UNIVERSIDADE DE LISBOA

Faculdade de Ciências

Departamento de Informática

EXODUS – EXERGAMES FOR UBIQUITOUS SCENARIOS

Paulo Alexandre André Ribeiro

DISSERTAÇÃO

MESTRADO EM ENGENHARIA INFORMÁTICA

Especialização em Sistemas de Informação

2013

UNIVERSIDADE DE LISBOA

Faculdade de Ciências

Departamento de Informática

EXODUS – EXERGAMES FOR UBIQUITOUS SCENARIOS

Paulo Alexandre André Ribeiro

DISSERTAÇÃO

MESTRADO EM ENGENHARIA INFORMÁTICA

Especialização em Sistemas de Informação

Trabalho orientado pelo Prof. Doutor Luís Manuel Pinto da Rocha Afonso Carriço

e co-orientado por Luís Miguel Santos Duarte

2013

Agradecimentos

Gostaria de expressar a minha gratidão ao meu orientador, o Professor Luís Carriço

e ao meu co-orientador Luís Duarte pelos comentários, observações e compromisso

prestados durante todo o processo de aprendizagem e trabalho desenvolvido com esta

tese de mestrado. Além disso, gostaria também de agradecer-lhes por me terem

apresentado ao tópico, bem como pelo apoio dado ao longo do caminho.

Adicionalmente, gostaria de agradecer ao Professor Tiago Guerreiro por todo o

contributo e aconselhamento providenciado no decorrer deste trabalho. Também

gostaria de agradecer a todos os participantes que contribuíram para o meu estudo e

sessões de teste, que voluntariamente compartilharam o seu precioso tempo durante

todo este processo.

Eu devo a minha mais profunda gratidão à minha família, aos meus pais e avós,

que me apoiaram durante todo o processo, tanto mantendo-me harmonioso como

ajudando-me a “juntar as peças”. Estarei eternamente grato pelo vosso amor e apoio

bem como pela paciência com que toleraram o meu egoísmo ao longo do caminho.

Não posso terminar estes agradecimentos sem mencionar o meu mais sincero

agradecimento para todos os meus colegas com quem partilhei estes últimos anos,

procurando arduamente concluir com êxito o curso. Nós derramámos suor, sangue e

lágrimas, mas valeu a pena!

Finalmente, também gostaria de expressar a minha gratidão à FCT (Fundação para

a Ciência e Tecnologia) pelo seu apoio financeiro.

Acknowledgements

I would like to express my gratitude to my supervisor, Professor Luís Carriço and

to my co-supervisor Luís Duarte for the useful comments, remarks and engagement

through the learning process and work developed with this master thesis. Furthermore I

would like to thank them for introducing me to the topic as well for the support given on

the way. In addition, I would also like to thank Professor Tiago Guerreiro for all the

useful input and advice provided during the course of this work. Also, I would like to

thank all the participants in my survey and testing sessions, who have willingly shared

their precious time during the process.

I owe my deepest gratitude to my loved ones, my parents and my grandparents,

who have supported me throughout the entire process, both by keeping me harmonious

and helping me putting pieces together. I will be forever grateful for your love and

support as well as your patience for putting up with my selfishness along the way.

I cannot end these acknowledgements without mentioning my heartfelt

appreciation for all my colleagues with whom I shared these last years, striving to

successfully conclude the course. We shed sweat, blood and tears together but it paid

off!

Finally, I would also like to express my gratitude to the FCT (Fundação para a

Ciência e Tecnologia) for their financial support.

Dedicatória.

i

ii

Resumo

A indústria de videojogos é uma indústria que está sempre em expansão. As

pessoas gostam de jogar porque isso proporciona-lhes momentos de diversão e serve

como uma forma de interação social com outros jogadores. Nos últimos anos, os

avanços de hardware possibilitaram a criação de novos e mais envolventes jogos que

abriram as portas do mundo dos jogos para jogadores não tradicionais e permitiu a

criação de novos sistemas de interação. Exergames, jogos que requerem atividade física,

fazem uso destas novas formas de interação onde normalmente é necessário o uso do

corpo inteiro dos jogadores como um controlador. Este desenvolvimento abriu uma

gama mais ampla de possibilidades e modelos para usar em cenários ubíquos. Fazendo

uso de periféricos e dispositivos de última geração, os exergames podem ter contextos

fixos e móveis, objetivos diferentes e proporcionar melhores interações sociais, através

do uso de diferentes modalidades.

Exergames traduzem-se por um subconjunto de sistemas de esforço, ao qual

denominamos aplicações móveis de esforço (MEA). Estas aplicações têm como

finalidade auxiliar atletas amadores e profissionais durante as suas sessões de atividade

física. As MEA ajudam ainda os seus utilizadores a monitorizar a sua localização

durante as suas atividades de esforço, aproveitando a sua natureza ubíqua e o conjunto

de recursos disponibilizados pelos dispositivos móveis modernos, agora equipados com

sensores GPS, câmaras de alta qualidade ou sensores complementares de saúde (por

exemplo um eletrocardiograma para monitorização de batimento cardíaco). Estes factos

permitem aos programadores e investigadores visionar a criação de aplicações que

oferecem a possibilidade de seguir a localização de um jogador no mundo real. Também

possibilitam a distribuição espacial de vários jogadores por espaços virtuais mapeando a

sua localização real, bem como oferecer suporte a uma variedade de recursos que

permitem aos jogadores analisar o seu desempenho, durante ou após as suas sessões de

treino. Possibilitam ainda a comparação entre os seus dados estatísticos com os dados

estatísticos de outros jogadores. Estes dispositivos também disponibilizam várias

modalidades de input e output fazendo uso de mensagens áudio, feedback visual e

modalidades baseadas em toque que podem ser utilizadas de forma independente ou em

combinação.

Aplicações que requerem esforço físico por parte dos seus utilizadores necessitam

de mecanismos que lhes permitam influenciar os seus utilizadores, motivando-os a

executar as diversas atividades de esforço requeridas, adotando e mantendo os

comportamentos esperados para este tipo de atividades. Para alcançar este objetivo,

diversas técnicas de persuasão podem ser incorporadas nestas aplicações. Geralmente, a

iii

persuasão está presente no nosso dia-a-dia, seja através de anúncios, propaganda

politica, mass media, entre outros. Estas áreas de persuasão operam na tentativa de nos

influenciar ou convencer de algo, o que se traduz no objetivo primordial da persuasão

enquanto conceito: convencer um público-alvo a interiorizar o argumento persuasivo e a

adotar essa nova atitude como perfazendo parte das nossas crenças. Áreas de

entretenimento, como a de jogos, apontam a um público-alvo tendencialmente mais

aberto a elementos persuasivos. Persuasão torna-se numa muito útil ferramenta quando

acoplada com sistemas de esforço. A motivação dos jogadores pode ser influenciada por

elementos persuasivos, implementados neste tipo de sistemas, recorrendo a

contingências de comportamento e atores sociais persuasivos.

Compreender as variáveis que envolvem o comportamento é essencial para a

conceção de técnicas persuasivas. Contingências de comportamento são compostas por

três elementos principais relacionados linearmente (antecedentes, comportamento e

consequências). Antecedentes traduzem-se por estímulos que precedem um

comportamento, ou melhor, que levam a um comportamento específico. Consequências

são o resultado de um comportamento e também dos antecedentes que o precederam,

essencialmente, sendo estímulos que se seguem ao comportamento. A análise desta

relação leva a perceber que tanto os antecedentes como as consequências ajudam na

formulação de determinados comportamentos, o que é a própria finalidade das técnicas

de persuasão.

Um sistema informático é capaz de desempenhar o papel de um ator social

persuasivo convincente, capaz de criar e manter um relacionamento com os seus

utilizadores. Combinando a forma como as pessoas respondem aos computadores,

influências sociais e dinâmicas podem desempenhar um papel significativo com os

fatores de persuasão aplicados a esses meios tecnológicos, especialmente quando estão

envolvidos jogos. Como tal, a pressão de outros intervenientes (colegas ou

desconhecidos) e comparação social (entre outros) torna-se num forte motivador para os

jogadores / utilizadores desses sistemas. Assim sendo, o design destas aplicações deve

procurar refletir a melhor qualidade e repetição possíveis para as diversas sugestões

sociais (social cues) utilizadas, de modo a evitar o risco de deixar os jogadores

aborrecidos ou mesmo zangados.

ExodUS é um trabalho que tem como objetivo estudar as características

persuasivas presentes em exergames e aplicações móveis de esforço no geral. Estas

técnicas podem contribuir para o sucesso comercial deste tipo de aplicações, bem como

para a motivação dos seus utilizadores ajudando-os a obterem melhores desempenhos e

resultados. Estudar como a interação social contribui para a motivação dos jogadores

explorando os casos onde os jogadores competem e / ou cooperam entre si é um dos

iv

objetivos deste trabalho. Além disso, perceber o que promove este tipo de interações é

outro dos propósitos. Dado que o tópico gira em torno de exergames e aplicações de

esforço, torna-se impossível dissociar os aspetos físicos inerentes ao tipo de atividades

incitadas por estas aplicações / jogos. Assim, o uso de dados de esforço, mais

concretamente a forma como são utilizados e transmitidos aos utilizadores, torna-se

num dos pontos secundários abordados neste trabalho. Durante atividades de esforço

tradicionais, é usualmente apresentada ao utilizador informação relacionada com a

atividade específica realizada. No nosso âmbito, este procedimento pode ser replicado,

selecionando cuidadosamente os dados e a forma de como são apresentados ao jogador,

uma vez que pode, potencialmente, influenciar o seu comportamento antes, durante e

após a execução das suas sessões de treino.

Para alcançar os nossos propósitos foi realizada uma análise abrangendo duas

etapas: a criação de um questionário cujo objetivo traduzia-se na formulação de uma

caracterização da relação dos utilizadores com este tipo de aplicações e respetivas

atividades e uma revisão das aplicações disponíveis no mercado. Pontos aos quais

demos especial enfâse prendem-se com a proficiência dos utilizadores com dispositivos

móveis (incluindo as funcionalidades disponibilizadas por esses mesmos dispositivos), a

sua frequência de uso (tanto do dispositivo, como das aplicações em questão) e a

perceção das funcionalidades persuasivas empregues nos produtos comerciais

disponíveis no mercado. O resultado desta análise permitiu a definição de um espaço de

desenho para aplicações móveis de esforço persuasivas, consolidando o conhecimento

existente sobre persuasão e aplicações móveis de esforço.

De modo a validar o nosso modelo de desenho, foi concebido um novo sistema

apoiado sobre uma aplicação móvel de esforço focada em atividades físicas semelhantes

a jogging. Este sistema faz uso de diversas técnicas persuasivas concebidas para estudar

a sua influência na motivação dos seus utilizadores. Uma experiência foi realizada,

recorrendo ao nosso sistema, cujo principal objetivo abrangeu testar os diferentes

fatores persuasivos implementados no mesmo, analisando se os jogadores se sentiam

mais motivados com os modos de jogo disponíveis. Os resultados permitiram identificar

alguns dos tipos de técnicas de persuasão usadas em aplicações móveis de esforço, bem

como provam que o uso destas técnicas pode influenciar a motivação dos utilizadores.

Adicionalmente, podemos também afirmar que aplicações móveis de esforço que não

empreguem qualquer tipo de técnica de persuasão não afetam ou prejudicam a

motivação dos utilizadores.

Palavras-chave: Interface de esforço, Exergames, Computação Móvel, Interfaces

Persuasivas, Entretenimento.

v

vi

Abstract

Exergames translate into a subset of effort systems, which we call mobile exertion

applications (MEA). These applications aid users in tracking their exertion activities by

capitalizing on the ubiquitous nature and feature sets of modern mobile devices.

Applications that require physical effort from their users need mechanisms that allow

them to influence their users, motivating them to perform the various effort activities,

adopting and maintaining behaviors expected for this type of activities. To achieve this

goal, various persuasion techniques can be incorporated into these applications.

ExodUS is a work that aims to study persuasive features present in exergames and

MEA´s in general. These techniques contribute to motivate users to perform better.

Studying how social interaction contributes to the motivation of the players by

exploring cases where players compete / cooperate with each other is an aim. Since this

work revolves around exergames and MEA´s, it becomes impossible to separate the

physical aspects inherent to the type of activities instigated by these applications. Thus

the use of exertion data, more specifically how it is used and transmitted to users

becomes a secondary point of this work.

For our purposes, we performed an analysis whose results allowed the definition

of a design space for MEA´s, consolidating existing knowledge about persuasion and

MEA´s. To validate this design model, we developed a mobile exertion application

system that makes use of several persuasive techniques designed to study their influence

on players´ motivation. Experimentation with ExodUS allowed for the identification of

some of the types of persuasive techniques used in MEAs, as well as to prove that the

use of these techniques can influence the motivation of users. Additionally, we can also

state that mobile exertion applications that do not employ any type of persuasive

technique do not affect or hinder the motivation of its users.

Keywords: Exertion Interface, Exergames, Mobile Computing, Persuasive Interfaces,

Entertainment

vii

viii

Content

List of Figures ......................................................................................................... xii

List of Tables ......................................................................................................... xiv

Chapter 1 Introduction .............................................................................................. 2

1.1 Motivation .................................................................................................. 3

1.2 Benefits of Exergaming .............................................................................. 4

1.2.1 Physical Standpoint Effectiveness ........................................................... 4

1.2.2 Cognitive Benefits ................................................................................... 4

1.3 Goals ........................................................................................................... 5

1.4 Contributions .............................................................................................. 6

1.5 Planning ...................................................................................................... 7

1.5.1 Schedule ................................................................................................... 7

1.6 Document Structure .................................................................................... 9

Chapter 2 Related Work .......................................................................................... 10

2.1 Persuasion Definition ............................................................................... 10

2.2 Persuasion Factors .................................................................................... 11

2.2.1 Contingencies and Behavior .................................................................. 11

2.2.2 Persuasive Models and Technology ....................................................... 14

2.2.3 Persuasion Approach.............................................................................. 20

2.3 Exergame Design ...................................................................................... 20

2.3.1 Design Considerations ........................................................................... 20

2.3.2 Persuasive Driven Design ...................................................................... 22

2.3.3 Design Approach ................................................................................... 27

2.4 Exertion Applications ............................................................................... 28

2.4.1 Exertion Applications Overview ............................................................ 29

2.4.2 MEA´s and Exergames .......................................................................... 29

2.4.3 Bridging Exercise and Entertainment .................................................... 35

2.5 Literature Summary .................................................................................. 36

ix

Chapter 3 A Design Space for Persuasive MEA .................................................... 38

3.1 Assessing Mobile Exertion Applications Usage ...................................... 38

3.1.1 Mobile Technology Usage ..................................................................... 39

3.1.2 Exertion Habits ...................................................................................... 42

3.1.3 Mobile Exertion Applications Usage ..................................................... 43

3.1.4 Dealing with Persuasion ........................................................................ 43

3.1.5 Survey Discussion .................................................................................. 44

3.2 Mobile Exertion Application Design Space ............................................. 45

3.2.1 Summary of MEA Features ................................................................... 45

3.2.2 Defining the MEA Design Space ........................................................... 47

3.2.3 Scenarios ................................................................................................ 52

3.2.4 Design Space Discussion ....................................................................... 54

Chapter 4 ExodUS .................................................................................................. 56

4.1 System Overview ...................................................................................... 56

4.1.1 System Components ............................................................................... 56

4.1.2 System Architecture ............................................................................... 57

4.1.3 Dispatch Server Overview ..................................................................... 58

4.1.4 Databases Overview ............................................................................... 61

4.1.5 Component Communication .................................................................. 63

4.1.6 Web Platform Overview ........................................................................ 66

4.1.7 Mobile Application Overview ............................................................... 69

4.2 System Review ......................................................................................... 86

Chapter 5 Experimentation with ExodUS ............................................................... 88

5.1 Experiment ............................................................................................... 88

5.1.1 Goals ...................................................................................................... 89

5.1.2 Methodology .......................................................................................... 89

5.1.3 Participants ............................................................................................. 89

5.1.4 Responsibilities ...................................................................................... 92

5.1.5 Equipment .............................................................................................. 92

x

5.1.6 Procedure ............................................................................................... 92

5.2 Results ...................................................................................................... 94

5.2.1 Findings .................................................................................................. 94

5.2.2 Anecdotal Evidence ............................................................................. 103

5.3 Experiment Limitations .......................................................................... 104

5.4 Experiment Discussion ........................................................................... 104

Chapter 6 Conclusions & Future Work ................................................................ 108

6.1 Future Work ............................................................................................ 109

Bibliography ......................................................................................................... 110

Annex I Assessing MEA Usage Questionnaire .................................................... 114

Annex II ExodUS Experiment Questionnaire ....................................................... 122

xi

xii

List of Figures

Figure 1 – Three-term contingency relationship ..................................................... 12

Figure 2 – Persuasive Technology Design in Eight-Steps ...................................... 18

Figure 3 – The Dual Flow Model for exergaming .................................................. 24

Figure 4 – Dual Flow Model Exergame Framework .............................................. 24

Figure 5 – The Exertion Framework ....................................................................... 26

Figure 6 – ExodUS Survey mobile device usage .................................................... 39

Figure 7 – ExodUS Survey mobile device provided technologies ......................... 40

Figure 8 – ExodUS Survey participants´ exercise frequency ................................. 42

Figure 9 – ExodUS Survey participants´ reasoning behind exercising activities ... 42

Figure 10 – ExodUS System Architecture illustrating all components and

interactions between them .............................................................................................. 58

Figure 11 – Generic RPC communication Sequence Diagram illustrating a request

from the systems´ front-end / client side ........................................................................ 58

Figure 12 – Server managed Database Class Diagram ........................................... 62

Figure 13 – Mobile App managed Database Class Diagram .................................. 63

Figure 14 – ExodUS Live Broadcast Sequence Diagram example ........................ 64

Figure 15 – ExodUS Website social interface options ........................................... 67

Figure 16 – ExodUS Website Player History ......................................................... 68

Figure 17 – ExodUS Mapping of the Tracked Workout Session ........................... 70

Figure 18 – ExodUS Exertion Data Structure representation ................................. 74

Figure 19 – 2-axis graph representation of a players´ distance metric ................... 75

Figure 20 – 2-axis graph representation of a players´ velocity metric.................... 76

Figure 21 – Macroscopic view of the Data Normalization Process ........................ 77

xiii

Figure 22 – ExodUS Settings Screen showing the Player Profile .......................... 78

Figure 23 – ExodUS Friend / Buddy list Screen ..................................................... 79

Figure 24 – ExodUS playable Activities and Workout Types ................................ 80

Figure 25 – ExodUS Player History Screen ........................................................... 81

Figure 26 – Selectable options for the various workout sessions recorded on the

Player History ................................................................................................................. 82

Figure 27 – ExodUS Player History Data Synchronization Screen ........................ 82

Figure 28 – ExodUS hierarchy of feedback types and respective modalities ......... 84

Figure 29 – ExodUS Experiment participants´ gender distribution ........................ 90

Figure 30 – ExodUS Experiment participants´ exercise frequency ........................ 90

Figure 31 – ExodUS Experiment participants´ reasoning behind exertion activities

........................................................................................................................................ 91

Figure 32 – ExodUS Experiment participants´ reasoning behind their use of MEA´s

........................................................................................................................................ 91

Figure 33 – ExodUS Experiment Location ............................................................. 93

Figure 34 – MEA Presence only impact on player motivation ............................... 95

Figure 35 – Audio performance updates´ impact on player motivation ................. 96

Figure 36 – Co-located Real World Opponent plus MEA Audio performance

updates´ impact on player motivation ............................................................................ 96

Figure 37 – Co-located Real World Opponent plus MEA Audio performance

updates´ impact on player emotional state ..................................................................... 97

Figure 38 – Virtual Partner ‘Ghost’ Challenge impact on player motivation ......... 98

Figure 39 – Virtual Partner ‘Ghost’ Challenge impact on player emotional state .. 98

Figure 40 – Live Broadcast Challenge with the presence of spectators´ impact on

player motivation ............................................................................................................ 99

Figure 41 – Presence of spectators´ impact on player emotional state ................... 99

Figure 42 – Audio Cue effectiveness .................................................................... 100

Figure 43 – Live Broadcast Challenge Audio Cue effectiveness ......................... 101

Figure 44 – Virtual vs. Real Partner impact on players ........................................ 103

xiv

List of Tables

Table 1 - Behavior Qualities and their effects ........................................................ 13

Table 2 – Foggs´ Primary Types of Social Cues .................................................... 15

Table 3 – Features of historical exergame products ............................................... 33

Table 4 – ExodUS survey participants´ age group distribution .............................. 39

Table 5 – MEA distribution in our design space .................................................... 52

Table 6 – ExodUS system distribution in our design space .................................... 86

Table 7 – ExodUS Experiment Scheduled Tasks ................................................... 93

1

2

Chapter 1

Introduction

In the old days people used to play games with each other using physical artifacts

such as cardboards and pieces. Other popular games like Hide and Seek, did not require

the use of items, could be played almost everywhere and did not have a limit for the

number of players. These old traditional games encouraged players to spend time

outside their homes and gather with friends to play. These characteristics lead players to

have fun, exercise and build relationships that otherwise they might have not built. The

introduction of personal computers gave users other possibilities for gaming, such as

racing with each other without being in the same physical space. Older games were

replicated in digital versions able to run on computers and players were able to play

either against an artificial opponent or against other people. But although personal

computers took away the need to physically gather with friends and to go outside to

play with networked games (where players can interact with each other and enjoy

playing together), it still led to the majority of players to a sedentary lifestyle with weak

social relations.

Nowadays people take their smartphones or their tablets and play games wherever

they may be. This fact gives us new windows of opportunity, new directions in the

design of games, especially in games which promote exertion activities. One of the

latest trends in the industry is the concept of exergames. Videogames can be played

through a gaming console or a computer. Players usually interact with them using a

hand-held controller. Exergames consist of digital games in which its results are

predominantly determined by the players’ physical effort, while partially / fully using

their body as the controller. When combined with modern smartphones, these kinds of

games present a high potential for ubiquitous scenarios due to the nature of the

3

aforementioned devices. Ultimately, this can be viewed as a possible solution to

promote healthy lifestyles and gaming beyond the couch.

1.1 Motivation

Many people, especially people that live in cities, lead a sedentary lifestyle. This

originates health conditions that could be prevented or even minimized by exercising a

few hours a day [1, 2, 3, and 4]. Unfortunately, many do not have the time to spend

exercising, either because they are working late or because there are other matters that

take that little time away. Another reason for this nonexistent exercising routine is the

physical discomfort that comes associated with these particular kinds of activities [5].

Among other reasons, in their essence games are played for entertainment which

often provides fun and enjoyment1. They can be played alone or with others

cooperatively in teams or competing against one or many rivals. Player goals while

playing vary from pure entertainment, to gaining rewards or achievements, depending

whether the player is a casual one, an amateur one or even a professional one1. Some

games can affect players, developing their agility and motor skills1, by means of

exercise or by developing their mental prowess, by means of puzzle and number

challenges. As such, games are supposed to have a fun factor associated with them and

this fact engages people to play and have fun. Considering this, games that require

physical effort have the potential to make physical activities more enjoyable, giving

another dimension to simple exercising. Exergames do not take away all of the

discomfort associated with physical activity but they bring a stimulus and engagement

that makes these games perceived as more fun [4]. Exergames provide an excellent set

of advantages that range from contributing to weight loss and physical health to

facilitating social behavior / engagement [3, 4, 5, and 6]. For this particular instance,

these games allow for new and captivating social experiences either by inducing

competitive edge or cooperation between players [5, 7].

1 http://en.wikipedia.org/wiki/Entertainment#Games

4

1.2 Benefits of Exergaming

Exergames consist of digital games in which its results are predominantly

determined by the players’ physical effort and provide the activity needed to counter

sedentary lifestyles and the problems that originate from that lifestyle.

1.2.1 Physical Standpoint Effectiveness

Exergames capitalize on the virtues of both games and physical activity, by

incorporating the fun factor of games in physical exertion. As such, exergames have a

strong goal in granting players with some physical benefit [1, 2, 3, 4, and 8]. Research

shows that in terms of energy expenditure2, exergames prevail over sedentary activities,

effectively increasing said expenditure [8] when movement is involved. Also, factors

like motivation, physical characteristics and experience contribute to the level of

exertion [8]. Energy expenditure is difficult to measure [8] and can be defined as a

combination of internal heat produced in living beings and their physical activity level2.

It is commonly measured in Calories3 (thermochemical calorie energy unit) where 1C ≈

4.184 kJ. However, research also indicates that there is a strong correlation between

energy expenditure and heart rate independently of physical factors such as age, height

and weight [8]. This is relevant because it simplifies the method of measurement for

energy expenditure in studies of exergames and their physical effectiveness.

1.2.2 Cognitive Benefits

Regular physical activity grants cognitive benefits such as improvements in mood,

concentration and memory [2, 4], as well as diminishing anxiety, stress and depression

symptoms [2, 4]. Although studies regarding the cognitive benefits of exergames are

limited, work comparing these benefits between a casual exergame, a standard treadmill

exercise and a sedentary version of the same exergame proved that they can grant

cognitive benefits on a temporary basis [4]. Casual exergames do not have a definite

definition but they should be perceived as engaging, with limited time commitment to

play and easy to establish. In that work, the casual exergame used the Kinect peripheral

2 http://en.wikipedia.org/wiki/Energy_balance_(biology)#Energy_expenditure

3 http://en.wikipedia.org/wiki/Calorie

5

granting a natural interaction approach and allowing for an easy set up procedure. The

sedentary version of the game could be played with a traditional mouse. Results state

that cognitive benefits found on the casual exergame were comparable to exercising on

the treadmill. However the sedentary version of the exergame did not achieve such

benefits. Additionally, the results indicate that the affective state of players improved

after playing the exergame. Another point relating to this work indicates that the

exertion levels provided from the casual exergame and the exercise on the treadmill are

similar but the first is perceived as more fun. This blends with what was stated in the

previous section, as exergames aim to incorporate the fun factor associated with games

and provide suitable energy expenditure when movement is involved [8].

1.3 Goals

The main objective of this work is:

To study persuasive features present in exergames and mobile exertion

applications in general.

These features should contribute to motivate users to perform better [8] and may

contribute to commercial success of these kinds of applications. Secondary objectives

are:

Studying how the use of different modalities in ubiquitous scenarios can

influence social behavior and user performance in exergames and mobile

exertion applications. It is mainly intended to research how these aspects

influence players to engage in these games (competing or working in

cooperation to encourage one another) and how they contribute to single player

behavior.

Studying how to promote these interactions is also an aim. Relevant aspects to

be considered include current devices and their possible contribution to

exergaming and the design of exergames related to target audiences, possible

benefits and disadvantages.

Another objective of this work regards physical exertion data and its handling.

Collecting physical exertion data in-game and carefully selecting which data

should or not be shown to the player might encourage him to try harder. The

6

same might happen if the game shows the other players data as well, in a group

or community game.

The object of study will be an exergame related with jogging. Jogging4 is known to

be a form of trotting or running at a slow pace which allows practitioners to increase

their physical fitness and their life span4, decrease the effects of aging

5 and benefits the

cardiovascular system4. Jogging is typically a sport practiced outdoors by athletes and

enthusiasts and shares a concept of an organized activity that promotes fitness and

sociability. As such, the use of mobile devices is favorable, not only because of their

practicality and proportions which matches well with a physical activity such as

jogging, but also for their computational power. Current mobile devices grant us the

possibility to track the location of a player and spatial distribute many players across

virtual spaces by mapping their real world locations. They also allow for the use of

input and output modalities such as audio, visual and touch-based modalities that can be

used independently or in combination.

In sum, this works research goals are as follows:

RG1 [Primary] – Provide evidence that the use of persuasive cues affects

players´ motivation

RG2 [Primary] – Research how persuasive social cues affect the players´

perceived competitive levels (e.g. stir competition or display cooperative play,

resulting from the used persuasive social cues)

RG3 [Primary] – Understand how the delivery of exertion data contributes to

influence players

RG4 [Secondary] – Study how the use of different feedback delivery

modalities in ubiquitous scenarios can influence social engagement when using

exertion systems

1.4 Contributions

This works´ major contribution will lie in the implementation of an exertion system

(including a mobile exertion application) featuring two specific operating modes, the

4 http://en.wikipedia.org/wiki/Jogging

5 http://med.stanford.edu/news_releases/2008/august/running.html

7

“ghost” mode and the “live” mode. The first consists of a single player variation where

the player can match his performance with a ‘ghost’ jogger. This ‘ghost’ can represent a

previously recorded performance from the player or even from other players. This mode

is inspired with similar features, normally present in video games of the racing genre

where players drive a virtual car around a virtual track and when repeating the course

are confronted with a ‘ghost’ virtual opponent that represents the players´ previous

performance.

The aims of the study with the ghost mode center around its influence on the

player, possibly persuading the player to achieve better performances and gain incentive

to keep playing and exercising. The “live” mode consists of a broadcast function that

allows players to share their workout sessions with friends who play the role of

spectators that can interact with the player in a real time fashion. Studying impact and

influence on social interaction and respective player performance is the main goal

regarding this mode. This contemplates the use of a communication channel that should

make joggers engage in competitive (e.g. by messaging ‘trash-talk’ to other joggers) or

cooperative (e.g. by messaging incentive lines to raise morale) play.

Another important contribution relates to the creation and proposal of a design

space correlating all mobile exertion application related variables with persuasion

factors. The aim of this design space is to consolidate the existing knowledge regarding

persuasion and mobile exertion applications into a single design space, providing a

broad enough coverage of the dimensions related to persuasion techniques employed in

applications of the genre. The channels used to convey that persuasion, the instant the

persuasive cue is created and delivered to the interested parties and the actors involved

in the process are specially emphasized.

1.5 Planning

1.5.1 Schedule

Review of the related work

T1: from 01-09-2012 to 31-10-2012

8

An extensive investigation of the related work regarding exergames and mobile

exertion applications which is the object of investigation was compiled. The purpose of

this investigation was to achieve a satisfactory level of familiarization with all aspects

regarding the theme under investigation, including technology and methodology

applicable to the work in progress.

Writing of the preliminary report

T2: from 01-11-2012 to 31-11-2012

A report containing all gathered information about the projects´ subject was

written. In it is described the state of the art in the area of exergames and mobile

exertion applications and the projects´ goals, challenges and expectations. Also any

work phase already accomplished was included.

Solution development

T3: from 01-12-2012 to 31-07-2013

A prototype of the Exodus exertion system was programmed. This prototype

featured all the designed features. All system components were developed in this

period.

Planning and execution of experiments

T4: from 01-08-2013 to 15-09-2013

Establishment of a series of experiences and trials using the implemented prototype

of the Exodus exertion system was performed. These tests were meant to verify whether

the implemented features achieve the projects purpose of study.

Writing of the Masters´ Thesis

T7: from 16-09-2013 to 25-10-2013

9

The masters´ thesis document was written, containing detailed description of all

development phases of the project, examination of the results obtained and some critical

evaluation of the overall work accomplished.

1.6 Document Structure

This document is organized as follows:

Chapter 2 – Related Work

This chapter will describe the synthesis of the whole process of research on

persuasion factors and design guidelines of exergames and mobile exertion

applications.

Chapter 3 – A Design Space for Persuasive Mobile Exertion Applications

This chapter will elaborate on a survey regarding mobile exertion applications,

specifically how they are used and perceived by people. Also, resulting from the

survey and related literature, a design space accommodating all persuasive

variables regarding these kinds of applications is introduced.

Chapter 4 – ExodUS

This chapter will present the elaborated work, detailing the developed mobile

exertion application system.

Chapter 5 – Experimentation with ExodUS

This chapter will focus an experiment devised for making use of the developed

solution.

Chapter 6 – Conclusions & Future Work

This chapter will present a discussion of the full work accomplished.

10

Chapter 2

Related Work

This chapter will expand upon persuasion factors linked to exergames. The study

will focus on persuasion techniques and instruments that can potentially influence

players to achieve better performances and higher motivation levels.

Being that exergames are the central point in this work, all work related to the

study, design and development of exergames, including persuasion factors employed

with exergames, will be warranted in the next subchapters.

Finally, this chapter will conclude with an extended overview of Mobile Exertion

Applications (from here forth referred to as MEA), bridging the existing conceptual

designs and techniques with our work.

2.1 Persuasion Definition

People are usually confronted by persuasion in their lives. Advertisements, politics,

mass media, among other areas, are always present in our everyday lives as a way to

influence us, or rather convince us of something. This is exactly what persuasions´ goal

is: to convince the target audience to internalize the persuasive argument and adopt this

new attitude as a part of their core belief system6. Persuasion can result in positive

outcomes for the persuasion target (e.g. public campaigns that urge people to quit

smoking or to recycle their garbage), but they might also not favor them (e.g. a pushy

salesman trying to sell you a car)7. Psychologists perceive persuasion as a:

6 http://psychology.about.com/od/socialpsychology/a/persuasiontech.htm

7 http://psychology.about.com/od/socialinfluence/f/what-is-persuasion.htm

11

“Symbolic process in which communicators try to convince other people to change

their attitudes or behaviors regarding an issue through the transmission of a message in

an atmosphere of free choice” [25]

Persuasion exists since ancient times and has evolved since then as well, being

nowadays characterized by the massive number of persuasive messages we encounter

daily (e.g. the number of commercials and advertisements we are faced with every day),

the speed in which these messages are delivered (e.g. through the Internet, television,

etc.), the existence of businesses which work solely around persuasive messages (e.g.

public relations companies or marketing firms), the refined subtlety employed and the

complexity level of the persuasive cues (e.g. careful selection of the persuasive medium

and message in order to stand out from the more diverse choices target audiences are

allowed to choose from).

2.2 Persuasion Factors

The hypothesis to use technology to change peoples´ behavior is the core belief of

persuasion factors incorporated into exergames and games alike. Persuasion techniques

aim to influence the players to adopt certain strategies and actions, coping with an

intended target behavior. To understand this topic, we need to consider how technology

can influence a person and how behavior works.

2.2.1 Contingencies and Behavior

Persuasion techniques have a pre-defined purpose, a target behavior to which the

players should adopt. Understanding behavior is essential when designing persuasive

techniques. Prior reading divides exergame benefits in two different dimensions, one for

psychological outcomes and another for physiological outcomes. The second deals with

aspects relative to exertion and physical activity. The first, however, deals mainly with

abstract variables such as enjoyment and engagement. This last one gains a specific

importance for exergames, since it is directly responsible for its success [8]. Examples

like the Yourself! Fitness8,9,10

exergame can provide insight on this issue. Although this

8 http://www.healthgamers.com/2010/exergaming/the-history-of-exergames/

9 http://gadgets.boingboing.net/2008/05/15/from-atari-joyboard.html

12

game provided high quality physical routines it couldn´t match the Wii Fits´6,7,8,11,12

success, because it was not perceived as fun and engaging7,8

. This leads designers to

seek an understanding of player behavior, which in turn relates to principles and

contingencies that affect players´ emotional states and likelihood of adoption and

maintenance of the games´ required physical activity [9].

Contingencies are present in ordinary games, not just exergames. In order to

perceive them a behavior needs to occur. Contingencies are composed of three main

elements in a linear relationship. These elements are antecedents, consequences and

behavior. Their relationship translates into the following: [9]

Figure 1 – Three-term contingency relationship

This is a temporal relationship that shapes simple or complex behavior.

Antecedents are essentially stimuli that precede a behavior, or rather leads to a specific

behavior. In games, these stimuli can affect the players’ emotional state and the

adoption and maintenance of game play. Consequences are the result of behavior and

also of the antecedents that preceded the behavior, essentially being the stimuli the

follows behavior. Analyzing this relationship we can discern that antecedents and

consequences help in shaping particular behaviors and this is the very aim of persuasion

techniques. Thus persuading using antecedents, signaling target behavior requirements

(e.g. compel players to learn and perform a certain move or play; push for an imitation

or a match of a certain behavior model), and consequences, decreasing the probability

of behavior (e.g. applying penalties) or increasing the probability of behavior (e.g.

positive or negative reinforcement), translate into persuasion techniques´ inner

functioning, coding specific antecedent and consequence elements incorporated within a

game. A specific example of this relationship within an exergame would be playing

10 http://www.todaysfitnesstrainer.com/making-exercise-fun-the-evolution-exergaming-part-two/

11 http://en.wikipedia.org/wiki/Wii_Fit

12 http://www.nintendo.com/games/detail/hoiNtus4JvIcPtP8LQPyud4Kyy393oep

13

Dance Dance Revolution6,7,13

. With this exergame, when a player steps and hits one of

the four arrows present in the respective dance platform (behavior), a stimuli may be

presented by means of an audio feedback mechanism or a graphical representation

(consequence) informing the player of the success or failure in the task executed. One

behavior can result in more than one consequence [9], for instance, if both an audio

feedback mechanism and a graphical representation occur then the behavior resulted in

two consequences instead of just one. Antecedents complete the three-term contingency

relationship and can be identified when, on the games´ on-screen interface the virtual

arrows scrolled upwards approach the “step zone” (antecedent). This translates as a

visual stimuli indicating that the player should step on the corresponding platform arrow

(behavior). Equally to consequences, antecedents can also be of multiple instances.

Finally, behavior is an inherent part of games and exergames (e.g. running, jumping,

dancing…), making up the players´ responses to the game and though interactions with

several consistencies lead to a change in dimensions such as rate and duration, among

others [9]. Games usually influence players´ behavior by permitting certain qualities of

behavior to affect it. Table 1 depicts these qualities (based on the research presented in

[9]).

Table 1 - Behavior Qualities and their effects

Quality of

Behavior Effect

Topography Form of the behavior (e.g. repertoire of paddle movements in a

Table Tennis game)

Frequency Number of times a behavior occurs

Rate Frequency within a specified period

Intensity Strength or force applied in the execution of a behavior

Latency Time between a signal for response and the resulting response

Duration Time from the onset to the offset of the response without

interruption

13 http://en.wikipedia.org/wiki/Dance_Dance_Revolution

14

The three-term contingency relationship can also reinforce behavior [9]. This

reinforcement (intermittent or continuous) occurs based on the aforementioned qualities

of behavior.

“Within each game, multiple levels can exist and each level presents multiple types

of stimuli, rules, and behavioral requirements. Game designers program these stimuli,

their functional effects, and the behavioral requirements for specific purposes, with or

without knowledge of their influence on adoption and maintenance of game play and

emotional states.” [9]

From this topic we learn that contingencies, the stimuli prior to and following

behavior, affect behavior. To reinforce this, carefully looking into the case of casinos´

slot machines will uncover how contingencies can persuade players. These machines are

designed to influence players to adopt and continue their gambling behavior [9]. These

techniques can be used in order to increase success probability for exergames (and

games alike).

2.2.2 Persuasive Models and Technology

Prior reading indicates that behavior and contingencies are intrinsically related with

persuasion elements. Behavior often leads to social interaction. Persuasive elements in

computer systems have originated from the belief that people often respond to

computers as though they are living beings [10]. This fact makes computer systems

suitable to play the role of a persuasive social actor, capable of creating and maintaining

a relationship with its users. An example would be the Tamagotchi14

, a handheld digital

pet simulation game, which consisted in the existence of a digital creature that could

grow, developing depending on the care provided by the player. This gameplay lead

players to interact with the digital pets, as strongly as if they were “real” pets. Some

other examples of such cues applied to exergames exist, such as the use of a virtual

trainer character in the Yourself! Fitness6,7,8

exergame, an exergame which operated

with the sole purpose of guiding the user to perform an exercising routine to improve

their shape with the guidance of MAYA, a virtual trainer character that was always

14 http://en.wikipedia.org/wiki/Tamagotchi.

15

present on-screen and that served as the proxy between the program and the user. The

use of such character implies that the game designers relied, on some level, on physical

(3D human model) and social role (role playing the physical instructor / teacher

authority role) cues. Another example is the successful Wii Fit exergame6,7,8,9,10

and its

use of cute virtual characters that take on social roles of opponents and manifest other

persuasive social cues.

Persuasive Social Actors

Foggs´ work suggests that when perceived as social actors, computer products can

leverage principals of social influence to motivate and persuade [10]. Combined with

the way people respond to computers, social influences and dynamics can play a

significant role in persuasion factors applied to technological means, particularly with

games. As such, peer pressure and social comparison (among others) become a strong

motivator for players / users of such systems. Fogg proposed five types of primary

social cues appliable to computer systems: [10]

Table 2 – Foggs´ Primary Types of Social Cues

Social Cue Example of an element regarding the Social Cue

Physical Face, eyes, body, movement

Psychological Preferences, humor, personality, feelings, empathy

Language Interactive language use, spoken language, language recognition

Social

Dynamics

Turn taking, cooperation, praise for good work, answering questions,

reciprocity

Social Roles Doctor, teammate, opponent, teacher, pet, guide

Persuading using physical cues aims to prompt social presence through physical

characteristics. Physical attractiveness tends to have a considerable impact on social

influence [10]. High-persuasive tasks are often fulfilled by more attractive people (e.g.

salesman, advertising or spokesman), which confirms that more attractive people tend to

16

be more persuasive than unattractive people. Mapping this to technology, we can

assume that technology that is visually attractive is likely to be more persuasive than an

unattractive one.

Although physical cues can prompt social presence this is not necessarily required,

as computer systems can also convey it without using physical cues altogether [10].

Psychological cues are intrinsically related with emotions and personality concepts.

Products that aim to persuade through psychological cues should try to convey simple

cues such as empathy and emotion or / and complex cues that portray personality [10]

(e.g. a computer that keeps crashing may convey a personality of being uncooperative).

Psychology studies indicate that people we think are similar to us in personality,

preferences or in other characteristics can persuade us more easily than those we deem

as not being similar to us [10]. This research provided insight on this matter, concluding

that similarity between computers and users (both with dominant or submissive

personalities) can influence, persuasion wise. Also, the notion of group affiliation (e.g.

teammates) contributes to persuasion through psychological cues.

Language also acts as a powerful tool with persuasion in mind. Delivering

feedback cues, whether they are written text or audio pieces, using spoken language can

influence people to adopt certain behavior (e.g. e-commerce sites use certain language

patterns to persuade users to keep buying). This approach often empowers the systems

with distinct personalities which lead to achieving goals such as acquiring and

maintaining users. This usage of language coupled with the recurring use of praise also

has a strong impact on users´ motivation and openness. When praised, people often

react more positively, resulting in better mood, interactions and attachment to the used

systems [10].

Culture influences behavior, as people tend to adopt certain behavior patterns that

are considered as a norm within their culture. This is also valid for social interaction

(e.g. shaking hands as a way of greeting). Modeling social dynamics is another way to

persuade using technology [10]. A successful application of this type of persuasive cue

are e-commerce sites that model traditional shopping social dynamics “behaving” just

as a traditional clerk would. Systems´ can adopt persistent approaches, as well as exert

peer pressure which ultimately should influence users. A powerful social dynamic is

based on the principle of reciprocity, which states that favors should be returned. Foggs´

work [10] proves that this principle is also replicated when interacting with computers,

17

meaning that if people find the systems´ feedback useful they will, most likely, try to

reciprocate the favor.

Lastly, persuading recurring to social roles can lead to compelling interactions

between systems and its´ users. The Tamagotchi12

is a good example of this persuasion

cue, as the game played the role of a pet which compelled users to take care of it as a

real, live pet. The role of a pet is good enough, persuasion wise, when the target

audience is composed of a young population. Social figures of authority such as police

officer, judges, teachers or doctors have enhanced powers of persuasion as people will

associate them with intelligence and influence. Some applications could also benefit

from other type of social role persuasion, such as the role of a teammate / friend or

opponent for games.

Careful handling of social cues is needed if designers want their systems to be

effective in persuasion [10]. The viability of persuasive social actors is always related

with the target audience and their specifics. Areas such as games and entertainment are

a prime target for the use of such actors because players are generally a target audience

which is more willing to accept and perhaps adopt the kind of social cues these actors

provide. Designers should also search for the best possible quality and repetition of a

social cue, to avoid risking players to become annoyed or even angry.

“When is it appropriate to make the social quality of the product more explicit? In

general, I believe it’s appropriate to enhance social cues in leisure, entertainment, and

educational products (smart toys, video games, kids’ learning applications). Users of

such applications are more likely to indulge, accept, and perhaps even embrace an

explicit cyber social actor — either embodied or not.” [10]

Persuasive Technology Design

Even though much work has been done regarding the use of persuasive factors in

technology, persuasive technologies are a relatively new reality in peoples´ lives and

this reality is changing with the advancing years, with many opportunities for the use of

this kind of technology arousing [11]. However, its design process has a big problem as

it mostly faces failure [11]. Work addressing this issue [11], proposes a process

consisting of eight basic steps that should be performed by design teams wishing to

develop persuasive technology. This method applies to the early stages of development

of such technologies and aims to solve the failure of most of these projects, whose most

18

likely cause has been identified to be because design teams have a goal that is too

ambitious or lack in experience to develop this kind of technology. The process is

described as follows: [58]

Figure 2 – Persuasive Technology Design in Eight-Steps

The first four steps are inextricably linked and might not necessarily be followed in

the order presented in the figure above. They are nevertheless essential to the eventual

success of the project. The first step advises to choose an appropriate behavior for

change and is the most important step. This behavior should be the easiest and smallest

relevant behavior. The second step recommends choosing an easy target audience,

meaning an audience that has demonstrated a desire to change the behavior targeted.

19

These kinds of projects usually fail, especially in its early stages, with selection of a

wrong audience [11]. The third step emphasizes on the issue of determining the reasons

which prevent the targeted audience from performing the targeted behavior. These

reasons normally fall under one of three categories, them being the lack of motivation

(e.g. to perform a physical activity), the lack of ability (e.g. not enough fitness level to

perform a physical activity with a certain intensity that requires a minimum fitness

level) or the absence of a well-timed trigger to perform the response (e.g. an audio

feedback mechanism alerting a jogger to dash during a fixed period). The fourth step

pertains to the choice of a familiar (in relation to the targeted audience) and adequate (in

relation to the targeted behavior) transmission channel, technology wise. The fifth step

states that the design team should search for relevant instances from companies which

are succeeding because their solutions are likely to be effective or in other words, the

designers should find successful examples relevant to their project (having in

consideration the decisions made from the previous four steps). A common issue

regarding this task is that information pertaining these examples might not be accessible

(e.g. companies generally don’t share data with outsiders) so making educated guesses

is a sensible approach [11] to overcome this problem. The sixth step relates to the old

concept of ‘no need to reinvent the wheel’, meaning that after obtaining some successful

examples imitating them is a reliable method to create the projects´ persuasive

technology. Identifying and adapting existing solutions grants benefits, time wise and

reliability wise, but to conduct this step, insight is required. The ability to explore an

existing successful example and extract whatever key components that make up its

effective success is the challenge. Real innovations, like all activities that foresee the

creation of something, require a solid foundation, which can be built recurring to these

“imitations”. The seventh step expresses the need to perform tests on the persuasive

experiences rapidly and repeatedly. Undergoing a series of small, rapid tests is better

than doing a big test [11] and these tests are not meant to be scientific experiments but

rather quick trials that allow for prototyping and to test peoples´ reactions by measuring

behavior. Each test, whether it is successful or not, will result in gain of more insight

into rapid testing and what´s likely to work next time. The final eighth step can be

considered once a small scale success is achieved. This moment defines a milestone and

the project can then be expanded. This expansion can either aim for a new audience,

less adoptive to the behavior and more difficult cases, or it can aim to target a tougher

behavior. This expansion should only vary one or two attributes form the previous

20

success and should be systematic. This step can also be considered for bigger tests, like

academic research, as a starting point [11], unlike the tests performed in the seventh

step.

2.2.3 Persuasion Approach

For this work, persuasion is a tool that should be used to achieve our research

goals. Persuasion techniques are used to mold and make players adopt specific

behaviors which in our case is to increase motivation while achieving better

performances and engaging in social interactions while playing. Programming

contingencies to target such behaviors is encouraged as they help shape our target

behavior. Antecedent stimuli can me conveyed recurring to a set of features ranging

from the use of different modalities for feedback messages to providing tutorials for the

players (e.g. teaching the correct way to run or breathe during the exertion activity).

Consequence stimuli can either be achieved through feedback messages, in-game

challenges and reward / penalty schemes. Social interaction plays a part in this work, as

some contingencies can originate from the use of persuasive social cues. Ultimately, the

design of our product will try to somewhat follow the eight-step design process

proposed by Fogg [11], both for the research phase and the testing phase.

2.3 Exergame Design

Following persuasion factors and their specifics, we need to address their standing

within the design considerations proposed for exergames. Aspects relating to persuasion

factors, social play, interfaces and audience specifics should be regarded in exergames´

design approaches.

2.3.1 Design Considerations

Social Design

It is perceived that, the design of exergames is easier to understand when we

associate them with sports [12], since exergames provide the same kind of experiences,

benefits and disadvantages that sports aim to provide [5, 13]. Sports normally involve

physical exertion and comprise activities that obey a set of rules and oblige practitioners

21

to acquire skills in order to be successful. These activities can originate different

behaviors and physical changes to its´ practitioners depending on the rules and goals of

gameplay. Sport players develop their physiques as they undergo training sessions

gaining skills in the process and engage in social interactions such as competing and

playing against others or cooperatively play with others in a team [12].

Some work suggests that movement in itself can be a trigger to emotional

responses and increase social ties between players [14]. Results hint at an increase in

social interaction when movement is present. Also, the use of a motion-based input

device increases social interaction [15].

Work regarding networked exergames and their influence on social play as been

done [3, 13] and suggest that exergames can facilitate social experiences within these

games. In regard to that work, a sport, Table Tennis, was augmented through the use of

telecommunication technologies. This helps to validate the claim that exergames can be

is easier to understand when we associate them with sports [12]. The results regarding

this work indicate various pointers towards social interaction. The focus of the

gameplay is shifted from the virtual world to the physical space, since most of the

attention of the players is directed at the events occurring in the real world physical

space. As such, this shift in focus introduces several aspects that allow for the

performance of unexpected behaviors that don´t occur in traditional games or are more

difficult to perform (which become new persuasion opportunities). Also, social play

interactions can be affected by previous relations between players [13].

Designing for Specific Audiences

Player needs are different when taking in consideration aspects like age and

physical condition. Usually, games that require physical activity are associated with a

young target audience, boasting their player skills and as such, for their testing it is also

usual for children or adolescents to participate in testing [8]. For older adults [6],

problems pertaining illness, lack in cognitive ability (e.g. deficit in attention, memory

loss) and diminished physical attributes (e.g. poor strength, low energy, less muscle

mass) should to be addressed and because common exergames are designed in a way

that does not address these issues, they result in elderly players´ exposure to injuries.

For these cases and for overall exergaming systems, using adaptive interaction

paradigms seems to be benefic in order to accommodate to players individual

22

differences [6]. Exergaming has been appointed as a way to improve existing

rehabilitation programs and grant other health benefits [4, 6, and 16]. From other works

we can assess that the leading consumer exergaming systems have sufficient speed and

accuracy for such tasks [16].

2.3.2 Persuasive Driven Design

The study of exergames has led to the creation of some frameworks regarding the

design process of exergames. These frameworks give valuable insight on how to create

successful exergames, accounting for generic game related points, including persuasive

factors.

Dual Flow Model

Exergames aim to provide two goals: incentive and physical benefit. However

these goals are tough to obtain simultaneously [8]. Incentive should be perceived as the

capability to retain a sufficient level of interest from the players and motivate them to

keep playing. As such, incentive is difficult to quantify given that every single player

has its own uniqueness, be it in preferences, physical attributes or how experiences are

perceived. But incentive is an integral part of the exergaming concept because it relates

to the fun factor associated with games1, which in turn counters the issue of the physical

discomfort associated with physical activity [5]. It also relates with persuasive factors as

it blends with the previously studied concepts of persuasive social cues and the use of

contingencies to persuade. These aspects tie with the factors that make exergames

succeed such as frequency [8, 17], enjoyment1 and attractiveness [17, 18]. Relating to

frequency, players expect to see results which should only show when the game is

played for an extended period of time thus requiring a firm motivation to continue to

play [8]. Attractiveness deals with the “flow” concept [17, 18], originally proposed by

Csikszentmihalyi´s work which revolves around the level of absorption on the activity

being performed, focus of awareness and actions/skills taken [19, 20]. The concept is

applied in various works, including with games [21]. A modified version of the concept

directed at video games treat the “flow” as “game flow” and lists a few characteristics

which are commonly identifiable in games, and not just exergames, that contribute to

their attractiveness and enjoyment.

23

During the design phase of a game, some characteristics can be explicitly built-in

the game in order to affect the player state [17]. Attractiveness can also be influenced by

the input device features [17], as in order for a player to gain a sense of immersion

while playing, their concentration needs to be focused and can either be set upon the

game itself or on the input device being used. Physical benefit [8, 17, and 18] should be

perceived as the effectiveness of the exercise benefits provided to the players. For the

successful design of exergames, other aspects have to be taken into account such as the

fact that players´ fitness is not static and the ability to deal with the physical exercise

and its intensity, which also varies [17]. These aspects are responsible for the existence

of challenge and skill and contrary to incentive, can be measured.

Based on the previously described aspects of attractiveness (incentive) and physical

benefit, the Dual Flow Model [17, 18] models both with the aim to deliver exergames

that meet a required level of exercise in order to maximize the benefits provide by

physical exercise while maintaining the activity interesting and engaging.

“One of the constructs of flow is the balance between perceived skills and

perceived challenge. An activity needs to be neither too easy nor too difficult.” [18]

Balancing players´ perceived skill versus perceived challenge makes a

psychological model, related to gameplay and aspects of the game where four different

states can be determined. Players lacking in both skill and challenge experience a state

of apathy, while players lacking in one of these variables experience boredom (skill >

challenge) or anxiety (skill < challenge). Likewise, balancing intensity versus fitness

makes a physiological model, related to exercise and physical activity where, again,

four states can be determined. Players lacking in both intensity and fitness experience a

state where no benefit is achieved whereas players with a high level of intensity and low

level of fitness experience failure (inability to continue the physical activity) and players

with low level of intensity and high level of fitness experience deterioration (fitness

level decrease). For both these models, the ideal state is the ‘flow’ state where both,

skill / challenge for the psychological model and intensity / fitness for the physiological

model, are high resulting in more engagement and better fitness through continued

exercise [17, 18].

24

Figure 3 – The Dual Flow Model for exergaming

An exergame created with the Dual Flow Model in mind [18], shows that it is

possible to deliver the necessary level of exercise across players, maintaining a good

heart rate control system. The fact that heart rate seems to be essential to achieve the

models´ purpose is expected [8]. Furthermore, from this work we can identify the need

to create control systems for both intensity and challenge and a few framework

components [18]. These components are responsible for maintaining player

performance history records and planned exercise, as well as gameplay and feedback

methods.

Figure 4 – Dual Flow Model Exergame Framework

25

Exertion Framework

The Dual Flow Model is one of the frameworks available for the design of

exergames. Other frameworks have been created with exergame specifics in mind, such

as the Exertion Framework [5]. The authors also mention persuasion factors with this

framework that aims to describe a design process that uses technology to create games

with more captivating exertion experiences that are mediated by that same technology.

To test this framework several works were used, all having a few common identifiable

characteristics, them being game support for physical exertion, compromise from

players to physical and virtual spaces / objects and favor digitally mediated social play.

In this framework, the human body perspective is ruled by a structured view in the

form of four lenses, which may overlap in some cases: the Responding Body, the

Moving Body, the Sensing Body and the Relating Body.

“A ‘four lens view’ provides sufficient detail and abstraction to analyze new

systems that feature the human body.” [5]

The Responding Body lens [5] refers to the view of the body from "inside",

representing the internal state of the body and its changes during the exertion activity.

The body reacts to physical activity by responding internally in order to maintain

homeostasis (a regulation process by which an organism obtains constancy of its

equilibrium), typically increasing heart rate, more frequent breathing, and occurrence of

sweat (these conditions are not initiated by the person but the person is aware of them).

The body also responds after the exertion activity has ended, resulting in weight loss,

increased muscle mass, torn tissue repair, among others. These changes are typically

persistent and can affect the body's reaction to a particular exercise done previously

when it is performed again. These reactions may be negative (pain, soreness or injury).

These reactions can be exploited as a captivating mechanism for games and do not

necessarily require movement.

The Moving Body lens [5] focuses on the displacement of the body’s muscle parts

in relation to each other during physical activity. Spatiality and temporality are

combined in this field as motion causes a response of the body. Deals with movement

characteristics such as intensity, as movement may have a "weight" associated with it

(equally regarded in the Dual Flow Model), continuity, as motion presents preparation

26

and monitoring phases and variety, which translates into the "richness" of the movement

performed. It also addresses the kinesthetic sense of the body.

The Sensing Body lens [5] resolves around the fact that motion is related to the

way the body understands and acts within the environment where it is inserted, either

virtual or physical. It describes how the body experiences the environment by

interacting with various artifacts in order to play (e.g. ball, bicycle…) and the

environment in which play is undergone (e.g. backyard, stadium, running in a park or

on a treadmill…). It aims to provide a contextual perspective, emphasizing the body's

interactions with the environment. This approach differentiates conventional exergames

from sports since these consist of physical and virtual spaces and objects.

Finally, the Relating Body lens [5] which encompasses the way people and their

bodies are related through digital technology. These social interactions are very diverse,

mediated by a wide variety of roles such as co-players, opponents and audiences. A

social vision about exertion also helps to identify concrete motivation and barriers

related to exercise practice.

These lenses are linked to characteristics similar to those present by the “flow”

concept [17, 18]. These characteristics become relevant because exergames focus

around digital games, which have rules, context and gameplay. These aspects are

established in the formal structure of games, in the experiences of the people involved

and in the larger context in which the game takes place [5].

Figure 5 – The Exertion Framework

27

Particularly, two of these characteristics are closely related to persuasion:

Rules (Awareness of exertion) [5] - players aim to overcome their body´s

limitations and with the introduction of digital technology, designers can choose to hide

or display information regarding the players´ body so that he can benefit from these

actions. The consciousness of expended exertion by the player can lead to comparisons

over time, with self-performance results or with other players’ results which in turn

leads to the promotion of competition and motivation and results in players the

investing more effort. Distractions, such as background music can help to reduce the

awareness of exertion allowing for more enjoyable experiences by hiding the discomfort

of the physical experience.

Gameplay (Expression of exertion) [5, 13] - exertion is a form of self-expression

that highlights the richness and expressive power of the human body. In exergames this

expression can be considered metagaming ("what happens within a game beyond the

game itself" - e.g. trash talking in a game supported by VOIP technology). These

actions require energy expenditure, which may not be favorable to achieve the objective

of the game but contribute to its experience.

2.3.3 Design Approach

These design guidelines are consistent with some interesting points that we intent

to exploit. Sports, in our case jogging, help to obtain a better understanding of

exergames, which imply movement. There are two main details deserving special focus

among these guidelines, necessary to achieve a flow state. Physical associated

characteristics which encompass the handling of exertion data and social play.

Physical condition can change due to exertion activities and players should be

aware of such changes. It becomes a requirement that designers consider

accommodating their solutions to the players’ individual differences and provide them

with information regarding this issue. The ability to provide a player profile and the

display of the exertion data collect through the workout sessions is an effective way to

meet this requirement.

Guidelines directed at social interaction heavily contribute to persuasive factors

applied to these kinds of applications. Incentive is related with motivation which

28

contributes to the subscription and maintenance of the exertion activity. Affecting

players´ emotional state by inducing a competitive edge or cooperative stance (e.g.

prerecorded challenges or workout session exposure to other people), adds to the

immersive feel of the gameplay and strengthens relations through technology.

Shifting focus to the activity at hand instead of the input device used can

potentially originate unintended behaviors (e.g. players performing actions that do not

favor the achieving of the games´ goal) which enriches the experience.

2.4 Exertion Applications

The commitment and will to perform better while working out is a goal which is no

longer the exclusive responsibility of the athlete, with the introduction of key elements

by technology, which boost an individual´s need to surpass themselves. By capitalizing

on the ubiquitous nature and feature sets offered by modern smartphones such as GPS

tracking, high quality cameras or health sensor add-ons, developers and researchers

envisioned the creation of applications which aid users in tracking their exertion

activities. We refer to these tools as mobile exertion applications (MEA). MEA translate

into a set of effort systems from where exergames subset from, assisting amateur and

professional athletes during exercise activities, with a pivotal role being attributed to

persuasive mechanisms. Examples of such applications are Endomondo Sports Tracker

[28], Nike+ [29], Adidas miCoach [30] or Runtastic15

[31]. MEA´s feature persuasive

mechanisms, from tracking our energy expenditure on the move to sharing exertion

results with friends. MEA´s are typically used to support exercise activities, such as

running, cycling, and hiking, among others. They also sport a variety of features for

users to analyze their performance during or after workout sessions as well as

comparing their data with their friends’. Ultimately, these features can be classified as

persuasive elements in a sense that they effectively lead the user to motivate him /

herself for future exertion activities. This subchapter reinforces these statements

resorting to an in-depth analysis of the various mobile exertion applications available

and their usage. Furthermore, a broad analysis of commercial exergame systems will be

addressed.

15 http://google.about.com/od/googlereviews/fr/runtastic-app-android.htm

29

2.4.1 Exertion Applications Overview

Persuasive elements are afforded by MEA´s by employing a variety of features

such as challenge modes, cooperative play, music playing and audio message delivery

which are based on the users’ efforts during workout. These applications sport a

significant feature set, allowing the users to configure options such as the course they

desire to tackle or defining personal challenges. Among these characteristics, the usage

of persuasive mechanisms is a common trait – if the application is not able to motivate

users into exertion activities, then it failed its purpose. We will now delve into the

specificities of some of the most popular MEA´s, considering aspects such as features

included and, most importantly for this research, how persuasive elements are

intermingled with the application itself.

MEA are tools that grant the user ways to keep track of their exertion activities,

recording their statistics and progress over time. They also possess a widespread social

context, allowing users to share and compare their results and granting tem the

possibility to interact with one another competitively, by defining challenges, or

cooperatively, by making possible the creation of teams and clubs. These applications

often work in conjunction with their respective web platforms, which grant users with

more features and more powerful functionality than those that are available in their

mobile devices. But focusing on persuasive features, these mobile tools gain special

relevance since they accompany users during their activities - they are literally with the

user. This fact makes them the first choice conduit for persuading the users, generating

some impact on their motivation, giving incentive to continue with the exertion

activities and to achieve better performances.

2.4.2 MEA´s and Exergames

Taking advantage of the new smartphones and tablets, exergames have recently

experienced a major trend increasing their reach to outdoor contexts and activities,

instead of the traditional indoor context (e.g. play at locations such as arcades, fitness

centers and schools or simply at home).

Endomondo Sports Tracker

Endomondo Sports Tracker [28] is one of the most popular MEA´s. It features a

variety of options: for instance, in the “Beat Yourself” mode a player can choose a

30

previous recorded workout as a challenge; in the “Beat a Friend” a user races against a

friend’s performance; in the “Calorie Goal” the players are able to set a calorie value to

burn in the subsequent physical activity. Users are able to establish workout routes via

GPS coordinate input.

As far as persuasion is concerned, Endomondo encompasses implicit and explicit

persuasive cues. The first concerns the usage of challenges which entice the user into

performing better by:

a) Beating his / her former performances;

b) Or beating a friend’s performance.

Explicit persuasion assumes the form of a feature labeled as “Audio Coach” by the

developers. Through it, users are informed about their performance in real time (e.g.

whether they are burning more or less calories at a certain point in the workout session

compared to previous ventures).

Cardio Trainer

Similarly to Endomondo, Cardio Trainer16

is an application that plays music amidst

a workout session. Here, users are able to establish their preferences regarding this type

of explicit persuasive feedback. In particular, they can define whether the informative

messages are delivered on a time basis or based on the distance covered.

Cardio Trainer capitalizes on GPS to track the users’ courses but also enables the

utilization of pedometers. The application tracks distance covered, workout time and

based on these variables it calculates the calories burned. As for the usage modes, the

emphasis goes to the “Race against Yourself” feature which mimics Endomondo’s

“Beat Yourself”, enabling users to challenge themselves and challenge their previous

performances.

RunKeeper

RunKeeper17

[36] shares most of its features with the two previously described

MEA´s. It provides a richer workout summary to its users, namely displaying top speed

information along with the travelled distance and elapsed workout time.

16 http://google.about.com/od/googlereviews/fr/cardio-trainer-android-app.htm

17 http://google.about.com/od/socialtoolsfromgoogle/fr/The-Runkeeper-App-For-Android.htm

31

However, its differentiating characteristic is the possibility of sharing the user’s

workout data and all GPS tracked activities into the application’s website. With the

user’s permission, this data is then shared among the community allowing for spectators

(individuals who may not be partaking in the workout activity themselves) to track the

workout session.

Nike+

Nike+ [29] sports most of the features already covered in the previous MEA´s. One

of its innovations pertains to the introduction of another form of persuasion: power-

songs. According to the developers, power-songs are pieces of music which due to their

characteristics (e.g. emotional affinity the user possesses towards it, high rhythmic

nature of the song itself, etc.) are capable of inspiring and motivating users into

pursuing better performances. In Nike+, users can create playlists and label some of

those songs as power-songs. While exercising, the user is then capable of touching a

button in the application which stops the song being currently played to start

reproducing one of the available power-songs.

Keep Running

Keep Running18

, one of the MEA´s mentioned in our survey, possesses a more

specialized nature: its goal is to ensure users maintain a regular speed while running.

Like previous entries, Keep Running allows users to listen to music while exercising.

This plays a key role in what concerns the user’s motivation and on how the application

persuades the user. Prior to engaging in the exertion activity, users can establish speed

thresholds for their desired performance. The application is then able to monitor the

user’s speed and, if his / her speed falls outside the boundaries of the said thresholds,

the application forcefully stops the music being played. Unlike other MEA´s, this

strategy employs an approach which is not based in natural language cues, but rather on

the usage of coded hints concerning the user’s performance.

Other Applications

Choosing which MEA to use is just not a matter of the feature set it provides:

personal preference for the interface, social pressure and how they are marketed factor

18 http://google.about.com/od/socialtoolsfromgoogle/tp/free-jogging-running-android.htm

32

into this decision. Up until this point we addressed MEA´s which showed unique

features by themselves. Yet a few more applications than those listed exist.

For instance, Runtastic15

[31] covers most the already mentioned features.

Additionally, it provides information regarding altitude and allows users to manually

configure their workout sessions from a list of predefined cardio-exercise templates.

Competitive challenges (against the user him / herself or against friends) along with live

workout monitoring are some of the features also present in this MEA. Strava [32] is

another application which shares most of the characteristics of other MEA´s.

Additionally, and given the user possesses a small device capable of measuring his / her

heartbeat rate, the application allows for the definition of a “suffer score”. These are

thresholds which indicate periods in which the user is overexerting him / herself.

Non-MEA Exergames

Some exergames have been developed, building upon MEAs´ features and their

typical usage. However, these exergames expand on MEAs´ characteristics, giving them

game related aspects, which in turn forcefully include other dimensions related with

games. Zombies Run [33] exemplifies an exergame which was built around MEAs´

characteristics. It provides tracking for running activities, measuring distance, duration,

average speed, average pace and calories burned, much like other MEA. It even

provides training programs. Players can analyze statistics regarding their performances

and share those statistics with other players. While playing, they are guided by audio

instructions and are accompanied by music and power songs. All these features are the

same as the ones present in other traditional MEA. Then what is the difference? These

exergames are heavily focused in gaming mechanics such as storyline and game

objectives (e.g. completing missions or assignments), which other MEA cannot offer.

Thus these exergames serve as an expansion of MEA and their capabilities.

Exergame Systems

Exergames possess a rich history which describes how they have been

conceptualized and played throughout time. Existing exergame systems can be coupled

in three major trends identified throughout the years, one of them being MEA´s. The

other trends pertaining exergames relate to the existence of specialized hardware and

home console exergaming. The former consists in the development and use of specific

hardware peripherals that aim to introduce new ways to interact with games and provide

players with new and more interesting gaming experiences. Systems related to this trend

33

encompass concepts like the use of exercising bikes, step machines and traditional

arcade machines. The latter consists in the development of exergames for contemporary

home video game console platforms such as Sony´s PlayStation19

. These games allow

for different experiences than those that traditional videogames provide. The identified

trends can mingle with one another. This fact is reinforced by systems which make use

of devices capable of detecting player movement - the PlayStation Move6,20

and

Microsoft Kinect6,21

.

From the studied exergame systems, all offer different characteristics in terms of

the number of players allowed to play (single vs. multi vs. community), the nature of

the location for playing them (indoor vs. outdoor) and the systems´ orientation (device

vs. activity). These characteristics are relevant because they can help us understand

what kind of playing modes are available (number of players), the ubiquity of such

systems (location) and their level of effectiveness / performance (orientation). The

following table resumes the systems taking into consideration these characteristics:

Table 3 – Features of historical exergame products

System Number of Players Location Orientation

Single Multi Community Indoor Outdoor Device Activity

Joyboard7,22

× × ×

Puffer7,23

× × ×

CompuTrainer7,24

× × × ×

Power Pad7,25

× × ×

Power Glove7,26

× × ×

19 http://en.wikipedia.org/wiki/PlayStation_%28console%29

20 http://en.wikipedia.org/wiki/PlayStation_Move

21 http://en.wikipedia.org/wiki/Kinect

22 http://en.wikipedia.org/wiki/Joyboard

23 http://www.atarihq.com/othersec/puffer

24 http://www.racermateinc.com/computrainer.asp

25 http://en.wikipedia.org/wiki/Power_Pad

34

Tectrix VR7,27

× × × ×

Prop Cycle7,28

× × ×

Dance Dance

Revolution6,7,29

× × × ×

Exertris Interactive

Exercise Bike6,30

× × × ×

GameBike6,7,31

× × × ×

Yourself! Fitness6,7,8

× × ×

Bodypad6,7,35

× × ×

Gamercize6,7,35,32

× × ×

Wii Fit6,7,8,9,10

× ×* × × ×

EA Sports Active6,33

× ×** × × ×

Nexersys35

× × ×

Move6,24

× × × ×

Kinect6,25

× × × ×

Endomondo × × × ×

Cardio Trainer16

× × × ×

Runkeeper17

× × × ×

26 http://en.wikipedia.org/wiki/Power_Glove

27 http://tulrich.com/tectrixvr/

28 http://www.arcade-museum.com/game_detail.php?game_id=9139

29 http://en.wikipedia.org/wiki/Dance_Dance_Revolution

30 http://www.gizmag.com/go/2123/

31 http://www.todaysfitnesstrainer.com/making-exercise-fun-the-evolution-exergaming-part-two/

32 http://www.gamercize.net/gamercize.htm

33 http://en.wikipedia.org/wiki/EA_Sports_Active

35

Runtastic15

× × × ×

Keep Running18

× × ×

* - playing simultaneously, each player with different consoles or one at a time

** - playing two at a time, using a different set of controllers each

Analyzing this table, we can state that most exergame systems explore

characteristics such as single player modes, indoor location and different orientation

styles (with a slight advantage to device orientation). Regarding the ‘number of players’

characteristic, multiplayer seems to be available in systems where an exercising bike or

a home video game console is present. Also, community based exergames are few in

number and commonly present in mobile exergames. MEA´s imply outdoor location in

contrast to every non-mobile system which is indoor. Finally, systems whose aim is to

provide physical activity (especially those that include the fun factor) are not dependent

on hardware and appear to be more successful. One explanation might be that device

dependent systems have risks of unsatisfactory performance and acceptance from

players.

2.4.3 Bridging Exercise and Entertainment

The popularity of MEA´s has propelled designers and researchers into identifying

even more ways to motivate and persuade users into exercising. The existing trend of

serious games was a decisive factor to identify how entertainment aspects could be

integrated in MEA´s in order to provide users with a more fun and appealing

experience. This category of MEA´s present videogame characteristics, such as a

compelling storytelling, to motivate users into exercising.

It is important to state that we do not cover videogame related elements of

persuasion (e.g. compelling storylines, game oriented tasks such as chasing after virtual

or real enemies). However, we are still interested in analyzing existing persuasive

elements which are related with MEA´s in general.

Since the object of study for this work relates to a mobile exergame the previous

analysis, supports our choice in testing an exergame aiming to incorporate

characteristics such as featuring a multiplayer mode, with outdoor orientation and the

overall system not being device dependent but activity-oriented. From all the collected

36

information from these systems, some details are worth a further mention. Most mobile

applications, akin to exergames, share common features and platforms. Also, recording

a history of the players’ workout sessions is heavily used in these applications, allowing

players to keep track of their performances and routes. Finally another feature

commonly shared among these applications and some other systems is the existence and

maintenance of a player profile which helps with a multiplayer feature and the

aforementioned player history log.

2.5 Literature Summary

Persuasion is used to influence people to adopt some attitudes, translating them into

specific behaviors. Entertainment areas, such as gaming, target audiences that are more

open to persuasion. MEA´s and exergames in particular, are physically demanding and

require high motivation and commitment levels from their players to achieve success.

Persuasion becomes a very useful tool when used in these systems. Motivation can be

influenced by persuasive elements implemented into the designed systems using

antecedents and consequences of behavior and persuasive social actors. These elements

are present in the majority of the reviewed systems, especially in MEA´s.

MEA´s are the latest trend in exergaming, capitalizing on their ubiquitous nature,

available playing modes, sport association and social interaction ability. These systems

are ideal to provide ways to motivate and persuade players to maintain and improve

their exercise sessions.

37

38

Chapter 3

A Design Space for Persuasive MEA

In this chapter we propose a design space for persuasive MEA´s. The respectable

number of products of this type available in different mobile application stores and the

continuous interest of researchers in exploring this area testify its importance. However,

to our knowledge, no one has attempted at consolidating the existing expertise about

persuasion and MEA´s into a design space. It is our goal to provide a broad enough

coverage of the dimensions related to persuasion featured in MEA´s. Special emphasis

is given to the channels used to convey that persuasion, actors involved in it and the

instant the persuasive cue is delivered to the interested parties. The process of creation

for this design space encompassed two steps: the creation of a survey which aimed at

assessing MEA usage habits and how useful individuals perceive the persuasive features

to be; a review of existing MEA´s with the intent of identifying the categories and

dimensions which would comprise the design space.

3.1 Assessing Mobile Exertion Applications Usage

Taking into account prior reading regarding MEA and our focus on motivation and

persuasive factors in MEA´s, we distributed a questionnaire online [see Annex I]. The

goal was to assess:

i) Under which conditions MEA´s are used;

ii) The popularity of these applications among individuals who exercise regularly;

iii) Their perception about persuasive elements;

iv) How these features affect their exercising practices or motivate them to pursue

further goals.

39

The questionnaire was answered online and disseminated through a variety of

channels (namely social networks, internal mailing lists, or by direct request). A total of

194 participants responded to the call (table 4). Most subjects are young adults between

18 and 35 years old (80% of the subject base), with a gender distribution of 56% males

and 44% females. The results are presented and discussed in the following subsections,

reflecting the topics addressed in the questionnaire and the features identified from

existing products and research.

Table 4 – ExodUS survey participants´ age group distribution

Age Group Number of participants Percentage

17 or younger 0 0%

18 to 25 86 44%

26 to 35 69 36%

36 or older 39 20%

3.1.1 Mobile Technology Usage

Smartphones were the most used mobile device by our participants, with more than

half using one (56%). As for other mobiles devices, participants also used tablets and

basic phone(s) (21% for each device) and more specific devices (2%) such as Amazons´

Kindle34

, Apples´ iPod Touch35

, Garmin’s´ GPS devices36

, Pebble Technology’s Pebble

E-Paper Watch37

or MP3 Players.

Figure 6 – ExodUS Survey mobile device usage

34 https://kindle.amazon.com/

35 http://www.apple.com/ipod/

36 http://www.garmin.com/en-US

37 https://getpebble.com/

40

In order to run mobile exertion products, these devices need to provide a few set of

technologies. Most of these applications heavily use the devices´ integrated GPS sensors

and connections to the Internet. As such, we asked our participants if their devices

provide such technologies. 136 participants stated that their devices provided them with

GPS functionality (42%). As for Internet connections, 150 participants stated that their

devices had 3G telecommunication networks (46%). As for other connections, 2

participants stated were able to use Wi-Fi connections and 1 participant made use of the

more recent telecommunication network 4G/LTE. 1 participant also stated that his

device provided him with NFC technology. Furthermore, 34 participants stated to not

have access to any of these technologies (10%) with most of them using only basic

phone(s) (91%). Finally, smartphones and tablets were the devices with the most

answers that provided users with both GPS sensors and 3G telecommunication

networks, with 123 participants stating their smartphone provided them with both

technologies and 46 participants stating the same for their tablets. This makes

smartphones and tablets suitable devices for running mobile exertion applications.

Figure 7 – ExodUS Survey mobile device provided technologies

Next, we asked the participants about their usage of their mobile devices,

particularly with what frequency they used their devices for communication, social,

productivity, entertainment and navigation:

Communication

As expected, all participants used their mobile devices for communication purposes

with the most typical frequency being that of several times a day (44%).

41

Social

Regarding Social purposes (Facebook, Twitter, Google+…), 39 participants stated

to never use their devices for this purpose (20%) with 28 participants using a basic

phone (71%), only 9 participants who use a basic phone have 3G telecommunication

network access and only 18 participants who use a basic phone have no internet

connection. Most participants stated to use their devices for this purpose with a

frequency of several times a day (29%).

Productivity

In terms of Productivity purposes (calendar, notes…), almost half of the

participants stated to use their devices for this purpose with a frequency of several times

a week (48%), with most of them using smartphones (68%). Interesting to note that

from those participants, basic phones are used by 34 participants, while only 18 use

tablets. Also, 17 participants stated to never use their mobile devices for productivity

purposes (9%), with most using basic phones (59%).

Entertainment

Most participants stated to use their devices for Entertainment purposes (Games,

YouTube…). 87 participants stated to have a typical frequency of several times a week

(45%) and 51 participants stated to have a typical frequency of several times a day

(26%). Yet, 44 participants claimed to never use their mobile devices for entertainment

purposes (23%). The bigger cluster (participants who stated to have a typical frequency

of several times a week) used more smartphones (87%), followed by tablets (28%).

These participants also claimed to use basic phones and a Pebble smart watch. As for

the cluster of participants who did not use their mobile devices for entertainment

purposes, the majority used basic phones (77%).

Navigation

The usage of the mobile devices for Navigation purposes (maps…), was greater in

a frequency of several times a week (62%) with 120 participants stating so. 57

participants declared that they did not use their mobile devices for navigation purposes

(29%). Of the 120 participants that declared a several times a week frequency in using

their devices for navigation purposes, 80% had GPS and 3G as a provided technologies

from their device, while 10% had only 3G as a provided technology from their device

and approximately 7% had only GPS as a provided technology from their device.

42

3.1.2 Exertion Habits

Most participants led a good relationship with exertion activities, with 51%

reporting to at least exercise 2 or 3 times a week. Out of all the participants, only 20%

reported lesser exercise frequency, with 6% only exercising at least once a month, 14%

exercising sporadically. Only 8% stated to not exercise at all.

Figure 8 – ExodUS Survey participants´ exercise frequency

The average time spent per workout sessions was between 1 and 2 hours (65%).

This trend is also verifiable considering the total number of participants where 66%

spent between one and two hours in their workout sessions, while only 6% spent more

than two hours on their workouts. The favored sport practiced by the participants was

Running / Jogging (38%). As far as social habits are concerned, 42% of the participants

stated to exercise by themselves, while 32% stated to doing it in the company of a friend

or colleague and 26% within a group. As for the reasons behind their exercising, the

indicated motivation was to “keep fit” (31%), followed by “as a hobby” (20%) and to

“deal with stress” (19%).

Figure 9 – ExodUS Survey participants´ reasoning behind exercising activities

43

3.1.3 Mobile Exertion Applications Usage

When asked how familiar they were with MEA´s, most participants stated to have

no knowledge of them (38%), while others knew about them but did not use them

(32%). Only 59 participants stated that they either have used them (16%) or use them

regularly (14%). For these, Endomondo Sports Tracker [28], Nike+ [29] and

RunKeeper17

[36] were their MEA´s of choice. Some users preferred other non-

mainstream applications such as Strava [32], GetRunning [34] or FitBit [35].

Participants with MEA experience use these products because they display

personal progress (40%) and because they are able to motivate users to do better (32%).

Other reasons chosen by participants ranged from “It makes exercising more fun”

(15%), “It allows me to share my results with other people” (8%) and “It allows me to

compare my results with other people” (8%), to more specific reasons such as “It allows

me to keep a log of my exercises”, “Allows me to see info about my tracks” and “Turns

it into a game”. For MEA power users, the main reason to use the software is the ability

to motivate them (36%), followed by the ability to show the users´ progress (34%). We

were also interested in finding how often participants used MEA´s. For those that stated

to either use them regularly or to have used them, the typical frequency was of at least

once a week (15%).

3.1.4 Dealing with Persuasion

Finally, we tried to assess the impact of motivational strategies which may or may

not be present in MEA´s, but which users recur to, prior to or during workout activities.

For our purposes, we disregarded participants who stated to either having no knowledge

of such strategies or that they have never used them before. We also looked into

persuasive cues usage frequency. The following list covers the most common

persuasive approaches:

Performance Messages – The presence of performance messages containing

information about the user’s progress made 57 participants feel more motivated (29%).

From the 51 participants that answered as such, 27 stated to use it frequently (54%) and

17 stated to use it rarely (29%).

Music Players – Listening to music while working out is a very common feature

used in mobile exertion products. From this study, 74 participants stated that using

44

music player features made them feel more motivated (38%). 63% of the users use

persuasive music features frequently.

Pep Talks – This surveys´ results suggest pep talks were the least popular

feature regarding the impact on user’s motivation. From the 92 participants that replied

to having engaged in pep talks prior to or during a workout session, 28% referred to

having felt more motivated for the exertion activity. Of these, 42% stated by engaged in

pep talking with their peers and adversaries frequently.

Competing against yourself – The existence of competition against oneself

proved to be the feature with higher positive influence on the participants´ motivation.

From the 108 participants that replied to this strategy having some influence on their

motivation, 78% stated they felt more motivated. Of these, 49% compete against

themselves frequently. Competition against oneself is a natural benchmark for simple

performance comparisons. MEA´s often call this feature “ghost” mode since a user is

challenging a “ghost” recording of him / herself or of another person in the past.

Competing against others – 95 participants responded they use another person

as a benchmark for their workout activities and it has some influence on their

motivation. 51% of these participants perceive this strategy as being a positive

persuasive approach, increasing their motivation. Of these, only 28% recurred to this

strategy with some frequency.

3.1.5 Survey Discussion

From this sample we can presume that smartphones are the preferred mobile

device, followed by tablets. These devices seem very compatible with exergames and

MEA´s, not only because they provide the necessary technologies for their use, but also

because of their practicality, proportions and computational power.

People use their devices with a higher frequency (several times a day) for

Communication and Social purposes. For Productivity purposes, participants have a

lesser frequency of usage (several times a week). Most use their devices for

Entertainment purposes several times a week (also, many use their devices for this

purpose several times a day), which is not a surprise considering the preferred device is

a smartphone. If the device is a basic phone, Entertainment purposes are discarded.

45

Finally, for Navigation purposes, the typical frequency of usage is several times a week

and mainly with devices that provide both 3G and GPS technologies.

Only a few participants did not partake in exertion activities. The most typical

frequency is that of 2 or 3 times a week, spending between one and two hours in each

workout session. The favored exertion activity was Running / Jogging. As for the main

reason for exercising the most typical answer is to keep fit. Most people claimed to

exercise alone, which blends well with the favored sport and typical frequency.

MEA´s were mostly unknown to the participants or they were known but not used.

Top reasons for their usage were because they show the users’ progress and because

they motivate users to do better. For participants that stated to either use these products

regularly or to have used them, the typical frequency of usage is / was of at least once a

week (fits with the typical frequency to execute exertion activities). For participants

who stated that they either not know any of these products or that they knew about them

but did not use them, when asked how likely they were to use these products in the next

six months, answered mostly with very unlikely.

Finally, regarding motivator features and their influence on users, the listed

motivator features had, overall, some positive contribution to the users’ motivation.

Features that made users feel more motivated or greatly motivated are often used

frequently. Competition against oneself was the most effective motivator feature, while

pep talks were deemed the most ineffective one (unexpectedly), followed by the Sharing

feature.

3.2 Mobile Exertion Application Design Space

Following all our research on MEA´s and our survey results, we formulated a

theoretical design space, with special emphasis on persuasion features, for application

of this genre. This subchapter will introduce that design space.

3.2.1 Summary of MEA Features

In light of our research of some of the most popular MEA´s available, we will

summarize the primary persuasion related features we were able to identify. This

46

process will fuel the conceptualization of a set of dimensions which will comprise the

MEA design space we propose.

One of the aspects we identified is the entity a user challenges during the exertion

activity. This entity is a clear persuasive element which increases the commitment of a

user into beating his / her previous performances. Two entities were identified: the user

him / herself and another user as an opponent. Endomondo Sports Tracker [28], Cardio

Trainer16

, RunKeeper17

[36], and others allow the user to record his / her workout

settings. These recordings can then be used as a benchmark for future exercise sessions,

effectively assuming the role of a challenge. In addition to being able to challenge

themselves, it is also possible to pick a challenge from another user (a friend, for

instance) and attempt at beating that performance. Of course we are required to address

an important aspect of these challenges: their location. While it is common for users to

engage in exercise sessions in the same vicinity, the possibility of picking a challenge

which took place in a totally different place exists. MEA´s adopt certain approaches to

tackle this issue: they typically use location-independent metrics such as calories

expenditure based on distance, exertion time, altitude, among other variables, to

compare user performance. In short, MEA´s enable users to tackle geographically co-

located or distributed challenges using their own performances or a third party’s as

benchmark.

Still related with the entities involved in a workout session, some MEA´s allow

users to share their data online. A new type of entity partakes in the activity assuming

the role of a spectator. This individual is able to monitor the user exercising and, in

some cases, cheer for him / her, effectively creating a persuasive channel between both

of them.

Another type of feature which emerged clearly when reviewing existing literature

is the way persuasive workout feedback is conveyed to users. Natural language is a

common element in almost all of the studied MEA´s. When the user is failing to attain

his / her goals of is clearly surpassing previous marks, the application typically issues an

auditory message using natural language to motivate the user. Music also has a strong

presence in these applications. The usage of personalized playlists appears to be a

common denominator across MEA´s. Some extend this feature to include power songs –

pieces of music which, through emotional affinity or their intrinsic rhythmic nature are

able to reach the user from a motivational perspective. Other less clear approaches are,

47

for instance, employed by Keep Running18

, in which musical interruptions are

employed to metaphorically signal loss of performance. The usage of applauses is also

gaining momentum given the connectivity with social networks and the presence of

spectators monitoring the workout sessions.

One last element is intermingled with every other one already discussed, without

emerging as a notorious one – time. Although the presence of this concept was not

evident in the studied MEA´s, almost all of the discussed features contemplate it under

various forms. For instance, partaking in a challenge is related with time: a user chooses

to tackle a workout session which took place previously and was ether performed by

him / herself or another person. Time is also present in other entities’ activities. For

instance, a spectator can cheer a user while he / she is working out (i.e. persuasion in the

same time frame), prior to the workout event taking place (i.e. pre-emptive motivation)

or after it (i.e. acknowledgement of the user’s abilities and motivation for future

sessions). The time concept is highly complex and is in fact related with every other

concept discussed in this summary. Next, we will address our design space proposal for

MEA´s, taking into account most of the discussed features in this section.

3.2.2 Defining the MEA Design Space

The design space for persuasive MEA´s is a macroscopic view on the general

dimensions which encompass persuasive and motivation factors for such applications.

For this design, we built on the valuable input provided by our survey’s participants and

on the review of existing MEA´s and literature in this topic. We aimed at identifying a

set of design space dimensions through the elicitation of features and requirements from

existing MEA´s, state-of-the-art research in the area and from the short survey we

performed. Ideally, these dimensions should be broad enough to cover the MEA

spectrum of persuasion possibilities while simultaneously it should also be clear in

understanding how, when and where persuasion connects different stakeholders. It is

important to note that this design space’s point of origin is always the runner him /

herself (as everything else is describe in relation to him / her).

Delivery

MEA´s sport a variety of mechanisms to deliver persuasive messages: pre-recorded

speeches or step sounds are two examples of this approach. The categorization we were

48

able to abstract is related with the perception of the content being delivered – for

instance, if natural language is included or if abstract and rhythmic sounds are

employed. Three categories were identified:

[1] Natural Language

This category includes persuasive cues delivered using natural language (i.e.

content clearly perceptible by the user without the need of recurring to metaphors or

other mnemonics). Persuasive messages (e.g. audio or textual descriptions such as

“you’re doing great”, “you’re catching up on your last performance”, etc.) or music (e.g.

songs which contain subliminal or explicit persuasive content) are examples of natural

language persuasion in the delivery dimension.

[2] Rhythmic Patterns

Existing MEA´s often modify a song’s tempo to make it sound accelerated or

give the impression of being slowed down. This mechanism is known to be an effective

persuasive mechanism based on the alteration of rhythmic patterns. This mechanism is

known to be an effective persuasive mechanism based on the alteration of rhythmic

patterns. Alternatives approaches involve mimicking easily identifiable and translated

sounds such as steps, or a heartbeat. These mechanisms often require some kind of

interpretation or mnemonic to associate the rhythm’s pace with a positive or negative

persuasive cue.

[3] Other (Languages)

Outside the previous two categories we can categorize other delivery methods

which no not fit in their entirety the aforementioned classifications. While, to our

knowledge, there are no examples of MEA´s employing Morse code to convey

persuasive cues, the sheer possibility of their existence must be accounted for in a broad

design space.

Stakeholders

The second dimension of this design space deals with the possible intervening

entities in relation to the exercise’s central object – the user. These stakeholders may

range from individuals who are not directly participating in the exercise, but whose

influence is directly related with that action to those who have influence on the user,

especially if the activity is taking place in the same vicinity. In light of these

possibilities, we identified three types of stakeholders in MEA´s:

49

[1] Spectators

This entity assumes an indirect influence on the user, since it is not

participating directly in the exertion activity itself. Spectators often interact with the

user through social networks or, if supported, via the MEA. Their interventions may be

framed in the activity timeline: spectators are able to cheer the user during the exertion

activity (e.g. likes on Facebook which are then conveyed through audio cues), before

the activity takes place (e.g. pre-emptive motivation or pressure for the user) or after the

activity (e.g. acknowledgment of the user’s actions and/or performance after the

exercise).

[2] Opponent

Individuals performing an equivalent exercise, not necessarily an adversary but

in most cases a companion. The existence of an opponent is often a strong driving force

behind one´s motivation. Their interventions are related with two other dimensions:

time and space. The user may be persuaded by opponents in real time, while exercising,

or be challenged by “ghosts” of those opponents, i.e. recordings of their previous

performances. Orthogonally, the design space covers co-located or spatially distributed

opponents. Either “ghost” or simultaneously the exercising could: progress on the same

path (e.g. a group of individuals, including the user, running together); or at completely

different places (e.g. the opponent running in a different region).

[3] Self

The last category is the user him / herself as an intervenient in the exercise.

This category is also linked with space and time dimensions with an obvious restriction:

it cannot happen in different places at the same time. The same-time/same-place, i.e.,

users’ self-persuasion while exercising, is possible and may include the MEA (e.g.

explicitly or implicitly change the persuasion cues or intensity). Different-time pertains

to events that occurred in the past (i.e. “ghost” recordings), independently of location.

Time

Time is a key dimension in this design space. The time dimension can be divided in

two subcategories, each with further parcels: one pertaining to the persuasive cue’s

instant of creation and another related to the instant of reception.

Instant of Creation

50

The instant of creation pertains to the moment in time in which a persuasive

message is created (regardless of message type):

[1] Before Exercise

Messages created before the exercise takes place are typically authored by

spectators and generally are aimed at increasing the adrenaline prior to the exertion

activity. Oppositely, they can be a source of anxiety or can be classified as peer

pressure. Opponents can also create persuasive messages beforehand, acting as a

stimulator or a challenge for the exercise itself.

[2] During Exercise

Persuasive cues created during the exercise can stem from two sources:

spectators cheering a live event; challenge cues exchanged with the opponent. The latter

is mediated through technology: while in co-located exercises there might not be a

mandatory usage of a MEA to challenge one another, such presence is required for the

delivery in spatially distributed activities.

[3] After Exercise

Messages created after the exercises do not influence the user on the moment

of activity, but rather act as a catalyst for the next exercises. These cues have an

acknowledgement nature (positive or negative) and can be originated from spectators or

from the opponent(s).

Instant of Reception

This subcategory is related with the moment a user receives a persuasive

message stemming from any source. Similarly to the instant of creation, the reception

may occur at three distinct times:

[1] Before Exercise

Persuasive messages may be received prior to engaging in an exertion activity.

Such messages add to the motivation or pressure felt by the receiving subject (e.g.

receiving social network notification regarding the exercise activity). These messages

may stem from both spectators and the future opponent(s).

[2] During Exercise

51

Spectators can cheer the subject as he / she carries on with the exertion activity.

Opponents may also intervene and send persuasive cues: albeit in co-located settings

such delivery may completely overlap the technology involved, in distributed settings

the latter becomes the vehicle of dissemination.

[3] After Exercise

The reception of messages after finishing the exertion activity is often a case of

acknowledgment. These post-exercise messages may fuel the motivation and

anticipation for the next exercise. Related sources may be both spectators and

opponent(s).

Space

The last dimension affecting persuasion in MEA´s pertains to the stakeholders’

spatial distribution:

[1] Co-Located

Without technological support, same-time/same-place persuasion is guaranteed

through the perception of how the opponents running with the user are performing.

With the introduction of MEA´s this reality transverse time and may become augmented

with different delivery mechanisms.

[2] Distributed

Existing applications support challenges from different locations than the one

the user is working out on. This enables persuasive message exchange regardless of user

location. The origin of these messages may cover all stakeholder categories: spectators

can be geographically distributed (e.g. someone cheering at home); the opponent may

be on a different place (e.g. running at the same time in distinct places); the user may

capitalize on an opponent’s “ghost” recording in a different course (e.g. a simple

challenge exercise); the user may capitalize on a “ghost” log to compare his / her

performance in different locations (e.g. altitude training).

Summary

In light of the envisioned design space, we will now address how a subset of the

existing applications is mapped into each dimension. A total of 11 MEA´s were chosen

from existing literature and from our questionnaire results: Endomondo Sports Tracker

[28], Nike+ [29], RunKeeper17

[36], Runtastic15

[31], Keep Running18

, Cardio Trainer16

,

52

Adidas miCoach [30], MapMyRun [37], GetRunning [34], Strava [32] and Fitbit [35].

Table 5 depicts the applications’ distribution in our design space:

Table 5 – MEA distribution in our design space

Delivery

Natural Rhythmic Other

All but “Keep Running” “Keep Running” Nike+, “RunKeeper”,

“Runtastic”, “MapMyRun”

Stakeholders

Spectators Opponent Self

5 8 “RunKeeper”, “Runtastic”,

“MapMyRun”

Time (Instant of Creation)

Before During After

11 6 1

Time (Instant of Reception)

Before During After

8 9 4

Space

Co-Located Distributed

11 10

3.2.3 Scenarios

A quick glance at the design space and, in particular, Table 5’s MEA mapping

shows that all dimensions and respective categories are represented in existing software.

Still, some caveats emerge, such as the lack of spectator support or using rhythmic

patterns to persuade the user. In addition, if we perform a multi-dimension analysis by

cross-referencing each category, new design challenges appear. We will now discuss

some scenarios generated from the cross analysis between the design space’s

dimensions, potential interest and feasibility with existing technology.

53

Distributed Real-Time Spectators

A substantial number of applications support live monitoring a workout session.

The user, while exercising, is able to receive applauses and cheers from spectators

(users in a remote place viewing workout statistics through a specialized interface).

Despite the innovative nature of this feature, it mostly only offers the monitoring of one

user at a time or several users in the same vicinity.

Our design space covers applications which allow for the monitoring of several

users in geographically distributed settings. As such, two users in geographically

distributed areas of the world can work out at the same time, while their friends cheer,

tease and motivate them from another place. In an increasingly global society in which

relatives and friends get in touch remotely, this scenario is capable of fostering social

bonds between them and improving mutual commitment to workout activities.

Distributed Real-Time Team Effort

The second scenario we envisioned is aimed at those groups of individuals who

enjoy working out in groups. We believe that the creation of group challenges which

require the cooperative effort of several individuals to meet the necessary requirements.

Instead of building on the efforts of a single person, this approach would create a sense

of community and responsibility between group members, requiring their added

motivation towards the fulfilment of these challenges. Multiple persuasive cues could

be used, namely capitalizing on the still unexplored rhythmic approach. For instance, to

motivate group members during the exercise, a song could be played. The particularity

of this approach is that each member would be assigned to an instrument being played.

If a member underperforms, then his / her instrument is removed from the song, raising

awareness to the group that one or more users are not meeting the challenge’s

expectations.

Rhythmic Spectator Awareness

Existing MEA´s such as RunKeeper+17

[36] allow users who are not participating

in the exertion activity to cheer the individuals partaking in the activity. While this

feature’s support is steadily increasing, there are still opportunities to explore it,

especially taking into account the intersection with other dimensions of our design

space. One characteristic of this feature is that the feedback is disruptive, only being

issued whenever a spectator takes any action with the system, despite possibly being

54

monitoring for the full duration of the exercise. Instead of capitalizing on a solely

disruptive approach, we envision the employment of a continuous persuasive audio

stream which does not require spectator interaction. The MEA user would receive a

continuous stream of cheers or applauses which would get more or less intense based on

the number of spectators monitoring him / her. This rhythmic strategy can be enhanced

with additional audio feedback as a metaphor for spectator activity while monitoring

(replacing the more disruptive applauses current MEA´s adopt). This approach allows

users to maintain awareness about the number of spectators monitoring them, through

the immediate assessment of the intensity of the virtual cheering they are able to hear.

3.2.4 Design Space Discussion

This design exercise resulted in two distinct contributions. On the one hand we

were able to define a design space for mobile exertion applications, contemplating a set

of dimensions and categories related with stakeholders and exercise location, among

other. On the other hand, the mapping of a set of relevant MEA´s in this design space

allowed us to identify some caveats that existing software possess. Furthermore, these

caveats are able to be turned into interesting research opportunities, as exemplified by

the depicted scenarios.

Current MEA´s fail to thoroughly explore rhythmic persuasive cue delivery, usage

of spectators to live monitor exercises in remote locations and the acknowledgement

through reinforcement messages after the exercise activity. While we do agree that

some of these unaddressed challenges are tackled via the integration of the MEA´s

functionalities in social networks, we also believe that this integration is responsible for

the lack of bold approaches towards persuasions in MEA´s.

The validation of this design space is undergoing, through the creation of a MEA

whose primary goal is to cover the scenarios we discussed in this chapter. Another goal

concerns testing this application with end-users. The intent is to perform a comparative

study on which types of persuasive delivery mechanisms impacts more intensively on

the user. This will be detailed in the coming chapters.

55

56

Chapter 4

ExodUS

In this chapter we will focus on the developed system, based on the information

described in the previous chapters. We envisioned a ubiquitous MEA that takes on the

most contemporary characteristics regarding the existing systems, as described in

chapters 2 and 3. Like the many MEA´s reviewed in this document, this work shares

some of their characteristics.

This chapter primarily serves as a report about the intrinsic details regarding our

MEA and their relation to our main research aim, the impact on motivation and

persuasive factors regarding these kinds of systems. Details regarding our entire system

components and their relations will be looked into.

4.1 System Overview

In this subchapter the entire system regarding the ExodUS mobile exertion

application will be detailed.

4.1.1 System Components

The system is composed of the following components:

Mobile Application

The main focus lies in the Mobile Application which is responsible for the

exergame, meaning the collection of exertion data is done by the app and it is the most

57

important component since it is with the user while he/she is performing the workout

sessions.

Dispatch Server

Importance wise, the Dispatch Server is next in line. It is responsible for managing

all information (exertion data, user accounts, user profiles, among other data) pertaining

the system. Thus, the back-end management of the system is handled by the server.

Web Platform [Website]

For the front-end, complementing the mobile application, there is the Web platform

(Website) in which the users can review/view all information pertaining them or other

members.

Database [×2]

All relevant information (user authentication credentials, user profile, user social

ties with other members, exertion data and challenges, among others) is stored in two

databases, one managed by the dispatch server that deals with all information relevant

for the system and the other managed by the mobile app pertaining workout session data

only.

4.1.2 System Architecture

Figure 10 illustrates the systems´ architecture and provides a basic understanding

of how the components intermingle with one another. Each database is only managed

and accessible by one other component, them being the dispatch server and the mobile

application, respectively. These databases are both implemented using the SQLite

software library. The dispatch server listens to requests from both the mobile

application and the website, responding to the received request accordingly. This server

is implemented in Java. The website is a web platform which offers exclusive

functionality within the system. It is implemented recurring to HTML, JavaScript and

CSS style sheets. Finally the Mobile Application, as stated earlier, is responsible for the

collection of exertion data, among other functionality. It is implemented for the Android

Platform, with support of older versions of the operating system such as Froyo (API

Level 8) and Gingerbread (API Level 10). Other technology used, revolves around the

integrated sensors present in mobile devices such as the GPS and vibration sensors.

58

Figure 10 – ExodUS System Architecture illustrating all components and interactions between

them

4.1.3 Dispatch Server Overview

For the ExodUS system, a bastion component for managing all system information

and assume control of its back-end side is needed. This component is the Dispatch

Server. Its basic functionality can be explained in a three step procedure:

1. Listen for incoming requests;

2. Process the requests;

3. Notify the clients who made the requests.

This working method translates into a Remote Procedure Call service (RPC), to be

explained in more detail in the coming sections.

Communication Sequence Diagram

Figure 11 – Generic RPC communication Sequence Diagram illustrating a request from the

systems´ front-end / client side

59

The previous sequence diagram illustrates a generic communication sequence

between a front-end/client component and the dispatch server (and its managed

database). The client kick starts the communication by sending a request to the server

which in turn computes a response to be sent back to the client. During the process of

constructing a response, the server may have to access its database for stored

information.

Server Handling

Configuration

Configuration wise, the server has a hardcoded port number for its listening socket

to operate. This ports´ number is 12345. The databases´ location on the file system is

also hardcoded and known from the start of execution.

Server Interaction

In order to interact with the administrative personnel this server maintains a screen

object which creates a windowed interface.

Database Handling

The server connects to its managed database at the start of execution (its location is

known). This is done through a Database object which acts as a proxy for all database

queries.

Data Structures

The server needs to provide a few data structures that enable the intended

functionality. The use of the Google Cloud Messaging service (GCM) requires the

creation and maintenance of a device list in which to store the unique ids that identify

the clients´ mobile devices. Also, maintaining the connected clients´ information and

pending or “in progress” requests is needed. Furthermore, management of social data

(e.g. clients´ buddy lists and pending friend requests) is addressed with its own data

structure.

60

Multithreading

The server is always available to answer any new incoming connection. In order to

avoid congestion, threads are used. As such, the server is always listening for incoming

connections. When a connection arrives, the server automatically starts a thread for that

requests´ processing and stores it. This new thread is only responsible of addressing a

specific client and it only terminates when the client logs out from the mobile

application / website.

Client Representation

Communication originating from the server is done through the use of the GCM

service. To work, this service requires a device id identifying the clients’

communication “address”. This means that a client can only be contacted if a valid

device id is registered within the server. As explained earlier these device ids are stored

within their own data structure, but it becomes necessary to relate a device id to its

specific client. Thus a client representation is needed.

To represent a client, three properties must be obtained: a username, a password

and a device id [used for the MEA clients only]. With these properties a device id can be

correlated with a specific client who mainly uses a unique username and a password for

authentication. The use of such representation also allows for the comparison between

clients within the server.

Push Notification Handling

Once the server has a valid registered device id and a request from a client arrives,

the thread responsible for its resolution creates and maintains a Push Notification

object which handles the creation and sending of a GCM push notification message

back to the client. This object contains a hardcoded SENDER_ID, necessary for the

usage of the GCM service. Furthermore, this object offers creation and sending methods

for the GCM response notifications.

Thread Processing

Reviewing the previous workings description of the server we know that:

61

1. The server creates and executes a thread for every request;

2. The server creates and maintains a virtual representation of a client;

3. Incoming messages are the first step in a three step resolution procedure, the

RCP service;

4. Outgoing messages are conveyed using the GCM service;

The preceding sections explain the details behind 1, 2.and 4. This section will now

focus on 3, the RCP service.

RCP Service Implementation

As stated before, the basic functionality for this RCP service can be explained in a

three step procedure: 1. Listen for incoming requests; 2. Process the requests; 3. Notify

the clients who made the requests. The first step is handled by the servers´ main

function (it listens for a request and when it receives one, creates a thread for its

processing). The other two steps are done in the individual threads´ functions.

Processing the requests involves identifying its operation code (opcode for short) and

executing the desired function.

4.1.4 Databases Overview

The ExodUS system makes use of two, distinct but similar, databases. One

database stores all information manageable by the system and acts as the main database.

The other one is kept within the players devices and stores the collected data only. This

data is only accessible on that same database, until the player synchronizes the content

of his / hers device database with the main database.

Dispatch Server Database [Main database]

Database Description

This database is responsible for holding all data relevant for system operation. The

dispatch server is the only component that has access to it, meaning all data that is

stored, removed or updated will be managed by the server. The data itself mainly

concerns workout sessions and user related information.

62

Data Description [plus Relational Model]

As stated above, the data can be mainly divided into two different clusters, one

cluster pertaining user information and another one dealing with workout session

statistics. For the first, the involved dimensions are as follows:

Account (user account, used mainly for authentication purposes)

Profile (user profile, used for storing user profile information)

Buddies (user buddies, used for social interactions within the systems´ members)

As for the second cluster, regarding exertion data, the involved dimensions and

attributes translate into:

Session (user workout sessions)

Entry (user workout session entry – each entry is equivalent to a second of the

overall workout session)

Data (user workout session entry data – each workout session entry has seven

properties to store: Latitude, Longitude, Altitude, Effort, Distance, Velocity and Time )

Figure 12 – Server managed Database Class Diagram

For each user account only one profile and buddy list may be associated.

Each account (user) may record many workout sessions. Each workout session will

have n entries (where n corresponds to the total seconds spent by the user in the

workout session) and each entry will have 7 data associated with it.

63

Mobile Application Database [Device database]

Database Description

This database is responsible for holding all data pertaining to workout session

information. Unlike the previous database, this one accommodates fewer tables

(exertion specific only). It is accessed only through the mobile application and is

supposed to be stored in the mobile devices´ SD card. Thus the application is

responsible for all data that is stored, removed or updated to this local database.

Data Description [plus Relational Model]

Unlike the previous database, this one deals with only one cluster, pertaining

information on workout session statistics. The involved dimensions and attributes are

the same as described for the previous database.

Figure 13 – Mobile App managed Database Class Diagram

4.1.5 Component Communication

In this system communication is carried out using two different methods: Google

Cloud Messaging (GCM) and standard TCP sockets. The idea is to use the GCM service

provided by Google to send push notifications to our Android client Application and use

the standard TCP sockets to send requests to our remote server. The server aims to offer

a Remote Procedure Call service (RPC), where a client can invoke a series of operations

on the data held on the server, which then sends a response back to the client, fulfilling

the request. For the RPC service to work the server must have hardcoded the procedures

64

and functions that will be executed following the reception of the remote calls from the

clients and the client must send the remote calls with a known and fixed format. This

format translates into the usage of an operation code (opcode) and the necessary

arguments for the procedures´ / functions´ execution. Finally, the server executes the

procedure / function specified by the received opcode and responds to the client by

using the GCM service. The usage of this service allows clients to reduce battery

consumption, since the need to constantly keep listening to a socket for incoming

responses disappears.

System Messages

The following Sequence Diagram illustrates one of the important successful use

cases of the system. This allows for a better perception of the communication process

between the systems´ components.

MEA Live Broadcast Execution

Figure 14 – ExodUS Live Broadcast Sequence Diagram example

65

1: Mobile Client Request:

Opcode: 004

Arguments: username, password

Request Description: “Begin Broadcast” – Mobile App signals the beginning of a

live broadcasted session to the server.

3: Server Request Processing:

Opcode: 025

Arguments: “Server”, “LIVEBUDDY_BEGIN”, live_friend

Response Description: “Web Client Broadcast Begin” – Server alerts the web client

that the mobile client (live_friend) started a live broadcast.

5: Mobile Client Request:

Arguments: username, password, data

Request Description: “Update broadcast data” – Mobile App sends the server the

latest exertion data to be broadcasted.

7: Server Request Processing:

Opcode: 027

Arguments: “Server”, “LIVEBUDDY_DATA”, live_friend, public_data

Response Description: “Web Client Broadcast Update” – Server sends the web

client the latest exertion data (public_data) to be displayed in the web platform.

9: Web Client Request:

Arguments: web_client, live_friend, message

Request Description: “Send Live Message” – Web Platform sends the server a

written message to be conveyed to the live mobile client.

66

10: Server Request Processing:

Opcode: 004

Arguments: “Server”, message

Response Description: “Web Client Live Message” – Server sends the mobile

client the received message from the web client, which is then “read out loud” by the

mobile app.

11: Mobile Client Request:

Opcode: 005

Arguments: username, password

Request Description: “End Broadcast” – Mobile App signals the ending of a live

broadcasted session to the server.

13: Server Request Processing:

Opcode: 026

Arguments: “Server”, “LIVEBUDDY_END”, live_friend

Response Description: “Web Client Broadcast End” – Server alerts the web client

that the mobile client (live_friend) ended a live broadcast.

4.1.6 Web Platform Overview

The ExodUS Web Platform aims to build an accessible platform via the web

enabling the users of our Mobile Exertion Application to manage their workouts and

interact with other users. The platform developed is in fact a website which will enable

users to access some services, such as managing workout data and engage in social

interactions with other players. The platform is available for all users identified through

a username and password.

67

User Driven Features

Player Profile Management

The existence of a player profile within the system is derived from the website sign

up service. During this process, the new user provides eight attributes (name, country,

gender, age, weight, email, username and password) which compose the players´

account. This data is stored within the database managed by the ExodUS dispatch server

only (the username attribute is checked, for it needs to be unique). The website then

allows players to manage their data, allowing for modifications to it (for instance, a

player might want to update their weight). The only attribute that is unchangeable is the

username attribute.

Social Biding

Much like social networks, the platform allows for the players to associate

themselves with other users. The system keeps and manages buddy lists that are filled

with “friends” associated with the users´ account. To add a friend to the buddy list the

website offers a player search function (by the unique username) which creates

friendship requests that are sent to the respective users. The latter, are presented with

any unanswered friend requests, deciding whether to accept or reject the request.

Once player X adds player Y onto Xs´ buddy list, Y is now able to follow any live

broadcasting done by player X with the ExodUS mobile application.

Figure 15 – ExodUS Website social interface options

68

Workout Session History

Similarly to the mobile application, the website offers the possibility to review all

synchronized workout session data with the dispatch sever. This service loads and

presents the stored workout sessions with the same format as the mobile application.

The difference is in the actions that the user as when selecting a workout session.

Unlike the application, the website compiles all the exertion data collected and

normalized from the database and presents it to the user. All location updates and

exertion data are posted in a map (again, using the Google Maps API). Checking a

specific location, the users can also check what were his / hers exertion data at that

precise moment in the workout session. All totals are also presented to the user, outside

the map. The ability to erase the session is the only function that is shared between the

mobile application and the website.

Figure 16 – ExodUS Website Player History

Workout Live Viewing

Whenever a friend broadcasts a workout session, the website informs the user,

giving him / her the option to follow it, in a live fashion. When viewing a live session

the location of the friend is displayed on a map and his / hers latest exertion data is

presented as well. Furthermore, the website offers the possibility to interact you the

friend performing the broadcast sending pep talks or pre-determined sounds (applause

or booing) which are reproduced on the mobile application also in a live fashion. This

further reinforces the social interactions available to the system and it also as the

69

potential to help motivate the player that is executing the workout session, through

incentives or taunts from their friends.

4.1.7 Mobile Application Overview

The game features three different modes of play, briefly explained as:

Single Mode(s) – where players can execute jogging sessions and benefit from

a series of specific information transmitted through the use of some modalities. Ghost Mode – where players are confronted with other performances (from the

player himself or from other players) in comparison to the current in-progress

performance. Spectator Mode – where online friends interact with the players, while they

jog, in a real-time fashion.

The ExodUS mobile exertion application feature set can be divided into two

different categories: one consisting of the inner workings of the application and another

related to user driven features. While the latter consists of the possible actions the

application provides the user, the first is of an extreme importance since it deals with

data collection and management, the core of the application.

Internal Functionalities

Location, Tracking and Mapping

A key feature for the system is the ability to perceive the players physical location.

All studied MEA utilize data obtained from the mobile devices´ built-in GPS to pinpoint

the real world location of the player. There are two aspects that should be highlighted

for the location acquisition process. First is the option to configure our GPS to “rest”

between location updates with the condition being either be a distance restraint, a time

restraint or both. This is useful mainly to conserve battery time but has the drawback of

diminishing the accuracy of the location fixes. Second is the criterion applied when

receiving a new location from the sensor. Since usually the sensor will return a couple

of locations when it has a fix, a small buffer is used to store them. When the buffer is

full the app will discard the remaining locations it receives and focuses on the ones

already stored inside the buffer. The next step is to select the “better” of the stored

70

locations. Every location object has an accuracy value of the fix in meters (0.0 if the

location object has no accuracy value associated) and by using this we can easily get the

fix with the best accuracy, in other words, the fix whose accuracy value is the lowest.

Finally, the resulting location object is used. The following is the pseudo-code

regarding this criterion:

FOR EACH LOCATION{

IF Current Location Better Than Best Location {

Current Location is Best Location

}

}

Two other steps are necessary to complete the tracking procedure within ExodUS.

The second step is giving visual feedback, illustrating the tracking, to the player. This is

done by mapping the course the player traverses. In ExodUS, mapping is done using the

Google Maps API, by providing the location data received from the GPS sensor and

place markers and overlays on a map, effectively translating the traversed route by the

players.

Figure 17 – ExodUS Mapping of the Tracked Workout Session

The final step in the tracking procedure translates into the exertion data acquisition.

The metrics gathered by ExodUS are Latitude, Longitude, Altitude, Effort, Distance,

71

Velocity, Acceleration and Time. The first three metrics can be directly obtained

through the GPS sensor.

Distance, Velocity and Acceleration Metrics

Through the use of the GPS sensor we can acquire the players´ location. This

location is described in the form of the latitude value and longitude value. To be more

precise, it is described in the form of a geopoint. Besides accurately placing a users´

location on a map these geopoints can be used to measure the distance between two of

them. The Spherical Law of Cosines [22, 23], allows for the calculation of the distance

between two geopoints or rather, the shortest distance over the earths´ surface:

D = acos( sin(φ1).sin(φ2) + cos(φ1).cos(φ2).cos(Δλ) ).R

Where φ is latitude, λ is longitude, R is earths’ radius

(approximately 6,371km).

With this formula we can calculate the distance between the different geopoints

given by the GPS sensor and by adding the different “segments” we can calculate the

total distance traversed by a player during their workout session. This means that the

more GPS location updates, the more accurate the calculated total distance value will

be. However this value will always come short of the real distance traversed by the

player since it uses straight lines between geopoints and does not take into account any

possible curves that might have been traveled (this is illustrated in figure 17).

Unlike distance, velocity and acceleration are instantaneous metrics. Velocity is the

derivative of the position of a moving object with respect to time. Acceleration is the

derivative of the velocity of a moving object with respect to time, or rather the second

derivative of distance. Given that we calculate the distance between one GPS location

update with another we can calculate the velocity and acceleration on the very same

instant by applying the following formula:

VelocityP2 = (Distance

p2 – Distance

p1) / (Time

p2 – Time

p1)

AccelerationP2 = (Velocity

p2 – Velocity

p1) / (Time

p2 – Time

p1)

Where p2 is the newest GPS location update and p1 is the former

GPS location update

72

Effort Metric

Effort is measured by adding up the number of calories burned, the energy

expenditure, while performing the exertion activity. The real challenge here is figuring

out how to accurately calculate this number. What are the variables we need to

consider? One formula was usable38

[24]. The authors of this workout devised a

formula taking into consideration aspects like gender, age, weight, time and heart rate,

as follows:

(MEN) Calories Burned = [(AGE * 0.2017) – (WEIGHT * 0.09036) + (HEART

RATE * 0.6309) – 50.0969] * TIME / 4.184

(WOMEN) Calories Burned = [(AGE * 0.074) – (WEIGHT * 0.05741) +

(HEART RATE * 0.4472) – 20.4022] * TIME / 4.184

Unfortunately, the heart rate variable required the use of a heart rate monitor,

meaning additional hardware. This was not, however, the reason to scrap the use of this

formula in ExodUS. Our main issue with it was the lack of consideration for the

distance traversed during the exertion activity. Distance reflects movement, which in

turn provides suitable energy expenditure39

[8]. So a formula that disregards distance

might not be ideal for our system. Thus another approach was needed. That approach

was to investigate how treadmills calculate the burned calories by its users. Finally we

decided to reach a compromise. Accurately calculating our players’ energy expenditure

is not the aim of this work. Instead we want to grant our players with a level of

comparison between different workout sessions (mostly using virtual ghost partners). So

we tried one last method to reach our formula. Searching the web produced another

result in the form of the following formula40

:

Calories Burned = WEIGHT * 0.75 * DISTANCE

38 http://fitnowtraining.com/2012/01/formula-for-calories-burned/

39 http://en.wikipedia.org/wiki/Energy_balance_(biology)#Energy_expenditure

40 http://www.myfitnesspal.com/topics/show/134478-accurate-formula-to-determine-calories-burned-jogging

73

This formula is suggested as the simplest way to accurately determine how many

calories a person is burning while jogging. It only takes into account the persons´

weight (in pounds) and the distance traversed (in miles) in the jogging workout session.

With ExodUS, both variables are usable as the players´ weight is stored on his / hers

player profile and the distance can be obtained throughout the tracking procedure. We

hypothesized that when a jogger is running uphill, the necessary effort needed to

perform it should be greater than when running on an even field or downhill. To

reproduce this behavior we use the altitude value given to us by the devices´ in-built

GPS sensor. Comparing the current location altitude value with the altitude value from

the previous established location we can discern field inclines and used them to translate

an increase or decrease of the effort value, depending on the case. This resulted in the

final formula used in the ExodUS Mobile Application:

Calories Burned Depending on DISTANCE = (WEIGHT * 0.75 * DISTANCE)

Calories Burned Depending on INCLINE = (WEIGHT * 0.75 * INCLINE)

Total Calories Burned = CB_DISTANCE + CB_INCLINE

Exertion Data Structure

Latitude, Longitude, Altitude, Effort, Distance, Velocity, Acceleration and Time

are the exertion data manipulated within ExodUS. Management of this data is essential

in order to provide the game with its intended functionality. ExodUS creates and

manages an object that is responsible for storing this exertion data. This data refers to

single point in time, or rather, it refers to one single second belonging to a workout

session, since our chronometer is displayed in the format of “HH:MM:SS”, meaning our

smallest time instance is equal to one second. This also allows for a fast and easy access

to any specific instance we want to access. Exertion data is obtained through the use of

the previous features only in a few seconds that compose the entirety of the workout

session (instants when the device receives location updates). For the remaining seconds

“blank” entries are stored (see figure 18).

74

Figure 18 – ExodUS Exertion Data Structure representation

Data normalization

When considering basic game modes, ignoring the “blank” entries is entirely

possible. But considering our devised virtual ghost challenge, these entries are no longer

ignorable. This is because we want to make a comparison, as directly as possible,

between our players´ performance and the respective ghost partners´ performance. So if

the player gets a location update in the instant 00:05:32, we want to relay feedback

pertaining the direct comparison between the newly created exertion data and the stored

exertion data whose instant is “00:05:32”, which should not be a “blank” entry. Thus,

some sort of data normalization process is needed in order to replace these “blank”

entries with other entries whose exertion data needs to be estimated. In ExodUS, this

process is done at the end of each workout session, in two intertwined steps:

1. Search the exertion data structure for the entries´ interval between the entries

whose exertion data where already known

2. Once an interval is located, normalize the “blank” entries by storing proper

exertion data on to them

Distance, Velocity, Acceleration and Effort are the metrics that go through this

normalization process, due to two reasons. The first is that these are the metrics that

have value regarding our feedback and workout session reviewing purposes. The second

is that Latitude, Longitude and Altitude are metrics which are normally given by the

GPS sensor and are much harder to extrapolate (it would hypothetically involve the

75

notion of direction and the resulting accuracy might be an issue). Contrary to these

metrics, the remaining are obtainable using equations of a straight line to try and

approximate the values for the different points belonging to the identified interval of

entries (in between the known exertion data metrics). Considering entries as points

belonging to a 2-dimensional axis graph, a function associated with the players´

distance variation can be outlined:

Figure 19 – 2-axis graph representation of a players´ distance metric

Figure 19 illustrates two known points, P1 and P3, representing entries which have

obtained exertion data stored. The aim is to obtain the normalized distance value for

point P2, a point in between the two known points. Since P1 and P3 are defined (we

have their distance and time values), the illustrated function can be defined by:

Y = MX + B ⇔ B = Y - MX

Which is the Slope-Intercept Equation

In this case, to find D2:

D2 = [ (D3 – D1) / (T3 – T1) × T2 ] + D3 – [ (D3 – D1) / (T3 – T1) × T3 ]

So using the Slope-Intercept Equation, we can calculate the distance values of any

given points between the two known points. But velocity and acceleration also need to

76

be normalized. The procedure is more or less the same for velocity, as the use of the

Slope-Intercept Equation in a similar fashion to the one used for distance is employable:

Figure 20 – 2-axis graph representation of a players´ velocity metric

Figure 20 also illustrates two known points, P1 and P3, representing entries which

have obtained exertion data stored. To obtain the normalized velocity value for point P2

function can be defined by:

Y = MX + B ⇔ B = Y - MX

Which is the Slope-Intercept Equation

In this case, to find V2:

V2 = [ (V3 – V1) / (T3 – T1) × T2 ] + V3 – [ (V3 – V1) / (T3 – T1) × T3 ]

For acceleration, the reasoning is a little different. Acceleration is the derivative of

velocity and with that in mind, the resulting value of a normalized acceleration value

should be equal to the value of the slope for its Slope-Intercept Equation. This means

that for any point P2, located in between two known points P1 and P3, the acceleration

value is a constant, given by:

A2 = [ (V3 – V1) / (T3 – T1) ]

77

Effort is the last metric which needs to be normalized. It is also the easiest to

normalize simply because it does not require any additional procedure. It is processed

as described earlier, with the peculiarity of using the normalized distance value for the

effort calculating formula. As for the necessary altitude value for the slope calculation,

the altitude values of the two known points which encapsulate the interval subject to

normalization are used.

Lastly, to conclude the normalization process description, it is necessary to refer

two particular cases that occur during the process itself. Those are the points originated

from the root (from the beginning of the workout session until the reception of the first

GPS location update) and the points ranging from the last GPS location update and the

very end of the workout session:

Figure 21 – Macroscopic view of the Data Normalization Process

The first batch are left as “blank” entries since the application had yet to find more

than two location updates to use as the target interval for normalization. The last batch

are treated as if the player did not move from the last location provided by the GPS

sensor and will always store the last distance value obtained and set their velocity and

acceleration values to zero.

78

User Oriented Functionalities

Player Profile

The player profile consists of 6 player attributes, namely: Name, Country, Gender,

Age, Weight and Email. Each of these attributes has its own purpose within the

exergame, some with a more relevant importance than others. The Name attribute is

divided into first and last names of the player and serves only for viewing / social

interaction purposes. The same can be said about the Country, Gender, Age and Email

attributes. Furthermore, it is possible to configure user characteristics such as Country

and Age. Finally, the Weight attribute is the most important attribute from the player

profile as it is used in the calculation of the players´ effort in-game and is also

configurable.

Figure 22 – ExodUS Settings Screen showing the Player Profile

Player Friends

ExodUS incorporates a Friend / Buddy list feature similar to those present in social

networks like Facebook or Google+ which aims to create, maintain and strengthen

social ties between players. These buddy lists are maintained by the systems´ Dispatch

Server and serve the purpose of allowing the players´ friends to role play the Spectator

role. This version of the app only allows for the viewing of the players friends. Further

functionality (friend search, friend requests, and unfriend actions) is available on the

Web Platform instead. Social interaction is promoted in two ways with ExodUS. The

79

first is explicit, the aforementioned Friend / Buddy list functionality. The second is

more implicit and regards the different game modes, namely Ghost challenges and Live

Broadcasts (more details in the coming sections).

Figure 23 – ExodUS Friend / Buddy list Screen

Workout Modalities

ExodUS offers the possibility track four different activities, them being: Biking,

Hiking, Running / Jogging and Walking. Besides offering different activities to track,

ExodUS also offers a broad set of workout types. Common uses such as executing a

simple workout session or trying to reach a specific goal are available to the player.

Working out alongside virtual partners or broadcasting sessions are not so common

workout types that are also available for the players. The following list presents the

playable scenarios with ExodUS:

Going out for a jogging session

Trying to reach a specific goal (time, distance or effort goal)

Workout alongside a virtual partner (representing the same person)

Working out while broadcasting the session for others to watch

Workout alongside a virtual partner (representing another person) [only

available if the same device is used for both players]

80

Figure 24 – ExodUS playable Activities and Workout Types

Data Storage

ExodUS has two types of data storage. One of them concerns the storage of the

workout session data into a local SQL Database (located on the devices´ SD card). The

other pertains the logging ability of the game, which stores different types of data, with

different purposes as well. Two types of logs are created and kept by the application

also in the devices´ SD card. One type of log is the debug log file. This type of log was

used throughout the development stage of the application in order to store any value that

might have been useful to the developer. The output file is a text file (.txt extension),

located on the “../exodus/debug/” folder, with no particular format employed. The other

type of log serves an important role for applications of this nature (MEA in general).

Most applications of this genre have limited capabilities when it comes to the ability to

present detailed information regarding exertion data and workout sessions in general.

Showing the user visual information such as graphs and tables is difficult and is mostly

available on their respective web platforms. To solve this issue, the application makes

use of the second type of log file, the workout session log file. This log stores all

exertion data obtained and normalized regarding a specific workout session executed by

the player. The output file is a comma-separated values file (.csv format), located on the

“../exodus/log/” folder and enables the players to use all the available functionality of

Microsoft Excel, for instance, to make graphs and tables regarding the workout

81

sessions´ data, with the disadvantage of this functionality not being automatic and

putting the work on the players´ hands.

Player History

Maintaining a history record allows players to keep track of their performances,

routes and overall activity within the application. For ExodUS, the player history

provides the game with much functionality, highlighting the option to select a ghost

partner from the available poll of recorded workout sessions, among others options.

Figure 25 illustrates the games´ player history screen. In it the various recorded

workout sessions are displayed offering not only information regarding those workout

sessions but also the option to manage them. The typical information conveyed to the

player regards the sessions´ metrics and definition. The available written description

shows the sessions´ date, effort, distance, time, selected workout type, points awarded

and player data. The icon shows the selected workout activity.

Figure 25 – ExodUS Player History Screen

Besides offering an overview of the several performed workout sessions, this

screen also offers management options regarding them. Three different actions can be

selected for each of the recorded workout sessions. These options concretize the ability

to erase a specific workout session (from the local database located in the devices´ SD

card), to review the comment written by the player at the end of the target workout

82

session and finally the possibility to undertake a ghost challenge with the specific ghost

partner.

Figure 26 – Selectable options for the various workout sessions recorded on the Player History

The ExodUS Player History feature also grants the players with a data

synchronization function. This function makes it possible to synchronize all the workout

session data stored in the local database with the remote database managed by the

ExodUS Dispatch Server. It will only process the unsynchronized workout sessions

once commanded by the player to do so, by means of the “Scan History” button (see

figure 27), which contacts the server in order to compile the sessions that should be

marked for syncing. The sessions are them shown to the players in the list located at the

bottom of the screen, with the selectable and available actions concretizing the ability to

erase the session, to read the recorded comment regarding said session and to start the

syncing process of the target workout session.

Figure 27 – ExodUS Player History Data Synchronization Screen

83

Feedback Modalities and Types

The usage of different modalities in ubiquitous scenarios is regarded as an

important issue when studying MEA´s and exergames. How can different modalities

influence social behavior and user performance? For our application, tree different

modalities are offer feedback:

1. Visual

Visual feedback is achieved by various means. Small popup messages in the form

of customized Android Toasts or animations on the workout maps are some examples of

this feedback modality.

2. Audio

Audio modality is done recurring to a speech synthesizer which provides the

application with a powerful tool that “speaks out loud” any given text.

3. Haptic

Haptic feedback is conveyed by using the smartphones´ vibrator (implemented

through the Android APIs´ Vibrator class).

ExodUS provides its players with two distinct types of feedback: performance

updates and pep talks. The latter are relayed using audio cues only, taking advantage of

the speech synthesizer to “read out loud” a message sent to the player. Performance

updates strictly regard effort metrics and should be subdivided into three different

feedback subtypes:

[1] Basic

Basic performance updates regard performance fluctuation of the player while

executing the workout session. These updates make use of all three modalities.

[2] Goal

Goal performance updates are only used if the player chooses to select a specific

metric goal to beat, which can either be a distance goal, a time goal or an effort goal.

These updates make use of the audio modality only.

[3] Comparison

Comparison performance updates, are updates used while performing ghost

challenges and report direct comparison between the current performance and the ghost

84

partners´. These updates make use of all three modalities similarly to the basic

performance updates.

Figure 28 – ExodUS hierarchy of feedback types and respective modalities

As stated previously, the effort metric is used to determine which feedback

message to deliver to the player. Whether the effort value regards a ghost partner or a

basic setting, a weighted comparison is used. This is done to enable lesser fit players to

still be able to challenge ghost partners recorded by more fit players in a way which the

difference is not so easily noted. This weighted comparison is done by defining a

threshold value that is used to define an effort interval, broadening the comparison

between the current, single effort value and the target / previous effort value. The

threshold value can be defined as 5%, 10%, 25% or 50%.

Lastly, the configuration of the weighted comparison threshold value and the usage

of the audio and haptic modalities (visual modality is always enabled by default) are

available on the applications´ Settings screen.

Point Awarding

Point awarding is the only pure gaming concept implemented within ExodUS (e.g.

points are awarded depending on player performance). It was originally thought of as a

85

way to persuade players to push themselves further trying to obtain more points from

their workout sessions. This was however not pursued because it turned away from our

research intent of using persuasion through specific challenges, using different

modalities and careful handling of exertion data.

Main Study Features

Aside from the already described features, the mobile application offers the Ghost

and Live playing modes. These workout types are our main study features, designed to

influence players to improve their performances and motivation, through the use of

persuasive cues. These cues were employed based on the research studied in our related

work, meaning they take into account how persuasive technology is designed and make

use of persuasive social cues.

Ghost Challenges

Ghost challenges are selectable using the apps´ player history screen. Upon

selecting a target workout session, it is loaded from the local database. This loading

process occurs immediately prior to the exertion activity (e.g. before running) and

results in the creation of a dedicated exertion data structure for the ghost partner. During

play, the app compares the exertion data of the player and the ghost partner. This

comparison is triggered by the location updates the player obtains throughout the

session. This guarantees that both exertion data nodes are not “blank” and makes the

comparison as directly as possible. If the player outlasts the ghost partner, the app

informs the player that the ghost partner has finished the session.

Live Broadcasting

Live broadcasting is a feature that, although present in some MEA´s, is still

relatively unexplored. The players can decide to broadcast their sessions to their friends

by selecting the option on the start screen of the app. By doing so, the players´ friends

are informed of the happenings and can watch the players´ session, in a live fashion.

This feature relies on two main persuasive factors. One regards the awareness of the

player that he might be being watched. This can either add pressure or further the

players´ motivation to perform better (e.g. showing off to friends). The other factor is

86

the ability of the system that allows for the players´ friends to interact with him / her

while the activity is being performed. The spectators can either send written messages

that are “read out loud” to the player or send cheering / booing sounds (both require

audio cues to be enabled). The latter functions similarly to a common like feature from

social networks. The former, gives opportunity to convey persuasive messages that are

beyond the scope of designers and programmers such as particular jokes and other

meaningful interactions only known by the stakeholders. From the mobile application

perspective, broadcasting the session triggers the process of sending each location

update to the dispatch server, which then handles the distribution of the relevant data to

the spectators and the possible incoming interaction from them.

4.2 System Review

The developed system encompasses the common feature set of the studied MEA´s,

as well as accommodates our intended study aims. Persuasive cues are employed by

making use of the apps´ playing modes, as well as the feedback modalities incorporated

within the system. This system also covers most of the dimensions proposed in our

design space for MEA´s, as can be seen in table 6:

Table 6 – ExodUS system distribution in our design space

Delivery

Natural Rhythmic Other

˅ ˅ ×

Stakeholders

Spectators Opponent Self

˅ ˅ ˅

Time (Instant of Creation)

Before During After

˅ ˅ ×

Time (Instant of Reception)

Before During After

˅ ˅ ×

Space

87

Co-Located Distributed

˅ ˅

The Stakeholders dimension is covered with the possibility to role play either the

core player (basic, workout goal challenges and every other playing mode), the

opponent (ghost challenges) and the spectator (live broadcasting and viewing from the

website). Virtual partners and live spectators have the ability to intervene in the players´

workout sessions, potentially influencing their motivation through social interaction

(e.g. live cheering or competition with the “ghosts”). Real world opponents also exert

persuasion on the players when considering a co-located context.

Regarding the Delivery dimension, this system makes use of natural language

(visual and audio feedback), as well as rhythmic patterns (haptic feedback). The use of

natural language allows the app to apply persuasive social cues, such as playing a social

role and conveying a personality.

Spatiality wise, both co-located (same-time/same-place persuasion is guaranteed)

and distributed settings (spectators can cheer in geographically distributed locations and

opponents can represent ghost challenges executed on different locations) are available

through ExodUS.

Lastly, regarding the Time dimension, some persuasive cues are created and

delivered before the workout session takes place (e.g. ghost challenges), as well as

during the exertion activities (e.g. spectators input). Similarly, these cues are delivered

before (in the case of ghost challenges and messages associated with a particular

workout session) and during (in the case of the spectators input) the workout activities.

Performance messages are created and delivered both while the player is performing his

/ her workout session and co-located real world opponents challenge cues are the same

for this case.

88

Chapter 5

Experimentation with ExodUS

An experiment was conducted using ExodUS, whose primary goal encompassed

the testing of the persuasive factors built in the system. Analyzing whether players felt

more motivated using the available gaming modes was the aim. Although this system

deals with exertion activities, the hard physical data regarding performance was not

analyzed. Co-located and collaborative play can be evaluated through the use of

physiological measures [26, 27]. However, for this experiment, we are primarily

interested in seeing how persuasive cues affect motivation, through subjective reported

experience data. As stated earlier, motivation is tied with most of the fundamental

requirements for exergames´ success, them being persistence and incentive.

5.1 Experiment

The experiment consisted in obtaining a better understanding of the designed

modules, gauging their acceptance for an assisted jogging session. The main objective

was to take note of the possible benefits it could bring to the participants´ overall

motivation levels. This experimental period lasted for approximately two weeks,

involving one researcher supervising the tests and providing support to subjects as

required.

89

5.1.1 Goals

The primary goal of the experiment was to assess how players react to different

persuasive cues in the same situational context. We were particularly interested in

analyzing:

If feedback cues or lack of thereof, can influence the players´ motivation

How the different types of feedback cues impacted the players

Effectiveness of social persuasion inducing features

Compare the different persuasive in terms of the players´ perception

The feedback cues´ aim was to inform the players of their performance statistics

and exertion data, varying from simple performance updates to direct comparison in the

case of the presence of a ghost partner. Essentially, several persuasive mechanisms were

used, each addressing different facets of persuasion. Some inherit characteristics of

social persuasion (e.g. live broadcast mode and ghost challenges), others make use of

natural language so as to bring the players and system closer to each other (e.g. using a

speech pattern that give the idea of cheering and of a close relationship).

The secondary goal pertained with the identification of which of the kind and type

of persuasive cues were perceived by the players as more effective.

5.1.2 Methodology

A questionnaire was distributed among the subjects [see Annex II] with the goal of

accessing how they related to exertion experiences and mobile exertion applications.

Furthermore, the questionnaire focused on persuasive cues used in other commercial

MEA and on our own. This allowed for the mapping of these types of cues, as well as to

access the level of impact they had on players´ motivation and emotional state.

5.1.3 Participants

The participants were composed by 15 subjects, with ages ranging from 21 to 40

years old ( ̅ = 27.7) and mostly by male players, with 12 male participants (80%) and 3

female ones (20%). All participants freely volunteered for the experiment and since the

trials were physically demanding, they were offered a bottle of water.

90

Figure 29 – ExodUS Experiment participants´ gender distribution

Most had a distant relationship with exertion activities with six players exercising

sporadically (40%) and one who did not exercise altogether (7%). The remaining, were

more involved with exertion activities with five participants exercising with a frequency

of 2 or 3 times a week (33%) and three participants stating to exercise at least once a

week (20%).

Figure 30 – ExodUS Experiment participants´ exercise frequency

Regarding the preferred exertion activities, most players took part in Collective

games (41%), followed by Running / Jogging (35%) which was the preferred choice of

six players. Aikido, Cycling and Gym attendances were other activities with which the

players were involved. Another relevant aspect to note was the reasoning behind their

exercising activities. Keeping fit (26%), to deal with stress (26%) and as a social

activity (26%) were the more popular reasons to engage in exertion activities.

91

Figure 31 – ExodUS Experiment participants´ reasoning behind exertion activities

MEA´s were not used by most of the participants. Twelve players admitted to not

using MEA´s (75%) and only three used these applications. For these three players the

MEA´s in question were the Nike+ [29] and RunKeeper17

[36]. Analyzing this

relationship further, the three players that stated to use MEA´s appointed the ability to

show their progress as one of the reasons to use these kind of applications. The other

two reasons appointed for their use were that it motivated them to do better and that it

made exercising more fun.

Figure 32 – ExodUS Experiment participants´ reasoning behind their use of MEA´s

Furthermore, considering the answers provided by the remaining twelve

participants, six had no prior experience with MEA´s since additionally to not using

one, they also had no experience with the usual motivator features provided by these

apps. Also from this twelve player pool, the favored motivator features were the ones

92

that involve challenges (e.g. beat yourself and beat others) and the use of music (e.g.

music player).

5.1.4 Responsibilities

Responsibility wise, the participants were committed to jog along a pre-determined

course that the researcher decided on, while being assisted by the designed application.

All participants performed the same tasks and only assumed the role of the player

stakeholder. The researcher had the responsibility to clearly explain the tasks, later

assuming the spectator stakeholder during each trial. The researcher also took on the

opponent stakeholder role in two different contexts: one as a real world partner and

another as the virtual partner for the ghost challenge used in this experiment.

5.1.5 Equipment

Subjects were handed an Android mobile phone (HTC Desire), with the ExodUS

application. Players´ carried the smartphone on their waists via an accessory for this

purpose. Headphones were also used in the course of this experiment, to avoid possible

environmental noise. A notebook was used to access the web platform during the

experiment. The dispatch server was running on a cloud service.

5.1.6 Procedure

The experiment was conducted on a pre-determined course, namely at the Estádio

Universitário in Lisbon. This location was chosen because it is a place where people

usually jog and therefore met the criteria for the experiment. The course had a full

length of approximately 600 meters, located around a football field and an athletics

track (see figure 33).

93

Figure 33 – ExodUS Experiment Location

The scheduled tasks for each of the participants are depicted in the following table:

Table 7 – ExodUS Experiment Scheduled Tasks

Task Description Duration (approx.)

Briefing Explanation of the experiment 5 min

Jogging Laps Series of runs, testing different variables 25 min

Post Survey The participants filled out a survey

regarding the experiment 10 min

Total 40 min

The main tasks´ procedure was as follows: players were randomly assigned to a

series of running laps around the course, each with their own unique settings. This was

done to better assess the influence of the designed persuasive cues while making each

one “stand out” in a controlled context. Each participant ran a full five laps, resulting in

a total of seventy five laps traversed.

94

The five mandatory laps had the following settings:

Players ran a full lap without any kind of interference from the application.

They were just required to carry it alongside them. The purpose was to measure

how the presence of an assisting tool could influence them.

The second test consisted in running while the app provided audio and visual

feedback regarding the participants’ performance. This test was meant to

disclose how providing some kind of feedback influences the player.

Another test aimed to gauge if the app had any influence when the player is

running alongside a real world partner. Does the app work just as a medium to

augment the jogging experience?

Following the previous test, the next one focused on “ghost” partners. Players

were asked to run alongside a virtual partner representing a previous recording

done by the researcher (e.g. participants where effectively trying to beat or

cooperate with someone else).

The last test pertained the presence of spectators. The players broadcasted their

lap and received real time feedback from the researcher. Messages such as “Go

faster!” or “You are too slow!” and cheering / booing sounds where used.

To diminish fatigue influence and give them a rest period between tests, players

rotated among themselves. The tests´ order was assigned randomly for each player.

5.2 Results

Following the planned experiment documented in the previous subchapter, this

subchapter is dedicated to the report of the resulting findings obtained by conducting the

said experiment.

5.2.1 Findings

The results will be addressed separately for each lap as each had a specific study

aim.

95

MEA Presence only [no direct persuasive technique]

In this test, players carried the application with them while they ran, but no

feedback was conveyed to them. This meant that no direct persuasive cues were

employed in this context. As such, the objective was to assess how just the awareness of

having a MEA tracking the players could influence their motivation.

Figure 34 – MEA Presence only impact on player motivation

Figure 34 shows that the presence of a MEA tracking the players by itself is of no

consequence to the players´ motivation. For the majority of the players, this base

context did not impact their motivation (60%). From these results we can also conclude

that it can potentially have negative effects since four players stated to either feel very

little or no motivation at all while performing this test.

Performance updates [through audio cues]

This test featured the execution of a basic workout. While running, the players

were informed through audio cues about their performance. This context introduced

persuasive social cues with the clear goal of persuading the players to try harder:

Using a speech pattern that give the idea of cheering and of a close

relationship;

Playing the role of an advisor.

96

Figure 35 – Audio performance updates´ impact on player motivation

This test indicated that these persuasive cues had mainly a positive effect on the

players´ motivation with nine players increasing their motivation in this test (60%). Six

of them stated to feel more motivated (40%) and the remaining three expressed that they

felt greatly motivated (20%) with the implemented audio cues. In informal

conversations, a few participants stated that the “audio coach” was really harsh

(meaning it played its social role well) and that it made them feel competitive towards

it, making them strive to “prove it wrong”.

Co-Located Real World Opponent [with MEA presence]

Evaluating motivation impact when a co-located real world opponent is present

additionally to the use of the MEA in a basic workout mode was the aim of this test.

Without technological support, co-located persuasion is guaranteed through the

perception of how the opponents running with a player is performing and the

introduction of the MEA does not change that reality. It can however augment the

experience by offering other channels of communication between those involved in the

exertion activity.

Figure 36 – Co-located Real World Opponent plus MEA Audio performance updates´ impact on

player motivation

97

In figure 36 we can visualize that players motivation increased while performing

this test. Only two participants stated that their motivation was unaffected while

performing it (13%), while to the remaining participants it had a positive effect.

However these results are not conclusive when considering the effects caused by the

MEA only. In our previous survey, participants´ exertion habits study showed that

although people preferred to exercise by themselves, a significantly percentage stated to

prefer to exercise either with a friend / colleague (32%) or in a group (26%). Also,

competition is a powerful motivator when addressing these kind of applications. This is

testified by the players (who participated in this experiment) themselves, whom the

majority answered with “It made me feel more competitive” (60%) when asked how

they related to this context (running alongside a real world partner).

Figure 37 – Co-located Real World Opponent plus MEA Audio performance updates´ impact on

player emotional state

Virtual Opponent [‘Ghost’ Challenge]

Ghost challenges present themselves an important persuasive mechanism used in

MEA´s. This argument was evident in our previous survey, with competition against

oneself proving to be the most effective motivator feature available in MEA´s. For this

test, a ghost recording had been prepared beforehand, following the same course.

Throughout the lap, the players´ performance was compared with the ghost partners´

performance and this information was conveyed recurring to audio cues, much like in

the previous tests. Although the audio cues content was different than in other tests, the

persuasive social cues remained present: the app continued to play the social role of an

advisor, in a friendly manner, using even taunts to persuade the players to try harder

(e.g. when the players´ performance was worse than the ghost partners´ the audio cue

would feature the taunt “Don´t you want to win?”).

98

Figure 38 – Virtual Partner ‘Ghost’ Challenge impact on player motivation

This tests´ results were similar to the results obtained when running alongside a

real world partner, meaning that motivation increased for the majority of the players.

Seven stated to feel more motivated (47%) and four felt greatly motivated (27%) when

running alongside a virtual partner. However this test featured more players whose

motivation remained the same (27%). Once again competition proved to be a powerful

motivator in this context, as 14 players (93%) felt more competitive when performing

this ghost challenge.

Figure 39 – Virtual Partner ‘Ghost’ Challenge impact on player emotional state

Live Broadcast [with the presence of Spectators]

The final test involved a more pronounced persuasive element, introducing

spectators on to the action. This is intended as a form of persuasion through social

interaction that has much potential to influence the motivation values of the players, as

documented throughout this document. During this test, players broadcasted their

performance data, which was displayed in the web platform where the researcher was

watching. The messages sent where typically “Go faster!”, “Faster, faster, faster!” or

“You are too slow!” accompanied by cheering or boing sounds.

99

Figure 40 – Live Broadcast Challenge with the presence of spectators´ impact on player motivation

Results regarding this test were quite surprising with the majority of people stating

to feel no impact on their motivation levels while performing these tests (47%).

Furthermore one player stated that he felt very little motivation and another one

indicated that he felt almost no motivation while performing this test. Sorting this issue

further, the players were asked how their performance was impacted while performing

this test.

Figure 41 – Presence of spectators´ impact on player emotional state

The results shown in figure 41 are interesting when coupled with the results

presented in figure 40. Peer pressure seems to be an issue for four players (27%),

although not always a negative influence. The player that previously stated to feel

almost no motivation at all also felt pressured with being watched by spectators. The

same happened to the player that felt very little motivation during the test, who also

declared to feel pressured. On the other hand, pressure seemed to have a positive effect

on two other players. One expressed to be greatly motivated during the lap and the other

more motivated while feeling pressured from the presence of spectators.

100

Additionally, of the seven players that had their motivation unaffected during the

test, four remained oblivious to being watched by spectators while the remaining 3 felt

more motivation with just their presence. Lastly, another point worth noting is the fact

that a player who stated to feel more motivation during the test also answered with “It

was irrelevant” when asked the second question, meaning motivation must have been

influenced by the content of the messages used during the test a not from the presence

of spectators per se.

Audio Cue Effectiveness

The results gathered from the analysis regarding the five different tests performed

hint at the possibility of the audio cues´ content having an independent impact on

players´ motivation levels regardless of other persuasive elements being incorporated

into the context (e.g. presence of spectator or a virtual ghost partner), our hypothesis.

Figure 42 – Audio Cue effectiveness

We asked our players to judge the audio cues they heard during testing, by

choosing the type of cue that they thought was more helpful or interesting to them. As

expected, audio cues relaying trivial information were judged poorly. Interestingly, both

performance updates and comparison information were chosen by ten players as the

more effective audio cue (45%). This proves consistent with our hypothesis since both

message types make use of persuasive social cues but the comparison updates imply

another type of persuasive element in the presence of a ghost partner and they are still

judged to be equally effective by the players.

The live broadcasting mode had its very own audio cue subsystem. The spectators

interacted with the players using two different types of audio cues: written messages

and cheering/booing sounds. From the results gathered in the study of this live feature

101

we obtained an instance where the player felt more motivated but stated that the

presence of the spectators was irrelevant for him. In this case the motivation increase

should be the direct result of the audio cues relayed to the player during the test run.

Figure 43 – Live Broadcast Challenge Audio Cue effectiveness

Figure 43 shows how the two types of messages influenced players´ motivation

during the tests. Written messages had mixed results with seven players experiencing a

motivation increase (46%) and another seven stating that they made no impact on their

motivation (47%). Only one player felt that these messages hindered his motivation

level. Rhythmic sounds had a broader result set, with a slight positive influence. Seven

players had their motivation increased from hearing these messages (47%), while for

four of them motivation remained the same (27%). The remaining four participants had

their motivation dropped with this type of messages (26%). These results cannot prove

that, in the case that the presence of spectators is irrelevant and motivation still

increases, is a direct cause of the messages received by the players but in some cases, in

informal conversations, the players would say that hearing a cheer or an inside joke

gave them a boost during the trials.

102

Virtual vs. Real Opponent

Both tests regarding the use of a partner, virtual and real, showed a motivation

increase to a significant portion of the test subjects. In order to understand the

differences between the two, the players were asked which of them influenced them the

most in terms of three different aspects:

Motivation

Considering all of our research on this topic, the virtual ghost partner was expected

to be the playing mode that created more impact on the players´ motivation levels, since

it is intrinsically linked with exercising alone and creating competition (most times

against oneself). The results, although leveled, show that for this sample, jogging with a

real world partner impacted players´ motivation further. For this particular sample,

running alongside a real world partner also instilled a competitive edge comparable to

the other playing mode. Another interesting characteristic regarding the two is that

running alongside a real partner made three participants feel more relaxed while that did

not happen with the ghost challenge for any of the participants.

Feedback

Feedback messages were different on both of these playing modes. Virtual ghost

challenges had comparison updates only, while running alongside a real world partner

returned feedback in the form of performance updates independently from the presence

of the partner in the workout session. In this case, players valued more the comparison

feedback when challenging a ghost partner.

Perception (objective awareness)

Following on the results regarding feedback, the players´ objective awareness

proved to be more influenced with the use of a ghost partners. Direct comparison

between partners gives the players perception of how they are matching against / with

another partner. This perception is often guaranteed in co-located scenarios where the

comparison tasks are shifted to the players themselves (e.g. direct comparison by visual

input, etc.). In these trials, this also happened, with the app being oblivious to the

presence of a real world partner could not provide comparison information, unlike with

the virtual partner.

103

Figure 44 – Virtual vs. Real Partner impact on players

5.2.2 Anecdotal Evidence

All players enjoyed participating in the trials (“it was a good experience”), but

pointed out a few faults in the current version of the application. Informal conversations

held with the participants during trials exposed a few interesting points.

The first relates to visual feedback, with a few subjects admitting to only peek at

the screen one or two times during the laps. The reason for this was that it is not

practical to stare at the smartphones´ screen while running. This did not hinder our goals

since, apart from the first described test (e.g. no feedback), all messages used audio cues

whether they were performance updates, comparison messages in regards to the virtual

partner or messages from the spectators. It does however raise questions regarding the

effectiveness of visual cues (with persuasive factors included) in MEA´s.

The absence of music was also a big issue for the participants, which is reasonable

since most players indicated that their favored motivator feature available in MEA´s

were the ones that involve challenges (e.g. beat yourself and beat others) and the use of

music (e.g. music player). Also, one player indicated that having some kind of

104

background music would help when he is performing his exertion activities. This is

consisted with the previous research done regarding MEA existing systems.

5.3 Experiment Limitations

Exertion data handling, namely how it was displayed, raised a few concerns.

During the tests a few players expressed their concern that it was important to them to

have a few statistics always available to them (e.g. chronometers´ time, among others)

and even requested that the pace statistic should be added to the app (it is not available

in this version), not only for reference but as well as “a way to improve objective

awareness”.

The live broadcast feature proved not be limited to the context tested in this

experiment because the magnitude of variables that play a part in this scenario cover a

wide enough research spectrum to be granted further testing and investigation (e.g.

nature of the relationship between player and spectators, number of simultaneous

spectators, message content, …). One particular variable identified by the players who

took part in the experimental procedure is the process of identifying a spectator and

their messages. Some players stated that they found difficult to distinguish between the

“audio coach” messages and the spectator written messages, even though the app used

the default notification received sound from the Android operating system to signal the

spectators´ messages followed by the username of the spectator who wrote the message.

The audio updates frequency was an issue, to the point that one player considered it

annoying. Other players expressed the need to have more feedback when on a steady

pace.

5.4 Experiment Discussion

Taking into consideration the data gathered from this experiment, we can verify

that the use of persuasive cues effectively influences players´ motivation, contributing

to its increase. The different tested persuasive elements proved to add to the motivation

of the players, with the same relative success.

MEA´s are appreciated because of three main reasons: they show players their

progress, motivate them to do better and make exercising more fun. The first is in

105

agreement with our idea that careful handling of the data displayed to the player may

influence him / her in some way.

Contributing to motivation, is the responsibility of the available motivator features

present in the app. Focusing on our main study aim features, the ghost challenges and

live broadcasting modes, both were able to contribute to the players´ motivation, which

is coherent with our previous research as well as our hypothesis.

Regarding social interaction, the results seem to be consistent with other research

regarding persuasive social cues [13]. In this document we had already stated that social

interaction is affected by previous relations between players and this was transitive to

this work, when the usage of some private jokes in between the programed social

messages used for the live broadcast feature appeared to have a better impact than the

rest. Social pressure is also a very strong point to considerer in these cases but the result

should depend entirely on the individuals using the app (on how they deal with

pressure). This is linked to how, in this test, some of the players that felt pressured had a

negative influence on their motivation level and others enjoyed an increase in

motivation thanks to it.

The live broadcasting essentially accomplished its goal, which was to persuade

players to do better and increase motivation. The use of rhythmic sounds as a way of

delivery had mixed results but similarly to the social pressure issue, it seems to be

individual dependent as one of the players confided that she liked the cheering sound

and it contributed greatly to her motivation.

The ghost challenge mainly contributed to incite competition with no signs of

cooperative play being shown. This can be explained by the fact that they were

accompanied by a prerecorded workout session executed from another person and the

fact that the majority of the players admitting to have a very competitive personality (in

informal conversations).

Performances messages also impacted motivation, which blends well with our idea

of careful handling of the exertion data, as well as the use of persuasive social cues.

This is another scenario which covers a very wide research area focusing on persuasive

social cues (refer to Foggs´ works and chapter 2). One particular player stated that “all

the audio cues didn´t have any effect on me because I was already doing as much as I

could and because of that my motivation remained the same”, which implies that just

reporting a few metrics and trying to use persuasion to influence players to do better

106

might turn into an “asking the impossible” situation, meaning that these elements should

be made “aware” of the players stated for both effectiveness and safety measures.

Real world, co-located opponents are a scenario in which we can witness

motivation increase. This is however ambiguous when evaluating the MEA

responsibility. From the results gathered, this increase seems to be responsibility of the

presence of the real world partner as a persuasive mechanism and not because of direct

influence of the MEA and its persuasive cues. Multiplayer features and other features

that might augment social interaction might contribute further in this case as well as

splitting the responsibility to increase motivation levels between the normal social

interactions and the action of the MEA.

The results obtained through this experiment clearly show that the presence of an

assisting tool, without the employment of any additional persuasive cue is insufficient

when aiming to increase the players´ motivation. One player even remarked that “I

could feel the difference having the cheering messages and not having anything at all.

The run that I didn´t have any updates from the app, was the run that I had the least

motivation”.

107

108

Chapter 6

Conclusions & Future Work

This work addressed the comprehension of persuasive features present in

exergames and mobile exertion applications. We performed an analysis of persuasion

techniques available on existing MEA´s and from its results we defined a design space

for persuasive mobile exertion applications consolidating the existing expertise about

persuasion and MEA´s. The analysis enabled the characterization of how people interact

with exertion activities using this kind of mobile tools. This characterization was

derived from the creation of a survey, aimed at assessing MEA usage habits and how

useful individuals perceive the persuasive features. The review of existing MEA´s, with

the intent of identifying the categories and dimensions which would comprise the

design space, completed the analysis. For MEA´s (and exergames) the ability to

persuade players to adopt and maintain exertion activities is a key requirement for

success. This requirement is met by the employment of persuasion techniques. To study

this assumption we developed ExodUS, a MEA for Android devices. It utilizes several

persuasive techniques designed to influence players´ motivation. The application allows

for the use of several playing modes, each with their own persuasive cues enabled. An

experiment using ExodUS was conceived intending to test the implemented persuasive

factors within it, aiming to analyze whether players felt more motivated using the

available gaming modes.

All in all, the diversity among persuasive features found in MEA´s contributes to

the lack of a formal structure to classify them. Despite more persuasive features being

found in the course of this work, it allowed us to identify at least four commonly used

types of persuasive features in MEA´s: performance updates, real world partner

challenges, virtual partner challenges and live broadcast events. The conclusions drawn

109

from using these cues are that players´ motivation can be influenced by persuasive cues

independently of the type used. Most persuasive cues make use of social interactions

and social behavior to exert their influence on players´ emotional states and incite a

desired behavior which often leads to either a positive or negative impact on the

players´ motivation. Through this study we can state that MEA´s which do not employ

any persuasive cues either do not affect or hinder the players´ motivation.

6.1 Future Work

Persuasive techniques possess a very large scope of possibilities which warrant

designers and researchers with many possibilities to work with. Focusing on the diverse

set of modalities granted by current and new mobile devices, social presence, persuasive

social acting or exergaming are fields of research that could expand this work.

The developed MEA and this works results are a necessary iteration in the process

of designing a persuasive technology that aims to improve exertion experiences. Our

next endeavors will encompass the gathered data and attempt to improve upon this

work. Overcoming the designed systems´ limitations is a step forward in this research.

Further validation of our proposed design space is also an ongoing process.

ExodUS did not cover the entire spectrum presented in our design space. Furthermore,

the scenarios present in section 3.2.3 are yet to be tested.

In this work we also approached physical outcomes, which are in their genesis,

connected to MEA´s and exergames alike. Our designed solution barely touched upon

the (quantifiable) exertion data related to this field. Our intent is to make use of

additional hardware in order to extend this research to cover both psychological and

physical outcomes.

110

Bibliography

[1] Weinberg, R.S. and Gould, D. Foundations of Sport and Exercise Psychology.

Human Kinetics, 2006.

[2] Physical Activity and Health: A Report of the Surgeon General. HHS, 1996.

[3] Mueller, F., 2009. Exertion in Networked Games. In Proceedings of the 4th

International Conference on Foundation of Digital Games (FDG, 2009). ACM, New

York.

[4] Gao, Yue, Mandryk, Reagan, 2012. The Acute Cognitive Benefits of Casual

Exergame Play. In Proceedings of the SIGCHI Conference on Human Factors in

Computing Systems (CHI, 2012). ACM, New York.

[5] Mueller, F., Edge, Darren, Vetere, Frank, Gibbs, Martin R., Agamanolis,

Stefan, Bongers, Bert, Sheridan, Jennifer G., 2011. Designing Sports: A Framework for

exertion Games. In Proceedings of the SIGCHI Conference on Human Factors in

Computing Systems (CHI, 2011). ACM, New York.

[6] Gerling, Kathrin, Livingstone, Ian, Nacke, Lennart, Mandryk, Regan, 2012.

Full-body Motion-based Game Interaction for Older Adults. In Proceedings of the

SIGCHI Conference on Human Factors in Computing Systems (CHI 2012). ACM, New

York.

[7] Lindley, Siân E., Le Couteur, James, Berthouze, Nadia L., 2008, Florence,

Italy. Stirring Up Experience Through Movement in Game Play: Effects on

Engagement and Social Behavior. In Proceedings of the SIGCHI Conference on Human

Factors in Computing Systems (CHI 2008). ACM, New York.

[8] Whitehead, Anthony, Johnston, Hannah, Nixon, Nicole, Welch, John.

Exergame effectiveness what the numbers can tell us_2010_New York In Proceedings

111

of the 5th ACM SIGGRAPH Symposium on Video Games (Sandbox, 2010). ACM, New

York.

[9] Adams, Marc A., Marshall, Simon J., Dillon, Lindsay, Caparosa, Susan,

Ramirez, Ernesto, Phillips, Justin, Norman, Greg J., 2009. A Theory-based Framework

for Evaluating Exergames as Persuasive Technology. In Proceedings of the 4th

International Conference on Persuasive Technology (PERSUASIVE, 2009).

[10] Fogg, BJ, 2002. Persuasive Technology: Using Computers to Change What

We Think and Do. In Ubiquity Magazine (2002).

[11] Fogg, BJ, 2009. Creating Persuasive technologies: An Eight-Step Design

Process. In Proceedings of the 4th International Conference on Persuasive Technology (PERSUASIVE, 2009).

[12] Ishii, Hiroshi, Wisneski, Craig, Orbanes, Julian, Chun, Ben, Paradiso, Joe,

1999. PingPongPlus: Design of an Athletic-Tangible Interface for Compputer-

Supported Cooperative Play. In Proceedings of the 1999 ACM SIGCHI Conference on

Human factors in computing systems (ACM, New York, USA, 1999).

[13] Mueller, F., Vetere, Frank, Gibbs, Martin R., 2009. Design Influence on

Social Play in Distributed Exertion Games. In Proceedings of the SIGCHI Conference

on Human Factors in Computing Systems (CHI, 2009). Boston, USA.

[14] Isbister, Katherine, Schwekendiek, Ulf, Frye, Jonathan, 2011. Wriggle an

exploration of emotional and social effects of movement. In Extended Abstracts on

Human Factors in Computing Systems (CHI EA, 2011). New York.

[15] Lindley, Siân, Le Couteur, James, Bianchi-Berthouze, Nadia, 2008. Stirring

up Experience through Movement in Game Play: Effects Engagement and Social

Behavior. In Proceedings of the SIGCHI Conference on Human Factors in Computing

Systems (CHI, 2008). New York.

[16] Kazumoto, Tanaka, Parker, Jim, Baradoy, Graham, Sheehan, Dwayne, Holash,

John R., Katz, Larry, 2012. A comparison of Exergaming Interfaces for Use in

Rehabilitation Programs and Research. In Journals.sfu.ca, 2012.

[17] Sinclair, Jeff, Hingston, Philip, Masek, Martin, 2007. Considerations for the

design of exergames. In Proceedings of the 5th

International Conference on Computer

112

Graphics and Interactive Techniques in Australia and Southeast Asia (GRAPHITE,

2007).

[18] Sinclair, Jeff, Hingston, Philip, Masek, Martin, 2009. Exergame development

using the Dual Flow Model. In Proceedings of the 6th

Australasian Conference on

Interactive Entertainment (IE, 2009).

[19] Csikszentmihályi, Mihály, 1975. Beyond boredom and anxiety. Jossey-Bass

publishers, 1975.

[20] Csikszentmihályi, Mihály, 1990. Flow: The Psychology of Optimal

Experience. Harper and Row, New York.

[21] Sweetser, Penelope, Wyeth, Peta, 2005. GameFlow:a model for evaluating

player enjoyment in games. In Computers and entertainment (CIE) – Theoretical and

Practical Computer Applications in Entertainment, 2005.

[22] Chris Veness, 2012. Calculate distance, bearing and more between Latitude /

Longitude points. In Movable-Type.co.uk, 2012.

[23] Weisstein, Eric W., 2012. Spherical Trigonometry. From MathWorld – A

Wolfram Web Resource, 2012.

[24] Keytel, L. R., Goedecke, J. H., Noakes, T. D., Hiiloskorpi, H., Laukkanen, R.,

van der Merwe, L., Lambert, E. V, 2005. Prediction of energy expenditure from heart

rate monitoring during submaximal exercise. In Journal of Sports Sciences, 2005.

[25] Perloff, R. M., 2003. The Dynamics of Persuasion: Communication and

Attitudes in the 21st Century. New Jersey, Lawrence Erlbaum Associates, Inc.

[26] Mandryk, Reagan L., Inkpen, Kori M., 2004. Physiological indicators for the

evaluation of co-located collaborative play. In Proceedings of the 2004 ACM

Conference on Computer Supported Collaborative Work, 2004.

[27] Mandryk, Reagan L., Inkpen, Kori M., Atkins, M. Stella, 2006. A Continuous

and Objective Evaluation of Emotional Experience with Interactive Play Environments.

In Proceedings of the SIGCHI Conference on Human Factors in Computer Systems

(CHI, 2006).

[28] Endomondo. In www.endomondo.com, 2012.

[29] Nike+. In nikeplus.nike.com/plus, 2012.

113

[30] Adidas miCoach. In www.adidas.com/us/micoach, 2012.

[31] Runtastic. In www.runtastic.com/en/apps/runtastic, 2012.

[32] Strava. In www.strava.com, 2012.

[33] ZombiesRun. In www.zombiesrungame.com, 2012.

[34] GetRunning. In splendid-things.co.uk/getrunning, 2012.

[35] FitBit. In www.fitbit.com/android, 2012.

[36] RunKeeper. In runkeeper.com, 2012.

[37] MapMyRun. In www.mapmyrun.com/app, 2012.

114

Annex I

Assessing MEA Usage Questionnaire

115

116

117

118

119

120

121

122

Annex II

ExodUS Experiment Questionnaire

123

124

125

126

127

128