Conversations with Animals: Participatory design with dogs
Duncan Edwards with Winnie and Cosmo
Our first workshop took place on the 12th and 13th of April 2013 at the Animal-Computer Interaction Lab (ACI), part of the Computing Department at the Open University. Our brief was to explore how one particular method for co-designing research - Participatory Design - might be adapted to working with non-humans, more specifically dogs. Participants included members of the project team, researchers at the ACI lab, two dogs from Dogs for the Disabled, Winnie and Cosmo, their trainers Helen McCain and Duncan Edwards, as well as Helen's companion dog, Willoughby.
We had already found that the process of designing the workshop itself had pushed us to think in new directions, so the team were really excited about what the two days themselves would bring. You can find reflections on everyone's experiences on the blog, and on this page we want provide a sense of what we actually did in the workshop.
In participatory design processes users are invited to work with designers and researchers to identify design problems, develop responses to these problems and then evaluate and refine them. Of course in a two day workshop we would never be able to do justice to each of these steps, but we were interested in trying out a cut-down version of this process that would help us get an initial sense of what kinds of issues might arise as a result of trying to include dogs in these processes. We decided to focus particularly on exploring the issues, identifying design problems and then coming up with some initial prototypes.
On Day One most of us initially spent a little time getting to know each other and then getting a broad overview of the project and the work of the ACI team. The dogs didn't participate in this first session, as Helen and Duncan wanted to get a feel for the venue first and to check out what the rest of us were like, though they did have name tags made up for them like the rest of the participants. This led to discussions at the time (which continued across the rest of the workshop) about what this meant from a participatory point of view. Intermediaries are often involved in participatory research, for example, and perhaps a desire for what seemed like equality might actually put the dogs at a disadvantage by putting extra stress on them and their trainers. This remained an open issue throughout the workshop.
We had already found that the process of designing the workshop itself had pushed us to think in new directions, so the team were really excited about what the two days themselves would bring. You can find reflections on everyone's experiences on the blog, and on this page we want provide a sense of what we actually did in the workshop.
In participatory design processes users are invited to work with designers and researchers to identify design problems, develop responses to these problems and then evaluate and refine them. Of course in a two day workshop we would never be able to do justice to each of these steps, but we were interested in trying out a cut-down version of this process that would help us get an initial sense of what kinds of issues might arise as a result of trying to include dogs in these processes. We decided to focus particularly on exploring the issues, identifying design problems and then coming up with some initial prototypes.
On Day One most of us initially spent a little time getting to know each other and then getting a broad overview of the project and the work of the ACI team. The dogs didn't participate in this first session, as Helen and Duncan wanted to get a feel for the venue first and to check out what the rest of us were like, though they did have name tags made up for them like the rest of the participants. This led to discussions at the time (which continued across the rest of the workshop) about what this meant from a participatory point of view. Intermediaries are often involved in participatory research, for example, and perhaps a desire for what seemed like equality might actually put the dogs at a disadvantage by putting extra stress on them and their trainers. This remained an open issue throughout the workshop.
Dee Heddon doing the clicker excercise
|
The next two sessions were led by Helen and Duncan, who provided us with a mixture of theoretical overview and practical experience. First, we were given an overview of the work of Dogs for the Disabled, and more specifics about the kinds of dogs that do this work, how they are selected and trained, and some information on canine perception. We also got to meet Winnie and Cosmo in this session for the first time. Breed was important for a number of reasons, partly because of the different temperaments - a guard dog might not be as interested in picking up dropped items as a retrieval dog for example, but also because the public already have certain perceptions about what kind of dog can do service work and so might not be supportive of less expected or 'quirky' breeds. A point that we particularly picked up on later in our design sessions was the idea that the fundamental unit for Dogs for the Disabled is not the individual service dog, but rather the 'partnership' which consists of a dog and a human, or a 'team' which is a dog, a human-child who the dog works with, and a human adult who occupies the role of the team leader.
To test out the kinds of communication used between the members of the partnership, we then tried out a clicker training exercise where we divided into teams of 'trainers' and 'dogs'. The 'dogs' were sent out of the room while the rest of the team decided what tasks they would like the 'dog' to accomplish. The only way we could communicate with the 'dog' was through the use of a clicker. A click signalled positive reinforcement for the behaviour, but there was no explicit signal for doing something incorrectly. It was a really interesting exercise, which required us to think through how the set task could be broken down into its components and also what was even possible with the limited range of behaviours and communication tools available to us. After lunch, Helen and Duncan focused more on the kinds of challenges assistance dogs face in their work, which would be important for us in our design session on Day Two. Helen had her dog Willoughby with her in this session, in part so that we could see the difference in behaviour between him and Winnie and Cosmo. We learned about the variety of objects that assistance dogs need to become familiar with, and particularly the difficulties of negotiating the range of designs used for buttons (which the dogs are able to push) and for doors (which the dogs can pull or push). Retrieving objects could also range from small, thin items like credit cards to large items like walking sticks. They could also be of different materials, including metal which is unpleasant for the dogs to pick up with their mouths. We also discussed the difficulties dogs could encounter in the different environments they work in such as the shiny floors in shopping malls, or seeing where exits might be in a crowded space where the dog is much lower down than most of the people. |
|
We then had an hour or so working with Winnie and Cosmo to learn how they are trained to retrieve and to push. Using a wooden disc on a long handle, a clicker and small treats we had a lot of fun learning basic training methods. Winnie is one of the Dogs for the Disabled 'demo dogs' and helps to train new trainers and works with members of the public more generally. She was very forgiving of our fumbling attempts to point, click and give a treat in the right order. Cosmo is younger and has only recently started his training, which gave us a some sense of the issues the dogs face at different stages of their training.
After saying goodbye to Helen, Duncan, Winnie, Cosmo and Willoughby, our final session was spent mapping out some of the key learnings from the day in relation to the four key steps in the design cycle (requirements, designs, prototypes, evaluation) in preparation for Day Two. |
|
On Day Two Clara put us to work designing. Helen and Duncan had identified buttons and doors as two key challenges that assistance dogs faced in their work and so we divided into teams to try to develop some paper prototypes that might respond to each of these design issues.
|
Both teams realised that when designing for multi-species teams or partnerships we actually had to ask some really fundamental questions about what exactly is a door or a button. How do dogs experience a door for example? We also tried to draw on the 'toolkit' of trained behaviours that Winne and Cosmo had demonstrated for us, including retrieving, pushing, pulling and speaking. Other questions were raised around the scope of our design interventions - were we seeking to simply augment a human-centred world, or did we want to try to design within an idealised framework where the wider architectural landscape could be transformed for our multi-species teams. Finally, we also tried to think through how our prototypes might be evaluated, and particularly how assistance dogs would participate in the evaluation process and make decisions about which designs could hypothetically be sent forward for further development.
|
|
|
Throughout the design process we had tried to keep in mind the five requirements for participation that were set out in Finn Kensing and Jeanette Blomberg's paper "Participatory Design: Issues and Concerns" [pdf], and so in our final session we returned to the requirements and discussed how each one might apply to our experiences, but also how what we learned in the workshop might work back on these requirements and transform them in unexpected ways.
|
Requirements for Participation: |
Thanks to Helen, Duncan, Winnie and Cosmo for contributing so much to the workshop. Thanks also to Clara Mancini for designing and organising the workshop, and to the rest of the ACI team for hosting us!
The research team from a dog's point of view
