During my almost three-year journey leading an HCI project aimed at developing non-traditional user interfaces for children with cerebral palsy, I learned something critical: All the specialized literature on HCI and disability I had read had not prepared me for the extent and level of the challenges I faced over the course of the project.
Not only did those challenges give me insights into how to plan and coordinate a project involving HCI research in the field of disability, but they also changed my view about the phenomenon of disability, revealing why it is critical for scientists and engineers providing technological solutions to be aware of the coexisting—and potentially contradicting—models of disability.
This article is not about the methodological challenges of carrying out an HCI research project on disability—there is enough literature of this kind covering different layers of the research process. Rather, it is about how these challenges allowed me to draft a taxonomy for describing the solution-discovery process.
To present these thoughts, I will walk the reader through the setbacks and achievements of the solution-discovery process I experienced.
The idea for project NEXO—a euphonious acronym for New learning Environments for the multiplication (i.e., "X") of Opportunities—was born after an informal conversation I had in 2009 with the principal of a public school for children with physical disabilities in Uruguay. I had been approaching organizations looking for a research topic involving the use of non-traditional user interfaces in the context of disability.
The school, officially named Public School No. 200 "Dr. Ricardo Caritat" (School #200 for short), is unique for two reasons. First, it is the only public school in the country that educates children with a broad range of motor and cognitive dysfunctions, mainly cerebral palsy (90 percent of students), spina bifida (8 percent), and other motor-related pathologies (2 percent). Uruguay is a relatively small developing country with around 3.5 million people. While there are other private institutions and NGOs educating and taking care of children with physical disabilities, School #200 is the only state-funded one. Second, rather than being a school for "special needs," it is ruled by the national curriculum for public schools.
Since Uruguay adhered to the One Laptop per Child (OLPC) program, every child attending the school received an OLPC XO laptop from the government, as did children attending every other public school. But according to the school's principal, rather than being a benefit, the laptops were a burden: The design of the XO and most of the applications shipped with it weren't accessible to half the school's population.
Beyond the impact on the children's self-esteem, this problem was also hindering opportunities for introducing the XO into the classroom and exploring ways to use it as a tool for stimulation and rehabilitation. Appealing to the government for more accessible computers, the principal told me, wasn't a realistic option; a positive response would have taken—optimistically—years.
When the principal described the context to me, I knew I'd found the research topic I was looking for. We both agreed it was a good opportunity for me to study whether it was feasible to build a prototype of a user interface circumventing the XO's traditional interaction schema (keyboard, trackpad, screen) to reduce access barriers.
The one-hour introductory exchange between me and the principal triggered a chain of events leading, ultimately, to the project kick-off 18 months later. Then followed a sequence of actions, including defining the project scope, applying for several funding options, finding a partner and a team, a lot of planning, going through the ethics review process, carrying out a series of observations at the school—and, naturally, dealing with the unavoidable academic bureaucracy.
The circumstances at the beginning of NEXO have a characteristic shared by most of the engineering projects I have been involved in. I call this commonality the problem-solution development phase, or PS phase. Projects in the PS phase start with a problem description P, identified by a person or a group of people, and which is taken as a given. Then, the engineering team (in a broad sense, including developers and programmers) applies certain methodologies to find a suitable solution S that fits P. S belongs to the landscape of possible solutions to P, some of which are close to maximizing user satisfaction and some of which are doomed to failure. (Some say the reason why certain projects fail is that engineers tend to start from S and then try to find a suitable P fitting S. For the sake of simplicity, let's not consider that option.)
Back to NEXO, the goals of the first stage of the project (Stage One) were contingent on the initial problem definition. Our engineering team created a prototype implementing a new interaction schema for the XO. The main goal: to assess the potential of the prototype as an accessibility enabler.
The prototype worked like this: 1) the user places a card containing a particular drawing—e.g., a filled-in outline of an elephant—in front of the XO's embedded camera; 2) an algorithm running in the background recognizes the drawing, using a video-tracking technique to identify fiducial markers in real time; 3) the algorithm triggers an operating system event simulating keystrokes or mouse events. Such an ingenious mechanism not only made it possible to develop new applications based on the new interaction schema, but also to integrate the prototype with existing applications out of the box.
The prototype was based on the working hypothesis—supported to a certain extent by prior facts, experiences, and observations—that children diagnosed with motor dysfunctions would benefit from an interaction mechanism of this kind.
However, by the end of Stage One, the results were discouraging. The prototype worked quite well for children who were already able to use the XO, but, regardless of the infrastructure used to assist children in holding the cards in front of the camera, children with severe motor impairments were still unable to access their computers.
|NEXO's research team testing a non-traditional user interface aimed at reducing access barriers.|
Using the terminology above, one explanation for this drawback would be that, starting from P, we had chosen the wrong path and wound up with a suboptimal S. What else could it be?
It turned out there was indeed another explanation: Maybe the issue was the problem itself. During the course of Stage One, we overlooked one particular characteristic of the population that we ended up discovering when we analyzed the results and performed the post-mortem assessment.
Whereas we were ultimately interested in helping children with severe motor dysfunction, we had to content ourselves with an initial selection of participants with mild symptoms. The reason was that we needed to work with children who could understand directions from the researchers. Therefore, since we couldn't wait until we had finished running the psychometric tests over the entire school population to start building the prototype, we asked the teachers to select, among the children with low learning difficulties, the ones with the highest motor impairments. However, this criteria brought only children with mild motor impairments in absolute terms. Apparently, there was a correlation, confirmed by the experience of the school staff, between low-motor and perceived low-cognitive skills.
It turned out there was another explanation: Maybe the issue was the problem itself.
This pattern could be attributed to different factors. It could be a random particularity of the school's population at the time. As most of the children attending the school come from low-income families that can't afford treatements or special care, it could result from poor stimulation during early childhood. Moreover, it could be related to a general characteristic of populations with cerebral palsy, for whom motor coordination difficulties during early childhood may diminish perceptual, social, and cognitive abilities . Or it could be an intrinsic problem of using psychometric tests for evaluating children with cerebral palsy. There's evidence that commonly used standardized methods for assessing the intelligence of children with cerebral palsy might not be reliable . Finally, we could have committed procedural errors during the experiment design and/or execution. However, the school staff, including teachers, physiotherapist, and psychologist, told us our observations were consistent with their experience.
In other words, if it were true that half the school population lacked a certain degree of understanding and learning abilities, then half of the XOs would have always remained inaccessible to their owners no matter what technological bridge we could have built. That's because P was, to some extent, ill posed.
Given this scenario, we regrouped and decided to gather more observations from the environment in pursuit of new needs or problems that might have been affecting the children and that remained hidden until then.
While the first stage of NEXO focused on the immediate accessibility problems caused by the XO, the second stage (Stage Two) expanded the locus of action to address the classes of problems producing lack of participation. The development process and the roles of stakeholders were reformulated, shifting attention toward the children as guiding agents of design decisions.
I call this the EPS phase, where E stands for environment, of engineering projects. EPS projects base the problem definition on observable evidence rather than using interpretations from intermediaries. However, in the current context it was still too early to call E into question—in NEXO's case, E would be School #200.
During Stage Two we observed that children's ability to move around and explore their surroundings was limited not only by motor impairments. Some children in wheelchairs also seemed to be afraid of moving across the hallways. According to the psychologists, this was due to an excess of anxiety related to different causes, including growing up with overprotective parents and experiencing discrimination.
These observations led to a new class of problems, producing a rich variety of solutions, including a mechanism to turn bananas, or almost any object, into assistive push-buttons; a game to encourage children to move around the school's hallways and discover new places; and a new communication schema allowing children to write virtual messages and "upload" pictures to fiducial markers hanging on the school's walls, which could be read and later overwritten by other schoolchildren.
By the end of Stage Two, the project went through its last evolution. Up until then, we didn't have enough contextual information to call into question the school itself. However, after two years in the field, we started posing some existential questions about the environment. For example, School #200 was a one-size-fits-all institution, receiving children from all over the city. The school as a solution may have been perceived as great in the early '90s, but no current inclusive disability policy would recommend educating disabled children too far away from their communities. This particular characteristic of the school generated a series of shortcomings. For example, it was impossible for the children's parents to socialize with each other. It was also very difficult for them to engage in activities proposed by the school.
Once we were able to see this, we could articulate the right questions, unveiling a layer of untold problems. I call this the OEPS phase of engineering projects, with the O standing for out-there, accounting for a sort of external reality that exists prior to the filter of human interpretation. Recognizing that E—School #200—was one of many possible interpretations of the unknowable out-there, allowed us to bring new problems to the table and envision a newer class of solutions. For example, instead of focusing on the accessibility issues of the XO, we could focus on building a social network to foster communication between parents from distant neighborhoods.
It would have been impossible for us to skip one of the intermediary phases before arriving at the OEPS phase, but with hindsight we could have done better. We would have probably saved time and effort if we had known about this taxonomy. For example, we would have given more importance to the school's origins, historical background, or financial situation. We would have compared it to other schools addressing the same issues in other countries, and maybe reached out to them asking for advice. Being aware that we were gaining knowledge and experience to arrive at the OEPS phase would have helped us think further outside the box.
Our techniques, methodologies, language, and technologies are underpinned by traditions, conventions, and an institutional framework that brings a particular symbolic interpretation of the context being analyzed. As Paul Ricoeur says, "We cannot work without bringing in our traditions and our symbolic interpretation of the world" .
Knowing that the environment itself is an interpretation and that our praxis as solution providers cannot be separated from ideology doesn't make us immune from failure on the way from O to S. However, it keeps us alert to new opportunities for finding meaningful solutions to real issues and needs.
1. Leonard, H.C. The impact of poor motor skills on perceptual, social and cognitive development: The case of developmental coordination disorder. Frontiers in Psychology 7, 311 (2016); http://doi.org/10.3389/fpsyg.2016.00311
Gusa Armagno is a software engineer with more than 15 years of experience in the software industry. He has been involved in social-impact projects ranging from open data to creating and developing non-traditional user interfaces for children with physical disabilities. He holds a master's degree in HC and is currently pursuing a Ph.D. in the field. firstname.lastname@example.org
©2018 ACM 1072-5520/18/07 $15.00
Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. To copy otherwise, to republish, to post on servers or to redistribute to lists, requires prior specific permission and/or a fee.
The Digital Library is published by the Association for Computing Machinery. Copyright © 2018 ACM, Inc.