If there is one thing that 20 years of conducting research in child-computer interaction has taught me, it's that every time I work with children, something I've never even anticipated will happen.
Sometimes I face new challenges in trying to get children to follow a specific study protocol, or I find that a new understanding challenges my thoughts about the role that technology can have in supporting children's development. But in my recent program of research working with children living in poverty in Nepal, many things occurred that challenged my thoughts about the ethics of doing research with vulnerable populations. I went to Nepal to attend the Advances in Computing Entertainment (ACE) conference. There, along with one of my graduate students, I led a workshop to introduce design thinking to Nepali children at a private school in Kathmandu . After the conference, I went to the ecotourism town of Pokhara to vacation for a few days. Which is when I began working with children who live in poverty—instead of going on vacation!
There's a stark juxtaposition of Westerners trekking and paragliding in the Himalayas with the reality of young children who are too poor to go to free schools, who are beaten daily, who do not have enough to eat, and who are sold into the sex trade for only a few dollars. This contrast made it very difficult for me to do anything but try to figure out a way to help some of these children. But as a child-computer interaction researcher and designer, I have a fairly limited skill set. I know how to design and evaluate new technologies for children—and that's really all I can offer.
When an opportunity came up to work with children living in poverty who attended a non-profit school (http://www.nepalhousesociety.org), I leapt. Any donations I might make would be unlikely to significantly benefit these children due to agency overhead and political corruption. Instead, I worked with a team from Canada and Nepal to develop and deploy a tablet-based brain-computer interface (BCI) system to help these children, who had suffered multiple traumas, learn and practice the self-regulation of anxiety and attention. The goal was to improve their ability to self-regulate during school .
It was during my work in Nepal that I had a head-on collision with ethics. During the early months, I asked myself almost daily if I was doing any good in the world or if I was just assuaging my Western guilt. I'm a tenured professor at Simon Fraser University, a well-respected comprehensive university. I live in Vancouver, Canada, a city with one of the highest quality-of-life ratings in the world (https://www.mercer.ca/en/newsroom/mercer-2017-retirement-quality-of-living-rankings.html). I have kind and talented friends and colleagues. I have faced little prejudice in raising my two sons as a separated single mother. I had many of the privileges that my culture offered, so I was a little suspect of my own motivation. It's those doubts that led me to deeply examine my own ethical views and practices and bring them to the forefront of my research, rather than thinking of them as an afterthought or as some annoying paperwork that I had to get done. This commitment to stop and really think about what I was doing with my research in Nepal was critical to deepening my understanding of the ethical issues that arise when working with vulnerable populations.
Most ethics boards classify all children as a vulnerable population. The reason given for this is that children lack the necessary cognitive capacity to decide whether or not to participate in most research. It may be difficult for them to foresee the risks or potential benefits to their own well-being or to understand how the conditions of research may or may not be in their own best interests. Children who have special challenges, such as those with dyslexia, ADHD, developmental delays, or mental health issues, or children living in poverty, who may be illiterate or repressed, may have even less capacity to understand and give assent to participate in research. Working with and for children, which is the cornerstone of the child-computer interaction community, raises a number of ethical questions.
I've come to think that any discussion of the ethics of conducting research with vulnerable populations, such as children in the developing world, should be grounded in a three-part foundation. The first is that we must be willing to take the time and make the effort to be aware of and ask ourselves some of the difficult questions that arise when working with vulnerable populations. The second is that we must understand that we view the world through a lens of our own experience. In order to better understand vulnerable populations, we need to be willing to step away from this lens, ask questions, and listen deeply to the answers. The third is that we should consider the Belmont Report's tenets of welfare, justice, and respect (https://www.hhs.gov/ohrp/regulations-and-policy/belmont-report/index.html) in our dealings with vulnerable populations, keeping in mind that often what we ought to do is not always consistent with what we can do.
With this tripartite foundation in mind, there are five issues that I came up against in my work in Nepal. I pose them as questions here, which I urge you to consider when you work with any vulnerable population, whether that be children living in poverty or people of any age with physical or mental health challenges or the elderly in developed countries.
First, how can we feel relatively certain that we are providing benefits to the population we are working with? We may think that children can benefit from participating in our research or from using the computational systems that result from our research. But is this true? How do we know if the children we study are actually benefiting from our research? In answering this question, it is important to understand that we all make assumptions about what children's development, education, and even their lives should look like.
In Crazy Like Us, the Globalization of the American Psyche , Ethan Watters suggested that mental health professionals from industrialized countries have generated modern theories of the human psyche, created treatments based on these models, and then exported them to many of the world's cultures, sometimes with devastating consequences. Can we really take what we know about mental health and well-being and transfer it to other cultures? The assumption behind this commodification is that everyone is crazy like us. But is this true? Before we can assess potential benefit, we must be aware of the assumptions we use to determine what benefit means.
|Figure. Showing rather than telling what children will do in Mind-Full sessions.|
In Nepal, we did this by working closely with the community at the school, including Nepali counselors, to set goals for the benefits we thought might come from our research. Based on the results from informal interviews, as well as detailed observational notes and photographs taken on site visits to the school and to the surrounding communities, we returned to Canada and built a BCI application called Mind-Full (http://www.mindfullapp.ca). The system provided neurofeedback depicted in ways that we hoped the children could understand, and that would make their invisible brain states visible to them as they learned to self-regulate. Mind-Full comprised three games. Two of the games showed children how to use their bodies to create calmer brain states (i.e., self-regulate anxiety), and the other game helped them learn to focus their attention. The games were based on everyday activities that the children already knew how to do, like blowing on a pinwheel to make it spin. The act of blowing (or deep breathing) produces a calmer brain state. We also built a second application that let the counselors calibrate the EEG data, adjust the difficulty of the games, and see each child's brain state in real time. It would not have been possible to create this kind of system without trying to see the world through the eyes of the Nepali children and their caregivers.
Can we really take what we know about mental health and well-being and transfer it to other cultures?
Once we had the Mind-Full system designed and usability-tested in Canada, we went back to Nepal to conduct a controlled field experiment at the school. When we developed the assessment instruments for the study, our starting point was the report cards that the school used to assess the children. These gave us a further window into the school's goals for the children. Fully understanding the goals required us to ask questions about their notions of well-being and developmental milestones. We had to be willing to listen to the answers, even when it meant revising our research designs and protocols. We spent almost a week at the school before starting the study, redeveloping our survey-based assessment instruments and tuning the Mind-Full system and session protocol, which dictated how the counselors would work with the children. Many elements of the study design changed as we adapted it to a culture and environment that were very different from any in which we had ever worked.
Our study results indicated that all of the children were able to learn to use Mind-Full within a few minutes. Over time, through their sessions with their Nepali counselors, the children also improved their ability to self-regulate. They were able to calm down and focus, not just on playing the Mind-Full games but also on the playground and in the classroom . From a research point of view, the project was a success. But at the end of the day I had to ask myself if the children really benefited. For the answer, I turned away from scientific evidence and looked to the responses of the people who gave their time and energy to do the study with me. When the head counselor at the school asked us to continue working with them, I had some assurance that I had helped them provide benefits to their children.
Second, given the requirements of many university ethics boards, how can we work with children who may most need help but are least able to give assent? The challenge is how to work out ways of presenting research that are comprehensible to children. Otherwise we exclude the children who could benefit the most from the work we do, because they cannot easily give assent or because they may be difficult to access or work with.
In Nepal I faced the challenge of how to get assent from children who did not speak English, had not yet learned to read and write, had never used computers, had never seen an EEG headset, and, of course, had no idea what research was. Yet somehow I had to explain these things and ask them if they wanted to participate in a research study with neurofeedback to improve their ability to self-regulate!
Instead of telling the children what the study was about, we decided to show them. For instance, we demonstrated how to take pictures with a tablet, which some had seen tourists do. We showed them that they could tap on their photo, which served as their ID, to start the Mind-Full app. The first game began with an illustration of a young Nepali girl holding a pinwheel, which was something they could understand and mimic. My son, who was 12 years old at the time, demonstrated how to play the Mind-Full games. The children watched him take deep breaths to calm his mind and saw that this caused the pinwheel on the tablet to spin. All the games were based on activities that were familiar to the children. My son showed them how he used his body to cause digital effects that were represented in ways that made sense to them (e.g., sustain calmness to land a paraglider, focus visual attention to build a rock cairn). Although my son was foreign, male, and older than these children, he was not yet an adult and he had no existing power relationship with them. The entire assent process took place between the children, my son, and their counselor. I think this may have avoided the power imbalance that would have existed between me and them if I had been involved. I'm still not clear if the children really understood what it was they would do in the study, but I think they did understand that it would not harm them and that it would be different from anything they had ever tried. They were curious to give it a go.
Third, how do we find a balance between conducting rigorous research and ensuring that our research does not harm the children we work with? The challenge is understanding the tradeoff between rigor in research design and causing harm to the children. The gold standard for effectiveness of an intervention is a double-blind, random, controlled experiment. This design requires an active control group with a placebo (e.g., sham neurofeedback) that creates the same expectations of benefits as the real intervention. That's the right design for rigor. But is this ethical? Putting the children through an intervention that I was almost certain wouldn't help them seemed unethical. It also seemed unacceptable to me to use valuable resources and time to run sessions with the children using a placebo intervention instead of using those resources to help the children in other ways. I was also concerned that a placebo intervention might directly harm the children. For example, it might have negatively impacted their self-esteem if they failed to progress. And yet if we do not use a control group and resort to a case study, then it is difficult to be sure that the intervention is effective. Again, here were may have used valuable resources but may not be clear about the benefits.
The entire assent process took place between the children, my son, and their counselor.
In Nepal we decided to use a waitlist control-group design, so that all the children had the opportunity to use the intervention. We ran the first group of children through the intervention and used a pre-post assessment design to decide if we thought there was benefit . We saw large within-group improvements on survey measures of "calm" and "attention" and saw little change in the control group, thus reducing the likelihood that the cause of the improvement was maturation or learning. Since it seemed likely that there was benefit, we then ran the waitlist children through the intervention . This study design felt like a balance between being rigorous and being ethically sound.
Fourth, how do we manage children's expectations and attachment with the research team in ways that will minimize any emotional damage to the children? The challenge is determining how to manage children's expectations and attachments to us as researchers. When I was speaking with the head of a Nepali orphanage, I learned that they do not allow volunteers from other countries to work directly with the children at the orphanage. They had seen a pattern where a child would form an attachment to a foreigner. This often created an expectation in the child's mind that the foreigner would adopt the child, return to care for them, or provide funding (i.e., perhaps send them to a foreign boarding school). This expectation often reduced the child's motivation to try and make a better life for themselves. In addition, children at the orphanage often became emotionally attached to volunteers, and when they left there was attachment damage (https://childtrauma.org/wp-content/uploads/2013/11/Bonding_13.pdf).
|Figure. What happens when the research is over.|
In our study, to manage expectations and reduce attachment damage, none of the researchers on my team worked directly with the children in the Mind-Full intervention sessions. Instead, we trained and coached the counselors at the school to run the sessions. While this may have reduced the amount of control we had over the intervention protocol, ethically it seemed like the right choice. Together with a Western-trained therapist, I spent the first week of the sessions on site checking that the intervention protocol was being followed, being there to address questions and provide coaching. Our team then left the site, remaining in email contact for several weeks, returning to do another check-in and to be present for the post assessment. In this way, we avoided having the children form attachments to us.
Fifth, what will we leave behind when the research ends? The challenge is thinking ahead and addressing the legacy that should be left behind when the research is complete. It also includes asking if there are other contributions outside of research that can be made to the community in which we are working.
One of the principles of the non-profit organization that runs the Nepali school is that when people come from developed countries, their role is to do what is called capacity building. This means transferring skills, knowledge, and techniques to the Nepali staff who work at the school, rather than parachuting in, doing work, and then leaving. We also felt that it was important to ask the school what we could leave behind and what they wanted as a legacy.
We came up with a number of legacy items. First, we spent the better part of a week training all the teachers and counselors to assess the children rather than doing it ourselves. We co-created an assessment administration manual with examples of observable behaviors taken directly from the school context. The ability to assess the children using an observation-based survey instrument may improve the staff's capacity to set goals and monitor the children's progress in all elements of their schooling. We also spent a good deal of time with two of the counselors explaining how the software worked, how to upload log files, and how to conduct software upgrades, providing valuable technical training. At the end of the study, we left two versions of the Mind-Full system, including all the hardware. The staff could then continue to work with the children if they chose. We knew that even if they didn't, they could utilize the four tablets. For example, the tablets provided a way for counselors to take notes, access email, and collect data remotely when they were working with youth at remote jail sites. Back in Canada, we also raised money for an operation for one of the girls at the school, which had nothing to do with our research and everything to do with ethics.
By examining these ethical questions as we plan and conduct our research, we can, as a community, take the necessary care to ensure that our work has a positive impact and that we are contributing in ways that are important not just in our own minds but also in the lives of the vulnerable populations that our work is intending to serve.
1. Antle, A.N. and Bevans, A. Creative design: Exploring value propositions with urban Nepalese children. In Advances in Computer Entertainment, Lecture Notes in Computer Science 7624. A. Nijholt, T. Romão, and D. Reidsma, eds. Springer, Berlin/Heidelberg, 2012, 465–468.
2. Antle, A.N., Chesick, L., Levisohn, A., Sridharan, S.K., and Tan P. Using neurofeedback to teach self-regulation to children living in poverty. Proc. of Conference on Interaction Design for Children. ACM Press, 2015, 119–128.
4. Antle, A.N., Chesick, L., Sridharan, S.K., and Levisohn, A. Learning self-regulation: Can a brain computer interface help children living in poverty? Personal and Ubiquitous Computing. (under review).
Alissa Antle is a professor in the School of Interactive Arts + Technology at Simon Fraser University. Her work in child-computer interaction pushes the boundaries of computation to augment the ways in which children develop, think, and learn. In 2015 she was included in the Royal Society of Canada's College of New Scholars. firstname.lastname@example.org
Copyright held by author. Publication rights licensed to ACM.
The Digital Library is published by the Association for Computing Machinery. Copyright © 2017 ACM, Inc.