonna talk about it a little bit later on. But universal design for learning is this notion that you can use multiple means of representation within a game, and you can provide multiple options for students to demonstrate their comprehension of concepts or phenomena, and you can allow students to engage with materials in a diverse way. So a game is an environment where the student can experiment and go in different direction in a dynamic changing. So on this slide, you'll see a game in the background and then all these call-out boxes. And this is a representation of what UDL looks like in a game. It's kind of hard if you know UDL and you've looked at the guidelines, you know, basically bullet points within each of the columns. But to take that and translate it into what it looks like in gameplay is challenging. So I've tried to do that for you on the next couple of slides, and I'm not gonna talk, I think for students with ASD, there are a couple of these that are extremely important. The first one on this slide is in the bottom left, and it, modeling. They can provide modeling here options that the student is going to have with them, and so as an example of this in a non-STEM context is this game called the Social Express, which information related to social cues and it, and a student and gives them prompts, and they go through and do this. So there's in-game modeling, appropriate social skills that should be included, if you folks purchase related, also, if you, ability level, so if it's too challenging for the student, you want to make sure that either the game adjusts itself, or you can adjust it for the student so that you can present it at the student's zone of proximal development, right. So you want to makes sure that it's not too challenging or too easy for the kid, or they will not be engaged. The next slide has very similar call-out boxes, but it's a, so I've done this for you so that you could see what the same type of tools, these are cognitive tools that are included in the game, might look like in different, so this is a game called Resilient Planet, which is part of the JASON Project, and students would watch a video of undersea exploration, and then they could go play a game about that. The information here for students with autism that I think is most relevant is that you've got the expert avatar again who's the scientist who provides modeling and feedback and then you also have the camera, which our students with ASD really liked. And in addition, you have a database that provides additional information about the game environment if they wanted to go and learn something that was out scope of what was happening in the game, if they wanted to learn the reason behind why it's behaving the way it is. This is a screenshot of a game, a series of screenshots, from a game called You Make Me Sick. And in this game, the students design their own pathogens, and then they infect a host. They get to see what the pathogen looks like as it travels down into the alveoli of an individual, and then they get to see division at the cellular level. And so in about 17 minutes of gameplay, a student experiences something that is traditionally a six week unit in a middle school science class. It's pretty incredible. The interesting point about this is that gives players a different perspective than they would normally have, and many of our students with ASD would look at the content materials, the traditional content materials and see them as black and white and not understand the dynamic nature of what an infection is, and through the gameplay, they were able to articulate that. So really quickly, I will go over what the teachers, teachers really appreciated the learning tools and the games. This is something that you need to look for if you're looking for video games to use with your students as a means to either supplement their instruction or provide their instruction. There are games out there now that are strategically aligned with next generation science standards. I think that's highly important, and the important piece here is that the games spurred unanticipated discussions among the students. And so, the teachers were able to prompt kids both with and without disabilities, so with and without ASD, to stop for a minute their gameplay and have a conversation about what was happening in the game and how it related to what they were learning in their content materials. And so this is a great opportunity of instruction to these students. And with few exceptions, the students said they would rather play video games than take a traditional paper and pencil test. There was one student, very interesting from our perspective, with ASD and it was a female, and she said that she would rather take a test. And we asked her why she would rather take a test than play a game, and she said, "Because it's not fair "for me to take the game; it," even though it actually had proven it, and she had demonstrated mastery of the content objectives. Because it wasn't a paper and pencil test, she didn't think she was actually demonstrating mastery of the content. And so that's an interesting perspective. I wish we were in-person so that I could hear all your thoughts about that, but that's the way it was. So there's a few things left, and so, one of the items I wanted to discuss with you was the process for implementing in your classroom, and I would say you should put on your ASD hat for a while and do an analysis of barriers that students might experience with the content and then make sure that the games are actually going to help the student to circumvent those barriers. So that's the first step, is the barrier analysis. Then you go and you select a game based on that barrier analysis. You play it yourself as a teacher. Then you prepare the students. They play it together in teams, and when I say in teams, you need to make sure that it's not competitive but that everyone has a role, and that they're working together. And then you have a group debrief. This is a great time to have to model social skills for each other, and then you have an extension activity where they go and actually talk to someone about what's happened in the game and how they think that that's going to lead to changes in their lives. One piece that's critical here if you're going to use games as assessment tools is that you recognize, there are two different types of proficiency within a game. There's the gameplay mechanics proficiency, which means if I want to move from one spot on the screen to another spot on the screen, do I know how to do that? And then the next piece is, if I know how to do that, I can use that information to measure the students' performance across certain benchmarks. So one of the keys to teaching with games effectively in STEM-related fields is to be able to test first whether the students actually can navigate the game in an effective way, and then you can go on to use those as assessments. On this slide, you'll see that this is a mock-up of what a student report would look like, and you can see on the left hand side, all of the digital supports that are included in the game, and you can see whether the student has used those or not. And then you get a proficiency score at the bottom of each, and so where it says M14 and M15, if you're not familiar with that, those are science standards, and then it gives you an overall proficiency as well as the amount of time that it takes in the game. And some of our folks with autism really appreciated having a certificate at the end of the game that showed what their performance was. With that, I'm going to conclude because what I've included on these slides are links to a number of STEM games, and then there's some tables and resources for you guys. And so these tables, I apologize that they're small. They're from an article that we published in teaching exceptional children about using STEM games for students with disabilities. And so we conducted an analysis of all these different video games, STEM video games. They're included for you in this table, so you can download this and blow it up so that you can look at it. And then on table two, we have an additional, apps that are included that you can use, so.

- [Amelia] Okay, so today I'm gonna talk about, we talked about software, and we talked about universal design for learning. And so I'm gonna discuss the use of humanoid robots, which is just another step in the ways technology can be used with students with autism. And I am associate professor in early childhood and special education at the University of North Carolina Wilmington, and I run the Center for Assistive Technology here. So as Matt and Jeff mentioned, there are unique needs of students with autism, and they include complications processing facial features, attention to individual details versus overall pictures of things, difficulty understanding intonation, emotions, and feelings, issues maintaining the joint attention and engagement, difficulties self-regulating behaviors, and we know there are communication deficits. But we know they love interacting with computers, and that is what we've been discussing so far. We wanted to know how can we take that enthusiasm that our students have for games and robotics and bridge them to humans, and so what people have come up with to bridge software and humans is humanoid robots. There's a lot of development research on these robots, and what we know is researchers worked to address those unique needs that I went over earlier. And so they wanted to offer fewer avenues of communication, so they could have a more targeted efforts for understanding and comprehension. They allowed students greater focus. They wanted to provide increased feedback during social interactions, and they also wanted to, interactions to represent human characteristics, so we could try to bridge that gap again. And then they wanted to reduce facial features because they know that that was an issue for processing. So this is an example of a NAO robot on the screen, and some of the features include limited eye movement, so one of the ways that they've removed a lot of the complex features, and it has gentle movements of the eyelids to indicate emotions. It has human traits like arms, legs, hands. It has hands that move. It has tactile sensors for those with non-verbal communication. It has fewer, or for non-verbal communication. There are fewer simpler facial features that you'll notice. There are no eyelashes or eyebrows to reduce those complex features, and they deliver simple and straightforward communication. Here are two more examples that have been developed, and they do generally have internet accessibility and video features to record children's responses. They imitate human reactions. You can have them dance and wave. You might have seen robots on the internet doing the evolution of dance for you. So students can get engaged and involved. They also talk and listen, so there's voice recognition software included in many of the models, and they sense when children start to get frustrated in some models and react accordingly. So there's work to assist in modulating emotions for those students. And robots right now are ranging from about four to $500 up to $20,000 or so. So and probably up from there depending on the size and functioning. So research on the effectiveness of the design of the robots showed that some of the ideas they developed. This all research started in about 2002, and they kind of researched what's working, what's not for students. And so what they've come up with is they think it increases intrinsic motivation when using robotics during the intervention and offered high levels of engagement, and this is very consistent with what the other presenters were saying, that there's just lots of engagement for students, and the technology is very engaging for students. They did bridge a gap that they felt between software and humans. They thought it provided interesting visual displays, and they responded to children's behavior. In some therapies, if a child would get upset, for example, it would adapt what it was going to do next based on that frustration. And then, it promotes social behaviors is what the research has shown us. So here at UNCW, we did an exploratory study using humanoid robots to look at engagement and communicative initiation when compared to using software. We did a multiple baseline design. We had five participants. They were ages seven to nine. I will say that most of the research being done right now is on elementary age students, and so I think what is being done for younger children can evolve into some of the other students that we're talking about that need more development in STEM to be successful in their careers. Two subjects were lost due to attrition, so we had one African American and two Caucasian males. And the activities on their tablets included playing simple games such as Angry Birds and Mario, and on the human robotics side, we did simple games like Simon Says that were interactive but simple. We used two measures on engagement and communicative initiation and offered some visuals because the students, we did this at a after school program for students, high-functioning students with autism. They required visuals to process information. They were very highly verbal. However, they had difficulties with their social responses and reactions and that behavior. An adult mediated both of the games so that we had consistency, and interactions were scripted. What we saw if you look at these results is that when it came to persistence, it was high across the board with either, you know, iPad use or robotics use. But when we looked at communication, there was very little communication going on. And this is, we did not have online gaming, so I think that brings an element of communication into these games. But we did not have it, and so there was very little communication. They were usually in small, they were actually always in small groups when they were doing these activities. And so they might say, look, as a minimal communication, but really, there was a whole lot of nothing going on until we moved into the intervention phase with the robotics, and you can just see the communication increase across the board, which I thought was interesting. So just to follow-up, we started to consider robots, my students started to consider the robots as part of their community as we went on. And these are younger children, and so they really enjoyed dressing up and acting out performances. And they started to dress up the robots as you see on the right, so they really did, you know, accept them into their community as people to some extent. They expressed interest in programming after that. They wanted to see if they could make the robot do something different, and so we started to move towards simple programming with the students. And then we expanded out into camps and after school programs focusing on STEM education, and we did have a focus of robotics in there, and that included LEGO robots and humanoid robots. So I think there is potential for growth. Some additional research on the benefits of humanoid robots have shown that they elicit prosocial behaviors. They increase eye gaze and joint attention. They support therapies that improve social interactions, and generally speaking, that was communication and social interventions, which Jeff spoke about. Enhancing communication with adult partners as compared to software, which is really the key purpose of some of this development, and supporting social communication. There really did, in my experience, encourage interactions, and other researchers have found that as well. And interestingly enough, if there was an interaction with a human provider with a robot partner, the communication and interactions increased. So just the presence of the humanoid robot in one study seemed to show a difference in their willingness to just, to talk and interact with adults. So it attracts their attention for sure. I think some challenges and future research ideas would be, challenges are the the expense of robots today. They're gonna come down and have progressively been coming down, but that would be an issue. Ease of implementation would be something to think about. Programming requirements, most of these humanoid robots are fairly easy to program, and they're getting easier as they go because they use icons. And then there's limited research across age group and skills. Like I said, most of the research concentrated on younger children, mostly elementary. So I think what we have to ask ourselves is what ages are best to respond to the use of humanoid robots. Considering the minimal skills that children require for their effective use, really the key question is, are the skills generalizable to humans? Because if that is the goal and the approach isn't working, and so that's some research that I think needs to be done. It's not very clear. And we should do it across settings and across ages to build that accumulation of research. And then lastly, how will the evolution of those robots influence research in the future because they're continuously enhancing and making those humanoid robots much more autism-friendly. So that is all I have to talk about right now, and there's some references in there just in case.

- [Russell] Okay, great. Well, thank you very much, Amelia, especially for popping up early. I believe Maya is back online now.

- [Maya] I am, thank you,

- And hopefully.

- [Maya] sorry about that.

- [Russell] Not a problem.

- [Maya] All right, I'm gonna see whether I can, yup, let's see here. Oh and look at that; it all works, so for the moment, we're all good. Sorry and thank you for the

- Hooray!

- [Maya] patience so, I know, I know, I'm a just keep our fingers crossed. All right, so I'm gonna talk to you a little bit about Computer Science for All just because I'm assuming that some of the folks on this call are not as familiar with computer science education, and then quickly, I'll transition into what are some challenges that students with disabilities face within this context? I'll talk a little bit about students on the autism spectrum, but really the focus of my work is on kids at risk for academics failure, kids with learning disabilities, and kind of a broad range of students who are challenged. So I don't have the specific focus on students with autism, but we just finished a case study with three students on the autism spectrum, so I can certainly talk to that. I want to talk about methodology for studying computer programming and computation with students with disabilities, and then I'm gonna share four strategies that we are studying right now that are showing promise for including kids with disability in computer science education. So the big aha for me because I'm a, my background is in special education, and my work has been around technology to support kids with disability, is that when we think about computer science, we really aren't thinking about kind of a narrow view of programming and tech companies. Essentially, computer science is everywhere. Really, only about half of computing jobs are in tech companies. We're seeing computing across areas such as architecture, gaming that Matt spoke about, medicine, movies, education, advertising, and so because of that, computer science is being considered, it's starting to be considered as a foundational content area for all students. So when we think about STEM, science, technology, engineering, and mathematics, computing is part of the T for sure, and it's really about how do we empower children to not just consume technology and games and apps, how do we help them and empower them to be creators of those technologies for themselves? So there's a lot of statistics out there about how there are gonna be more computing jobs available than there are gonna be people to fill them, but really for us, it's beyond jobs. So thinking about who has access to computer science education from an equity perspective, traditionally, when we think about computer science in schools, it's been seen as an enrichment activity for students who are academically advanced, or it's after school or camp activities for students who are financially privileged, and so opening up computing to all children is an equity issue. Beyond that, the idea is that you can teach a lot of real world problem solving. You can teach persistence, collaboration, applied mathematics through computing. One of my unassessed projects is looking at integrating computing and mathematics, and so we're seeing a lot of promise for teaching mathematics through computing as well. So with all that, in the United States, computer science education is not new, but over the past year, there has been a flurry of new activity starting with January, where the President announced an initiative to bring computer science to schools across the nation. Recently, 26 governors across the country urged Congress to fund computer science education along with CEOs from both tech and non-tech industries, completely bipartisan. Both the National Science Foundation and the U.S. Department of Education have put in investments, financial investments to move computer science education to the next level. Every week, I'm hearing about more districts and states including computer science, and there's a new framework for K-12 computer science that's really looking at what the concept and practices that we want all students to be able to experience. The problem is among all of those is that there's very little discourse about where students with disabilities fit in. I think that there is a huge movement to include more students in CS, in computer science education, but the idea of how do we really meaningfully include students with disabilities is part of the hashtag csforall. It's still something that we need to advocate for. So my research is looking at the challenges that students with disabilities face, coming up with new methodologies, and actually the good news is that what we're seeing is that there are strategies that support kids, even those who are really disengaged on the surface from computing. This slide just kind of shows that when we're thinking about challenges the students face, they're both instructional challenges, and there are challenges inherent to computing. So the instructional challenges are things such as inaccessible curricula and software, so for example, if the software requires a lot of reading or doesn't offer opportunities to practice. If a student has a visual impairment and we're just using visual block-based programming, that software, that curricula is disabling. We're also seeing challenges around low expectations by teachers. Because computer science education is fairly new to most teachers, and they're not confident in their own pedagogical approaches, when a student is struggling and that student has a disability, there are often assumed that that disability is the cause for the struggle and not necessarily the curriculum or the pedagogical approaches and lack of instructional strategies specific to computer science education. So those are the instructional pieces of it, and then there are challenges specific to computing. So if we think about what happens if somebody sits down to do computer programming, often, they're trying to design something. And so it's an ill-defined problem, and it's not something that has a lot of prescribed steps that are given for the student. It requires multi-step, complex problem solving and a lot of open exploration. On the one hand, that's really exciting for students and motivating and fun. On the other end, because often students with disabilities struggle in these areas, we're setting them up for failure from the very beginning. And so what we do with our research is that we start with instructional strategies that we know work in other content areas. So for example, thinking about integrating some explicit instruction into these open activities is something that we know that there's a long body of literature around explicit instruction around universal design for learning. And so we're starting with what we already know from decades of research around what works for students who are struggling and students with disabilities. We also, all of our research is in classrooms that are highly diverse so that we can see whether what we're learning can be translated to other classrooms. And we're constantly working with teachers and administrators to make sure that the questions that we're asking are actually relevant to teachers and administrators. The programming environments, there are two types of programming environments. These two that I'm gonna talk about are visual, but we're, certainly, these aren't the only ones we look at. But we're using visually intuitive block-based programming for this particular study that I'm gonna talk about now, and essentially for this, students don't need to know the text-based languages. They only need to, they can still learn the concepts because those are programmed within the blocks, and so students drag and drop a pile of code and learn programming logic and problem solving without having to worry about the syntax in the programs such as Java or Python. Scratch, which is the one on the left, is one of the best known block-based programming environments. It's highly graphical. It allows users to create and manipulate two-dimensional sprites, you know, add music, animation, other kinds of interactivity, and it's also meant to be real collaborative. Code.org is on the right. It's also a block-based programming environment, but is has a more structured teaching approach that eventually leads to more open-ended activities, but it leads them through a more structured, puzzle-based environment that teach explicitly some of the computing concepts and processes. So I'm gonna talk a little bit about measurement, and I'm gonna try not to get stuck here, but I want to show you kind of the power of what we do and how we look at students in computing beyond just looking at the products that they are developing. We're looking at the process. We developed an instrument called the Collaborative Computing Observation Instrument, and we're using screen capture software, so we're able to see what the students are doing on the screen as well as listen to their conversation and then we code what they do, who they do it with, and with whom within this instrument. And the kind of research questions that we can ask are things like, how are the students asking for help or are they working individually? How long are they persisting? What kind of support are they getting? Is that resulting in any kind of success for them? And so then what ends up happening is we end up with a sequence of codes that start with when a student begins to work on any kind of task and ends when they either complete that task or they abandon that task. And then what we're able to do is take those codes and create a visual representation through a directed graph. On the right here is a real simple one where a student works with a teacher's aide, and this was a student with an autism spectrum disorder, and he was working and getting really frustrated, and so the teacher's aide who actually didn't have a lot of expertise in computing tried to help him when he was struggling, and he basically ignored her and eventually abandoned the task. It's not a very pretty picture, but it's one that we see fairly regularly. I'll show you a more complex one in a moment when we talk about collaborative problem solving. So the four instructional practices that we look at are collaborative problem solving, metacognitive self-regulation, and that's particular to when a student is stuck. What do I do when I don't know why my code isn't working? Universal design for learning, which is more classwide-based instructional approach and then balancing explicit instruction with open inquiry. So with collaborative problem solving, what the literature shows is that we really need to teach children how to collaborate and especially when we're talking about students with ASD and other disabilities where communication is a problem anyway, explicitly providing them the tools for collaboration is really important, modeling that, and then we're using a framework called a collaborative discussion framework, where essentially we're teaching students a conversation script. So often students with disabilities, when they get stuck and they've already experienced a bunch of failures, they're gonna almost insist that the teacher or their peer give them the answer. And so to combat that, what we've done is we've taught the students this script, so when a student says, "I need help," or "I've tried to do something; it's not working," the peer is supposed to say things like, "What are you trying to do?" or "What have you tried already? "What would happen if?" And then if students are not as verbal, the peers will say things like, "I notice that you did this. "I wonder what would happen if," dot, dot, dot. And so this has been very helpful for students, but I'm gonna show you in a moment that it's really not enough. I'm gonna share the story of a young man. We're gonna call him Bradley. He's got an autism spectrum disorder, and for him, it was very common to have very low persistence. So almost immediately, as soon as he started programming anything, he would immediately stop, either wait for somebody to offer him help or ask for help. So here is a directed graph of 11 minutes of Bradley's interactions, and I'll talk you through this, not so much so that you completely under the directed graph but just so that we can highlight some of the communication collaboration issues that students such as Bradley face. So this is the directed graph from the Collaborative Computing Observation Instrument, and essentially with this 11 minutes, Bradley tries to have six different interactions with his peers, with teachers, either with him asking for help or with a student offering help without his wanting to do so. He asked a peer for help twice. That's that 1A piece of it. But when he did, he didn't ask for help very explicitly. So when we say students ask for help, we then look at the level of adaptive help-seeking. Are they asking explicitly? Are they just expressing frustration and hoping somebody comes to them and offers help? So Bradley would say things like, "Oh, I need help," or "This isn't working," or "No one is helping me," and then somebody would come and help him. But because of that, his peers really didn't have a lot of guidance in how to help him because they didn't know where he was stuck. He did ask his teacher for help one time, and then three times, the peers just noticed that he was frustrated. So when we're watching the screen capture data, we're able to see that Bradley is just having a difficult time, and so the peers would just come to him and say, "What can I do to help you?" Here's a little conversation, so a student came with a collaborative discussion framework and says, "What are you trying to do?" And Bradley says, "Yes, I'm trying to go up, "but you put a circle?" And then the peer says, "What have you tried already?" He goes, "Yeah, that worked, it worked." And the peer kind of goes, "Really? Let me see." And it didn't work actually. So the peer then tries to engage with him some more, and Bradley just gets really frustrated and says, "Oh gosh, I can't do this." And so what we saw is that just the collaborative discussion framework wasn't enough for Bradley. These interactions did not result in any kind of real problem solving, and eventually, he abandoned the task. The 15B code essentially means that he ended up, as soon as one collaborative interaction ended, he immediately went to the another one, and he never actually did end up successfully solving the problem he was working on. And so the next piece that we went to is how do we get Bradley and students like him to think about how to debug, how to problem solve so that then when they go to a peer, they have some more tools. They both stopped and thought a little bit about what they were doing. So we're teaching them explicitly how to problem solve. We can't just say to Bradley or his peers, you know, try to debug your code. We've created some checklists with steps such as, okay, did you read your code block-by-block? Does any part of your code work? And they're just checking it off, saying, yes, they did this, or, no, I didn't. Are there pieces here that already worked that you don't need to break apart? Because what we're seeing is the students are stuck. They don't have any real strategies for problem solving, so they'll completely break apart their code and try again and break it apart and not really have a systematic way of doing. So this is a strategy that's starting to show some promise, especially if they take this tool and work with their peers with something in hand. The universal design for learning Matt spoke about just a little bit, but this is a schoolwide, classwide intervention where we want teachers to create an environment that support students like Bradley and his peers more proactively. So for example, for multiple means of representation, we want to make sure that the teacher is modeling the computing lesson, creating templates, using video tutorials. And then for multiple means of action and expression, we hope that the teacher either offers, you know, either a partially completed code or offers a template or offers some kind of way for a student to participate even if they are stuck. And for multiple means of engagement, can they offer them choices in what they're programming. Can they consider the barriers to learning from the get-go and then create a learning environment that is more universally designed? When we work with teachers, what's important, you know, at least with computing is because it's such an emerging pedagogy, what we tell teachers is, think about bringing in some elements of universal design for learning into instruction. Evaluate for engagement and learning, and then we know that for some students such as Bradley, we're gonna have to layer on top of that his individualized support, and what our research is showing is that those supports that work in other content areas typically work within this environment as well. So for example, since Bradley needed some verbal prompting in math class, in english class, he would also need some verbal prompting during computer science instruction. And typically, that seems to work. So for example, if a student has a behavior plan, and that behavior plan is needed to keep that child on task in other content areas, that same behavior plan would probably work within this environment. Or if a student needs a one-on-one aide to support, you know, just general problem solving, that's gonna probably be needed in this environment too. And it's really helpful to teachers to know that when we tell them, you know, you know your students. You know about good intervention, and so just think about applying those within the context of CS and then evaluate for engagement. The very last thing I'm gonna talk about is balancing explicit instruction with open inquiry, and this is kind of a dichotomy, right, because explicit instruction, we know, can reduce barriers, can reduce frustration for students with disabilities because it's providing the steps of a process. It is explaining details, and the teachers have been able to monitor student progress and give feedback. On then the other hand, we just, as a computing is very open-ended, so how do we actually balance these two approaches is challenging. We don't want to give too much explicit instruction so that the students will never learn to problem solve on their own, but we don't want to give too little so that they're experiencing too much frustration and failure, and they're giving up. So here's a quick example. I'm gonna talk a little bit about a teacher who wanted to teach students about conditionals, just providing step-by-step demonstration, breaking down the paths, modeling those, and then providing lots and lots of examples. We used Archer and Hughes' 2011 Explicit Instruction: Effective and Efficient Teaching, so those 11 explicit instruction models. And so for example, with conditionals, which are essentially logic statements that determine program flow, so if a condition is true, you do one thing. If a condition is false, you do another thing, and that's kind of a simplification. And so what this particular teacher did is she created a worksheet of different conditionals, and the students would say, "If I do my homework at night, "then I'll get to play outside." or "If I clean up my room, "then I'll get my allowance," or whatever it is. There were even, by the water fount, there was this picture of, if I'm thirsty, drink for less than three seconds. Then go back to the end of the line, which cracked me up. I just thought it was wonderful. But it's really after learning how to do this in a concrete way that you move to the abstract. So those are the four different strategies that we're working on. Just like Matt, I put some resources in the back. This is my email address. Everyday Computing is one of our unassessed projects. We've got another one starting up, really looking at these strategies. This is a video to, it's a link to a video that we created for the STEM for All Showcase for the NSF. It's two minutes that provides a little more information, and here are some references. So that is that, thank you.

- [Russell] Well thank you, Amelia, and the rest of our panel. I think this really starts a nice discussion about the use of technology in STEM education for autistic students or students with autism and also, the whole issue of computer science and how to teach computer science and robotics, With that, I'd like to actually open it up for questions from our audience for the remainder of the time. So you can either ask questions, I guess, verbally across the phone line here, or type them in, and we'll read them out loud and reply to you. And while we wait for our first question, do we have any other comments from the other panelists? So you've all had an opportunity to listen to each other. Do you have any additional comments?

- [Jeff] This is Jeff. I just appreciate, you know, hearing everyone's subsequent talk, how, you know, focusing in on kids with disabilities and kids with autism in particular, just highlights the incredible detail of the learning process itself that unfolds so naturally for most kids. And with this group, we're just having to, you know, focus that natural process that's not happening and augment every little part and just be testing and trying each little component. And, you know, it's really encouraging for me to see everything that's being done.

- [Russell] Okay. Well, I have a general question from Jane Brown. She says, "I work with students with autism "in colleges and universities. "Do you have any suggestions for technology "besides what we usually use, "such as smartpens, voice input, et cetera?"

- [Matt] So Russ, this is Matt, and if you go to projectican.net, that's one of the projects that I'm principal investigator of. It has a whole list of resources for college students with different types of executive function disorders, and you can probably find some resources in that website.

- [Russell] Great.

- [Maya] And I think what I would add to that too is there's been a lot of work about universal design for learning in higher education, so looking at some of that work as well in terms of how can I proactively address that variability. How can I provide information in different ways and allow students to express their understanding? My area isn't higher education, but I certainly, I mean, we could certainly put up some of those resources too.

- [Russell] Okay. Well, I'll ask a general question now too. So we've actually had a fairly wide-ranging discussion today. If we could wave a magic wand and create some research programs to answer things we don't know, what would be the major things we should be looking at?

- [Matt] This is Matt. I would jump in and say that we really need to do some research to determine whether we can use, and of course this has got a games plan to it, right, because of the research that I do. But I think that we could do virtual training programs for students with autism, and you and I discussed this, Russ, before everybody else came on. But I think there needs to be some specific research looking at what aspects of career preparation can lead to effective transition plans and then effective transition outcomes for students with autism.

- [Russell] Mm-hmm.

- [Maya] Yes, I also would add to that that's so, assessment is such a big thing, so now with the ESSA, universal design for learning being part of the law, both from the technology perspective as well as from an assessment perspective, how do we design assessments that really measure learning without that paper and pencil task? So these technologies, you know, Jeff spoke to this at the beginning and Matt, I mean, I think all of us spoke to this, are able to show what kids know in a way that doesn't require them to write or doesn't require them to demonstrate understanding in a way that is a traditional way. So thinking about assessments and learning and technology-mediated solutions is another area that I think would be really helpful.

- [Russell] So let me, let's see, do we have another question? Still waiting for another question, so I'll ask it. So if we have any software developers, say, on the call here today, do you think we have enough information to help them design software applications better for students with autism? And what other information do you think we really need to be looking at that we really don't know a lot about right now?

- [Russell] Well, maybe that's the wrong question.

- [Jeff] Well, no, I think it's a good question, but this is Jeff. But I would just want to echo what Maya said. And perhaps, one thing I know I want to leverage is the expertise of teachers and clinicians who work with these populations and understanding what they do that's effective and successful and giving them tools that maximize their own impact. So to some degree, we don't need to build, as we're building the software to be smart, we need to use software to capture what these experts are doing implicitly. So that's one angle because a lot of the, I think the Bradley example was a great one for how it was just so challenging for him even as people were trying to help him. So assessment is so crucial. In some ways, we gotta do that before we even try to solve the problem.

- [Maya] Yeah, yeah, and part of what I'm thinking with the C-COI and other instruments is, can we somehow cue the teacher that a student is struggling because he would just get more and more frustrated. There are other students who would just sit there and actually persist unsuccessfully and not want go and offer, you know, ask for help. And unless there's an adult right next to them, you wouldn't know that that child is just sitting there because, you know, it's a technology-mediated environment, so students are in front of their computers, right, so it's very difficult as a teacher to know what's happening as students are working independently

- Right.

- [Maya] and collaboratively.

- [Russell] Yes,

- Maya, I think that's

- [Russell] across the whole student populations, really.

- [Maya] Correct.

- [Russell] Yeah, well, let's move on; we do have a,

- I have a question.

- [Russell] Oh, go ahead.

- [Jackie] Russ, this is Jackie Peth. First, let me thank all the presenters for all that great information. I'm wondering, we've covered a lot about the teacher role and the peer role and student role. I'm wondering about the parent role. Parents of students with ASD often have to or want to, but they wind up getting particularly involved in supporting the students at a lot of levels. Do any of you have thoughts about recommendations for how the parents can best support the student learning.

- [Amelia] This is, can you hear me; this is Amelia.

- [Maya] Yes.

- [Amelia] I would say

- Yes

- [Amelia] that parents play an integral part in this because you are experts on your children, and I think that you are aware of evidence-based strategies that work for your children. Is it visuals? Is it prompts? So how can we consider what's working for your student and translate that to educators in the school, so we make sure that the interventions match what they are getting in school.

- [Maya] Yeah, I would agree, and I would also add to that, we work a lot with students who come from low socioeconomic backgrounds, so it's doubly challenging with technology. So having opportunities for parents and students to work together and to have, you know, after school programs and tech time type of experiences where the technology could be made less scary and less other is something that is helpful. I mean, there's a whole lot of issues with how do you get students and parents to come after hours. But with some creativity, it can be successful because, you know, especially with new technologies, it's very intimidating, especially to families who are from lower socioeconomic homes.

- [Jeff] And I appreciate the question and agree it's such an important piece, any parent able to help us understand what doesn't work and the implications of trying to incorporate technology, you know, at home when the power button goes off or, you know, whatever the constraint is because it's a battle. I mean, probably many of us as parents feel this tension, this push, pull with technology's role in our kids' lives, period, and I think that challenge can even be more complicated with kids with ASD. So that's not a, there's no solution there, but I think parents, we need to know what they want solved, you know, because often technology can fill a role of just occupying a child in a predictable, safe way while the parent has to take a shower or cook a meal or something. And so, you know, there's really some really tangible challenges that thinking creatively about technology can help us. And designing without much thought can exacerbate a problem potentially like when the power goes off, and the kid melts down every time, even though the app they're working with is a real benefit.

- [Amelia] Jeff, that's a really good point because transitions away from technology often cause issues for parents and teachers.

- [Russell] Yes, as a parent of two autistic students, I can vouch for that.

- [Russell] So.

- [Jeff] It's hard for me.

- Could we,

- [Russell] Well, do we have, do we have any other questions coming up here? We've got Linda Tsantis who asked something about virtual coaching for teachers. You see any roles for helping create tools to help teachers adapt?

- [Matt] Russ, this is Matt. So we're doing this project at UCF called iCan, and what we're doing is pairing the folks who are, they're certified special education teachers, and they take these secondary curriculum class, methods class with me, and we pair them virtually with college students who are in STEM majors who have registered with disability services and have an executive function disorder. And what happens is they record their interactions with the college students, and they act as coaches, and then we coach them on how they coach the college students. So it's a really interesting dynamic. By using Adobe Connect, we're able to video capture that interaction, and then, we are able to review it with the secondary teacher and provide them with some feedback in a timely manner about the explicit strategies they're using to help the students. And then at the conclusion of the semester, they backwards map that, so they say, "Okay, if I know that in college, "this student might struggle in this area, "I can provide compensatory strategies "at the secondary level to prepare them to be successful "when they go on to post-secondary school." So it's an interesting approach.

- [Russell] Yeah, and I think we are actually at our time limit now. So I would like to thank all of our panelists and everyone who's participated. This has been some really nice active discussions on the sidebar. And please finish or fill out our survey and questionnaire here at the end that's posted so that we can really have feedback on how to make these events better in the future. And hopefully, this is the beginning of a conversation. We really ask for your help in pushing this particular area forward. So thank you very much for your time.

- [Amelia] Thank you, everybody.

- [Maya] Thank you.

- [Matt] Thank you.

- [Jeff] Thank you.