Quantified Self - An Exploratory Study on the ... - Semantic Scholar

Quantified Self - An Exploratory Study on the Profiles and Motivations of Self-Tracking

Bachelor Thesis

by Marcia Nißen January 31st, 2013

Karlsruhe Institute of Technology (KIT) Department of Economics and Business Engineering Institute of Information Systems and Marketing (IISM)

ii

“Identity is our mystery. We have no idea who we are – what humans are, and what humans are good for. […] Self-tracking and the Quantified Self movement are contemporary probes into this mystery, part of our feeble attempt to figure out who we are – as individuals and a collective. Quantifying your self is an act of self-assertion. All this attention is not a narcissist adoration of the self, but a self-definition in an age of great uncertainty about who we are.”

QS co-founder Kevin Kelly, 2011

Abstract

iii

Abstract The revolutionary rise of smartphones worldwide since the iPhone’s release in 2007 together with today’s ubiquitously available internet has changed the relationship between humans and their technical devices and changes still. All of the new possibilities through technology has opened up a world that offers new accesses to get to know oneself. The Quantified Self (QS) movement evolved alongside these technological developments, which is a community of people who use the capabilities of technical devices to gain a profound understanding of collected self-related data. With the help of smartphones, tablets or desktop computers, selftrackers keep record of everything in their lives and thus try to increase their self-knowledge.

This bachelor thesis sheds light on the profiles and motivations for self-tracking - within the scope of the Quantified Self (QS) development. The paper will start by giving a detailed theoretical introduction into the QS community since 2007 and a literature-based review of the latest preliminary findings from motivational theories. Afterwards, the methods and findings from a survey on self-tracking profiles and motivations conducted on 150 self-trackers worldwide will be described and eventually critically evaluated.

With due regard to the findings, a fundamental distinction of two different self-tracker profiles helps to highlight the differences between various self-tracking motivations. On one hand, WELL-BEING ONLY self-trackers are healthy in general but nonetheless try to improve for example their fitness, looks, well-being, performance or effectiveness (etc.) apart from their good health condition. On the other hand, WELL-BEING AND HEALTH self-trackers usually report chronic conditions and focus on tracking health data in addition to their tracking of daily-life activities.

Concerning the self-tracking motivation, a self-tracking motivation model with five motiatoin has been developed with the help of an Evolutionary Factor Analysis (EFA). This Five-Factor Framework of Self-Tracking Motivations may help future researchers to understand and study the most important motivations of self-tracking. The model includes a question set of 19 questions by which the motivational aspects of self-designing, self-disciplining, entertaining, self-associating and self-healing prospects of self-tracking can be measured. Keywords: Quantified Self, community, self-tracking, survey, self-tracking profiles, well-being, health-related, motivations, self-discipline, entertainment, self-association, HealthStyle, SelfDesign, hurdles, deployed technologies

Table of Contents

iv

Table of Contents

Abstract ........................................................................................................................ iii Table of Contents .......................................................................................................... iv List of Abbreviations ..................................................................................................... vii List of Figures .............................................................................................................. viii List of Tables .................................................................................................................. x

1

Introduction ............................................................................................................ 1 1.1 Motivation ................................................................................................................... 1 1.2 Outline of this work ..................................................................................................... 2

2

What is Quantified Self? .......................................................................................... 3 2.1 The credo “Self-knowledge through numbers” / “Know thyself” ............................... 4 2.2 Community development............................................................................................ 4 2.3 Self-tracking activities ................................................................................................. 6 2.4 Quantified Self in comparison to LifeHacking and LifeLogging ................................... 6 2.5 Market penetration of smartphones and health apps ................................................ 7

3

Theoretical Background: Motivational Theories ....................................................... 9 3.1 Motivational Theories in general ................................................................................ 9 3.1.1 Intrinsic motivation within the scope of self-tracking activities ............................ 9 3.1.2 Extrinsic motivation within the scope of self-tracking activities.......................... 10 3.2 Motivations to participate in online communities .................................................... 11 3.3 Presumed motivations for self-tracking .................................................................... 15

4

Survey: Methodology and Evaluation..................................................................... 18 4.1 Methodological approach and survey design ........................................................... 18 4.1.1 Theoretical approach ........................................................................................... 18 4.1.2 Expert interviews ................................................................................................. 19

Table of Contents

v

4.1.3 Questionnaire conceptualization ......................................................................... 19 4.1.4 Online implementation and distribution of the survey ....................................... 23 4.1.5 Data evaluation and representation .................................................................... 24 4.2 Evaluations ................................................................................................................ 25 4.2.1 Data retrieval........................................................................................................ 25 4.2.2 General demographic factors............................................................................... 26 4.2.3 Object of tracking ................................................................................................. 31 4.2.4 Scope of activity ................................................................................................... 46 4.2.5 Technology ........................................................................................................... 55 4.2.6 Motivations .......................................................................................................... 57 4.2.7 Hurdles ................................................................................................................. 73 5

Conclusion ............................................................................................................ 74 5.1 Synopsis ..................................................................................................................... 74 5.2 Discussion .................................................................................................................. 80 5.3 Critical review of self-tracking motivations............................................................... 83 5.4 Prospect..................................................................................................................... 85 5.5 Limitations of this thesis and future work ................................................................ 86

References ...................................................................................................................... I Appendix........................................................................................................................V I.

Questionnaire “Quantified Self - Self-Tracking Motivations” ..................................... V Qualifying questions … ..................................................................................................... V Object of tracking ............................................................................................................ VI Scope of activity ............................................................................................................. VII Motivation for self-tracking .......................................................................................... VIII Technologies .................................................................................................................... X Hurdles ............................................................................................................................. X Personality ....................................................................................................................... XI

Table of Contents

vi

Demographics ................................................................................................................ XII Eventually … .................................................................................................................. XIII Thank you! ..................................................................................................................... XIII II.

Free Text Input Field Answers (Question 11) ........................................................... XIV

III.

Additional Tables ..................................................................................................... XXV

Trademarks ..............................................................................................................XXVII Statement of authorship ..........................................................................................XXVIII

List of Abbreviations

List of Abbreviations BMI

Body Mass Index

EFA

Exploratory Factor Analysis

FZI

Forschungszentrum Informatik

FB

Facebook

IISM

Institute of Information Systems and Marketing

KIT

Karlsruhe Institute of Technology

PA

Parallel Analysis

QS

Quantified Self

vii

List of Figures

viii

List of Figures Figure 1 - The official Logo of the Quantified Self Community includes the community’s credo “Self-knowledge through numbers” (www.quantifiedself.com, retrieved November 17, 2012) 3 Figure 2 - Quantified Self Meetup Groups worldwide (http://quantified-self.meetup.com/, retrieved January 28, 2013). ......................................................................................................... 5 Figure 3 - The Reader-to-Leader Framework by PREECE & SHNEIDERMAN (2009). “The thickness of the green arrows and smaller shapes indicate the decreasing number of people who move from one form of participation to another. The thin grey arrows indicate how people can also move in a non-linear fashion to participate in different ways”, (Preece & Shneiderman, 2009, p. 16). .............................................................................................................................................. 13 Figure 4 - Classification of the Objects of Tracking based on the categorization of self-tracking possibilities by Nißen (2012, p. 14). ............................................................................................ 21 Figure 5 - Composition and derivation of the finally used total of responses after extracting all 411 responses from LimeSurvey® on December 1, 2012. .......................................................... 26 Figure 6 - Age distribution of all 150 respondents ...................................................................... 27 Figure 7 - Gender distribution of all 150 subjects ....................................................................... 28 Figure 8 - Status of employment of all 150 respondents ............................................................ 28 Figure 9 - Average origin of all 150 respondents ........................................................................ 29 Figure 10 - Histogram of the distribution of the disposable monthly household income of all 150 respondents. The default income classes have been provided based on an income classification by Destatis (Statistisches Bundesamt, 2008, S. 4). ................................................ 31 Figure 11 - Venn-Diagram for people that only track well-being related parameters (group WELL-BEING ONLY, here: light green) and people that track well-being as well as health-related parameters (WELL-BEING AND HEALTH, here: dark green) .............................................................. 35 Figure 12 - Coherence of people that indicated that they suffer from a chronic disease (‎Question 4) and respondents that indicated tracking health-related parameters (‎Question 5) ..................................................................................................................................................... 37 Figure 13 - Comparison of tracked parameters by WELL-BEING ONLY and WELL-BEING AND HEALTH self-trackers ................................................................................................................................. 39 Figure 14 - Tracked body parameters ......................................................................................... 40 Figure 15 - Comparison of tracked medical parameter for WELL-BEING ONLY and WELL-BEING AND HEALTH self-trackers (all

-Test; df = 1; p < 0.05). ...................................................................... 43

Figure 16 - Histogram of years of conducted self-tracking activities .......................................... 49

List of Figures

ix

Figure 17 - Minutes spent on own daily self-tracking in comparison for the two reference groups WELL-BEING ONLY and WELL-BEING AND HEALTH ................................................................... 52 Figure 18 - Comparison of time for actual self-tracking and time for self-tracking related activities by median and mean for all respondents, WELL-BEING ONLY and WELL-BEING AND HEALTH self-trackers ................................................................................................................................. 54 Figure 19 - Screeplot of unrotated eigenvalues of all 31 items after EFA and PA. The point of intersection can be found at five factors and indicates the optimum amount of factors for the factor analysis. ............................................................................................................................ 63

List of Tables

x

List of Tables Table 1 - First structure of question categories for the later questionnaire in preparation for the expert interviews .................................................................................................................. 20 Table 2 - Presumed underlying motivations of self-tracking (sorted into extrinsic and intrinsic motivations). Order and applicability are under consideration and part of the research of this bachelor thesis ............................................................................................................................ 22 Table 3 - Age distribution in age groups ..................................................................................... 27 Table 4 - Total itemization of the origin of all respondents ........................................................ 30 Table 5 - Comparison of reasons of starting self-tracking for people that have started selftracking before and after 2010 (multiple answers possible) ...................................................... 32 Table 6 - Amount of people suffering from a chronic disease or not suffering from a chronic disease ......................................................................................................................................... 33 Table 7 - Manually counted answers from free input field in ‎Question 4 “Do you suffer from a chronic disease?” ........................................................................................................................ 33 Table 8 - Self-assessment of tracked parameters by all 150 respondents ................................. 34 Table 9 - Statistical measures on the age distribution of WELL-BEING ONLY and WELL-BEING AND HEALTH (T-Test; p = 0.003). .......................................................................................................... 36 Table 10 - Comparison of triggers to start self-tracking of WELL-BEING ONLY and WELL-BEING AND HEALTH trackers, multiple answers possible (

-Test; df = 4; p = 0.005)..................................... 36

Table 11 - Statistical measures on the amount of tracked parameters per participant for all reference groups (T-Test; p < 0.001). .......................................................................................... 38 Table 12 - Tracked physical activities’ parameter for all reference groups ................................ 38 Table 13 - Tracked body parameters for all reference groups.................................................... 40 Table 14 - Tracked nutrition parameters for all reference groups ............................................. 41 Table 15 - Tracked well-being related parameters for all reference groups .............................. 41 Table 16 - Tracked addiction parameter for all reference groups .............................................. 42 Table 17 - Tracked medical parameters for all reference groups ............................................... 43 Table 18 - Tracked environmental parameters ........................................................................... 44 Table 19 - Tracked relationship related parameters ................................................................... 44 Table 20 - Tracked parameter in the category "Other" .............................................................. 45 Table 21 - Average respondents per parameters for each tracking category ............................ 45 Table 22 - Comparison of average tracked parameters of WELL-BEING ONLY and WELL-BEING AND HEALTH .......................................................................................................................................... 46

List of Tables

xi

Table 23 - Kind of self-tracking activities for all reference groups (multiple answers possible) 48 Table 24 - Descriptive analysis of so far length of self-tracking for all respondents, WELL-BEING ONLY and WELL-BEING AND HEALTH (Mann-Whitney U-Test, z = -0.617, p = 0.537, 2-tailed) ......... 49 Table 25 - Descriptive analysis of daily time spent with actual self-tracking for all respondents, WELL-BEING ONLY and WELL-BEING AND HEALTH (Mann-Whitney U-Test, z = -2.584, p = 0.010, 2tailed) inclusive respondents that indicated tracking more than 60mins per day ..................... 50 Table 26 - Frequency scale of minutes spent with daily self-tracking for all respondents, WELLBEING ONLY and WELL-BEING AND HEALTH self-trackers ................................................................... 51

Table 27 - Descriptive analysis of daily time spent on self-tracking related activities for all respondents, WELL-BEING ONLY and WELL-BEING AND HEALTH (Mann-Whitney U-Test, z = -3.082, p = 0.002, 2-tailed) inclusive all respondents that indicated spending more than 60mins per day on self-tracking related activities ................................................................................................ 53 Table 28 - Time spent on self-tracking related activities per day (classified) ............................. 53 Table 29 - Comparison of the time spent with actual self-tracking and time spent on selftracking related activities by means of median and mean for all reference groups (MannWhitney U-Test, z = -3.644, p < 0.001 , 2-tailed)......................................................................... 54 Table 30 - Mean of total time exposure to self-tracking activities (means of the sum of selftracking and self-tracking related activities) for all reference groups (Mann-Whitney U-Test, z = -3.644, p < 0.001 , 2-tailed). ........................................................................................................ 54 Table 31 - Deployed technologies for self-tracking .................................................................... 55 Table 32 - Self-tracking expenditures for all reference groups ................................................... 56 Table 33 - Self-tracking expenditures: Comparison of all references groups (Mann-Whitney UTest, z = -1.016, p = 0.310, 2-tailed). ........................................................................................... 57 Table 34 - Mean overall motivation for all reference groups (T-Test; p = 0.1139). .................... 58 Table 35 - General Statistics for all motivation items at a glance for all 150 respondents ........ 59 Table 36 - Correlation matrix of all motivation items ................................................................. 61 Table 37 - Rotated component matrix of the exploratory factor analysis on all motivations after five iterations of filtering. Extraction Method: Principal Component Analysis. Rotation Method: Varimax with Kaiser Normalization. ............................................................................................ 66 Table 38 - Mean motivation for all five factors and comparison of all reference groups .......... 72 Table 39 - Satisfaction with own amount of self-tracking .......................................................... 73 Table 40 - Hurdles of self-tracking .............................................................................................. 73 Table 41 - Synopsis of demographical backgrounds for all reference groups ............................ 74 Table 42 - Synopsis for "Objects of Tracking" for all reference groups ...................................... 75 Table 43 - Synopsis of the Scope of Activities for all reference groups ...................................... 76

List of Tables

xii

Table 44 - Synopsis of deployed technologies for all reference groups ..................................... 77 Table 45 - Synopsis of Motivations for all reference groups ...................................................... 78 Table 46 - Synopsis of the mean accordance for all reference groups with every motivational factor of self-tracking as retrieved by an Exploratory Factor Analysis on all motivations’ question that had been scored on a Likert scale ranging from 1 (Strongly Disagree) to 5 (Strongly Agree). ......................................................................................................................... 79 Table 47 - Synopsis of all "Hurdles" ............................................................................................ 79 Table 48 - Frequency scale of tracked items per person for all respondents compared to those of the WELL-BEING ONLY group as well as the WELL-BEING AND HEALTH group .............................. XXV Table 49 - Deployed technologies of respondents using exclusively only one single type of technology ................................................................................................................................. XXV Table 50 - Detailed frequency Scale of the distribution of time spent on self-tracking related activities with changing class widths........................................................................................ XXVI

Introduction

1

1 Introduction 1.1 Motivation In recent years, the possibilities of keeping records of daily activities have increased remarkably, alongside the technological development and spread of smartphones. “Quantified Self” is the name of an internet community and refers to a movement focused on selfquantifying. Participants of the Quantified Self (QS) movement measure everything in their lives from: heart rate, running exercises, body weight, blood pressure, clouds, relationships, coffee consumption, sleep quality and duration, etc. - in effect almost everything that can be put into numbers.

One motivation for this thesis has been to understand the various types of self-tracking: Creative underlying questions such as “Do I sleep better, when there are more clouds in the afternoon” are as typical, as health-oriented questions such as “Do my chronic symptoms get better when I lose a few pounds?”. For example NICHOLAS FELTRON is a known self-tracker who has completed several comprehensive annual reports of more or less random facts about his life1 or SETH ROBERTS who uses the methods of self-tracking for self-experimentations in order to understand the coherencies in his life for example between his sleep, mood, health, and weight (Roberts, 2004, p. 1).

Another motivation has been to understand what self-trackers are driven by. There are selftrackers that suffer from chronic diseases and conduct self-tracking activities not for fun but as a means of improving their health conditions. Meanwhile, there are millions of people commonly tracking their running exercises or recording their body measurements and mood factors – all with the help of smartphones and numerous health- and well-being self-tracking apps. So, who actually are these people and what are they driven by?

Until today there has only been a little research done on the QS movement and respectively on self-tracking. To get an overall idea of what QS and self-tracking are all about, this thesis presents the results of an exploratory survey on profiles and motivations of self-trackers which were conducted in November 2012.

1

http://feltron.com/ar11_01.html, retrieved January 19, 2013.

Introduction

2

1.2 Outline of this work For a common understanding of this bachelor thesis, the first section (chapter 2) gives a detailed introduction into the Quantified Self (QS) community by reflecting the idea behind QS, its development, its virtual and real-life organization as well as its relation to the evolvement of new technologies such as smartphones and apps since 2007.

The following section, chapter 3, gives an overview of the theoretical fields of motivations in general and those for participating in online communities such as e-learning communities, knowledge sharing platforms or social networks.

Finally chapter 4 deals with the methodological process of the survey and all evaluations on the profiles and motivations of self-trackers. Therefore a fundamental distinction of selftracking types will be made and two reference groups will be introduced for all further analyses: WELL-BEING ONLY self-trackers and WELL-BEING AND HEALTH self-trackers. The distinction of these types highlights the existence of self-trackers that suffer from chronic disease and other self-trackers that are healthy in general but still try to improve their health/wellbeing/performance/effectiveness/etc. by self-tracking. Apart from that also an Exploratory Factor Analysis (EFA) of all questions concerning the motivations of self-tracking will help to develop a Five Factor Framework of Self-Tracking Motivations. This motivation model is designed to help researchers, designers and managers to understand what self-trackers are motivated by. By understanding the motivations of self-tracking several further approaches can be explored in future research: whether economically, socially or psychologically.

The last section, chapter 5, summarizes the evaluations, discusses all findings critically and gives suggestions for testing the framework and for further research on self-tracking.

What is Quantified Self?

3

2 What is Quantified Self? The term of Quantified Self (QS) evolved in 2007, when the American journalists and publishers of The Wired Magazine, Gary Wolf and Kevin Kelly, founded the Blog QuantifiedSelf.com, which remains the most important website and central hub within the QS Community today (Economist, 2012). In 2008 they started organizing meetings to discuss the personal and social impact of the smartphone development and other new technologies: “The intended focus of QS was, how people are using new technologies, how they are incorporating them into their lives, and the changes that occur. Rather than being about the technology itself (the fact that there are new devices that do new things), they were interested in what it means when people use the technology” (Butterfield, 2012).

Today, the term Quantified Self addresses two dimensions of this movement. It first includes the idea of garnering knowledge about oneself by quantifying and analyzing self-related data as it is declared in the community-wide accepted credo “Self-knowledge through numbers” (Wolf, 2008).

Figure 1 - The official logo of the Quantified Self community includes the community’s credo “Self-knowledge through numbers” (www.quantifiedself.com, retrieved November 17, 2012)

Second, people that keep records of certain aspects of their lives form a worldwide connected community under the name “Quantified Self”. This community as an entity is connected via the central website quantifiedself.com, via official and private Blogs, in social networks as in Facebook groups, but also in real-life Meetup groups2 and regional and global conferences (Kelly, Wolf, Carmichael, & Ramirez, 2007).

2

QS Meetup Groups are also sometimes called „Quantified Self Show&Tells“. They are organized and held face-to-face by QS participants to share experiences with self-tracking projects, methods or experiments and bring QS to a personal level apart from the networking possibilities in the internet (www.meetup.com, retrieved January 24, 2013).


4

2.1 The credo “Self-knowledge through numbers” / “Know thyself” Whereas governments and companies always measured success by setting up quantifiable goals and ratios, the idea of using metrics for quantifying daily life only became accepted for individuals recently. Of course, to some degree many people self-track and -measure: For a long time, many people have known their body height, average body weight, have had more or less an idea of what they eat and drink and have kept short diary entries in their notebooks.

However, although the idea of self-measurement is not new, the quality and quantity of quantifying activities have been changing in the past years. The increased perception of selftracking being a key factor to individual personal achievements and success, has helped building up a worldwide community of people that share a common belief in the power of selfquantifying. In his article “The Data-Driven Life” QS co-founder GARY WOLF explains this phenomenon with the commonly shared belief that the “numbers [which they are tracking] hold secrets that they can’t afford to ignore, including answers to questions they have not yet thought to ask”, which means that numbers do not only help to objectify certain observations in a particular context but also help to gain a profound understanding of self-related information (Wolf, 2010). According to WOLF, this is also the main reason, why self-quantifying activities are not necessarily conducted to achieve a specific goal. In contrast, self-trackers “may take up tracking with a specific question in mind”, but they probably continue because they are curious in the information content of the numbers itself (Wolf, 2010).

Partly, (new) goals may even evolve from the content of this information as a logical inference to improve this data – respectively to improve oneself. “Self-tracking breeds selfexperimenters“, (Kelly, 2011). Self-quantifiers are interested in the meaning of quantifiable information and consequently start searching for ways to manipulate them: “By intimately monitoring themselves, they increase control over their own lives, liberate themselves from unwanted habits, pursue their goals with maximum insight – and thus better their chances of success”, (Beato, 2012).

2.2 Community development Since GARY WOLF and KEVIN KELLY started the QS-Blog quantifiedself.com in 2007, various Meetup groups have been established around the globe. In the beginning of 2012, only 50 cities were counted organizing local Meetups (The Economist, 2012), one year later yet by January 2013 there were already 108 local Meetup groups installed worldwide (Figure 2).


5

Figure 2 depicts the distribution of all these Meetup groups around the globe. Apparently, QS still seems to be an occidental phenomenon that is widely spread in Northern America and Europe.

Figure 2 - Quantified Self Meetup Groups worldwide (http://quantified-self.meetup.com/, retrieved January 28, 2013).

The QS Community is mainly connected virtually but also physically in real-life. The network’s global centrum is the official QS Blog quantifiedself.com3, but there also exist official national websites, for example the German QS Blog qsdeutschland.de4. They are accepted communitywide as central hubs and provide their members with general information on QS, self-tracking tools and apps. They also reblog articles about QS from around the world or announce the next meetups within their geographical reach (Kelly, Wolf, Carmichael, & Ramirez, 2007).

The QS forum, that is located on this website, is also the central platform for all of its members to share articles, information or data and discuss the latest apps and tools5.

The local QS groups use the online platform Meetup.com, wherein self-quantifiers living within certain geographical areas may find companions to share information on QS, discuss certain 3 4 5

QuantifiedSelf.com, http://www.quantifiedself.com/about, retrieved November 17, 2012. QSDeutschland.de, http://qsdeutschland.de/, retrieved January 28, 2013. Quantifiedself.com: Guide to Self-Tracking Tools, http://quantifiedself.com/guide/, retrieved January 28, 2013.


6

fields of interest or organize local real-life meetings. Normally between one and five members (that are most of the time also the initiators of these groups) voluntarily take charge of the moderation and administration of the group website and of the organization of the local meetups.

Additionally, there exists an unknown number of individual, personal blogs that often depict personal experiences with self-tracking activities.

Apart from these online possibilities to connect, self-quantifiers regularly meet physically in local meetups or global conferences. These meetings include speeches by members and organizers as well as presentations by tool and app developers. They use these physical platforms to discuss new technologies and developments of self-tracking tools or share individual experiences and learning points.

2.3 Self-tracking activities According to SWAN (2009), the term Self-Tracking refers to the action of regular, voluntary elicitation and collection of all kind of metrics that can be related to a person. These metrics may include biological, psychological as well as physical variables. This means either healthrelated data like body weight, sleep quantity and quality, blood results, blood pressure, nutrition habits and mood, but also data that can be linked to environmental or behavioral information as for example tracing one’s location, work time or weather and finances. The evaluation of this data is often realized graphically by diagrams, graphs or (especially for qualitative data) by tagclouds (Swan, 2009, p. 509).

2.4 Quantified Self in comparison to LifeHacking and LifeLogging As the self-quantifying phenomenon is still developing, there also exist designations for some marginalized groups such as “LifeLogging”, “Sousveillance6” and “Habit Design”. The members of these groups only focus on certain aspects of the QS idea. “Self-tracking is part of an exploratory worldview in which the key goal is learning through the process of data collection and interpretation”. (Fajans, 2013). LifeLogging for example refers to the idea of generally recording data and journaling one’s life. LifeLoggers and Self-Trackers in general build immense archives of life data with the aid of technical devices: “This includes all text, all visual information, all audio, all media activity, as well as all biological data from sensors on one’s 6

http://en.wikipedia.org/wiki/Sousveillance#cite_note-1; accessed on January 15th, 2013.


7

body”, (Kelly, 2007). However, LifeLoggers do not necessarily intend to use this data as a resource for future decisions: “The first person I met doing this was Ted Nelson in the mid1980s who recorded every conversation he had, no matter where or what importance. To my knowledge his archives have never been revised, even by him”, (Kelly, 2007).

Meanwhile “LifeHacking”, “HealthHacking”, “BioHacking” and “MindHacking” refer to an extended version of the Quantified Self idea. LifeHackers consciously conduct multifarious selfexperiments, intending to creatively optimize (“hacking”) various areas of their lives. “[…] Selfexperimentation repeatedly generated ideas. It did so by producing “accidents” (unexpected observations) and making me think”, (Roberts, 2004, p. 257). They base their self-experiments on data they can collect about themselves. Thus, they try to justify their findings by the use of scientific standards. Seth Roberts e.g., Professor of Psychology at Tsinghua University in Beijing and Emeritus Professor of Psychology at the University of California at Berkeley, conducted self-experiments over a period of twelve years on several areas of his life and scientifically published his data-based results in Behavioral and Brain Sciences.

As self-experimentation has always played a big role in medical research (Weisse, 2012), neither the ideas of Self-Tracking, LifeHacking nor LifeLogging are completely new. LifeLogging may lead to self-quantification; self-tracking activities may lead to LifeHacking experiments. “A goal-oriented tracking project can shift over time into a curiosity-driven exploration, which then may shift into a new way of life”, (Fajans, 2013). Most of the individuals within these movements share a common curiosity in their individuality (Kelly, 2011) and try to “find their way” apart from the means or averages that can be found in official scientific studies (Fajans, 2013).

2.5 Market penetration of smartphones and health apps A nationwide study of the PEW RESEARCH CENTER’S INTERNET & AMERICAN LIFE PROJECT among 3,014 adults in the U.S. in 2012 proved evidence that 85% of all U.S. citizens own a cell phone and 53% of them a smartphone (Fox & Duggan, 2012, p. 3). The situation in Germany trails a little bit in the polls with only every third cell phone being a smartphone, although every German citizen owns averagely more than one cell phone (Bundesnetzagentur, 2012). Worldwide 1.038 billion smartphones had been sold by the end of Q3/2012 which makes every seventh human being in the world a smartphone user (Strategy Analytics, 2012).


8

At the same time, a new market of health-, self-tracking- and LifeLogging-apps as well as the development of affordable tracking devices are facilitating the tracking of daily life activities. In 2012, “some 19% of smartphone owners have at least one health app on their phone. Exercise, diet, and weight apps are the most popular types”, (Fox & Duggan, 2012, p. 2). By the latest PEW study from January 2013, it has been revealed that already “seven in ten U.S. adults keep track of a health indicator”, (Fox & Duggan, 2013, p. 2).

Synchronically to the smartphone ubiquity emerged this internet phenomenon and community that comes together under the terms of “Quantified Self”, “Self-Tracking” and “LifeLogging”.

Theoretical Background: Motivational Theories

9

3 Theoretical Background: Motivational Theories After an introductory presentation of Quantified Self, this chapter will help understand what motivation is. What motivates people to do certain things and what prevents them from doing others? As this bachelor thesis is primarily about the motivational reasons to conduct selftracking the following chapter depicts the general theoretical basis of motivational theories and reviews the common definitions for intrinsic and extrinsic motivation. The second part of chapter 3 reviews the basic theories of motivations to participate in online communities such as e-learning or knowledge-sharing platforms and thus especially highlights the Reader-toLeader-Framework by JENNIFER PREECE & BEN SHNEIDERMAN (2009).

3.1 Motivational Theories in general The word “motivation” originates from the Latin word ‘movere’ - to move. To be motivated to do something literally means to be moved to do it (Deci & Ryan, 2000, p. 54). Moreover, the “concept of motivation” is not only used to understand the correlation between the performance of an actual action and a biological need (such as hunger or sleep), but also human’s behavioural variability (Zimbardo & Gerrig, 2008). According to ZIMBARDO & GERRIG, a reason that leads us to decidedly perform an action one day may be the same reason we refuse to perform the same action the next day. However, this term may also refer to the different amounts of motivation that individuals need to eventually do something. “People have not only different amounts, but also different kinds of motivation”, (Deci & Ryan, 2000, p. 54).

An interest in “the reasons that give rise to perform an action” (Deci & Ryan, 2000, p. 55) did not only lead to the focus of interest of this bachelor thesis, but also to the fundamental questions in the Motivational Theories: Why are people motivated? And why do these reasons differ from person to person? As explored in the next two subchapters, the motivational theories generally bring into account two different types of motivations: intrinsic and extrinsic motivation. 3.1.1

Intrinsic motivation within the scope of self-tracking activities

Intrinsic or internal motivation refers to an action that is performed for its own sake (Heckhausen, 1989, p. 465) or “doing something because it is inherently interesting or


10

enjoyable”, (Deci & Ryan, 2000, p. 55). Intrinsically motivated actions generally follow the pursuit of a specific goal, but consist of an experience that is inherently satisfying in itself: “It [intrinsic motivation] underlies activities that are performed purely for the joy gained from the activities themselves”, (Levesque, Copeland, Pattie, & Deci, 2010, S. 618) According to CSIKSZENTMIHALYI (1999) this mental condition is characterized by the impression of a lack of time, full control over the running task and a positive self-awareness. The notation “flow” or “flow-experience” illustrates the fact, that in this state “the sense of time is distorted; hours seem to pass by in minutes. […] A person that is in flow is completely focused, there is no space in consciousness for distracting thoughts, irrelevant feelings. Self-consciousness disappears, yet one feels stronger than usual”, (Csíkszentmihályi, 1999, p. 33). Of course, intrinsic motivations may also include tasks that serve “basic [physiological] need satisfaction” (Deci & Ryan, 2000, p. 57).

However, in the context of self-tracking and especially in the context of this bachelor thesis, it has to be remarked that an elevated activity such as self-tracking normally will not originate in a desire to satisfy a basic physiological need. A self-tracker ought not to be likely to try satisfying basic physiological needs (such as breathing, eating or drinking) by tracking his daily footsteps or his TV and food consumption. On the contrary, intrinsic motivation accrues from basic “psychological needs - namely, the innate needs for competence, autonomy, and relatedness” (Deci & Ryan, 2000, p. 57). A person tracking haphazard life data – for example the amount of days per year spent abroad like NICHOLAS FELTRON7 or the colors of daily consumed food like MIMI O CHUN8 – might be likely to find satisfaction in the activity of doing so without necessarily pursuing an ultimate goal. .

In the context of this bachelor thesis the following of the presumed motivations (as seen in Table 2, p. 22) are considered to be intrinsically motivated: entertainment, LifeLogging, selfreflection, self-determination, contribution to an online community. 3.1.2

Extrinsic motivation within the scope of self-tracking activities

While intrinsically motivated actions are performed for the experience of the action itself, for the actual performance of an extrinsically motivated action, there needs to be an underlying external stimulus. This might be to “attain a goal, to obtain a reward, or to avoid a penalty or a negative consequence” (Levesque, Copeland, Pattie, & Deci, 2010, S. 619). Such an external 7 8

Official website of the self-tracker Nicholas Feltron: www.feltron.com, retrieved January 15, 2013. Official blog of Mimi O Chun: http://mimiochun.see.me/, retrieved January 15, 2013.


11

trigger may be either sowed by a different person, by an external event or by the prospect of any kind of internally felt reward or penalty. In contrast to intrinsic motivation, extrinsically motivated actions are not performed for the enjoyment of the action itself, but focuses on reaching (or averting) a desired (or unwanted) outcome. Individuals perform such activities because they see these activities as expedient instruments that serve as a means to an end (Levesque, Copeland, Pattie, & Deci, 2010, S. 619). A patient suffering from a chronic disease that has been asked by their physician to track their blood pressure and weight every day would obviously be motivated externally to pursue this task. The same patient would be said to be motivated extrinsically if they themselves saw an advantage to tracking certain vital parameters and started these activities in order to understand how their blood pressure or body weight affected certain symptoms of the disease. However, it seems unlikely the same patient would start tracking such parameters just for the enjoyment of the tracking-activity itself.

Within the scope of self-tracking activities the following presumed motivations as listed in Table 2 (on page 22) could be considered being extrinsically motivated:

LifeHacking,

overcoming self-deception, self-comparison, self-efficacy, self-gratification, self-optimization, self-awareness, self-determination, sense of belonging, sustaining relationships and the search for therapy alternatives.

3.2 Motivations to participate in online communities Motivations to engage in daily-life communities such as sport clubs, public charity or political associations differ in many ways. Most of them hark back to personal convictions and beliefs, political interests or favored kinds of leisure time activities: The differences are as multifarious as there are individual personalities. What they have in common is that humans have a need to belong and be affiliated with others (Ridings & Gefen, 2005): In a social environment individuals get help to achieve goals, they get rewarded and get a unique chance to form their own “social identity of values, attitudes and behavioral intentions” (Ridings & Gefen, 2005) with respect to a group and its purpose (Ludford, Cosley, Frankowski, & Terveen, 2004, p. 6).

Based on this, the motivation to (pro-)actively participate in online communities can be regarded from a different point of view. The underlying motivations to participate in a knowledge sharing platforms (e.g. Wikipedia®) are different to the motivations to participate


12

in social networks (e.g. Facebook©), health social networks9 (e.g. PatientsLikeMe©) or professional social networks (e.g. LinkedIn©, XING©). For example, a student that intensively shares his knowledge on e-learning platforms with a team acts from different convictions than a student who shares their knowledge by writing articles for Wikipedia or a student that writes articles for a professional blog in XING.

In the following subchapter the different levels of participations in online communities will be explained by the Reader-to-Leader-framework by JENNIFER PREECE & BEN SHNEIDERMAN (2009). Afterwards, the literature on motivations of participating in online communities will be reviewed.

The Reader-to-Leader Framework There are many types of distinctions that can be made when it comes to different kinds of virtual participation in online communities: There are people submitting video posts to YouTube and other people that write articles for Wikipedia, there are bloggers, photographers, forum administrators, etc. All of them contribute in different ways to their chosen community.

However, from another perspective one could also simply differentiate between active and inactive members or as AMY JO KIM names them: elders and lurkers (Kim A. J., 2006). The Reader-to-Leader-framework by PREECE & SHNEIDERMAN (2009) explains the different ways of participation in online communities by the different roles that members go through in an evolutionary process. The authors indicate that users are comparatively cautious in the beginning and do not necessarily interact with the online community. With the passing of time however, users may pass through a natural evolution process: from a passive reader they will become a contributor. Some will then become a collaborator with only a few finally become a leader (Preece & Shneiderman, 2009, p. 15)

Figure 3 depicts the different stages in the Reader-to-Leader Framework. In the beginning many users may look at a community’s website but only a few will decide to come back on a regular basis and thereby become a so-called reader. In the second stage, some of the readers start contributing content to the community as a contributor; later as a collaborator they start developing relationships and work together with other members. Only a few will then become a leader: A leader is an outstandingly active member that not only contributes content-wise to 9

SWAN has been the first introducing the notion “health social networks” for a special type of social networks that – besides their typical functions of nowadays online communities - deal with health-related topics. (Swan, 2009)


13

the community but also “promotes participation, mentors novices and helps setting and upholding policies”, (Preece & Shneiderman, 2009, p. 23). The arrows in Figure 3 indicate that a member does not necessarily need to pass through every stage in order, but can skip one or two stages or even go back to an earlier one. They can also take on two roles at once. A leader that administers discussions and user rights can still be a contributor that writes articles or submits blog posts on his own.

Figure 3 - The Reader-to-Leader Framework by PREECE & SHNEIDERMAN (2009). “The thickness of the green arrows and smaller shapes indicate the decreasing number of people who move from one form of participation to another. The thin grey arrows indicate how people can also move in a non-linear fashion to participate in different ways”, (Preece & Shneiderman, 2009, p. 16).

The assumed underlying motivations by PREECE & SHNEIDERMAN for the different levels of participation will now be explained in detail:

For the group of readers, the authors assume that the personal benefit from reading usergenerated content and fun are the most important motivations: “Most readers of health discussions, consumer product reviews, or Wikipedia articles presumably are looking for information that they plan to use. Sometimes people may read to gain information that benefits others, such as family members, friends, or colleagues”, (Preece & Shneiderman, 2009, p. 17). Some research refers to the group of readers as lurkers. Lurkers are not actively contributing to a network but only benefiting from community content (Bishop, 2007). They are attracted to online communities, “because of their desire for information that is credible, relevant, and easy to find”, (Tedjamulia, Olsen, Dean, & Albrecht, 2005).


14

Looking at the next higher stage, the group of contributors is – according to PREECE & SHNEIDERMAN – especially motivated by goal-oriented, self-presenting and self-rewarding aspects of contributing (Preece & Shneiderman, 2009, p. 19).

Goal-oriented contributors, who have a greater need-to-achieve, are more likely to consider their participation and contribution to be important; furthermore “they find it enjoyable to work hard, to be compared to a standard, and to be challenged”, (Tedjamulia, Olsen, Dean, & Albrecht, 2005, p. 4).

With the second motivation “self-presenting” there are several additional underlying motivations that go with it: For example LUDFORD ET AL. list a human’s “desire to stand out in a crowd [and to] feel special with respect to the group and its purpose” (Ludford, Cosley, Frankowski, & Terveen, 2004, p. 6). The rewarding factor of getting recognition for the quality and quantity of contributions is another motivator to contribute. “Reality shows, talent competitions, YouTube, blog, and microblog revelations of intimate personal details and Flickr posts of provocative pictures are all manifestations of the need to be noticed. They also fuel and make real the belief that anyone can be famous”, (Preece & Shneiderman, 2009, p. 19) According to TEDJAMULIA ET AL. on the one hand intrinsic altruistic motivations such as “community citizenship, generalized reciprocity, moral obligation and pro-social behavior” play a major role while on the other hand extrinsic non-altruistic motivations such as self-interest or self-benefit also have an impact (Tedjamulia, Olsen, Dean, & Albrecht, 2005, p. 2).

To become a collaborator, who develops relationships within a community and works together with others members, it requires “a mutual understanding, shared beliefs and assumptions”, (Preece & Shneiderman, 2009, p. 9). Additionally, altruism and a belief in helping others have been identified to explain the major motivations of collaborators. The prospect of name recognition as a form of reputation also plays a large role as a motivator for participation (Huffaker & Lai, 2007).

Finally, leaders are the most active participants in online communities. According to PREECE & SHNEIDERMAN leaders are driven by power, the respect received for helping others, the visibility within a community and rewards (Preece & Shneiderman, 2009, p. 24). In their essay “Why do people write for Wikipedia”, ANDREA FORTE & AMY BRUCKMAN emphasize that the prospect of an increased credibility especially motivates people to take on the responsibilities of a leadership role. As soon as someone is recognized as a leader, his credibility increases. By implication,


15

along with a rise of one’s position, one’s credibility increases as well. “In fact, what seemed to drive scientists and motivate them was a sense of credibility that allowed them to assume more and more central roles in the scientific community. In its fullest sense, “credit” is not something that is given or received by individuals in the community, but a measure of power and efficacy”, (Forte & Bruckman, 2005, p. 7).

3.3 Presumed motivations for self-tracking In his article “The Data-Driven Life”, one of the QS founders GARY WOLF classifies three important types of self-tracking motivations. The first type of self-quantifier is motivated by the optimization possibilities of something in his life, for example his training performance or weight loss.

“We use numbers when we want to tune up a car, analyze a chemical reaction, and predict the outcome of an election. We use numbers to optimize an assembly line. Why not use numbers on ourselves?” GARY WOLF, 2010

Self-quantification helps self-trackers keep track of their progress and is esteemed to be more accurate and reliable than just journaling (or memory): For example one of the participants of the survey quoted in Question 11 (view Appendix I) “I like to see how active I was at a certain time. Having a calendar where I write down with whom I caught up and what we did together also supports my memory. I like the fact being organized and to be able to revitalize my memories”, Participant ID #62.

According to an article in THE ECONOMIST (2012), the QS community exists of “[…] early adopters, fitness freaks, technology evangelists, personal-development junkies, hackers and patients suffering from a wide variety of health problems”, (The Economist, 2012). The article confirms WOLF’s assumption, that self-tracker primarily share a persuasion that the collection and analysis of their data of daily life activities helps them to improve their lives, which can also be referred to as self-optimization. Self-tracking also helps increase control over one’s life, pursuing goals with maximum insight and improves one’s chances of success (Beato, 2012).

The second type of self-quantifying originates from curiosity: “For many self-trackers, the goal is unknown. Although they may take up tracking with a specific question in mind, they


16

continue because they believe their numbers hold secrets that they can’t afford to ignore, including answers to questions they have not yet thought to ask”, (Wolf, 2010). WOLF does not refer to curiosity in general humankind or curiosity for other human beings’ condition but curiosity in one’s own life. Self-trackers are curious in how certain things in their life correlate, as for example SETH ROBERTS who searched for connections between his sleep, mood, health and weight and realized that for example 8h of standing per day (instead of sitting) made him sleep less but more restoratively (Roberts, 2004, p. 1). Self-quantification thus serves as a means to an end to overcome improvising, guessing, forgetfulness and unconscious alterations of underlying conditions (Wolf, 2010).

As a third motivation, WOLF refers to the restrictions of human beings to really understand who they are. This lack of self-knowledge can be overcome with the help of machines, “Behind the allure of the quantified self is a guess that many of our problems come from simply lacking the instruments to understand who we are”, (Wolf, 2010). More and more human beings want to understand in which ways they are different from others. This need for self-individualization in today’s standardized meritocracy thus increases the need for an enhanced self-knowledge (Langwieser & Kirig, 2010). This enhanced self-knowledge can be obtained for example with the help of sensors and trackers. In WOLF’s eyes, health self-trackers often try to find their own way of improving symptoms away from the seductions of pharmaceutical marketing and the errors of common opinions. “People are not assembly lines. We cannot be tuned to a known standard, because a universal standard for human experience does not exist”, (Wolf, 2010). In the same way, fitness trackers try to adapt their trainings to their individual conditions and goals while trying to understand their strengths and weaknesses.

Self-optimization, curiosity and self-knowledge: Besides the motivations explained by THE ECONOMIST and WOLF, in the survey for this bachelor thesis, there have been more motivations taken into consideration and tested. Additionally, a combination of all preliminary findings from the theoretical research on psychological motivational theories influenced the final questionnaire on self-tracking motivations. As depicted in the previous chapters 3.1 and 3.2, fun, self-presentation, self-benefit, power, community citizenship, self-rewarding, sense-ofbelonging, altruism, goal-orientation, self-efficacy, etc. and a few more may be motivational reasons to participate in online communities. But which of these presumed motivations for participating in communities in general apply especially to the motivational reasons for participating in the self-tracking movement?


17

The survey conducted to answer these questions – the conceptualization process as well as all findings from the survey – will be now presented in the following chapter.

Survey: Methodology and Evaluation

18

4 Survey: Methodology and Evaluation This chapter deals with the concrete process of an exploratory study on the motivational reasons and general profiles of self-trackers and the findings from the survey. The whole survey process took place from September 2012 till January 2013. At first, the methodical approach and the survey conceptualization and design will be explained in detail. The second part of this chapter contains all comprehensive evaluations and all findings from the survey.

4.1 Methodological approach and survey design This section gives an outline of the development process of this study. The research for this bachelor thesis primarily consisted of five main phases: 

Theoretical preparation and outlining of research questions



Expert interviews at the 4th QS Meetup in Berlin in September 2012



Conceptualization and pre-testing (qualitative & quantitative) of the questionnaire



Online implementation and distribution of the survey (November 1, till December 1, 2012)

 4.1.1

Evaluation and representation of the data Theoretical approach

In the first phase the psychological fields of motivational theories and more specifically the motivations for participating in online communities have been skimmed. This first phase provided a foundation of background information for the first hypothetical applications of motivations of self-tracking and participating in the QS community. Also the insights from an earlier written seminar paper about “The Meaning of the Quantified Self Movement in a health-oriented context”, written in summer 2012, have been used to build up questions concerning self-tracking activities such as objects of tracking or scope of activity (Nißen, 2012).

The theoretical presumptions framed the first layout of the later questionnaire (Appendix I) and supported the preparation of the upcoming expert interviews.


4.1.2

19

Expert interviews

As I had the chance to visit the 4th QS Meetup of one of the Berlin Meetup groups Quantified Self Berlin10 on September 8, 2012, the goal of the expert interviews has been to compare the first presumptions from the theoretical research approach (see chapter 4.1.1) with the lived experience of six self-quantifiers and to discuss their individual self-tracking motivations. Altogether the attendance of the 4th Quantified Self Meetup in Berlin did not only ought to broaden the general understanding of self-tracking motivations, but additionally helped building up a network of self-trackers and QS members for later support (as for example the online distribution of the survey via QS blogs, forums and Meetup groups).

The expert interviews took place in an informal atmosphere at the Meetup location right after the official program of presentations and keynote speeches and lasted four hours in total. The six interview partners have been chosen almost haphazardly, three of them had been holding stage presentations on their past self-tracking experiences during this event. By directly linking to some details from the presentations as a conversation starter, the individual interviews did not stringently follow a formal questionnaire that needed to be answered right through. With some questions in mind from the theoretical research, the interview partners have been engaged in discussions sharing their opinions on how the preliminary findings from the theoretical research might actually apply to the specific motivations of self-trackers. 4.1.3

Questionnaire conceptualization

In the third phase, the gathered knowledge from the theoretical research and the expert interviews led to a comprehensive questionnaire. In order to give a general outline of the survey concept, the questionnaire will now be explained step by step.

The final survey consists of 25 questions that are grouped in nine questions sections based on the categories as outlined in Table 1.

10

http://www.meetup.com/Quantified-Self-Berlin/, retrieved December 29, 2012.


Underlying question 1. 2.

Introduction Objects of Tracking

“What?”

3.

Activity level

“How much?”

4.

Motivations

“Why?”

5.

Hurdles

6.

Technology used

7.

Personalities

“Who?”

8.

Demographics

“Who?”

9.

Other

“Why not (more)?” “How?”

…

20

Explanation Is the respondent an actual self-tracker or not? Are the participants either tracking well-being or health-related parameters or both or not directly self-related? Which parameters are they tracking concretely? Do they maybe suffer from a chronic disease? How much time do self-trackers spend on selftracking activities? For how long are they already self-tracking? What are the underlying and psychological motivations to include self-tracking activities in their daily lives? What could be possible hurdles for self-trackers to keep on self-tracking? How to self-trackers measure and record their life? Do they use specific tools and spend a lot of money on it or are they trying to keep it as simple as possible? Implies being a self-tracker several specific characteristics of the personality’s dimensions? Standard questions on the demographic background of the participants: age, origin, income, occupation. Did you answer honestly? Are you paying attention? Do you want to participate in the lottery drawing for one out of three Amazon vouchers?

Table 1 - First structure of question categories for the later questionnaire in preparation for the expert interviews

Besides the questions on motivations and hurdles for self-tracking (the “Why”), the questionnaire also ought to shed light on the general composition of self-tracking types and characters (“Who” and “What”). Therefore, the following topics as summed up in Table 1 had been identified: a general classification of the tracking goals (such as a differentiation between wellbeing- or health-related objects), the amount of self-tracking activities and the engaged technologies as well as demographic factors of the participating self-trackers. Additionally, an abbreviated 10-item-short version (BFI-10) of the Big Five Personalities Tests (NEO-PI-R) by Rammstedt & John has been added to identify typical self-tracking personalities (Rammstedt & John, 2007).

In the following, the development process of the question sections Objects of Tracking and Motivations will be explained in detail. Objects of Tracking In order to evaluate the type of tracking activities and to question the objects and goals of tracking of the respondents of this study, several insights from a prior seminar paper “The


21

meaning of the Quantified Self movement in a health related context” (Nißen, 2012) have influenced the final set of questions in this section. This work generally distinguishes between tracking ambitions that are either health-related or not-health-related. Furthermore, the health-related ambitions can be divided into medically motivated (“Health”) and selfexperimental motivated (“Healthy Living”) activities (Nißen, 2012, p. 14). For the use in this study, several adaptions have been made based on the expert interviews at the 4 th QS Meetup in Berlin (see chapter 4.1.2) and further research within the QS community. Figure 4 depicts the final structure of questions on tracked parameters of this research section.

Figure 4 - Classification of the Objects of Tracking based on the categorization of self-tracking possibilities by Nißen (2012, p. 14).

Motivations The gathered knowledge from the theoretical research (chapter 3) and the expert interviews led to a first comprehensive set of presumed self-tracking motivations: with reserve and prudently sorted into extrinsic and intrinsic motivations as follows in Table 2 (order and applicability are under consideration and part of the research of this study).


Intrinsic motivation

Extrinsic motivation

Fun & entertainment

Overcoming self-

22

? Contribution to QS

deception Interest

Self-optimization

Therapy alternative

LifeHacking

Self-efficacy

Self-presentation

LifeLogging

Self-control

Need for affiliation

Sense of belonging

Self-determination

Self-reflection

Self-responsibility

Flow

Self-gratification Self-regulation

Table 2 - Presumed underlying motivations of self-tracking (sorted into extrinsic and intrinsic motivations). Order and applicability are under consideration and part of the research of this bachelor thesis

As this bachelor thesis on self-tracking motivations enters a new field of research, there have been almost no possibilities of using approved standard questions. To understand the general idea of formulating questions to define certain topics, several online and psychological questionnaires have been reviewed. The finally implemented questions can be found in the Appendix I.

To ensure the general validity of the question set, all of the questions (Appendix I) and especially the motivations section have been discussed in several iterations at the Institute of Information Systems and Marketing (IISM) at the Karlsruhe Institute of Technology (KIT) on a qualitative basis. They have also been reviewed and feed-backed by members from the QS community. In the final questionnaire, each item of the motivation questions has been scored on a 1 to 5 Likert scale, with scale scores ranging from “Strongly Disagree”, “Disagree”, “Neutral”, “Agree” to “Strongly Agree” (view Question 13).

By the quantitative analysis of a pre-test of the survey with 20 self-trackers among the Quantified Self community it had been ensured that the questions would be easily understandable and would not lead to ambiguous results by implying two diametrically opposed questions on the same presumed motivation.

Also, to identify typical self-tracking personalities, an abbreviated 10-item short version (BFI10) of the Big Five Personalities Tests (NEO-PI-R) by RAMMSTEDT & JOHN (2007) has been deployed.


23

A paper-based final version of the questionnaire can be found in the Appendix I. 4.1.4

Online implementation and distribution of the survey

This chapter outlines the process of the implementation and distribution of the survey online. With the intention to get a representative sample of self-trackers worldwide, the survey has been implemented and provided online via the free and open online survey tool LimeSurvey®. The introductory question “Are you keeping records of or tracking anything that occurs in your life?” has been included to ensure recognizing the actual self-trackers of all survey participants. Participants that would answer this question with “No” only would be asked to answer the last part of the questionnaire starting Question 16 in the “Hurdles”-section “If you are not tracking something at all or if you would like to track more, why don’t you do so?”

The survey has been tested for consistency and functionality in several iterations by colleagues and friends in the week before it has been released on November 4, 2012.

Distribution of the Survey and Participant Recruitment The official online phase started on November 4, 2012, and had been officially announced this day

via

the

German

QS

Facebook

Group

“Quantified

Self

Deutschland”,

(https://www.facebook.com/groups/qs.germany/permalink/504703469554890/,

retrieved

December 17, 2012). To motivate people to participate, the possibility to be putted into a draw of one out of three Amazon vouchers of 50$ each has been advertised. Marketing efforts into the distribution of the survey have been putted in till the end of November via several channels: 

Meetup.com groups: 43 Meetup groups worldwide with a total of 9469 members (http://quantified-self.meetup.com/, accessed and calculated November 21, 2012) have been approached to ensure getting a representative set of self-trackers of the QS community. Usually the admins of the groups have been approached and been asked to spread the word about the survey.



Facebook (FB) groups: QS FB groups as well as health related community groups (such as the FB groups of PatientsLikeMe11, Wellsphere12, the Patients Voice13, etc.) with an average community size of 8000 members (measured by the amount of FB “Likes”)

11 12 13

https://www.facebook.com/PatientsLikeMe?fref=ts, retrieved December 29, 2012. https://www.facebook.com/pages/Wellsphere/10916387499?ref=ts&fref=ts, retrieved December 29, 2012. https://www.facebook.com/thepatientsvoice?fref=ts, retrieved December 29, 2012.


24

have been approached via the personal messaging function of FB and with posts on their respective FB “walls”. 

Twitter: Via the news messaging platform Twitter several Twitter hashtags (#) as well as several user accounts (@) related to QS have been used to spread the word about the survey, e.g. @Quantified Self14, #Quantified Self Forum15, @QS_Berlin16, #big data17, #health-data18 etc.



Blogs: Official as well as private QS blogger have been approached and asked to spread the word about this survey in a post or via a blogroll link in their respective sidebars. Especially the official German QS blog helped distributing the survey by adverting the survey link in an individual post (http://qsdeutschland.de/teilnehmer-fur-studie-zuquantified-self-gesucht/, retrieved December 29, 2012) as well as several times indirectly within announcement posts.



Individual contacts: As I had been attending the 4th QS Meetup on September 8, 2012 of on the Berlin QS Meetup groups, Quantified Self Berlin19, I used the chance to share the survey individually via e-mail with several there-met self-trackers.



Attendance of the 1st Meetup in Berlin: On November 22, 2012, at the 1st Meetup of another QS Meetup group in Berlin “The Berlin Quantified Self Meetup Group20” with more than 85 signed in attendants, I had the chance to give a short presentation on the actual status of the survey and to call for participation among all present attendants.

4.1.5

Data evaluation and representation

To evaluate and represent the data, especially the spreadsheet application Microsoft® Excel 2010 as well as some specific functions of the statistical analysis package IBM® SPSS Statistics 21 came to use. For all currency based statistics within this bachelor thesis, all data has been translated to US-Dollar on the base of the official conversion rate of October 17, 2012: 1€ = 1.3132$.

14

https://twitter.com/quantifiedself, retrieved November 15, 2012. https://twitter.com/search?q=%23QuantifiedSelf&src=hash, retrieved November 15, 2012. https://twitter.com/QS_Berlin, retrieved November 15, 2012. 17 https://twitter.com/search?q=%23bigdata&src=hash, retrieved November 15, 2012. 18 https://twitter.com/search?q=%23healthdata&src=hash, retrieved November 15, 2012. 19 http://www.meetup.com/Quantified-Self-Berlin/, retrieved November 15, 2012. 20 http://www.meetup.com/Quantified-Self-Berlin-International/, retrieved November 15, 2012. 15 16


25

4.2 Evaluations This chapter will present all findings of the final survey and thereby concentrate on general demographic factors of all self-trackers, objects of tracking, scope of activity, deployed technologies, motivations and hurdles.

At first, the data retrieval and selection process after the official end of the survey on December 1, 2012 will be explained in detail. In the second part all findings of the survey will be presented by dimension: 

General demographic factors



Object of tracking and introduction of two reference groups: WELL-BEING ONLY and WELLBEING AND HEALTH-RELATED self-trackers



Scope of activity



Deployed technology



Underlying motivations of self-tracking



Hurdles

The data and findings from the BFI-10 (the short version of the Big Five Personalities Tests (NEO-PI-R) by Rammstedt & John, 2007) will not be evaluated in this paper.

4.2.1

Data retrieval

By the official end of the survey on December 1, 2012, a total of 411 respondents had seen the survey. As reconstructed in Figure 5, thereof a total of 217 participants has fully completed the questionnaire and thereof a total of 167 could be identified as actual self-trackers by answering the introductory question “Are you keeping records of or tracking anything that occurs in your life?” with “Yes” (Appendix I, Question 2).


26

Figure 5 - Composition and derivation of the finally used total of responses after extracting all 411 responses from LimeSurvey® on December 1, 2012.

Additionally, for the final analysis 15 people that failed the “Paying Attention”-Test (Appendix I, Question 13) as well as two participants that answered the question “Did you answer honestly throughout the questionnaire?” with “No” (Appendix I, Question 38), have been excluded from the finally analyzed responses. Eventually, a total of 150 responses could be evaluated within the scope of this research.

Remark: Six people that answered Question 16 with “I am not tracking anything” have been included in almost all analysis as a manual analysis of their given data showed that they are self-trackers in general but not involved in own self-tracking experiments at the moment. 4.2.2

General demographic factors

This chapter will provide some general statistics on the personal information of the respondents and evaluate all compiled demographic factors such as gender and age distribution, status of employment, origin and monthly income.

Age Distribution The age distribution of all respondents ranges from 14 to 76 years as depicted in Figure 6. Classifying all respondents in age groups with a range of 10 years as in Table 3, the biggest age group can be found within the 20 to 29 years old respondents with a total of 42% of all participants.


27

Number of Subjects

Age Distribution 14 12 10 8 6 4 2 0 14 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44 46 48 50 52 54 56 58 60 62 64 66 68 70 72 74 76 Age

Figure 6 - Age distribution of all 150 respondents

The average age is approximately 34 and the 50%-median is at the age of 30. In total, 75% of the respondents are younger than 39 years. The standard deviation amounts to 11.90 years.

Age Groups = 80 Total

Frequency 6 63 43 16 14 7 1 0 150

Percent 4% 42% 29% 11% 9% 5% 1% 0% 100%

Table 3 - Age distribution in age groups

The astonishing amount of respondents older than 50 years may be due to the fact that several health social networks21 have been approached in order to recruit some health related tracking respondents.

Gender Distribution Regarding the gender, 58.00% of all participants claimed being male, a total of 37.33% being female. Seven participants did choose answering “No answer”.

21

Explained in chapter 3.2


28

Gender Distribution 4,67% Male 37,33%

58,00%

Female No answer

Figure 7 - Gender distribution of all 150 subjects

Status of Employment To get an idea of the professional background of the participants, a set of answer options had been prepared including an open space field for different answers. As it can be trailed in Figure 8, almost a third of the participants are students and approximately 40% are employees. A remarkable amount of almost 20% of the respondents indicated being self-employees. Among the nine participants that answered “Other” (6.00%), four of them indicated being disabled and two of them being currently either pregnant or on maternal leave.

Status of Employment 2,00%

0,67% 0,67%

6,00%

Employed Student 17,33%

40,67%

Self-employed Other Unemployed

32,67%

Housewife/-husband Retired

Figure 8 - Status of employment of all 150 respondents


29

Origin Figure 9 depicts the average geographic origin of the participating respondents. 52.67% of all participants are of European origin, whereby thereof 68.35% are German. The remainder of all European participants spread throughout participants from France, Belgium, The Netherlands, UK, Italy, Spain, Ireland, Greece and Hungary – each represented within ranges from 0.66 to 3.33%.

All of the 39.33% respondents that are originally from North-America are either inhabitants of the United States (83.06% of all North American respondents) or of Canada (16.97%). The remaining twelve respondents are from Brazil, Australia, New Zealand, Pakistan, Singapore and Taiwan. A fully itemization can be found in Table 4.

Origin 8,00%

Europe 39,33%

52,67%

North-America Other

Figure 9 - Average origin of all 150 respondents

This study might not represent the average geographic distribution of self-trackers in the world. Due to the fact, that the QS community counts the biggest Meetup groups in North America and Europe (also compare Figure 2, p. 5) with partly more than 500 registered members, during the Distribution of the Survey and Participant Recruitment (chapter 4.1.4) more people from North America and Europe might have had the chance to get to know of the survey.


Origin

Country

Europe

Belgium France Germany Greece Hungary Ireland Italy Netherlands Slovakia Spain Switzerland United Kingdom Canada United States Australia Brazil Israel New Zealand Pakistan Singapore Taiwan Total

North-America Other

30

Frequency 5 3 54 1 1 1 4 3 1 1 2 3 10 49 2 1 2 1 1 4 1 150

Percent

Total Frequency

3,33% 2,00% 36,00% 0,67% 0,67% 0,67% 2,67% 2,00% 0,67% 0,67% 1,33% 2,00% 6,67% 32,67% 1,33% 0,67% 1,33% 0,67% 0,67% 2,67% 0,67% 100,00%

79

59 12

150

Table 4 - Total itemization of the origin of all respondents

Monthly disposable household income To get an idea of the financial situation of the average self-tracker, participants of the survey had the possibility to answer a question on their monthly disposable household income (Question 23). The participants could choose between their preferred currency, either USDollar ($) or EURO (€).22 Additionally, default income classes have been provided based on an income classification by DESTATIS (Statistisches Bundesamt, 2008, S. 4). 11.33% choose not to answer at all. 46.00% of the respondents are in monthly funds of less than 2,000$. The 50%Median can be found between 1,700$ and 2,000$, the average monthly income amounts to 2,697.33$.

22

For all currency based statistics in this bachelor thesis, currency data (such as the monthly income or the money spent on self-tracking related technology as in Question 15 of the survey) has been translated to US-Dollar on the base of the official conversion rate of October 17, 2012, that has been measured with 1€ = 1.3132$ (see chapter 4.1.5).


31

Monthly disposable household income 30

Number of respondents

25 20 15 10 5 0

Monthly disposable household income [US-$]

Figure 10 - Histogram of the distribution of the disposable monthly household income of all 150 respondents. The default income classes have been provided based on an income classification by Destatis (Statistisches Bundesamt, 2008, S. 4).

4.2.3

Object of tracking

The objects of tracking within the self-trackers differ in many ways. Some self-quantifier are only tracking their daily steps or the running distances, others are recording their body weight or BMI23 development on a regular base and try to put everything into correlation with environmental influencing factors such as weather or GPS-tracked location. Still others suffer from chronic diseases and therefore keep records of their medication or occurring symptoms in different situations. As outlined in chapter 4.1.3 several dimensions of possible tracking objects came to interest of research and will now be presented question-by-question. Especially the medical background and the self-assessment of the individual self-tracker how they estimate their tracked parameters will be evaluated as well as the actually tracked parameters.

23

BMI = Body Mass Index = body weight/(body height)² = [kg/m²]


32

Reasons for starting self-tracking A given set of multiple eligible answer possibilities has been presented to get to know the reasons of starting self-tracking. As 84 of all 150 respondents (56.00%) answered “I just thought I should”, self-encouragement seems to be an important factor of starting selftracking. 28% stated that they have heart about QS before. Only 7.33% are influenced by friends that started doing so as well and only 8.00% have been asked by their physician to start tracking some of their vital parameters or symptoms. Comparing respondents that have started self-tracking before 201024 and after 2010 (Table 5) clearly indicates that respondents that started tracking before QS became popular in the media are less influenced by friends or news but are more likely to have being recommended conducting self-tracking by their physician (

-Test; df = 4; p = 0.042). With a p-value of 0.042, which is smaller than the

significance level o α = 0.05, the null hypothesis that all rows are and columns are independent can be rejected.

Started tracking Started tracking Why did you start self-tracking? after 2010 before 2010 Answer Frequency Percent Frequency Percent I just thought I should. 46 30.67% 38 25.33% I heard about self-tracking (tools, apps, …) in the news. 31 20.67% 11 7.33% Friends of mine started doing so as well. 11 7.33% 0 0.00% My physician recommended tracking some of my vital parameters / symptoms. 4 2.67% 8 5.33% Other 26 17.33% 18 12.00% Total 118 75 Table 5 - Comparison of reasons of starting self-tracking for people that have started self-tracking before and after 2010 (multiple answers possible)

On the other hand since 2010 there have been more people that have heard about selftracking in the news (20.67% after 2010 compared to 7.33% of all respondents before 2010).

Medical records and background of self-tracking The respondents have been directly asked whether they suffer from a chronic disease or not. If they were suffering from a chronic disease, they either could choose to indicate explicitly from which chronic disease they suffer or to not disclose the chronic disease. The respondents also had the chance not to answer this question at all by indicating “No answer”. 24

As in 2010 the first journalistic reviews of the self-measurement movement in the US can be found in the internet. For this bachelor thesis we presumed that in this time a media hype aroused around the QS movement and more people got to know of QS and self-quantifying possibilities.


33

Answer Frequency Percent Yes, I suffer from … 36 24.00% Yes, but I don’t want to disclose the disease. 13 8.67% No, I do not suffer from a chronic disease. 100 66.67% No answer. 1 0.67% Total 150 100.00% Table 6 - Amount of people suffering from a chronic disease or not suffering from a chronic disease

As it can be seen in Table 7, two thirds (66.67%) of all self-trackers do not suffer from a chronic disease at all whereas almost the rest of the respondents (except one respondent choosing “No answer”) suffers from a chronic disease. From the second group, 36 respondents (73.46% of all chronically ill persons) added their diseases verbally. These entries have been counted manually (see Table 7); the fully, unsorted itemization of all free text input fields can be found in Appendix II.

Yes, I suffer from … Frequency Rheumatoid Arthritis 7 Diabetes 6 Thyroid Disorders 4 Morbus Crohn 2 Multiple Sklerose 3 Fibromyalgia 2 Asthma 3 Insomnia 2 Other 23 Total 52 Table 7 - Manually counted answers from free input field in Question 4 “Do you suffer from a chronic disease?”


34

Self-assessment of tracked parameters In order to understand the self-perception of all self-tracking respondents, in Question 5, the respondents have been asked how their tracking parameters can be described the best. To categorize the type of self-tracking three categories have been preset. All respondents had been obliged to decide between either “Yes”, “Uncertain” or “No” for each of the three following categories:

(1) health-related and directly linked to a chronic disease they suffer from (later referred to as “health-related”) (2) related to themselves, their body, their general health and “well-being” (later referred to as “well-being”) (3) “related to environmental or daily-life activities but not fitting into one of the two first mentioned categories” (later referred to as “other”)

Table 8 illustrates the distribution of all answers of Question 5 in which respondents had been asked to self-assess their own tracking activating in all three categories mentioned above: health-related, well-being or other.

Self-assessment of tracked parameters health-related Frequency Percent well-being Frequency Percent other Frequency Percent

Yes 49 32.67% 128 85.33% 79 52.67%

Uncertain 9 6.00% 8 5.33% 19 12.67%

No 92 61.33% 14 9.33% 52 34.67%

Total 150 150 150

Table 8 - Self-assessment of tracked parameters by all 150 respondents

83.33% of all respondents track parameters that they consider to be related to well-being related, 32.67% track health-related parameters. The relatively high amount of people that indicated tracking other parameter (52.67%) might be due to the fact, that this field could be comprehend widely.

Intersection and combination of the categories health-related, well-being and other Due to the question mode, respondents had the possibility to apply tracking multiple kinds of parameters, for example by answering both (1) health-related and (2) well-being with “Yes”. A


35

comprehensive overview of all combinations and intersections of “Yes”-answers is depicted in the Venn-Diagram in Figure 12.

An amount of 47 respondents only tracks well-being-related parameters; the groups that are exclusively tracking health-related or other parameter are with only three and twelve respondents relatively slow and will not be considered any further (see Figure 11)

Introduction of two reference groups: WELL-BEING ONLY and WELL-BEING AND HEALTH Based on the answers in Question 5, two groups have been formed for further comparative evaluations. Figure 11 depicts in light green the first group that contains all 84 respondents that track well-being but not health-related related parameters (group “WELL-BEING ONLY”), the second group in dark green contains all 44 respondents that track well-being as well as healthrelated parameters (group “WELL-BEING AND HEALTH”).

Figure 11 - Venn-Diagram for people that only track well-being related parameters (group WELL-BEING ONLY, here: light green) and people that track well-being as well as health-related parameters (WELL-BEING AND HEALTH, here: dark green)

How and in which fields might respondents that are tracking health-related data differ from people that are tracking well-being related data? These two groups will be compared several times in spots. With an average age of 38.95 years, respondents of the WELL-BEING AND HEALTH group are significantly older than those of the WELL-BEING ONLY group with 32.65 years (T-Test, p = 0.003).


Age Min Max Mean Median Standard Deviation

36

WELL-BEING WELL-BEING AND ONLY HEALTH 18 14 62 76 32.65 38.95 30 34.5 10.10 13.66

Table 9 - Statistical measures on the age distribution of WELL-BEING ONLY and WELL-BEING AND HEALTH (T-Test; p = 0.003).

They also started self-tracking out of different reasons (see Table 10). Significantly more WELLBEING AND

HEALTH self-trackers started self-tracking because their physician recommend doing

so (18.18%) compared to the WELL-BEING ONLY self-trackers (2.38%). Vice versa, 59.52% of all WELL-BEING ONLY just thought themselves they should start compared to 45.45% of all WELLBEING AND HEALTH

self-trackers. Both of these answers might indicate that WELL-BEING ONLY self-

trackers are more intrinsically motivated than WELL-BEING AND HEALTH self-trackers (

-Test; df

= 4; p = 0.005).

Why did you start self-tracking?

WELL-BEING ONLY

WELL-BEING AND HEALTH

Answer

Freq. Percent

Freq. Percent

I just thought I should. I heard about self-tracking (tools, apps, etc.) in the news. Friends of mine started doing so as well. My physician recommended tracking some of my vital parameters / symptoms.

50

59.52%

20

45.45%

27

32.14%

13

29.55%

6

7.14%

5

11.36%

2

2.38%

8

18.18%

Other

24

28.57%

13

29.55%

Total

109

59

Table 10 - Comparison of triggers to start self-tracking of WELL-BEING ONLY and WELL-BEING AND HEALTH trackers, multiple answers possible ( -Test; df = 4; p = 0.005).

Coherence of health data tracking people and people that are suffering from chronic diseases From the 49 respondents that indicated that they are tracking health-related parameters, only 37 (= 75.51%) also answered that they actually suffer from a chronic disease in Question 4 “Do you suffer from a chronic disease?”. This coherence is depicted in the Venn-Diagram in Figure 12.


37

Figure 12 - Coherence of people that indicated that they suffer from a chronic disease (Question 4) and respondents that indicated tracking health-related parameters (Question 5)

Furthermore, 12% of all self-trackers that indicated that they are not suffering from a chronic disease, stated that they nonetheless track health-related data.

The study “Tracking for health” by the PEW RESEARCH CENTER’S INTERNET & AMERICAN LIFE PROJECT by SUSANNE FOX & MAEVE DUGGAN undergird these findings: In a telephone survey with 3,014 participants they find out that 40% of U.S. adults reporting one chronic condition are selftrackers and 62% of all U.S. adults that are reporting two or more chronic conditions. They also found out, that 19% of U.S. adults that are not reporting a chronic condition nonetheless say that they track health data or symptoms (Fox & Duggan, 2013, p. 2).

The findings of their study and the herein presented survey are not comparable unconditionally, as the study setting and demographic factors of the participants differs remarkably. The study by PEW RESEARCH CENTER’S INTERNET & AMERICAN LIFE PROJECT only has been conducted on U.S. adults whereas the survey for this study has been responded partly by Northern Americans and European citizens.

As a result, it neither can be said that a health-related self-tracker suffers from a chronic disease, nor could one say that a person suffering from a chronic disease necessarily also conducts self-tracking activities that would be linked to this disease.


38

Quantity and kind of tracked parameters As a last question in the “Object of Tracking” section, the participants chose as many parameters as they wanted out of a list of 54 default parameters that have been defined and classified in advance. This classification developed as explained in chapter 4.1.3.

As depicted in Table 11, respondents of the group WELL-BEING ONLY are significantly tracking less parameters (8.11 items per person) than those of the WELL-BEING AND HEALTH group (12.95 items per person) (T-Test; p < 0.001).

Statistical measure

All respondents

Min Max Mean Median Standard Deviation

1 39 8.94 8 6.84

WELLBEING ONLY

1 24 8.11 7 5.79

WELL-BEING AND HEALTH 2 39 12.95 11 7.36

Table 11 - Statistical measures on the amount of tracked parameters per participant for all reference groups (T-Test; p < 0.001).

A detailed frequency scale can be found in the Appendix III, Table 48. Regarding all chosen single parameters, all observations will now be explained per category. All answers of the free text input fields “Other” for each category can be found in the Appendix II.

Physical Activities The parameter tracked the most often is “Exercises” (51.33%), followed by Running (40.00%) and Steps (35.33%). The last one mentioned (steps) is interesting in so far as to track this it needs special equipment, for example pedometers such as of the Fitbit®.

Tracked physical activity parameter Exercises Running Steps Floors climbed Biking Other

All respondents Freq. 77 60 53 29 27 25

WELL-BEING ONLY

Percent Freq. 51.33% 46 40.00% 41 35.33% 33 19.33% 20 18.00% 17 16.67% 13

Percent 54.76% 48.81% 39.29% 23.81% 20.24% 15.48%

WELL-BEING AND HEALTH Freq. Percent 28 63.64% 14 31.82% 19 43.18% 9 20.45% 8 18.18% 10 22.73%

X²-Test

Table 12 - Tracked physical activities’ parameter for all reference groups

p 0.334 0.065 0.670 0.667 0.780 0.310


39

Comparing the two groups WELL-BEING ONLY and WELL-BEING AND HEALTH in Table 12, more than half of the respondents of the WELL-BEING ONLY group (54.76%) track their exercises compared to 63.64% of the WELL-BEING AND HEALTH group. For this parameter as well as for counted steps, the WELL-BEING AND HEALTH respondents are more active than WELL-BEING ONLY respondents. For the activities running and floors climbed the WELL-BEING ONLY group is tracking more parameters than the WELL-BEING AND HEALTH reference group. All parameters are comparatively depicted in Figure 13.

Percent of respondents [%]

Physical activities 80%

Well-being only

60%

Well-being and healthrelated

40% 20% 0% Exercises

Running

Steps

Floors climbed

Biking

Other

Tracked parameters

Figure 13 - Comparison of tracked parameters by WELL-BEING ONLY and WELL-BEING AND HEALTH self-trackers

The high amount of people that inserted own answers in the “Other”-section might be caused by the question mode and by clustering prelisted parameters in several categories without having explained all categories in before. Free answers such as medication, blood work, heart rate, sleeping behavior, distance and mediation could have been ticked off in another category as well.

Body The three most important tracked body parameters are body weight (60.67%), heart rate (34.00%) and the BMI (24.00%). For body weight, blood pressure and stamina there can also be found a significant difference for the two reference groups with 61.90% of all WELL-BEING ONLY respondents compared to 79.55% of all WELL-BEING AND HEALTH respondents for the body weight tracking, 33.33% compared to 47.73% for BMI tracking and 7.14% compared to 20.45% for the tracking of stamina (

-Test; df = 1; p < 0.05). Table 13 lists all prelisted parameters

descending sorted by the first frequency column.


Tracked body parameter Body weight Heart rate BMI Body fat Blood pressure Body measurements Muscle Strength Stamina Other Body water

40

All respondents

WELL-BEING ONLY

Freq. 91 51 36 35 30 28 23 18 13 8

Freq. 52 28 25 22 13 14 13 6 5 5

Percent 60.67% 34.00% 24.00% 23.33% 20.00% 18.67% 15.33% 12.00% 8.67% 5.33%

WELL-BEING AND HEALTH Freq. Percent 35 79.55% 21 47.73% 11 25.00% 13 29.55% 17 38.64% 13 29.55% 10 22.73% 9 20.45% 8 18.18% 3 6.82%

Percent 61.90% 33.33% 29.76% 26.19% 15.48% 16.67% 15.48% 7.14% 5.95% 5.95%

X²-Test p 0.042 0.112 0.569 0.686 0.003 0.090 0.310 0.026 0.030 0.848

Table 13 - Tracked body parameters for all reference groups

Among the answers in the “Other”-section three notes have been found, that might had been missing in the prelisting and should be later included in the comprehensive outline of tracking


objects: skin structure, menstruation and pregnancy related factors (Appendix II).

Body 90% 80% 70% 60% 50% 40% 30% 20% 10% 0%

Well-being only Well-being and healthrelated

Tracked parameters

Figure 14 - Tracked body parameters


41

Nutrition Calories intake is tracked by 29.33% of all respondents, water consumption by 23.33% and calories balance by at least 19.33%.

Tracked nutrition related parameter Calories Intake Waterconsumption Calories Balance Nutritional supplements Other

All respondents

WELL-BEING ONLY

Freq. 44 35 29 29

Percent 29.33% 23.33% 19.33% 19.33%

Freq. 23 19 11 12

Percent 27.38% 22.62% 13.10% 14.29%

10

6.67%

4

4.76%

WELL-BEING AND HEALTH Freq. Percent 19 43.18% 14 31.82% 17 38.64% 15 34.09% 6

X²-Test p 0.071 0.258 0.001 0.009

13.64%

0.076

Table 14 - Tracked nutrition parameters for all reference groups

As it can be comprehended in Table 14, there are significant differences between the reference groups when it comes to the tracking of calories balance ( and nutritional supplements (

-Test; df = 1; p = 0.001)

-Test; df = 1; p = 0.009). In both cases, the WELL-BEING AND

HEALTH self-trackers rank higher than the WELL-BEING ONLY self-trackers

Well-being In this section, sleep time and quality as well as mood are the three most important tracked parameters as it can be comprehended in Table 14. 44.00% of all respondents track their sleep time, 32.00% also track their sleep quality, and the tracking of mood can be counted with 19.33%.

Tracked wellbeing related parameter Sleep time Sleep quality Mood General well-being Dreams Other

All respondents

WELL-BEING ONLY

Freq. 66 48 38 29 18 1

Freq. 37 26 20 8 10 1

Percent 44.00% 32.00% 25.33% 19.33% 12.00% 0.67%

Percent 44.05% 30.95% 23.81% 9.52% 11.90% 1.19%

WELL-BEING AND HEALTH Freq. Percent 27 61.36% 22 50.00% 16 36.36% 19 43.18% 7 15.91% 0 0.00%

X²-Test p 0.063 0.034 0.134 0.000 0.526 0.467

Table 15 - Tracked well-being related parameters for all reference groups

Comparing the reference groups for the parameter general well-being, it attracts attention that the WELL-BEING AND HEALTH group is significantly more involved into tracking their general well-being (43.18%) compared to only 9.52% in the WELL-BEING ONLY group (

-Test; df = 1; p
10


50

Daily time spent with own self-tracking In Question 9, all respondents have been asked how much time they spend daily on their actual self-tracking activities. The question mode allowed free answers ranging from 0 to 3660 minutes. The six people that answered Question 16 with “I am not tracking anything” (see chapter 4.2.1), have been excluded from this analysis as they shouldn’t distort the picture of current self-trackers.

WELL-

Minutes spent on actual self-tracking per day

All respondents

Number of observations

144

80

44

Min

0.00

0.00

2.00

Max

1,440.00

1,440.00

1,410.00

Mean

52.42

51.73

67.45

Median

15.00

10.00

20.00

176.26

135.89

117.37

Standard Deviation

BEING

ONLY


Table 25 - Descriptive analysis of daily time spent with actual self-tracking for all respondents, WELL-BEING ONLY and WELL-BEING AND HEALTH (Mann-Whitney U-Test, z = -2.584, p = 0.010, 2-tailed) inclusive respondents that indicated tracking more than 60mins per day

As calculated in Table 25 the average time spent on daily self-tracking amounts to 51.42min for self-trackers in general, to 51.73min for WELL-BEING ONLY self-trackers and to 67.45min for WELL-BEING AND HEALTH self-trackers. The 50% median for self-trackers in general amounts to 15.00min, as for WELL-BEING ONLY self-trackers to 10.00min, and for the WELL-BEING AND HEALTH group to 15.00min. This is to the say that 50% of all participants of this survey invest 15.00min and less in their own daily self-tracking. All differences between WELL-BEING ONLY and WELLBEING AND

tailed).

HEALTH self-trackers are significant (Mann-Whitney U-Test, z = -2.584, p = 0.010, 2-


Minutes spent with daily selftracking Upper class limit [min]

51

All respondents

WELL-BEING ONLY


Freq.

Freq.

Freq.

Percent

Percent

Percent

1

8

5.33%

6

8.00%

0

0.00%

2

9

6.00%

5

6.67%

2

5.26%

3

1

0.67%

0

0.00%

0

0.00%

4

0

0.00%

0

0.00%

0

0.00%

5

20

13.33%

15

20.00%

3

7.89%

6

1

0.67%

1

1.33%

0

0.00%

7

0

0.00%

0

0.00%

0

0.00%

8

0

0.00%

0

0.00%

0

0.00%

9

2

1.33%

1

1.33%

1

2.63%

10

29

19.33%

15

20.00%

8

21.05%

> 10

80

53.33%

32

42.11%

24

72.73%

Total

150

100.00%

75

100.00%

38

100.00%

Table 26 - Frequency scale of minutes spent with daily self-tracking for all respondents, WELL-BEING ONLY and WELL-BEING AND HEALTH self-trackers

Regarding the frequency scale for this question in Table 26, it is from special interest, that 11.33% of all self-trackers invest only two minutes or less for the daily tracking of their selves. 25.33% needs five minutes or less.

Comparing the two reference groups, 14.67% of all WELL-BEING ONLY self-trackers invest two minutes or less compared to 5.26% of all WELL-BEING AND HEALTH self-trackers. 34.67% of all WELL-BEING ONLY self-tracker also only need five minutes or less per day compared to 13.15% of the WELL-BEING AND HEALTH group (Mann-Whitney U-Test, z = -2.584, p = 0.010, 2-tailed).


52


Time spent on own daily selftracking 35% 30% 25% 20% 15% 10% 5% 0%

Well-being only Well-being and health-related

Time in Minutes [min/day]

Figure 17 - Minutes spent on own daily self-tracking in comparison for the two reference groups WELL-BEING ONLY and WELL-BEING AND HEALTH

Daily time spent on self-tracking related activities In order to differentiate between the time that people need to actually record their data and parameters, Question 10 forced the respondents to think about the time they spent on activities that are related to self-tracking but do not mean the self-tracking itself: For example research on self-tracking in the internet, attending QS conferences or meetups, discussing results with friends or writing blogs and tweets about self-tracking would be regarded as selftracking related.

As calculated in Table 27, the average time spent on daily self-tracking related activities amounts to only 44.93min for self-trackers in general, to 24.46min for WELL-BEING ONLY selftrackers but to 80.07min per day for WELL-BEING AND HEALTH self-trackers.

The 50% median for self-trackers in general amounts to 15.00min, as for WELL-BEING ONLY selftrackers to 10.00min, and for the WELL-BEING AND HEALTH group to 30.00min. Testing the differences for the two references group with a Mann-Whitney U-Test and a confidence interval of 90%, all differences between WELL-BEING ONLY and WELL-BEING AND HEALTH selftrackers are significant (Mann-Whitney U-Test, z = -3.082, p = 0.002, 2-tailed).


53

Minutes spent daily on self-tracking related activities Number of observations

All respondents

WELL-BEING ONLY


144

80

44

1,420.00

480.00

1,420.00

Mean

44.93

24.46

80.07

Median

15.00

10.00

30.00

131.54

43.62

120.48

Max

Standard Deviation

Table 27 - Descriptive analysis of daily time spent on self-tracking related activities for all respondents, WELLBEING ONLY and WELL-BEING AND HEALTH (Mann-Whitney U-Test, z = -3.082, p = 0.002, 2-tailed) inclusive all respondents that indicated spending more than 60mins per day on self-tracking related activities

Regarding the frequency scale in Table 28 (a detailed frequency distribution can be found in Appendix III, Table 50) it is apparent, that 28.47% of all respondents are not occupied with selftracking related activities at all. Comparing WELL-BEING ONLY and WELL-BEING AND HEALTH selftrackers, more respondents of the WELL-BEING AND HEALTH group are interested in self-tracking related activities besides their own actual self-tracking.

Time spent on selftracking related activities per day

All respondents Freq.

Percent

WELL-BEING ONLY Freq. Percent

WELL-BEING AND HEALTH Freq. Percent

0 min

41

28.47%

24

30.00%

8

18.18%

1 - 5 min

17

11.81%

13

16.25%

3

6.82%

6 - 10 min

12

8.33%

7

8.75%

4

9.09%

11 - 15 min

14

9.72%

8

10.00%

2

4.55%

16 - 30 min

27

18.75%

16

20.00%

10

22.73%

31 - 60 min

15

10.42%

8

10.00%

6

13.64%

1 - 1.5 hours

2

1.39%

1

1.25%

1

2.27%

1.5 - 2 hours

8

5.56%

2

2.50%

6

13.64%

2 - 4 hours

5

3.47%

0

0.00%

3

6.82%

more than four hours

3

2.08%

1

1.25%

1

2.27%

Table 28 - Time spent on self-tracking related activities per day (classified)

Comparison of the time for actual self-tracking and for self-tracking related activities Putting the time exposure for all actual self-tracking into relationship with the time for all selftracking related activities, some differences become clear. Regarding Figure 18 with the help of Table 29, it is obvious that WELL-BEING AND HEALTH self-trackers are more active as soon as it comes to their own self-tracking as well as the engagement in self-related activities.


54

All WELLrespondents BEING ONLY Mean 52.42 51.73

Sort of activity own self-tracking self-tracking related activities


67.45

Median

15.00

10.00

20.00

Mean

44.93

24.46

80.07

Median

15.00

10.00

30.00

Table 29 - Comparison of the time spent with actual self-tracking and time spent on self-tracking related activities by means of median and mean for all reference groups (Mann-Whitney U-Test, z = -3.644, p < 0.001 , 2-tailed).

Time [min/day]

Comparison of the scope of activity for all reference groups 100 80 60 40 20 0 Mean

Median

Mean

actual self-tracking Total

Median

self-tracking related activities

Well-being only

Well-being and health-related

Figure 18 - Comparison of time for actual self-tracking and time for self-tracking related activities by median and mean for all respondents, WELL-BEING ONLY and WELL-BEING AND HEALTH self-trackers

WELL-BEING AND HEALTH self-trackers spend 73.76min per day as well with their own selftracking as with self-tracking related activities. This is significantly almost as double as much as the reference group of WELL-BEING ONLY self-trackers with only 38.09min per day (MannWhitney U-Test, z = -3.644, p < 0.001, 2-tailed).

Total time exposure to self-tracking activities [min/day] All respondents

Mean of total time exposure to all kinds of Self-Tracking 48.68

WELL-BEING ONLY

38.09


73.76

Table 30 - Mean of total time exposure to self-tracking activities (means of the sum of self-tracking and selftracking related activities) for all reference groups (Mann-Whitney U-Test, z = -3.644, p < 0.001 , 2-tailed).


4.2.5

55

Technology

To get an idea of the deployed tools and technologies that self-trackers use, all respondents have been asked to indicate how they are keeping track of everything they are tracking (Question 14) and how much money they spent on self-tracking activities in the last year (Question 15).

Sort of deployed technologies For this question several antagonisms have been of interest: Does the average self-tracker use mobile, online or local software and hardware? Is he more likely to use ready-to-use or selfmade software and does he spend money for self-tracking by purchasing special tools or does he only use abundant tools and devices like his smartphone.

Percent of Freq. all 150 respondents 86 57.33%

Deployed technologies for self-tracking Mobile hardware (Smartphone) and software (apps) Web- and desktop applications

82

54.67%

Self-Tracking dedicated hardware (Fitbit®, heart rate monitor)

72

48.00%

Self-made desktop tools (spreadsheets)

58

38.67%

Pen and Paper

52

34.67%

9

6.00%

359

-

Other Total

Table 31 - Deployed technologies for self-tracking

The average of deployed technologies per user amounts to 2.33. 57.33% of all respondents use mobile hardware and software such as Smartphones and corresponding self-tracking Apps, followed by 54.67% respondents using online web- and ready-to-use local desktop applications. It is striking that 48.00% of all respondents spend money on extra self-tracking dedicated hardware such as the Fitbit® or heart rate monitors.

Nonetheless, whereas the average respondent uses 2.33 technologies, 27.33% of all participants use only one exclusive technology for self-tracking, thus 8.76% respondents indicated using only Pen and Paper. A complete itemization can be found in Appendix III, Table 49.


56

Personal self-tracking expenditures As at least 48.00% of all self-trackers indicated in Table 31 that they do not only invest timewisely but also financially in self-tracking (by using self-dedicated devices such as the Fitbit®), a distribution of all self-tracking expenditures will be given in this section.

In contrast to these 48.00% indicating that they spend money on self-tracking dedicated devices (Table 31), 39.33% said they spent no money ($0) on self-tracking at all in the past year as it can be found in Table 32. The remaining 12.67% actually do spent money for self-tracking, but not on self-tracking dedicated hardware as exampled in the question.

All respondents

Expenditures on selftracking activities

Freq.

Percent

WELL-BEING ONLY Freq. Percent

WELL-BEING AND HEALTH Freq. Percent

$0

59

39.33%

32

38.10%

12

27.27%

$1 - 50

23

15.33%

9

10.71%

10

22.73%

$51-250

35

23.33%

28

33.33%

7

15.91%

$251 - 500

17

11.33%

8

9.52%

7

15.91%

$501 - 750

6

4.00%

2

2.38%

4

9.09%

$751 - 1,000

4

2.67%

2

2.38%

1

2.27%

more than $1,000

6

4.00%

3

3.57%

3

6.82%

Total

150

100.00%

84

100.00%

44 100.00%

Table 32 - Self-tracking expenditures for all reference groups

Besides the 39.33% that did not spend extra money in the past year on self-tracking at all, 50.00% of all respondents spent between $1 and $500 on self-tracking in the last year. The remaining 10.66% spent up to $7,000 in the past year. Although the mean expenditures of the two reference groups are seemingly different the differences won’t be discussed any further as the Mann-Whitney U-Test doesn’t show significance for α = 0.05 (Mann-Whitney U-Test, z = -1.016, p = 0.310, 2-tailed) any ocularly differences of WELL-BEING ONLY and WELL-BEING AND HEALTH self-trackers that can be seen in Table 33 or Table 32 won’t be regarded any further.


Self-tracking expenditures Number of observations

57

All WELLWELL-BEING respondents BEING ONLY AND HEALTH 146 84 44

Min

$0,00

$0,00

$0,00

Max

$7,000.00

$2,539.40

$7,000.00

$250.75

$187.07

$438.30

$32.83

$65.66

$75.00

$705.76

$301.43

$650.07

Mean Median Standard Deviation

Table 33 - Self-tracking expenditures: Comparison of all references groups (Mann-Whitney U-Test, z = -1.016, p = 0.310, 2-tailed).

4.2.6

Motivations

Providing a better understanding of the motivations of self-tracking had been the leading motivation for this bachelor thesis. Not only the interest to get to know who these selftrackers are and what they are exactly doing, but also why they are spending so much time, money and efforts on self-tracking activities. Chapter 4.2.3 already gave some clue concerning the objects of self-tracking; this is to say concrete goals of self-tracking. This chapter aims to understand the deeper underlying motivations: What exactly keeps a self-tracker working on his goals? Is he rather driven by the fun and entertaining aspects of self-tracking or rather by a strong self-responsibility and willingness to reach his goals?

As explained in chapter 4.1.3, theoretical research as well as expert interviews, discussions and feedback loops have led to a final set of 31 items covering 15 presumed types of motivations (see chapter 4.1.3, Table 2). Deeper extrinsic and intrinsic psychological motivations (e.g. selfresponsibility, self-optimization …) have been questioned as well as more superficial, tangible and obvious underlying motivations (e.g. fun, LifeLogging, …). In the final questionnaire, each item of the motivation questions has been scored on a 1 to 5 Likert scale, with scale scores ranging from “Strongly Disagree”, “Disagree”, “Neutral”, “Agree” to “Strongly Agree” (view Question 13). They will now be evaluated in detail in several steps.

To begin the following evaluation with, all single items will be evaluated individually on their general statistics such as minima, maxima, means and correlations for each reference group. In a second step, an exploratory factor analysis (EFA) will be conducted. An EFA helps breaking down the total amount of 31 single items into a determined number of factors and thereby helps classifying the statements into categories of a motivation factor model. The complete EFA process as well as the deriving motivation factor model, the Five Factor Framework of Self-


58

Tracking Motivations, will be discussed and introduced in the second part of this chapter. In the last part of this chapter, further evaluations on the reference groups will be provided based on the developed five factor framework of self-tracking motivations. 4.2.6.1 General Statistics on Motivations Before having a look at the interrelated structure of all items, this chapter will discuss several statistics that are helpful to get an idea of the overall motivations of self-tracking.

All items at a glance Table 35 depicts all items of the motivations section, as well as their scored minimum and maximum on a 1 to 5 ranging Likert scale, as well as the mean accordance and the 50% median. The items are all sorted descending by their mean. The highest accordance for all selftrackers for item (#27) “I’m tracking because I feel responsible for my life” amounts to 90.00% of all self-trackers indicating that they either “Agree” or even “Strongly Agree”. All self-trackers are also interested in “… optimizing the way they are living” (#20) with a mean of 4.20 on the Likert Scale and in “… controlling what they are doing with their life” (#25) with a mean of 4.10.

Only the last two items score below the mean Likert scale score of 2.50: item #3 “… I enjoy getting lost totally in self-tracking activities” with a mean accordance of 2.42 and item #29 “… many of my online friends are doing so as well” with a mean accordance of only 2.29 on the Likert scale.

Mean accordance of each reference group Comparing the mean accordance on the Likert scale for all items for both reference groups, differences in the mean accordance as in Table 34 are not significant (T-Test; p = 0.1139).

Reference group All respondents WELL-BEING ONLY WELL-BEING AND HEALTH

Mean accordance 3.38 3.33 3.57

Table 34 - Mean overall motivation for all reference groups (T-Test; p = 0.1139).


59

Frequency distribution of Likert answers Strongly Agree (5)

Total

2.00%

8.00% 41.33% 48.67%

1.00

2.00%

9.33% 50.00% 37.33%

1.00

0.67%

4.67% 10.67% 52.00% 32.00%

1.00

4.09

1.33%

4.67% 10.00% 51.33% 32.67%

1.00

4

4.05

0.67%

4.67% 17.33% 44.00% 33.33%

1.00

5

4

3.99

1.33%

5.33% 16.67% 46.00% 30.67%

1.00

1

5

4

3.91

2.00%

8.00% 16.67% 44.00% 29.33%

1.00

1

5

4

3.85

2.00% 10.67% 16.00% 43.33% 28.00%

1.00

1

5

4

3.78

3.33% 12.67% 16.67% 37.33% 30.00%

1.00

... I enjoy being my own master. ... otherwise I would forget what I have done in the past. ... I try to manipulate certain aspects in my life.

1

5

4

3.77

0.67% 10.67% 23.33% 41.33% 24.00%

1.00

1

5

4

3.74

2.00% 12.00% 18.67% 44.67% 22.67%

1.00

1

5

4

3.73

2.00% 12.67% 16.00% 49.33% 20.00%

1.00

... it is fun and entertaining. ... I like playing around with my smartphone/ technical device etc.. ... I like playing around with numbers/statistics etc. ... it helps me to overcome selfdeception. ... I enjoy being part of the Quantified Self movement.

1

5

4

3.65

4.00% 12.67% 18.67% 43.33% 21.33%

1.00

1

5

4

3.63

6.00% 12.67% 20.00% 34.67% 26.67%

1.00

1

5

4

3.57

4.67% 16.00% 18.67% 38.67% 22.00%

1.00

1

5

4

3.51

5.33% 14.00% 25.33% 34.67% 20.67%

1.00

1

5

3

3.27 10.00% 12.00% 36.00% 25.33% 16.67%

1.00

22

... I want to be someone.

1

5

3

3.26

9.33% 16.67% 30.00% 26.67% 17.33%

1.00

19

... It allows me to reward myself. ... the way I'm doing is interesting for others/might help others

1

5

3

3.25

7.33% 16.67% 32.00% 32.00% 12.00%

1.00

1

5

3

3.04

9.33% 24.67% 32.67% 19.33% 14.00%

1.00

... I want to help/inspire others. ... I want to present myself to others. ... I want to compare my results to others. ... I enjoy being someone in the Quantified Self community. ... I enjoy forgetting about time while doing so. ... I found a unique technique doing so. ... I want to be independent from traditional medical treatments. ... I don't trust in the health care system/classic therapies. ... I want to see wherein I am better than others. ... I enjoy getting lost totally in selftracking activities. ... many of my (online) friends are doing so as well.

1

5

3

3.03 12.00% 24.00% 26.00% 24.67% 13.33%

1.00

1

5

3

2.92 11.33% 23.33% 36.00% 20.67%

8.67%

1.00

1

5

3

2.89 11.33% 26.00% 34.00% 19.33%

9.33%

1.00

1

5

3

2.86 14.67% 20.00% 40.00% 15.33% 10.00%

1.00

1

5

3

2.84 10.00% 26.00% 40.00% 18.00%

6.00%

1.00

1

5

3

2.75 11.33% 34.00% 32.67% 12.67%

9.33%

1.00

1

5

3

2.73 13.33% 32.67% 26.00% 23.33%

4.67%

1.00

1

5

3

2.71

8.00% 42.00% 28.00% 14.67%

7.33%

1.00

1

5

2

2.52 18.00% 38.00% 26.00% 10.00%

8.00%

1.00

1

5

2

2.42 19.33% 37.33% 29.33% 10.00%

4.00%

1.00

1

5

2

2.29 22.00% 40.00% 27.33%

2.67%

1.00

# 27 20 25 24 17 11 16 9 18 26 12 10 1 5 4 13 28

8 6 23 14 21 2 7 31 30 15 3 29

Question: I am self-tracking because …

Strongly Disagree (1)

Min

Max

Median

Mean

... I feel responsible for my life. ... it helps me to optimize the way I'm living. ... I want to control what I'm doing with my life. ... numbers help me to reflect on what I'm doing. ... it motivates me to keep on working for a goal. ... I like keeping track of what I'm doing.

2

5

4

4.37

0.00%

1

5

4

4.20

1.33%

1

5

4

4.10

1

5

4

1

5

1

... it facilitates my self-discipline. ... I'm interested in how certain things in (my) life interact. ... I'm the only one who is able to change myself.

Disagree (2)

Neutral (3)

Agree (4)

8.00%

Table 35 - General Statistics for all motivation items at a glance for all 150 respondents


60

Correlations between all items It is necessary to understand, which items correlate with other items. This generally helps getting an idea, which items (motivations) go together and might have something in common. The correlation coefficients in the correlation matrix as depicted Table 36 in attain values between -1 and +1. Correlations above 0.50 are considered to proof a correlation between the crossed items.

The highest correlation amounts to 0.82 and can be found for item #21 “… I enjoy being someone in the QS community” and #28 “… I enjoy being part of the Quantified Self movement. The second highest correlation with a correlation coefficient of 0.77 can be found for item #14 “… I want to compare my results to others” and #15 “… I want to see wherein I am better than others”.

It is remarkably that there are no lower values below -0.09 (item #4 “… I like playing around with numbers/statistics” and #18 “… I’m the only one who is able to change myself” as well as item #17 “… it motivates me to keep on working for a goal” and #2 “… I enjoy forgetting about time while doing so”).


Table 36 - Correlation matrix of all motivation items

61


62

4.2.6.2 Processing of the Exploratory Factor Analysis (EFA) An EFA comes to use when a set of variables needs to be reduced by detecting structures in the relationship between the variables. Thereto, an EFA classifies these variables into interrelated factors. “Psychologists and educators have used the technique to determine how people perceive different “stimuli” and categorize them into different response sets.” (Kim & Mueller, 1978, p. 6). Classically, an EFA presumes that all observed variables (here: the statements) are linear combinations of some underlying factors (here: the assumed underlying motivations), whereby these underlying factors can be either hypothetical or unobservable (Kim & Mueller, 1978, p. 9).

Generally, a factor analysis consists of three steps: “(1) the preparation of an appropriate covariance (or correlation) matrix, (2) the extraction of initial (orthogonal) factors [and determination of number of factors]; and (3) rotation [here: Varimax rotation] to a terminal solution” (Kim & Mueller, 1978, p. 10). All of these steps are automatically covered by IBM® SPSS Statistics and won’t be derived in detail any further. The final rotated component matrix with five factors (Table 37) – that derived from a Parallel Analysis and after a Varimax rotation – will be examined in detail in this chapter.

Step-by-step description of the EFA: Conduction of a Parallel Analysis (PA) As always for EFAs, one of the most important questions a researcher has to take into consideration is the optimal number of factors. For this bachelor thesis, a Parallel Analysis (PA) has been executed to derive the optimum amount of factors. The PA has been developed by HORN in 1965 (Horn, 1965) and is generally recognized as “… the best available alternative for solving number-of-factors-retain-problem in EFA …” (Ledesma, 2007, p. 4).


63

8 7

Factor loading

6 5

EFA: Unrotated Eigenvalues

4

PA: Unrotated Eigenvalues

3 2 1 0 1

3

5

7

9

11 13 15 17 19 21 23 25 27 29 31

Number of factors Figure 19 - Screeplot of unrotated eigenvalues of all 31 items after EFA and PA. The point of intersection can be found at five factors and indicates the optimum amount of factors for the factor analysis.

The comparison of the eigenvalues from the PA with 31 fictional items and 150 fictional participants with the eigenvalues of the real survey data proofed that the best number of factors should be at five factors (Figure 19). The consequently conducted second factor analysis with five factors will be discussed by a step-by-step derivation of the filtering algorithm.

Step-by-step derivation of the final rotated component matrix by several filtering iterations The complete EFA has been conducted in several iterations by the principal components method. In each iteration, a Varimax rotation26 has been applied to make the factors better interpretable. Additionally, in each iteration the following four criteria (filters) helped filtering and dropping not-well-correlating questions:

26



Filter #1: No major factor loadings27 (all factor loadings