IT 16 034

Examensarbete 30 hpJuni 2016

The effects of push vs. pull notifications on overall smartphone usage, frequency of usage and stress levels.

Marin Mikulic

Institutionen för informationsteknologiDepartment of Information Technology

Teknisk- naturvetenskaplig fakultet UTH-enheten Besöksadress: Ångströmlaboratoriet Lägerhyddsvägen 1 Hus 4, Plan 0 Postadress: Box 536 751 21 Uppsala Telefon: 018 – 471 30 03 Telefax: 018 – 471 30 00 Hemsida: http://www.teknat.uu.se/student

Abstract

The effects of push vs. pull notifications on overallsmartphone usage, frequency of usage and stresslevels.Marin Mikulic

Consumers of media (e.g. television, the Internet, smartphones) have been found toexperience both media overuse and the development of negative media habits due todifferent personality traits, life events, operant conditioning processes and deficientself-regulation (Ifinedo, 2016; LaRose, Lin, & Eastin, 2003; McIlwraith, 1998;Oulasvirta, Rattenbury, Ma, & Raita, 2012). Recent research found that smartphones,too, seem to be conducive to both overuse and habit formation (Haug et al., 2015;Lee et al., 2014; Oulasvirta et al., 2012). This study turns its attention toward asmartphone staple functionality – smartphone notifications, in order to determinetheir effect on overall usage and frequency of usage, as they were already found tocause stress and inattention (Kushlev, Dunn, & Proulx, 2016). The study recruitedthree participant groups. Each group had different notification settings (none,maximum, control group) during a ten-day study period. Analysis of variance statisticalmethod (ANOVA) was used and results indicate that manipulating the daily amount ofsmartphone notifications does not affect overall smartphone usage, possibly due toalready established device usage habits and past experience.

Tryckt av: Reprocentralen ITC IT 16 034Examinator: Anders JanssonÄmnesgranskare: Mats LindHandledare: Edward Peter Greenwood White

1

TABLE OF CONTENTS: ABSTRACT .................................................................................................................................... 1

TABLE OF FIGURES: ................................................................................................................... 2

1 INTRODUCTION ................................................................................................................... 3

2 LITERATURE REVIEW ........................................................................................................ 5

2.1 Theory behind smartphone usage ..................................................................................... 5

2.1.1 Uses and Gratifications (U&G) ................................................................................ 5

2.1.2 Technology Acceptance Model (and related models) .............................................. 6

2.1.3 Individual differences and psychological predictors ................................................ 7

2.2 Theory behind unregulated smartphone usage ............................................................... 10

2.2.1 The Disease Model of Addiction ............................................................................ 10

2.2.2 Social Cognitive Theory (SCT) .............................................................................. 12

2.2.3 Habit formation ....................................................................................................... 13

2.3 Smartphone ubiquity and habit formation ...................................................................... 14

2.3.1 Applications ............................................................................................................ 14

2.3.2 Notifications ............................................................................................................ 15

3 Literature review findings ..................................................................................................... 18

4 AIM AND RESEARCH QUESTIONS ................................................................................. 20

5 METHODOLOGY ................................................................................................................ 21

5.1 Literature review ............................................................................................................ 21

5.2 Data collection and analysis ........................................................................................... 21

5.2.1 Quantitative Questionnaires .................................................................................... 21

5.2.2 Usage Tracking ....................................................................................................... 21

5.2.3 Statistics .................................................................................................................. 22

5.2.4 Sample..................................................................................................................... 22

5.2.5 Relevant variables ................................................................................................... 23

5.2.6 Study Period ............................................................................................................ 23

5.2.7 ......................................................... 23

6 RESULTS .............................................................................................................................. 28

6.1 Results for all participants .............................................................................................. 28

6.1.1 Group 1 (No Notifications) ..................................................................................... 28

6.1.2 Group 2 (Max Notifications) .................................................................................. 29

2

6.2 Results for participants with complete datasets ............................................................. 29

6.2.1 Group 1 (No Notifications) ..................................................................................... 29

6.2.2 Group 2 (Max Notifications) .................................................................................. 30

6.3 Questionnaire Results ..................................................................................................... 31

7 DISCUSSION ........................................................................................................................ 32

8 CONCLUSION ..................................................................................................................... 33

8.1 Limitations ..................................................................................................................... 33

9 FUTURE WORK .................................................................................................................. 35

10 REFERENCES .................................................................................................................. 36

11 APPENDIX A Stress Questionnaire ............................................................................... 42

12 APPENDIX B Happiness Questionnaire ........................................................................ 45

13 APPENDIX C Demographics Questionnaire ................................................................. 47

14 APPENDIX D Participant Demographics ...................................................................... 51

TABLE OF FIGURES: Figure 1 - Comparison of theories of technology usage and their findings .................................................................. 9Figure 2 - Quality Time initial screen, showing total smartphone daily usage and timeline activity ......................... 24Figure 3 - Total smartphone daily usage and usage per app ...................................................................................... 25Figure 4 - Total smartphone usage frequency per app ................................................................................................ 26Figure 4- Total Daily Usage, Group 1 compared to Group 3 (control group) ........................................................... 28Figure 5 - Total Daily Usage, Group 2 compared to Group 3 (control group) .......................................................... 29Figure 6 - (complete dataset) Total Daily Usage, Group 1 compared to Group 3 (control group) ............................ 30Figure 7 - (complete dataset) Total Daily Usage, Group 2 compared to group 3 (control group) ............................. 30Figure 8 - pre/post Stress Questionnaire, Group 1 ..................................................................................................... 31Figure 9 - pre/post Stress Questionnaire, Group 2 .................................................................................................... 31

3

1 INTRODUCTION The capacity to exercise control over one's own thought processes,

motivation, and action is a distinctively human characteristic. Because judgments and actions are partly self-determined, people can effect change in

(Bandura, 1989, p. 1175)

There is no doubt technology has improved the general quality of life (Littleton & Light, 2002). It has affected our leisure, working life, education and nearly every other aspect of everyday life

(Littleton & Light, 2002; Oulasvirta et al., 2012). We have delegated physical and menial tasks to machines and processors (Säljö, 1998), thus enabling ourselves to dedicate precious cognitive capacity to yet unsolved problems we face. We now offload data and arithmetic calculations to super powered computers (Scannapieco, 2013; Wang & Harris, 2013) that can fill rooms with their processing circuits and petabytes of data so we would not be bogged down with enormous amounts of information (Scannapieco, 2013).

(Soukup, 2015). Laptops, tablets and smartphones are now a part of many interactions (Oulasvirta et al., 2012) because of their immense usefulness: they remember for us, remind us, provide necessary computing power and help us communicate over enormous distances (Soukup, 2015). Technology, especially mobile technology, has given us a level of interconnectedness never before possible (Soukup, 2015). Not only are we able to connect with others, but we are also able to do it almost instantly (Dery, Kolb, & MacCormick, 2014). Many of the needs we have can be

- the smartphone (Oulasvirta et al., 2012). Whether we require working tools, contact, play or entertainment, the smartphone provides all there, truly at our fingerprints (Soukup, 2015).

However, smartphones and the instant gratification they provide are conducive to unconscious habit-formation our brains are prone to (Oulasvirta et al., 2012; Park & Lee, 2012). Because smartphones offer a wealth of content (Soukup, 2015) and aim to always re-engage the user through various notifications (Duan, Dong, & Gopalan, 2005), we tend to use these devices in a repetitive manner (Deloitte, 2015; Oulasvirta et al., 2012). We check social media, over and over again (Ifinedo, 2016; Pempek, Yermolayeva, & Calvert, 2009), because it is something we have done before. We begin to use our devices more and more until we approach and even exceed the global average of 3.3 hours (Deloitte, 2015; Scott & Sale, 2014), each day.

But what facilitates this burgeoning interaction, besides interesting content and functionality? functionalities

notifications. s usefulness by seeking to have the consumer reengage with the device on a constant basis . Through this feature, new content is efficiently brought to the chaving to constantly, and intentionally, request it (Miajung, 2014). However, the lack of intentionality means that notifications are delivered on an irregular schedule, without any control on the part of the consumer (Duan et al., 2005; Stothart, Mitchum, & Yehnert, 2015). Only after

4

receiving the notification is the consumer able to process it, once he or she has already been interrupted via notification sounds and vibrations (Stothart et al., 2015). -ending flood of new content was found to facilitate further smartphone activity, often unrelated to the original content notification (Oulasvirta et al., 2012).

Notifications expand smartphone capabilities and turn them into complex communication devices that seek to attract and retain the ubiquitously, but may also facilitate unfavorable habit formation and overuse. This thesis proposes that notifications require intense scrutiny in relation to what we know about smartphones, habit formation and the way consumers use technology in order to discern whether a higher degree of control over the delivery of notifications is required.

5

2 LITERATURE REVIEW A wide array of different theories and model attempts to explain the diverse factors that influence how consumers use their smartphones, what best predicts future/continued use and how smartphone usage can become unregulated (Bian & Leung, 2015; Haug et al., 2015; Kushlev et al., 2016; Oulasvirta et al., 2012).

Unregulated usage can be defined as usage which is repetitive (Oulasvirta et al., 2012) and not actively attended t (LaRose et al., 2003, p. 234) in terms of conscious control. Well-regulated

usage, on the other hand, is intentional and marked by a purposeful consideration of needs, gratifications and outcomes (LaRose et al., 2003).

2.1 Theory behind smartphone usage The theories discussed below are: Uses and Gratification, Technology Acceptance Model, Unified Theory of Acceptance and Use of Technology, individual preferences, operant conditioning processes, personality traits, social cognitive theory and habit formation. These theories provide insights into how we use technology in general and smartphones in particular, including habitual or unregulated usage.

2.1.1 Uses and Gratifications (U&G) Uses and Gratifications (U&G) is a theory which examines how and why consumers use different forms of media to satisfy needs (Joo & Sang, 2013). The basic premise of the U&G is

rent media is intentional, purpose oriented and done to support or complete a socio-psychological need or desire (Weiyan, 2015). Additionally, the theory makes

prior involvement with media helps make a more informed choice and that c) various groups of users use different media as one of the many ways to support or complete a need they might have (Livaditi, Vassilopoulou, Lougos, & Chorianopoulos, 2002).

The U&G framework has been used to examine a wide variety of communications media affects such as the television (Livaditi et al., 2002), the Internet (LaRose, Mastro, & Eastin, 2001; Roy, 2009; Stafford, 2003) and social media (Chen, 2011; Ifinedo, 2016). Their findings explain media consumption as an attempt to satisfy perceived needs and gratifications (Livaditi et al., 2002). Additionally, media satisfies both instrumental and ritualistic usage (Livaditi et al., 2002; Rubin, 1984), explained below in greater detail.

In recent years, U&G research has examined smartphones (Joo & Sang, 2013; Kushlev et al., 2016; Miajung, 2014; Park & Lee, 2012;) and their relationship with the Internet. In addition to common telephony functionality, smartphones provide an immediate connection to the Internet and are, in a sense, a gateway to it (Soukup, 2015). Research actually regards smartphones as devices which serve as a point of transfer or as a springboard toward the immersive experience of the Internet (Soukup, 2015) (Soukup, 2015).

6

U&G research has focused on discovering and defining most common motives consumers have for using smartphones: interaction, diversion, instant gratification, portability and social sense of status (Miajung, 2014), as well as the motives for using more particular functions housed within the concept of a smartphone, e.g. Short Message Service (SMS). Motives to use a technology shape how we use it (Miajung, 2014) and U&G offers important insights into two distinct types of usage:

Instrumental usage - defined as a goal-oriented use of media content (Joo & Sang, 2013, p. 2513), employed to satisfy a need or a motive for information (Rubin, 1984).

Ritualized usage - defined as a habitual state employed for attention diversion and for (Livaditi et al., 2002, p. 2).

In other words, instrumental usage suggests utility and intention while ritualistic usage is diversionary and habitual (Livaditi et al., 2002), the latter of which may result in increased usage of media (e.g. increased smartphone usage through excessive, diversionary checking of new content on social media). Ritualistic usage can be defined as one where activities are generally

the activity in the first place (Joo & Sang, 2013). The smartphone and, by extension the Internet, allows the consumer to repeatedly choose (and come back to) a specific source or multiple sources of content (Joo & Sang, 2013). Consumers can choose content at will and receive it almost instantaneously to satisfy their ritualistic needs. According to (Miajung, 2014) and (Limayem, Hirt, & Cheung, 2007), ritualistic usage is the strongest predictor of future (continued) use more so than any other factor (e.g. motivational, social, etc.).

Since U&G research shows that ritualistic usage of quickly changing content can be conducive to habit formation (Haug et al., 2015; Oulasvirta et al., 2012; van Deursen, Bolle, Hegner, & Kommers, 2015), the aim of this study is to demonstrate that certain aspects of smartphone architecture (i.e. notification systems) also facilitate habitual behavior and contribute to unregulated behavior (van Deursen et al., 2015).

2.1.2 Technology Acceptance Model (and related models) Technology Acceptance Model explains individual behavioral intentions to use a technology (Miajung, 2014). The basic premise of the TAM is that a technology consumer, when considering choosing and adopting a new technology (e.g. a smartphone), forms two mental models about the technology: perceived usefulness (PU) and the perceived ease of use (PEU) (Kessler, 2013). The intention to use and adopt a technology depends on the relation between PU and PEU, according to TAM.

However, additional variables were necessary to complement the TAM because of a serious lack of social and other considerations inherent in the model. As seen in mobile research using the TAM, researchers had to build on the TAM by implementing other variables, e.g. anxiety, social pressure, social norms, self-efficacy, etc. (Kwon et al., 2000; in (Miajung, 2014). This demonstrates that consumer choices cannot solely be explained by PU and PEU.

To expand the TAM and provide for its shortcomings (Venkatesh, Morris, Davis, & Davis, 2003) expanded the original model into the Unified Theory of Acceptance and Use of Technology

7

(UTAUT). The new model integrated four new core factors to better explain technology adoption and use: (1) performance expectancy, (2) effort expectancy, (3) social influence, and (4) facilitating (Miajung, 2014, p. 18). The UTAUT model had a higher success explaining variance (e.g. behavioral intention, usage behavior) in comparison to the TAM model. When applied to smartphone adoption, UTAUT consistently proved that expectancy, effort expectancy, and social influence explained mobile phone usage inten(Miajung, 2014, p. 8). But, even though the UTAUT was more successful than TAM, both theories failed to consider the effects of previous experience and habit on technology adoption. Research found that prior experience (Agarwal & Prasad, 1997; Ally & Gardiner, 2012; Taylor & Todd, 1995) and habit (Kim, 2012; Ouellette & Wood, 1998; Oulasvirta et al., 2012) positively affected and best predicted continued technology usage and so must be considered when researching smartphones. Hence, both theories were superseded by the UTAUT2 model which included different consumer contexts: habits, hedonic motivations and the price value (Venkatesh, L. Thong, & Xu, 2012). In the context of smartphone acceptance and use, UTAUT2 better explained variances in behavioral intention. A study conducted by (Venkatesh et al., 2012) found that habits have a

(Miajung, 2014, p. 9), more so than the actual intent of use. Habits seem to have both direct and mediated effects on technology use, and these effects are moderated by individual differences (Venkatesh et al., 2012). Additionally, (Venkatesh et al., 2012) incorporated two important theoretical views (Ouellette & Wood, 1998) on habits: the stored intention view (Ajzen, 2002) and the automaticity view (Limayem et al., 2007). Based on the automaticity view, the multitude of changes that can occur in a smartphone

connection between what triggers a certain behavior and the behavior itself (Venkatesh et al., 2012). Similar to the U&G framework, UTAUT2 demonstrates the positive effect of habit on predicting continued smartphone use. Moreover, UTAUT2 also shows the significance of cues (i.e. notifications) in predicting overall smartphone usage (the automaticity view). This research proposes that a significant change their behavior and causes a change in overall usage, regardless of consumer intentions.

2.1.3 Individual differences and psychological predictors Previous studies have explored the effects of gender, age and the common personality traits (extraversion, agreeableness, conscientiousness, neuroticism, and openness to experience) on smartphone usage and problematic usage (Bian & Leung, 2015; Bianchi & Phillips, 2005; Butt & Phillips, 2008; Chittaranjan, Blom, & Gatica-Perez; Uffen, Kaemmerer, & Breitner, 2013).

Extraversion, defined as a person who thrives on excitement, makes relatively impulsive choices and is partial to taking chances (Bianchi & Phillips, 2005), was found to be a positive predictor of overall use as well as problematic use. Self-esteem, i.e. an evaluation of self-worth and

8

identity (Bianchi & Phillips, 2005), was found to predict problematic phone usage but not the overall usage. Since low-esteem has been consistently linked to addiction and addiction-like behavior, prior research supports the presumption that people with lower self-esteem have a greater likelihood toward developing unregulated smartphone use (Bianchi & Phillips, 2005).

Age was found to predict unregulated usage. Elderly individuals were found to be less capable when exploring a smartphone device (Al-Showarah, AL-Jawad, & Sellahewa, 2014) and to have a less positive attitude toward technology compared to young individuals (Bianchi & Phillips, 2005).

Neuroticism was not found to be a significant predictor of neither smartphone usage nor problematic usage, even though neuroticism has often been linked to some types of addictive behaviors (Bianchi & Phillips, 2005). Smartphone consumers who show characteristics of neuroticism seem to find less appeal in the smartphone. The fact that a smartphone device signifies constant connectedness may potentially make it less favorable in the eyes of neurotic people. This could well, compared to extroversion and lower self-esteem (Bianchi & Phillips, 2005).

While studies that compare personality traits and smartphone use do not directly touch on habits and habitual behavior, certain personality traits (i.e. extroversion, low self-esteem) and consumer demographics (i.e. age) can possibly cause an increase in smartphone usage and the development of unregulated smartphone usage (Bianchi & Phillips, 2005; Butt & Phillips, 2008; Chittaranjan et al.). The effect of personality traits can be further exacerbated by a torrent of notifications

(Kushlev et al., 2016; Park & Lee, 2012) since, in many cases, they attend to the various beeps and vibrations almost immediately (Kushlev et al., 2016). Reducing the amount of notifications is important not only for the potential decrease of smartphone usage in general population, but also in those segments of population that are more prone to prolonged usage due to specific personality traits or age.

Overall, the studies concerning smartphone adoption and continued use concluded that a) habitual behavior was the strongest predictor of continued use, more so than any other, and b) personality traits have an impact on overall smartphone use and may have a compounding effect toward unregulated use (Miajung, 2014). Habit in particular is an important factor as it has repeatedly been found to determine continued use of communications media more than other factors (social, motivational) (Table 1) and so necessitating further research into the particular components of smartphone devices (e.g. notifications) that may induce habitual behavior.

9

Approach Studies Study Purpose Findings

U&G

Livaditi, Vassilopoulou, Lougos, & Chorianopoulos, 2002;

Define needs and motives for media consumption.

Ritualized usage best predictor of continued use.

Limayem, Hirt, & Cheung, 2007;

Examine the role of habit in continued use of technology.

Past experience and frequency of experience affect continued use. Habit more predictive than usage intentions.

TAM

Kim, 2012; Role of habit in the context of smartphone applications.

Both the intention of continued use and habit play important role within the smartphone environment.

Davis, 1989; Examine what predicts continued technology usage.

Perceived usefulness (PU) and perceived ease of use (PEU) partially explain variance in technology use.

UTAUT Venkatesh, Morris, Davis, & Davis, 2003;

To extend TAM with expectations, social influence and facilitating conditions

Implementation of new variables explain more variance compared to the TAM.

UTAUT2 Venkatesh, L. Thong, & Xu, 2012;

To extend TAM and UTAUT with price value, hedonic motivations and habit.

Habit has both a direct and an indirect effect on continued technology use, (affected by individual differences). Defines habit as best predictor of use. Explains more variance than both antecedents.

Individual differences Bianchi & Phillips, 2005; Examine psychological predictors of continued smartphone use.

Age, extraversion and low self-esteem promote continued use, while neuroticism does not.

Figure 1 - Comparison of theories of technology usage and their findings

10

The forthcoming sections demonstrate theories and mechanisms by which consumers can develop unregulated smartphone usage, especially through the facilitating ability of smartphone notifications.

2.2 Theory behind unregulated smartphone usage Consumers of media and communication content via smartphones appear to be susceptible to unregulated use and habit formation theories(Lee, Chang, Lin, & Cheng, 2014; Oulasvirta et al., 2012; van Deursen et al., 2015; Verplanken, 2006). The aforementioned theories attempt to explain the development of unregulated smartphone usage namely, the addictive personality theory and the operant conditioning theory, both of which arise from the disease model of addiction (LaRose et al., 2003). Additionally, Social Cognitive Theory is discussed as a viable alternative to the disease model of addiction (LaRose et al., 2003; Miajung, 2014) .

2.2.1 The Disease Model of Addiction A significant amount of literature on excessive media consumption and unregulated smartphone use is grounded in diagnostic criteria of various known addictions, as defined by the American Psychiatric Association (LaRose et al., 2003). The studies define excessive media consumption as a disease which needs clinical intervention (LaRose et al., 2003). Previous research studied video game addiction through the criteria of pathological, addictive gambling (Griffiths, 1991). (McIlwraith, 1998) used substance abuse criteria to study television addiction and others did the same for internet addiction (Block, 2008; Weinstein & Lejoyeux, 2010). Other research defined the Internet Addiction Disorder using the diagnostic criteria for psychoactive substance abuse (Young, 1998).

It is important to note that the term addiction has become less frequently used in recent years as excessive mobile consumption habits do not fall under the category of addiction as defined by the (American Psychiatric Association, 2013), because consumers are able to sometimes overcome these addictions without clinical intervention (LaRose et al., 2003; Marlatt, 1988). The term itself has been phased out in favor of the term dependence (Kubey, 1996; LaRose et al., 2003). Mobile

. This research refers to the problem as unregulated media usage through a smartphone device, a term coined by (LaRose et al., 2003).

There is no one unifying theory that explain the origins and development of unregulated media usage. Based on the studies within the media addiction as a disease view, two models have dominated past research: the addictive personality model and the operant conditioning processes model (LaRose et al., 2003).

2.2.1.1 Addictive Personality The addictive personality model theory describes the origins of addiction and dependence through a specific type of personality, namely an oral, neurotic, sensation-seeking, or

(LaRose et al., 2003, p. 229) or a personality with defective ego self-sufficiency. This theory was originally supported by research that found common personality traits across the television, Internet, video games and alcohol addictions

11

(LaRose et al., 2003). Further research found moderate correlations between various addictions (i.e. television, alcohol, the Internet) and personality traits (McIlwraith, 1998).

However, additional studies found that similarities between different addictions can be a result of demographics, culture, lifestyle or other factors, rather than personality alone (LaRose et al., 2003; Rozin & Stoess, 1993). Common factors found in different addictions can be explained through the learning processes that occur across different substances and behaviors (Marlatt, 1988) and research shows that diction,

(LaRose et al., 2003; Nathan, 1988). Personality traits are one of many factors related to unregulated media usage, rather than the key factor and cannot alone account for its progression.

2.2.1.2 Operant Conditioning The operant conditioning model explains the progression of media consumption behavior in four phases: initiation, transition to ongoing use, addiction, and behavior change (LaRose et al., 2003, p. 230) (Dery et al., 2014) and plays a definitive role in the development of unregulated smartphone use. Outcome expectations regarding a behavior play a role in the transition to habitual use, but due to the process of classical operant conditioning, the behavior can spiral out of control and push the behavior toward the addiction extreme (LaRose et al., 2003).

For example, in the initiation phase, sending an instant message through a smartphone app is experienced as pleasurable and rewarding (Lee et al., 2014; Neo & Skoric, 2009). Then, because of perceived uses, gratifications and outcomes, the behavior may become more frequent (Haug et al., 2015; Miajung, 2014). At that point, coupled with attention-drawing notifications, sending instant messages can begin exhibiting habitual or ritualistic characteristics while still retaining a conscious quality. The transition to problematic usage can begin if resorting to sending instant messages becomes an important stress, loneliness, depression and/or anxiety relief system. Excessive usage can then cause "life problems, confrontations with significant others, and an inability to stop [...] consumption once started". The subsequent increased stress and anxiety may exacerbate the dependency on the behavior (e.g. instant messaging) as a means to achieve relief (LaRose et al., 2003).

Smartphone provide near instant gratifications of consumer needs through quick access to various media content (Oulasvirta et al., 2012), they are uniquely capable of facilitating the classical operant conditioning cycle, which when coupled with an excessive amount of notifications may serve as behavior triggers and so facilitate the strengthening of the trigger-behavior link (Miajung, 2014).

Ultimately, neither the addictive personality model nor the operant conditioning model fully account for the initiation and progression of unregulated smartphone usage. Research found that only about 6% of online population (Bandura, 1999) and around 12.5 % of smartphone users (Pearson & Hussain, 2015) ever exhibit addiction/dependency behavior a remarkably small percentage considering the salience of both the addictive personality traits in the general

12

population and of the cues we are all constantly exposed too. Additionally, neither model explains how are clinical intervention, a fact which puts some doubt on the basic premise of the disease model of addiction (LaRose et al., 2003). Despite this, both an addictive personality (McIlwraith, 1998) and operant conditioning processes can lead to increased smartphone usage (Lee et al., 2014), especially when coupled with smartphone notifications which can serve both as reward for the former and as a trigger for the for the latter.

2.2.2 Social Cognitive Theory (SCT) Social Cognitive Theory (SCT) is a psychological model of human behavior which combines reciprocal causality, internal factors (cognitive, affective, biological) various behavioral patterns and environmental influences into a multifaceted, dynamic model in which different parts influence each other bi-directionally and conscious interventions are possible and necessary (Bandura, 2001). In stark opposition to both addictive personality and operant conditioning processes, SCT places significant value on the notion of human agency (Bandura, 1989). According to the SCT, human beings are neither independent agents nor mechanical automatons that simply respond to contextual or personality cues. Instead, human beings are capable of direct intervention, able to add to their own motives and are able to directly cause effects within the triadic reciprocal causation system (Bandura, 1989, p. 1175). Triadic reciprocal causation system describes human behavior as a product of multiple factors: personal characteristics, behavior patterns and social environment, as opposed to it solely being a reflection of the environment.

The notion of human agency (i.e. intentional action) within the system of triadic reciprocal causation is a central to SCT. (Bandura, 1989) argues that people are able to monitor, judge and modify their behavior through the three mechanisms of personal agency, described as follows:

Exercise of Agency through Self-Belief of Efficacy - is the people belief in their own ability to exert influence over the events that have an impact on their lives. It is a central and most

(Bandura, 1989, p. 1179).

Exercise of Agency through Goal Representations - is a unique human ability of forethought which enables people to see the possible consequences of certain actions, adjust personal goals, and think of the best courses of action that will bring them closest to desired outcomes (Bandura, 1989).

Exercise of Agency through Anticipated Outcomes - is the human capacity to imagine the possible results of the available courses of action (Bandura, 1989), which is partly governed by the self-belief of efficacy because the outcomes people anticipate depend on the degree of self-belief the individual in question commands.

13

These mechanisms of personal agency, according to (Bandura, 1989), enable people to intentionally affect their internal and external environment without being at the complete mercy of, for example, addictive personality or operant conditioning processes. This aligns with the findings that many people are able to overcome unregulated technology use (e.g. the Internet) despite possible personality traits and cue salience (LaRose et al., 2003).

2.2.2.1 Self-regulation and deficiency The significance of SCT in terms of unregulated smartphone use lies self-regulation of behavior. Self-regulation is an internal mechanism to self-monitor a course of action, make a judgment based on personal standards and undertake corrective behavior if deemed necessary (Bandura, 2001).

Unregulated media consumption via smartphones may be caused by deficient self-regulation - a mental state where the ability for intentional self-governance is somewhat reduced (LaRose et al., 2003, p. 232) The unregulated behavior (e.g. sending an instant message) occurs repetitively as long as each particular instance of the behavior escapes the notice of attentive self-observation (LaRose et al., 2003, p. 234). Subsequently, as the strength of the behavior increases (operant conditioning), attentive self-observation diminishes (LaRose et al., 2003, p. 234). Deficient self-regulation occurs from various reasons (e.g. traumatic life events, social isolation) and can result in habit formation through the operant conditioning system, but not in the clinical definition of addiction per se. Thus, unregulated or habitual media usage is an existing problem of a relatively benign nature and does not require clinical intervention (LaRose et al., 2003). Decreasing the overall amount of notifications could then oral,

- personality (LaRose et al., 2003, p. 229), reduce the cue salience necessary of continued operant conditioning (Lee et al., 2014) and relieve the strain on the self-regulatory mechanism as defined by (Bandura, 1989).

2.2.3 Habit formation

(Verplanken & Wood, 2006, p. 91). They are processes at the neural level which protect individuals from extensive cognitive load related to routine events. Some tasks become automatized in order to relieve cognitive capacity for other, more important stimuli (LaRose, 2010). less cognitive effort. The control of the action itself transfers from a willing choice to cues and

(Lally, van Jaarsveld, Cornelia H. M., Potts, & Wardle, 2010, p. 998). Habit formation requires the establishment of associative links between actions and specific contextual features (e.g. location, time of day) during which the actions are executed (Verplanken & Wood, 2006). It is also necessary to state that habitual behavior is not uninterruptible and can be overridden by the prefrontal cortex (LaRose, 2010), which may explain why consumers are able to modify habits once they become aware of their existence.

Within the context of smartphones, a study by (Oulasvirta et al., 2012) found evidence that smartphones are indeed habit forming because they provide immediately accessible rewards to the consumer. They enable instant gratification in terms of goal-seeking behaviors (e.g. social

14

media, communication, news, games). Furthermore, findings provided evidence that users form mobile checking habits "automated behaviors where the device is quickly opened to check the standby screen or information content in a specific application" (Oulasvirta et al., 2012, p. 107). These quickly executed behaviors are important because they can drive further reengagement. They are characteristic of smartphone device use and gateway to other

(Oulasvirta et al., 2012, p. 107) and so further increase usage. Mobile checking habits and the toward increased usage are particularly interesting as little research has been done to evaluate which aspects of smartphone architecture in particular (e.g. notifications) increase overall usage, which is the angle this study explores. The following section describes smartphones in greater detail: their ubiquity, conduciveness to habit formation and health effects of unregulated smartphone usage.

2.3 Smartphone ubiquity and habit formation Mobile phones have undergone a dramatic change in the last decade, going from simple communication devices to multifaceted communications and computing smartphone devices (Böhmer, Lander, Gehring, Brumby, & Krüger, 2014). This allowed for tremendous growth in usage: smartphones especially have seen a surge in recent years. Statista (2016) predicts that

by 2017. Western Europe in particular has seen an enormous increase in smartphone penetration and usage, with the period from 2014 to 2015 marked by a 65% growth in usage (Kesiraju, 2015). Similarly, now almost two thirds of the U.S. population owns a smartphone (Smith, 2015).

Due to their ubiquity, they are changing the fabric of daily life and social interaction: consumers do things in new ways and increasingly rely on their smartphones to perform tasks like navigation, note-taking, emailing, etc they socialize in new ways; they do tasks in new ways, often interleaving and cross-pollinating

(Oulasvirta et al., 2012, p. 105).

2.3.1 Applications The sheer scope of smartphone pervasiveness is staggering and it is, in part, driven by support for application installation (Perez, 2014). Smartphone applications are self-contained programs that support and extend the capabilities of the device any feature a smartphone might not have out of the box can be acquired through a variety of specialized applications. These customizable applications make smartphone usefulness limitless. A variety of applications caters to user needs and almost anything can be accomplished through the bright interface. On average, the global smartphone user has 26 applications on his or her device (Richter, 2013). South Koreans are at the forefront with 40 applications, while Swedes and the Swiss both have 39 (Richter, 2013).

Because applications are commercial products, the statistics above are an important measure of their successfulness. According to Adjust (2014), the number of downloads used to be the main

, but 74.5% of applications are opened only once before subsequently being ignored or deleted (eMarketer, 2015). Because of this phenomenon, retention and reengagement metrics became prominent in discerning just how successful and popular an application is (Hoch, 2014b). Common ways to promote re-engagement are deep-

15

linking, push notifications and in-app advertisement (Adjust, 2014). Because of their interruptive power and irregularity, push notifications are particularly interesting and were the main focus of this study.

2.3.2 Notifications Notifications are various ways in which smartphone devices and applications try to engage our attention in order to accomplish something, e.g. check social media, reply to an email, read a news article, etc. There are various different types of notifications (visual, aural, tactile) which prompt the consumer of new content from sources relevant (and often irrelevant) to their lives. A common scenario is receiving an instant message via one of the many Instant Messaging applications: the smartphone displays a visual cue on the screen, plays a sound through the speakers and/or engages the vibration motor to provide tactile feedback as well. All of these cues are meant to draw the consumers attention and reengage them with their smartphone device. Another common scenario is receiving a social media update where the consumer reengages with the device in order to learn of new content on their favorite social media networks.

Smartphone notifications are interruptive when delivered on an irregular schedule (Duan et al., 2005; Stothart et al., 2015). They are delivered with little to no regard for context whether they are with family or doing sports or attending an important business meeting

(Duan et al., 2005; Stothart et al., 2015). Applications are most often able to deliver notifications at the will of the company behind it and the consumer can only decide what to do with the notification after it has been downloaded and has already drawn the his or her attention (Duan et al., 2005). interrupted to reactively assess the appropriateness (Turner, Allen, & Whitaker, 2015, p. 70). Moreover, the interruptive power and irregularity of notifications may give rise to contextually irrelevant thoughts and facilitate additional smartphone usage beyond what the consumer had intended (Oulasvirta et al., 2012; Stothart et al., 2015).

Notifications enhance the necessary cue salience for operant conditioning processes, provide rewards for sensation seeking personalities and place additional strain on possible deficient self-

due to their scheduling and cognitive demands (Stothart et al., 2015), all of which can contribute to an increase in overall usage.

2.3.2.1 Notification delivery protocols The key to better management of notifications to avoid irregularity and interruptions lies in the notification delivery protocol (Duan et al., 2005; Warren & Meads, 2014). Broadly speaking, there is only one notification delivery: the push delivery protocol (Warren & Meads, 2014). A

application. The user is then notified of new content via visual/aural/tactile notification banners, popups, sounds, etc. Facebook, for example, maintains massive interconnected networks of servers which push new content and notifications to connected devices almost instantaneously and at all times. Since more than half of smartphone consumers enable push notifications (Hoch, 2014a), the interruptive power and attentional cost of smartphone notifications is incredibly high

16

(Kushlev et al., 2016) and their effect on overall usage and psychological well-being is the central part of this study.

The push notification delivery protocol can be approached via two distinct, and common, models: sender-push and receiver-pull (Duan et al., 2005). Since this research argues that increasing or decreasing the amount of notification affects consumers and their behavior, both the sender-push and the receiver-pull model play an important role in understanding how notifications travel from their sources to the models can potentially increase and decrease overall usage.

The key component of the sender-push model is that the sender can push traffic to the receiver at will, and the receiver can only accept the traffic which often comes at irregular intervals (Duan et al., 2005) An example of this is receiving email or instant messaging notifications even when not using the smartphone. The content and notifications of new content are delivered at will, at irregular intervals, regardless of what the consumer may be doing at the time. An important aspect of sender-push model is that the entire message is received before any receiver-side

(Duan et al., 2005, p. 26), i.e. the consumer can choose to accept or discard a notification only after it has been received and has drawn the attention of the consumer (Turner et al., 2015).

In the receiver-pull model, instead of uncontrollably receiving notifications from various applications installed on the device, the user can choose to opt-out of the push (automatic) notifications (Duan et al., 2005). In terms of, for example email and instant messaging, this means the consumer would have to go into the app they are interested in and press the refresh button to receive new content. The action of acquiring new content would have to become increasingly deliberate, instead of passive and receptive. The receiver-pull model implies intentionality the user requests content when he or she wants it, instead of being notified by push notifications from various applications at irregular intervals.

This thesis proposes that pulling intentionally, instead of passively receiving content, can decrease the overall smartphone usage through reducing cue salience, unexpected rewards and the strain on self-regulatory mechanism. Both the sender-push model and the receiver-pull model are an integral part of the hypotheses that initiated this study.

2.3.2.2 Why is the choice of notification delivery protocol models important? While smartphones, along with applications, provide an incredible benefit to daily life (especially in terms of connectivity and Internet usage), there are also obvious adverse effects of their overuse: for smartphone users, their phone is the first thing they look at in the morning, and the last thing they look at before going to sleep (Lee et al., 2014, p. 373). Many studies report various health issues: thumb pains, neck pains and faulty posture such as forward neck posture, slouched posture or rounded shoulders (Jung, Lee, Kang, Kim, & Lee, 2016). In Sweden, research has found that smartphone overuse was associated with sleeping disturbances and depression symptoms (Thomée, Härenstam, & Hagberg, 2011).

17

Smartphones have also been linked to anxiety and addiction-like symptoms. Researchers at the University of Maryland conducted a study where 200 students were put on a media fast for 24 hours, i.e. they had to abstain from any and all forms of media for a full day. At the end of the 24 hours, students were asked to privately blog about their experience. Many students used terms closely related to addiction to describe their relationship to media, with some going as far as

these days are in a similar situation, for between having a Blackberry, a laptop, a television, and (Independent, 2010).

Recently, many studies have focused on compulsive smartphone usage, suggesting users tend to lose control over the time spent with their devices through development of involuntary habits (Oulasvirta et al., 2012; van Deursen et al., 2015). The use of smartphones can yield immediate gratification, but it can also be accompanied by a diminished sense of volitional control and induce persistent activity (Lee et al., 2014, p. 373). Research into smartphone pervasiveness found that smartphone users tend to develop a mobile checking habit, which is defined as brief usage sessions repeating smartphone use (Oulasvirta et al., 2012). These habits drive smartphone re-engagement and are usually non-intentional content on the home screen or in an application.

18

3 Literature review findings Adoption and use of technology has been the subject of extensive previous research in terms of communications media: the television (Shade, Kornfield, & Oliver, 2015), the Internet (LaRose et al., 2001; Roy, 2009; Stafford, 2003), social media (Chen, 2011; Ifinedo, 2016) and smartphone devices (Joo & Sang, 2013; Kushlev et al., 2016; Oulasvirta et al., 2012; Wei & Lu, 2014). The research was grounded in either the Uses and Gratification theory or the Unified Theory of Technology Acceptance and Use of Technology, both of which attempt to explain the adoption of different media and predict continued use (Miajung, 2014). The findings concluded that previous experience and already established habits predict continued use, more so than any other factors (e.g. perceived usefulness of media, consumer needs, etc.) (Miajung, 2014).

However, habitual behavior concerning communications media can also result in unregulated use of media (LaRose et al., 2003), due to different personality traits (Bianchi & Phillips, 2005), operant conditioning processes (Marlatt, 1988; Oulasvirta et al., 2012) and deficient self-regulation (LaRose et al., 2003).

Within the context of smartphone devices, research found a tendency toward habitual overuse (Oulasvirta et al., 2012) and detrimental, health-related side-effects (Bian & Leung, 2015; Haug et al., 2015; Jung et al., 2016; Thomée et al., 2011). most salient characteristics notifications, whose sole purpose is to draw the consumers attention to their devices.

Notifications can be divided into two models: sender-push and receiver-pull (Duan et al., 2005). The main difference between the two is that the former delivers notifications on an irregular sc the consumer goes to an application of his choice to intentionally check for new content, thus controlling the exchange of information (Duan et al., 2005).

Since notifications serve to reengage the user and facilitate usage, they may directly be involved in creating a context favorable toward overuse (Miajung, 2014; Stothart et al., 2015). They may be what provides cues for operant conditioning processes, rewards for addictive personality and what may further diminish consumer self-regulatory mechanisms. Notifications have also been found to cause stress and even ADHD-like symptoms (Kushlev et al., 2016), and to act as cues that enhance the operant conditioning behavior-cue link (Oulasvirta et al., 2012). Because of the above, it is necessary to examine the relationship between smartphone notifications and their effects on the time consumers spend on their devices. Due to smartphone ubiquity and pervasiveness (Smith, 2015; Soukup, 2015; Warren & Meads, 2014), consumers have changed their daily conduct many daily activities now depend on what smartphones provide (Oulasvirta et al., 2012). Smartphone notifications are usually delivered to consumers without their control and a closer examination of their direct influence on time spent with smartphone devices may further the understanding of what makes habitual behavior so relevant within the smartphone context. Habitual smartphone behavior seems to adversely impact consumer lives (postural problems, anxiety, stress, depression, etc.) and so diminishes the immense usefulness of these devices.

19

This study hypothesizes that eliminating or minimizing the sender-push model of notification delivery can reduce overall smartphone usage and in doing so help prevent health problems and unregulated smartphone usage.

20

4 AIM AND RESEARCH QUESTIONS The aim of this master thesis project is to examine the effects of different smartphone notifications protocols (i.e. sender-push, receiver-pull) on overall daily smartphone usage, frequency of usage and overall perceived stress levels.

HYPOTHESIS 1a: Setting notifications to sender-push protocol on smartphone devices increases overall

smartphone usage and increases the frequency of use. HYPOTHESIS 1b:

Setting notifications to sender-push protocol on smartphone devices increases overall participant stress levels, as measured by the Perceived Stress Scale.

HYPOTHESIS 2a: Setting notifications to receiver-pull protocol on smartphones decreases overall mobile phone usage and decreases the frequency of use.

HYPOTHESIS 2b: Setting notifications to receiver-pull protocol on smartphone devices decreases overall participant stress levels, as measured by the Perceived Stress Scale.

21

5 METHODOLOGY

In order to examine the effects of smartphone notifications on smartphone usage, both primary (recruited participants) and secondary (literature review) data was used.

5.1 Literature review Secondary data was examined through Uppsala University online library, which has access to a wide range of information sources (databases, journals and catalogues). Other academic and commercial databases (e.g. Wiley Online Library) were used, as well as Internet search engines (Google).

5.2 Data collection and analysis The study consisted of gathering participant data through quantitative questionnaires and smartphone usage tracking.

5.2.1 Quantitative Questionnaires Before the start of the study, participants signed a consent form to take part in the study, followed by three questionnaires:

1. Subjective Happiness Questionnaire (Lyubomirsky & Lepper, 1999) a 4-item self-report happiness questionnaire. (Appendix A)

2. Perceived Stress Scale Questionnaire (Cohen & Williamson, 1988) a 10-item self-report stress questionnaire. (Appendix B)

After the 10-day study period, all the participants completed the same questionnaires so that pre and post study results could be compared in relation to the different notification settings. All of the questionnaires have valid Cronbach Alpha variables, and can be found in the appendices.

5.2.2 Usage Tracking The purpose of this study was to examine consumer smartphone usage behavior under either no notification or maximum notification settings. The study followed a randomized between-subject non-crossover study model with two conditions related to notification settings. The 26 participants were divided into three groups using true random number generator (TRNG) (Chermaine, 2005; Haahr, 2016), which ensured no bias when creating the study groups. The group condition assignment was as follows:

Group 1 (9 participants): all participants in this group were required to mute all application notifications besides regular phone calls and SMS messages. This included navigating to each

settings to manually mute all notifications from making a sound, vibrating, flashing the notification light or appearing on the smartphone screen. In order for a participant to check new content in any application, he or she had to navigate to the application, open it and refresh content manually. The notification settings in group 1 correspond to the receiver-pull notification delivery model.

22

Group 2 (9 participants): In this group, all participants were required to maximize all visual, auditory and tactile notifications for each app. In other words, notification settings for each application located on the smartphone device were manually all set to be settings in group 2 correspond to the sender-push notification delivery model.

Group 3 (8 participants, control group): In this group, participants notification where not changed and use their smartphone as they normally would. This group was created in order to have a control data sample with which to compare data gathered from groups 1 and 2.

Following the data gathering and analysis, the mean, module and median for both groups 1 and 2 was compared to the mean, module and median value for the control group (group 3) in order to determine if there was any statistical significance through the ANOVA (analysis of variance) model.

It is important to note that the participants were not prohibited from e.g. turning off their phones before sleep or when working. Participants were encouraged to use their smartphones as they normally would.

The crucial data regarding actual, real-time smartphone usage was gathered using a usage-tracking Android-based application Quality Time.

5.2.3 Statistics

Acquired data was examined via the General Linear Model (GLM) and a one-way between-subjects analysis of variance (ANOVA) in order to possibly reject the null hypothesis (no differences between groups) and reveal any significant differences between the mean values of the condition groups (1 & 2) and the control group in a number of different scenarios, i.e. daily usage, frequency of usage, etc. StatSoft Statistica 13.0 suite of analytics software products and Microsoft Excel 2016 were used to conduct all statistics for this study.

5.2.4 Sample

The sample was recruited using social networks (Facebook) and additional participants were recruited friends and acquaintances. The data was in the end collected from a sample of 26 people across different demographics, all of which have been using an Android-based smartphone device for at least a year. The sample consisted of 46.15% females and 53.85% males. The sample was 7.69% Asian/Pacific Islander and 92.31% White Caucasian. Of the total 26 participants, 3.85% were Albanian, 3.85% Chinese, 23.07 % Croatian, 3.85% Finnish, 23.07% German, 7.69 % Italian, 7.69 % Dutch, 19.23 % Swedish, 3.85 % Greek and 3.85 %

23

Thai. Age wise, 80.77 % were between the ages of 18-29, 7.69% between the ages of 30-44 and 11.54% between the ages of 45-59.

5.2.5 Relevant variables In order to compare overall usage and usage frequency between the groups with different conditions, the following variables were of particular interest:

Total Daily Usage (TDU) - coded as a number representing minutes. Total Daily Application Interactions (TDAI) - coded as a number representing interaction

frequency. Total Daily Unlocks (TDUn) - coded as a number representing the frequency of

smartphone screen unlocks and subsequent locks.

age habits and patterns.

5.2.6 Study Period The study was conducted under the period between April 16th, 2016 and April 25th, 2016. Prior to the beginning of the study, all participants completed all three questionnaires, had installed the QualityTime application and had their notifications modified to the correct setting based on the group they belonged to. The notification modification for each participant was done on the day before the beginning of the study (April 15th), roughly around the same time to avoid confounding variables caused by different modification introduction times.

Following the conclusion of the study period (April 26th), all participant data was gathered within two days and all the participants completed the questionnaires anew. The following section will expand on the results of the acquired data and the subsequent analysis.

5.2.7 Usage tracking was necessary to test the hypotheses. To that effect, a comparison of available applications on the Google Play Store market was conducted and Quality Time was chosen among a multitude of other choices due to the following features and characteristics: a) stability, b) ease of use and setup, c) available data backup d) good experience from previous users and e) light on device battery and RAM/CPU resources. Other available applications (e.g. Offtime, UBhind) have similar features but were found lacking in one or more of the above stated categories.

Quality Time is an Android-based application developed by ZeroDesktop, Inc available on the Google Play Store market. The application, once installed, tracks smartphone usage which it then displays in a readable format within the application interface. The application tracks various different variables, e.g. total daily usage, total daily unlocks, total daily application interactions,

-

24

Additionally, the interface enables the consumer to keep track of other applications which consume the most time or are accessed most frequently in order to familiarize oneself with personal smartphone habits. The application allows the user to block certain apps and to enable overuse warning messages. These features are outside the scope of this study and were not used.

5.2.7.1 Application interface Once the application is setup on the smartphone and has run for some time, it begins to show usage data. The initial screen (Figure 1) shows total daily usage up to the moment of checking, as well as timeline activity on a per minute and per second basis.

Figure 2 - Quality Time initial screen, showing total smartphone daily usage and timeline activity

25

Swiping down on the display brings up a more detailed overview of the daily usage - namely in terms of most used applications (in minutes) (Figure 2, left), most accessed applications (frequency) (Figure 2, right) as well as the number of screen unlocks (frequency).

Figure 3 - Total smartphone daily usage and usage per app

The application keeps a detailed track of all smartphone usage and displays it in a readable format which allows for in depth examination of participant smartphone usage patterns and habits.

26

Figure 4 - Total smartphone usage frequency per app

5.2.7.2 Usage tracking limitations Two limitations of the Quality Time application and the Android operating system (OS) are important to note:

1. The application does not have a readymade export feature, so all data was exported from el 2016 spreadsheet and thoroughly

cross-checked for errors. 2. The Android OS does not support application auto-start after device power off. To

minimize the occurrence of the data gaps, all participants were instructed on how to keep the application running as a background process and how to turn it on should the application turn off. The participants every morning of the study received an email and social network messages reminding them to keep the application running on their phones. Despite the instructions, significant gaps in the data were still found in 11 out of the

processes on their smartphones.

smartphones, all notifications coming from the application itself were turned off for each

27

participant, regardless of the group they belong to, in order to make it as non-intrusive as possible.

28

6 RESULTS This thesis set out to examine the importance of notification delivery models in relation to overall smartphone usage. This was accomplished in two parts. The first part consisted of conducting a one-way between-subjects ANOVA to compare the overall pre and post study questionnaires for each group. The second part consisted of conducting a one-way between subjects ANOVA to compare the tracked usage data for each of the two condition groups in comparison to the control group. Additionally, each study group results will be discussed in two ways: for the entire group (9 participants) and then for the members who had a complete usage tracking data set, i.e. the Quality Time application was running for the full duration of the study with no significant data gaps.

6.1 Results for all participants A one-way between subjects ANOVA was conducted to compare the effect of notifications on overall smartphone daily usage and daily frequency of usage in notifications off (group 1), all notifications on (group 2) and no change (group 3) conditions.

6.1.1 Group 1 (No Notifications)

No Notification Total Daily Usage, Total Daily App Interactions and Total Daily Unlocks yielded no significant results at 9,18= 1.360.28, [F (9,36) = 1.20, p = 0.32] and [F (9,36) = 1.54, p = 0.17] respectively.

Figure 5- Total Daily Usage, Group 1 compared to Group 3 (control group)

29

6.1.2 Group 2 (Max Notifications) The effect of the Max Notification independent variable on Total Daily Usage, Total Daily App Interactions and Total Daily Unlocks yielded no significant results at [F (9, 27) = 2.15, p = .06], [F (9, 54) = .40, p = .93] and [F (9, 54) = .61, p = .78] respectively.

Figure 6 - Total Daily Usage, Group 2 compared to Group 3 (control group)

6.2 Results for participants with complete datasets 6.2.1 Group 1 (No Notifications) The effect of the No Notification independent variable on Total Daily Usage, Total Daily App Interactions and Total Daily Unlocks yielded no significant results at [F (9, 36) = 1.02, p = .18], [F (9, 36) = 1.20, p = .32] and [F (9, 63) = 1.27, p = .27] respectively.

30

Figure 7 - (complete dataset) Total Daily Usage, Group 1 compared to Group 3 (control group)

6.2.2 Group 2 (Max Notifications) The effect of the Max Notification independent variable on Total Daily Usage, Total Daily App Interactions and Total Daily Unlocks yielded no significant results at [F (9, 54) = 1.83, p = .08], [F (9, 54) = .40, p = .93] and [F (9, 63) = .79, p = .88] respectively.

Figure 8 - (complete dataset) Total Daily Usage, Group 2 compared to group 3 (control group)

31

6.3 Questionnaire Results Group 1: Stress Questionnaire [F (1, 8) = 0.33, p = 0.58], Happiness Questionnaire: [F (1, 8) = 0.41, p = 0.54].

Figure 9 - pre/post Stress Questionnaire, Group 1

Group 2: Stress Questionnaire [F (1, 8) = 2.28, p = 0.17], Happiness Questionnaire: [F (1, 8) = 0.05, p = 0.83].

Figure 10 - pre/post Stress Questionnaire, Group 2

32

7 DISCUSSION To further the understanding of smartphone habits this research proposed that push notifications and their particular subcategories (i.e. sender-push, receiver-pull) affect the total smartphone daily usage and frequency of usage. This research also proposed that the total amount of daily notifications would affect the general stress levels measured by the Perceived Stress Scale (PSS) questionnaire (Cohen & Williamson, 1988), as well as the general happiness levels measured by the Subjective Happiness Scale questionnaire (Lyubomirsky & Lepper, 1999).

To examine the effects of the sender-push and receiver-pull model of notifications, participants were divided into three groups. Group 1 had all their notifications turned off while group 2 had all notifications on (i.e. all applications had permission to push notifications). Additionally, a usage tracking application (Quality Time) was in order to gather usage statistics, namely in terms of daily use, number of interactions and the number of screen unlocks/locks.

The findings, however, uniformly disproved all three hypotheses. Close examination of gathered data revealed that silencing all smartphone notifications did not affect overall daily usage nor did it affect stress levels in any meaningful way with ANOVA p-values often showing very high values (p > 0.5). Both groups 1 (no notifications) and 2 (maximum notifications) revealed no significant differences compared to the control group of participants who received no modification to their devices.

Analysis of the data shows that, over a ten-day study period, participants in group 1 did experience a sharp increase in statistical variance of overall usage and frequency of usage during the first quartile. This can possibly be explained by feelings of anxiety caused by the lack of aural, tactile or visual notifications which prompted participants to check their devices even more often than usual to make sure they were not missing out on something. However, the increased usage, contrary to expectations, quickly subsided and subsequent levels were not significantly different compared to the control group. Frequency of usage and daily screen unlocks showed no spikes during the ten-day period and were not significantly different to the control group. Group 2 (maximum notifications) also revealed no significant differences compared to the control groups.

Most surprisingly, an ANOVA between groups 1 (no notification) and 2 (maximum notifications) also revealed no significant differences, further suggesting that changing the amount of notifications does not affect usage. However, one limitation of the study was revealed here: group 2 participants were required to enable push notifications on all applications prior to the start of the study. Since many group 2 participants already had push notifications enabled, it is possible that the overall number of notifications for this group did not increase significantly and thus had no effect of overall usage. While there is ample evidence that smartphones are conducive to habitual behavior (Oulasvirta et al., 2012; van Deursen et al., 2015), the amount of notifications and their frequency does not seem to significantly influence the overall time spent on the device. Except for a brief increase in usage at the beginning of the ten-day period, the

usage was fairly stable and did not differ significantly between groups.

33

8 CONCLUSION Previous research found that smartphones can be conducive to both habit formation (Oulasvirta et al., 2012) and unregulated usage (Lee et al., 2014), as defined by (LaRose et al., 2003). This study attempted to provide evidence that smartphone notifications can cause an increase in overall smartphone usage and stress levels, due to their salience and instant gratification two characteristics both of which have previously been linked to habit formation (Oulasvirta et al., 2012; Stothart et al., 2015) and increased stress levels (Kushlev et al., 2016).

However, the findings show no correlation between smartphone notifications and smartphone usage/frequency of usage. Even though smartphone notifications correspond to operant conditioning cues, offer reward- -traits and may further diminish self-regulative mechanisms, there seems to be no significant link connecting notifications to overall smartphone usage nor participant stress levels. The disproval of all of the hypotheses provides evidence and support to the claim that already established habits (of use and conduct) best predict continued use (Limayem et al., 2007; Miajung, 2014). This correlates with prior findings regarding habits and media consumption, grounded in Uses and Gratifications theory (Joo & Sang, 2013) and the Unified Theory of Acceptance and Use of Technology (Venkatesh et al., 2012).

Drastic changes to the amount of notifications received does not seem to affect already established, habitual smartphone behavior. Consumer smartphone habits and usage patterns are likely intimately connected to the value and usefulness a smartphone provides in everyday life (van Deursen et al., 2015), a connection that is resilient to direct changes to the smartphone notification context.

8.1 Limitations Number of participants due to time and organizational constraints, this study was completed by

Further studies should increase the number of participants to increase the reliability of the statistics.

Usage Tracking Application for this study, an existing Android-based application was used to track overall daily usage, overall daily interactions as well as per application usage. Due to firmware limitations, the application did not autostaresulted in an incomplete data for 11 of the participants studied. To avoid this, it is advised that further studies should develop a custom usage tracking application which runs with full privileges and is always on for the duration of the study.

Time period due to organizational constraints, this study was limited to a 10-day period. Even though this allows for a glimpse into the difference modifying notifications introduces in

e period of several months would be preferable so participants would have time to adjust to the new settings.

Participant location this study involved participants from several different countries, which necessitated an involved process of application installation (some participants installed and setup

34

the phone. This made the participants aware of all the changes to their smartphones as well as the goals of the study, which may affect the data.

Previous behavior and settings participants were not screened for previous smartphone behavior and notification settings, so it may be possible that participants in group 2 already had all or most of applications set to enable push notifications.

35

9 FUTURE WORK In order to acquire more in-depth data and minimize data gaps, a custom application running with full privileges (i.e. can autostart) on the would provide a better overview of the effects of notifications on smartphone usage.

Another recommendation is based on the length of this study - the time period was limited to a ten-day period which may have not been enough time for participants to fully adjust to the new notification environment. A future, longer study is recommended.

Furthermore, a significant effort was made to setup and install the tracking application on the , often over phone calls and remote connection software. Also,

notification modification was done in tandem with several participants by walking them through the process. This may have influenced participants and caused them to be fully aware of what the study aimed to accomplish, which may have affected the results. A future study with less participant input would be beneficial.

Lastly, any future studies should be able to manually increase and decrease the number of notifications participants receive during the study period as this will ensure that there is a measurable difference in notifications received (especially for the maximum notifications group), from which a more concrete conclusion may be drawn.

36

10 REFERENCES Adjust. (2014). Beyond the install: Retargeting and reengagement explained. Retrieved from

http://adjust.com/assets/downloads/beyond-the-install-re-engagement-retargeting_adjust.pdf Agarwal, R., & Prasad, J. (1997). The Role of Innovation Characteristics and Perceived

Voluntariness in the Acceptance of Information Technologies. Decision Sciences, 28(3), 557582. doi:10.1111/j.1540-5915.1997.tb01322.x

Ajzen, I. (2002). Perceived Behavioral Control, Self-Efficacy, Locus of Control, and the Theory of Planned Behavior1. Journal of Applied Social Psychology, 32(4), 665 683. doi:10.1111/j.1559-1816.2002.tb00236.x

Ally, M., & Gardiner, M. (Eds.) 2012. The moderating influence of device characteristics and usage on user acceptance of Smart Mobile Devices: er ACIS.

Al-Showarah, S., AL-Jawad, N.,. . . Sellahewa, H. (2014). Effects of User Age on Smartphone and Tablet Use, Measured with an Eye-Tracker via Fixation Duration, Scan-Path Duration, and Saccades Proportion, 8514, 3 14. doi:10.1007/978-3-319-07440-5_1

American Psychiatric Association. (2013). Diagnostic and Statistical Manual of Mental Disorders: American Psychiatric Association.

Bandura, A. (1989). Human agency in social cognitive theory. American Psychologist, 44(9), 1175 1184. doi:10.1037/0003-066X.44.9.1175

Bandura, A. (1999). A Sociocognitive Analysis of Substance Abuse: An Agentic Perspective. Psychological Science, 10(3), 214 217. doi:10.1111/1467-9280.00138

Bandura, A. (2001). Social Cognitive Theory: An Agentic Perspective. Annual Review of Psychology, 52(1), 1 26. doi:10.1146/annurev.psych.52.1.1

Bian, M., & Leung, L. (2015). Linking Loneliness, Shyness, Smartphone Addiction Symptoms, and Patterns of Smartphone Use to Social Capital. Social Science Computer Review, 33(1), 61 79. doi:10.1177/0894439314528779

Bianchi, A., & Phillips, J. G. (2005). Psychological Predictors of Problem Mobile Phone Use. CyberPsychology & Behavior, 8(1), 39 51. doi:10.1089/cpb.2005.8.39

Block, J. J. (2008). Issues for DSM-V: internet addiction. The American journal of psychiatry, 165(3), 306 307. doi:10.1176/appi.ajp.2007.07101556

Böhmer, M., Lander, C.,. . . Krüger, A. (2014). Interrupted by a phone call, 3045 3054. doi:10.1145/2556288.2557066

Butt, S., & Phillips, J. G. (2008). Personality and self reported mobile phone use. Computers in Human Behavior, 24(2), 346 360. doi:10.1016/j.chb.2007.01.019

Chen, G. M. (2011). Tweet this: A uses and gratifications perspective on how active Twitter use gratifies a need to connect with others. Computers in Human Behavior, 27(2), 755 762. doi:10.1016/j.chb.2010.10.023

Chermaine, K. (2005). Random Number Generators: An Evaluation and Comparison of Random.org and Some Commonly Used Generators.

Chittaranjan, G., Blom, J.,. . . Gatica-Perez, D. Who's Who with Big-Five: Analyzing and Classifying Personality Traits with Smartphones, pp. 29 36.

37

Cohen, S., & Williamson, G. M. (1988). Perceived Stress in a Probability Sample of the United States. In S. Spacapan & S. Oskamp (Eds.), "The Social Psychology of Health.", 31 66.

Deloitte. (2015). 2015 Global Mobile Consumer Survey: US Edition. Retrieved from http://www2.deloitte.com/content/dam/Deloitte/us/Documents/technology-media-telecommunications/us-tmt-global-mobile-executive-summary-2015.pdf

Dery, K., Kolb, D.,. . . MacCormick, J. (2014). Working with connective flow: How smartphone use is evolving in practice. European Journal of Information Systems, 23(5), 558 570. doi:10.1057/ejis.2014.13

Duan, Z., Dong, Y.,. . . Gopalan, K. (2005). Push vs. Pull: Implications of Protocol Design on Controlling Unwanted Traffic.

eMarketer. (2015). How Many Apps Do Smartphone Owners Use? - eMarketer. Retrieved from http://www.emarketer.com/Article.aspx?R=1013309&ecid=MX1086

Griffiths, M. D. (1991). Amusement machine playing in childhood and adolescence: A comparative analysis of video games and fruit machines. Journal of Adolescence, 14(1), 5373. doi:10.1016/0140-1971(91)90045-S

Haahr, M. (2016). Random.org - List Randomizer. Retrieved from https://www.random.org/analysis/

Haug, S., Castro, R. P.,. . . Schaub, M. P. (2015). Smartphone use and smartphone addiction among young people in Switzerland. Journal of Behavioral Addictions, 4(4), 299 307. doi:10.1556/2006.4.2015.037

Hoch, D. (2014a). 52% of Users Enable Push Messaging on their Mobile Devices. Retrieved from http://info.localytics.com/blog/52-percent-of-users-enable-push-messaging

Hoch, D. (2014b). App Retention Improves - Apps Used Only Once Declines to 20%. Retrieved from http://info.localytics.com/blog/app-retention-improves

Ifinedo, P. (2016). Applying uses and gratifications theory and social influence processes to understand students' pervasive adoption of social networking sites: Perspectives from the Americas. International Journal of Information Management, 36(2), 192 206. doi:10.1016/j.ijinfomgt.2015.11.007

Independent. (2010). Students' addiction to media akin to drugs. Retrieved from http://www.independent.co.uk/life-style/health-and-families/students-addiction-to-media-akin-to-drugs-5536767.html

Joo, J., & Sang, Y. technology acceptance model and uses and gratifications theory. Computers in Human Behavior, 29(6), 2512 2518. doi:10.1016/j.chb.2013.06.002

Jung, S. I., Lee, N. K.,. . . Lee, D. Y. (2016). The effect of smartphone usage time on posture and respiratory function. Journal of Physical Therapy Science, 28(1), 186 189. doi:10.1589/jpts.28.186

Kesiraju, L. (2015). The Europe Report: Early Adopters Continue to Evolve. Retrieved from http://flurrymobile.tumblr.com/post/133792415780/europereport

Kessler, E. H. (2013). Encyclopedia of management theory: Technology Acceptance Model.

38

Kim, B. (2012). The diffusion of mobile data services and applications: Exploring the role of habit and its antecedents. Telecommunications Policy, 36(1), 69 81. doi:10.1016/j.telpol.2011.11.011

Kubey, R. (1996). Television Dependence, Diagnosis, and Prevention:: With Commentary on Video Games, Pornography, and Media Education. Young.

Kushlev, K., Dunn, E., & Proulx, J. (Eds.) 2016. "Silence Your Phones":: Smartphone Notifications Increase Inattention and Hyperactivity Symptoms.

Lally, P., van Jaarsveld, Cornelia H. M.,. . . Wardle, J. (2010). How are habits formed: Modelling habit formation in the real world. European Journal of Social Psychology, 40(6), 998 1009. doi:10.1002/ejsp.674

LaRose, R. (2010). The Problem of Media Habits. Communication Theory, 20(2), 194 222. doi:10.1111/j.1468-2885.2010.01360.x

LaRose, R., Lin, C. A.,. . . Eastin, M. S. (2003). Unregulated Internet Usage: Addiction, Habit, or Deficient Self-Regulation? Media Psychology, 5(3), 225 253. doi:10.1207/S1532785XMEP0503_01

LaRose, R., Mastro, D.,. . . Eastin, M. S. (2001). Understanding Internet Usage: A Social-Cognitive Approach to Uses and Gratifications. Social Science Computer Review, Vol. 19 No. 4, Winter 2001 395-413.

Lee, U., Lee, J., Ko, M., Lee, C., Kim, Y., Yang, S.,. . . Song, J. (Eds.) 2014. Hooked on smartphones: an exploratory study on smartphone overuse among college students: ACM.

Lee, Y.-K., Chang, C.-T.,. . . Cheng, Z.-H. (2014). The dark side of smartphone usage: Psychological traits, compulsive behavior and technostress. Computers in Human Behavior, 31, 373 383. doi:10.1016/j.chb.2013.10.047

Limayem, M., Hirt, S. G.,. . . Cheung, C. M. K. (2007). How Habit Limits the Predictive Power of Intention: The Case of Information Systems Continuance. MIS Quarterly, 31(4), 705 737.

Littleton, K., & Light, P. (2002). Learning with computers: analysing productive interaction. New York, London: Routledge.

Livaditi, J., Vassilopoulou, K.,. . . Chorianopoulos, K. (2002). Needs and gratifications for interactive TV implications for designers, 9 pp. doi:10.1109/HICSS.2003.1174237

Lyubomirsky, S., & Lepper, H. S. (1999). A measure of subjective happiness (1999). Social Indicators Research, 46(2), 137 155. doi:10.1023/A:1006824100041

Marlatt, G. (1988). Addictive Behaviors: Etiology And Treatment. Annual Review of Psychology, 39(1), 223 252. doi:10.1146/annurev.psych.39.1.223

McIlwraith, R. D. (1998). `I'm addicted to television': The personality, imagination, and TV watching patterns of. Journal of Broadcasting & Electronic Media, 42(3), 371.

Miajung, K. (2014). The Effects of External Cues on Media Habit and Use:: Push Notification Alerts and Mobile Application Usage Habits.

Nathan, P. E. (1988). The addictive personality is the behavior of the addict. Journal of Consulting and Clinical Psychology, 56(2), 183 188. doi:10.1037//0022-006X.56.2.183

39

Neo, R. L., & Skoric, M. M. (2009). Problematic Instant Messaging Use. Journal of Computer-Mediated Communication, 14(3), 627 657. doi:10.1111/j.1083-6101.2009.01456.x

Ouellette, J. A., & Wood, W. (1998). Habit and intention in everyday life: The multiple processes by which past behavior predicts future behavior. Psychological Bulletin, 124(1), 54 74. doi:10.1037/0033-2909.124.1.54

Oulasvirta, A., Rattenbury, T.,. . . Raita, E. (2012). Habits make smartphone use more pervasive. Personal and Ubiquitous Computing, 16(1), 105 114.

Park, N., & Lee, H. (2012). Social implications of smartphone use: Korean college students' smartphone use and psychological well-being. Cyberpsychology, behavior and social networking, 15(9), 491 497. doi:10.1089/cyber.2011.0580

Pearson, C., & Hussain, Z. (2015). Smartphone Use, Addiction, Narcissism, and Personality:: A Mixed Methods Investigation. International Journal of Cyber Behavior, Psychology and Learning, 5(1), 17 32. doi:10.4018/ijcbpl.2015010102

Pempek, T. A., Yermolayeva, Y. A.,. . . Calvert, S. L. (2009). College students' social networking experiences on Facebook. Journal of Applied Developmental Psychology, 30(3), 227 238. doi:10.1016/j.appdev.2008.12.010

Perez, S. (2014). Majority Of Digital Media Consumption Now Takes Place In Mobile Apps. Retrieved from http://techcrunch.com/2014/08/21/majority-of-digital-media-consumption-now-takes-place-in-mobile-apps/

Richter, F. (2013). Infographic: The Average Smartphone User Has Installed 26 Apps. Retrieved from http://www.statista.com/chart/1435/top-10-countries-by-app-usage/

Roy, S. K. (2009). Internet uses and gratifications: A survey in the Indian context. Computers in Human Behavior, 25(4), 878 886. doi:10.1016/j.chb.2009.03.002

Rozin, P., & Stoess, C. (1993). Is there a general tendency to become addicted? Addictive Behaviors, 18(1), 81 87. doi:10.1016/0306-4603(93)90011-W

Rubin, A. M. (1984). Ritualized and Instrumental Television Viewing. Journal of Communication, 34(3), 67 77. doi:10.1111/j.1460-2466.1984.tb02174.x

Säljö, R. (1998). Learning As the Use of Tools: A Sociocultural Perspective on the Human-technology Link: Learning with Computers. In K. Littleton & P. Light (Eds.), Learning with Computers. Analysing Productive Interactions (1) (pp. 144 161). New York, NY, 10001: Routledge. Retrieved from http://dl.acm.org/citation.cfm?id=299844.299860

Scannapieco, C. (2013). Cosmological simulations of galaxy formation: Successes and challenges in the era of supercomputers. Astronomische Nachrichten, 334(6), 499.

Scott, M., & Sale, S. (2014). Consumers use smartphones for 195 minutes per day: Spend only 25% of that time on communications.

Shade, D. D., Kornfield, S.,. . . Oliver, M. B. (2015). The Uses and Gratifications of Media Migration: Investigating the Activities, Motivations, and Predictors of Migration Behaviors Originating in Entertainment Television. Journal of Broadcasting & Electronic Media, 59(2), 318 341. doi:10.1080/08838151.2015.1029121

40

Smith, A. (2015). U.S. Smartphone Use in 2015. Retrieved from http://www.pewinternet.org/2015/04/01/us-smartphone-use-in-2015/

Soukup, P. A. (2015). Smartphones. Communication Research Trends, 34(4), 3. Stafford, T. F. (2003). Differentiating between adopter categories in the uses and gratifications

for internet services. IEEE Transactions on Engineering Management, 50(4), 427 435. doi:10.1109/TEM.2003.819652

Stothart, C., Mitchum, A.,. . . Yehnert, C. (2015). The attentional cost of receiving a cell phone notification. Journal of experimental psychology. Human perception and performance, 41(4), 893 897. doi:10.1037/xhp0000100

Taylor, S., & Todd, P. A. (1995). Understanding Information Technology Usage: A Test of Competing Models. Information Systems Research, 6(2), 144 176. doi:10.1287/isre.6.2.144

Thomée, S., Härenstam, A.,. . . Hagberg, M. (2011). Mobile phone use and stress, sleep disturbances, and symptoms of depression among young adults - a prospective cohort study. BMC Public Health, 11(1), 66. doi:10.1186/1471-2458-11-66

Turner, L. D., Allen, S. M.,. . . Whitaker, R. M. (2015). Push or Delay? Decomposing Smartphone Notification Response Behaviour. In A. A. Salah, B. J. A. Kröse, & D. J. Cook (Eds.), LNCS sublibrary: SL6 - Image Processing, Computer Vision, Pattern Recognition, and Graphics: Vol. 9277. Human behavior understanding. 6th International Workshop, HBU 2015, Osaka, Japan, September 8, 2015, proceedings / Albert Ali Salah, Ben J.A. Kröse, Diane J. Cook (eds.) (Vol. 9277, pp. 69 83). Cham: Springer International Publishing.

Uffen, J., Kaemmerer, N.,. . . Breitner, M. H. (2013). Personality Traits and Cognitive Determinants An Empirical Investigation of the Use of Smartphone Security Measures. Journal of Information Security, 04(04), 203 212. doi:10.4236/jis.2013.44023

van Deursen, A. J., Bolle, C. L.,. . . Kommers, P. A. (2015). Modeling habitual and addictive smartphone behavior. Computers in Human Behavior, 45, 411 420. doi:10.1016/j.chb.2014.12.039

Venkatesh, V., L. Thong, J. Y.,. . . Xu, X. (2012). Consumer Acceptance and Use of Information Technology: Extending the Unified Theory of Acceptance and Use of Technology. MIS Quarterly, 36(1), 157 178.

Venkatesh, V., Morris, M. G.,. . . Davis, F. D. (2003). User Acceptance of Information Technology: Toward a Unified View. MIS Quarterly, 27(3), 425 478.

Verplanken, B. (2006). Beyond frequency: Habit as mental construct. British Journal of Social Psychology, 45(3), 639 656. doi:10.1348/014466605X49122

Verplanken, B., & Wood, W. (2006). Interventions to Break and Create Consumer Habits. Journal of Public Policy & Marketing, 25(1), 90 103. doi:10.1509/jppm.25.1.90

Wang, R., & Harris, C. (2013). Scaling radio astronomy signal correlation on heterogeneous supercomputers using various data distribution methodologies. Experimental Astronomy, 36(3), 433 449. doi:10.1007/s10686-013-9340-7

Warren, I., & Meads, A. (2014). Push Notification Mechanisms for Pervasive Smartphone Applications. IEEE Pervasive Computing, 13(2), 61 71. doi:10.1109/MPRV.2014.34

41

Wei, P.-S., & Lu, H.-P. (2014). Why do people play mobile social games?: An examination of network externalities and of uses and gratifications. Internet Research, 24(3), 313 331. doi:10.1108/IntR-04-2013-0082

Weinstein, A., & Lejoyeux, M. (2010). Internet addiction or excessive internet use. The American journal of drug and alcohol abuse, 36(5), 277 283. doi:10.3109/00952990.2010.491880

Weiyan, L. (2015). A Historical Overview of Uses and Gratifications Theory. doi:10.3968/7415 Young, K. (1998). Caught in the Net: how to recognize the signs of Internet addiction & a

winning strategy for recovery. The Globe and Mail (Index-only).

42

11 APPENDIX A Stress Questionnaire

43

44

45

12 APPENDIX B Happiness Questionnaire

46

47

13 APPENDIX C Demographics Questionnaire

48

49

50

51

14 APPENDIX D Participant Demographics