A New User Testing Methodology for Digitally Mediated Human-Animal Interaction

(1)

A NEW USER TESTING METHODOLOGY

FOR DIGITALLY MEDIATED

HUMAN-ANIMAL INTERACTION

MICHELLE WESTERLAKEN

Interaction Design Two-year MSc Thesis project 1 – 15 credits

Semester 2 – June 2014 Supervisor: Simon Niedenthal

(2)

2

Acknowledgements

This first thesis project for the MSc Interaction Design is the result of ten intense and informative weeks of reading, writing, testing, coding, analysing, and reflecting. Without a doubt, this has been the most interesting and fruitful period in the first year of this Master, as I was able to spend all my time on a topic which I feel endlessly passionate and enthusiastic about. It is my belief that we, as human beings, can use our abilities to develop technology to improve the lives of animals that live in our society. I have dedicated myself to this topic during my studies and in my free time for the last three years, and I hope that this thesis will form another small addition on the way to this great ambitious goal. This work would not have been possible without the help of several people and animals.

First and foremost, I would like to thank Alex Camilleri, for being on my side during every step along the way. Not only did he offer valuable help with data-analysis and proofreading, he is also the programmer, animator, and additional designer of Felino, the tablet game used in this research. On top of this, despite his anxiety for cat-like creatures, he was always there to join testing sessions in the wild.

I also thank Marcello Gomez Maureira, a dear friend with whom I share an interest in the academic field. His expertise in the field of data-analysis has helped me to make sense of my research. Marcello is always there to listen and provide unconditional and valuable help.

I would also like to thank Simon Niedenthal, my supervisor for this thesis. Despite my endless stories about cats and other animals, he was able to provide me with constructive feedback and valuable help for this thesis. While we mostly communicated over distance, his feedback was always clear and relevant. Next to this he always showed interest and enthusiasm for my work. My thanks are also extended to Stefano Gualeni. Even though he was not involved in this thesis, the work that we have done together formed the basis for this research. While he was writing his PhD he spent a great amount of his free time on helping me write my first research paper. Without this, I would not have been able to apply the methodologies and theories that are used in this thesis. Furthermore, I thank the people at the Animal Shelter Breda and the other cat owners that let their cats participate in this study for their voluntary help and openness towards my research. With this, I also thank Tijger, Bino, Petit, Wodan, Sita, Goofy, Siep, Sienna, Jackson, Pommie, Femke, Kaka, Sting, Sjaakie, Sunshine, Dorus, Amanda, Morris, and Missy for their participation in this study. Thank you for being yourself at all times.

Two other creatures I would like to thank are my dogs Jojo and Boogie. Looking into their eyes every day reminds me of the goals that I have and strengthens my ambition to achieve these. Throughout this year they kept me company in Sweden and eagerly joined in the early stages of the digitally mediated interactions that I developed (see michellewesterlaken.wordpress.com). The luxury of living with two enthusiastic little research participants is wonderful.

Finally I want to thank you, the reader, for taking your time to go through this work. I hope you will find it interesting and inspiring.

(3)

3

Abstract

This thesis evaluates a novel methodology for the user testing of digitally mediated human-animal interactions. The proposed method includes the structural analysis of video observations following a Grounded Theory approach. Complemented with more subjective human observations, this methodology aims to initiate a more informed iterative design and research process in which the animal’s experience with a playful artefact is analysed and reflected upon. The research involves the user testing of a prototype for an independently developed tablet game designed for cats and humans. With a focus on the user experience of the cat, the data analysis of this study results in new insights in the behaviour of the cat while interacting with the game. These outcomes are subsequently concluded in the form of design iterations that can help to improve the prototype. This study demonstrates how a new methodology can provide an initial focus on the perceptions and experience of the animal and lead to valuable insights that can advance the design of a digital artefact intended for animal use. Further research in this new area of interaction design can benefit from this study by expanding the theoretical framework and methodologies to different contexts and settings with the integration of playful technological artefacts and other animals that are known to engage in natural play.

(5)

5

1. Introduction

One fundamental difference between humans and animals is the ability to perform self-reflection. According to Plessner’s Theory of Positionality, animals have an experiential centre and a degree of self-awareness; they are conscious beings (Plessner, 1975). This gives animals an understanding of their body in space and, to a certain extent, enables them to make independent decisions. For example, the animal can decide to relocate, look for food, interact with other beings, and interpret certain signals such as body language or sounds. Plessner (1975) argues that humans have the same capabilities; however, on top of that, humans can form a cognitive relationship with their experiential centre, they are not only conscious, but also self-conscious. This enables them to perform self-reflection. This means that we as humans can perceive our own body and look at ourselves from distance. We can compare ourselves to others, and think about our experiences in time and space. This fundamental difference enabled humans to start making and using tools and eventually develop technology.

Nowadays technology affects our society and its surroundings on many different levels. We use it as a means in a number of ways such as to achieve goals, generate inputs, produce outputs, develop systems, solve problems, design interactions, craft, and control our natural environment. Our lives exist around technology and our global economy is dependent on it. Next to the design of technology for human life, we also develop indefinite amounts of technology that affect animal life, in areas such as agricultural engineering, animal tracking, breeding, animal research, veterinary, and the domestic animal market. In the last few decades, a considerable amount of research as well as commercial products have appeared in the field of interaction design. Even though this research field is still exploratory, technical innovation and an increased importance of improving animal welfare make the development of technological interactions for animals both possible and timely. This thesis further explores the possibilities of designing interactions that have animals as their intended users and focuses specifically on digitally mediated human-animal interaction that includes the domestic cat.

1.1 DIGITALLY MEDIATED HUMAN-ANIMAL INTERACTION

In 2011, Mancini was the first to introduce a research field within the area of Human Computer Interaction (HCI) that focuses on the design of interactions for animals, called Animal Computer Interaction (ACI). She claims that even though animals have been involved in machine interactions for many decades, the design of these technologies remains fundamentally human centred (Mancini, 2011). Therefore, Mancini proposes to start talking about ACI as a discipline in its own right, and to start working towards a systematic development taking a user-centred approach informed by the best available knowledge of animals’ needs and preferences. In short, ACI aims to:

“Study the interaction between animals and computing technology within the animals’ habitual contexts, design interactive technology that can support animals in their habitual tasks or daily lives, and that can foster the relationship between humans and animals, [and] develop a user-centred approach to the design of technology that is intended for animal use.” (Animal-Computer Interaction, n.d.)

(6)

6 From this context, I started to do further research towards the possibilities of designing interactions for animals in 2012. I wrote a thesis for my Master in Media Innovation that formed the basis for a paper published in 2013 together with Stefano Gualeni with the title “Digitally Complemented Zoomorphism: a Theoretical Foundation for Human-Animal Interaction Design” (Westerlaken & Gualeni, 2013). With this paper, we initiate a theoretical foundation that can facilitate further research. We propose a more informed and aware form of anthropomorphism that recognizes ‘play’ as a free and voluntary activity shared by both animals and humans. Next to this, we regard the designed technological or digital artefact as a mediator that can encourage playful interactions between humans and animals. It is therefore that, rather than ACI, I will refer to the design of digitally mediated human-animal interaction, which is a more specific term within the context of this thesis.

1.2 RESEARCH QUESTION AND KNOWLEDGE CONTRIBUTION

The aim of this thesis is to review a novel methodology in the field of ACI that is based on a more informed iterative design and research process. This method is approached from a Grounded Theory perspective, which includes a systematic way for looking at data without starting from strong preconceptions and hypothesis (Furniss et al., 2011).Within this research I will perform a series of user tests for an independently developed prototype of a tablet game that is designed for cats and humans. With a focus on the user experience for the cat, I will analyse audio/video observations in a systematic manner and try to find new design iterations. The research question therefore reads as follows:

How can the structural analysis of video observations using a Grounded Theory approach lead to a more informed iterative design process for digitally mediated human-animal interaction research and specifically, how can this methodology contribute to new design iterations for a tablet game that is designed for cats and humans?

This thesis aims to contribute knowledge to the field of interaction design and ACI in specific in the following ways:

- First, this study builds an overview of related research and commercial projects involving animals and digital systems by categorizing existing work and providing a critical stance towards some of the existing methodological approaches (Chapter 2 – Related Work); - Second, a theoretical framework provides an overview of concepts that are relevant within

the scope of this research and the current field of ACI. Together, these concepts propose a novel methodology that can be used for user testing digitally mediated human-animal interactions. This framework is structured around concepts related to the activity of ‘Play’, ‘User-Centred Design with Cats’, ‘Existing Design Guidelines’, and ‘Grounded Theory’ (Chapter 3 – Theoretical Framework);

- Third, Felino, the game that is used for user testing in this study is introduced and its main design process and decisions are outlined and explained (Chapter 4 – Introducing Felino); - Fourth, the research methodology and testing procedure explains how this research has

been conducted and how the data was analysed (Chapter 5 – Research Methodology and Procedure);

- Fifth, the research results and the data in which they are grounded are presented (Chapter 6 – Research Results), the research question is answered, and the implications for further research are proposed (Chapter 7 – Discussion);

(7)

7

2. Related Work

Throughout the last century, biological systems such as plants, microorganisms, and animals have been involved in machine driven interactions in a number of ways. From the 1930s, behaviourist B.F. Skinner started his famous experiments using operant conditioning methodologies involving animals such as rats and pigeons. For example, in Project Pigeon, a research program executed during World War II, Skinner proposed to train pigeons to peck at a target inside the nose of a missile to steer it in the right direction and thereby avoiding the development of more complicated technological solutions (Skinner, 1960).

Since then, more biological systems have been included in machine interactions, gradually evolving into complex and sophisticated environments. In this chapter I will outline and briefly review existing research and developments that appeared over the last few decades. Regarding the scope of this thesis I will focus solely on existing work involving animals in a digital environment, rather than other biological or non-digital systems. I divide these examples according to three categories: animals influencing digital systems, digital systems influencing animals, and playful digital animal interactions.

2.1 ANIMALS INFLUENCING DIGITAL SYSTEMS

A large amount of the existing research and practical implications in the field of ACI put the animal in control of the digital system. This makes the animal capable of providing input to a programmable environment, which in return, provides a certain output. In some cases, the animal is actively creating this input, for example by moving, and the output is generated accordingly. In 1989, researchers developed a system in which two Rhesus monkeys were trained to control a joystick and respond to computer-generated targets (Rumbaugh D. M., Richardson W. K., Washburn D. A., Savage-Rumbaugh E. S., Hopkins W. D., 1989). 15 years later, this experiment has been repeated successfully with albino rats (Washburn et al., 2004). In the last decade, this type of research has been advanced by medical researchers to the extent that monkeys can directly control robot arms using brain functions as input (Carmena et al., 2003; Velliste et al., 2008). In other words, the partial removal of the monkey’s skull and the application of neural implants allow researchers to measure cortical signals that are then manipulated into movements of the robot arm. Figure 1 shows a screenshot from a movie clip of the 2008 experiment (Velliste et al., 2008). Even though the ethical considerations of this research are outside the current objectives of the current ACI field, these studies fundamentally influenced medical studies in the field of prosthetics. Unfortunately the aims of these studies remain fundamentally human centred and currently provide no added value to the life of animals.

(8)

8 Another existing work in which the animal actively controlled a digital system includes an art installation by Ken Rinaldo called Augmented Fish Reality (Rinaldo, 2004). In this interactive installation, show in Figure 2, each sculpture contains a Siamese fighting fish. With the use of infrared sensors, the movements of the fish are tracked and translated into the movement direction of the sculpture. This allows the fish to move around in the environment and freely approach other fish and humans. A somewhat similar project was carried out in 2004 by Garnet Hertz. In this work, the designer created an experimental robotic system in which the bodily movements of a cockroach are translated into the physical locomotion of a three wheeled robot (Hertz, n.d.). See Figure 3 for an image of this robot. Both of these examples provide the animal with control over the digital system. However, the question that remains is to what extent the animal is aware of this interaction and if it has a certain understanding of the effects of its movements. In both cases, the designers claim that the animals are allowed to freely move around, but we cannot assume that the animal is aware of this without having enough understanding of the animal’s ontology.

Figure 2: Augmented Fish Reality by Ken Rinaldo Figure 3: Cockroach Controlled Mobile Robot by Garnet Hertz

In other existing work, the input and output might happen outside of the active embodied interaction of the animal with the digital system and does not require any training. These include the generation of automatic input such as GPS tracking systems and animal collars containing sensors. Over de last years, a few research projects in the area of HCI have looked at the use of tracking devices and sensor collars for domestic animals. In 2011, Paldanius et al. carried out an explorative study to start understanding the experiences and expectations of dog owners related to communication technology with animals (Paldanius et al., 2011). This study resulted in three

Figure 1: Monkey controlling a robot arm for self-feeding. In this video the camera angle does not show the partial removal of the monkey's skull in order to place the required brain implants.

(9)

9 different concepts. The first concepts proposed a dog collar for remote contact between dog and human, GPS location tracking, and biometric data generation of the dog (such as body temperature and the amount of barking). See Figure 4 for a visualisation of this concept. The second concept included a mobile service that stores activity data, such as exercises, of the dog and the owner in a diary. The third concept contained a dog-blog that is automatically updated with the use of data from a collar.

The researchers continued with a follow-up study in which these concepts were further evaluated and a combination of these three concepts was tested through a Wizard of Oz-type study (Paasovaara et al., 2011). Even though these studies might provide us with more insights on the expectations of dog owners and uncover usability issues of the concepts, they are fundamentally human centred, since they solely focus on the interaction with the human and mainly rely on data from interviews and focus groups. This results in exploratory research that follows a tentative approach without generating more understanding of the relationships that are mediated by the technology that is proposed. Mancini et al. also describe these underlying issues and invite researchers to question the sense-making mechanisms on both sides of the relationship (Mancini et al., 2012); furthermore, they call for a systematic approach to do research with animals through a user-centred approach.

2.2 DIGITAL SYSTEMS INFLUENCING ANIMALS

Another area of digital systems that mediate human-animal interactions are programmable environments or artifacts that control or influence animals. An example of this is a study from 1997 in which researchers use artificial electrical stimulation in order to control the movements of a cockroach (Holzer & Shimoyama, 1997). The cockroach is wearing an electronic backpack that forces it to walk in a predefined straight line. In 2013, a successfully backed Kickstarter project managed to collect over $12,000 to commercially produce cockroach control systems that allows the control of a cockroach with a mobile phone (The RoboRoach Kickstarter, n.d.). This so called first commercial cyborg can now be bought for $99 with complete hardware, firmware, and cockroaches (The Roboroach, n.d.). After the human installs all hardware and software, the cockroach can be controlled for approximately 20 minutes, after this, it needs a small break and after a couple of days, the cockroach device stops working and the cockroach can be allowed to retire. Figure 5 and 6 show an example of the product. These neurological systems have also been developed using other animals such as beetles (Maharbiz & Sato, 2010) and rats (Talwar et al., 2002).

(10)

10

Figure 5: The Roboroach mobile interface Figure 6: The Roboroach installed on a cockroach

Another, perhaps more well-intended, form of technologically mediated human-animal interaction that places the human in control involves humans and poultry. According to researchers Lee et al. touching and stroking is an important activity in human-pet relationships and therefore they developed and tested a wearable computer and mixed reality system for human–poultry interaction through the internet (Lee et al., 2006). In this system, the animal, in this case a chicken, is wearing a haptic jacket including vibration motors to simulate a stroking sensation. The human controls a plastic doll, as a representation of the chicken, which contains touch sensors. If the human touches the plastic doll, the jacket starts vibrating. Next to this, the human receives haptic feedback of this vibration. Figure 7 and 8 show the prototype.

Figure 7: A rooster wearing the haptic jacket Figure 8: The plastic representation of the chicken

The researchers conducted user tests with this system by conducting preference tests. Two chickens were observed separately during 28 days in which they could enter a small room with food and water, or a similar room with food, water, and the use of the haptic jacket. The chickens could make this choice over 200 times and the researchers found that in 73% of the cases, the animals chose the room in which the haptic jacket was used. The researchers therefore concluded that the animals did not experience negative feelings; however further studies have to be done in order to determine if the chickens actually liked the haptic interaction itself (Lee et al., 2006).

(11)

11

2.3 PLAYFUL DIGITAL ANIMAL INTERACTIONS

Another domain of digital environments that started to use animals in their interaction is the area of playful interactions and games. This area is discussed here separately, since it forms a general focus within the scope of this thesis. Even though the general aim of this work is to provide playful environments, the animal is often used as a component or mechanic to provide entertainment or play with a human focus, rather than a digital system which regards the animal as an actual user. This phenomenon can be observed in a project from 2006 in which researchers and designers built a simulation of the game Pac-Man in which human players could play against real crickets (Lamers & van Eck, 2012). The human players experience a regular digital Pac-Man game in which the ghosts are not controlled by computer code, but by real crickets that walk around in a physical maze. Another example includes the development of a series of biotic games, in which researchers at Stanford University built videogames such as Pong, Pac-Man, and Pinball in which micro-organisms form the moving elements of the game (Riedel-Kruse et al., 2011). In these games, the actions of the human player influence the movements of the micro-organisms and the aim of this project is to show the biological processes of these organisms and have fun at the same time (Riedel-Kruse et al., 2011).

Fortunately, there are also a few playful examples that would fit within the realm of ACI and approach a more user-centred approach in which the playful interaction is also intended for animal use. One of these examples is the Playing with Pigs project in which the prototype presents a game that allows humans and captive pigs to play together from distance (Alfrink, K., Van Peer, I., Lagerweij, H., Driessen, C., and Bracke, M., n.d.). The pigs can interact with a large touch sensitive display that is placed in the shed and shows moving objects. These moving objects are controlled by the human through a tablet device. The goal for the human is to guide the pig’s snout to a target on the screen. Figure 9 and 10 show the prototype of this game called Pig Chase.

Figure 9: Pig Chase – pig interaction Figure 10: Pig Chase – human interaction

During the past few years digital systems have also been implemented in captive environment such as zoos with the goal to enrich the animals’ environments or prevent problems related to boredom. One of these research projects is the ongoing TOUCH project which aims to provide environment enrichment for captive orangutans who cannot be reintroduced to their natural environment and to create new possibilities for playful cross-species communication (Wirman, 2013). After researching the organutan’s preferences in terms of their use of technology, custom made playful touch

(12)

12 interfaces are now being tested with two male orangutans under close human supervision in an experimental study (Wirman, 2013). In this research project, the activity of play facilitates the communication and meaningful engagement between species (Wirman, 2013).

Since the focus of this thesis situates around the design of digital mediation for the domestic cat, I will finally discuss the development of digital games that have cats as their intended user. Because of the cat’s natural interest in visual representations on digital screens and their physical capabilities of interacting with touch screen devices, the design of mobile and tablet applications for cats has become more popular over the last few years. Most of these include commercial developments that are available for Android and iOS devices and do not produce any published research on their design, development, or testing (Crazy Cat, n.d.; Game for Cats, n.d.). All of the cat game examples I came across generally present a moving object on screen that, in some cases, reacts to the input when the cat taps the screen, either with visual or auditory feedback. Unfortunately these games do not seem to include the experience and perceptions of the animal into account, since they include mechanics such as scoring systems and a lack of progression that is understandable for the cat. Furthermore, the objects in the game that the cat is supposed to tap often do not provide clear or consistent auditory of visual feedback. Lastly, the interfaces are designed for the human to start or stop the game are easily operatable by the cat, which might cause confusion and it breaks the play activity.

One research project that studied the possibilities for the design of a game for cats and humans includes a paper from 2011 in which a game was designed and tested (Noz & An, 2011). For this pilot study, the researchers carried out user tests with a prototype and conducted interviews with seven the cat owners and their cats (Noz & An, 2011). The data was analysed by coding audio, video, and transcripts from the cat owner’s interactions. Even though the research provided insights that might be helpful to improve the design of the game, sadly the research did not produce any results of the data generated from the cat’s interactions with the game. Instead it focused soley on human experiences with the game and therefore risked the adoption of superficial anthropomorphic statements about the cat’s perceptions and game interaction.

In summary, the advances in new technology and interaction possibilities facilitated a number of research projects and commercial developments that include animals in their digital systems. However, the small research area that follows a user-centred approach and takes into account the experiences and perceptions of the animals is still exploratory, tentative, and lacks a systematic approach. Despite the establishment of an ACI community that provides academic objectives and a research agenda (Mancini, 2011) and the formulation of initial design guidelines (Westerlaken & Gualeni, 2013), the design of new interactions and the results of research projects that fit into this approach still need to be presented.

(13)

13

3. Theoretical Framework

This chapter provides an outline of theoretical concepts that are relevant within the scope of this research. These concepts form the theoretical grounding for this study and guide the research. First, this chapter goes deeper into the theory behind the activity of play and consequently, I argue why this activity forms a suitable context for research towards digitally mediated human-animal interaction design. Second, I will propose the fundamental and unavoidable problem that user-centred design with animals brings forward which relates to the understanding of animal ontology. Third, I present existing design guidelines within the field of ACI that have been established by other researchers as well as my own work that try to cope with the issues of understanding non-human users. And fourth, I will provide a brief introduction to Grounded Theory as an approach for the methodological analysis of user-testing data.

3.1 PLAY

The activity of play is shared among many cultures and organisms and therefore can be considered of great influence in the lives of both animals and human beings. In 1938, Huizinga already described the independency of play from culture and human beings:

‘Play is older than culture, for culture, however inadequately defined, always presupposes human society, and animals have not waited for man to teach them playing. We can safely assert, even, that human civilization has added no essential feature to the general idea of play. Animals play just like men’ (Huizinga, 1950, p. 1).

Huizinga furthermore describes that play is by definition a voluntary activity as we can never force a human or animal to play (Huizinga, 1950). Play is considered an intrinsically motivated activity that provides pleasure or enjoyment and is carried out among animals, humans, and it can be a shared activity in interspecies-interaction. These characteristics of play: pre-cultural, voluntary, and shared between species, form a specifically suitable and safe context from where we can start designing and user-testing digital interactions (Westerlaken & Gualeni, 2013). The freeness of play provides us with a setting in which the animal does never have to be forced or artificially involved in the interaction.

In many animals, including the domestic cat, humans have found a series of behavioural characteristics from which we could derive that the animal is playing (Burghardt, 2005). However, it still remains difficult for human beings to establish the exact functions, boundaries, and rules to animal play, because we inevitably connect the activity with human perceptions and emotions. Throughout time, researchers have established different functions that play in animals might fulfil, such as discharge of energy, need for relaxation, exercise, imitative instinct, or training (Fagen, 1981). Furthermore, according to Burghardt, defining play itself has led to innumerable ways of characterizing the activity that mainly provide a list of elements that are involved in play, whereas a more valuable approach for systematic analysis would be an approach that separates play from other behaviours with which it might be confused (Burghardt, 2005). Following an ethological approach, Burghardt therefore states the following five criteria all have to be present in order to assign the label ‘play’ to the animal behaviour:

(14)

14 1. Limited immediate function: the behaviour does not seem to be fully functional or

contributing to the immediate survival of the animal;

2. Endogenous component: the behaviour is spontaneous, voluntary, intentional, pleasurable, rewarding, reinforcing, or ‘done for its own sake’ (at least only one of the concepts needs to apply);

3. Structural or temporal difference: the behaviour is different from ‘normal’ behaviour because it seems incomplete, awkward, exaggerated, precocious, or contains modified behavioural patterns;

4. Repeated performance: the behaviour is repeated during at least some part of the animal’s life span;

5. Relaxed field: the behaviour is performed when the animal finds itself in a relaxed, or low-stressed situation (Burghardt, 2005).

These elements provide a helpful checklist that can be used in analysing the behaviour of the playful and non-playful cats during the user-testing.

The types of play that will be applied in this research are defined by Burghardt as ‘object’ and (inter-species) ‘social’ play (Burghardt, 2005). The object serves as a digital mediator and the playful interaction will be taking place in the triangle of this mediator, the cat, and the human. This interaction triangle is visualised in Figure 11. Because of the digital component that specifically serves as a playful mediator in this case, I will also refer to the digital mediator as a (digital) game.

According to Salen and Zimmerman, successful game design is dependent on the extent to which meaningful play can be created:

‘[T]he process by which a player takes action within the designed system of a game and the system responds to the action. The meaning of an action in a game resides in the relationship between action and outcome’ (Salen & Zimmerman, 2004, p. 34).

(15)

15 However, I argue that in the proposed interaction triangle that forms the scope of this research, meaningful play is not only created in the relationship between the actions of the players and the outcomes of the system, but also in the playful interaction between the human and cat players. The purpose of the digital mediator is to facilitate meaningful play and the interaction triangle is the context for which the game is designed. In other words, the affordances and mechanics that are implemented in the game aim to stimulate playful interaction in all directions, not solely between the system and the player. This means that the game needs to avoid forcing the play and the players in a fixed direction. Therefore, rather than implementing for example strict rules, time pressure, winning/losing conditions, or scoring systems, the mediator embraces play in a more open setting. This eventually led to the creation of a game that merely serves as a digital toy, which is explained in more detail in Chapter 5 of this thesis. The biggest challenge of this design is to adopt a user-centred approach that integrates cats as users in an iterative design process.

3.2 USER-CENTRED DESIGN AND CATS

This research follows a User-Centred Design (UCD) approach that aims to involve the needs, wants, and limitations of the end-users into different phases in the design process. This design method is helpful in optimizing the development according to the preferences and experiences of the user. UCD is an iterative process. This means that the designer cycles through prototyping, testing, analysing, and refining a work-in progress (Zimmerman, n.d.). The design decisions are informed by the available knowledge on the user and their experience with the prototype.

INTERSUBJECTIVITY

Unlike a significant amount of the related work that was described in Chapter 2 of this thesis, this UCD research focuses on the user experience of the animal. However, when we want to design interactions for animals based on a UCD approach, a fundamental philosophical problem arises: How can we, as human beings, ever fully understand the experiences and perceptions of animals? In his 1974 essay What is it Like to Be a Bat, Nagel introduces this ontological problem and argued that the fact that animals have a conscious experience at all must mean that there is something it is like to be that animal. However, by trying to imagine what it is like to be an animal, in this case a bat, Nagel poses a conceptual issue:

“It will not help to try to imagine that one has webbing on one's arms, which enables one to fly around at dusk and dawn catching insects in one's mouth; that one has very poor vision, and perceives the surrounding world by a system of reflected high-frequency sound signals; and that one spends the day hanging upside down by one's feet in an attic. In so far as I can imagine this (which is not very far), it tells me only what it would be like for me to behave as a bat behaves. But that is not the question. I want to know what it is like for a bat to be a bat. Yet if I try to imagine this, I am restricted to the resources of my own mind, and those resources are inadequate to the task. I cannot perform it either by imagining additions to my present experience, or by imagining segments gradually subtracted from it, or by imagining some combination of additions, subtractions, and modifications” (Nagel, 1974, p. 324).

In other words, human beings are subjectively bound to their human experience and incapable of letting this go. The philosophical area of Phenomenology studies these differences of structures in

(16)

16 experience and consciousness (Dourish, 2002). According to Husserl (1913) the phenomenological method seeks to temporarily reduce and erase the world of speculation (subjectivity) by returning the subject to its primal experience to get to the core and essence of the thing, idea, feeling, perception, or experience. In other words, phenomenology can be understood as a study towards an objective understanding of consciousness by systematic removal of aspects until only the pure experience is left. Later, Heidegger altered this form of phenomenology (also called Cartesian Dualism) by regarding object and subject as one and therefore accepting the intersubjectivity of human beings. In his view, ‘being’, our most fundamental relation to the world, is practical rather than cognitive and embodied in our daily life (Dourish, 2002). Heidegger wrote that phenomenology is the art or practice of ‘letting things show themselves’ (Phenomenology (Stanford Encyclopedia of Philosophy), n.d.).

In other words, the intersubjectivity of human beings causes humans to interpret reality according to their own subjective experience. In the example of Nagel, in which he tries to imagine what it is like to be a bat, it becomes clear that this subjectivity might result in the projection of human ideas onto non-human life, which makes it theoretically impossible for a designer to fully understand the animal-user.

It is therefore that we need to accept that this intersubjectivity is unavoidable. However, as a designer following a UCD approach, the awareness of this ‘human projection-risk’ emphasizes the importance of informed design decisions and extensive user-testing. Chapter 3.3 presents the existing research in the field of ACI that propose methods that could further enable an informed UCD approach.

ETHICAL CONSIDERATIONS

A fundamental difference between studies involving humans and animals is the ability of humans to give informed consent to participating in a research. This emphasizes the importance to make ethical considerations in a UCD approach that includes animals as users.

In 2013, Väätäjä and Pesonen published a paper including a literature review and a total of 23 ethical guidelines that should be implemented in HCI studies that involve living animals (Väätäjä & Pesonen, 2013). In Appendix A, these 23 guidelines are stated and it is explained how this study adhered to these ethical considerations.

These guidelines are assembled from 13 sources in the fields of international animal welfare associations and animal behaviour societies as well as other relevant societies’ and organizations (Väätäjä & Pesonen, 2013). Furthermore, the guidelines are related to general guidelines, procedures prior to the study, during the study, and procedure related to reporting the study.

Even though this study is, so far, the only research proposing ethical guidelines for HCI studies involving animals, it forms a helpful and relevant addition to this study.

(17)

17

3.3 EXISTING GUIDELINES FOR RESEARCH AND DESIGN

Even though the research field of ACI is still exploratory, a few theoretical approaches to design and research have been carried out over the last few years. This chapter outlines three approaches that have been described including methodologies and guidelines in the field of Ethnography, Indexical Semiotics, and Digitally Complemented Zoomorphism.

ETHNOGRAPHY

Weilenmann and Juhlin argue that the field of ACI could advance by adopting an ethnomethodological perspective on animal computer interaction and thereby assessing the appropriate level of anthropomorphism (Weilenmann & Juhlin, 2011). In other words, since we cannot avoid our human subjectivity to assign projections on animal behaviour, we could make use of a more informed form of anthropomorphism by exploring and trying to understand the natural habitat and behaviour of the animals that are studied. Such an approach implies that the animal’s mental state is studied in particular social situations (Weilenmann & Juhlin, 2011). The authors furthermore write that ethnography is a specifically suitable approach for studying human-animal interaction, because of the focus on manifest, observable actions, rather than inner, mental states (Weilenmann & Juhlin, 2011).

However, the authors continue their study by researching human-dog interaction and the use of positioning systems (GPS devices) during hunting activities with a focus on the human side of the interaction. With this, they recommend to avoid doing research that requires the ethnographical analysis of the animal in order to avoid the projection of human subjectivity (Weilenmann & Juhlin, 2011). The result of this is the study towards the use of a device that involves both humans and dogs in the interaction, but only focuses on the human as a user. Even though the risk of human projection might be avoided, it does not provide us with any insights on the animal experience. Using this ethnographical study as an example, Mancini et al. later write that even though the approach of Weilenmann and Juhlin focuses on the immediate context of the interaction, the interaction itself might be defined by a broader relational context, which includes both the animal and the human (Mancini et al., 2012). They propose to advance the ethnomethodological approach with indexical interspecies semiotics (Mancini et al., 2012).

INDEXICAL SEMIOTICS

Semiotics is the study towards the making of meaning, including the study of signs and sign using behaviour (Merriam-Webster Dictionary, n.d.). Haraway described how we can start to understand other animals and the relationships we form with them by engaging with their material semiotics, even if they are not fully accessible by humans (Haraway, 2008).

On this basis, Mancini et al. propose a new approach in which the exchange of indexical semiotics forms the focus of understanding the relationships between animals, humans and technology (Mancini et al., 2012). There are three different kinds of communication signs: symbols, icons, and indices. Symbols (for example words) and icons (for example a no-smoking sign) are merely abstract signs that require linguistic abilities and cultural understanding. However, the authors write that indices are instead directly and physically grounded in a bodily relationship with the environment and other beings (Mancini et al., 2012). An example of an index is an animal releasing pheromones or an animal’s alarm call. This means that these indexical signs do not preclude or require shared mental abilities; instead they unfold themselves in the interaction. In other words, if

(18)

18 we are able to interpret an animal’s semiotic process on the level of understanding their indexical signs, we can connect meaning to them in the context of digitally mediated human-animal interaction (Mancini et al., 2012). An example of this could be playing with a dog. Through certain indexical signs such as body movement and playful signalling, both the dog and the human understand that they are playing and not actually fighting.

The paper continues with a study towards the relationships between humans, dogs, and tracking device technologies. In this study, it is researched how the dog-owner and the dog co-evolve in the interaction with each other (Mancini et al., 2012). Co-evolving, in this case, means that within the interaction, both human and animal might each attribute or express (their own) meaning (Mancini et al., 2012). By both trying to interpret the environment and make sense of the indexical signs of each other, the meaning of the interaction is created.

This work provides a first theoretical articulation towards the understanding of digitally mediated human-animal interaction in which both the human and the animal are involved as users. However, the study continues to rely solely on the subjective human interpretations gained from interview data. This is also clearly stated by the authors:

‘Although, even well-grounded insights may only crudely approximate the meaning of canine indexical associations, they may be good enough to be useful when developing or evaluating technology that is intended to support either human-animal interactions or animal-computer interactions’ (Mancini et al., 2012, p. 151).

This approach formed a valuable grounding for the development of a research paper published by myself and Gualeni in which we tried to further articulate a theoretical framework for further research with an approach we defined as digitally complemented zoomorphism (Westerlaken & Gualeni, 2013).

DIGITALLY COMPLEMENTED ZOOMORPHISM

Our research is based on the understanding of ‘play’ as a free and voluntary activity. Furthermore, it encompasses a more informed and compromising form of anthropomorphism that relocates the focus from the human perspective to the animal. The aim of this zoomorphic approach is to facilitate further research towards the design of artefacts that mediate the relationships between humans and animals. From this framework, the following three guidelines emerge and form the basis of this thesis:

1. Using external stimuli in the form of technological artefacts:

The natural curiosity of animals and their explorative behaviour can be used to stimulate their engagement with interactive technological artefacts in a research setting. This means that the animal is motivated by the artefact to engage in natural and voluntary play;

2. Analysing animal behaviour through ‘going along’ in a common embodied praxis: The understanding of indexical semiotics and common traits in the way bodily signals are produced and interpreted allows specific species to understand others to a certain degree. This ‘going along’ could be achieved in a common and free praxis such as play. This objective unfolds itself intuitively in the course of the interaction;

3. Digitally tracking metric and/or biometric data concerning the animal experience: In order to complement the subjective human approach that results from the first two guidelines, metric and/or biometric research can offer additional insights in the experiences

(19)

19 of the animals that are studied. This includes methods that can provide a quantifiable analysis to uncover the interaction with the artefact. Furthermore, we can discover changes in the animals’ bodily dimensions and/or movements and measure their psychophysiological changes with a higher degree of objectivity. These include the digital recording of changes in the movement patterns of the animal such as pace or position. Next to this, we can monitor physical dimensions such as heart rate, respiration, or body temperature. These methods allow us to understand and quantify the reaction of the animal to certain stimuli or experiences in a more systematic manner. Furthermore we can use metrics, for example through the logging data of the technical artefact, to define quantitative patterns in the user interaction that can be used to improve the artefact (Westerlaken & Gualeni, 2013).

These three guidelines form a combined interpretation of the relationship between humans, animals, and technology with the aim to provide insights and understandings that can help to improve the designs of playful technical and digital mediators. More details on how these guidelines are followed in this thesis can be found in Chapter 5 of this document. This study aims to evaluate these guidelines with a practical experiment. As an additional method for analysing the data that is obtained in this study, I will follow a Grounded Theory approach.

3.4 GROUNDED THEORY

Grounded Theory is a method that is often used within HCI research and proposes the iterative analysis and gathering of data without the formulation of a predefined hypothesis or other preconceptions about the expectations of the study (Furniss et al., 2011). The method typically includes the collection of data through interviews or video observations, which are then coded and identified in order to make sense of the data (Furniss et al., 2011).

There are many different forms of Grounded Theory studies and explaining each of these would go outside the scope of this thesis. Important is, that this study adopts a mixed methodology in which the data of video observations is coded, grouped into several concepts, and then translated into results. Furthermore, this study mixes both qualitative observations and quantitative data. As is often the case in Grounded Theory research, the results of this study are mostly presented as probability statements rather than statistical significance (Furniss et al., 2011).

The use of Grounded Theory for this study is helpful as it does not require me to start the experiments with any preconceptions about the findings. This is valuable since it avoids the projections of human statements about the interaction of the animal with the artefact a priori. Instead of focusing on a small element of the data that is gathered according to the hypothesis, all observations are recorded and coded accordingly. Chapter 5 describes this research methodology and procedure in more detail.

(20)

20

4. Introducing Felino

The tablet game that I used for the research in this thesis is called Felino. Felino has been independently developed together with programmer, animator, and additional designer Alex Camilleri. I am mainly responsible for the design and art of the game. The development started in August 2013 and the current prototype is the result of about 25 Sundays of work and a generous amount of additional readings on the available knowledge of cats regarding their sensory perceptions, experiences, and behaviour.

Felino can be considered a digital toy that allows humans and cats to play together. The main objective of Felino is to enjoy and share playful moments together with a cat. The game does not include mechanics such as highscores, time pressure, or game-overs, since these are most likely not understandable for the animal in cross-species play and do not facilitate open-play as explained in Chapter 3.1. Instead, Felino tries to provide an experience that is more recognizable for a cat: time spent together with the owner. Figure 12 shows a screenshot of the game’s prototype.

4.1 FELINO’S PLAYFUL ELEMENTS

The virtual environment of Felino is a top-down representation of an aquarium. This aquarium is populated by fish and sea creatures that the cat can catch by tapping on the tablet screen. At the bottom of the screen, a group of virtual controls allow the human player to alter specific attributes of game elements in order to align the game to the personal and playful preferences of the cat. The human player can for instance change the size, speed, colour, and brightness of the fish. Each fish has a specific look and behaviour that allows the cat to experience an unpredictable and varied activity of play.

(21)

21 One specific virtual control that the human can use consists of a small joystick that regulates the movement of a crab that is always present in the aquarium. With this crab, the human player can collect spheres that are released whenever the cat catches a fish. By collecting these spheres, other small crabs will be created behind the main one, generating a trail that the cat will be able to interact with (see Figure 13).

In this way we create a gameplay structure that allows cat and human to have a relevant role during the play. In addition to these gameplay elements, the human player can change the type of fish on the screen, find statistics of the gameplay, and information about the development of the game and the sensory perceptions and behaviour of cats by navigating the game menus.

4.2 INFORMED DESIGN DECISIONS

As developers, we tried to create a game that would be specifically suitable for further research according to the design guidelines proposed in Chapter 3.3. Thereby we intended to provide the human and the animal with the possibility to ‘go-along’ in the shared activity of play. Felino therefore invites the human and cat to play together, and provides the human with options to adjust the play according to the human’s interpretation of the interaction. Next to this, we decided to keep the cat-area free of elements that, we believe, are not understandable for the cat. These include for example introduction menus, textual explanations, or human buttons that can easily be operated by the cat.

Furthermore, we took certain design decisions based on how we could interpret the perceptions of the cat according to existing research. This includes for example information that we have about their perceptions of colour (cats mainly see blue and yellow colours), their eyesight (cats only see sharp from 30cm or more, but are very good at distinguishing patterns and contrasts), normal play behaviour (mainly using their paws), physical toy interaction (the use of different movements and paces), and sound perception (the use of high frequency sounds of prey-like animals) (Bradshaw, 2013).

4.3 DESIGN PROCESS

After we decided to develop a digital game for cats it was clear that tablet devices would offer us with the widest range of possibilities compared to fixed monitor systems (computers or consoles), or mobile devices with smaller screens (smartphones or portable game consoles). Next to this, the amount of cat owners that owns a tablet nowadays is growing which confirmed our choice to develop a digital game for tablet devices.

The first stage of the design process included the research towards existing tablet games for cats. Even though there are few, they are all built in a similar way: a moving object on the screen that (in some cases) reacts to the input when the cat taps the screen, either with visual or auditory feedback. After trying out some of these games with cats, and looking at gameplay videos we noticed that there was definitely a potential interest from the cat in these games. However, the existing games have an overly human focus and are lacking in many ways. They include scoring mechanics, game objectives that are not understandable for the cat such as time pressure, they sometimes lack visual or auditory feedback at all, they offer no progression, are highly repetitive, and the cat would often tap the human interface and thereby control the game unwillingly. Furthermore, these games are meant to be played by the cat only. We wanted to create a game that

(22)

22 takes a UCD approach, makes informed design decisions, and offers a playful experience for both cat and human.

We developed different game concepts with the aim to design game mechanics that provide both human and cat with a playful experience and the possibility to explore the activity of ‘going-along’ (see Figure 14). One of the problems that we came across in this phase was the difficulty to sync the playful interaction for the human and cat. For example, some of the concepts required the cat to kill a fish so that the human could carry out a new action (such as catching the fish). The concept for Felino was unique in the way that it provides playful interaction for both cat and human at the same time and allows the human to adjust the game environment in order to optimize play. With this premise in mind we decided to develop a prototype for Felino.

During the prototype development, a great amount of time was spent on the interface design (see Figure 15). We tested different control buttons and interface elements that were intuitive for the human, but not so easily operable for the cat. One of the main decisions was the division of the screen in a cat-area and human-area. This division allowed us to design a human interface that does not block the gameplay for the cat. The cat can continue playing even when the human player opens the menu or covers the interface with her hands.

During the development we continuously reflected the design on our knowledge about cat behaviour and perceptions. For example, by looking at regular playful behaviour of cats with toys, we tried to animate the movements, feedback, and swimming patterns of the fish accordingly. Occasionally we tried the prototype with a small amount of cats to analyse their reaction and find small confirmations or problems before the prototype was ready for extensive play-testing.

Figure 14: Different game concepts

(23)

23

5. Research Methodology and

Procedure

In this chapter I will outline the research methodology and testing procedure of this study.

5.1 METHODOLOGY

To answer the research question, I conducted user testing with a prototype and a total of 19 domestic cats. Most of these cats (15) were temporarily living in the Animal Shelter Breda, the Netherlands, because they were abandoned or stray cats. The other 4 participants were tested in their home environment. In total there were 9 male and 10 female cats that participated in the study and they formed a mix of appearances, and characters. The average age of the cats was 5,4 years old, the youngest cat was 7 weeks old and the oldest cat 12,5 years.

All user testing sessions were completely recorded using both audio and video. The length of each recording greatly diverted because it was dependent on the interest of the cat. On average the sessions lasted 5 minutes and 51 seconds. However, the shortest session was 57 seconds and the longest session lasted for 15 minutes and 54 seconds. Appendix B shows an overview of all the participants including their participant number, name, testing date, age, gender, and the length of the test session.

The data was then analysed according to the principles of Grounded Theory as explained in Chapter 3.4 of this thesis. With the use of Elan software, the audio/video recordings were coded according to six different categories: Cat behaviour, Human behaviour, Game state, External distractions, Unclear video, and Talking. Each of these categories contained a different number of annotations. After the first round of data analysis, another round of coding was carried out including two new categories: Cat killings, and Extra Interest. The complete codebook can be found in Appendix C of this thesis. Figure 16 shows an example of the software’s interface during the coding phase of this research. The research follows the three guidelines that are presented in Chapter 3.3 in the following manner:

- The tablet-game Felino forms the technological artefact that stimulates the engagement of the cat in natural and voluntary play.

- The human can subjectively analyse the behaviour of the cat by joining the play session and ‘going along’ in this common embodied activity. Thereby the human can use her understanding of indexical semiotics such as the body language of the cat to adjust the game in the course of the interaction. The human can for example change the speed, size, colour, and brightness of the different fish, control the crab, move the tablet, and use body language to play with the cat.

- With the use of coded video recordings a systematic quantifiable analysis can uncover the interactions with the artefact as well as the interactions of the human and cat. This coding is done on different levels including the coding of cat behaviour to analyse movement patterns and playful signals, the coding of human behaviour to analyse their interaction with both the cat and the game, and the coding of game content to analyse what happens in the game at certain points of interest. Initially, the game was programmed to record different types of

(24)

24 in-game metrics, such as the amounts of taps in specific areas and the type of fish that were killed. However, due to the large amount of human interaction in the cat area, the metrics were highly influenced by the human interaction and the video recording data proved to be more valuable for the analysis of game content.

5.2 USER TESTING PROCEDURE

The user testing sessions generally started with a short greeting to establish the mood and general character of the cat. This greeting involved some petting and talking to the cat in order to initiate the interaction and determine if the cat was not scared, aggressive, or otherwise uninterested in participating. After this short introduction, the tablet with the prototype of the game was presented to the cat. The screen was positioned in such a way that the area including the content for the cat was pointed in the cat’s direction and the human-area in my direction, as seen in Figure 17. Furthermore, I positioned myself on the same height of the cat whenever possible in order to avoid dominant body gestures as much as I could.

Figure 17: The positioning of the tablet Figure 18: Participant interaction with the prototype Figure 16: Elan software interface for participant 011

(25)

25 During the testing session, the cats were free to behave however they wanted. As an example, some cats decided to leave the interaction immediately after the prototype was presented, others solely looked at the screen with interest for different durations, some cats started touching the screen and showing playful interaction, and some cats showed a little more surprising behaviour (see Figure 18). All their behaviours were recorded with a mobile camera by the co-developer of the prototype. This second person did not interfere in any of the interactions with the cat or prototype. Nevertheless, his presence was often clearly noticed by the cats.

In principle, the prototype was presented to one cat at the time. However, because of the open nature of the shelter’s cat quarters, another cat interfered in the interaction during two different testing sessions. Next to this, the animal shelter formed an environment with many distractions for the cat and human which are visible in the recordings, such as barking dogs, visitors, other cats, and employees.

As mentioned before, the length of the sessions were determined by the interest of the cat. If the cat decided to definitely leave the interaction, the recording was ended. In this case, leaving the interaction should be interpreted in both physical and mental form. The cat could for instance actually walk away from the tablet, look away from the tablet for a long amount of time, turn its back to the game, or move over to another activity. Furthermore, the recording was ended after I determined that a certain data saturation had been reached. In other words, when the cat did not show any change in behaviour for an extended amount of time or when I felt it was unlikely that the cat would show any further interaction with the game. After the session ended, the cat was shortly greeted and petted again.

My role, during the testing session, was to encourage the cat to start interacting with the prototype. Even though I did not have any knowledge about the particular preferences of the cat-participants, I interacted in a similar way in which a human could encourage the cat to play with a regular toy, for example by talking to the cat, petting the cat, making sounds, or moving my hands over the tablet. Furthermore, the game is designed in such a way that it can help the human to facilitate play for the cat. The human can alter things such as the speed, size, brightness, and colour of the moving objects in order to encourage play and try to align the game to relate to the playful behaviour of the cat. Even though the game is designed for both cats and humans, for this study I chose to fulfil the role of human player myself during all play sessions, because of the single focus of this study on user-testing of animals. By including different human participants, the data could have been largely influenced by the differences in interactions with the cats. Furthermore, the game itself could have been played differently by each human, which could create inconsistencies among each testing session. Naturally, a follow up study could include both human and cat play-testers in order to include the evaluation of the game from the human perspective.

(26)

26

6. Research Results

This chapter presents the results of the user-testing sessions. These findings were derived from several iterations of coding and analysis of the video recordings. The results that are presented in this section have a distinct focus on informing design iterations of the tablet game Felino. The general evaluation of the methodology and the specific design iterations that can be derived from these results are discussed in Chapter 7 of this document. The behavioural analysis in this chapter uses the terms interest, extra interest, and rejection. Each of these terms is defined according to the following specific cat behaviour:

- Rejection: looking or walking away, grooming, being distracted, self-petting, and falling asleep. - Interest: looking, sniffing, looking at the side of the tablet, and changing position while looking. - Extra interest: looking at the game while following specific objects, tapping with the paws, killing fish, walking on top of the tablet, sitting on top of the tablet, walking on top of the tablet while playing, and controlling the interface.

1. All the cats that were tested showed an interest in the game at some point during the session and some of the cats showed playful behaviour.

Appendix D shows an overview of the interaction curve for each test session. From these graphs, it becomes clear that all cats showed an interest in the game at some point. Furthermore, in 5 cats (participants 005, 011, 015, 018, and 019) playful behaviour was recognized according to the Burghardt checklist that was presented in Chapter 3.1 of this thesis (see Figure 19).

2. The cats generally showed more interest in the different fish than in the crab.

Qualitative observations of the video recordings suggested that the cats were generally more interested in the fish than the crabs. This was noted because they seemed to follow the movements of the fish more often than the crab. Furthermore, a coding of in-game objects that were tapped by the cat shows a significant difference in killings of the fish versus the crab. Even though both objects were continuously on screen, the fish were killed a total of 63 times and the crab only 11 times. Appendix E shows the exact data of this finding (see Figure 20).

(27)

27 3. The physical interaction of the human with the tablet often generated interest from the cat.

Qualitative observations of the video recordings suggested that the physical embodied behaviour of the human, including playing with the hands, tapping a fish on the screen, ticking with the hands on the tablet, tapping on the water in the game, and moving the tablet, influenced the interest of the cat in the game in a positive way. Appendix F shows a quantitative overview of the frequencies with which the cat showed an interest, extra interest, or rejection towards the game while the human performed different types of physical interactions. From this graph, it becomes clear that the cat mostly reacted positively towards these physical interactions. Only ‘cat_petting’ seemed to generate more rejection than interest. From the observations, this is a logical outcome, because whenever I started petting, the cat mostly moved its attention away from the tablet. All other physical interactions have a relatively low percentage of rejection, compared to the percentage of interest (measured according to the total amount of annotations).

4. The digital sounds that were generated by tapping the fish and the water might cause extra interest in the game from the cat.

The physical human interactions including tapping the fish and tapping the water might create the high amounts of interest showed in Appendix F because of the digital sounds that they caused. Through qualitative observations it was noted that some cats directly looked at the side of the tablet containing the speaker. This might be because they clearly noted the sound (see Figure 21). 5. The cats all accepted and perhaps enjoyed the human presence. They remained calm and sometimes showed affectionate behaviour towards the human.

From qualitative observations of the video recordings, it was found that the cats generally seemed to accept the presence of the human being. Some of the cats did not show any continuous interest in the tablet or walked away from the game, but none of the cats showed any aggression, anxiety, or severe stress related body signals. Furthermore, some of the cats showed affectionate behaviour before, during, or after the test sessions such as rubbing their body against the human or purring, which seemed to indicate that they enjoyed the human’s presence (see Figure 22).

Figure 21: Participant 003 suddenly looking at the speaker after I tapped a fish

Figure 22: Participant 012 rubbing its head against my hand

(28)

28 6. The speed control slider had the highest impact on the cat compared to the sliders for size,

colour, and brightness.

When comparing the interaction with the four control buttons for speed, size, brightness, and colour, it was found that the speed slider had the highest impact (both positive and negative), generated the most interest, the most extra interest, and the lowest rejection. This can be seen in the graphs in Appendix G.

7. From these testing sessions, it remains unclear which of the control sliders was perceived as most positive by the cat.

Appendix G also shows that the frequencies of the human interaction with the control buttons that could be compared with the cat behaviour at the same time are quite low and the percentages of the remaining three sliders (size, brightness, and colour) showed variety in their impact, interest, and rejection. It therefore remains unclear how to further categorize these three forms of interaction. 8. In some cases, the cat controlled the human interface and thereby changed the game-state. Appendix C shows that the cats controlled the interface a total of 14 times during the sessions. With this, the cat closed the game several times. In some occasions the cats changed the colour, brightness, speed, and size of the fish. The cats did not manage to slide the menu up or down (see Figure 23).

9. The cats often showed extra interest in the game or the side of the tablet whenever a fish would swim outside of the screen.

Appendix H shows what happened on the screen of the tablet whenever a cat showed extra interest in the game. From this graph, as well as from qualitative observations, it becomes clear that the cat showed extra interest specifically when the fish moved off the screen. They often looked at the side of the tablet until the fish came back or until they focused their interest on something else (see Figure 24).

Figure 23: Participant 018 controlling the human

interface Figure 24: Participant 019 showing interest in the tablet’s side

10. Several game bugs were found.

During the testing sessions minor bugs were found regarding multi-touch input detection, sound triggers, and crab tail behaviour. These bugs did not cause major influence on the game-play.